Encyclopedia of Endocrinology 0124755704, 9780124755703

Citation preview

Editor-in-Chief

Luciano Martini University of Milan Milan, Italy

Associate Editors Eli Y. Adashi

Hiroo Imura

Ellis Samols

University of Utah School of Medicine Salt Lake City, Utah, United States

Prime Minister’s Office, Government of Japan Tokyo, Japan

University of Nevada, Las Vegas School of Medicine Las Vegas, Nevada, United States

John P. Bilezikian Columbia University New York, New York, United States

George P. Chrousos

Antonio Liuzzi Istituto Auxologico Italiano Verbania, Italy

Gustav Schonfeld Washington University, St. Louis School of Medicine St. Louis, Missouri, United States

National Institute of Child Health and Human Development Bethesda, Maryland, United States

Frank L. Moore

Junichi Fukata

Daniel Porte

Kochi Women’s University Kochi City, Kochi, Japan

VA San Diego Health Care System San Diego, California, United States

Julianne Imperato-McGinley

Jens F. Rehfeld

Wilmar M. Wiersinga

Weill Medical College of Cornell University New York, New York, United States

Copenhagen University Hospital Copenhagen, Denmark

University of Amsterdam Amsterdam, The Netherlands

Oregon State University Corvallis, Oregon, United States

Pierre C. Sizonenko University of Geneva, Hoˆpital la Tour Geneva, Switzerland

Foreword

The field of endocrinology is inextricably linked to physiology. The specialty was initially founded when it became clear that various glands produced hormones that exerted characteristic physiologic effects on growth, reproduction, and metabolism. Early descriptions of hormone deficiency syndromes such as Addison’s disease and myxedema were soon followed by hormone replacement strategies, often resulting in dramatic clinical effects. These observations unleashed intensive efforts to isolate and characterize the steroid and peptide hormones produced by the adrenal, thyroid, parathyroid, pituitary, and pancreatic islets, and other glands. The success of this era was epitomized by the isolation of insulin and the successful treatment of children with type 1 diabetes mellitus in 1922. The development of radioimmunoassays (RIAs) was a monumental advance that allowed hormones to be measured in various physiologic conditions. RIAs transformed endocrinology more than any other field. The ability to measure hormone levels during stimulation and suppression tests firmly established the principles of feedback regulation and formed the basis for many current diagnostic algorithms. RIAs also revealed the natural patterns of hormone secretion, including circadian rhythms and reproductive cycles, as well as hormonal responses to sleep, meals, stress, exercise, and other daily life events. Our understanding of hormone action has been accelerated by studies of their membrane and nuclear receptors, which convey hormone specificity in target tissues. The signaling pathways elicited by these receptors constitute intricate and complex networks that inform the cell about its external environment. Recombinant DNA technology has been essential for cloning the genes and cDNAs that encode large families of hormones and receptors. Growth hormone, chorionic gonadotropin, and

somatostatin were among the first mammalian cDNAs to be cloned. With completion of the human genome project, all of the genes that encode hormones and their receptors have, in principle, been identified. However, many of these receptors remain ‘‘orphans’’ with still-unknown ligands and incompletely defined functions. Not surprisingly, genetic advances have revealed remarkable insight into inherited endocrine disorders. Most physicians are attracted to the field of endocrinology because it so beautifully integrates physiology, biochemistry, and cell signaling with patient care. Clinical manifestations of endocrine disorders can usually be explained by understanding the physiologic role of hormones—whether deficient or excessive. The conceptual framework for understanding hormone secretion, hormone action, and principles of feedback control provides the clinician with a logical diagnostic approach that typically employs appropriate laboratory testing and/or imaging studies. The fact that many endocrine disorders are amenable to cure or effective treatment also makes the practice of endocrinology especially satisfying. Because most glands are inaccessible to physical examination, endocrinologists are trained to detect key features of the medical history and subtle physical signs that point toward true endocrine disease. Increasingly, the challenge is to identify endocrine disorders at their earliest stages rather than when the clinical manifestations are obvious. Terms such as subclinical hypothyroidism, impaired glucose tolerance, and incidental adrenal or pituitary adenoma have crept into our vocabulary and have changed our approach to patients. Laboratory testing takes on added importance as we attempt to diagnose more subtle forms of disease. Building on this strong foundation of basic science, the knowledge base in endocrinology continues to change rapidly. In addition to the dramatic advances

xxix

xxx generated from genetics and molecular biology, the field has benefited from the introduction of an unprecedented number of new drugs, particularly for the management of diabetes and osteoporosis. Common diseases such as diabetes, hypertension, obesity, and osteoporosis have also been the subject of numerous large-scale clinical trials that provide a powerful evidence base for medical decision-making. The rapid changes in medicine mandate that physicians continuously update their knowledge base and clinical skills. The Encyclopedia of Endocrine Diseases recognizes this challenge and provides a remarkable compilation of current knowledge in basic and clinical endocrinology. This ambitious fourvolume set provides nearly 500 articles on basic and clinical endocrinology. The topics range from classic endocrine subjects such as hypothyroidism and acromegaly to new dimensions of the field including adipocytokines, ghrelin, and the role of the aldosterone receptor in cardiovascular disease. The international group of authors are experts in their topics, which have been subdivided to provide in-depth coverage. Thus, acromegaly is separated into articles on clinical features, diagnosis, and therapy to provide the level of detail needed to manage the most challenging cases. The standardized format and clear illustrations help to quickly offer answers. It is difficult to imagine an endocrine topic not covered in this encyclopedia, which provides a new, comprehensive reference for the daunting body of knowledge in endocrinology. With a four-volume Encyclopedia of Endocrine Diseases in hand, it is interesting to speculate about

Foreword

the remaining big questions and future discoveries in endocrinology. What are the genetic and adaptive elements that cause such a broad normal range for hormone values? It is humbling to recognize that we still have an incomplete understanding of the hormonal control of fundamental processes such as the onset of puberty, appetite control, gonadal differentiation into testes or ovaries, islet cell regeneration, insulin resistance, and causes of autoimmune endocrine disease. We still have much to learn about the optimal way to deliver many hormone therapies to mimic normal physiology. This topic is prominent in our efforts to provide intensive insulin replacement, or to replace growth hormone or cortisol, such that the beneficial effects outweigh complications. New therapies, such as intermittent PTH for osteoporosis, will be subjected to additional clinical trials, and one can easily imagine emerging questions about how to cycle the therapy and how to use it in relation to other treatments that alter osteoblast and osteoclast function. Gene transfer and stem cell strategies provide promising treatments for disorders such as diabetes and osteoporosis. The Encyclopedia of Endocrine Diseases can help to foster these discoveries by keeping researchers and clinicians at the cutting edge of endocrinology. J. Larry Jameson, M.D., Ph.D. Irving S. Cutter Professor and Chairman Department of Medicine Feinberg School of Medicine Northwestern University Chicago, Illinois, United States

Preface

Among the most complex constructs in the body, the endocrine system comprises a group of glands that secrete hormones directly into the bloodstream, together with the receptors for these hormones and the intracellular signaling pathways they invoke. The endocrine system maintains and regulates stable functioning by using hormones to control metabolism, temperature, biological cycles, internal fluid volume, reproduction, growth, and development. Fortunately, the system is a marvel when functioning optimally. However, the ways in which its processes, actions, and functions may go awry are myriad. The Encyclopedia of Endocrine Diseases is not meant as a primer on the subject of endocrinology, but instead is intended to provide a comprehensive reference work on the extensive spectrum of diseases and disorders that can occur within the endocrine system. This groundbreaking encyclopedia is especially timely, as there have been dramatic discoveries in the field of endocrinology over the past 10 to 20 years, particularly with respect to diagnosis techniques and treatment methods. Indeed, during the time since the encyclopedia was conceived, new hormones have been named. To bring a major reference work of such broad scope from initial conception to final publication involved a great deal of planning, staging, and organization, together with the efforts of innumerable individuals. At the start, the broadest possible list of topics was compiled and a distinguished multinational panel of 14 associate editors was assembled. Throughout the editorial process, the editors supervised their subject area of expertise, recommended and corresponded with article contributors, reviewed the subsequent manuscripts, and continuously helped to refine the topics list. The encyclopedia is intended to serve as a useful and comprehensive source of information spanning

the many and varied aspects of the endocrine system. It consists of nearly 500 topics explored by some 800 eminent clinicians and scientists from around the world, a veritable who’s who of endocrine research. Here the interested reader can find articles on newly discovered hormones such as ghrelin and leptin; articles about such maladies as hypertension, hypoglycemia, diabetes, cancer, osteoporosis, kidney stones, Graves’ disease, Paget’s disease, Alzheimer’s disease, Noonan syndrome, Langerhans cell disease, Cushing’s syndrome, thyroid and pituitary disorders; and articles dealing with subjects ranging from the evolution of the endocrine systems, the mechanisms of hormone action, and the endocrine failure in aging to the integration between the nervous and the endocrine systems. Written to be accessible to both the clinical and nonclinical reader, all of the articles are formatted in similar fashion and each is intended as a stand-alone presentation. Beginning each article is a glossary list defining key terms that may be unfamiliar to the reader and are important to an understanding of the article. The body of the article begins with a brief introduction to the subject under discussion, bold headings lead the reader through the text, and figures and tables explain and illuminate most articles. Following the article are reference citations to provide the reader with access to further in-depth considerations of the topic and cross-references to related entries in the encyclopedia. A compilation of all glossary terms appearing in the complete four-volume work is presented in the final volume as a dictionary of subject matter relevant to the endocrine system and its disorders. It is my hope that the Encyclopedia of Endocrine Diseases proves to be a valuable resource to a deservedly diverse readership, and particularly to students, many of whom it may well attract to the rewarding field of

xxxi

xxxii endocrinology. The project would not have been possible without cooperation, coordination, and reliance on e-mail among the key people, who were located in Japan, The Netherlands, Denmark, Switzerland, Italy, and the United States. I am greatly indebted to the dedicated and unstinting efforts of my associate editors, as well as the diligence and generosity of spirit of the Elsevier/Academic Press personnel who shepherded the project: Tari Paschall, Chris Morris, Carolan Gladden, and Joanna Dinsmore. To all of our

Preface

contributors go profound thanks for investing time and energy to produce their articles, which together have made the encyclopedia.

Luciano Martini Editor-in-Chief Professor of Endocrinology University of Milan Milan, Italy

A

ABCA1 Defects Gerd Schmitz and Wolfgang Drobnik University of Regensburg, Regensburg, Germany

Glossary ATP-binding cassette (ABC) transporters A family of multispan transmembrane proteins that mediate the active uptake or efflux of specific substrates across biological membrane systems. high-density lipoproteins Plasma lipoprotein subclass mediating reverse cholesterol transport. lipid microdomain Plasma membrane compartments with a specific lipid composition characterized by a distinct physical state and insolubility in certain nonionic detergents.

p0005

ABCA1

belongs to the family of ATP-binding cassette transporters, which represent one of the biggest multispan membrane protein families described by the early 21st century.

INTRODUCTION p0010

ATP-binding cassette (ABC) transporters have attracted much attention because mutations in these molecules are the cause of various human inherited diseases. Functional ABC transporters usually consist of two transmembrane domains (TMDs) and two nucleotide-binding domains (NBDs) or ABCs that are present either in one polypeptide chain (full-size transporter) or in two polypeptides (half-size transporter). A signature motif located between both ABCs is characteristic of each of the seven ABC subfamilies (ABCA–ABCG) described by the early 21st century. The majority of these proteins mediate the active uptake or efflux of specific substrates across various biological membrane systems, whereby two different groups of ABC proteins can be distinguished by their mode of action. The one group of ABC transporters (e.g., ABCB1 (MDR1), ABCC1 (MRP1)) has strong ATPase activity, and the resulting free energy is directly coupled to the movement of molecules across membranes, whereas the other group (e.g., ABCC7 (CFTR), ABCC9 (SUR2), ABCA1) is characterized by very low

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

ATP hydrolysis but conformational change subsequent to ATP binding, which is linked to regulatory processes rather than direct transport-pump activity. A multitude of substrates is transported by the various ABC family members, including glutathione, glucuronate, or sulphate conjugates; xenobiotics; peptides; nucleotides; ions; and various lipid species. Members of the ABCA family especially are involved in the transport of steroids and various phospholipid and sphingolipid species, and the transcriptional control of at least seven ABCA members is controlled or influenced by lipids. These data indicate an important role of the whole ABCA subfamily in cellular lipid transport processes.

ABCA1 DEFECTS ABC defects cause high-density lipoprotein (HDL) deficiency syndromes. Lipid-rich a-HDL originates from lipid-poor discoidal apolipoprotein–phospholipid complexes, with apo AI being the main apolipoprotein (Fig. 1A). Apo AI binds to cells and promotes vesicle transport and exocytosis of cholesterol and phospholipids. Cholesterol acquisition followed by lecithin:cholesterol-acyltransferase (LCAT)-mediated esterification results in the formation of lipid-rich spherical a-HDL. The cholesteryl esters associated with these mature HDL particles could be removed from the circulation by a scavenger receptor BI-mediated selective uptake into hepatocytes. Disorders of HDL metabolism could result from mutations in various genes along this metabolic pathway, including LCAT, apo AI, and ABCA1. ABCA1 is a 2261 amino acid protein with a molecular weight of 220 kDa that is expressed in a multitude of human organs, including liver, adrenal tissues, placenta, and spleen. Mutations in the ABCA1 gene have been identified as a cause for Tangier disease and other HDL deficiencies. Tangier disease is a rare inherited disorder of HDL metabolism. It was initially described by D. S. Fredrickson in 1961 and has since been diagnosed in at least 70 patients from 60 families. The patients are characterized by a complete deficiency of a-HDL and a severe

1

2

ABCA1 Defects

Figure 1 (A) High-density lipoprotein metabolism. (B) ABCA1: A regulator of lipid rafts, vesicular transport, and filipodia formation.

reduction of apo AI to 1 to 3% of normal, accompanied by low plasma cholesterol and normal or elevated triglycerides. The main clinical signs include the accumulation of cholesteryl esters in various tissues, hyperplastic orange tonsils, splenomegaly, and relapsing neuropathy. In addition, some Tangier patients have premature coronary artery disease (CAD), whereas others, even those over 60 years of age, are without any clinical symptoms of CAD. The clinical phenotype of Tangier disease and the biochemical features (e.g., low HDL) are inherited in an autosomal recessive mode and a co-dominant mode, respectively. In 1998, the genetic defect in Tangier disease was confined to chromosome 9q31, followed by the

demonstration that ABCA1 that is contained within this candidate region is subject to sterol-dependent regulation. Subsequent studies have shown that homozygous mutations in ABCA1 are the underlying defect in Tangier disease, whereas heterozygous mutations are found in patients with the more frequent and less severe familial HDL deficiency, which is inherited in a dominant mode and lacks clinical features of Tangier disease (i.e., at least a subgroup of familial HDL deficiency patients are Tangier heterozygotes). The essential role of ABCA1 in the regulation of HDL metabolism was further supported by demonstrating that targeted disruption of the ABCA1 gene in mice produces a phenotype similar to human Tangier

3

ABCA1 Defects

disease, whereas overexpression of ABCA1 increased HDL cholesterol levels. ABCA1-deficient cells are characterized by a complete loss of apo AI-mediated cholesterol and phospholipid efflux, indicating that ABCA1 is critically involved in the first step of reverse cholesterol transport. Without the initial ABCA1dependent lipidation, apo AI undergoes rapid degradation, explaining the extremely low apo AI levels in Tangier patients and the increased catabolism of apo AI and HDL in Tangier patients infused with radiolabeled normal HDL. In BAC transgenic mice overexpressing ABCA1, a highly significant correlation was observed between the increase in cholesterol efflux from various tissues and the elevation of plasma HDL. Similarly, in patients with heterozygote ABCA1 mutations, the reductions in cholesterol efflux and plasma HDL were closely correlated. Together, these data provided direct evidence that ABCA1dependent cholesterol efflux is a major determinant of plasma HDL cholesterol levels.

ABCA1: A REGULATOR OF APO AI-MEDIATED LIPID EFFLUX The exact cellular processes facilitating and regulating ABCA1-dependent lipid efflux are the subject of intense investigation. Data suggest that apo AI interacts either directly with ABCA1 or with lipid domains in close proximity to ABCA1. This interaction stimulates a Golgi- and energy-dependent vesicular transport process, resulting in the translocation of intracellular cholesterol and phospholipids to sites accessible to the apolipoprotein. Data also suggest that this apo AI-mediated lipid efflux is a two-step mechanism, with an initial ABCA1-dependent efflux of phospholipids and a subsequent ABCA1-independent efflux of cholesterol to the newly formed apo AI– phospholipid complex. Moreover, it has been shown that ABCA1 rapidly recycles between the cell surface and the intracellular compartments, although it is currently unclear whether this recycling is involved in the lipid transport process or instead regulates synthesis and degradation of ABCA1. Regarding the direct function of ABCA1, it was initially assumed that ABCA1 functions as an active pump translocating cholesterol from the inner leaflet of the plasma membrane to the outer leaflet, where it is accessible to the uptake by apo AI. However, in contrast to ABC transporters that exert bona fide pump function (e.g., MDR-1), ABCA1, similar to the ABC regulator proteins CFTR and SUR, shows only marginal intrinsic ATPase activity. Thus, ABCA1 may act as a transport

facilitator rather than as an active pump. Interestingly, both the C terminus of CFTR and ABCA1 contain a PDZ domain-binding sequence. By using the yeasttwo-hybrid system, we demonstrated a direct interaction of ABCA1 with the PDZ domain-containing protein b2-syntrophin. Immunoprecipitation confirmed these results and identified utrophin as an ABCA1 interaction partner. In analogy to the function of b2-syntrophin–utrophin complexes in anchoring insulin-containing secretory granules, it is tempting to speculate that the interaction of ABCA1 with b2syntrophin–utrophin regulates the availability of ABCA1 at the cell surface. Additional ABCA1 interacting proteins include components of t-SNARE complexes, which are involved in targeted vesicle transport and are also known to interact with the N terminus of CFTR.

REGULATION OF ABCA1 EXPRESSION AND FUNCTION Several factors have been shown to control the expression of ABCA1. Since the initial finding that cholesterol influx into the cell potently induces ABCA1 expression, a number of transcriptional control elements have been characterized. Tissue-specific regulation of ABCA1 is controlled by the transcription factors Sp1/3, USF1/2, and HNF-1a, and considerable attention has been paid to nuclear liver X receptors (LXR) as inducers of ABCA1 expression in response to lipid loading. In addition, the zinc finger protein ZNF202 appears to function as a major repressor of ABCA1 transcription, and the oncostatin M-induced ABCA1 transcription provides a new concept for how members of the IL6 family of cytokines may regulate lipid transport proteins. Additional regulators of the ABCA1-dependent lipid efflux pathway include cAMP, phospholipase C, phospholipase D, and bioactive sphingolipids such as ceramide, sphingosine, and sphingosine-1-phosphate. The effects of these signaling pathways are probably cell type dependent, and further work is necessary to determine the exact mechanisms by which they control ABCA1-dependent cell function.

ABCA1 AND SUSCEPTIBILITY TO ATHEROSCLEROSIS Considering the known reverse relationship between HDL cholesterol levels and the risk of premature CAD, several groups have investigated the role of ABCA1 in atherogenesis. Thus, it has been reported

4 that homozygote and heterozygote mutations in ABCA1 are associated with an increased prevalence of premature CAD that correlates to the reduction in HDL cholesterol. Furthermore, it has been demonstrated that diet-induced development of atherosclerotic lesions is significantly reduced in transgenic mice overexpressing ABCA1. In contrast, complete inactivation of ABCA1 in apo E-/- and LDL-/- mice had no effect on the development of atherosclerotic lesions, although it markedly reduced HDL cholesterol. It was suggested that the proposed atherogenic effect of complete ABCA1 deficiency may be compensated by a less atherogenic lipid profile, a hypothesis that may also partially explain the lack of premature atherosclerosis in a significant number of Tangier patients. Importantly, two independent studies showed that targeted disruption of ABCA1 in leukocytes of LDL-/- or apo E-/- mice resulted in the development of more advanced atherosclerotic lesions without significantly affecting HDL levels. Together, these data indicate that ABCA1 clearly serves an anti-atherogenic function, although this may involve properties of ABCA1 that are independent of plasma lipids and HDL levels. Several factors may account for the protective effect of ABCA1 in atherogenesis. First, ABCA1-mediated cholesterol efflux may significantly compensate excessive cholesterol uptake by macrophages in the vessel wall without significantly influencing plasma HDL levels. Second, ABCA1 has been implicated in the engulfment of apoptotic cells by macrophages. Thus, it is conceivable that the ABCA1-mediated phagocytic activity of lesion macrophages may counteract excessive accumulation of apoptotic material that, in return, may stimulate the inflammatory response within the vascular wall. Finally, we have previously hypothesized that ABCA1 function regulates the differentiation, lineage commitment (phagocytic vs dendritic cells), and targeting of monocytes into the vascular wall or the RES. This concept has been substantiated by recent work from our laboratory demonstrating accumulation of macrophages in liver and spleen in LDL receptor-deficient mouse chimeras that selectively lack ABCA1 in their blood cells. The fact that the absence of ABCA1 from leukocytes is sufficient to induce aberrant monocyte recruitment into specific tissues identifies ABCA1 as a critical leukocyte factor in the control of monocyte targeting. An interesting clue as to how ABCA1 may be implicated in the control of monocyte/ macrophage trafficking at the cellular level comes from the observation that apo AI-mediated lipid efflux is paralleled by the down-regulation of the protein Cdc42 and filipodia formation, which may mitigate monocyte recruitment to the artery wall.

ABCA1 Defects

ABCA1: A REGULATOR OF MEMBRANE PROTRUSIONS AND LIPID MICRODOMAINS CDC42 is a member of the family of small GTPbinding proteins that controls a wide range of cellular functions, including cytoskeletal modulation, formation of filipodia, and vesicular processing. Similar to ABCA1, the protein expression of CDC42 is increased by cholesterol loading of monocytes, whereas deloading by apo AI and HDL has the opposite effect. These changes in CDC42 expression are paralleled by alterations in M-CSF-induced filipodia formation and fMLP-induced chemotaxis, with an increase on ELDL-mediated cholesterol loading and a decrease in response to apo AI and HDL. ABCA1-deficient monoyctes from Tangier patients showed reduced filipodia formation and decreased CDC42 expression. Thus, the ABCA1 pathway is linked to the formation of membrane protrusions, which may be of significant relevance for the anti-atherogenic effects of apo AI and HDL. Further evidence for this functional link was provided by Matsuzawa and colleagues, who demonstrated that overexpression of ABCA1 in HEK293 cells induces formation of filipodia and long membrane protrusions. These ABCA1 effects and the apo AI-mediated lipid efflux in MDCK cells were significantly reduced by a dominant negative form of CDC42. Together with the finding that ABCA1 coimmunoprecipitated with CDC42, the data suggest a role for CDC42 as a downstream mediator of ABCA1 function. Moreover, it could be imagined that the principal function of ABCA1 is to facilitate the supply of choline-phospholipids and cholesterol for newly emerging plasma membrane extensions, which in the presence of apo AI are transferred to the extracellular acceptor rather than being used for the formation of new membrane areas (Fig. 1B). This hypothesis would be in accordance with the slightly impaired intestinal cholesterol absorption observed in Abca1-/- mice, a process that involves the microvilli of the enterocyte brush boarder membrane. Further support is derived from findings that apo AI preferentially depletes cholesterol and phospholipids from a novel type of cholesterol-based microdomain called Lubrol raft. Ro¨per and colleagues showed that these lipid microdomains are building units for different forms of plasma membrane protrusions and that CDC42 and ABCA1 were partially localized to these domains. In fibroblasts, apo AI-induced lipid efflux also involved classical Triton X-100 rafts, and independent of the cell type, these domains were also modified by spherical HDL3 (Fig. 1B). Because Triton rafts are recognized as

ABCA1 Defects

platforms of signal transduction and cell regulation, the observed effect may be of major importance for the role of HDL in atherogenesis.

See Also the Following Articles Atherogenesis . Hypercholesterolemias, Familial Defective ApoB (FDB) and LDL Receptor Defects . Low HDL/High HDL Syndromes

Further Reading Aiello, R. J., Brees, D., Bourassa, P. A., Royer, L., Lindsey, S., Coskran, T., Haghpassand, M., and Francone, O. L. (2002). Increased atherosclerosis in hyperlipidemic mice with inactivation of abca1 in macrophages. Arterioscler. Thromb. Vasc. Biol. 22, 630–637. Bodzioch, M., Orso, E., Klucken, J., Langmann, T., Bottcher, A., Diederich, W., Drobnik, W., Barlage, S., Buchler, C., PorschOzcurumez, M., Kaminski, W. E., Hahmann, H. W., Oette, K., Rothe, G., Aslanidis, C., Lackner, K. J., and Schmitz, G. (1999). The gene encoding ATP-binding cassette transporter 1 is mutated in Tangier disease. Nat. Genet. 22, 347–351. Brooks-Wilson, A., Marcil, M., Clee, S. M., Zhang, L. H., Roomp, K., van Dam, M., Yu, L., Brewer, C., Collins, J. A., Molhuizen, H. O., Loubser, O., Ouelette, B. F., Fichter, K., AshbourneExcoffon, K. J., Sensen, C. W., Scherer, S., Mott, S., Denis, M., Martindale, D., Frohlich, J., Morgan, K., Koop, B., Pimstone, S., Kastelein, J. J., and Hayden, M. R. (1999). Mutations in ABC1 in Tangier disease and familial high-density lipoprotein deficiency. Nat. Genet. 22, 336–345. Costet, P., Luo, Y., Wang, N., and Tall, A. R. (2000). Steroldependent transactivation of the ABC1 promotor by the liver X receptor/retinoid X receptor. J. Biol. Chem. 275, 28240– 28245. Dean, M., Hamon, Y., and Chimini, G. (2001). The human ATPbinding cassette (ABC) transporter superfamily. J. Lipid Res. 42, 1007–1017. Diederich, W., Orso, E., Drobnik, W., and Schmitz, G. (2001). Apolipoprotein Al and HDL(3) inhibit spreading of primary human monocytes through a mechanism that involves cholesterol depletion and regulation of CDC42. Atherosclerosis 159, 313–324. Drobnik, W., Borsukova, H., Bottcher, A., Pfeiffer, A., Liebisch, G., Schutz, G. J., Schindler, H., and Schmitz, G. (2002). ATPbinding cassette transporter A1 (ABCA1) affects total body sterol metabolism. Traffic 3, 268–278. Drobnik, W., Lindenthal, B., Lieser, B., Ritter, M., Christiansen, W. T., Liebisch, G., Giesa, U., Igel, M., Borsukova, H., Buchler, C., Fung-Leung, W. P., Von Bergmann, K., and Schmitz, G. (2001). Apo Al/ABCA1-dependent and HDL3-mediated lipid

5 efflux from compositionally distinct cholesterol-based microdomains. Gastroenterology 120, 1203–1211. Joyce, C. W., Amar, M. J., Lambert, G., Vaisman, B. L., Paigen, B., Najib-Fruchart, J., Hoyt, R. F., Jr., Neufeld, E. D., Remaley, A. T., Fredrickson, D. S., Brewer, H. B., Jr., and SantamarinaFojo, S. (2002). The ATP binding cassette transporter A1 (ABCA1) modulates the development of aortic atherosclerosis in C57BL/6 and apoE-knockout mice. Proc. Natl. Acad. Sci. USA 99, 407–412. Klucken, J., Buchler, C., Orso, E., Kaminski, W. E., PorschOzcurumez, M., Liebisch, G., Kapinsky, M., Diederich, W., Drobnik, W., Dean, M., Allikmets, R., and Schmitz, G. (2000). ABCG1 (ABC8), the human homolog of the Drosophila white gene, is a regulator of macrophage cholesterol and phospholipid transport. Proc. Natl. Acad. Sci. USA 97, 817–822. Langmann, T., Porsch-Ozcurumez, M., Heimerl, S., Probst, M., Moehle, C., Taher, M., Borsukova, H., Kielar, D., Kaminski, W. E., Dittrich-Wengenroth, E., and Schmitz, G. (2002). Identification of sterol-independent regulatory elements in the human ATP-binding cassette transporter A1 promoter. J. Biol. Chem. 277, 14443–14450. Orso, E., Broccardo, C., Kaminski, W. E., Bottcher, A., Liebisch, G., Drobnik, W., Gotz, A., Chambenoit, O., Diederich, W., Langmann, T., Spruss, T., Luciani, M. F., Rothe, G., Lackner, K. J., Chimini, G., and Schmitz, G. (2000). Transport of lipids from golgi to plasma membrane is defective in tangier disease patients and Abc1-deficient mice. Nat. Genet. 24, 192–196. Porsch-Ozcurumez, M., Langmann, T., Heimerl, S., Borsukova, H., Kaminski, W. E., Drobnik, W., Honer, C., Schumacher, C., and Schmitz, G. (2001). The zinc finger protein 202 (ZNF202) is a transcriptional repressor of ATP binding cassette transporter A1 (ABCA1) and ABCG1 gene expression and a modulator of cellular lipid efflux. J. Biol. Chem. 276, 12427–12433. Roper, K., Corbeil, D., and Huttner, W. B. (2000). Retention of prominin in microvilli reveals distinct cholesterol-based lipid micro-domains in the apical plasma membrane. Nat. Cell Biol. 2, 582–592. Rust, S., Rosier, M., Funke, H., Real, J., Amoura, Z., Piette, J. C., Deleuze, J. F., Brewer, H. B., Duverger, N., Denefle, P., and Assmann, G. (1999). Tangier disease is caused by mutations in the gene encoding ATP-binding cassette transporter 1. Nat. Genet. 22, 352–355. Szakacs, G., Langmann, T., Ozvegy, C., Orso, E., Schmitz, G., Varadi, A., and Sarkadi, B. (2001). Characterization of the ATPase cycle of human ABCA1: implications for its function as a regulator rather than an active transporter. Biochem. Biophys. Res. Commun. 288, 1258–1264. Tsukamoto, K., Hirano, K., Tsujii, K., Ikegami, C., Zhongyan, Z., Nishida, Y., Ohama, T., Matsuura, F., Yamashita, S., and Matsuzawa, Y. (2001). ATP-binding cassette transporter-1 induces rearrangement of actin cytoskeletons possibly through Cdc42/N-WASP. Biochem. Biophys. Res. Commun. 287, 757–765. Wang, N., Silver, D. L., Thiele, C., and Tall, A. R. (2001). ATPbinding cassette transporter A1 (ABCA1) functions as a cholesterol efflux regulatory protein. J. Biol. Chem. 276, 23742–23747.

Abetalipoproteinemia John R. Wetterau and Richard E. Gregg Bristol-Myers Squibb, Princeton, New Jersey, United States

Glossary chylomicron Triglyceride-rich lipoprotein made in the intestine that transports dietary lipid to various tissues throughout the body. microsomal triglyceride transfer protein (MTP) Heterodimeric lipid transfer protein found within the lumen of the endoplasmic reticulum of enterocytes and hepatocytes that promotes the transport of triglyceride and cholesteryl ester between phospholipid membranes in in vitro lipid transfer assays. protein disulfide isomerase Multifunctional redox chaperone protein found within the lumen of the endoplasmic reticulum that promotes the proper folding of newly synthesized secretory proteins that contain disulfide bonds. very low-density lipoprotein (VLDL) Triglyceriderich lipoprotein made in the liver and intestine that transports lipid to peripheral tissues.

A

betalipoproteinemia is an autosomal recessive disease caused by mutations in the gene encoding the microsomal triglyceride transfer protein. Affected individuals have defects in the production of plasma lipoproteins that contain apolipoprotein B: chylomicrons, very low-density lipoproteins, and low-density lipoproteins. As a result of the defect, subjects have plasma cholesterol and triglyceride levels of approximately 40 and 10mg/dl, respectively. Neuromuscular and retinal degeneration occurs due to a secondary deficiency in vitamin E, a fatsoluble vitamin that depends on lipoproteins for its absorption and transport throughout the body.

INTRODUCTION p0010

In 1950, F. A. Bassen and A. L. Kornzweig diagnosed a patient with retinitis pigmentosa, malformed erythrocytes, celiac disease, and ataxia. This disease was later named abetalipoproteinemia when H. B. Salt and colleagues associated a similar syndrome with an absence of plasma lipoproteins with beta electrophoretic

6

mobility. Subsequently, it was found that many of the pathological consequences of the disease, particularly the neurological findings, were related to vitamin E deficiency. In the 1990s, the molecular basis of the disease was elucidated when a link between mutations in the gene encoding the microsomal triglyceride transfer protein [MTP; a lipid transfer protein located at the sites of chylomicron and very low-density lipoprotein (VLDL) assembly], a defect in lipoprotein assembly, and abetalipoproteinemia was established, thus demonstrating that the proximal cause of abetalipoproteinemia is a mutation in the MTP gene.

ABETALIPOPROTEINEMIA IS CAUSED BY MUTATIONS IN THE MTP GENE Overview of Lipoprotein Metabolism Plasma lipids are transported throughout the body in lipid–protein complexes. Chylomicrons and VLDL are triglyceride-rich emulsions surrounded by a monolayer of phospholipid, free cholesterol, and protein. They transport dietary and endogenously synthesized triglyceride, respectively, to peripheral tissues, where it is hydrolyzed to produce free fatty acids that can be used as an energy source or stored as fat in adipocytes. Following lipolysis, chylomicron remnants are rapidly cleared from plasma. Some VLDL remnants are cleared directly from plasma, but a portion are converted to cholesterol and cholesteryl ester-rich LDLs, which are subsequently removed from plasma by a receptor-mediated process. The primary protein component of VLDL is apolipoprotein B (apoB)-100 (4536 amino acids). The primary protein component of chylomicrons is apoB-48 (2152 amino acids), which is encoded by an edited version of the mRNA that encodes apoB-100. High-density lipoproteins (HDLs) are small, dense particles secreted directly from the liver or made from excess surface components of chylomicrons and VLDL following hydrolysis of their triglyceride

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

p0015

7

Abetalipoproteinemia

core. HDLs play an important role in the transport of cholesterol from peripheral tissues to the liver, where it can be processed and transported out of the body.

Genetic Link between a Defect in MTP, a Defect in Lipoprotein Assembly, and Abetalipoproteinemia p0025

The MTP is a heterodimeric lipid transfer protein found in the lumen of the endoplasmic reticulum and Golgi apparatus. One subunit is protein disulfide isomerase, a 58-kDa multifunctional redox chaperone protein that catalyzes the proper folding of newly synthesized proteins that contain disulfide bonds within the endoplasmic reticulum. The second subunit is a novel 97-kDa subunit that confers the lipid transfer activity to the protein complex. In lipid transfer assays utilizing synthetic membrane substrates, MTP accelerates the transfer of triglyceride, cholesteryl ester, and, to a lesser extent, phospholipid between membranes. Intestinal biopsies from abetalipoproteinemic subjects were found to be devoid of MTP activity and the unique large subunit of MTP. Following the cloning of the MTP large subunit, various missense, nonsense, frameshift, and splice site mutations were identified in the gene encoding the MTP large subunit in the patients studied. All mutations were either homozygous or compound heterozygous and explained the complete absence of functional MTP protein, consistent with the autosomal recessive transmission of the disease. These and subsequent studies of specific inhibitors of MTP lipid transfer activity established that MTP is required for the assembly of apoBcontaining lipoproteins, and that defects in MTP are the proximal cause of abetalipoproteinemia.

PATHOLOGICAL CONSEQUENCES OF A DEFECT IN MTP Clinical Description of Abetalipoproteinemia Abetalipoproteinemic subjects have fat malabsorption that results in abdominal discomfort, diarrhea, and steatorrhea following a meal with a normal fat content. Intestinal enterocytes are fat laden. Subjects also have malabsorption of fat-soluble vitamins—vitamins E, A, K, and, to a lesser extent, D. Plasma cholesterol and triglyceride levels are very low (approximately 40 and 10mg/dl, respectively), and plasma triglyceride levels do not increase following a meal. Apolipoprotein B,

the major protein component of VLDL and chylomicrons, is virtually absent. Although HDL levels may be only moderately decreased, they have an abnormal composition, including an elevated total cholesterol, free cholesterol/esterified cholesterol ratio, and sphingomyelin content. Half or more of the erythrocytes in abetalipoproteinemic subjects have irregular cytoplasmic projections. These abnormal erythrocytes, referred to as acanthocytes, are thought to arise from an altered membrane composition. A moderate anemia may occur due to hemolysis and a shortening of the red blood cell resident time in the circulation. Coagulation abnormalities secondary to a deficiency of vitamin K may occur. Liver biopsies show evidence of steatosis. Elevated transaminase levels and fibrosis have been reported. There are also rare reports of serious liver pathology, including cirrhosis. The neurological abnormalities are usually the most severe consequence of the syndrome and include spinal–cerebellar degeneration and peripheral neuropathies, which can lead to a loss of reflexes, altered sensation, muscle weakness, and ataxia that can progress to a point at which the affected individual is unable to walk. Degenerative pigmentary retinopathy, retinitis pigmentosa, leads to decreased night and color vision. If left untreated, this will progress to blindness.

p0040

Treatment The gastrointestinal side effects of the disease can be controlled by avoiding high-fat meals. This is particularly important early in life, when severe gastrointestinal side effects and malabsorption can result in a failure to thrive. Therapy for abetalipoproteinemic subjects includes fat-soluble vitamin supplements, including massive levels of vitamin E (up to 20,000mg/ day) and more usual replacement doses of vitamins A and K. Vitamin D does not usually need to be supplemented. The neurological and ophthalmological manifestations of abetalipoproteinemia are similar to those found in animals with vitamin E deficiency. Vitamin E, which plays an important role in preventing lipid peroxidation, requires chlyomicrons and VLDL for efficient absorption and transport throughout the body. Although plasma vitamin E levels remain far below normal following vitamin E therapy, tissue levels may increase to near normal levels. Vitamin E and A therapy may slow the progression or even stabilize the neurological and retinal consequences of the disease.

p0050

8

See Also the Following Articles Anderson’s Disease (Chylomicron Retention Disease) . Dysbetalipoproteinemia and Type III Hyperlipidemia . Familial Low Cholesterol Syndromes, Hypobetalipoproteinemia . Hypercholesterolemias, Familial Defective ApoB (FDB) and LDL Receptor Defects . Low HDL/High HDL Syndromes

Further Reading Berriot-Varoqueaux, N., Aggerbeck, L. P., Samson-Bouma, M.-E., and Wetterau, J. R. (2000). The role of the microsomal triglyceride transfer protein in abetalipoproteinemia. Annu. Rev. Nutr. 20, 663–697.

Abetalipoproteinemia

Gordon, D. A., and Jamil, H. (2000). Progress towards understanding the role of microsomal triglyceride transfer protein in apolipoprotein-B lipoprotein assembly. Biochim. Biophys. Acta 1486, 72–83. Kane, J. P., and Havel, R. J. (1995). Disorders of the biogenesis and secretion of lipoproteins containing the B apolipoproteins. In “The Metabolic and Molecular Bases of Inherited Disease” (C. R. Scriver, A. L. Beaudet, W. S. Sly, and D. Valle eds.), 7th ed., Vol. 2 pp. 1853–1885. McGraw-Hill, New York. Rader, D. J., and Brewer, H. B., Jr. (1993). Abetalipoproteinemia: New insights into lipoprotein assembly and vitamin E metabolism from a rare genetic disease. J. Am. Med. Assoc. 270, 865–869. Wetterau, J. R., Lin, M. C. M., and Jamil, H. (1997). Microsomal triglyceride transfer protein. Biochim. Biophys. Acta 1345, 136–150.

Acetylation Paolo Portaleone University of Turin, Turin, Italy

Glossary arylalkylamine Ammonia (NH3) derivative with aromatic (aryl) and aliphatic (alkyl) radicals. choline Alcoholic molecule derived from phosphatidylcholine (or lecitine) supplied by food. enzyme A protein or proteinaceous substance produced by a living cell that catalyzes a specific reaction. histone Eight proteic sequences (octamer) constituting the core of the nucleosome (chromatin unit) around which the DNA double helix is wrapped. peptide hormone Proteic endocrine messenger with low molecular weight, from 2 to several 10s of amino acid sequence linked by amino acid bond.

A

cetylation is a biochemical reaction, catalyzed by specific enzymes (acetyltransferases), that consists in the transfer of an acetyl radical (CH3–CO–) from a donor (e.g., acetyl coenzyme A) to an acceptor molecule.

INTRODUCTION Acetylation is a general metabolic reaction common to both the plant and animal kingdoms. Nevertheless, acetyltransferase enzyme systems are involved not only in the biotransformation of xenobiotics but also in activating–deactivating processes of endogenous active agents such as proteins (e.g., histones, peptide hormones), alcohols, amino acids, and amines (e.g., choline, 5-hydroxy-tryptamine). In this regulatory function, it is evident that acetylation holds a crucial role in basic processes of cellular life, from DNA transcription, replication, and repair to the control of selective messenger information through neurohormones and neurotransmitters. For the relevance of the acetyltransferase regulatory role, its involvement in several pathologies, from functional or degenerative diseases (e.g., headache, depression,

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

Parkinson’s disease, Alzheimer’s disease) to neoplastic growth, has been suggested.

HISTONE ACETYLATION The nucleosome, the fundamental unit of chromatin in eukaryotic cells, holds a core particle consisting of a proteic histone octamer, two copies each of H2A, H2B, H3, and H4, around which 146 bp of DNA is wrapped. The whole chromatin components are responsible for DNA dynamic behavior, but histone modifications have the main influence on DNA transcription, repair, and replication. Histones are modified by means of the addition of several chemical radicals such as phosphate, methyl, acetyl, ribosyl, and ubiquitin groups. Acetylation of core histones is associated with transcriptional activation. In contrast to cotranslational N-terminal a-acetylation of many proteins, histone acetylation occurs posttranslationally and reversibly on the e-NH3þ groups of highly conserved lysine residues of the N-terminal tails of core histones. The enzymes catalyzing reversible histone acetylation are histone acetyltransferases (HATs) and deacetylases (HDACs). HATs catalyze the transfer of the acetyl moiety from acetyl coenzyme A (Ac-CoA) to the e-NH3þ of lysine residues, whereas in the opposing deacetylation reaction, HDACs remove the acetyl groups and thereby reestablish the positive charge in the histones. Thus, acetylation neutralizes the positive charge increasing the hydrophobicity of the histones, leading to ‘‘opening up’’ the chromatin complex through a reduced affinity of acetylated N-terminal domains of histones to DNA. Therefore, the modification in nucleosomal structure is considered to play a causative role in activating transcription. Another possibility, or rather an additional effect, has been suggested: histone acetylation could represent a signal ‘‘flag’’ recognized by other molecules linked to transcriptional activation. Two different classes of HATs have been described: type A and type B. Type A HATs are localized in the

9

10 nuclei and acetylate nucleosomal histones leading to transcriptional activation, whereas type B HATs can be found in cytoplasmic fractions and acetylate newly synthesized histones before chromatin assembly during DNA replication. Several known transcriptional regulators in mammals have been found to possess intrinsic type A HAT activity, and among them the best understood family is the Gcn5-related N-acetyltransferase (GNAT) family. No type B HAT has been identified and characterized in mammals. Not only nuclear histones are substrates for the acetyltransferase, but also numerous nonhistone proteins involved in transcription regulation, such as p53, E2F1, EKLF, TFIIEb, TFIIF, TCF, GATA1, HMGI(Y), and ACTR, or even non-nuclear proteins, such as a-tubulin, are substrates for the acetyltransferase. There are three possible consequences of the acetylation processes, depending on where acetylation takes place within the protein: increased or decreased DNA binding, protein–protein interaction regulation, and protein stability. If there has been an explosion of studies on HAT activity during the past decade or so, only a few lines of evidence indicate the regulation of the enzymatic activity of acetylases. A bromodomain (a specific HAT protein structure) is present in many transcriptional activators with HAT activity, and it seems to be a requisite for targeting the enzyme to the substrate. The regulation of HAT activity is carried out by proliferation and differentiation signals by means of phosphorylation or hormonal signaling. Because acetylation can regulate such wide and different cellular functions, both nuclear and cytoplasmic (including the circadian clock in DNA transcription), it is obvious that its dysfunctioning could be at the origin of the different pathologies.

PEPTIDE HORMONE ACETYLATION N-terminal acetylation is a nearly selective posttranslational processing event among peptide hormones, and among the end products of the pro-opiomelanocortin (POMC) biosynthetic pathway, only a-melanocytestimulating hormone (a-MSH) and b-endorphin undergo this posttranslational modification. The relevance of the POMC-derived peptide N-acetylating mechanism, under a phylogenetic point of view, is supported by its persistence as an ‘‘ancestral’’ mechanism throughout vertebrate evolution. In mammals, a-MSH and b-endorphin, as final products of a set of cleavage reactions of POMC, were found not only in

Acetylation

secretory granules of pituitary neurointermediate lobe but also in anterior pituitary lobe and in neurons, mainly of the hypothalamic arcuate nucleus. The N-acetylation reaction, which requires acetylCoA as a coenzyme, occurs on the serine–NH2 terminal for a-MSH and on the tyrosine–NH2 terminal for b-endorphin. On the basis of the various aminoterminal targets of N-acetyltransferase (NAT) and its different regional distributions in the pituitary and brain, two distinct enzymes, an a-MSH-acetyltransferase (MAT) and a b-endorphin-acetyltransferase (EAT), have been proposed. Actually, two forms of NAT have been found. One enzyme, specifically localized in secretory granules of the neurointermediate lobe (NIL) of the pituitary gland with an optimal pH of 6.0 to 6.6, is inhibited by several solubilizing detergents and possesses similar characteristics in the acetylation process of both the serine of a-MSH and the tyrosine of b-endorphin. Therefore, this single NATcapable of acetylating the opioid and melanotropic peptides has been termed opiomelanotropin acetyltransferase (OMAT). A second enzyme, with an optimal pH of 7.4, is inhibited by Mg2þ, shows different anatomical and subcellular (cytosol) distribution, and has a more general acetyltransferase activity (GAT). The POMC-derived peptide NAT is coexpressed with the POMC gene and undergoes the same regulatory control of POMC synthesis by inhibitors and activators (e.g., glucocorticoids, sex steroids, and dopamine as inhibitors; adrenalectomy, castration, and dopamine receptor antagonists as activators). The N-acetylation of des-acetyl-a-MSH and b-endorphin substantially alters the physiological responses produced by both peptides. The acetylated form of a-MSH, in fact, is about 10 to 100 times more effective than its des-acetylated form in increasing arousal, memory, and attention in the Y-maze visual discrimination task and in eliciting excessive grooming. Moreover, the in vitro lipolytic activity on rabbit adipose tissue slices and the in vitro melanotropic activity in frog skin are markedly reduced after removing the acetyl group from monoacetyl-a-MSH. Conversely, the acetylation of b-endorphin completely eliminates the opiate analgesic activity of the peptide and markedly reduces its affinity in binding to opiate receptors. Furthermore, a substantial decrease of NAT activity has been observed during the lifetime, fitting in with the decreased concentrations of a-MSH found in rat aged brain. The physiological role of POMC-derived peptide NAT, with its ambivalent effects, activating des-acetyla-MSH, and deactivating b-endorphin, is very difficult to interpret, and the lacking characterization of the

11

Acetylation

enzyme, very unstable and ubiquitous, makes the task arduous. Considering that the POMC-derived peptides (adrenocorticotropin hormone [ACTH], a-MSH, and b-endorphin) are the main effectors in the ‘‘organized stress response’’ coordinating the biological and behavioral adaptive effects, we suggest that the deactivating– activating N-acetylation regulates the chronology of the adaptive response sequence with a rapid inactivation of ACTH and b-endorphin effects and with a potentiation of the a-MSH long-lasting adaptive activity. The progressively lower synthesis and activating– deactivating NAT activity of POMC-derived peptides during aging could be related to the reduced adaptive capabilities of aged individuals.

CHOLINE ACETYLATION Acetylcholine (ACh) is the neurotransmitter of the parasympathetic nervous system and is synthesized from choline and Ac-CoA in a single-step reaction catalyzed by the enzyme choline acetyltransferase (ChAT) that is expressed selectively in cholinergic neurons, where it serves as a phenotypic marker. The substrate choline, derived from phospholipids and the ACh hydrolysis, is taken up by the highaffinity sodium (Naþ)-dependent choline transporter. The newly synthesized neurotransmitter is accumulated in synaptic vesicles by means of a specific vesicular acetylcholine transporter (VAChT), a 12-transmembrane domain protein that uses the electrochemical gradient generated by a proton ATPase to exchange two protons by one ACh molecule. ChAT is encoded by a single gene and is coexpressed with VAChT. The gene of VAChT is embedded in the first intron of the ChAT gene. This unique organization was named ‘‘cholinergic gene locus,’’ suggesting reciprocal posttranslational regulation between the two proteins. In humans, four of the six identified transcripts translate to the same 69-kDa protein. The fifth and sixth transcripts yield 82- and 74-kDa forms of ChAT. Until now, the mechanisms regulating production of these different transcripts and their physiological roles have not been elucidated. ChAT exists in two forms in cholinergic nerve terminals: a soluble form (80–90% of the total enzyme activity) and a membrane-bound form (10–20%). Moreover, the 82-kDa ChAT has been localized to the nucleus, whereas the 69-kDa enzyme is largely cytosolic. Because it is unclear whether ACh synthesis occurs in the nucleus, different functional roles for 82-kDa ChAT must be considered. The 69-kDa ChAT is the form more represented and responsible for the majority of ACh biosynthesis.

The fact that ChAT is not saturated at the substrate concentrations in in vitro kinetic studies means that the enzyme would be in kinetic excess. Based on these in vitro data, it is accepted that the neuronal ChAT levels might not be rate limiting in ACh synthesis; therefore, the availability of the substrates would regulate ACh production. On the other hand, some data support the regulatory role of ChAT to maintain the homeostatic levels of ACh under some conditions of neuronal activity and demand for ACh synthesis. Short- and long-term regulation of ChAT activity has been described. Thomas Dobransky and R. Jane Rylett reviewed the role of phosphorylation in ChAT short-term regulation, showing that ChAT is phosphorylated by several protein kinases at various sites of ChAT primary structure in response to different functional states of neurons. Taken together, these lines of evidence indicate that phosphorylation of ChAT is physiologically significant and could serve as a regulatory mechanism. Neurotropic factors (neurotrophins) such as nerve growth factor (NGF) are responsible for long-term ChAT regulation, enhancing the enzyme expression and/or its activity. Clarifying the molecular mechanisms of ChAT regulation and its dysfunctions may be helpful in explaining the possible cellular mechanisms responsible for the loss of cognitive attributes associated with cholinergic deficit in Alzheimer’s disease.

ARYLALKYLAMINE N-ACETYLATION The acetylation of serotonin (5-hydroxy-tryptamine [5-HT]) in the pineal gland is another acetylation reaction essential to regulating the biological circadian rhythms in vertebrates by means of the production of the pineal hormone melatonin. Melatonin is synthesized from serotonin, which is acetylated into N-acetylserotonin (NAS) by the arylalkylamine N-acetyltransferase (AA-NAT) and then is methylated by the hydroxyindole-O-methyltransferase enzyme (HIOMT). The rate of O-methylation is largely a function of substrate availability, whereas the N-acetylation step is regulated by the amount and activation of AA-NAT protein. Melatonin synthesis occurs in darkness under the control of norepinephrine (NE) that is released from sympathetic fibers originating in the superior cervical ganglia and that is regulated, through a polysynaptic neuronal pathway, by a circadian oscillator located in the suprachiasmatic nuclei (SCN). NE activates pinealocytes acting on a1and b1-adrenergic receptors and subsequent increases in the intracellular concentration of calcium ions ([Ca2þ]i) and cyclic AMP (cAMP), respectively, leading

p0115

12

Acetylation

a strong AA-NAT activation at both the trascriptional and posttrascriptional levels. cAMP-increased levels induce the phosphorylation protein kinase Adependent of the stimulatory transcription factor cAMP response element-binding protein (CREB). Phosphorylated CREB binds to a c-AMP response element (CRE) element in the AA-NAT gene, resulting in increased transcription and accumulation of AA-NAT mRNA and then in increased expression and activity of the AA-NATenzyme. The activation of a1-adrenergic receptors induces increased concentrations of Ca2þ and diacylglycerol (DAG), leading to activation of protein kinase C (PKC) potentiating b1adrenergic receptor stimulation of adenylate cyclase (AC) through a postreceptor mechanism. Furthermore, there is posttranslational regulation of AA-NAT protein levels by means of cAMP-dependent inhibition of proteosomal proteolysis. This mechanism involves phosphorylation-dependent binding of AANAT to 14-3-3 proteins, shielding the enzyme from proteolysis. In rodents, cAMP stimulation causes AANAT mRNA to increase more than 150-fold at night, whereas in ungulates and primates, the night/day ratio is approximately 1.5. This difference may be explained by the fact that in ungulates and primates, AA-NAT protein levels are regulated primarily at the posttranslational level by controlled proteosomal proteolysis. AA-NAT is a member of a large superfamily of proteins referred to as the GNAT family and acts through catalysis of the transacetylation of serotonin to Nacetylserotonin with Ac-CoA as a donor. AA-NAT is a globular protein with a molecular weight of 23 kDa and consisting of eight stranded b-sheets containing AcCoA-binding sites. It is described in two different conformational states influencing its functional efficiency. In addition to NE-induced activation of the melatonin-generating system, several other neuroactive substances have been shown to influence melatonin synthesis. These include vasoactive intestinal peptide (VIP), pituitary adenylate cyclase-activating polypeptide (PACAP), and ACh acting on AA-NAT

Achondroplasia see Chondrodysplasias

mRNA expression and/or on second messengerinduced AA-NAT stimulation. It has been hypothesized that alterations in AANAT synthesis and/or activity can represent the pathophysiological basis not only of circadian rhythm disturbances but also of several pathologies such as migraine, depression, and insomnia.

See Also the Following Articles ACTH, a-MSH, and POMC, Evolution of Disease and Hormones

.

Alzheimer’s

Further Reading Cangemi, L., Adage, T., Morabito, A., and Portaleone, P. (1995). N-acetyltransferase mechanism for a-melanocyte stimulating hormone regulation in rat ageing. Neurosci. Lett. 201, 65–68. Davie, R. J., and Spencer, V. A. (1999). Control of histone modifications. J. Cell. Biochem. Suppl. 32/33, 141–148. Dobransky, T., and Rylett, R. J. (2003). Functional regulation of choline acetyltransferase by phosphorylation. Neurochem. Res. 28, 537–542. Dores, R. M., Stevenson, T. C., and Price, M. L. (1993). A view of the N-acetylation of a-melanocyte-stimulating hormone and b-endorphin from phylogenetic perspective. Ann. NY Acad. Sci 680, 161–174. Ganguly, S., Coon, S. L., and Klein, D. C. (2002). Control of melatonin synthesis in the mammalian pineal gland: The critical role of serotonin acetylation. Cell Tissue Res. 309, 127–137. Hasan, S., and Hottiger, M. O. (2002). Histone acetyl transferases: A role in DNA repair and DNA replication. J. Mol. Med. 80, 463–474. Kuo, M-H., and Allis, C. D. (1998). Roles of histone acetyltransferases and deacetylases in gene regulation. BioEssays 20, 615–626. O’Donohue, T. L., Handelmann, G. E., Chaconas, T., Miller, R. L., and Jacobowitz, D. M. (1981). Evidence that N-acetylation regulates the behavioral activity of a-MSH in the rat and human central nervous system. Peptides 2, 333–344. Prado, M. A. M., Reis, R. A. M., Prado, V. F., de Mello, M. C., Gomez, M. V., and de Mello, F. G. (2002). Regulation of acetylcholine synthesis and storage. Neurochem. Intl. 41, 291–299. Schomerus, C., Laedtke, E., Olcese, J., Weller, J. L., Klein, D. C., and Korf, H-W. (2002). Signal transduction and regulation of melatonin synthesis in bovine pinealocytes: Impact of adrenergic, peptidergic, and cholinergic stimuli. Cell Tissue Res. 309, 417–428.

Acromegaly, Clinical Features of Ariel L. Barkan University of Michigan, Ann Arbor, Michigan, United States

Glossary g9000

g9005

g9010

g9015

g9020

arthropathy Abnormality in the condition of a joint of the body; enlargement, swelling, degeneration, or other such disturbance in a joint. growth hormone (GH) A polypeptide hormone that is secreted mainly by the pituitary gland, although it is also produced by other cell types, such as lymphoid cells. Its actions are related mainly to growth (soft tissues, long bones, etc.) and to metabolism. It belongs to a family of hormones that includes prolactin and placental lactogens as well as other placental factors. Acromegaly results from GH hypersecretion. insulin-like growth factor (IGF) One of a class of hormones structurally related to insulin, but exhibiting proliferative and differentiative, rather than metabolic, effects. IGF-1 overproduction is associated with acromegaly. multiple endocrine neoplasia (MEN) syndromes A group of genetically distinct familial diseases in which two or more endocrine glands develop excess normal tissue (hyperplasia) and/or adenoma (tumor). sleep apnea A sleep disorder in which the subject has intermittent periods of a failure to automatically control respiration; these involuntary pauses in breathing may occur repeatedly during a given period of sleep.

F

rom the purely clinical perspective, acromegaly is the most spectacular endocrine disease. Humanity has always been fascinated with giants, and from the biblical Goliath to James Bond’s foe, Jaws, they populate the lore of virtually every era and culture.

INTRODUCTION Acromegaly is due to growth hormone (GH) hypersecretion and the resultant secondary overproduction of insulin-like growth factor-1 (IGF-1). In virtually all cases (>99%), the source of excessive GH is a benign pituitary tumor of purely somatotroph or mixed cellular origin. Rarely, somatotroph tumors arise in an ectopic pituitary, a remnant of the primitive

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

Rathke’s pouch, and are found in the posterior pharynx, sphenoid bone, or sphenoid sinus or even within the sella, separate from the normal pituitary gland. Ectopic production of GH-releasing hormone (GHRH) leads to pituitary somatotroph hyperplasia with subsequent adenoma formation. Carcinoids and islet cell tumors are the most frequent sources of ectopic GHRH. Hypothalamic/pituitary gangliocytomas or choristomas, or in one case the pituitary adenoma itself, were shown to be a source of excessive GHRH production. Ectopic GH secretion was documented in only two cases: one by a malignant islet cell tumor and another by a non-Hodgkin’s lymphoma. Acromegaly may be a part of well-defined MEN syndromes such as MEN-1 (parathyroid, pituitary, pancreas), McCune–Albright, and Carney syndromes. In some instances, the clinical syndrome of acromegaly may be overshadowed by other manifestations of a malignant or polyglandular disease. Clinical manifestations of acromegaly correlate better with the prevailing levels of IGF-1 than with GH, and the duration of GH/IGF-1 excess may play a major role. A clinical and biochemical syndrome of acromegaly may be transiently expressed during normal puberty or pregnancy. This is due to physiological overproduction of GH by the normal pituitary gland during sexual maturation or by the placental synthesis of the GH variant. In most cases, acromegaly is an insidious disease, and its early clinical manifestations usually go unnoticed by the patient, the patient’s family, and/or the patient’s family physician. Retrospective questionnaires and the inspection of old photographs usually set the clinical onset of disease at 5 to 10 years prior to the diagnosis. Clinical presentation of acromegaly consists of the mass effects of the tumor itself, the manifestations of the abnormal growth affecting virtually all organs and tissues, and the metabolic derangements effected by GH itself. Despite the seemingly straightforward pathophysiological mechanisms, the clinical picture of acromegaly is often protean, and the correct

13

14

Acromegaly, Clinical Features of

and timely diagnosis requires significant clinical acumen.

MASS EFFECTS OF THE TUMOR p0030

Most pituitary somatotropinomas (60–80%) are large at the time of diagnosis (macroadenoma, >10 mm in the largest diameter) and are often invasive. Visual field defects were present at diagnosis in 90% of patients in the past, but this figure decreased to
10 to 20% due to earlier recognition of the disease. Ophthalmoplegia ( III, IV, VI, and V1, V2 nerves) is rare and, if present, suggests recent rapid expansion of the tumor by a hemorrhage. Headache is present in 50 to 60% of patients and may be severe in about half of them. Headache may be present even in patients with relatively small tumors, where mass effect is unlikely to provide an explanation. Hypopituitarism (ACTH and TSH deficiency) in acromegaly occurs less frequently than in patients with nonfunctioning tumors of similar size. However, hypogonadism is frequent (
50%), likely as a consequence of coexisting hyperprolactinemia and inherent lactogenic effect of GH itself. Symptomatic pituitary hemorrhage occurs in less than 5% of patients, but asymptomatic events may occur in 30 to 40% of the cases.

ABNORMAL GROWTH p0035

Onset of pathological GH hypersecretion before puberty results in an augmented statural growth. Concomitant hypogonadism prevents epiphyseal closures, and gigantism ensues. Postpubertal onset of disease leads to disproportionate growth and dysmorphic features, that is, true acromegaly (“large extremities”). Eventually, even pituitary giants develop an acromegalic appearance.

Face p0040

A combination of bone and soft tissue overgrowth leads to a typical “acromegalic” face: large nose, thick lips, exaggerated nasolabial and frontal skin furrows, mandibular overgrowth and prognathism, teeth separation, and frontal bossing. These features are seen in 98 to 100% of patients (see Fig. 1).

Extremities Hands and feet become very “fleshy.” An increase in finger circumference and a widening of the hands and feet develop in 98 to 100% of patients. Patients

Figure 1 Appearance of a patient with florid acromegaly. Courtesy of Professor Stefan S. Fajans.

f0005

routinely recall repeat resizing of rings and changes in shoe size (mostly widening).

Skin and Appendages Skin is characteristically thickened because of excessive deposition of the glycosoaminglycans, hyaluronic acid, chondroitin sulfate, and dermatan sulfate in the papillary and upper reticular dermis. These compounds are very hydrophylic, causing the appearance of a nonpitting edema. Skin thickening at the vertex

p0050

Acromegaly, Clinical Features of

causes a peculiar appearance of cutis verticis gyrata (skin folds at the top of the head). Hair growth is increased, and women complain of hirsutism. As opposed to the androgen-related hirsutism, this is pronounced even on the forearms and forelegs. Many women with acromegaly have exceedingly thick scalp hair growth. Hair loss after successful therapy is often a cause of concern but is essentially a physiological return to normalcy. The functional capacity of sweat and sebaceous glands is increased, resulting in excessive perspiration, often with offensive odor, and in oily skin. Skin tags are frequently present, particularly on the neck. Whether their presence and number can be a marker for colonic neoplasia is uncertain.

Neuromuscular The muscle mass is increased, but this is primarily due to increased intracellular water, so that muscle strength is either normal or low. In fact, many patients have clinically obvious proximal myopathy. Muscle biopsy often shows hypertrophy of type I and/or atrophy of type II fibers. Compression neuropathies are common (30–50%); the median nerve is most often affected. It was believed that carpal tunnel syndrome was due to external compression by the components of the wrist compartment, but MRI data showed that the nerve itself is swollen. Occasionally, distal symmetric polyneuropathy may be present.

Oral Progathism and widening of the interdental spaces have already been mentioned. Importantly, the size of the tongue is usually enlarged and contributes to the obstruction of the pharynx, with the resultant sleep apnea and impaired mastication. The abnormal oral anatomy often results in speech disturbances. The diameter of the trachea is increased, and the vocal cords are thickened. Together with grossly enlarged sinus cavities, this results in a low and hollow voice pitch. Salivary glands are typically enlarged, and their size is a convenient measure of the GH effect on parenchymatous organs.

Articular GH receptors are present on all major cell types comprising the skeletal system: fibroblasts, chondrocytes, and osteoblasts. These cells readily produce IGF-1 and are targets for both endocrine and autocrine IGF-1

15 effects. Thus, arthropathy is a frequent (60–80%) symptom of acromegaly, affecting both axial and appendicular skeleton. The degree and severity of arthropathy best correlate with the duration of disease. Joint pain and low back pain may be experienced soon after the clinical onset of acromegaly but are often fully reversible with successful therapy. However, clinical duration of acromegaly in excess of 10 years is often associated with clinical and radiographical joint deformities that are only minimally affected by the GH-lowering therapy. Appendicular Arthropathy The knee is the joint most frequently affected, followed by the shoulder, hip, ankle, elbow, and small hand joints. Radiographic changes are usually seen even in clinically unaffected joints. Initially, GH excess causes cartilage hypertrophy and laxity of the ligaments. The combination of altered geometry of the joint and its instability leads to repeat trauma to the cartilage. The ensuing cartilage fissures are filled by regenerative fibrocartilage, with the subsequent calcification, formation of osteophytes, and exposure of the subchondral bone. Eventually, the articular cartilage becomes thinned and the joint space narrows. The end-stage acromegalic arthropathy looks essentially like degenerative osteoarthritis. Clinical and radiological reversibility of arthropathy can be seen after successful therapy in the joints exhibiting cartilage thickening, but late stages may require joint prosthesis for relief of pain and functional mobility. Axial Arthropathy Lumbar involvement is most common, followed by thoracic and cervical arthropathy. Overall, approximately 50% of patients complain of back pain and limitation of movements. Thickened intervertebral discs and lax paraspinal ligaments contribute to abnormal joint mobility. End-stage arthropathy is characterized by the narrowing of the intervertebral space. Ossification of the anterior aspect of the vertebral bodies with exuberant osteophyte formation often bridges the disc space, mimicking diffuse idiopathic skeletal hyperostosis (DISH). Vertebral deformities often lead to kyphosis that may become almost grotesque in some patients. Bone Metabolism Biochemical markers of bone remodeling are increased in patients with acromegaly, but histomorphometric data are conflicting; cortical bone shows predominance of bone formation over resorption, whereas trabecular bone has the opposite pattern. Because most patients

16 with acromegaly have concomitant hypogonadism, these data should be interpreted with caution. Bone density studies are equally controversial, but overall it appears that cortical bone mineral density may be normal or even increased, whereas trabecular bone mineral density is normal or decreased. The “normalcy” of the latter is questionable because of the interference by osteophytes. In any case, patients with acromegaly do not have altered fracture rates, and the assessment of bone mineral density in acromegaly may be of academic interest only.

Cardiovascular p0100

Increased mortality of untreated or poorly treated acromegaly is almost completely attributable to cardiovascular disease. Hypertension is seen in 20% of patients and does not appear to be reversible by GH/ IGF-1-lowering therapy. Concentric left ventricular hypertrophy is seen in approximately two-thirds of patients, and its occurrence is best related to the duration of disease. Hypertrophy frequently occurs even in normotensive patients. Early stages of acromegaly, typical of young patients with short disease duration, are characterized by tachycardia and increased systolic output (hyperkinetic syndrome). Progressively, cardiac hypertrophy and diastolic dysfunction ensue, and end-stage disease is characterized by impaired systolic function and heart failure. The development of hemodynamic abnormalities is augmented by valvular involvement in approximately 20% of patients. Normalization of GH and IGF-1 can reduce the degree of left ventricular hypertrophy within 2 to 4 weeks. Rhythm abnormalities, occurring in 40% of patients, are more ominous. Ectopic beats, paroxysmal atrial fibrillation, paroxysmal supraventricular tachycardia, sick sinus syndrome, ventricular tachycardia, and bundle branch blocks all are seen with increased frequencies in patients with acromegaly and are exacerbated by physical exercise. Disturbingly, biochemical control of acromegaly might not improve conduction abnormalities. Late ventricular potentials (low-amplitude, high-frequency waves in the terminal phase of QRS complexes) are strong predictors of future arrhythmic events and are seen in 50% of patients with active acromegaly.

Sleep Apnea Sleep apnea occurs in 75% of patients. In most, it is of an obstructive nature due to soft tissue hypertrophy of the pharynx. Interestingly, in approximately one-third

Acromegaly, Clinical Features of

of patients, there is a central component of sleep apnea with decreased ventilatory drive. The pathogenesis of central sleep apnea in acromegaly is unknown, but both subtypes may be ameliorated by GH-lowering therapy. Sleep apnea is manifested as snoring, daytime sleepiness, fatigue, and headache. It is a strong predictor of future cardiovascular events, hypertension, or stroke. Many patients are unaware of their snoring; family members provide markedly more reliable information.

Renal Kidney size and glomerular filtration rate are characteristically increased. Hypercalciuria may lead to kidney stone formation (in approximately 10% of patients).

Gastrointestinal Liver and spleen sizes are normal. Hepatomegaly in patients with acromegaly always should be assumed to result from another disease process and should be vigorously investigated. Similarly, the incidence of cholelithiasis is not increased relative to that in the general population. Patients with acromegaly often suffer from constipation due to long and tortuous colon. The incidence of colonic polyps appears to be increased.

METABOLIC Impaired glucose tolerance is present in 40% of patients, and frank diabetes (DM-2) is present in 30% of patients. These often improve or vanish altogether after successful GH-lowering therapy. Other metabolic abnormalities include hypertriglyceridemia, hypercalciuria, and hyperphosphatemia. Hypercalcemia is not a feature of acromegaly per se, and its presence should suggest another pathological process.

NEUROPSYCHIATRIC Similar to any chronic disease associated with physical discomfort and lifelong therapy, acromegaly is associated with decreased quality of life. However, there are no specific neuropsychiatric features attributable to the disease.

CANCER Because patients with acromegaly have increased incidence of colonic polyps, some investigators assumed that the incidence of colonic carcinoma will also be increased. This was supported by the data from several small studies. Large-scale data, however, failed to

17

Acromegaly, Clinical Features of

document increased incidence of colon cancer in acromegaly. Similarly, the incidence of breast and prostate cancers was not increased. This, of course, does not apply to patients with genetic forms of acromegaly, such as MEN-1, in which there is a predictable component of coexisting neoplastic diseases.

See Also the Following Articles Acromegaly, Diagnosis of . Acromegaly, Therapy for Growth Hormone (GH) . Pituitary Tumors, Clonality Pituitary Tumors, Molecular Pathogenesis

. .

Further Reading Barkan, A. (1997). Acromegalic arthropathy and sleep apnea. J. Endocrinol. 155(Suppl. 1), S41–S44. Colao, A. (2001). Are patients with acromegaly at high risk for dysrhythmias? Clin. Endocrinol. 55, 305–306. Colao, A., Marzullo, P., Di Somma, C., and Lombardi, G. (2001). Growth hormone and the heart. Clin. Endocrinol. 54, 137–154. Ezzat, S., Forster, M. J., Berchtold, P., Redelmeier, D. A., Boerlin, V., and Harris, A. G. (1994). Acromegaly: Clinical and biochemical features in 500 patients. Medicine 73, 233–240. Melmed, S. (2001). Acromegaly and cancer: Not a problem? J. Clin. Endocrinol. Metab. 86, 2929–2934.

Acromegaly, Diagnosis of Klaus von Werder Humboldt University, Berlin, Germany

Glossary acromegaly Clinical consequence of growth hormone hypersecretion. growth hormone (GH) Hormone of the anterior pituitary lobe that is essential for normal growth in children. GH is also an important metabolic hormone in adults. insulin-like growth factor-1 (IGF-1) Peptide hormone produced in the liver and other tissues following activation of the GH-receptor.

E

ndocrinological investigation using hormone measurements and function tests to document active acromegaly is indicated whenever symptoms and clinical stigmata suggestive for the disease are present. Thus, the diagnosis of acromegaly is made by measuring elevated, often glucose nonsuppressible, growth hormone levels and elevated insulin-like growth factor-1 levels.

MEASUREMENT OF GROWTH HORMONE LEVELS p0010

Growth hormone (GH) levels are measured by immunoassay or immunoradiomatric assay. Most centers use polyclonal antibodies in their assay systems which measure higher GH levels compared to assay systems that use oligo- or monoclonal antibodies. Cutoff values that differentiate between normal and abnormally elevated GH levels are based on polyclonal assay systems according to the consensus conference on diagnostic aspects of acromegaly held in Cortina d’Ampezzo in 1999. Using these systems, fasting morning or random GH levels are usually clearly elevated in patients with acromegaly of at least more than 0.4 ng/ml (Fig. 1). Using the more sensitive and better standardized assay system with oligo- or monoclonal antibodies, these cutoff values will have to be redefined in the future. Since GH is episodically secreted in normal subjects, GH levels considerably

18

higher than the cutoff levels are not indicative for acromegaly unless the insulin-like growth factor-1 (IGF-1) level is also elevated. Because of the spontaneous fluctuation of GH levels in nonacromegalic subjects, many investigators recommend an oral glucose tolerance test as a suppression test for GH secretion. Acromegaly is documented in patients in whom GH levels are higher than 1.0 ng/ml and who have elevated IGF-1 levels simultaneously. Blood sugar measurements during the oral glucose tolerance test demonstrate the different degrees of carbohydrate intolerance that are observed in active acromegalics. Measurement of insulin levels demonstrating insulin antagonism in active acromegalics is not required for making the diagnosis. Furthermore, the cumbersome measurement of GH 24-h profiles is not necessary, nor is it helpful to document the inappropriate stimulation of GH after thyrotropin releasing hormone (TRH) or gonadotropin releasing hormone (GnRH) injection that occurs in 70 and 40% of all active acromegalics, respectively. Measurement of urinary GH is also not useful for documentation of a GH hypersecretory state.

MEASUREMENT OF IGF-1 LEVELS Serum IGF-1 levels are invariably elevated in active acromegaly, showing a certain degree of correlation with GH levels. Since IGF-1 levels are also elevated during puberty and pregnancy, they are specific for active acromegaly in the nonpregnant adult. Since IGFBP-3 is concomitantly produced and secreted with IGF-1, IGFBP-3 levels are also elevated in acromegalic patients. Measurement of the binding protein, however, does not add to the accuracy of the diagnosis. This pertains also to the acid labile subunit (ALS), which is also elevated as part of the ternary complex consisting of IGF-1, IGFBP-3, and ALS. In addition, measurement of free IGF-1, which is only possible in some laboratories, is not required for making the diagnosis of active acromegaly.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

p0015

19

Acromegaly, Diagnosis of

f0005

Figure 1 Algorithm for the diagnostic workup of a suspected acromegalic patient. According to the consensus statement (Giustina et al., 2000.)

MEASUREMENT OF PROLACTIN AND OTHER ANTERIOR PITUITARY HORMONES Since some GH-producing tumors also produce and secrete prolactin (PRL), basal PRL should be measured in all acromegalics. PRL levels higher than 200 mg/liter are usually indicative of a somatomammotrophic adenoma. The identification of a somatomammotrophic tumor may be important with regard to medical therapy since the PRL-cosecreting adenomas respond particularly well to dopamine agonists. Slightly elevated PRL levels can result from impingement of the somatotroph tumor on the pituitary stalk leading to unrestrained PRL release from the nontumorous pituitary. Tests to evaluate the other pituitary functions—the gonadal, adrenal, and thyroid axis—are also indicated.

DIFFERENTIAL DIAGNOSIS OF ACROMEGALY p0035

In more than 99% of all patients with acromegaly, a monoclonal GH-producing tumor is the cause of the

disease (Fig. 2). The imaging method of choice is magnetic resonance imaging (MRI), which in most cases allows visualization of a macroadenoma and, less frequently, a microadenoma. This situation has been referred to by Losa and von Werder as classical acromegaly. In contrast to classical acromegaly, there are rare cases who have extra pituitary lesions causing GH hypersecretion, leading to the clinical picture of acromegaly. The ectopic growth hormone-releasing hormone (GHRH) syndrome is most common (Fig. 2). In this syndrome, GHRH is secreted from benign or malignant, often neuroendocrine, tumors, leading to somatotroph hyperplasia with consecutive GH hypersecretion. GH and IGF-1 levels are not different from those of classical acromegaly. On MRI, no pituitary adenoma can be demonstrated, although sometimes these patients have suprasellarly extending lesions due to somatotroph hyperplasia. The diagnosis of ectopic GHRH syndrome cannot be made using function tests, although many of these patients do not respond to exogenous GHRH administration but show hypersecretion of GH after TRH. The diagnosis of ectopic GHRH secretion is made by the measurement of elevated peripheral GHRH levels (Fig. 1), which are in the nanogram range, in contrast

20

Acromegaly, Diagnosis of

In contrast, patients with hypoptholamic GHRHsecreting tumors, so-called eutopic GHRH syndrome, do not have elevated GHRH levels in the periphery (Fig. 2). This diagnosis is usually made after transsphenoidal operative therapy. In these cases, no tumors which are separated from the anterior pituitary lobe, but in which somatotroph hyperplasia is found, are sometimes intermingled with neuroendocrine tissue expressing GHRH. Occasionally, GHRH-producing cells are found in the pituitary specimen interspersed within the somatotroph cells. The ectopic GH syndrome is extremely rare. Only two cases, one with pancreatic cancer and the other with non-Hodgkin lymphoma (both shown to produce and secrete GH), have been reported. The diagnosis can be suspected when GH levels are elevated, there is no evidence of a pituitary lesion (either adenoma or somatotroph hyperplasia), and GHRH levels are normal. When there is a family history of acromegaly, one has to consider familial acromegaly or multiple endocrine neoplasia type 1. In the latter case, one needs to exclude primary hyperparathyroidism and islet cell tumors. Occasionally, acromegaly occurs as part of the McCune–Albright syndrome, with very severe bone deformations caused by acromegaly accompanied by osteofibrotic bone disease, which is the hallmark of this rare syndrome.

See Also the Following Articles Acromegaly, Clinical Features of . Acromegaly, Therapy for . Growth Hormone (GH) . Insulin-like Growth Factors . McCune-Albright Syndrome . Prolactin (PRL)

Further Reading f0010

Figure 2 Different causes of acromegaly. In more than 99% of all cases, a monoclonal pituitary tumor can be demonstrated (classical acromegaly). The ectopic GHRH syndrome is the second most common cause of GH hypersecretion. Eutopic GHRH secretion from hypothalamic lesions cannot be differentiated from classical acromegaly before operative therapy. The extremely rare ectopic GH syndrome has been demonstrated unequivocally in only two cases. SRIF, somatotropic hormone release inhibiting factor.

to classical acromegaly, in which GHRH levels do not exceed 100 pg/ml. Often, the ectopic GHRH source can be localized by somatostatin receptor scintigraphy since these tumors seem to express somatostatin receptor types 2 and 5.

Giustina, A., Casanueva, F. F., et al. (2000). Consensus: Criteria for cure of acromegaly: A consensus statement. J. Clin. Endocrinol. Metab. 85, 526–529. Harris, A. G. (1996). “Acromegaly and Its Management.” Lippincott–Raven, Philadelphia. Lamberts, S. W. J. (1998). Acromegaly. In “Clinical Endocrinology” (A. Grossman, ed.), 2nd ed., pp. 170–183. Blackwell, Oxford, UK. Landolt, A. M., Vance, M. L., and Reilly, P. S. (eds.) (1996). “Pituitary Adenomas.” Churchill Livingstone, New York. Losa, M., and Werder, K. von (1997). Pathophysiology and clinical aspects of the ectopic GH-releasing hormone syndrome. Review. Clin. Endocrinol. 47, 123–135. Melmed, S. (ed.) (2002). “The Pituitary.” Blackwell, Oxford, MA. Wass, J. (ed.) (2001). “Handbook of Acromegaly.” Bioscientifica, Bristol, UK.

p0045

Acromegaly, Therapy for Anat Ben-Shlomo and Shlomo Melmed University of California, Los Angeles School of Medicine, Cedars–Sinai Research Institute, Los Angeles, California, United States

Glossary arthralgia Pain in a joint. bradycardia A slower than normal contraction of the heart muscle causing a pulse of less than 60 beats per minute. dopamine A hormone-like substance that is an important neurotransmitter in both the central and peripheral nervous systems; occurs as an intermediate in epinephrine and norepinephrine biosynthesis. gallstone An amalgamous concretion, commonly of cholesterol crystals, bilirubin, and protein, that is formed in the gallbladder or bile duct.

A

cromegaly, a somatic growth and proportion disorder, is a rare and insidious disease caused by growth hormone (GH)secreting pituitary tumors or (rarely) extrapituitary disorders. Elevated levels of GH and insulin-like growth factor-1 (IGF-1) are the hallmarks of this syndrome. Clinical manifestations include skeletal and soft tissue growth and deformations as well as cardiac, respiratory, neuromuscular, endocrine, and metabolic complications. Early diagnosis and aggressive control of the disease significantly attenuate or even abolish the increased morbidity and mortality from the disease. Although transsphenoidal surgery is considered the primary treatment of choice, it is apparent that complete macroadenoma resection, especially if invasive, is difficult even for skilled surgeons, hence their low rate of biochemical control.

INTRODUCTION When using strict criteria for disease control to define a cure for acromegaly (e.g., random serum growth hormone [GH] levels < 2.5 ng/ml, GH levels after oral glucose tolerance test [OGTT] < 1 ng/ml, genderand age-matched insulin-like growth factor-1 [IGF-1] levels in the normal range), approximately 70% of patients were in remission during the short period after surgical intervention alone and 40% were in remission up to 16 years postsurgery. Radiotherapy,

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

either conventional external deep X-ray or heavyparticle (proton-beam) irradiation, is reserved mainly as adjuvant therapy for patients who have failed surgical or medical treatment. Therefore, pharmacological treatment has assumed more importance in managing patients with acromegaly. This article briefly discusses the pharmacotherapy available for acromegaly, including somatostatin analogues, GH receptor antagonists, and dopamine agonists.

SOMATOSTATIN AND ITS SYNTHETIC ANALOGUES Somatostatin, a peptide that suppresses GH secretion, was discovered in the hypothalamus some 30 years ago and was shown to reduce GH serum levels. The two major somatostatins that are enzymatically cleaved from the large preprosomatostatin precursor molecule are somatostatin-18 and somatostatin-28. These are produced in multiple tissues and induce their variable actions through a family of five major receptors, SSTR1 to SSTR5. Because GH cell pituitary adenomas express SSTRs (mainly SSTR5 and SSTR2 subtypes) that regulate GH secretion, the pharmacological use of somatostatin binding to these receptors has been employed for treating these tumors. Endogenous somatostatin-14 has a high affinity to all SSTRs, produced locally and rapidly degraded to prevent systemic effects. Exogenous administration of the peptide is not applicable for prolonged treatment of acromegaly due to an extremely short half-life and multiple side effects. The synthetic compounds octreotide (SMS201-995) and lanreotide (BIM23014) offer a greater metabolic stability and subtype specificity for the treatment of acromegaly (e.g., high affinity to SSTR2 and SSTR5 and moderate affinity to SSTR3 vs somatostatin-14). Octreotide is a short-acting molecule, with a halflife of 2 h, that binds with high affinity to SSTR2 and, to a lesser extent, to SSTR5. This analogue was the first to be used clinically and has been proved safe and effective for medical therapy of acromegaly.

21

22

Acromegaly, Therapy for

Reduction of GH levels is achieved with octreotide doses ranging from 50 to 500 mg given subcutaneously every 8 h (in most patients, 100 mg trice daily). The drug will achieve its maximal suppressive effect on GH levels 2 to 6 h after injection. With continuous treatment, the extent by which GH levels rise between injections is reduced. After 6 months of octreotide treatment, integrated mean GH levels decreased to less than 5 ng/ml in 53% of 115 patients with acromegaly, and IGF-1 levels were normalized in 68% of the patients. After 30 months of treatment, GH levels were reduced to less than 5 ng/ml in 65% and to less than 2 ng/ml in 40% of 97 patients. In a summary of 11 studies, serum IGF-1 levels were normalized in 53% (range 42–80%) of 417 patients with acromegaly treated with 100 to 1500 mg octreotide daily for a period of 3 to 57 months. Improved clinical symptoms, including headache, perspiration, fatigue, arthralgia, and cystic acne, were observed in up to 78% of 115 patients treated with 750 mg octreotide for 6 months and in up to 95% of 103 patients in another study. Adverse drug effects include diarrhea (60%), abdominal discomfort (45%), loose stool (32%), nausea, headache, dizziness, flatulence, and constipation. These symptoms usually subside after 1 to 3 weeks of treatment. Clinically nonsignificant bradycardia develops in approximately 25% of treated patients. Asymptomatic cholesterol gallstones form in approximately 25% of treated patients, usually during the first 2 years of therapy. Octreotide LAR is a depot intramuscular preparation (D,L-lactide-coglycolide-glucose) that provides sustained slow release of the drug peaking at 28 days.

Injections of 20 to 30 mg octreotide LAR at 28-day intervals suppresses GH levels and generally achieves steady-state drug levels after two or three injections. As demonstrated in Table I, GH levels are reduced to less than 2.5 ng/ml in 70 to 90% of patients, and IGF-1 levels are normalized in approximately 65% or even as many as 88% of patients treated with 20 to 40 mg octreotide LAR monthly injections for 12 to 30 months. Octreotide responders respond well to the LAR form. Moreover, reduction of GH levels to less than 5 ng/ml was observed in more patients treated with octreotide LAR (94%) than in those having received subcutaneous octreotide (82%). IGF-1 levels were normalized in 65% of patients treated with octreotide LAR, as compared with 50% of those treated with subcutaneous octreotide. Marked improvement of carpal tunnel syndrome, paresthesias, perspiration, arthralgia, headache, and fatigue was reported, as was approximately 20% tumor shrinkage. Adverse effects are mostly mild, last for 1 to 2 days, and include diarrhea (45%), abdominal pain (32%), and flatulence (35%). These effects are transient, and their frequency decreases with treatment extension to 11, 3, and 8% of patients, respectively. Gallbladder abnormalities that were apparent in 26% of patients treated for 30 months included asymptomatic cholelithiasis, sediment, sludge, and biliary or gallbladder dilatation. More recently, a case of probable partial tachyphylaxis to both depot preparations of somatostatin was reported along with demonstration of antisomatostatin analogue antibodies. Slow-release lanreotide is a cyclic octapeptide somatostatin analogue administered intramuscularly

Table I Summary of Prospective Open-Label Studies on the Effects of Long-Term Treatment (3–89 months) of Patients with Acromegaly with Octreotide LAR, Lanreotide SR, and Pegvisomant on Serum GH and/or IGF-1 Levels Treatment outcome Number of patients Lanreotide SR

GH < 5 ng/ml

a

298

GH < 2.5 ng/ml

225 (76)

b

197 (53)

369

c

Octreotide LAR

538 a 360 b

439

c

Pegvisomant

Normalized IGF-1

297 (55) 337 (94) 260 (59)

449

308 (69)

208

193 (93)

Note. Percentages are in parentheses. a Total number of patients who achieved GH levels of < 5 ng/ml. b Total number of patients who achieved GH levels of < 2.5 ng/ml. c Total number of patients who achieved normal IGF-1 levels.

23

Acromegaly, Therapy for

at a dose of 30 mg every 10 or 14 days. To summarize 10 open-labeled prospective studies, 71% of patients treated with lanreotide for at least 1 year exhibited GH levels less than 5 ng/ml (43% < 2.5 ng/ml) and 58% exhibited normal IGF-1 levels. Of 23 newly diagnosed, nonoperated or irradiated patients treated with lanreotide SR, 5 (22%) had significant tumor shrinkage (> 20%) after 6 months of therapy. Clinical improvement of perspiration, headache, and tissue swelling was reported in 4 of 13, 4 of 6, and 5 of 5 patients, respectively, after 1 month of treatment. Arthralgia improved in 3 of 5 patients after 9 months of therapy. Patients treated with long-term subcutaneous octreotide and then subsequently with lanreotide reported similar improvement with both therapies, including improved headache (81% of 22 patients), paresthesia (73%), and soft tissue swelling (61%). Lanreotide Autogel, a new delivery formulation of lanreotide administered by deep subcutaneous injection at a dose of 60 to 120 mg every 28 days, was introduced recently. Studied in 107 patients with acromegaly for 3 months, lanreotide Autogel proved to be at least as efficacious and well tolerated as 30 mg lanreotide. Improvement of clinical symptoms, including headache (30%), night sweats (20%), and joint pain (40%), was reported after 3 months of treatment with both preparations. Similar clinical improvement was reported with lanreotide SR during long-term treatment. Side effects include mainly diarrhea (38%), abdominal pain (22%), gallbladder lithiasis or sludge (27%), and sludge (11%).

GROWTH HORMONE RECEPTOR ANTAGONIST Pegvisomant (B2036-PEG) is a genetically engineered analogue of human GH. This mutated GH not only prevents the homodimerization of the two GH receptors, thereby preventing signaling, but it also has a higher affinity to growth hormone-binding protein (GHBP), thereby prolonging its circulating half-life. This GH receptor antagonist is not yet approved for clinical use in the United States. Subcutaneous daily injection of up to 40 mg normalized serum IGF-1 levels in 89% of 80 patients treated for 12 weeks and in 97% of 152 patients treated for 12 to 18 months. The biochemical improvement was accompanied by a clinical alleviation of soft tissue swelling, excessive perspiration, and fatigue. Serum GH levels were increased nearly twofold in the treated patients, and serum anti-GH antibodies were detected in 8% of the patients after 12 weeks of therapy and in 17%

after longer treatment. Adverse effects included headache (26%); infection (mainly upper respiratory tract infection) (33%); injection site reaction (11%); pain in scalp, neck, shoulders, arms, and/or legs (23%); diarrhea (13%); asthenia (13%); arthralgia (12%); sinusitis (10%); and hypercholesterolemia (14%). In addition, one patient exhibited a significant increase in transaminases levels. No tumor shrinkage was observed; moreover, two patients had an increase in tumor size under treatment (1.6- and 1.8-fold increases). Despite being the most effective of all drugs at reducing IGF-1, and despite the promising results when used as adjuvant therapy, there is insufficient data supporting the use of pegvisomant as primary pharmacotherapy for acromegaly and safety for long-term treatment.

DOPAMINE AGONISTS Dopamine agonists decrease GH levels in some patients with acromegaly in contrast to those in normal individuals. Bromocriptine, long-acting bromocriptine LAR, pergolide, and carbegoline are ergot derivative dopamine agonists, whereas quinagolide is a nonergot derivative. Only bromocriptine and quinagolide are approved for clinical use in the United States. In contrast to their dramatic effect on prolactinomas, their effect on GH-secreting adenomas is limited and largely less effective than somatostatin analogues. GH levels decreased in 20% and IGF-1 levels decreased in 5 to 10% of patients treated with high doses of bromocriptine (up to 60 mg daily) and in 34 and 43% of those receiving carbegoline (up to 7 mg weekly) and quinagolide (up to 0.6 mg daily), respectively. Because of high drug doses required to achieve hormone reductions, the incidence of adverse effects is high and often unacceptable to patients. Side effects include gastrointestinal discomfort, including transient nausea and vomiting, dizziness (due to postural hypotension), headache, nasal congestion, and mood disorders. Addition of dopamine agonists, especially longer acting ones, may improve therapeutic efficacy of other modes of therapy for the treatment of acromegaly.

CONCLUSION The main goal of acromegaly therapy is tight control of GH and IGF-1 levels, hence reducing GH serum levels to less than 1 ng/ml after OGTT, normalizing age- and gender-matched IGF-1 levels, and reducing tumor mass as much as possible. Treatment of the disease’s complications is also of great importance.

p0045

24 The preferred primary treatment is transsphenoidal surgery by an experienced surgeon. Patients who were not cured by surgery, or who cannot or do not wish to be operated on, may be controlled by pharmacotherapy. First choice for medical treatment is the long-acting somatostatin analogues octreotide LAR and lanreotide SR. Other somatostatin analogues, such as SOM203, are at different stages of development and are not yet in clinical use. Dopamine agonists can be added, especially if the patient has coexisting hyperprolactinemia. Until side effects and efficiency of GH receptor antagonist are better established, this treatment is reserved for patients who cannot be cured by surgery and/or somatostatin and dopamine agonists.

See Also the Following Articles Acromegaly, Clinical Features of . Acromegaly, Diagnosis of . Growth Hormone (GH) . Insulin-like Growth Factors . Pituitary Tumors, Clonality . Pituitary Tumors, Molecular Pathogenesis

Further Reading Abs, R., Verhelst, J., Maiter, D., van Acker, K., Nobels, F., Coolens, J. L., Mahler, C., and Beckers, A. (1998). Carbegoline in the treatment of acromegaly: A study in 64 patients. J. Clin. Endocrinol. Metab. 83, 374–378. Ambrosio, M. R., Franceschetti, P., Bondanelli, M., Doga, M., Maffei, P., Baldelli, R., Tamburrano, G., Sicolo, N., Giustina, A., and degli Uberti, E. C. (2002). Efficacy and safety of the new 60-mg formulation of the long-acting somatostatin analog lanreotide in the treatment of acromegaly. Metabolism 51, 387–393. Attanasio, R., Barausse, M., and Cozzi, R. (2001). GH/IGF-1 normalization and tumor shrinkage during long-term treatment of acromegaly by lanreotide. J. Endocrinol. Invest. 24, 209–216. Ayuk, J., Stewart, S. E., Stewart, P. M., and Sheppard, M. C. (2002). Long-term safety and efficacy of depot long-acting somatostatin analogs for the treatment of acromegaly. J. Clin. Endocrinol. Metab. 87, 4142–4146. Baldelli, R., Colao, A., Razzore, P., Jaffrain-Rea, M. L., Marzullo, P., Ciccarelli, E., Ferretti, E., Ferone, D., Gaia, D., Camanni, F., Lombardi, G., and Tamburrano, G. (2000). Two-year follow-up

Acromegaly, Therapy for

of acromegalic patients treated with slow release lanreotide (30 mg). J. Clin. Endocrinol. Metab. 85, 4099–4103. Caron, P., Beckers, A., Cullen, D. R., Goth, M. I., Gutt, B., Laurberg, P., Pico, A. M., Valimaki, M., and Zgliczynski, W. (2002). Efficacy of the new long-acting formulation of lanreotide (lanreotide Autogel) in the management of acromegaly. J. Clin. Endocrinol. Metab. 87, 99–104. Colao, A., Ferone, D., Marzullo, P., et al. (2001). Long-term effects of depot long-acting somatostatin analog octreotide on hormone levels and tumor mass in acromegaly. J. Clin. Endocrinol. Metab. 86, 2779–2786. Herman-Bonert, V. S., Zib, K., Scarlett, J. A., and Melmed, S. (2000). Growth hormone receptor antagonist therapy in acromegalic patients resistant to somatostatin analogs. J. Clin. Endocrinol. Metab. 85, 2958–2961. Ho, K. K. (2001). Place of pegvisomant in acromegaly. Lancet 358, 1743–1744. Kendall-Taylor, P., Miller, M., Gebbie, J., Turner, S., and Al-Maskari, M. (2000). Long-acting octreotide LAR compared with lanreotide SR in the treatment of acromegaly. Pituitary 3, 61–65. Melmed, S. (1998). Tight control of growth hormone: An attainable outcome for acromegaly treatment. J. Clin. Endocrinol. Metab. 83, 3409–3410. Melmed, S., Casanueva, F. F., Cavagnini, F., Chanson, P., Frohman, L., Grossman, A., Ho, K., Kleinberg, D., Lamberts, S., Laws, E., Lombardi, G., Vance, M. L., von Werder, K., Wass, J., and Giustina, A. (2002). Consensus: Guidelines for acromegaly treatment. J. Clin. Endocrinol. Metab. 87, 4054–4058. Pradhananga, S., Wilkinson, I., and Ross, R. J. M. (2002). Receptor antagonists: Pegvisomant—Structure and function. J. Mol. Endocrinol. 29, 11–14. Shimon, I., Taylor, J. E., Dong, J. Z., Bitonte, R. A., Kim, S., Morgan, B., Coy, D. H., Culler, M. D., and Melmed, S. (1997). Somatostatin receptor subtypes specificity in human fetal pituitary cultures: Differential role of SSTR2 and SSTR5 for growth hormone, thyroid-stimulating hormone, and prolactin regulation. J. Clin. Invest. 99, 789–798. Shimon, I., Yan, X., Taylor, J. E., Weiss, M. H., Culler, M. D., and Melmed, S. (1997). Somatostatin receptor (SSTR) subtypeselective analogues differentially suppress in vitro growth hormone and prolactin in human pituitary adenomas: Novel potential therapy for functional pituitary tumors. J. Clin. Invest. 100, 2386–2392. Trainer, P. J., Drake, W. M., Katznelson, M. B. L., Freda, P. U., Herman-Bonert, V., van-der Lely, A. J., Dimaraki, E. V., Stewart, P. M., Friend, K. E., Vance, M. L., Besser, G. M., and Scarlett, J. A. (2000). Treatment of acromegaly with growth hormone-receptor antagonist pegvisomant. N. Engl. J. Med. 342, 1171–1177.

ACTH (Adrenocorticotropic Hormone) George Mastorakos University of Athens, Athens, Greece

Evangelia Zapanti Alexandra Hospital, Athens, Greece

Glossary g0005

adrenocorticotropic hormone (ACTH) A 39amino-acid peptide hormone (MW 45,000) that is part of the proopiomelanocortin precursor molecule. It controls the function of the adrenal cortex. proopiomelanocortin (POMC) A precursor molecule (MW 28,500) that, in the anterior lobe of the pituitary, is processed to adrenocorticotropic hormone and b-lipotropin and further processed to b-endorphin, an endogenous opioid peptide of 31 amino acids.

A

drenocorticotropin is the hypophyseal hormone that controls the function of the adrenal cortex.

BIOCHEMISTRY p0010

Adrenocorticotropic hormone (ACTH) is a 39-aminoacid peptide hormone (MW 45,000) that is part of the proopiomelanocortin (POMC) precursor molecule (MW 28,500). The PMOC gene is located on chromosome 2. Proopiomelanocortin is expressed in the brain, skin, and immune system and in the anterior and intermediate lobes of the pituitary gland. In the anterior lobe, POMC is processed to ACTH and b-lipotropin (b-LPH), which is further processed to b-endorphin, an endogenous opioid peptide of 31 amino acids (Fig. 1). In the intermediate lobe, ACTH is processed to a-melanocyte-stimulating hormone (a-MSH) (ACTH 1-13) and corticotropin-like intermediate lobe peptide (CLIP) (ACTH 18-39). In species with developed intermediate lobes (rat and sheep), these fragments are secreted, whereas in humans they are normally found only during fetal life. The NH2terminal amino acids of ACTH are identical in all species thus far studied, but there are species differences in the COOH terminus of the molecule.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

Interestingly, only the first 18 amino acids beginning with the NH2-terminal amino acids of ACTH are required for biological activity. In the brain, POMC is expressed in the arcuate nucleus of the hypothalamus and the nucleus tractus solitarius of the caudal medulla. In the brain, POMC generates a range of bioactive peptides, including ACTH, b-endorphin, and a-, b-, and g-MSH. b-Lipotropin, a fragment with 91 amino acids, which contains b-MSH (41-58), g-LPH (1-58), and b-endorphin, is secreted in equimolar quantities with ACTH. The endogenous opioid peptides a-, b-, and g-endorphin are derived from POMC and are composed of amino acids 61–76, 61–91, and 61–77 of b-LPH, respectively. The biological effects of POMC-derived peptides are diverse and are largely mediated through melanocortin (MC) receptors (R), five of which have been described. MC1R, MC2R, and MC5R have established roles in the pigmentation in the skin, adrenal steroidogenesis, and thermoregulation, respectively. Proopiomelanocortin-derived peptides are also found in the placenta, pancreas, testes, and gastric antrum.

PATHOLOGY AND EMBRYOLOGY The synthesis and processing of POMC to ACTH and to the other POMC derivatives are executed within the corticotroph cells of the anterior pituitary lobe. This cell type was originally considered chromophobic and later was shown by light microscopic studies to be basophilic (Fig. 2). Corticotrophs represent 15–20% of adenohypophyseal cells. The embryological origin of corticotrophs is the intermediate lobe, but groups of cells migrate during development into regions of the anterior and posterior lobes. Ultrastructural studies have shown different morphologies in these two groups of cells, reflecting differences in the processing enzymes responsible for the production of the above-mentioned hormones. The cells in the anterior lobe appear irregularly shaped with

25

26

ACTH (Adrenocorticotropic Hormone)

POMC

N-terminal fragment

ACTH

β −Lipoprotein β −MSH

γ−MSH

ACTH

α −MSH

CLIP

β −Lipoprotein

γ −Lipoprotein

β −Endorphin

Figure 1 Enzymatic process of POMC in the pituitary.

sparse secretory granules and poor staining, whereas those of the intermediate and posterior lobes exhibit dense granulation. In states of glucocorticoid excess (endogenous or exogenous), corticotrophs undergo degranulation and microtubular hyalinization, known as Crooke’s hyaline degeneration. In adrenal insufficiency, the corticotroph cells in the anterior lobe increase in number and the densely granulated cells of the intermediate and posterior lobes decrease in number.

MEASUREMENT OF ACTH The radioimmunoassay for the measurement of plasma ACTH is a highly sensitive and reliable assay for clinical use. Assays using other methodologies are also widely employed (i.e., enzyme-linked immunosorbent assay, chemiluminescence). The basal morning concentration of ACTH ranges from 9 to 80 pg/ml. The episodic secretion of ACTH causes fluctuations in plasma ACTH and cortisol levels. ACTH secretion has a distinct diurnal rhythm, with peak levels in the early morning and the lowest levels at approximately midnight. The interpretation of ACTH values requires a simultaneous cortisol determination. Provided that adrenocortical function is intact, plasma cortisol measurements, from a practical viewpoint, are a reliable index of ACTH secretion. There are other assays available for measuring ACTH (bioassays, radioreceptor assays, cytochemical assays), but measurement of ACTH by these techniques is performed only for research purposes, because of complexity and cost. The half-life of ACTH depends on the assay used for its measurement. Bioactive ACTH disappears from the circulation more rapidly (half-life of 3–9 min) than does immunoreactive ACTH (half-life of 7–12 min).

PHYSIOLOGY Control of ACTH Secretion Figure 2 Corticotroph cells in the anterior pituitary, positively stained immunohistochemically with a basophilic stain.

The secretion of ACTH is controlled by an inherent diurnal rhythmicity; it is augmented by noxious stimuli that are neurally, hormonally, and biochemically

ACTH (Adrenocorticotropic Hormone)

mediated, which is termed stress (open-loop component), and is inhibited by glucocorticoids (closed-loop, negative feedback). In normal circumstances, the daily secretion of ACTH and cortisol is episodic and variable. The highest burst of activity is observed in the early morning hours. Thereafter, the release of both ACTH and cortisol occurs only in 7–15 episodes per day and the levels of both hormones gradually decrease, reaching a nadir at approximately midnight. Diurnal Rhythmicity The basis of the diurnal rhythm is poorly understood. There is evidence for the involvement of at least three factors in the regulation of the diurnal rhythmicity of ACTH: (1) intrinsic rhythmicity of the secretion of the corticotropin-releasing factor (CRH) as well as vasopressin (AVP); (2) light–dark exposure; and (3) feeding times. Diurnal variation frequently disappears during periods of stress and depression and is also changed by conditions that affect cortisol metabolism (liver disease, chronic renal failure, alcoholism). Intrinsic Rhythmicity of Hypothalamic CRH and AVP Tests for the evaluation of the CRH secretion pattern have shown a diurnal intrinsic rhythmicity that persists even in hypophysectomized animals that are deprived of ACTH and glucocorticoid feedback. This hypothalamic rhythmicity appears to be neuronal but not hormonal. On the other hand, the presence of diurnal rhythmicity of ACTH in women during pregnancy when the increased circulating levels of placenta-derived CRH do not show a nychthemeral rhythm strongly suggests that hypothalamic AVP secretion plays a role in this diurnal regulation. Light–Dark Cycles Secretory episodes of ACTH increase between the third and fifth hours of sleep and peak in the morning during the first several hours of wakefulness. This rhythm usually appears after the first year of life but may not be established until the age of 8 years. Reversal of the normal asleep–awake patterns, as occurs when an individual moves to a distant time zone, is followed by a corresponding change in the diurnal pattern of ACTH secretion over the course of 2 to 3 weeks. Feeding Cycles Experiments in rats have shown that feeding schedule is more important than the light–dark cycle in determining the glucocorticoid diurnal secretory pattern. It appears that glucocorticoids tend to be released during fasting and decrease with feeding. Less is

27 known about the effects of feeding schedule on ACTH release in humans. Open-Loop Control (Stress) The open-loop component of ACTH control may be initiated by noxious stimuli of various sorts, all of which represent types of physical or emotional stress, such as pain, fever, trauma, hypoglycemia, hypoxia, surgery, anxiety, and depression. All of these stimuli stimulate the secretion of ACTH via the release of CRH. Corticotropin-releasing hormone and AVP are probably the two major physiologic secretagogues of hypophyseal ACTH. Immunoreactive CRH is found in the human hypothalamus in the paraventricular, supraoptic, and infundibular nuclei and also in the human thalamus, cortex, cerebellum, and pons. However, most human CRH-secreting neurons are located in the anterior portion of the paraventricular nucleus and their nerve endings project to the external layer of the median eminence, where CRH is released into the portal hypophyseal circulation. The same neuronal bodies in the paraventricular nucleus also produce AVP. These AVP neurons are probably the most important in vasopressin control of ACTH release, but the major site of vasopressin neurons is the supraoptic nucleus. These vasopressin neurons are usually of the magnocellular type and most of them project to the neural lobe. There is strong evidence that both a- and b-adrenergic stimuli, cytokines, angiotensin II, and opiates are involved in the regulation of ACTH secretion. During infection, autoimmune processes, or trauma, a complex cascade of events ensues, characterized by fever, circulation of cytokines, and alterations in acute-phase proteins in plasma that are important to initiate, propagate, and terminate host defense mechanisms. In addition, it has been known for several decades that activation of the hypothalamic–pituitary–adrenal (HPA) axis occurs in parallel. It has become apparent that several mediators of inflammation play a major role in this phenomenon. Among all cytokines, three [tumor necrosis factor a (TNFa), interleukin-1 (IL-1), and interleukin-6 (IL6)] are responsible for most of the stimulation of the HPA axis that is associated with the immune/inflammatory response. These three cytokines are produced at inflammatory sites and elsewhere in response to inflammation. Tumor necrosis factor a, which has a tumoricidal activity and is responsible for cachexia, is the first to appear in the inflammatory cascade of the events and stimulates both IL-1 and IL-6; similarly, IL-1 stimulates both TNFa and IL-6. In contrast, IL-6, which participates in a major fashion in the

28 acute-phase reaction, inhibits the secretion of both of the other cytokines. All three inflammatory cytokines have been shown to activate the HPA axis, i.e., ACTH secretion in vivo, alone or in synergy with one another. This effect can be blocked significantly with CRHneutralizing antibodies, glucocorticoids, and prostanoid synthesis inhibitors. When administered to humans, both IL-1 and TNFa have significant toxicity, including fever, general malaise, and hypotension, at the doses needed to activate the HPA axis. In the past, it has been demonstrated that IL-6, with its ability to inhibit the two other inflammatory cytokines and its modest toxicity in experimental animals, was a potent stimulator of the HPA axis in humans, causing an impressively marked and prolonged elevation of plasma ACTH and cortisol when administered either subcutaneously or intravenously. The elevations of ACTH and cortisol attained after stimulation with IL-6 were well above those observed with maximal stimulatory doses of CRH, suggesting that parvocellular AVP and other ACTH secretagogues were also stimulated by this cytokine. In a dose–response study, maximal levels of ACTH were seen at doses at which no peripheral AVP levels were increased. At higher doses, however, IL-6 stimulated peripheral elevations of AVP, indicating that this cytokine might also be able to activate magnocellular AVPsecreting neurons. This suggested that IL-6 might be involved in the genesis of the syndrome of inappropriate secretion of antidiuretic hormone, which is observed in the course of infectious or inflammatory diseases or during trauma. It has been shown that IL-6, in patients with head trauma (an aseptic inflammatory state) and a syndrome of inappropriate secretion of antidiuretic hormone, is quantitatively correlated with AVP. In addition to their hypothalamic effects, the inflammatory cytokines can apparently directly stimulate pituitary ACTH and adrenal cortisol secretion. This may be related to the chronicity of the elevation of the inflammatory cytokines or may be a doserelated phenomenon. It is noteworthy that IL-1 and IL-6 are themselves produced in the anterior pituitary and adrenal glands, where they may have autocrine/ paracrine effects. Closed-Loop Feedback The negative feedback control of ACTH secretion is mediated by cortisol, which exerts inhibitory effects on both the central nervous system and the pituitary. Negative feedback occurs via three mechanisms: (1) fast feedback, which is sensitive to changes in the levels of circulating cortisol, (2) intermediate

ACTH (Adrenocorticotropic Hormone)

feedback, and (3) slow feedback, which is sensitive to the absolute cortisol level. Increased concentrations of glucocorticoids accelerate the progression from fast to slow feedback. Both fast feedback and intermediate feedback appear to be mediated by inhibition of the release of the existing CRH and ACTH rather than by inhibition of their synthesis. Slow feedback is characterized by decreased synthesis of ACTH, complete suppression of POMC gene transcription, and a lack of responsiveness of the corticotroph to the administration of CRH. The last result is mediated by the direct inhibitory effect of cortisol on the pituitary and there is evidence suggesting that this inhibitory effect of cortisol at the level of the pituitary constitutes the most important cortisol negative feedback in the physiological regulation of ACTH secretion. Glucocorticoids decrease the hypothalamic content of CRH and AVP and also decrease AVP mRNA content, but they have only a minimal effect on CRH mRNA. Electrophysiological studies have shown that there are hypothalamic as well as extrahypothalamic sites of cortisol feedback, which serve to suppress the release of CRH.

Action of ACTH The adrenal cortex is the principal target organ for ACTH. ACTH stimulates the synthesis and release of steroids by binding to high-affinity plasma membrane receptors of adrenocortical cells. The ACTH–receptor interaction then activates adenyl cyclase and therefore stimulates the production of intracellular cyclic AMP (cAMP). The cAMP formed activates a number of intracellular phosphoprotein kinases that mediate both acute and chronic effects on steroidogenesis. Acute and Chronic Actions of ACTH ACTH stimulates the synthesis and release of cortisol within 2 to 3 min by increasing free cholesterol formation as a consequence of increased cholesterol esterase activity and decreased cholesteryl ester synthetase activity. ACTH rapidly promotes the transport of cholesterol across the mitochondrial membranes, facilitates the binding of cholesterol to the cytochrome P450scc, and facilitates the release of newly synthesized pregnenolone from the mitochondria. ACTH also stimulates the release of adrenal mineralocorticoids and androgens, as well as the release of various intermediate products. Chronic actions of ACTH are exerted on both adrenal architecture and steroidogenesis. ACTH chronically stimulates low-density lipoprotein (LDL) uptake and metabolism and the synthesis of the LDL receptor and of other factors, so

ACTH (Adrenocorticotropic Hormone)

it has tropic effects on all known early steps in steroidogenesis. Chronic effects of ACTH on steroidogenesis occur mainly by promoting the transcription of the genes that encode steroidogenic enzymes and other factors. ACTH increases the transcription of the genes for P450scc, P450c17, P450c21, and P450c11 and stimulates the accumulation of human P450scc mRNA and human P450scc activity. The exact mechanisms of ACTH stimulation of the side-chain cleavage enzyme P450scc remain to be elucidated. ACTH promotes both adrenal cellular hypertrophy and hyperplasia. ACTH at physiologic concentrations can promote the synthesis of insulin-like growth factor-II (IGF-II) and also the synthesis of basic fibroblast growth factor and epidermal growth factor, which may act with IGF-II to stimulate adrenal growth.

See Also the Following Articles ACTH, a-MSH, and POMC, Evolution of . Adrenal Cortex, Anatomy . Adrenal Cortex, Physiology . Adrenal Suppression . Circadian Rhythms: Hormonal Facets . Corticotropin-Releasing Hormone, Family of . Glucocorticoids, Overview . Stress and Endocrine Physiology

Further Reading Chrousos, G. P. (1995). The hypothalamic–pituitary–adrenal axis and the immune/inflammatory reaction. N. Engl. J. Med. 332, 1351–1362. Desir, D., Van Cauter, E., Fang, V., et al. (1981). Effects of ‘‘jet lag’’ on hormonal patterns. I. Procedures, variations in total plasma proteins and disruption of adrenocorticotropin–cortisol periodicity. J. Clin. Endocrinol. Metab. 52, 628–641. DiBlasio, A. M., Voutilainen, R., Jaffe, R. B., and Miller, W. L. (1987). Hormonal regulation of mRNAs for P450scc (20,22 desmolase) and (17a-hydroxylase/17,20 lyase) in cultured fetal adrenal cells. J. Clin. Endocrinol. Metab. 65, 170–175.

29 Favrod-Coune, C. A., Gaillard, R. C., Langevin, H., Jaquier, M. C., Doci, W., and Muller, A. F. (1986). Anatomical localization of corticotropin-releasing activity in the human brain. Life Sci. 39, 2475–2481. Gionis, D., Ilias, I., Moustaki, M., et al. (2003). Hypothalamic– pituitary–adrenal axis and interleukin-6 activity in children with head trauma and syndrome of inappropriate secretion of antidiuretic hormone. J. Pediatr. Endocr. Metab. 16, 49–54. Grino, M., Dakine, N., Paulmyer-Lacroix, O., and Oliver, C. (2001). Ontogeny of the hypothalamic–pituitary–adrenal axis. In ‘‘Adrenal Disorders’’ (A. N. Margioris and G. P. Chrousos, eds.), pp. 1–10. Humana Press, Totowa, NJ. Jefcoate, C. R., McNamara, B. C., and DeBartolomeis, M. S. (1986). Control of steroid synthesis in adrenal fasciculata cells. Endocr. Res. 12, 315–350. Krieger, D. T. (1979). Rhythms in CRF, ACTH and corticosteroids. In ‘‘Endocrine Rhythms’’ (D. T. Krieger, ed.), pp. 123–142. Raven Press, New York. Malee, M. P., and Mellon, S. H. (1991). Zone-specific regulation of two distinct messenger RNAs for P450c11 (11/18-hydroxylase) in the adrenals of pregnant and non-pregnant rats. Proc. Natl. Acad. Sci. USA 88, 4731–4735. Mastorakos, G., Chrousos, G. P., and Weber, J. S. (1993). Recombinant interleukin-6 activates the hypothalamic–pituitary– adrenal axis in humans. J. Clin. Endocrinol. Metab. 77, 1690–1694. Mesiano, S., Mellon, S. H., Gospodarowicz, D., DiBlasio, A. M., and Jaffe, R. B. (1991). Basic fibroblast growth factor expression is regulated by corticotropin in the human fetal adrenal: A model for adrenal growth regulation. Proc. Natl. Acad. Sci. USA 88, 5428–5432. Miller, W. L., and Tyrell, J. B. (1995). The adrenal cortex. In ‘‘Endocrinology and Metabolism’’ (P. Felig, J. D. Baxter, and L. A. Frohman, eds.), International Edition, pp. 555–680. McGraw Hill, New York. Smith, A. L., and Funder, J. W. (1988). Proopiomelanocortin processing in the pituitary, central nervous system, and peripheral tissues. Endocr. Rev 9, 159–179. Tanaka, K., Nicholson, W. E., and Orth, D. N. (1978). The nature of the immunoreactive lipotropins in human plasma and tissue extracts. J. Clin. Invest. 62, 94–104. Whitfeld, P. L., Seeburg, P. H., and Shine, J. (1982). The human proopiomelanocortin gene: Organization, sequence, and interspersion with repetitive DNA. DNA 1, 133–143.

ACTH, a-MSH, and POMC, Evolution of Robert M. Dores and Phillip B. Danielson University of Denver, Denver, Colorado, United States

Glossary adrenocorticotropin (ACTH) An anterior pituitary polypeptide hormone that induces the production of cortisol by cells of the adrenal cortex. b-lipotropin (b-LPH) A biosynthetic intermediate derived from the proopiomelanocortin precursor that contains the sequence of b-melanocyte-stimulating hormone and b-endorphin.

g0015

melanocortins Adrenocorticotropin, a-melanocytestimulating hormone (MSH), b-MSH, g-MSH, and d-MSH. a-melanocyte-stimulating hormone (a-MSH), b-MSH, g-MSH, d-MSH Polypeptide hormones produced in the intermediate pituitary that share the melanocortin core sequence HFRW.

g0010

proopiomelanocortin (POMC) The common precursor for melanocyte-stimulating hormone-related polypeptides and b-endorphin.

P

roopiomelanocortin is the common precursor for the melanocortins and b-endorphin. This gene is expressed in corticotropic cells of the anterior pituitary and melanotropic cells of the intermediate pituitary. As a result of differential posttranslational processing, these two regions of the pituitary will yield distinct sets of melanocortin-related and b-endorphinrelated end products.

INTRODUCTION During the middle portion of the 20th century, prior to the implementation of immunocytochemistry and subsequent molecular biology procedures to analyze endocrine cells, several pituitary polypeptide hormones were biochemically characterized from extracts of the anterior and intermediate pituitary. These analyses revealed the surprising observation that adrenocorticotropin (ACTH) and b-lipotropin (b-LPH), polypeptides characterized from the anterior pituitary, shared the amino acid sequence motif HFRW with

30

a-melanocyte-stimulating hormone (a-MSH) and b-melanocyte-stimulating hormone (b-MSH), polypeptides characterized from the intermediate pituitary. The fact that ACTH contained the complete sequence of a-MSH and that b-LPH contained the complete sequence of b-MSH led to the hypothesis that all of these polypeptides had a common origin. Later immunocytochemical studies would show that ACTH-related immunoreactivity and b-LPHrelated immunoreactivity were colocalized in the corticotropic cells of the anterior pituitary, whereas a-MSH-related immunoreactivity and b-MSHrelated immunoreactivity were colocalized in the melanotropic cells of the intermediate pituitary. From these observations, several research groups in the 1970s arrived at two conclusions: (1) ACTH, b-LPH, a-MSH, and b-MSH must be derived from a common precursor and (2) the common precursor must undergo differential posttranslational processing to yield distinct sets of end products in the anterior pituitary and intermediate pituitary. The cloning and characterization of proopiomelanocortin (POMC) mRNA from the bovine pituitary by Nakanishi and colleagues in 1979 confirmed the common precursor hypothesis. In addition, that study revealed the presence of a third MSH sequence (g-MSH) in the precursor. Since that time, the POMC gene has been analyzed in representative species from nearly every major group of vertebrate. This article focuses on the changes that have occurred in the POMC gene during the evolutionary radiation of the vertebrates.

ORGANIZATION OF POMC IN TETRAPODS The organization of the POMC precursor is remarkably conserved in tetrapods as diverse as amphibians and mammals. Tetrapod POMC can serve as a good model to illustrate the distribution of biologically active sequences and spacer regions within this precursor and to outline the major issues with respect to the origin and evolution of the POMC gene within

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

31

ACTH, a-MSH, and POMC, Evolution of

Figure 1 The amino acid sequence of Xenopus laevis POMC. The deduced amino acid sequence for the pre-pro form of POMC (Accession No. X59370) is presented. Proposed endoproteolytic cleavage sites are marked by asterisks. MSH core sequences are underlined. The opioid core sequence for b-endorphin is Y202G203G204F205. NTS, N-terminal sequence; JP, joining peptide. the phylum Chordata. Tetrapod POMC can be divided into three major regions: the 16K fragment region, the ACTH region, and the b-LPH region (Fig. 1). The 16K fragment is located in the Nterminal region of POMC immediately following the signal sequence. This region can be subdivided into the N-terminal sequence (NTS), the g-MSH sequence, and the joining peptide sequence. The NTS (Q1–I47) contains four cysteine residues at positions 2, 8, 20, and 24. The number and location of these cysteine residues represent a feature common to all gnathostome POMC sequences and undoubtedly play a role in influencing the three-dimensional structure of the precursor. In vitro studies on developing mammalian pituitary cells indicate that the NTS can serve as a maturation factor to promote the development of prolactin-producing cells. Studies have not been performed in vivo to verify this activity. However, it is coincidental that in teleosts, vertebrates in which the anterior pituitary cell types segregate into zones, ACTH-producing cells are found in close proximity to prolactin-producing cells. The joining peptide region (E79–N111) is one of the most variable regions in POMC and is one of the spacer regions found in this precursor. No function has been found for this sequence. Located between the NTS and the joining peptide is the sequence of g-MSH (K49–Y75). g-MSH is one of three sequences within POMC that contains the MSH core sequence HFRW. Bioassay studies indicate that g-MSH has weak

melanocyte-stimulating activity at nonphysiological concentrations. However, studies on rats indicate that this peptide may be involved in regulating blood pressure at sites outside of the central nervous system. The g-MSH sequence is flanked by paired basic amino acids (R48K49 and K76R77), and these sites undergo endoproteolytic cleavage in the intermediate pituitary but not in the anterior pituitary. In both mammals and anuran amphibians, the R48K49 site is a functional monobasic cleavage site. Hence, for the Xenopus laevis POMC sequence presented in Fig. 1, the first residue in the g-MSH sequence is K49. The reason for this particular cleavage event is not clear. Often, the presence of a proline residue C-terminal to a basic amino acid prevents the removal of the basic amino acid. However, for the R48K49 sequence, the position C-terminal to K49 is a tyrosine residue. In fact, as seen in Fig. 2, the tyrosine residue is conserved at this position in the g-MSH sequences of vertebrates as diverse as sharks and amphibians. In contrast to the R48K49 site, proteolytic cleavage at K76R77 follows the conventional endoproteolytic mechanism and both amino acids are removed. Many gnathostome gMSH sequences have a potential internal cleavage site at R62R63 that can be cleaved to yield an amidated form of g-MSH that would correspond to K49–F60. The ACTH region in the X. laevis POMC sequence extends from A113 to L152. ACTH is a potent stimulator of cortisol production by adrenal cortical cells. This hormone plays a critical role in

32

ACTH, a-MSH, and POMC, Evolution of

Figure 2 Comparison of various vertebrate POMC sequences. The deduced amino acid sequences for the POM from POMC in lamprey (Petromyzon marinus; Accession No. 55629) and the POMC sequences for dogfish (Squalusacanthias; Accession No. AB017198), gar (Lepisosteus osseous; Accession No. U59910), sockeye salmon POMC A (Oncorhynchus nerka), Australian lungfish (Neocerodatus forsteri; Accession No. AF141926), and Xenopus laevis (Accession No. X59370) are presented. Sites that are identical in at least five of the six taxa are shown in boldface type. MSH core sequences are underlined. The opioid core sequence is located at Y319F322.

the hypothalamus–pituitary–adrenal axis in response to chronic stress. The ACTH sequence is flanked by sets of paired basic amino acid proteolytic cleavage sites (K111R112 and R153R154). In the anterior pituitary, both of these cleavage sites are removed to yield ACTH (1–39) as a major end product.

An interesting feature of ACTH is the presence of four basic amino acids located in the interior of the ACTH sequence (R128K129R130R131). In the intermediate pituitary of all vertebrates, endoproteolytic and exoproteolytic cleavage mechanisms remove R128K129R130. However, the presence of P132, a residue

33

ACTH, a-MSH, and POMC, Evolution of

found in all vertebrates at this position, apparently prevents cleavage at R131. The result of these proteolytic cleavage events yields ACTH(1–13) amide (A113– V126) and corticotropin-like intermediate lobe peptide (R131–L152) (CLIP) as products. ACTH(1–13) amide will undergo N-terminal acetylation to form a-MSH, the second polypeptide sequence in amphibian POMC that has the MSH core sequence (H119F120R121W122). a-MSH is a potent stimulator of physiological color change in several species of amphibians and reptiles. The function of CLIP is still an enigma. The final major region of the POMC precursor is the b-LPH region (Fig. 1). Located in the C-terminal portion of POMC, this sequence has traditionally been divided into the g-LPH sequence (E154–D199) and the b-endorphin sequence (Y202–Q232). The former sequence contains the b-MSH sequence (N183–D199), the third polypeptide with a MSH core sequence (H190F191R192W193) in tetrapod POMC. b-MSH has melanocyte-stimulating activity and appears to work in concert with a-MSH to promote physiological color change in amphibians and reptiles. In all vertebrates in which MSH has been studied, b-MSH is an end product of the intermediate pituitary, but not of the anterior pituitary. The other major product derived from b-LPH is b-endorphin. This polypeptide is an endogenous opiate-like chemical signal that functions as an inhibitory neurotransmitter when released from neurons located in the central nervous system. In the mammalian anterior pituitary, b-LPH is a major end product and b-endorphin is a minor end product. In the mammalian intermediate pituitary, just the opposite result is observed. In this tissue, b-LPH serves as a biosynthetic intermediate that is processed to yield Nterminally acetylated, C-terminally truncated forms of b-endorphin. The later peptides lack opiate receptor-binding activity. The organization of tetrapod POMC is interesting for a number of reasons. This precursor contains three polypeptide sequences that have very distinct functions (ACTH, a-MSH, and b-endorphin). The repeat of the MSH core sequence appears to be the result of a series of domain duplication events likely resulting from unequal crossover. This feature is not unique to POMC. Several neuropeptide precursors, such as proTRH or proenkephalin, contain repeats of a biologically important sequence. However, the presence of three MSH sequences in tetrapod POMC leads to the questions of when these sequences appeared and which of the polypeptides was the ancestral MSH sequence. To address these questions, a phylogenetic comparison of POMC sequences is required.

PHYLOGENETIC ANALYSIS OF POMC SEQUENCES The POMC gene has been analyzed in a jawless fish, two cartilaginous fishes, several ray-finned fishes, two lobe-finned fishes, several amphibians, and several mammals. Figure 2 provides a comparison of a few representative species. These species were selected because they represent five major taxonomic groups of vertebrates (i.e., jawless fish, cartilaginous fish, rayfinned fishes, lobe-finned fish, and tetrapods). The jawless fish were well established over 500 million years ago (mya) and are represented today by lamprey and hagfish. In the lamprey, Petromyzon marinus, two distinct POMC genes are expressed. The POM form of POMC, a gene expressed in the intermediate pituitary of this species, is presented in Fig. 2. The jawed vertebrates (gnathostomes) appear in the fossil record approximately 420 mya and two distinct groups were established by 400 million years ago: the cartilaginous fish, represented in Fig. 2 by the dogfish, and the bony fish. The bony fish radiated into two major groups: the ray-finned fish, represented by the gar and sockeye salmon, and the lobe-finned fish, represented by the Australian lungfish. The lungfish lineage can be traced back in the fossil record to 390 mya. Approximately 20 million years later, the first tetrapods were clearly present in the fossil record. X. laevis was included in Fig. 2 to provide a representative tetrapod POMC sequence. Two ray-finned fish POMC sequences are also included in Fig. 2 to illustrate some unique features associated with the radiation of the POMC gene in this group. The gar belongs to an older lineage of ray-finned fish that can be traced in the fossil record back to the Jurassic era. The sockeye salmon is a teleost, a relatively recent group of bony fish that first appears in the fossil record during the Cretaceous period. A striking feature of the sequences presented in Fig. 2 is the presence of a-MSH-like, b-MSH-like, and b-endorphin sequences in all taxa. There has been considerable divergence of the POMC sequence in vertebrates and without these conserved sites it would have been very difficult to align the agnathan POM sequence to the gnathostome POMC sequences. Even with these conserved sites, it was still necessary to insert 11 gaps to facilitate the alignment of the sequences. The presence of a-MSH-like, b-MSHlike sequences in the lamprey precursor indicates that the POMC gene must have been in the ancestral chordate. Indeed, this gene may have been in the ancestral eucoelomates. Based on this data set, it is not possible to ascertain when the a-MSH/b-MSH duplication event occurred. One hypothesis would be

34

ACTH, a-MSH, and POMC, Evolution of

that the ACTH sequence appeared early in eucoelomate evolution and that b-MSH is the result of the duplication of the a-MSH sequence. In the case of the lamprey, nucleotide insertions occurred in both the aMSH- and the b-MSH-coding regions of the lamprey POM gene. The gnathostome a-MSH and b-MSH sequences are highly conserved in the taxa presented in Fig. 2. In addition, the b-endorphin region contains a remarkable number of conserved sites in all taxa. There must be selection pressure to retain certain features of the b-endorphin sequence that is not being exerted on other regions of the gene. A sequence that is conspicuously absent from the lamprey POM sequence is g-MSH. Based on this observation, it is reasonable to propose that the duplication event that yielded the g-MSH sequence must have occurred in the ancestral gnathostomes. Although a gMSH-like sequence can be detected in dogfish, gar, and Australian lungfish POMC, this sequence is shorter than the tetrapod g-MSH sequence. In addition, paired basic proteolytic cleavage sites are not found C-terminal to the g-MSH sequence in the gar or lungfish sequences. In addition, the gar g-MSH-like sequence does not have an intact MSH core sequence and the entire g-MSH sequence is absent in teleost POMC. Apparently during the radiation of the ray-finned fish, selection pressures favored the gradual degeneration and eventual loss of the g-MSH sequence. Whereas the presence of a-, b-, and g-MSH is a feature common to many tetrapods, the dogfish POMC sequence has a fourth MSH sequence: d-MSH (D223–P237; Fig. 2). This sequence is located in a large insertion (V204–A242; Fig. 2) that appears to have occurred during the radiation of the cartilaginous fish and hence is limited to this group of gnathostomes. The uniqueness of the d-MSH sequence is balanced by the high degree of primary sequence conservation observed for ACTH among the gnathostome taxa. For the gnathostome species presented in Fig. 2, 66% of the positions are identical in the ACTH sequence. In fact, 88% of the first 25 positions in gnathostome ACTH are

Addison’s Disease see Adrenal Insufficiency

identical. For the b-endorphin region, 45% of the positions are identical in all taxa. This percentage rises to 67% when all gnathostome fish sequences are compared. Clearly, ACTH and b-endorphin must be physiologically important in all vertebrates.

ORIGIN OF POMC POMC is a member of the opioid/orphanin gene family that includes the proenkephalin gene, the prodynorphin gene, and the proorphanin gene. A unifying feature of this family is the presence of at least one Y(F)GGF sequence in each precursor. To date, members of this gene family have not been found in any prokaryotes and it appears that these genes are not found in pseudocoelomates such as Caenorhabditis elegans. Hence, the gene family appears to have evolved in the eucoelomates. Although POMC is clearly a member of this family because of the b-endorphin sequence, the presence of the melanocortin sequences (ACTH, a-MSH, b-MSH, g-MSH, and d-MSH) is a feature that is unique to the POMC gene. Since POMC is found in all vertebrate groups and perhaps is common to all chordates, the origin of the melanocortin sequences may provide the key to deciphering the genesis of this opioid/orphanin gene family.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Pituitary Tumors, ACTH-Secreting

Further Reading Cone, R. D. (ed.) (2000). ‘‘The Melanocortin Receptors.’’ Humana Press, Totowa, NJ. Danielson, P. B., and Dores, R. M. (1999). Molecular evolution of the opioid/orphanin gene family. Gen. Comp. Endocrinol. 113, 169–186. Hadley, M. E. (ed.) (1988). ‘‘The Melanotropic Peptides,’’ Volumes 1, 2, and 3. CRC Press, Boca Raton, FL. Norman, A. W., and Litwack, G. (1997). ‘‘Hormones,’’ 2nd ed. Academic Press, San Diego, CA.

Adenylyl Cyclase Ferenc A. Antoni University of Edinburgh, Edinburgh, United Kingdom

Glossary cyclic nucleotide phosphodiesterases (PDEs) Enzymes that break down cyclic nucleotides into nucleotide monophosphates. guanosine-50 -triphosphate (GTP) Required for the activity of G proteins to modulate effectors such as adenylyl cyclase; hydrolyzed by G proteins into guanosine-50 -diphosphate (GDP) before the GDP is released and new GTP is bound and effector activation is resumed (‘‘GTP-ase switch’’). heterotrimeric G proteins A group of proteins that consist of three subunits (a, b, g.) and couple cell surface receptors to their effector enzymes; on activation of cell surface receptors by their ligands, heterotrimeric G proteins dissociate into the a-subunit and the bg–complex, both of which regulate the activity ¨ of effector enzymes such as adenylyl cyclase in the cell membrane. protein kinases Enzymes that tag proteins with phosphoryl residues derived from ATP; activity may be regulated by a variety of intracellular messengers, protein–protein interactions, or phosphorylation by other protein kinases. protein phosphatases Reverse the action of protein kinase and remove phosphoryl groups from proteins. second messengers Generated by extracellular stimuli arriving at the cell surface; transduce the extracellular signal toward the interior of the cell.

A

denylyl cyclases (EC 4.6.1.1) are enzymes that generate the second messenger adenosine-30 50 -monophosphate from ATP.

INTRODUCTION Adenosine-30 50 -monophosphate (cAMP) was the first ‘‘second messenger’’ to be discovered in 1956 by Earl Sutherland. As a prototype and, for a long time, as the only identified intracellular signaling pathway, the

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

cAMP cascade was the subject of a large number of studies until the mid-1980s, when interest shifted toward novel signaling mechanisms such as polyphosphoinositides and receptor tyrosine kinases. The fact remains, however, that cAMP is a ubiquitous second messenger from gametogenesis to cognition. The application of molecular genetics to the field has highlighted the molecular diversity in the biosynthetic and catabolic arms (cyclic nucleotide phosphodiesterase, PDE) of cAMP metabolism. Ten adenylyl cyclase (AC) genes have been identified in mammals. Diversification of AC genes is already apparent in Drosophila melanogaster and Caenorhabditis elegans, indicating a distinct adaptational advantage of this feature. Taking into account the currently known molecular variants of AC and PDE, the turnover of cAMP in a mammalian cell may occur in up to 500 different ways. The biological significance of this diversity is only slowly becoming apparent and is currently under investigation in numerous biological systems.

ADENYLYL CYCLASE GENE AND PROTEIN STRUCTURE Two types of ACs have been identified in mammals: a membrane-bound species representing 9 of the 10 known ACs and a soluble enzyme. All of the membrane-bound ACs have the same predicted structure (Fig. 1), resembling that of ion channel/transporter proteins. The N-terminal end is cytoplasmic and continues in the M1 intramembrane segment consisting of six membrane-spanning helices. This is followed by a large cytoplasmic loop (C1), which can be subdivided into C1a and C1b on the basis of the conservation of the C1a segment in other ACs as well as a corresponding region (C2a) in the second cytoplasmic domain. The C1b portion of the cytoplasmic loop is a nonconserved isotype-specific segment. The second half of the molecule is analogous to the first half. The six transmembrane helices of the M2

35

36

Adenylyl Cyclase

Figure 1 Schematic representation of mammalian ACs. N: N-terminal cytoplasmic loop; M1 and M2: intramembrane domains; C1a and C2a: homologous catalytic domains (thick lines); C1b and C2b: nonhomologous regulatory domains (thin lines). The C1a and C2a domains must come into physical contact to form the catalytic core. This is facilitated by Gsa that binds to the C2a domain. Most ACs are inhibited by Gsa that binds to a site on the C1a domain. intramembrane segment are followed by a cytoplasmic tail designated C2, which is usefully divisible into C2a and C2b because C2a is relatively well conserved in all known ACs and also shows significant homology to C1a. Both intramembrane domains are predicted to be glycosylated on the extracellular surface. The C1 and C2 segments of ACs also show highly significant homology to guanylyl cyclases. In fact, a point mutation in the C1a region is sufficient to change an AC into a guanylyl cyclase enzyme. Note that for historical reasons, the sequence motif corresponding to the conserved catalytic core in ACs is called guanylyl cyclase motif in protein structure databases. Mini-protein heterodimers consisting of the C1a and C2b domains of ACs are catalytically active and retain numerous regulatory properties of the holoenzyme. Furthermore, high-resolution X-ray crystallographic maps of the secondary structure of C1a and C2a heterodimers have been reported. The results concur that the catalytic core of the enzyme, including the substrate binding site, requires both domains. The heterodimer C1a:C2a is a symmetrical array; guanosine-50 -triphosphate (GTP)–Gsa is bound by the C1a domain, whereas forskolin, an alkaloid known to stimulate ACs, binds to C1a and C2a in the vicinity of the Gsa binding site in C1a (Fig. 2). The C1–C2 complex also contains two metal-binding sites, which under physiological circumstances are occupied by Mg2þ. One of these sites attracts Mg2þ

in complex with ATP, whereas the other site binds Mg2þ and has a preference for Mn2þ, explaining early observations on the stimulatory action of Mn2þ on ACs. The 10th mammalian AC is a soluble enzyme that is structurally more closely related to cyanobacterial ACs than to the membrane-spanning family of mammalian ACs. The crystal structure of this enzyme was not yet published at the time of this writing.

REGULATION OF ADENYLYL CYCLASE General Considerations There are currently 10 mammalian AC genes, each of which encodes a different protein. These are differentially controlled by heterotrimeric G proteins, intracellular Ca2þ, and protein phosphorylation (Table I). In addition, ACX is uniquely controlled by bicarbonate ions. Hence, several signaling pathways converge on ACs, which act as molecular signal integrators. In a further aspect, AC isotypes have distinct tissue distributions, indicating nonredundant physiological roles. At the single-cell level, expression of multiple AC isotypes is the rule rather than the exception. Thus, polarized distribution of ACs within cellular microdomains or compartments is yet another element of functional diversity in the cAMP signaling cascade.

37

Adenylyl Cyclase

f0010

Figure 2 The structure of the AC catalytic core. The C1a and C2a domains of dog ACV (VC1) and rat ACII (IIC2) in complex with GTP–Gsa are shown. Note the pseudo-symmetrical arrangement of a-helices (represented by cylinders) numbered 1 to 7 and b strands (represented by arrows) of the C1a and C2a domains. ATP and forskolin (FSK) are drawn as stick models; Gsa is represented as a cylinder. The S and T denote serine and threonine residues that are putative phosphorylation sites in ACIII and ACII, respectively. Note that residues from both polypeptide chains are required to form the substrate-binding site. Reproduced from Tesmer, J. J., and Sprang, S. R. (1998). The structure, catalytic mechanism, and regulation of adenylyl cyclase. Curr. Opin. Struct. Biol. 8, 714, Elsevier.

G Proteins p0035

The activity of all transmembrane ACs is stimulated by Gsa combined with GTP. The structure–activity analysis of the AC catalytic core by various methods has identified the site of interaction with GTP–Gsa (Fig. 2) in the cytosolic C2a domain. The binding of GTP–Gsa to C2a is thought to facilitate the association of C1 and C2. The activity of some but not all ACs is inhibited by Gia (Table I), thereby opposing the stimulation by Gsa. In addition to G protein alpha subunits, Gßg subunits have a direct influence on the activity of some ACs. Above all, in the presence of Gsa, activity of ACII and ACIV is enhanced further by Gßg. Thus, ACII and ACIV are coincidence detectors for receptors coupled to Gs and Gi/o, respectively. It is important to note that Gßg derived from Gs or Gq is unlikely to exert this type of effect given that the EC50 of Gßg for this action is approximately 100 nmol/L, whereas the stimulatory effect of Gsa is already maximal at approximately 10 nmol/L and the tissue levels of this G protein are low compared with those of Gi and Go. The potential for Gßg to stimulate ACs is remarkable because it shows that the direction as well

as the magnitude of the effects of hormones on cAMP synthesis can be dependent on the context of the stimulus applied. In contrast to ACII and ACIV, Gßg subunits reportedly inhibit the activation of ACI by Gsa. A summary of the actions of G proteins is shown in Fig. 3.

Calcium ACI was the first AC protein to be purified and the first AC cDNA to be cloned. It is an enzyme stimulated by Ca2þ–calmodulin and is highly abundant in the brain. The calmodulin modulatory site is in the C1b domain of ACI. The Ca2þ–calmodulin complex and GTP–Gsa produce synergistic stimulation of ACI. ACVIII is also activated by Ca2þ–calmodulin, but apparently through a different, IQ-type, calmodulin-binding domain, and no synergy with Gsa has been reported. Thus, these enzymes are capable of directly responding to a rise of intracellular-free Ca2þ from a variety of sources, especially plasma membrane calcium channels. Another mode of regulation by Ca2þ is an apparently direct inhibition of activity. ACVI is the best characterized example of

Table I Mammalian AC Genes

Cyclase

G proteins

ACI

Gsa stimulates, Gbg inhibits

ACII

Calcium–Calmodulin Calmodulin-dependent stimulation, synergy with Gsa

Phosphorylation

Facets of tissue distribution

Human chromosomal localization

Other

Inhibition by CaMKIV

Mainly in the brain

7p13-p12

Gsa stimulates, Gbg synergizes with Gsa

Stimulation by protein kinase C in synergy with Gsa

Brain, lung, uterus pituitary

5p15.3

ACIII

Gsa stimulates

Inhibition by CaMKII

Brain, kidney, testis, vasculature, pancreas beta cells

2p24-p22

Important for olfaction, involved in insulin release

ACIV

Gsa stimulates

Inhibition by protein kinase C

Lungs

14q11.2

Very little known about this enzyme

ACV

Gsa stimulates, Gia inhibits

Direct inhibition by Ca2þ

Stimulation by protein kinase C

Brain, heart

3q13.2-q21

Effect of protein kinase C not dramatic in vivo, has splice variant, required for EGF stimulation of cAMP

ACVI

Gsa stimulates, Gia inhibits

Direct inhibition by Ca2þ

Inhibition by protein kinase A and C, enhancement by tyrosine kinase

Widespread, including pituitary, adrenal cortex

12q12-q13

ACVII

Gsa stimulates

Not known

Brain, retina, pituitary, hematopoetic system

16q12-q13

Has splice variant

ACVIII

Gsa stimulates

Brain specific

8q24.2

Has splice variants

ACIX

Gsa stimulates, Gia inhibits

Inhibited by dephosphorylation pathway involving calcineurin

Widespread, including major endocrine glands, prostate, uterus

16p13.3

Participates in adrenal corticosteroid feedback

Not known

Testis, also present in other tissues

1q24

Modulated by bicarbonate, highly expressed in spermatogonia

ACX N/A (soluble)

Calmodulin-dependent stimulation

Not known

Participates in synaptic plasticity

Adenylyl Cyclase

f0015

39

Figure 3 Summary of the control of AC isotypes by G proteins. Arrows indicate stimulation; T bars indicate inhibition; R indicates G-protein-coupled receptor.

this control. The inhibition is apparent between submicromolar (>100 nmol/L) to low-micromolar concentrations of Ca2þ. The structural requirements for direct inhibition of ACs by Ca2þ are not known.

Protein Phosphorylation p0050

Typically, ACs contain a multitude of potential consensus phosphorylation sites for a variety of protein kinases. Because these are large molecules that cannot be assayed reliably in single-cell assays, the characterization of control by phosphorylation has been very slow to appear. The most striking effects have been reported for ACII and ACVII. The activity of both enzymes is markedly stimulated by protein kinase C. Whether protein kinase C phosphorylation alone is sufficient to activate the enzymes in normal cells is unclear, but synergistic stimulation by GTP–Gsa and

Figure 4 Summary of the modulation of AC activity by protein kinase C. The a and j variants of protein kinase C have been implicated in the control of ACs.

Figure 5 Distribution of mRNAs for cytochrome CYP11B2, which is responsible for the synthesis of aldosterone in the zona glomerulosa, cytochrome CYP11B1,which in turn is responsible for the synthesis of corticosterone in the zona fasciculata–reticularis and various ACs in the rat adrenal gland detected by in situ hybridization histochemistry. Animals were treated with vehicle or ACTH for 10 days before processing of the adrenal gland to in situ analysis. Reproduced from Endocrinology (1997), Vol. 138, p. 4596, with permission of The Endocrine Society, Bethesda, MD.

activators of protein kinase C is well established. In the case of ACII, the site of phosphorylation has been localized to the C2b domain. Inhibition of ACVI by protein kinase C has also been reported; here there are multiple sites of phosphorylation, one of which is in the N-terminal segment of the protein. The effects of protein kinase C activation on ACs are summarized in Fig. 4. ACVI is also inhibited by protein kinase A (cAMP-dependent protein kinase), and mutagenesis studies indicated that the site of phosphorylation is in the C1b segment. Thus, ACVI is inhibited by

40

p0060

Adenylyl Cyclase

cAMP-dependent negative feedback. In a similar vein, ACI is inhibited by calmodulin-dependent protein kinase IV. Calmodulin-dependent protein kinase II is inhibitory to ACIII, which appears to be important in the olfactory system of rodents where very rapid restoration of cellular responsiveness is vital. Yet another mode of Ca2þ-mediated inhibition is apparent in the case of ACIX, which is inhibited by a pathway involving the Ca2þ–calmodulin-dependent protein kinase calcineurin (protein phosphatase 2B). The nature of the presumed stimulatory phosphorylation is not known.

aldosterone-synthesizing zona glomerulosa, which expresses mainly ACVI and the zona fasciculata, where the predominant AC appears to be ACIX. These characteristics are retained and accentuated after treatment of the animals with adrenocorticotropic hormone. It was also shown that sodium restriction causes a marked increase of ACVI mRNA expression in the zona glomerulosa. Other studies of AC expression in endocrine tissues have not been consistent; however, there is little doubt that similar segregation of functionally distinct ACs is not restricted to the adrenal gland.

ADENYLYL CYCLASES IN ENDOCRINE SYSTEMS

See Also the Following Articles

It is quite clear that without ACs, hormonal control would simply not function. Several hypothalamicreleasing hormones exert their actions through cAMP, and the effects of the majority of pituitarytropic hormones that regulate the peripheral endocrine glands are mediated, at least in part, by cAMP. Interestingly, no obvious endocrine abnormalities have emerged from the three gene deletion studies published as of this writing that targeted ACI, ACIII, and ACVIII, respectively. This is plausibly due to the fact that these ACs are found mainly in the central nervous system or that, as in the cases of ACI and ACVIII, a double gene deletion was necessary to demonstrate a clear behavioral phenotype. Data on the role of AC isotypes in endocrine systems are very limited indeed. The rat adrenal gland is the best explored tissue (Fig. 5) from this respect. There is striking compartmentalization of the ACs in the medulla and cortex. Furthermore, within the cortex, there are differences between the

G Protein-Coupled Receptors . G Proteins and Effectors . Lipid Second Messengers and Receptors . Receptor Tyrosine Kinase

Further Reading Antoni, F. A. (2000). Molecular diversity of cyclic AMP signaling. Front. Neuroendocrinol. 21, 103–132. Cooper, D. M. F., Karpen, J. W., Fagan, J. W., and Mons, N. E. (1998). Ca2þ-sensitive adenylyl cyclases. In ‘‘Advances in Second Messenger and Phosphoprotein Research’’ (D. M. F. Cooper, ed.), Vol. 32, pp. 23–52. Lippincott–Raven, Philadelphia, PA. Patel, T. B., Du, Z. Y., Pierre, S. C., Cartin, L., and Scholich, K. (2001). Molecular biological approaches to unravel adenylyl cyclase signaling and function. Gene 269, 13–25. Tang, W.-J., Yan, S., and Drum, C. L. (1998). Class III adenylyl cyclases: Regulation and underlying mechanisms. In ‘‘Advances in Second Messenger and Phosphoprotein Research’’ (D. M. F. Cooper, ed.), Vol. 32. Lippincott–Raven, Philadelphia, PA. Taussig, R., and Gilman, A. G. (1995). Mammalian membranebound adenylyl cyclases. J. Biol. Chem. 270, 1–4. Tesmer, J. J., and Sprang, S. R. (1998). The structure, catalytic mechanism, and regulation of adenylyl cyclase. Curr. Opin. Struct. Biol. 8, 713–719.

Adipocytokines Tohru Funahashi and Iichiro Shimomura Osaka University Graduate School of Medicine, Osaka, Japan

Yuji Matsuzawa Osaka University Graduate School of Medicine, Sumitomo Hospital, Osaka, Japan

Glossary adipovascular axis Adipose tissue secretes various vasoactive substances. Dysregulation of these adipocytederived vasoactive molecules in obesity directly contributes to the development of vascular diseases. lipodystrophy A disease in which there is disturbed differentiation of adipocytes. Human generalized lipodystrophy and bioengineered mice lacking fully differentiated adipocytes exhibit phenotypes resembling that of obesity, such as insulin resistance, dyslipidemia, and fatty liver. metabolic syndrome The condition characterized by upper body obesity, insulin resistance, dyslipidemia, and hypertension. The metabolic syndrome is a common cause of atherosclerotic vascular diseases. visceral fat Adipose (fat) tissue located in the omentum and mesenterium. Accumulation of visceral fat is a common cause of the metabolic syndrome.

O

besity is a common risk factor for type 2 diabetes and cardiovascular diseases, and it is a major health problem in industrialized countries; however, the molecular basis for the link between obesity and obesity-related diseases has been unclear. Traditionally, adipose tissue has been regarded as an organ passively storing excess energy. However, research on adipocyte biology has found that adipocytes produce and secrete a variety of biologically active molecules, including growth factors, cytokines, and complement factors, in the immune system.

INTRODUCTION The adipocyte-derived factors (Fig. 1) affect the function of adipocyte in an autocrine and a paracrine fashion and affect whole body homeostasis through the bloodstream. These findings indicate that adipose tissue is an endocrine organ and the adipocyte-derived bioactive substances are adipocytokines. Because these adipocytederived substances include the molecules belonging to

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

the strict cytokines, they are also called adipokines. Every cell type secretes various cytokines and bioactive substances to the surrounding milieu to maintain normal functions. The importance of adipocytokines is highlighted by the fact that adipose tissue is the largest organ in the body. Although the amount of adipocytokines produced by a single adipocyte is small, the total amount in the body greatly affects whole body functions. Another notable feature is that adipose tissue is supplied by the bloodstream, and adipocytokines released from adipocytes easily flow into the systemic circulation. Findings obtained from bioengineered mice with undifferentiated adipose tissue revealed the importance of adipocytokines in maintaining metabolic homeostasis. Adipocytokines play a significant role in the self-defense system against metabolic overload and probably also in the immune system. There is substantial evidence that dysregulated production of adipocytokines is involved in the development of obesity-related diseases. Tumor necrosis factor-a (TNF-a), leptin, and resistin affect insulin sensitivity in the whole body. Overproduction of TNF-a has been suggested to contribute at least in part to the development of insulin resistance, which is a major clinical feature of obesity. Furthermore, adipose tissue also produces various vasoactive substances, such as plasminogen activator inhibitor Growth factors

PAI-1 HB-EGF

Cytokines

Complement factors

TNF-α Leptin

Adiponectin

Figure 1 Adipocytokines.

41

42 (PAI-1), an inhibitor of the fibrinolytic system, and heparin-binding epidermal growth factor-like growth factor (HB-EGF), a potent growth factor for vascular smooth muscle cells. These factors are also overproduced in obesity. These findings suggest the existence of an adipovascular axis in which adipocytederived factors affect vascular functions independent of the coronary risk factors frequently associated with obesity. Although the total amount of adipose tissue is increased in obesity, not all adipocytokines are overproduced. For example, adipsin is factor D of the complement system. The expression of adipsin in adipose tissue is severely impaired in obesity and results in the reduction of plasma concentration. The physiological meaning of the decreased plasma adipsin in obesity is unclear. Next, adiponectin belongs to the soluble defense collagen superfamily. Although the protein possesses antiatherogenic and antidiabetic properties, its plasma concentration is decreased in obesity. The proteome analysis proved that there is impaired secretion of various molecules from the adipocytes of obese subjects. Not only hypersecretion of offense adipocytokines but also hyposecretion of defense adipocytokines affects the development of obesity-related disorders.

LEPTIN Leptin is a 16-kDa protein secreted primarily from adipocytes. Rodents defective in leptin synthesis (ob/ob mice) or leptin receptor function (db/db mice, Zucker fa/fa rats, and Koletsky rats) are obese and develop hyperinsulinemia and insulin resistance. Leptin suppresses food intake and increases energy expenditure by enhancing thermogenesis and metabolic rate. These functions seem to be mediated mainly by the central nervous system because intracerebroventricular injection of leptin produced significant effects with much smaller amounts than those required by systemic injection. Leptin receptors (OB-R) are single membranespanning receptors with homology to members of the cytokine receptor superfamily. Seven different leptin receptors, produced by alternative splicing, have been identified. The receptors containing transmembrane domains can be divided into two groups. One group has a short amino acid residue intracellular domain (OB-Ra, OB-Rc, OB-Rd, and OB-Rf). The other group has a long intracellular domain (OB-Rb). OB-Rb is mainly expressed in hypothalamus, whereas the short forms are expressed in a variety of tissues.

Adipocytokines

The long form of the receptor has two Janus kinase sites in the intracellular domain and the short form has one. Only the long form can activate the signal transducers and activators of the transcription family (STAT). C57B1/Ks db/db mice, which lack the long form of the receptor and have intact short forms, exhibited almost identical phenotype to ob/ob mice, which lack leptin. There is accumulating evidence indicating that leptin mimics some of the insulin actions in liver, adipose tissue, and muscle. In diabetic rats, leptin increased glucose uptake in muscle and brown adipose tissue and normalized hyperglycemia. In liver and hepatocytes, leptin lowered hepatic glucose production by decreasing glycogenolysis and increasing glycogen synthesis. The physiological significance of leptin was emphasized by the successful administration of leptin to treat the metabolic disorders of lipodystrophy in mice and humans. Generalized lipodystrophy is a disorder characterized by a paucity of adipose (fat) tissue accompanied by a severe resistance to insulin, leading to hyperinsulinemia, hyperglycemia, and an enlarged fatty liver. In the two genetically engineered mouse models of lipodystrophy, the plasma concentration of leptin was very low due to the loss of mature adipose tissues. Leptin supplementation overcame insulin resistance, diabetes, hyperlipidemia, and fatty liver in both models. These effects were not observed with chronic food restriction. The results support the theory that leptin exerts favorable effects on glucose and lipid metabolism independently of its effect on food intake. Based on these observations, human lipodystrophy patients were recently treated with leptin, resulting in an improvement in metabolic disorders and fatty liver.

TNF-a It has been shown that the adipose TNF-a mRNA and plasma TNF-a protein are increased in most animal models and human subjects with obesity and insulin resistance. Neutralizing the blood TNF-a in obese rats with a soluble TNF-a receptor–IgG fusion protein markedly improved insulin resistance. These results indicated that the higher production of TNF-a in accumulated adipose tissue was causative for obesity-associated insulin resistance. TNF-a treatment reduced insulin-stimulated autophosphorylation of the insulin receptor and IRS-1 phosphorylation in various tissue cultured cells, including adipocytes, fibroblasts, and hepatoma cells. These disturbances of insulin signaling were also

43

Adipocytokines

observed in the muscle and adipose tissues of the obese and insulin-resistant fa/fa rats. It has been reported that TNF-a induces serine phosphorylation of IRS-1 in cultured adipocytes and hepatocytes. The serine phosphorylated form of IRS-1 was also increased in the muscle and adipose tissues of fa/fa rats. The modified IRS-1 inhibited the autokinase activity of the insulin receptor, resulting in the deterioration of insulin signaling. Hotamisligil and colleagues fed wild-type and TNF-a knockout mice with a high-fat diet. Both types of mice became similarly obese, but obese TNF-a(
/
) mice maintained high insulin sensitivity. These results demonstrated that TNF-a deficiency blocked the development of insulin resistance associated with diet-induced obesity. In the ob/ob mice with targeted mutations in both p55 and p75 TNF-a receptors, the signaling and function of TNF-a were completely abolished. Without the changes in body weight, the ob/ob mice with null mutations of TNF-a receptors showed higher insulin sensitivity than the ob/ob mice with normal TNF-a receptors. These results further support the theory that the enhanced TNF-a protein and TNF-a signaling are involved in obesity-associated insulin resistance.

PAI-1 PAI-1 is an inhibitor of plasminogen activators and fibrinolytic activity. Although PAI-1 is synthesized in endothelial cells and liver, adipose tissue is also a main source of plasma PAI-1. Expression of PAI-1 is augmented in adipose tissue, especially intraabdominal visceral fat in obesity. The amount of visceral fat is positively correlated with the plasma level of PAI-1 and is one of major determinants of plasma PAI-1. It is well-known that the plasma concentration of PAI-1 is elevated in subjects with type 2 diabetes and hypertriglyceridemia, although the precise mechanism is unclear. Type 2 diabetes and hypertriglyceridemia are often associated with visceral fat accumulation. PAI-1 produced by accumulated visceral fat may explain the high plasma PAI-1 in these conditions. The reason why adipocytes secrete PAI-1 has not been clarified. Adipocytes dramatically change cell size in response to nutritional conditions. Plasmin works to destroy the basement membrane to facilitate cell expansion. PAI-1 may control the activity of plasminogen activators to prevent the overproduction of plasmin.

ADIPONECTIN RESISTIN Resistin is a 12.5-kDa, cysteine-rich protein identified by screening for the genes that were induced during the differentiation of the adipocytes but were downregulated in mature adipocytes exposed to glitazone, an insulin-sensitizing drug. Mouse resistin contains 114 amino acids and circulates as a homodimer of two peptides. Administration of resistin to mice impaired glucose tolerance and insulin action. Plasma concentrations of resistin were higher in genetic and diet-induced obese mice with insulin resistance. Administration of the neutralizing antibody against resistin increased insulin sensitivity in obese mice. These results suggest that resistin is a fat-derived factor causing insulin resistance in obesity and that the insulin-sensitizing effect of glitazone can be attributed to its inhibition of resistin expression. One study, however, showed contradictory results. It was reported that resistin expression was decreased in obese mice and increased in response to glitazone. The human homologue of resistin is located on chromosome 19p13.3, a region not previously implicated in the susceptibility to obesity, insulin resistance, or diabetes. Its significance in humans remains to be clarified.

Adiponectin is an adipocyte-derived factor identified through the extensive search of adipose tissue transcripts in the human genome project. The expression of adiponectin mRNA is exclusive in adipose tissue. The protein is composed of two structurally distinct domains—the C-terminal collagen-like fibrous domain and the complement C1q-like globular domain. Adiponectin is abundant in the circulating plasma in a multimeric form. Interestingly, plasma concentrations of adiponectin are decreased in obese subjects despite its restricted expression in adipocytes. Plasma adiponectin levels are also lower in patients with coronary artery disease and type 2 diabetes than those in body mass index-matched subjects. Physiologically, adiponectin inhibits the differentiation of adipocytes in a paracrine manner via the COX-2 pathway. When the endothelial barrier is injured, adiponectin accumulates in the subendothelial space of the vascular walls. The protein has antiatherogenic properties, such as suppression of monocyte attachment to vascular endothelial cells via the reduced expression of adhesion molecules, suppression of foam cell formation and TNF-a secretion of macrophages, and suppression of the growth factorinduced proliferation of vascular smooth muscle cells. Administration of adiponectin also improves fatty

44 oxidation and insulin resistance in dietary-induced and genetically obese animals. Adiponectin is cleaved between fibrous and globular domains. The globular form of adiponectin has more potent activity for insulin sensitization than the whole protein. Thus, adiponectin has a dual function on insulin sensitivity and vascular functions. Hypoadiponectinemia in obesity may be a key factor in the metabolic syndrome, which is often accompanied by insulin resistance and cardiovascular diseases. Interestingly, patients with genetic hypoadiponectinemia caused by a missense mutation in the adiponectin gene also exhibit the clinical phenotype of the metabolic syndrome. Studies of mice genetically lacking adiponectin confirmed the significance of adiponectin in the metabolic syndrome. Two cohort studies of a specialized population report that hypoadiponectinemia is a risk for cardiac death or the developments of type 2 diabetes. The genetic variation in the adiponectin gene is also associated with an increased risk of type 2 diabetes in the Japanese population. Further

Adipocytokines

studies are necessary to verify the significance of adiponetin in the metabolic syndrome and obesityrelated diseases.

See Also the Following Articles Diabetes, Type 2 . Leptin Necrosis Factor (TNF)

.

Obesity Regulation

.

Tumor

Further Reading Friedman, J. M., and Halaas, J. L. (1998). Leptin and the regulation of body weight in mammals. Nature 395, 763–770. Funahashi, T., Nakamura, T., Shimomura, I., Maeda, K., Kuriyama, H., Takahashi, M., Arita, Y., Kihara, S., and Matsuzawa, Y. (1999). Role of adipocytokines on the pathogenesis of atherosclerosis in visceral obesity. Intern. Med. 38, 202–206. Matsuzawa, Y. (1997). Pathophysiology and molecular mechanisms of visceral fat syndrome: The Japanese experience. Diabetes Metab. Rev. 13, 3–13. Spiegelman, B. M., Choy, L., Hotamisligil, G. S., Graves, R. A., and Tontonoz, P. (1993). Regulation of adipocyte gene expression in differentiation and syndromes of obesity/diabetes. J. Biol. Chem. 268, 6823–6826.

Adrenal Androgens Lucia Ghizzoni and Alessandra Vottero University of Parma, Parma, Italy

Glossary adrenal gland Triangular, endocrine, hormone-secreting gland located on top of each kidney. androgens Sex hormones that stimulate the development of secondary masculine characteristics (i.e., deep voice, facial hair). receptor Molecule on the cell surface or inside the cell that recognizes a specific antigen or hormone.

T

he adrenal androgens (AAs), normally secreted by the fetal adrenal zone or the zona reticularis, are steroid hormones with weak androgenic activity. Although AAs do not appear to play a major role in the fully androgenized adult man, they seem to play a role in the adult woman and in both sexes before puberty. Girls, women, and prepubertal boys may be negatively affected by AA hypersecretion, in contrast to adult men. This article discusses AA biosynthesis, regulation, physiology, and biological action.

ADRENAL ANATOMY AND ANDROGEN BIOSYNTHESIS Adrenal Gland Anatomy The adrenal glands, consisting of the cortex and medulla, have a roughly pyramidal shape and lie above the upper poles of the kidneys. The two zones receive their blood supply from branches of the phrenic arteries, aorta, and renal arteries. Arterial blood enters from the outer cortex, flows through fenestrated capillaries between the cords of cells, and drains into venules in the medulla. On the right, the adrenal vein directly enters the inferior vena cava; on the left, it usually drains into the left renal vein. The adrenal cortex is divided into three histologic and functional zones: the outer, aldosterone-secreting zona glomerulosa; the intermediate, cortisol-secreting zona fasciculata; and the inner, androgen-secreting zona reticularis. Whereas the zona glomerulosa is primarily

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

regulated by angiotensin II, both the zona fasciculata and the zona reticularis are regulated by adrenocorticotropic hormone (ACTH).

Fetal Adrenal and Development The fetal adrenal cortex arises from mesodermal cells migrating from the celomic epithelium very early in the embryonic period. It consists of both the adult adrenal zona glomerulosa and zona fasciculata and an inner large adrenal zone, which virtually disappears within weeks after birth. The active secretion of steroids occurs by week 6 from the provisional zone, which represents the functional cortex in the fetal period. Remaining cell foci from the fetal adrenal zone presumably give rise to the adrenal zona reticularis, starting at the age of 4 to 5 years in both sexes. This zone continues to grow until young adulthood (20 to 25 years), remains at a plateau for 5 to 10 years, and regresses gradually after the age of 35 years. Aging results in a reduction in the size of the zona reticularis and a relative increase in the outer cortical zones with no significant difference in the total width of the cortex.

Biosynthesis Like all human steroid hormones, AAs are derived from cholesterol, which can be synthesized within the adrenal from acetyl coenzyme A, but mainly (80%) is derived from circulating plasma lipoproteins (lowdensity lipoproteins). The major androgens secreted by the adrenals are dehydroepiandrosterone (DHEA), DHEA sulfate (DHEAS), and androstenedione (D4-A) (Fig. 1). Production of testosterone (T) by these glands is minimal. DHEA and DHEAS are mainly products of the zona reticularis; D4-A and T are secreted by both the zona reticularis and the zona fasciculata. The enzymes responsible for the synthesis are hydroxylases, dehydrogenases, isomerases, and desmolases (Fig. 1), most of which require NADPH or NADþ as

45

46

Adrenal Androgens

Steroidogenesis

Cholesterol CYP11A1

Pregnenolone 3β-HSD

Progesterone

CYP17

17α-Hydroxypregnenolone

CYP17

3β-HSD

3β-HSD CYP17

DHEA

CYP17

17α-Hydroxyprogesterone

CYP21A2

Deoxy-corticosterone (DOC)

CYP21A2

11-Deoxycortisol (S)

CYP11B1

Corticosterone (B)

∆4-Androstenedione

CYP11B1

Cortisol (F)

CYP11B2

Aldosterone

Figure 1 Adrenal androgen biosynthetic pathway. CYP11A1, cholesterolmside-chain cleavage enzyme; desmolase; CYP17, 17ahydroxylase/17,20-lyase; 3b-HSD, 3b-hydroxysteroid dehydrogenase; CYP21A2, 21-hydroxylase; CYP11B1, 11b-hydroxylase; CYP11B2, aldosterone synthase, corticosterone 18-methylcorticosterone oxidase/lyase. cofactors. The anatomical alterations of the adrenal cortex, occurring mainly in the age groups of 20–30 years and 50–60 years, result in a marked decline in circulating adrenal C19 steroids and their metabolites.

BIOCHEMISTRY The steroid hormones produced by the adrenal cortex are members of a large family of compounds derived from the cyclopentanoperhydrophenanthrene ring structure that comprises three cyclohexane rings and one cyclopentane ring. Three 19-carbon compounds are the principal androgens secreted by the adrenals: DHEA, DHEAS, and D4-A.

and T are secreted synchronously with cortisol both in secretory episodes and in a circadian pattern. The levels of plasma DHEAS do not exhibit a circadian rhythm because of its much longer circulating half-life. Numerous other endocrine signals were proposed as regulators of adrenal androgen secretion. Glasow et al. reported the presence of prolactin (PRL) receptors in the human adrenal gland and suggested a direct effect of PRL on adrenal steroidogenesis that may be of particular relevance in clinical disorders characterized by hyperprolactinemia. Interestingly, adults with

ADRENAL ANDROGEN REGULATION AND PHYSIOLOGY Regulation Adrenal androgens are secreted by the adrenal glands in response to ACTH. ACTH is a 39-amino-acid peptide derived from proopiomelanocortin, synthesized and secreted by the anterior pituitary under the regulation of corticotropin-releasing hormone (CRH) and arginine–vasopressin (AVP) (Fig. 2). Both CRH and AVP are produced by parvocellular neurons of the paraventricular nucleus of the hypothalamus and act in synergy with each other (Fig. 2). Under ACTH regulation, adrenal androgens such as DHEA, D4-A,

Figure 2 Schematic representation of adrenal androgen regulation.

47

Adrenal Androgens

hyperprolactinemia have increased secretion of AAs by the zona reticularis, which is corrected by reduction of PRL secretion with bromocriptine. In women with PRL-secreting tumors, there is a correlation between PRL levels and DHEAS. Together with ACTH and/or PRL, other factors including estrogen, epidermal growth factor, prostaglandins, angiotensin, growth hormone, gonadotropins, b-lipotropin, and b-endorphin may act as stimulators of androgen secretion. Interleukin-6 (IL-6) is also known to stimulate mineralocorticoid, glucocorticoid, and androgen production by acting through specific receptors expressed in the adrenals, mainly in the zonae fasciculata and reticularis, but also to a lesser extent in the zona glomerulosa. The ability of IL-6 to stimulate mineralocorticoid, glucocorticoid, and androgen production suggests that IL-6 might play a role in coordinating the responses of all adrenocortical zones and the interaction of the adrenal function with the immune system. Both ACTH and PRL stimulate AA secretion by the fetal adrenal zone. Placental CRH production, which rises exponentially during human pregnancy, may also play a key role in promoting DHEAS production by the fetal adrenals, leading to an increase in placental estrogen synthesis and contributing to the process of parturition in humans.

Physiology AAs are secreted in small amounts during infancy and early childhood and their secretion gradually increases with age, paralleling the growth of the zona reticularis. Adrenarche is the appearance of pubic hair (pubarche) resulting from a rise in adrenal androgen levels. The mechanisms by which the zona reticularis develops with age and by which adrenarche is regulated are not fully known. It was shown that children with premature pubarche have hormonal responses to a CRH stimulation test that are similar in magnitude to those of prepubertal children of comparable age, ruling out a prominent role for CRH in premature pubarche. Gell et al. suggested that as children mature, a decrease in 3b-hydroxysteroid dehydrogenase activity in the adrenal reticularis occurs, resulting in the shift of pregnenolone through the 17a-hydroxylase/ 17,20-lyase pathway, leading to increased production of DHEA and DHEAS, as seen during adrenarche. Locally produced insulin-like growth factor type II (IGF-II) modulates fetal adrenocortical cell function by increasing responsiveness to ACTH via activation of the IGF type I receptor and increases the capacity of those cells for androgen synthesis by directly

augmenting the expression of P450c17. Thus, IGF-II may play a pivotal role in AA production, both physiologically in utero and at adrenarche, as well as under conditions of hyperandrogenemia. Experiments by Miller and co-workers suggested that an increased serine phosphorylation of human P450c17 might play a role in the development of both human adrenarche and hyperandrogenism of polycystic ovary syndrome (PCOS), resulting in a substantial increase in 17,20-lyase activity. P450c17 is the key enzyme that regulates androgen synthesis. It is the only enzyme known to have the ability to convert C21 precursors to the androgen prehormones, the 17-ketosteroids. It is a single enzyme with two activities, 17-hydroxylase and 17,20-lyase, and serine phosphorylation appears to modulate its activity. In particular, it promotes 17,20-lyase activity and at the same time inhibits the activity of the insulin receptor. It was postulated that a single abnormal serine kinase might hyperphosphorylate both P450c17 and the insulin receptor, accounting for the hyperandrogenism and the hyperinsulinism responsible for both the premature pubarche and later in life for PCOS. In vitro studies, however, failed to find evidence for hyperphosphorylation of insulin receptor-b and P450c17 in PCOS.

CIRCULATION Circulating steroid hormones are largely bound to plasma proteins (binding globulins and albumin). Approximately 90% of DHEA, DHEAS, and D4-A is bound to albumin and 3% is bound to sex hormone-binding globulin. The binding globulins have high affinity and low capacity, whereas albumin has a low affinity and high capacity for steroids.

ANDROGEN RECEPTOR The inactive androgen precursors secreted by the adrenal, after conversion to Tand 5a-dihydrotestosterone (DHT), exert their effects in most peripheral tissues by interacting with high-affinity receptor proteins. The androgen receptor (AR), encoded by the AR gene on the X chromosome, is a member of the steroid receptor superfamily. This gene contains a polymorphic CAG microsatellite repeat within exon 1, which codes for a variable length of polyglutamine chain at the amino terminus, the transactivation domain of the AR protein. Triplet-repeat DNA sequences can be sites of genetic instability and their expansion in a variety of genes has been associated with human genetic diseases, such as

48

p0080

fragile X syndrome and myotonic dystrophy. In the case of the AR gene, an inverse correlation between the number of CAG repeats and the risk for prostate cancer was described and its expansion was documented in Kennedy’s disease (spinal and bulbar muscular atrophy), a disorder associated with primary hypogonadism. In vitro studies showed that progressive expansion of the repeat length in the AR is associated with a linear decrease in its transactivation function. These observations support the idea that there is an optimal number of repeats that varies in the population from 11 to 31 (average size: 21  2 repeats). Methylation of deoxycytosine residues is another factor involved in the modulation of gene expression. Belmont et al. showed that the methylation of HpaII and HhaI sites near the polymorphic CAG repeats in the first exon of the human AR (HUMARA) locus correlated with X inactivation. Patients with idiopathic hirsutism were shown to have a normal number of CAG repeats but with a preponderance of the shortest and most active alleles. These patients also had a preferential methylation of the longer AR allele compared to normal subjects, leading to inactivation of the functionally weaker gene. This skewing could allow the shorter, more active AR allele to be preferentially expressed, explaining the peripheral hypersensitivity to androgens in hirsute patients. Multiple coactivators that enhance the transcription of the AR gene have been identified, including AP-1, Smad3, nuclear factor kB, sex-determining region Y, and the Ets family of transcription factors. The relative importance of these coactivators for any particular cell type remains unclear, since a putative coregulator that can alter transcriptional activity is typically included in transient transfection experiments. Although AR is normally thought to function as a homodimer, it was also shown to heterodimerize with other nuclear receptors including the estrogen receptor, glucocorticoid receptor, and testicular orphan receptor 4. One of the major mechanisms through which coregulators might function is by forming a bridge between the DNA-bound nuclear receptor and the basal transcriptional machinery (type I regulators). Coactivators may also facilitate ligand binding, promote receptor nuclear translocation, or mediate signal transduction (type II coregulators). The role of corepressors in AR function is poorly defined. Three corepressors of androgen-bound AR have been identified thus far: cyclin D1, calreticulin, and HBO1 [histone acetyltransferase binding to ORC (origin recognition complex)]. However, relatively little is known about the mechanism(s) of their repressive effects.

Adrenal Androgens

PERIPHERAL CONVERSION AND METABOLISM DHEA, DHEAS, and D4-A are converted to the potent androgens T and DHT in peripheral tissues. Major conversions are those of D4-A to T and of T to DHT as carried out by the enzymes 17-hydroxysteroid dehydrogenase (17b-HSD) and 5a-reductase, respectively. Major peripheral sites of androgen conversion are the hair follicles, the sebaceous glands, the prostate, and the external genitalia. The active uptake of androgens and in situ estrogen synthesis occur in peripheral adipose tissue and are carried out by the enzymes 17b-HSD and aromatase, respectively. Peripheral conversion contributes significantly to circulating T levels in women, but not in men, in whom T is largely produced by the testis. The AAs and their metabolites are inactivated or degraded in various tissues, including the liver and kidneys. Major biochemical routes for inactivation and excretion are conjugation of androgens to glucuronate or sulfate residues to produce hydrophilic glucuronides or sulfates that are excreted in the urine.

BIOLOGICAL EFFECTS In adult men, the conversion of adrenal D4-A to testosterone accounts for less than 5% of the production rate of the latter; thus, its role in the physiological androgenization of the male is negligible. Excessive AA secretion appears to have no major clinical consequences in the adult man, although this may be a matter of debate. AA hypersecretion in prepubertal boys, on the other hand, has clearly been associated with isosexual precocious puberty. In adult women, adrenal D4-A and D4-A generated from the peripheral conversion of DHEA contribute substantially to total androgen production and its effect. In the follicular phase of the menstrual cycle, adrenal precursors account for two-thirds of testosterone production and one-half of dihydrotestosterone production. At midcycle, the ovarian contribution increases and the adrenal precursors account for 40% of testosterone production. In women, increased AA production may be manifested as cystic acne, hirsutism, male type baldness, menstrual irregularities, oligo-ovulation or anovulation, infertility, and/or frank virilization. Excessive adrenal androgen secretion in prepubertal or pubertal girls can cause heterosexual precocious puberty.

Adrenal Androgens

See Also the Following Articles Adrenal Cortex, Anatomy . Adrenal Cortex, Physiology . Adrenal Insufficiency . Adrenarche, Premature . Androgen Biosynthesis and Gene Defects . Androgen Biosynthesis Inhibitors and Androgen Receptor Antagonists . Androgen Insensitivity Syndrome . Androgens, Gender and Brain Differentiation

Further Reading Auchus, R. J. (1998). The regulation of human P450c17 activity: Relationship to premature adrenarche and the polycystic ovary syndrome. Trends Endocrinol. Metab. 9, 47–50. Ghizzoni, L., Virdis, R., Ziveri, M., Lamborghini, A., Alberini, A., Volta, C., and Bernasconi, S. (1989). Adrenal steroid, cortisol, adrenocorticotropin, and b-endorphin responses to human corticotropin-releasing hormone stimulation test in normal children and children with premature pubarche. J. Clin. Endocrinol. Metab. 69, 875–880. Heinlein, C. A., and Chang, C. (2002). Androgen receptor (AR) coregulators: An overview. Endocr. Rev. 23, 175–200. Longcope, C. (1986). Adrenal and gonadal androgen secretion in normal females. In ‘‘Clinics in Endocrinology and Metabolism’’

49 (R. Horton and R. A. Lobo, eds.), pp. 213–228. W. B. Saunders, Philadelphia, PA. Mastorakos, G., Weber, J. S., Magiakou, M. A., Gunn, H., and Chrousos, G. P. (1994). Hypothalamic–pituitary– adrenal axis activation and stimulation of systemic vasopressin secretion by recombinant interleukin-6 in humans: Potential implications for the syndrome of inappropriate vasopressin secretion. J. Clin. Endocrinol. Metab. 79, 934–939. McKenna, T. J., Fearon, U., Clarke, D., and Cunningham, S. K. (1997). A critical review of the origin and control of adrenal androgens. Bailliere Clin. Obstet. Gynecol. 11, 229–248. Nicholson, W. E., Liddle, R. A., Puett, D., and Liddle, G. W. (1978). Adrenocorticotropic hormone biotransformation, clearance, and catabolism. Endocrinology 103, 1344–1351. Parker, L. N. (1991). Control of adrenal androgen secretion. Endocrinol. Metab. Clin. North Am. 20, 401–421. Vale, W., Spiess, J., Rivier, C., and Rivier, J. (1981). Characterization of a 41-residue ovine hypothalamic peptide that stimulates secretion of corticotropin and b-endorphin. Science 213, 1394–1397. Vottero, A., Stratakis, C. A., Ghizzoni, L., Longui, C. A., Karl, M., and Chrousos, G. P. (1999). Androgen receptor-mediated hypersensitivity to androgens in women with nonhyperandrogenic hirsutism: Skewing of X-chromosome inactivation. J. Clin. Endocrinol. Metab. 84, 1091–1095.

Adrenal Cortex Development, Regulation of Michel Grino Universite´ de la Me´diterrane´e, Marseille, France

Glossary adrenocorticotropic hormone (ACTH) A 39amino acid peptide secreted by the anterior pituitary that acts primarily on the adrenal cortex, stimulating its growth and its secretion of corticosteroids. corticotropin-releasing hormone (CRH) A 41amino acid peptide secreted by the hypothalamus and the placenta; hypothalamic CRH acts primarily on the anterior pituitary to stimulate adrenocorticotropic hormone (ACTH) synthesis and secretion, whereas placental CRH acts on the fetal adrenals to regulate corticosteroid secretion and on the reproductive tract to regulate the processes of parturition. estrogens Female sex hormones that are responsible for the development of the female secondary sex characteristics, the regulation of the menstrual cycle, the production of an environment suitable for the implantation of the early embryo, the maintenance of pregnancy, and the development of fetal and maternal tissues. growth factors Secreted proteins that exert diverse effects on cell growth, metabolism, and differentiation.

p0005

S

teroids synthesized by the adrenal cortex, in particular glucocorticoids, play an important role in the maintenance of pregnancy and the maturation of maternal and fetal tissues during fetal life and in the adaptation to extrauterine life after birth. An adequate exposure to glucocorticoids relies on coordinate regulation of adrenal growth, maturation, and steroid biosynthesis. Anterior pituitary adrenocorticotropic hormone (ACTH) is believed to be the major trophic regulator of adrenal development and glucocorticoid synthesis and secretion. However, because ACTH is not a mitogen per se, it is now generally accepted that some of the trophic actions of ACTH on the fetal adrenal cortex are mediated indirectly via tissue growth factors. In addition, other hormonal factors originating from the placenta, such as corticotropin-releasing hormone (CRH) and estrogens, are important for fetal adrenal maturation.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

HORMONAL REGULATION OF ADRENAL GROWTH AND STEROID BIOSYNTHESIS Adrenocorticotropic Hormone Adrenocorticotropic hormone (ACTH) is the main stimulator of glucocorticoid synthesis and secretion in the adult. Exposure to robust ACTH secretion is required for normal adrenal growth and maturation. Indeed, disruption of the hypothalamo–pituitary– adrenal (HPA) function in the human fetus, experimental anencephaly in the fetal rhesus monkey or the fetal rat, or fetal hypophysectomy in sheep inhibits growth of the fetal adrenal cortex. ACTH acts through a specific adrenal cortical cell surface G protein-coupled receptor that activates adenylate cyclase, leading to an increase in intracellular cyclic adenosine monophosphate (cAMP) that in turn activates protein kinase A and initiates the cascade of intracellular signaling events. In midgestation human fetal adrenals, ACTH receptor mRNA is localized in cells from all cortical zones; abundance is higher in the definitive zone than in the fetal zone. The effect of ACTH on adrenal growth and/or metabolism is zone specific. In anencephalic fetuses, the definitive zone appears to be normal. In fetal monkeys, blockade of endogenous fetal ACTH secretion by glucocorticoid treatment decreases expression of 3b-hydroxysteroid dehydrogenase (3b-HSD) and eliminates the transitional zone but has no effect on the size of the definitive zone. Conversely, ACTH administration or stimulation of endogenous ACTH secretion by metyrapone treatment results in stimulation of 3bHSD expression and in an increase of the width of the transitional zone but not of the size of the definitive zone. The growth and the steroidogenic activity of the fetal zone are also controlled by ACTH. In fetuses with congenital adrenal hyperplasia, the fetal zone is hypertrophied and adrenal androgen concentrations are stimulated dramatically. In monkeys, inhibition of

61

p0015

62

p0030

endogenous ACTH decreases adrenal 17a-hydrolase/ 17–20 lyase (P450c17) expression, whereas metyrapone-induced stimulation of ACTH secretion results in stimulation of P450c17 gene transcription in the fetal zone. The stimulatory effect of ACTH on adrenal growth and steroid biosynthetic capacities seems to depend both on the circulating levels of the hormone and on the sensitivity of the adrenal gland. In addition, these phenomena could be associated given that it is established that ACTH up-regulates the expression of its own receptor gene in both adult and fetal adrenocortical cells. In human fetuses, the maximal rate of adrenal growth occurs during the time the plasma ACTH concentration is the lowest. In baboons, there is a temporal parallelism between ACTH receptor mRNA expression in the fetal zone and ACTH-stimulable dehydroepiandrosterone (DHEA) formation. It has been suggested that the increase in 3b-HSD expression and cortisol production observed during late gestation results from the ACTH receptor-mediated development and enhanced functional capacity of the transitional/definitive zone. In sheep, between embryonic day 90 (E90) and E120, adrenal growth and adrenal cell division occurs at the lowest rate compared with any other time during gestation, and there is a decreased expression of steroidogenic enzymes. Between E130 and postnatal day 2 (P2), there is a phase of rapid adrenal growth and functional maturation. Circulating ACTH increases steadily between E110 and E145, whereas plasma cortisol concentrations are low between E110 and E130 and increase abruptly between E130 and E145. Interestingly, the lowest level of adrenal expression of the ACTH receptor occurs at E90, whereas there is increased ACTH receptor mRNA, ACTH binding, and ACTH-induced adenylate cyclase activity after E125. The ACTH sensitivity of fetal adrenal cells incubated in vitro parallels the evolution of the density of ACTH binding sites. Ovine fetuses that have undergone hypothalamo–pituitary disconnection at E115 show decreased ACTH and cortisol secretion at E126 but normal ACTH and decreased cortisol secretion at E145. Hypophysectomy in ovine fetuses induces a decrease in adrenal expression of the side-chain cleavage enzyme (P450ssc), 3b-HSD, and P450c17 mRNAs that is reversed by ACTH1–24 infusion. Other pro-opiomelanocortin (POMC)-derived peptides can participate in adrenal maturation because, in fetal sheep, infusion of N-POMC 1–77 stimulates adrenal growth and 21-hydrolase (P450c21) mRNA expression.

Adrenal Cortex Development, Regulation of

In fetal rats, plasma corticosterone levels rise progressively from E16 to E19. The pattern of plasma ACTH levels parallels that of corticosterone, indicating that the observed adrenal hyperactivity during late gestation is driven by increased ACTH secretion from the corticotropes. In newborn rats, circulating ACTH levels are low during the first 10 days after birth, the so-called stress hyporesponsive period, and the density of adrenal ACTH-binding sites and basal and ACTH-stimulated adenylate cyclase activity are decreased in the adrenals of P7 rats. Chronic administration of ACTH or POMC-derived peptides (e.g., Lys-g3-melanocyte-stimulating hormone) during this period has a trophic effect on the adrenal and potentiates the subsequent corticosterone response to stress or to ACTH injection.

Angiotensin II In adults, angiotensin II, acting through membrane G protein-coupled receptors, is one of the most important factors involved in the regulation of aldosterone biosynthesis and secretion. Two subtypes of angiotensin II receptors have been described: AT1 and AT2. AT1 and AT2 receptors are present on human gestational week 16 (W16) to W18 fetal adrenocortical cells. AT1 receptors, which mediate most of the known actions of angiotensin II, are located in the definitive zone, whereas AT2 receptors are present throughout the gland, with a predominant labeling in the fetal zone. It has been proposed that AT2 receptors are involved in the apoptotic process observed in the human fetal adrenal gland and could participate, after birth, in the involution of the fetal zone. In sheep, isolated adrenal cells obtained from E40 to E90 fetuses secrete aldosterone and respond to relatively high doses of angiotensin II. At E100 to E130, adrenal glands become unresponsive to angiotensin II. In vivo, during late gestation, fetuses are less responsive to infused angiotensin II, in terms of aldosterone secretion, than are adult sheep. In addition, in vitro, angiotensin II inhibits ACTH-induced cortisol secretion and P450c17 expression. This phenomenon may be mediated via AT1 receptors. The mRNA coding for the AT1 receptor is first detected in the unzoned gland as early as E40 and is present at high levels in the zona glomerulosa and, to a lesser extent, in the zona fasciculata at E60 to E105. During late gestation (E120– E135), the AT1 receptor hybridization signal decreases before showing a further increase to reach adult values by P2. The fall in AT1 receptor expression during late gestation could allow ACTH to override angiotensin

p0035

Adrenal Cortex Development, Regulation of

63

II-induced inhibition of cortisol secretion and P450c17 expression. The mRNA for the AT2 receptor is present in the same location from E40 to E130 and declines to extremely low levels after E140. In the rat adrenal zona glomerulosa, angiotensin II receptor content decreases from very high levels at birth to adult levels by P20. This decrease is due to a decline in AT2 receptors. In P7 rats, aldosterone secretion is more sensitive to the stimulatory effect of ACTH than of angiotensin II.

Placental Factors p0060

p0065

Functional interactions exist between the placenta and the fetal adrenals. The placenta synthesizes corticotropin-releasing hormone (CRH), the major hypothalamic neuropeptide that stimulates anterior pituitary POMC synthesis and ACTH secretion. Estrogens, which originate mainly from the placental conversion of fetal DHEA sulfate (DHEA-S), have a direct effect on fetal adrenal activity and regulate placental metabolism of maternal glucocorticoids (Fig. 1). Corticotropin-Releasing Hormone In primates, CRH and POMC synthesized in the placenta are released in both the maternal and fetal circulations. Plasma CRH increases exponentially during gestation, peaking at labor. The role of placental CRH in the regulation of fetal adrenal function is not clear. A circulating binding protein is present in the human fetal circulation and is capable of inactivating a large amount of CRH. CRH may modulate fetal adrenal function indirectly by stimulating placental or fetal anterior pituitary ACTH and POMC-derived peptide release, or it may do so directly by regulating the fetal adrenal cortex. Indeed, human midgestation fetal adrenals express CRH receptor mRNA. CRH stimulates DHEA-S and cortisol production, stimulates P450ssc and P450c17 expression, and increases ACTH responsiveness in cultured human fetal adrenocortical cells. Estrogens Estrogens play an important role in regulating, either directly or indirectly, cortisol and DHEA synthesis and secretion. The primate fetal adrenal begins to synthesize cortisol de novo from cholesterol between mid- and late gestation. As a consequence, at midgestation, most if not all of the glucocorticoids in the fetal circulation originate from the mother. Maternal glucocorticoids are metabolized in the placenta, which

Figure 1 Hormonal regulation of fetal adrenal growth and steroid biosynthesis. It is proposed that fetal anterior pituitary adrenocorticotropic hormone (ACTH), either directly or mediated by growth factors, and pro-opiomelanocortin (POMC)-derived peptides have a stimulatory effect on fetal adrenal growth and steroid biosynthesis. Placental estrogens, originating from local conversion of the fetal adrenal androgen precursor dehydroepiandrosterone (DHEA), can regulate fetal adrenal metabolism directly or can act through a modulation of placental 11b-hydroxysteroid dehydrogenase (11b-HSD) activity. Between early and midgestation, 11b-HSD shows preferential reduction of cortisone to cortisol, leading to decreased pituitary ACTH secretion and subsequent reduced adrenal maturation and cortisol biosynthesis. With advancing gestation, the increase in estrogen production enhances placental 11b-HSD oxidation of cortisol to cortisone, leading to decreased cortisol feedback at the fetal pituitary level and acceleration of adrenal maturation and stimulation of cortisol and DHEA secretion, thereby creating a positive feedback loop. Placental corticotropin-releasing hormone (CRH) may modulate fetal adrenal function indirectly by stimulating fetal anterior pituitary ACTH and POMC-derived peptide release or directly by regulating fetal adrenal cortex. IGF-II, insulin-like growth factor-II; EGF, epidermal growth factor; VEGF, vascular endothelial growth factor; FGF-2, fibroblast growth factor-2; TGF-bs, members of the transforming growth factor-b family; DHEA-S, dihydroepiandrosterone sulfate; E2, estradiol.

synthesizes the two isoforms of 11-b-hydroxysteroid dehydrogenase (11b-HSD). 11b-HSD-1, which is expressed in placental intermediate trophoblast cells and in the vascular endothelium, has both oxidase (active cortisol-to-inactive cortisone) and reductase (cortisone-to-cortisol) activities. 11b-HSD-2, which is present in placental syncytiotrophoblast cells, exhibits only oxidase activity. Because 11b-HSD activity

f0005

64

t0005

Adrenal Cortex Development, Regulation of

Table I Main Characteristics of the Major Growth Factors Involved in Fetal Adrenal Growth and Steroidogenesis Growth factor

Site of synthesis

Receptors location

Role in growth

Role in steroidogenesis

Regulated by

IGF-II

C, DZ, TZ, FZ

DZ, TZ, FZ

Stimulates FZ proliferation

Stimulates P450ssc and P450c17 activity

ACTH

EGF

Mediated by TGF-a

DZ, FZ

Stimulates FZ and DZ proliferation; synergistic with IGF-II

Stimulates 3b-HSD acivity in DZ and TZ; stimulates cortisol and aldosterone secretion

ND

VEGF

Mainly FZ

ND

Angiogenic factor

ND

ACTH forskolin

FGF-2

Cortex

ND

ND

ACTH cAMP

TGF-b1

ND

Neocortex

Inhibits basal and EGF-stimulated growth of FZ and TZ

Inhibits P450ssc and P450c17 activity; inhibits cortisol and DHEA secretion

ND

Activin

DZ, TZ

DZ, FZ

Inhibits basal and EGF-stimulated FZ proliferation

Inhibits cortisol secretion

ACTH

Stimulates FZ and DZ proliferation; synergistic with IGF-II

Note. IGF-II, insulin-like growth factor-II; EGF, epidermal growth factor; VEGF, vascular endothelial growth factor; FGF-2, fibroblast growth factor-2; TGF-b1, transforming growth factor-b1; C, capsule; DZ, definitive zone; TZ, transitional zone; FZ, fetal zone; P450ssc, side-chain cleavage enzyme; P450c17, enzyme complex having 17a-hydrolase and 17,20-lyase activities; 3b-HSD, 3b-hydroxysteroid dehydrogenase; DHEA, dehydroepiandrosterone; ACTH, adrenocorticotropie hormone; cAMP, cyclic adenosine monophosphate; ND, not determined.

shows preferential reduction of cortisone to cortisol between early and midgestation, increased circulating fetal cortisol leads to decreased pituitary ACTH synthesis and secretion and to subsequent reduced adrenal maturation and cortisol biosynthesis capacity. With advancing gestation, the increase in estrogen production enhances placental 11b-HSD oxidation of cortisol to cortisone, leading to decreased cortisol feedback at the fetal pituitary level. The subsequent increase in anterior pituitary ACTH synthesis and secretion induces acceleration of adrenal maturation and stimulation of cortisol secretion by the definitive zone and of DHEA production by the fetal zone, creating a positive feedback loop. In addition to their effects on placental glucocorticoid metabolism, estrogens can act directly at the adrenal level. Indeed, the primate fetal adrenal contains estrogen receptors b. Also, in vivo, estrogen treatment increases the responsivity of the fetal adrenal gland to ACTH, presumably through stimulation of protein kinase A activity. In sheep, during late gestation, estrogens also have a stimulatory effect on fetal ACTH and cortisol secretion. However, unlike in primates, the stimulatory effect of estrogens on the HPA axis takes place at the central nervous system level, that is, on the hypothalamic ACTH secretagogues arginine vasopressin and CRH.

ROLE OF GROWTH FACTORS As mentioned previously, ACTH is the primary regulator of adrenal growth and steroid biosynthesis. However, because ACTH is not a mitogen per se, it is now generally accepted that some of the trophic actions of ACTH on the fetal adrenal cortex are mediated indirectly via tissue growth factors such as insulinlike growth factors (IGF-I and IGF-II), epidermal growth factor (EGF), vascular endothelial growth factor (VEGF), fibroblast growth factor-2 (FGF-2), members of the transforming growth factor-b (TGF-b) family, and adrenomedullin (Table I).

p0080

Insulin-like Growth Factors IGF-I and IGF-II are mitogenic peptides, structurally related to proinsulin, that affect growth and function in a wide variety of cell types and can act as autocrine, paracrine, or endocrine factors. The biological actions of IGF-I and IGF-II are modulated by insulin-like growth factor-binding proteins (IGFBPs). IGF-I mediates many of the somatotropic actions of growth hormone, whereas IGF-II is important in the regulation of fetal development. The effect of IGFs on adrenal cortical cells is most likely mediated through

p0085

65

Adrenal Cortex Development, Regulation of

p0090

p0095

IGF type 1 receptors because this subtype has been identified in both adult human and fetal nonhuman primate adrenal glands and because IGF type 2 receptor does not bind IGF-II. Circulating concentrations of IGF-II are high during fetal life in several species (e.g., primates, sheep, rodents) and decrease after delivery. IGF-I and IGF-II mRNAs are present in adrenals from midgestation human fetuses. IGF-I mRNA is detected only in the capsule and not in the cortical zones, whereas IGF-II mRNA is detected in the definitive and fetal zones as well as in the capsule. The IGF-II mRNA is present in high abundance in human fetal adrenals and is barely detectable in adult adrenals, whereas the mRNA encoding IGF-I is expressed at low levels in fetus adrenals and at high levels in adult adrenal glands. In adrenals obtained from nonhuman primates, IGF-II mRNA is abundant from early (E60) to late (E165) gestation and is localized in the definitive, transitional, and fetal zones. The mRNAs coding for IGFBP-2 and IGFBP-6, which have been shown to bind preferentially IGF-II, are expressed in adrenals obtained from monkey fetuses in the definitive, transitional, and fetal zones. In the developing ovine adrenal, IGF-II mRNA is highest in E60 fetuses, decreases slightly between E60 and E100, remains relatively constant until term, and decreases significantly after birth, with IGF-I mRNA being expressed at very low levels. At all gestational ages, IGF-II mRNA and protein are localized in the capsule and mesenchymal cells surrounding the gland and in the steroidogenic cells of the zona glomerulosa and zona fasciculata. IGFs have profound effects on adrenal growth. Transgenic mice that are IGF-I, IGF-II, or IGF type 1 receptor-deficient have intrauterine growth retardation. However, IGF-II-deficient mice have nearnormal postnatal growth due to IGF-I. In ovine fetuses, chronic infusion of either IGF-I or IGF-II between E120 and E130 results in an increase in adrenal growth. In vitro, IGF-I and/or IGF-II stimulate proliferation of fetal adrenocortical cells obtained from sheep or human fetuses. As mentioned previously, IGF-II may mediate the trophic actions of ACTH on the fetal adrenal gland. Treatment of rhesus monkey fetuses with metyrapone, which likely increases pituitary ACTH secretion, induces in the adrenals hypertrophy of all cortical zones together with an increase in the concentrations of IGF-II and IGF type 1 receptor mRNAs. In vitro, in human fetal adrenal cells in culture, ACTH increases IGF-II mRNA levels. In addition to its effects on adrenal growth,

IGF-II regulates adrenal steroidogenesis. It has been demonstrated, using primary cultures of adrenal cortical cells obtained from human midgestation fetuses, that IGF-I and IGF-II stimulate basal and ACTH-, forskolin-, or cAMP-induced cortisol and DHEA production. Under the same experimental conditions, IGF-II increases adrenocortical responsiveness to ACTH without any effect on ACTH receptor mRNA, suggesting that IGF-II modulates ACTH sensitivity in fetuses by increasing ACTH signal transduction at some point distal to the ACTH receptor. In addition, IGF-II increases ACTH-stimulated abundance of P450ssc and P450c17 mRNAs, thereby augmenting the potential for adrenal androgen synthesis.

Epidermal Growth Factor Human cord blood EGF levels have been shown to increase with progressive gestation, suggesting a functional role for EGF during the perinatal period. Knockout mice for the EGF receptor have intrauterine growth retardation. However, EGF and EGF precursor mRNA are expressed as low levels, and tissue EGF immunoreactivity appears late in rodent fetuses. Therefore, it has been proposed that TGF-a, a member of the EGF family, is the ligand for the fetal EGF receptor. TGF-b has been identified in steroidogenic cells of adult adrenal cortex. EGF receptor concentrations increase steadily from E15 until birth in several mouse tissues. EGF receptors have been identified in both the definitive and fetal zones from midgestation human adrenal fetuses. EGF has been shown to stimulate the proliferation of human fetal adrenal cells in vitro and to act cooperatively with IGF-I and IGF-II. Late-gestation EGF-treated monkey fetuses show an increase in adrenal weight due to a hypertrophy of the definitive zone and stimulation of 3b-HSD immunoreactivity in the definitive and transitional zones. In EGF-infused fetal sheep, adrenal cortical hypertrophy is accompanied by increased cortisol and aldosterone secretion. EGF may act directly at the adrenal level or indirectly through stimulation of the hypothalamo–pituitary axis because EGF increases the secretion of CRH from the hypothalamus and of ACTH from the anterior pituitary. It has been demonstrated that, in fetal monkeys, EGF acts on the hypothalamic–pituitary axis to modulate adrenal cortical growth and functional maturation of the transitional zone, whereas EGF can act independently of the hypothalamic–pituitary axis to stimulate functional maturation of the definitive zone.

p0100

66

Vascular Endothelial Growth Factor VEGFs are a family of direct-acting endothelial cell mitogens and angiogenic factors that derive from a single gene by alternative splicing. The mRNAs coding for the four isoforms of VEGF and for the two VEGF receptors have been identified in the adult mouse adrenal gland. VEGF mRNA and protein have been detected in the adrenal cortex from midgestation human fetuses, predominantly in the fetal zone. The predominant staining for VEGF in the fetal zone correlates with the extensive vasculature of this zone. In primary cultures of human fetal adrenal cortical cells, ACTH and forskolin increase both VEGF mRNA levels and VEGF protein secretion.

Fibroblast Growth Factor-2 p0110

FGF-2 (also called basic FGF) is a potent angiogenic molecule that belongs to the family of the FGF and interacts with four cell surface receptor subtypes. FGF-2 is synthesized by human W16 fetal adrenal cortex and is stimulated by ACTH and cAMP. FGF-2 mRNA increases steadily during gestation in adrenal glands obtained from nonhuman primate fetuses. In vitro, FGF-2 stimulates the proliferation of human fetal adrenal cells from the definitive or fetal zone, with its effect being additive to those of IGF-I and IGF-II.

Adrenal Cortex Development, Regulation of

basal P450ssc expression and on ACTH-induced cortisol secretion and P450c17 activity. The interplay between TGF-b1 and ACTH in regulating adrenal growth and steroid production does not involve mutual control of their respective receptors because TGF-b1 does not influence ACTH receptor gene expression and because ACTH increases TGF-b1 binding in the fetal adrenal cortex.

Inhibins Inhibins are dimers of an a-subunit and either a bAor a bB-subunit, whereas activins are homo- or heterodimers of either a bA- or a bB-subunit. a-, bA-, or bB-subunit immunoreactivity and mRNAs, as well as activin type I/II receptor and inhibin receptor mRNAs, have been detected in the definitive and fetal zone of adrenals from midgestation human fetuses. In vitro, activin A inhibits basal and EGFstimulated fetal zone proliferation, possibly through enhanced apoptosis, whereas inhibin A has no apparent mitogenic effect. The inhibitory effect of activin A on fetal adrenal cell proliferation is additive to that of TGF-b. In addition, ACTH has a stimulatory effect on a and bA subunit mRNA levels and inhibin A /B secretion from primary cultures of human fetal adrenal cells.

Transforming Growth Factor-b

Adrenomedullin

TGF-bs are a family of potent multifunctional cytokines that modulate a wide variety of cellular activities. Members of the superfamily include TGFb-1, activin, and inhibin.

Adrenomedullin is a multifunctional peptide, initially purified from an adrenal tumor of the medulla, that shows structural homology with calcitonin generelated peptide (CGRP). In mice, adrenomedullin mRNA and protein are present in the adrenal primordia as early as E12. The observations that antagonism of adrenomedullin function during rat pregnancy causes fetal growth restriction, and that proadrenomedullin N-terminal 20 peptide enhances proliferation of adult rat zona glomerulosa cells by acting through CGRP1 receptors, suggest that proadrenomedullin-derived peptides may be important in regulating adrenal growth during development.

TGF-b1 TGF-b1 immunoreactivity has been detected in the adrenal cortex from adult and neonatal mice. However, expression of TGF-b1 in human adrenals during development remains to be determined. TGF-b1binding sites have been identified in fetal human cortical cells. Several reports indicate that TGF-b1 inhibits basal and EGF-stimulated growth of human fetal cortical cells, in both the definitive and fetal zones, possibly through a stimulation of apoptosis. This inhibitory effect is significantly blunted by ACTH. TGF-b1 is also a potent inhibitor of adrenal steroidogenesis. Incubation of primary cultures of human fetal cortical cells with TGF-b1 results in a decrease of both basal and ACTH-induced DHEA secretion and P450c17 gene expression. In fetal sheep, adrenal TGF-b1 has an inhibitory effect on

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . ACTH, a-MSH, and POMC, Evolution of . Adrenal Cortex, Anatomy . Adrenal Cortex, Development . Adrenal Insufficiency . Angiotensin, Evolution of . EGF and Related Growth Factors . Fibroblast Growth Factor (FGF) . Glucocorticoids, Overview . Insulin-like Growth Factors

Adrenal Cortex Development, Regulation of

Further Reading Challis, J. R. G., Sloboda, D., Matthews, S. G., Holloway, A., Alfaidy, N., Patel, F. A., Whittle, W., Fraser, M., Moss, T. J. M., and Newnham, J. (2001). The fetal placental hypothalamic– pituitary–adrenal (HPA) axis, parturition, and postnatal health. Mol. Cell. Endocrinol. 185, 135–144. Coulter, C. L., Ross, J. T., Owens, J. A., Bennett, H. P. J., and McMillen, I. C. (2002). Role of pituitary POMCpeptides and insulin-like growth factor II in the developmental biology of the adrenal gland. Arch. Physiol. Biochem. 110, 99–105.

67 Mesiano, S., and Jaffe, R. B. (1997a). Developmental and functional biology of the primate fetal adrenal cortex. Endocrine Rev. 18, 378–403. Mesiano, S., and Jaffe, R. B. (1997b). Role of growth factors in the developmental regulation of the human fetal adrenal cortex. Steroids 62, 62–72. Naaman Re´pe´rant, E., and Durand, P. (1997). The development of the ovine fetal adrenal gland and its regulation. Reprod. Nutr. Dev. 37, 81–95. Pepe, G. J., and Albrecht, E. D. (1995). Actions of placental and fetal adrenal steroid hormones in primate pregnancy. Endocrine Rev. 16, 608–648.

Adrenal Cortex, Anatomy Ilias Vrezas, Holger S. Willenberg, and Stefan R. Bornstein University of Du¨sseldorf, Du¨sseldorf, Germany

Glossary adrenal cortex The outer portion of the adrenal gland. It produces glucocorticoid and mineralocorticoid hormones and adrenal androgens. adrenal gland A pair of small glands, each of which is located on top of one of the kidneys. cortisol The major natural glucocorticoid in humans and the primary stress hormone.

T

he adrenal gland was first described by Bartholomeus Eustachius in 1563. However, it was Thomas Addison in 1855 who first recognized the importance of the adrenal glands, and in 1856 Charles Edward Brown-Se´quard showed by bilateral adrenalectomy in experimental animals that the function of these glands was necessary for life. Using histochemical techniques developed in the mid-19th century, it was demonstrated that the adrenal medulla and the adrenal cortex have divergent cellular and functional properties.

HISTOLOGY Whereas the fetal cortex mainly consists of the zona fetalis, the adult adrenal cortex consists of at least three anatomically distinct zones: the outer zona glomerulosa, which is the site of mineralocorticoid production (e.g., aldosterone); the central zona fasciculata, which is predominantly responsible for glucocorticoid production; and the inner zona reticularis, where adrenal androgens [predominantly dehydroepiandrostenedione (DHEA), DHEA sulfate, and androstenedione] are located and some glucocorticoid synthesis (cortisol and corticosterone) occurs (Fig. 1). In rats, a fourth zone, the zona intermedia, can be discerned that is believed to contain adrenocorticoid stem cells and to be the region in which adrenocyte differentiation begins. The adrenal cortex synthesizes exclusively steroid hormones, which are derived from various

50

modifications of the precursor cholesterol. Glucocorticoids are secreted during the course of stress regulation and act mainly on intermediary metabolism and the immune system. Mineralocorticoids exert their main action on salt and water homeostasis. Adrenal androgens show a testosterone-like effect but can also be precursors for aromatization to estrogens. The human adrenal cortex is able to synthesize more than 50 steroids, but not all of them are secreted into the blood circulation or are biologically active. In addition, adrenocortical cells can secrete active peptides, cytokines, and other hormones. At the ultrastructural level, cells of the zona glomerulosa are rounded and smaller than the polyhedral cells of the zona fasciculata, which gradually extend into the zona reticularis. The latter zone consists of cells that appear identical to those of the zona fasciculata and also those of another type of smaller cells with a dark-staining nucleus. The cells of the adrenal cortex are arranged in a cord-like manner, extending from the adrenal capsule to the medulla and are surrounded by a capillary network. In adrenocortical cells, the mitochondria are particularly numerous, and the smooth endoplasmic reticulum is especially

Figure 1 Cross-section of a human adrenal gland double immunostained for 17a-hydroxylase and chromogranin A. zM, adrenal medulla; zR, zona reticularis; zF, zona fasciculata; zG, zona glomerulosa; C, adrenal capsule.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

51

Adrenal Cortex, Anatomy

Figure 3 Adrenocortical endothelial cells (arrow) immunostained for CD31.

Figure 2 Adrenal gland double immunostained for 17a-hydroxylase (zona reticularis and zona fasciculata, gray) and for tyrosine hydroxylase (adrenal medulla, black). Black islets of medullary cells are located in the gray adrenal cortex, and gray islets of adrenocortical cells are located in the black medulla.

abundant and forms a network of anastomosing tubules. The classic view of a strict separation between the steroid-producing adrenal cortex and the catecholamine-producing medulla has been shown to be an oversimplification; displaced chromaffin cells have been found in all zones of the adult adrenal cortex and, similarly, cortical cells are found in the medulla (Fig. 2). The close anatomic colocalization of the cortical and medullary cells has been suggested to be a prerequisite for paracrine interactions.

central adrenal vein it is drained directly or indirectly (via the renal vein) into the inferior vena cava (Fig. 3).

INNERVATION The adrenal cortex receives extensive afferent innervation, with evidence of direct contact between nerve terminals and cortical cells. A possible efferent innervation has been reported, with the presence of baroreceptors and chemoreceptors in the adrenal cortex. Adrenal innervation influences compensatory adrenal hypertrophy and has been implicated in the regulation of the diurnal variation of cortisol secretion. Moreover, splanchnic nerve activation has been demonstrated to regulate adrenal steroid release (Fig. 4).

BLOOD SUPPLY p0030

Relative to its small size, the adrenal gland is one of the most extensively vascularized organs in the body, with an estimated flow rate of 5 ml per minute. Each gland may be supplied by as many as 50 arterial branches, or arterioles, that arise directly from the aorta, the renal arteries, and the inferior phrenic arteries. The subcapsular arteriolar plexus receives this blood supply and distributes it via two types of vessels. Through the sinusoids, both the adrenal cortex and medulla are supplied, and through the medullary arteries there is a direct blood supply to the medulla. Blood converges at the corticomedullary junction, and through the

Figure 4 Nerve cells and fibers (arrow) silver stained.

52

Adrenal Cortex, Anatomy

Figure 5 Lymphocytes (arrows) immunostained for CD45.

IMMUNE CELLS Macrophages are found within the adrenal cortex. Not only do they possess the ability to act as phagocytic cells but also they are able to produce and secrete a variety of different compounds, such as cytokines [interleukin-1 (IL-1), IL-6, and tumor necrosis factor-a) and neuropeptides (VIP), that influence adrenocortical function. It has been demonstrated that lymphocytes infiltrate the adrenal cortex and have the ability to produce ACTH-like substances (Fig. 5). In addition, a cytokine-independent cell– cell-mediated regulation of adrenal androgen release has been demonstrated.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Adrenal Androgens . Adrenal Cortex, Development . Adrenal Cortex Development, Regulation of . Adrenal Cortex, Physiology . Adrenal Insufficiency . Adrenal Suppression

Further Reading Bornstein, S. R., Ehrhart-Bornstein, M., Scherbaum, W. A., Pfeiffer, E. F., and Holst, J. J. (1990). Effects of splanchnic nerve stimulation on the adrenal cortex may be mediated by chromaffin cells in a paracrine manner. Endocrinology 127, 900–906. Bornstein, S. R., Gonza´lez-Herna´ndez, J. A., Ehrhart-Bornstein, M., Adler, G., and Scherbaum, W. A. (1994). Intimate contact

of chromaffin and cortical cells within the human adrenal gland forms the cellular basis for important intraadrenal interactions. J. Clin. Endocrinol. Metab. 78, 225–232. Dallman, M. F., Engeland, W. C., and McBride, M. H. (1977). The neural regulation of compensatory adrenal growth. Ann. N. Y. Acad. Sci. 297, 373–392. Dijkstra, I., Binnekade, R., and Tilders, F. J. H. (1996). Diurnal variation in resting levels of corticosterone is not mediated by variation in adrenal responsiveness to adrenocorticotropin but involves splanchnic nerve integrity. Endocrinology 137, 540–547. Dinarello, C. A. (1992). The biology of interleukin 1. In “Interleukins: Molecular Biology and Immunology” (T. Kishimoto, ed.), pp. 1–32. Karger, Basel. Ehrhart-Bornstein, M., Hinson, J. P., Bornstein, S. R., Scherbaum, W. A., and Vinson, G. P. (1998). Intraadrenal interactions in the regulation of adrenocortical steroidogenesis. Endocr. Rev. 19(2), 101–143. Gonza´lez-Herna´ndez, J. A., Bornstein, S. R., Ehrhart-Bornstein, M., Geschwend, J. E., Adler, G., and Scherbaum, W. A. (1994). Macrophages within the human adrenal gland. Morphological data for a possible local immune–neuroendocrine interaction. Cell Tissue Res. 278, 201–205. Ottaway, C. A. (1991). Vasoactive intestinal peptide and immune function. In “Psychoneuroimmunology” (R. Ader, D. L. Felten, and N. Cohen, eds.), pp. 225–262. Academic Press, San Diego. Vinson, G. P., Hinson, J. P., and To´th, I. E. (1994). The neuroendocrinology of the adrenal cortex. J. Neuroendocrinol. 6, 235–246. Wolkersdo¨rfer, G. W., Lohmann, T., Marx, C., Schro¨der, S., Pfeiffer, R., Stahl, H.-D., Scherbaum, W. A., Chrousos, G. P., and Bornstein, S. R. (1999). Lymphocytes stimulate dehydroepiandrosterone production through direct cellular contact with adrenal zona reticularis cells: A novel mechanism of immune–endocrine interaction. J. Clin. Endocrinol. Metab. 84, 4220–4227.

Adrenal Cortex, Development Michel Grino Universite´ de la Me´diterrane´e, Marseille, France

Glossary embryogenesis Differentiation of the fertilized ovum during the period of most rapid development, i.e., after the long axis appears until all major structures are represented. nuclear receptor Ligand-inducible transcription factor that specifically regulates the expression of target genes involved in metabolism, development, and reproduction. pregnancy or gestation The condition of having a developing embryo or fetus in the body. Duration of pregnancy is 266 days in women, 184 days in baboons, 165 days in rhesus monkeys, 145–150 days in sheep, 21 days in rats, and 18.5 days in mice. steroid hormones Lipophilic molecules, having a C17 ring as the basis of their chemical structure, that freely cross the cell membrane and interact with nuclear receptors. transcription factor Protein that directly affects the initiation of transcription of specific genes.

mammals, including primates, ruminants, and rodents. In particular, the role of nuclear receptors and transcription factors in the regulation of adrenocortical organogenesis and steroidogenesis is examined.

EMBRYOGENESIS, DEVELOPMENT AND GROWTH Figure 1 depicts the major milestones of adrenal cortex development in primates, sheep, and rodents.

In Primates Human adrenal development begins at approximately the fourth week of gestation and continues into adult life. Adrenocortical cells derive from a single cell lineage that originates in the celomic epithelium in the notch between the primitive urogenital ridge and the dorsal mesentery. These cells are also the origin of gonadal and kidney structures. Five landmark phases have been described: Condensation of the celomic epithelium (3–4 weeks of gestation). . Proliferation and migration of celomic epithelial cells (weeks 4–6) that stream medially and cranially, accumulating at the cranial end of the mesonephros, forming the adrenal blastema. . Morphological differentiation of fetal adrenal cortical cells into two distinct zones (weeks 8–10): the fetal zone and the definitive zone. The fetal zone is an inner cluster of large, eosinophilic cells and represents the largest part (80–90%) of the adrenal cortex. The definitive zone is a thin outer band of small basophilic cells, densely packed, showing structural characteristics of proliferative cells that appear to function as a reservoir of progenitor cells that may populate the remainder of the gland. A third zone, located between the fetal and definitive zones, has been called the transitional zone. By week 30 of gestation, the definitive and transitional zones resemble the adult zona glomerulosa and the zona fasciculata, .

M

aturation of the hypothalamic–pituitary–adrenal (HPA) axis, which is characterized by increased activity during late gestation, is essential for the development of the fetus and plays a critical role in preparing for its transition to extrauterine life. The fetal adrenal cortex synthesizes and secretes androgens and glucocorticoids. In primates, androgens are necessary for placental conversion to estradiol, a hormone that is crucial for the maintenance of pregnancy, the maturation of the fetal and maternal tissues, and immunosuppression, leading to implantation of the placenta and the fetus. Glucocorticoids are essential for the maturation of brain, lung, liver, gut, kidney, and the adrenal itself. In some species, a surge in fetal glucocorticoid secretion has been suggested as being integral to the cascade of events leading to the onset of parturition. However, premature or abnormal exposure of fetuses or newborns to high levels of glucocorticoids permanently programs the HPA axis, leading to an increased prevalence of metabolic and cardiovascular disease. This article details both morphological and functional aspects of adrenal cortex development in several types of

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

53

p0015

54

Adrenal Cortex, Development

Figure 1 The adrenocortical development cascade in human (A), sheep (B), and mouse (C), showing progression from the urogenital ridge to final zonation. W, week of gestation; PW, postnatal week; M, month; Y, year; E, day of gestation; P, postnatal day; ZG, zona glomerulosa; ZF, zona fasciculata. Adapted from Trends in Endocrinology and Metabolism 13, Keegan, C. E., and Hammer, G. D. Recent insights into organogenesis of the adrenal cortex, pp. 200–208, copyright 2002, with permission from Elsevier.

respectively. A period of rapid growth begins at approximately week 10 and continues to term. The fetal zone grows by hypertrophy and limited proliferation, whereas growth in the definitive zone occurs mainly by hyperplasia. Several lines of morphological evidence indicate that the human fetal adrenal gland is a dynamic organ, in which proliferating cells located at the periphery migrate, differentiate, and finally undergo senescence in the central part of the gland. . Decline and disappearance of the fetal zone (first 3 postnatal months); by this period, the primate adrenal cortex remodeling involves apoptosis of the fetal zone and expansion of the preexisting zona glomerulosa and zona fasciculata. Following the involution of the fetal zone, chromaffin elements, derived from the fetal ectoderm, begin to cluster around the central vein. The medulla acquires an adult-like pattern by 12–18 months.

. Establishment and stabilization of the adult zonal pattern (10–20 years of age); this leads to individualization of the three distinct cell layers: the outer zona glomerulosa, the central zona fasciculata, and the inner zona reticularis.

In Sheep In the ovine fetus, the adrenal gland can be identified as early as embryonic day (E) 28. Its development occurs during three phases, reflecting the interactions of the cellular kinetic phenomena of hyperplasia and hypertrophy: . Establishment of functional zonation (E50– E90): this first growth phase is characterized by the separation of the medullary and cortical portions and the presence of a well-defined zona glomerulosa at

55

Adrenal Cortex, Development

approximately E50 and zona fasciculata at approximately E90. During this period, the zona fasciculata grows mainly by hyperplasia. . Quiescence and reactivation (E90–E125): during this period, adrenal growth and adrenal cell division occur at the lowest rate compared to any other time in gestation. . Structural and functional maturation [E130 to postnatal day 2 (P2)]: this second growth phase is characterized by an increase in the rate of cell growth and cell division of the zona fasciculata, with the rate of increase in the volume of steroidogenic cells becoming greater than at any other stage. Cellular hypertrophy precedes cellular hyperplasia. The zona reticularis does not become apparent before 1 month of life.

In Rodents In mice, the adrenal gland starts to develop on E11 from the celomic epithelium, when the anlage of the adrenal cortex is formed. On E12, the cortical cells are found close to the adrenal medulla sympathoblasts. The capsule and the cortical capillaries have completed their development by E15, when the adrenal gland has become a morphologically distinct structure. The individualization of the different layers of the adrenal cortex is nearly completed by birth. The

mouse adrenal possesses a transient developmental zone between the cortical zones and the adrenal medulla: the X zone, which becomes histologically distinct at P10–P14 and enlarges until P21. The X zone subsequently degenerates. The function of the rodent X zone remains unclear.

ONTOGENY OF STEROID BIOSYNTHESIS The adrenals synthesize several classes of steroids: androgens, glucocorticoids, and mineralocorticoids. The major human adrenal steroidogenic pathways are summarized in Fig. 2. Schematically, the process of adrenal steroidogenesis has two major components. The first is quantitative; i.e., it regulates how much steroid can be made at a given moment. The second is qualitative; i.e. it regulates which particular steroid is made. .

First step: Conversion of cholesterol to pregnenolone. This step involves the delivery of the substrate cholesterol to the inner mitochondrial membrane, driven by steroid acute regulatory protein (StAR), and the conversion of cholesterol to pregnenolone, catalyzed by P450 steroid chain cleavage (P450ssc), adrenodoxin, and adrenodoxin reductase.

Figure 2 Schematic view of human adult adrenal and peripheral steroidogenic pathways. StAR, steroid acute regulatory protein; P450ssc, side-chain cleavage enzyme; P450c17, enzyme complex having 17a-hydrolase and 17,20-lyase activities; 17bHSD, 17b-hydroxysteroid dehydrogenase; 3b-HSD, 3b-hydroxysteroid dehydrogenase; P450c21, 21-hydroxylase; P450AS, aldosterone synthase (contains 11b-hydroxylase, 18-hydroxylase, and 18b-hydroxysteroid dehydrogenase activities); P450c11b, 11b-hydroxylase; P450arom, aromatase.

56

t0005

Adrenal Cortex, Development

Table I Localization and Relative Expression of the Enzymes Involved in Fetal Adrenal Steroid Biosynthesis Human

< Week 25

> Week 25

DZ

TZ

FZ

DZ

TZ

FZ

P450ssc

þ

þ

þ

þ

þ

þ

3b-HSD







þþ

þþ



P450c17



þþþ

þþ



þþþ

þþ

P450c21

þþ

þþ

þ

þþ

þþ

þ

P45011b



þ



þþ

þþþ

þþ

Sheep

E50

E90

E130

Whole cortex

ZG

ZF

ZG

ZF

P450ssc 3b-HSD

þþþ þ

þ þ

þ þþ

þþþ þþ

þþþ þþ

P450c17

þþþ



þþ



þþþ

P450c21

þ

þþ

þþ

þþþ

þþþ

þþ

þþ

þþ

þþ

P10

P25

P45011b

Whole Cortex Rodents

E16.5

E18.5

E21

P1

P450ssc

þ

þþ

þþþ

þ

þþþ

þþþ

3b-HSD

þ

þþ

þþþ

þ

þþ

þþþ

P450c17 P450c21

þ ND

 þ

 ND

 þ

 þ

 þþ

P45011b

ND

þ

þþ

þ

þþ

þþþ

Note. E, day of gestation; P, postnatal day; DZ, definitive zone; TZ, transitional zone; FZ, fetal zone; P450ssc, side-chain cleavage enzyme; 3b-HSD, 3b-hydroxysteroid dehydrogenase; P450c17, enzyme complex having 17a-hydrolase and 17,20-lyase activities; P450c21, 21-hydroxylase; P450c11b, 11b-hydroxylase; ND, not determined.

.

the fetal zone produces dihydroepiandrosterone (DHEA), the bulk of which appears to be secreted as a 3-sulfoconjugate DHEA sulfate (DHEAS) that is formed by the action of DHEA sulfotransferase and the cofactor 30 -phosphoadenosine 50 -phosphosulfonate. DHEAS is used by the placenta for conversion to estrogens. By term, the fetal zone produces approximately 200 mg DHEA per day. In fetal sheep and rodents, the adrenals do not synthesize DHEA.

Second step: Transformation of pregnenolone to active hormones. The coordinate regulation of several enzymes will direct the transformation of pregenenolone toward a given class of steroid hormone—androgens, glucocorticoids, and mineralocorticoids. During adrenal cortex development, the synthesis of these various steroids is not temporally coordinated, is zone-dependent, and is species-specific. Table I summarizes the localization and relative expression of the enzymes involved in fetal adrenal steroid biosynthesis.

Androgens The conversion of cholesterol to pregnenolone is mature in the fetal zone during early pregnancy in humans. High levels of 17a-hydroxylase/17,20-lyase (P450c17) are present in the fetal zone very early (as soon as 44 days postconception). As a consequence,

Glucocorticoids In Primates Assays of steroids in cord blood have suggested that human fetus can synthesize cortisol as early as the 10th week of gestation. This hypothesis is supported by clinical and biological observations in patients with congenital adrenal hyperplasia (CAH) caused by deficiency of 21-hydroxylase (P450c21),3b-hydroxysteroid dehydrogenase (3b-HSD), or 11b-hydrolase (P450c11b). Patients with this disorder have impaired glucocorticoid production with a compensatory increase in pituitary adrenocorticotropic hormone (ACTH) secretion and subsequent stimulation of DHEA synthesis. Interestingly, female infants with CAH have ambiguous external genital development, indicating that, in unaffected female fetuses, cortisol is synthesized by the adrenal by the time of external genital differentiation, a process that is sensitive to androgens and begins at approximately week 10. These observations do not necessarily indicate that the week 10 fetal adrenal is able to synthesize cortisol de novo from cholesterol. Although StAR and P450ssc immunoreactivities are expressed in the cytoplasm of human fetal adrenocortical parenchymal cells of the transitional zone as soon as week 14, in early gestation the fetal adrenal can synthesize cortisol using placental progesterone as substrate, since high levels of progesterone are present in the fetal circulation. P450c21 and P450c11b immunoreactivities have been detected in the transitional zone of human fetal adrenals as soon as weeks 13–14 and P450c17 mRNA is present in the transitional zone of week 22 human fetal adrenal. This is consistent with the observation that progesterone infusion into human fetuses between weeks 16 and 18 results in cortisol synthesis. As a consequence, the key enzyme for the production of cortisol de novo from cholesterol during early gestation is 3b-HSD. It has been reported that 3b-HSD immunoreactivity and mRNA could not be detected in human adrenals before weeks 22–24, suggesting that the transitional zone of human fetus adrenals is

57

Adrenal Cortex, Development

able to synthesize de novo cortisol beginning at week 24. Similarly, in nonhuman primates, 3b-HSD mRNA is undetectable during early gestation, starts to increase at midgestation, and is further stimulated at late gestation. 3b-HSD protein expression in the transitional zone of adrenals from fetal rhesus monkeys follows a comparable developmental pattern. In Sheep Adrenal cortisol synthesis and secretion in ovine fetuses follow a triphasic pattern. P450ssc and P450c17 are expressed at a relatively high level in the whole adrenal cortex as early as E40–E60. In the zona fasciculata, the levels of P450ssc and P450c17 decrease between E90 and E120 before showing a further elevation between E130 and P2. At E90, P450c17 expression is confined to the zona fasciculata. 3b-HSD is present at uniformly moderate levels in the zona fasciculata at mid and late gestation. P450c21 mRNA shows a steady increase throughout gestation. Intense P450c11b immunoreactivity is consistently detected throughout the adrenal cortex as early as E90 and remains constant until birth. Cortisol is secreted from adrenal cells from E50 fetuses. Basal cortisol secretion decreases at E100 and increases subsequently, between E130 and E145. In Rodents Adult mouse and rat adrenal cortex, which lack P450c17, produce mainly corticosterone. P450ssc and adrenodoxin are expressed at midgestation in the rodent adrenal cortex (E15–E16). 3b-HSD mRNA and protein have been detected in the fetal rat adrenal as early as E16; their labeling in the reticular and fascicular zones is at a higher level than in the glomerular zone at E18. P450c21 and P450c11b immunoreactivity and mRNA are present in the fetal rat adrenal at E18. P450c17 mRNA and activity are detectable in the adrenals of mouse fetuses at E12.5, increase in abundance from E12.5 to E14.5, and are then lost between E16.5 and E18.5, suggesting that the fetal mouse adrenal is able to synthesize cortisol during late gestation. Although the ontogeny of StAR expression in the rodent embryonic adrenal remains to be elucidated, it is known that fetal rat adrenals synthesize and secrete corticosterone as early as E13. The corticosterone contents of the fetal rat adrenal are high from E16 to E20 and plasma corticosterone concentrations rise progressively from E16 to E19. This phenomenon occurs following activation of pituitary ACTH and the hypothalamic neuropeptides that control ACTH synthesis and secretion. During the first 10 days after birth, there is a marked decrease in adrenal corticosterone contents

and both basal and stress-induced circulating corticosterone levels. This period has been called the ‘‘stress hyporesponsive period’’ (SHRP). These low-circulating glucocorticoid levels are believed to be essential for normal brain and behavioral development. A decrease in the steroidogenic capacity of the newborn adrenal cortex may account, at least in part, for the SHRP. It has been demonstrated that StAR mRNA and protein are highly expressed in the adrenals at birth, decrease subsequently until P14, and increase thereafter. The level of expression of P450ssc is comparable to that of P1 in adults. At birth, adrenal 3b-HSD activity is low at P1, increases at P10, and remains stable until adulthood. P450c21 activity is low (approximately half of the adult values) on P1 and P10. P450c11b immunoreactivity has been detected in the adrenals of newborn rats and did not change during neonatal development. In P7 rats, P450c11b mRNA is present at high levels only in the zonae fasciculata and reticularis.

Mineralocorticoids In Primates Regarding the ontogeny of aldosterone secretion, assays of steroids in cord blood have suggested that human fetus can synthesize aldosterone as early as weeks 16–20. However, in vitro experiments have demonstrated that at midgestation human fetal adrenal tissues do not produce detectable levels of aldosterone, under basal or stimulated conditions. The primary steps in aldosterone synthesis, i.e., those driven by StAR, P450ssc, 3b-HSD, and p450c21, are mature in the definitive zone in human fetal adrenals at the end of the midgestation period. However, in human fetal adrenals obtained from second-trimester abortuses, aldosterone synthase (P450AS) immunoreactivity is absent in the definitive zone and P450AS mRNA is weakly detectable in the whole cortical zone. Similarly, P450AS immunoreactivity is absent in the definitive zone of fetal rhesus monkey adrenal until near term. Activation of the late gestation fetal rhesus monkey HPA axis (obtained after treatment with metyrapone, a compound that inhibits P450c11b) was able to induce P450AS expression. Taken together, these data indicate that in the primate fetal adrenal gland, the definitive zone has the capacity to synthesize aldosterone, but not until term. In Sheep In the fetal sheep, levels of aldosterone synthesis and secretion are very low until the final part of gestation.

58

Adrenal Cortex, Development

Although P450AS is expressed as early as E90 in the fetal zona glomerulosa, there is weak 3b-HSD staining until term. In Rodents In rats, P450AS immunoreactivity is detected in E16 adrenals, in small clusters of cells, dispersed throughout the gland. By E18–E19, the number of P450ASlabeled cells increases and these cells become localized in the outer cortex. P1 adrenals have a pattern of P450AS staining comparable to that of the adult gland. In P7, rats, P450AS mRNA is confined to the subcapsular zona glomerulosa. Aldosterone content in fetal adrenal homogenates increases between E17 and P1 and remains stable thereafter.

ROLE OF NUCLEAR RECEPTORS AND TRANSCRIPTION FACTORS The complex cascade of the development and differentiation of the adrenal cortex, from the formation of the urogenital ridge to the zonation of the fetal adrenal, involves several nuclear receptors and transcription factors (Table II).

WT1 and WNT4 Two genes are crucial for the development of the urogenital ridge: WT1 (Wilm’s tumor suppressor gene 1), which encodes a protein having many characteristics of a transcription factor, and WNt4 (wingless-related mouse mammary tumor virus integration site 4), a member of the WNt family of developmentally regulated signaling molecules. WT1 is developmentally one of the earliest known genes that specifies kidney, gonadal, and adrenal cell lineages. Mice fetuses knocked out for Wt1 and rescued with a YAC (yeast artificial chromosome) construct spanning the Wt1 locus have adrenal-like structures that are greatly reduced in size and express lower levels of P450ssc mRNA than wild-type mice. WnT4 is involved in the development of the kidney, pituitary gland, female reproductive system, and mammary gland. Wnt4 is expressed next to the anterior site of the mouse mesonephros on E11.5 and in the developing adrenal cortex from E12.5 onward. Adrenals of Wnt4 knockout mice are morphologically comparable to those of wild-type animals. However, Wnt4 knockout animals have reduced adrenal P450c21 and P450AS mRNA concentrations and decreased aldosterone production.

Table II Transcription Factors and Nuclear Receptors Involved in Adrenal Differentiation and Steroidogenesis Transcription factor or nuclear receptor

First detected on

Phenotype of rodent KO

Genes that are regulated in adrenal cells

WT1

Mouse: E9

Lethal: lacks adrenals, kidneys, and gonads

DAX-1

WnT4

Mouse: E11.5; human: E33

Abnormal kidney and adrenal development

P450c21 P450AS

SF-1

Mouse: E11; human: E33

Lethal: lacks adrenals and gonads

StAR P450ssc 3b-HSD P450c17 P450c21 P450AS P450c11b ACTH-R DAX-1

DAX-1

Mouse: E10.5

Lack of X zone regression

StAR P450c17 P450c21 ACTH-R

Note. WT-1, Wilm’s tumor suppressor gene 1; WnT4, wingless-related mouse mammary tumor virus integration site 4; SF-1, steroidogenic factor-1; DAX-1, dosage-sensitive sex reversal adrenal hypoplasia congenita critical region on the X chromosome, gene 1; E, day of gestation; StAR, steroid acute regulatory protein; P450ssc, side-chain cleavage enzyme; P450c17, enzyme complex having 17a-hydrolase and 17,20-lyase activites; 17b-HSD, 17b-hydroxysteroid dehydrogenase; 3b-HSD, 3b-hydroxysteroid dehydrogenase; P450c21, 21-hydroxylase; P450AS, aldosterone synthase; P450c11b, 11b-hydroxylase; ACTH-R, adrenocorticotropic hormone receptor.

p0080

59

Adrenal Cortex, Development

SF-1, DAX-1, and GATA Proteins p0085

p0090

Several transcription factors, such as steroidogenic factor-1 (SF-1), dosage-sensitive sex-reversal adrenal hypoplasia congenita critical region on the X chromosome, gene 1 (DAX-1), and GATA proteins, play major roles in the development of the adrenal primordium, its functional zonation, and the regulation of fetal adrenal steroidogenic capacities. SF-1 SF-1, also called Ad4BP or NR5A1, belongs to the orphan receptor class of nuclear receptors. It has been proposed that SF-1 acts through recruitment of cofactors, including both coactivators (CREB-binding protein, WT1, and nuclear receptor coactivator 1) and corepressors (nuclear receptor corepressor, DEAD/H box polypeptide 20, COP9 constitutive photomorphogenic subunit 2, nuclear receptor-interacting protein 1, and DAX-1). In mouse, Sf-1 mRNA was detected in the adrenal primordium from E11. From E14–E14.5 onward, Sf-1 mRNA is restricted to the steroidogenic cells in the cortex. In humans, SF-1 was expressed as early as E33 in the presumptive adrenal primordium. Newborn Sf-1 knockout mice are devoid of adrenals and die from adrenocortical insufficiency shortly after birth. However, Sf-1 heterozygous mice (SF-1þ/) are viable but show adrenal disorganization (reduced adrenal size and hypoplastic zona fasciculata and adrenal medulla), altered adrenal gene expression (increased StAR mRNA and decreased ACTH receptor mRNA), and impaired basal and stress-induced glucocorticoid secretion. Studies of SF-1 gene mutations in humans demonstrated that SF-1 regulates adrenal development in a dose-dependent manner. A number of studies have demonstrated that SF-1 acts a global regulator of the proteins involved in adrenal steroidogenesis (StAR, P450ssc, 3b-HSD, P450c21, P450AS, P45011b, and P450c17). In addition, Sf-1 regulates adrenal ACTH sensitivity through a cell-specific modulation of both constitutive and cyclic AMP-induced expression of the ACTH receptor gene. In the late-gestation ovine fetus, Sf-1 mRNA is present in adrenal extracts; its expression is positively regulated by a pituitary-dependent factor, which is not ACTH. DAX-1 DAX-1 is an orphan nuclear receptor that regulates both adrenal development and functional zonation. Mutations in the DAX-1 gene are responsible for X-linked adrenal hypoplasia congenita, an inherited disorder in humans that is characterized by hypoplasia

of the fetal adrenal glands with absence of the definitive zone and persistence of the fetal zone. DAX-1 has been postulated to bind to hairpin loops of the StAR promoter and/or to function as an RNA-binding protein. In human fetuses, DAX-1 is expressed in the adrenal primordium as early as E33. In mouse, it has been shown that Dax-1 is colocalized with Sf-1 in the developing adrenal and that Sf-1 expression precedes or coincides with expression of Dax-1, suggesting that these two molecules could cooperate in regulating adrenal development. In vitro, Dax-1 inhibited Sf-1induced stimulation of transcription of the adrenal promoter of StAR, P450c17, Dax-1 itself, possibly through interactions with Sf-1, and transcriptional corepressors such as NcoR and Alien, whereas Sf-1 stimulated expression of the Dax-1 promoter. In vivo, male mice with a single Dax-1-deleted allele (Dax-1/y) have normal zona glomerulosa, but the zona fasciculata is less well developed and shows decreased staining for 3b-HSD and the X zone fails to regress. Dax-1/y animals have increased stress-induced corticosterone secretion and enhanced adrenal responsiveness to exogenous ACTH stimulation, most probably following increased adrenal P450c21 and ACTH receptor expression. Dax-1/y animals have normal expression of Sf-1 and expression of Dax-1 is maintained in Sf-1þ/ mice. The absence of Dax-1 partially reverses adrenal growth defects in Sf-1þ/ mice. However, the precise mechanisms governing the interplay between Sf-1 and Dax-1 in regulating adrenal development and steroidogenesis are not known. GATA Proteins GATA proteins are transcription factors that bind to the consensus sequence (A/T)GATA(A/G) in the promoter and enhancer regions of their target genes. GATA-4 is present in human adrenocortical carcinoma and in a transgenic mouse model developing adrenocortical tumors. GATA-4 and GATA-6 are detectable from E14 and week 19 onward in the mouse and human adrenal cortex, respectively. After birth, GATA-4 expression decreases, whereas GATA-6 continues to be expressed. Studies in chimeric mouse embryos demonstrated that GATA-4 is not essential for early adrenocortical differentiation. The exact role of GATA-6 in adrenal development remains to be determined.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Adrenal Androgens Adrenal Cortex, Anatomy . Adrenal Cortex Development,

.

60 Regulation of . Adrenal Cortex, Physiology . Adrenal Insufficiency . Adrenal Suppression . Thyroid Gland Development, Molecular Biology

Further Reading Challis, J. R. G., Sloboda, D., Matthews, S. G., Holloway, A., Alfaidy, N., Patel, F. A., Whittle, W., Fraser, M., Moss, T. J. M., and Newnham, J. (2001). The fetal placental hypothalamic– pituitary–adrenal (HPA) axis, parturition and post natal health. Mol. Cell. Endocrinol. 185, 135–144. Grino, M., Dakine, N., Paulmyer-Lacroix, O., and Oliver, C. (2001). Ontogeny of the hypothalamo–pituitary–adrenal axis. In ‘‘Contemporary Endocrinology: Adrenal Disorders’’ (G. P.

Adrenal Cortex, Development

Chrousos and A. N. Margioris, eds.), pp. 1–9. Humana Press, Totowa, NJ. Keegan, C. E., and Hammer, G. D. (2002). Recent insights into organogenesis of the adrenal cortex. Trends Endocrinol. Metab. 13, 200–208. Matthews, S. G. (2002). Early programming of the hypothalamo– pituitary–adrenal axis. Trends Endocrinol. Metab. 13, 373–380. Mesiano, S., and Jaffe, R. B. (1997). Developmental and functional biology of the primate fetal adrenal cortex. Endocr. Rev. 18, 378–403. Naaman Re´pe´rant, E., and Durand, P. (1997). The development of the ovine fetal adrenal gland and its regulation. Reprod. Nutr. Dev. 37, 81–95. Parker, K. L., and Schimmer, B. P. (1997). Steroidogenic factor 1: A key determinant of endocrine development and function. Endocr. Rev. 18, 361–377.

Adrenal Cortex, Physiology Christian A. Koch University of Leipzig, Leipzig, Germany

Glossary adrenocorticotropic hormone or corticotropin (ACTH) Synthesized as part of a 241-amino acid precursor, proopiomelanocortin. Usually, ACTH derives from the pituitary gland to stimulate the adrenal glands through the melanocortin-2 receptor (rarely, ACTH is produced ectopically, i.e., in ectopic Cushing’s syndrome). corticotropin-releasing hormone (CRH) Hypothalamic 41-amino acid peptide that usually stimulates the pituitary to release ACTH. hypothalamic–pituitary–adrenal (HPA) axis Hypothalamic-releasing factors including CRH are influenced by central nervous system afferents. For instance, stress can trigger CRH release. The major secretagogue of CRH is ACTH, which subsequently stimulates the adrenals to produce and release cortisol. Cortisol, on the other hand, can inhibit further release of CRH and ACTH in a negative feedback loop. steroids Hormones produced by the adrenal gland. Members of a large family of compounds that are derived from the cyclopentanperhydrophenanthrene ring structure that consists of three cyclohexane rings and one cyclopentane ring. The nomenclature denotes rings by a letter and the individual carbon atoms by a number. Gonane is the unsaturated 17-carbon ring structure. Estranes are steroids with 18 carbons (C18 steroids) by adding a methyl group at C13. Androstane is a C19 steroid with two methyl groups. Pregnane is a C21 steroid with methyl and ethyl groups. The endingene denotes a double bond of the parent compound.

T

he Tabulae Anatomicae mention Bartholomeo Eustachius, who apparently first described the anatomy of the adrenal glands in 1563. A central physiologic role for the adrenals was presented in 1849 by Thomas Addison, and today Addison’s disease refers to adrenal insufficiency, a life-threatening condition. To better understand pathophysiologic states of the adrenal glands, one must recall their normal anatomy and physiology. The adrenal glands are divided into cortex and medulla. The adrenal cortex is composed of three zones— the glomerulosa, fasciculata, and reticularis. The largest and

68

most important zone is the fasciculata, where glucocorticoids including cortisol are produced. Extracellular volume status is influenced by aldosterone, the hormone of the zona glomerulosa. Androgens such as dehydroepiandrosterone are produced by the zona reticularis, which starts to grow at approximately age 4, shortly before adrenarche occurs. Corticotropin, usually coming from the pituitary, is the principal stimulus for cortisol secretion and is released under stress and other stimuli through hypothalamic corticotropin-releasing hormone (CRH) secretion. Corticotropin can also stimulate the zona glomerulosa (to a minor extent) and reticularis to secrete aldosterone and adrenal androgens. A feedback loop exists between the hypothalamus (H), pituitary (P), and adrenal gland (A)—the HPA axis. High peripheral cortisol levels inhibit further release of CRH and corticotropin. At the extreme (e.g., in conditions of supraphysiologic exogenous glucocorticoid administration), the secretion of hypothalamic CRH and pituitary corticotropin becomes severely suppressed, with subsequent atrophy of the adrenal glands due to the lacking stimulus corticotropin. The adrenal medulla forms postnatally and exerts effects on the adrenal cortex and vice versa. Adrenomedullary chromaffin cells are intermingled with the adrenal cortex, facilitating an interaction between the two.

FETAL ADRENAL GLAND Fetal Adrenal Steroidogenesis The fetal adrenal cortex plays a critical role in regulating intrauterine homeostasis and the maturation of fetal organ systems that are necessary for extrauterine life. The important mediators for these functions are steroid hormones from the fetal adrenal. Throughout gestation and postnatally, the fetal adrenal gland undergoes morphological and functional changes during its transformation to the adult adrenal gland. The progenitor cells of the adrenal cortex stem from a cell lineage that also leads to steroid-secreting cells of the gonads. By week 8 of gestation, the adrenal cortexforming progenitor cells build an inner cluster (the fetal zone) consisting of large eosinophilic steroidsecreting cells that express high levels of steroid 17a-hydroxylase (CYP17) and an outer zone (the definitive zone) consisting of cells that do not express

Encyclopedia of Endocrine Diseases, Volume 1. Published by Elsevier Inc.

69

Adrenal Cortex, Physiology

CYP17. Between the fetal and definitive zone lies the transitional zone, which is composed of cells similar to those of the zona fasciculata of the adult adrenal gland. At gestational week 8, chromaffin cells enter the rudimentary adrenal gland and remain there as discrete islands until day 8 postnatally, before they form a rudimentary adrenal medulla. During gestation, the fetal zone represents 85% of the cortical volume and produces large amounts of dehydroepiandrosterone (DHEA) sulfate, a hormone that serves as a precursor for other androgens synthesized in the periphery. Before birth, DHEAS is used by the placenta to synthesize estrogen. Placental estrogen, on the other hand, supports the fetal adrenal to synthesize cortisol. Several endocrine, paracrine, and autocrine factors influence the steroidogenesis of the fetal adrenal cortex. Within the first postnatal year, the fetal zone degenerates. During the third trimester, aldosterone production starts by the zona glomerulosa, derived from the outer zone of the fetal adrenal cortex (without CYP17). A transitional area between the inner and outer zones gives rise to the zonae fasciculata and reticularis, which secrete glucocorticoids and DHEA sulfate under stimulation and control by corticotropin. Certain transcription factors (e.g., steroidogenic factor 1, DAX1, and WT1) are essential in adrenal development. Adrenal dys- or agenesis with subsequent adrenal insufficiency and death can result from nonfunction of these factors, for instance, by mutations in the steroidogenic factor 1.

Growth Factors During the first trimester of pregnancy, human chorionic gonadotropin (HCG) regulates the growth of the fetal adrenal. Adrenocorticotropic hormone (ACTH) is critical for growth, steroidogenesis, and differentiation of the fetal adrenal gland and becomes the main force for further growth after the fifth month of gestation. ACTH deficiency leads to increased apoptosis and subsequent atrophy of the adrenal glands, whereas ACTH excess (e.g., in Cushing’s syndrome or congenital adrenal hyperplasia) can cause hyperplastic adrenal glands. In addition to HCG, ACTH, and its receptor ACTHR, local growth factors are important for steroidogenesis, growth, and development of the adrenal gland. Insulin-like growth factors 1 and 2 (IGF-1 and IGF-2), their receptors, and binding proteins are all expressed in the fetal adrenal. IGF-1 not only amplifies the effect of ACTH on the adrenal but also enhances steroid production of the adrenal gland by increasing

the activities of 17a, 21-, and 11b-hydroxylase. Similarly, IGF-2 promotes ACTH action. In addition, IGF-2 helps the fetal adrenal to synthesize cortisol and androgen by regulating the enzymes p450scc, p450c17, and 3b-hydroxysteroid dehydrogenase. Basic fibroblast growth factor is a mitogenic protein and is more effective (stimulating proliferation) on adrenal cells of the definitive zone than those of the fetal zone. Epidermal growth factor (EGF) and transforming growth factor-a (TGF-a) are growth factors with sequence homology that both activate the EGF receptor. This may lead to hypertrophy of the fetal adrenal. Although EGF stimulates hypothalamic corticotropin-releasing hormone (CRH) release, it does not cause subsequent ACTH secretion from the pituitary. The TGF-b family of growth factors, including activin, inhibin, and TGF-b1, are paracrine/autocrine regulators of growth and steroidogenesis in the fetal adrenal cortex. Activin increases ACTH-stimulated cortisol production but not DHEAS production in fetal zone cells. In the adult adrenal cortex, activin has no effect on growth or steroidogenesis. In fact, activin may lead to apoptosis and involution of the fetal adrenal cortex postnatally. TGF-b1 appears to decrease fetal and definitive zone cell proliferation and steroidogenesis.

Nuclear Receptors Steroidogenic factor 1 (SF-1), DAX-1, and the estrogen receptor (ER) belong to the nuclear receptor superfamily. Members of this family are transcription factors that are important for regulating expression of genes involved in cellular growth control and differentiation. SF-1 is classified as an orphan receptor because its ligand is unknown. The human cDNA sequence of SF-1 is highly homologous (>95%) to murine and bovine sequences. Human adrenal cortex, ovaries, testes, and spleen show high SF-1 mRNA expression. In human placenta, SF-1 is not or only minimally expressed. SF-1 plays an essential role in the organogenesis of the fetal adrenal gland and also in regulating genes that code for steroidogenic enzymes. SF-1 stimulates the promoter activities of genes encoding steroidogenic acute regulatory (StAR) protein, the scavenger receptor-type class BI (SR-BI), and the ACTH receptor. StAR protein is critical in the translocation process of cholesterol from the outer to the inner mitochondrial membrane. In contrast to the adult adrenal gland, the fetal adrenal uses lowdensity lipoprotein (LDL) rather than high-density lipoprotein cholesterol as the main source for steroid

70

p0035

Adrenal Cortex, Physiology

biosynthesis. It appears that SR-BI binds to LDL with high affinity. SF-1 influences the constitutive activity of the human ACTH receptor gene promoter and regulates steroid hydroxylase enzymes. In addition, SF-1 regulates the genes coding for the b-subunit of luteinizing hormone, the a-subunit of the glycoprotein hormones, gonadotropin-releasing hormone receptor, prolactin receptor, oxytocin, Mullerian inhibiting substance, and aromatase. Furthermore, SF-1 interacts with other proteins and cofactors. Gene deletions at Xp21/22, where the DAX-1 gene is located, may lead to X-linked adrenal hypoplasia congenita, glycerol kinase deficiency, hypogonadotropic hypogonadism, and/or Duchenne’s muscular atrophy. DAX-1 (dosage-sensitive sex reversal, adrenal hypoplasia congenita, X-linked) is highly expressed in the fetal and adult adrenal gland, hypothalamus, pituitary, testes, and ovaries. Together with SF-1, it may coregulate steroidogenesis as well as adrenal and gonadal organogenesis. DAX-1 can block steroidogenesis by inhibiting the activity of StAR and the expression of p450scc and 3b-hydroxysteroid dehydrogenase. Estrogens are important in cell differentiation, growth, and function of various tissues. Estrogen receptors are members of the steroid receptor superfamily Z Glom

Z Fas

and mediate the action of estrogens. ERb is highly expressed in the fetal adrenal gland, in contrast to ERa. As mentioned previously, estrogen is critical in fetal steroidogenesis and organogenesis.

ADULT ADRENAL CORTEX Steroid Biosynthesis and Regulation of Cortisol Production The normal adult human adrenal gland weighs approximately 5 g. Ninety percent of this weight is due to the adrenal cortex, which is composed of three zones (from outside to inside): the zona glomerulosa, the zona fasciculata, and the zona reticularis. The adrenal medulla forms postnatally and is composed of chromaffin cells, some of which may still be intermingled and spread within the adrenal cortex. The adult adrenal cortex produces glucocorticoids, mineralocorticoids, and adrenal androgens (Fig. 1 and Table I). The blood flow in the adrenal gland is centripetal (from outside to inside), which exposes the inner zones and the adrenal medulla to increasing concentrations of adrenal steroids. High cortisol levels in the medulla are needed to induce enzymes for epinephrine biosynthesis. Seventy-five percent Z Ret

Cholesterol P450scc (Desmolase)

P450c17 (17-Hydroxylase) Pregnenolone

P450c17 (17/20 Lyase) 17-OH-Pregnenolone

Dehydroepiandrosterone

3β-OH-steroid dehydrogenase I/II

5α-Red I & II

17β-HSD III/II Progesterone

17-OH-Progesterone

P450c21 (21-Hydroxylase)

Deoxycorticosterone

11-Deoxycortisol

Androstenedione P450aro (19-H) Estrone

Testosterone 17β-HSD I/II

Dihydrotestosterone

P450aro Estradiol

P450c11 (11-Hydroxylase)

Corticosterone

Cortisol

P450c11 (18-Hydroxylase)

18-OH-Corticosterone P450c11 (18-Oxidase)

Aldosterone

Figure 1 Adrenal steroidogenesis. Z Glom, zona glomerulosa; Z Fas, zona fasciculata; Z Ret, zona reticularis; 19-H ¼ 19hydroxylase; HSD, hydroxysteroid dehydrogenase; P450aro, aromatase; 5a-Red, 5a-reductase. The three adrenal cortex zones Z Glom, Z Fas, and Z Ret are listed above the hormones that are produced in the respective zones. The steroidogenic enzymes on the left, starting with P450scc (desmolase), are listed in order of vertical and horizontal reading—that is, desmolase converts cholesterol to pregnenolone, 3b-OH-steroid dehydrogenase I/II convert pregnenolone to progesterone, 17-OH-pregnenolone converts progesterone to 17-OH-progesterone, P450c11 converts deoxycorticosterone to 18-OH-corticosterone and 11-deoxycortisol to cortisol, etc.

71

Adrenal Cortex, Physiology

Table I Key Human Steroidogenic Enzymes and Cofactor Proteinsa Protein

Gene

Chromosomal locus

Substrate

Activities

StAR

STAR

8p11.2

Cholesterol flux within mitochondria

Sterol delivery to P450scc

P450-OR

CPR

7p15-q35

Microsomal P450s

Electron transfer

Cytochrome b5

CYB5

18q23

Aids P450c17

Augments 17,20-lyase activity

Adrenodoxin reductase

ADR

17q24–25

Mitochondrial P450s

Electron transfer

Adrenodoxin 11 b-HSDI

ADX HSD11B1

11q22 1

Mitochondrial P450s Cortisol, cortisone, corticosterone, 11-De B

Electron transfer 11 b-Ketosteroid dehydrogenase

11b-HSD II

HSD11B2

16p22

Cortisol, corticosteron

11b-Hydroxysteroid dehydrogenase

5a-Reductase Type I

SRD5A1

5p15

Testosterone, C21-steroids

5a-Reduction 5a-Reduction

5a-Reductase Type II

SRD5A2

2p23

Testosterone, C21-steroids

17b-HSDI

HSD17B1

17q21

Estrone

17b-Ketosteroid reductase

17b-HSD II 17b-HSD III

HSD17B2 HSD17B3

16q24 9q22

Testosterone, estradiol, DHT Androstenedione

17b-Hydroxysteroid dehydrogenase 17b-Ketosteroid reductase

17b-HSD V

HSD17B5

10p14–15

Androstenedione, DHT, 3a-asdiol, asone

17b-Ketosteroid reductase

Asdione, 3a-androstan

3a-Hydroxysteroid

Preg, 17OH-Preg, DHEA, adiol

3b-Dehydrogenase

Dehydrogenase 3b-HSD I

HSDB1

1p13

DHEA, asdiol

5/4-Isomerase

3b-HSD II

HSDB2

1p13

Preg, 17OH-Preg, DHEA, asdiol

3b-Dehydrogenase

P450scc

CYP11A1

15q23–24

Cholesterol

5/4-Isomerase 22R-Hydroxylase

Hydroxysterols

20, 22-Lyase

Preg, 17OH-Preg

17a-Hydroxylase

20R-Hydroxylase P450c17

CYP17

10q24.3

Prog, DHEA

17,20 Lyase

P450c21

CYP21B

6p21.1

Prog, 17OH-Prog

21-Hydroxylase

P450c11b

CYP11B1

8q21–22

11-Deoxycortisol

11-Hydroxylase

8q21–22

11-DOC Corticosterone

11-Hydroxylase

P450c11AS P450aro

CYPB2 CYP19

15q21.1

11-DOC

18-Hydroxylase

Androstenedione

19-Hydroxylase

Testosterone

19-Oxidase, aromatization

a

Modified from Auchus and Miller (2001). Androstan, androstanediol; Preg, pregnenolone; Prog, progesterone; DOC, deoxycorticosterone; HSD, hydroxysteroid deydrogenase; DHT, dihydrotestosterone; 11-De B, 11-dehydrocorticosterone B; 17OH-Prog, 17a-hydroxyprogesterone; 3a-Asdiol, 3a-androstandediol; Asdione, androstanedione.

of the adrenal gland weight is due to the zona fasciculata, the largest zone and the one that synthesizes glucocorticoids. The zona fasciculata also produces DHEA and DHEAS, whereas the zona reticularis also produces cortisol. In contrast to the zona fasciculata, the zona reticularis is small and not very involved in adrenal androgen production until adrenarche (approximately 6 years of age). The precursor for glucocorticoid production is cholesterol, which in a first step is converted to pregnenolone in the adrenal cortex. Steroids derive from the cyclopentanoperhydrophenanthrene four-ring hydrocarbon nucleus, a relatively inert structure. Depending on the presence

of several enzymes in the respective adrenal cortex zone, several steroid hormones can then be synthesized. Cytochromes P450 are categorized into two classes: type 1 enzymes that reside in the mitochondria and type 2 enzymes located at the smooth endoplasmic reticulum (Table II). The secretion and synthesis of cortisol are regulated by the hypothalamic–pituitary–adrenal (HPA) axis. Certain stimuli including stress lead to the release of CRH in the hypothalamus. CRH then stimulates ACTH release from the pituitary (Fig. 2). ACTH binds to ACTH receptors located on adrenocortical cells and stimulates them to release cortisol

72

Adrenal Cortex, Physiology

Hypothalamus/Hippocampus CRH CRH

ACTH ACTH

ACTH

Cortisol Apoptosis

Adenoma

Atrophy Cortisol

Cortisol Normal adrenal gland

Patient with a cortisol-producing adrenal adenoma

Figure 2 The HPA axis. Modified from Koch and Chrousos (2001).

through cAMP. Cortisol can then increase energyproviding compounds, including glucose, free fatty acids, and free amino acids. As mentioned previously, ACTH is also growth promoting on the adrenal cortex; that is, continuous stimulation by ACTH may cause adrenal hypertrophy, whereas a lack of ACTH may lead to adrenal atrophy. The HPA axis is very sensitive to exogenous glucocorticoid administration, which can easily lead to ACTH suppression through a negative feedback loop on CRH. In normal individuals who are not working in (night) shifts, there is a diurnal variation of cortisol production, with serum cortisol being highest in the morning and lowest at midnight. In patients with Cushing’s syndrome (hypercortisolism), these normal physiologic circuits are disturbed (i.e., there is no suppression of

the HPA axis and cortisol production by administration of exogenous glucocorticoids such as dexamethasone). Prolonged (7–48 h) increases in ACTH lead to an increased synthesis of all the steroidogenic enzymes, especially P450scc, as well as an increased uptake of cholesterol from the circulation. Chronic lack of ACTH (e.g., through exogenous glucocorticoid administration) leads to adrenal atrophy. Therefore, the exogenous glucocorticoid has to be tapered to allow the pituitary and adrenal gland to recover in order to synthesize normal levels of cortisol on its own. Depending on the level of suppression, this may take weeks or months.

Table II Locations of Steroidogenic Proteins

Aldosterone, the major human mineralocorticoid, is produced in the zona glomerulosa of the adrenal cortex. Its secretion is stimulated mainly by angiotensin II (and III) through the renin–angiotensin– aldosterone system and, to a lesser extent, by ACTH. Chronic infusion of ACTH stimulates aldosterone secretion for only 24 h. Less potent stimulators of aldosterone secretion are endothelin and serotonin. Also, increases in potassium concentrations stimulate aldosterone production. An increase in serum potassium of 0.1 mmol/liter can elevate plasma aldosterone by 35%. On the other hand, a decrease in serum potassium of 0.3 mmol/liter can reduce plasma aldosterone by 46%. Aldosterone decreases the absorption

Endoplasmic reticulum

Cytoplasm

Mitochondria

11b-HSD I and II

3a-HSD

3b-HSD II

5a-Reductase I and II

17b-HSD V

StAR

17b-HSD I–III

17b-HSD I

Adrenodoxin reductase

3b-HSD II

3b-HSD II

P450c11AS

Cytochrome b5 P450 oxidoreductase

StAR Adrenodoxin

P450c11b P450scc

P450aro P450c21 P450c17 Modified from Auchus and Miller (2001).

Biosynthesis and Regulation of Aldosterone Production

73

Adrenal Cortex, Physiology

of potassium and promotes sodium reabsorption and fluid retention, thereby increasing the extracellular fluid volume. However, after a few days of extracellular fluid expansion by increased aldosterone levels, the individual will be protected from continuous expansion through a so-called “escape” mechanism that denotes attaining a new sodium balance and the formation of a new steady state. Target tissues of aldosterone, including kidney (distal tubules and cortical collecting ducts), colon, and salivary glands, have mineralocorticoid receptors that bind aldosterone. In the distal nephron, cortisol is a potent agonist at the mineralocorticoid receptor. Among inhibitors of aldosterone secretion are atrial natriuretic peptide (ANP) and dopamine. ANP strongly inhibits stimulated (e.g., low sodium intake) aldosterone secretion, with much less effect on basal (e.g., normal or high sodium intake) activity. Chronic sodium restriction leads to increased activity of aldosterone synthase and a higher content of this enzyme in the zona glomerulosa. The first steps of aldosterone biosynthesis are identical to those of cortisol biosynthesis (Fig. 1). The synthesis of cortisol, however, depends on 17a-hydroxylation of pregnenolone by 17a-hydroxylase (P450c17), which is exclusively expressed in the zona fasciculata. On the other hand, aldosterone synthase is normally expressed only in the zona glomerulosa.

Regulation of Adrenal Androgen Production At approximately 4 years of age, in both sexes the zona reticularis forms and continues to grow until the mid20s. After age 40, this zone gradually regresses. Corticotropin and prolactin stimulate adrenal androgen secretion in the fetal adrenal zone. Postnatally, the zona reticularis responds to ACTH, as exemplified in congenital adrenal hyperplasia in which ACTH and androgen hypersecretion can occur. During infancy, only small amounts of androgens are secreted, and it is unknown how adrenarche, the time point at which a slight amount of pubic hair develops, is regulated. Seventy percent of circulating testosterone in women with normal menstrual cycles derives from the conversion of adrenal DHEA. The principal androgens secreted by the adrenals are DHEA, DHEAS, androstenedione, and (minimally) testosterone. DHEAS per se has only weak androgenic effects. Peripheral conversion of the aforementioned precursors leads to more potent androgens, such as testosterone and dihydrotestosterone. Major conversion sites include

the hair follicles, sebaceous glands, external genitalia, and prostate. Peripheral adipose tissue can convert androgens into estrogens by the highly active enzymes aromatase and 17-ketosteroid reductase. Glucocorticoids stimulate aromatase. Inactivation or degradation of androgens and their metabolites occur at different sites, including the liver and kidneys. Exogenous adrenal androgen administration can suppress gonadotropin secretion. Excess endogenous androgen production can be caused by several conditions, including congenital adrenal hyperplasia and adrenal tumors.

Impact of the Sympathoadrenal System on the Regulation of Adrenocortical Function Adrenocortical steroid hormones influence the differentiation and hormone production of adrenal chromaffin cells. On the other hand, the sympathoadrenomedullary system modulates diurnal variations of steroidogenesis in the adrenal cortex. The adrenal cortex is innervated by neurons originating in cell bodies within the adrenal medulla and by nerves that have cell bodies outside the adrenal, reaching the cortex via blood vessels. Adrenal chromaffin cells contain many neuropeptides that regulate adrenocortical steroid production in many species. Adrenomedullary cells are found throughout the adrenal cortex, which facilitates the paracrine action of their products. Another avenue for adrenomedullary secretory products reaching the adrenal cortex is the lymphatics.

CONCLUSION The adrenal cortex fulfills important functions before and after birth. Prenatally, the fetal adrenal cortex is large and mainly consists of the fetal zone, which produces high amounts of DHEAS, a hormone that serves as a precursor for other androgens. DHEAS is used by the placenta to synthesize estriol and to regulate intrauterine homeostasis as well as maturation of fetal organ systems that are necessary for life after birth. Postnatally, the adrenal cortex becomes a three-zoned structure, with the largest zone being the zona fasciculata. From outside to inside, the three zones are the zona glomerulosa, the zona fasciculata, and the zona reticularis. The zona glomerulosa produces aldosterone. Production of this hormone is stimulated by angiotensin II and, to a lesser extent, ACTH, potassium, endothelin, and serotonin. Aldosterone secretion is inhibited by dopamine and atrial natriuretic peptide. The zona fasciculata produces mainly cortisol, the stress hormone.

74

Adrenal Cortex, Physiology

Secretion of this hormone is regulated by the HPA axis. Chronic stress can overstimulate the HPA axis, leading to depression. The zona reticularis produces adrenal androgens and can be stimulated by ACTH. All three zones gradually regress during the life span. Interaction between the adrenal cortex and the adrenal medulla exists and is facilitated by the intermingling of adrenomedullary chromaffin cell islets with the adrenal cortex. The adrenal medulla matures within the first 18 months postnatally and produces the stress hormones epinephrine and norepinephrine as well as other catecholamines.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Adrenal Androgens . Adrenal Cortex, Anatomy . Adrenal Cortex, Development . Adrenal Cortex Development, Regulation of . Adrenal Insufficiency . Adrenal Suppression

Further Reading Auchus, R. J., and Miller, W. L. (2001). The principles, pathways, and enzymes of human steroidogenesis. In “Endocrinology” (L. J. DeGroot and J. L. Jameson, eds.), 4th ed. Saunders, Philadelphia.

Adrenal Hypoplasia see Congenital Adrenal Hypoplasia

Crowder, R. E. (1957). The development of the adrenal gland in man, with special reference to origin and ultimate location of cell types and evidence in favor of the cell migration theory. Contemp. Embryol. 251, 195–209. Erhart-Bornstein, M., Hinson, J. P., Bornstein, S. R., Scherbaum, W. A., and Vinson, G. P. (1998). Intraadrenal interactions in the regulation of adrenocortical steroidogenesis. Endocr. Rev. 19, 101–143. Koch, C. A., and Chrousos, G. P. (2001). Is the diminuto/dwarf1 gene involved in physiologic adrenocortical size regulation and tumor formation? J. Clin. Endocrinol. Metab. 86(11), 5127–5129. Kvetnansky, R., Pacak, K., Fukuhara, K., Viskupic, E., Hiremagalur, B., Nankova, B., Goldstein, D. S., Sabban, E. L., and Kopin, I. J. (1995). Sympathoadrenal system in stress: Interaction with the hypothalamic–pituitary–adrenocortical system. Ann. N. Y. Acad. Sci. 771, 131–158. Margioris, A. N. and Chrousos, G. P. (eds.) (2001). “Adrenal Disorders.” Humana Press, Totowa, NJ. Mesiano, S., and Jaffe, R. B. (1997). Development and function of the primate fetal adrenal cortex. Endocr. Rev. 18, 378–404. Mortensen, R. M., and Williams, R. H. (2001). Aldosterone action. In “Endocrinology” (L. J. DeGroot and J. L. Jameson, eds.), 4th ed. Saunders, Philadelphia. Orth, D. N., and Kovacs, W. J. (1998). The adrenal cortex. In “Williams Textbook of Endocrinology” ( J. D. Wilson, D. W. Foster, H. M. Kronenberg, and P. R. Larsen, eds.), 9th ed. Saunders, Philadelphia. Parker, K. L., and Schimmer, B. P. (2001). Genetics of the development and function of the adrenal cortex. Rev. Endocr. Metab. Dis. 2, 245–252.

Adrenal Insufficiency Evangelia Charmandari and George P. Chrousos NIH/National Institute of Child Health and Human Development, Bethesda, Maryland, United States

Glossary Cushing’s syndrome A condition caused by increased adrenocortical secretion of cortisol resulting from adrenocortical hyperplasia or tumor and characterized by obesity, acne, amenorrhea, hypertension, abdominal pain, and weakness. HIV infection Human immunodeficiency virus, the causative agent of AIDS. thrombosis The formation, development, or presence of a thrombus, or blood clot, in a blood vessel or the heart.

A

drenal insufficiency is a disorder characterized by impaired adrenocortical function and decreased production of mineralocorticoids, glucocorticoids, and/or adrenal androgens.

INTRODUCTION Adrenal insufficiency can be caused by diseases affecting the adrenal cortex (primary), the pituitary gland and the secretion of adrenocorticotropic hormone (ACTH) (secondary), or the hypothalamus and the secretion of corticotropic-releasing hormone (CRH) (tertiary). This article provides a brief overview of the etiology, clinical manifestations, diagnosis, and treatment of adrenal insufficiency.

CAUSES OF ADRENAL INSUFFICIENCY Primary Adrenal Insufficiency Autoimmune Adrenalitis This condition is the result of an autoimmune process that destroys the adrenal cortex. Both humoral and cell-mediated immune mechanisms directed at the adrenal cortex are involved. Antibodies that react with several steroidogenic enzymes as well as all three zones of the adrenal cortex are detected in 60–75%

Encyclopedia of Endocrine Diseases, Volume 1. Published by Elsevier Inc.

of patients with autoimmune primary adrenal insufficiency but only rarely in patients with other causes of adrenal insufficiency or normal subjects. Approximately 50% of patients with autoimmune adrenal insufficiency have one or more other autoimmune endocrine disorders, whereas patients with the more common autoimmune endocrine disorders, such as type 1 diabetes mellitus, chronic autoimmune thyroiditis, or Graves’ disease, rarely develop adrenal insufficiency. The combination of autoimmune adrenal insufficiency with other autoimmune endocrine disorders is referred to as the polyglandular autoimmune syndromes types I and II (Table I). Infectious Adrenalitis Many infectious agents may affect the adrenal gland and result in adrenal insufficiency, including tuberculosis, disseminated fungal infections, and HIV infection. Hemorrhagic Infarction Bilateral adrenal infarction caused by hemorrhage or adrenal vein thrombosis may lead to adrenal insufficiency. Adrenal hemorrhage has been mostly associated with meningococcemia (Waterhouse– Friderichsen syndrome) and Pseudomonas aeruginosa infection. Drugs Drugs may cause adrenal insufficiency by inhibiting cortisol biosynthesis, particularly in individuals with limited pituitary and/or adrenal reserve, or by accelerating the metabolism of cortisol and most synthetic glucocorticoids following induction of hepatic mixed function oxygenase enzymes.

Secondary and Tertiary Adrenal Insufficiency Secondary adrenal insufficiency may be caused by any disease process that affects the anterior pituitary and

75

76

Adrenal Insufficiency

Table I Etiology of Adrenal Insufficiency Primary adrenal insufficiency Autoimmune (polyglandular failure) Tuberculosis

whether mineralocorticoid production is preserved. The onset of adrenal insufficiency is often gradual and may go undetected until an illness or other stress precipitates an adrenal crisis.

Sarcoidosis, amyloidosis, hemochromatosis Hemorrhage (meningococcemia, anticoagulants, and trauma) Fungal infections

Adrenal Crisis

Metastatic neoplasia/infiltration

Adrenal crisis or acute adrenal insufficiency may complicate the course of chronic primary adrenal insufficiency and may be precipitated by a serious infection, acute stress, bilateral adrenal infarction, or hemorrhage. It is rare in patients with secondary or tertiary adrenal insufficiency. The main clinical manifestation of adrenal crisis is shock, but patients may also have nonspecific symptoms, such as anorexia, nausea, vomiting, abdominal pain, weakness, fatigue, lethargy, confusion, or coma. Hypoglycemia is rare in acute adrenal insufficiency but more common in secondary adrenal insufficiency. Hypoglycemia is a common manifestation in children and thin women with the disorder. Hyperpigmentation due to chronic ACTH hypersecretion and weight loss are indicative of long-standing adrenal insufficiency, and additional symptoms and signs relating to the primary cause of adrenal insufficiency may also be present. The major factor precipitating an adrenal crisis is mineralocorticoid deficiency and the main clinical problem is hypotension. Adrenal crisis can occur in patients receiving appropriate doses of glucocorticoid if their mineralocorticoid requirements are not met, whereas patients with secondary adrenal insufficiency and normal aldosterone secretion rarely present in adrenal crisis.

Congential adrenal hyperplasia Congenital adrenal hypoplasia Congenital unresponsiveness to ACTH (glucocorticoid deficiency and ACTH resistance) Adrenoleukodystrophy/adrenomyeloneuropathy Acquired immunodeficiency syndrome Bilateral adrenalectomy Steroid synthesis inhibitors (e.g., metyrapone, ketoconazole, and aminoglutethimide) Adrenolytic agents (o,p0 DDD, suramin) Glucocorticoid antagonists (RU486) Secondary and tertiary adrenal insufficiency Following discontinuation of exogenous glucocorticoids or ACTH Following the cure of Cushing’s syndrome Pituitary and hypothalamic lesions Tumors Inflammation Infections Autoimmune lesions Granulomatous infilitration Trauma Congenital aplasia, hypoplasia, dysplasia, ectopy Pituitary–hypothalamic surgery Pituitary–hypothalamic radiation Pituitary–hypothalamic hemorrhage (apoplexy) Acquired isolated ACTH deficiency Familial corticosteroid-binding globulin deficiency

interferes with ACTH secretion. The ACTH deficiency may be isolated or occur in association with other pituitary hormone deficits (Table I). On the other hand, tertiary adrenal insufficiency can be caused by any process that involves the hypothalamus and interferes with CRH secretion. The most common causes of tertiary adrenal insufficiency are abrupt cessation of high-dose glucocorticoid therapy and treatment of Cushing’s syndrome.

CLINICAL MANIFESTATIONS OF ADRENAL INSUFFICIENCY The clinical manifestations of adrenal insufficiency depend on the extent of loss of adrenal function and

Chronic Primary Adrenal Insufficiency Patients with chronic primary adrenal insufficiency may have symptoms and signs of glucocorticoid, mineralocorticoid, and androgen deficiency. In contrast, patients with secondary or tertiary adrenal insufficiency usually have normal mineralocorticoid function. The onset of chronic adrenal insufficiency is often insidious and the diagnosis may be difficult in the early stages of the disease. The most common clinical manifestations of chronic primary adrenal insufficiency include general malaise, fatigue, weakness, anorexia, weight loss, nausea, vomiting, abdominal pain, diarrhea that may alternate with constipation, hypotension, electrolyte abnormalities (hyponatremia, hyperkalemia, or metabolic acidosis), hyperpigmentation, autoimmune manifestations (vitiligo), decreased axillary and pubic hair, and loss of libido and amenorrhea in women.

77

Adrenal Insufficiency

Secondary or Tertiary Adrenal Insufficiency p0065

The clinical features of secondary or tertiary adrenal insufficiency are similar to those of primary adrenal insufficiency. However, hyperpigmentation is not present because ACTH secretion is not increased. Also, because the production of mineralocorticoids by the zona glomerulosa is mostly preserved, dehydration and hyperkalemia are not present, and hypotension is less prominent. Hyponatremia and increased intravascular volume may be the result of an “inappropriate” increase in vasopressin secretion. Hypoglycemia is more common in secondary adrenal insufficiency, possibly due to concomitant growth hormone insufficiency, and in isolated ACTH deficiency. Clinical manifestations of a pituitary or hypothalamic tumor, such as symptoms and signs of deficiency of other anterior pituitary hormones, headache, or visual field defects, may also be present.

DIAGNOSIS The clinical diagnosis of adrenal insufficiency can be confirmed by demonstrating inappropriately low cortisol secretion, determining whether the cortisol deficiency is secondary or primary and, hence, dependent or independent of ACTH deficiency, and detecting the cause of the disorder.

Cortisol Secretion The diagnosis of adrenal insufficiency depends on the demonstration of inappropriately low cortisol secretion. Serum cortisol concentrations are normally highest in the early morning hours (4:00–8:00 am) and increase further with stress. Serum cortisol concentrations of less than 3 mg/dl (80 nmol/liter) at 8:00 am are strongly suggestive of adrenal insufficiency, whereas values less than 10 mg/dl (275 nmol/liter) make the diagnosis likely. Basal urinary cortisol and 17-hydroxycorticosteroid excretion is low in patients with severe adrenal insufficiency but may be lownormal in patients with partial adrenal insufficiency. Generally, baseline urinary measurements are not recommended for the diagnosis of adrenal insufficiency.

ACTH Secretion Inappropriately low serum cortisol concentrations in association with increased plasma ACTH concentrations determined simultaneously are suggestive of

primary adrenal insufficiency. On the other hand, inappropriately low baseline morning cortisol and ACTH concentrations indicate secondary or tertiary disease. Given that plasma ACTH measurements depend on proper preparation of the sample and may not be readily available, confirmation of the diagnosis requires stimulation of the adrenal glands with exogenous ACTH.

Short ACTH Stimulation Tests A short ACTH stimulation test should be performed in all patients suspected of having adrenal insufficiency. It involves the intravenous administration of synthetic ACTH(1–24) (cosyntropin), which has the full biologic potency of native ACTH(1–39), and subsequent measurement of serum cortisol concentrations at regular intervals for up to 1 h (usually at 0, 30, and 60 min). High-Dose ACTH Stimulation Test This test consists of determining serum cortisol responses immediately before and 30 and 60 min after intravenous administration of 250 mg of cosyntropin. This dose of cosyntropin results in pharmacologic plasma ACTH concentrations for the 60-min duration of the test, which may be too high to detect cases of chronic partial and mild pituitary ACTH deficiency. Therefore, this test may miss mild cases of adrenal insufficiency. Also, in early acute secondary or tertiary adrenal insufficiency, as in Sheehan syndrome, the test is not reliable because it takes several days for the adrenal cortex to atrophy. The advantage of the high-dose test is that the cosyntropin can be injected intravenously or intramuscularly since pharmacologic plasma ACTH concentrations can be achieved by either route. An increase in serum cortisol concentration after 30 or 60 min to a peak of 18–20 mg/dl (500–550 nmol/ liter) or more is considered a normal response to the high-dose ACTH stimulation test and excludes the diagnosis of primary adrenal insufficiency and almost all cases of secondary adrenal insufficiency. However, if secondary adrenal insufficiency is of recent onset, the adrenal glands will have not yet atrophied and will still be capable of responding to ACTH stimulation normally. In these cases, a low-dose ACTH test or insulin-induced hypoglycemia may be required to confirm the diagnosis. Low-Dose ACTH Stimulation Test This test theoretically provides a more sensitive index of adrenocortical responsiveness because it

78 results in physiologic plasma ACTH concentrations. It is performed by measuring serum cortisol concentrations immediately before and 30 min after intravenous injection of cosyntropin in a dose of 1.0 mg (160 mIU) per 1.73 m2. This dose stimulates maximal adrenocortical secretion up to 30 min postinjection and, in normal subjects, results in a peak plasma ACTH concentration approximately twice that of insulin-induced hypoglycemia. A value of 18 mg/dl (500 nmol/liter) or more at any time during the test is indicative of normal adrenal function. The advantage of this test is that it can detect partial adrenal insufficiency that may be missed by the standard high-dose test. The low-dose test is also preferred for patients with secondary or tertiary adrenal insufficiency.

Prolonged ACTH Stimulation Tests Prolonged ACTH stimulation tests are rarely performed because the history and physical examination, the CRH test, and/or determination of cortisol and ACTH concentrations in association with the lowdose ACTH test may provide all necessary information. Prolonged stimulation with exogenous ACTH is used to differentiate between primary and secondary or tertiary adrenal insufficiency. In secondary or tertiary adrenal insufficiency, the adrenal glands display cortisol secretory capacity following prolonged stimulation with ACTH, whereas in primary adrenal insufficiency, the adrenal glands are partially or completely destroyed and do not respond to ACTH. Eight-Hour ACTH Stimulation Test This test consists of administering 250 mg (40 IU) of cosyntropin intravenously as an infusion for 8 h and determining serum cortisol and 24-h urinary cortisol and 17-hydroxycorticoid (17-OHCS) concentrations before and after the infusion. In normal subjects, the 24-h urinary 17-OHCS excretion increases three- to fivefold above the baseline. Serum cortisol concentrations reach 20 mg/dl (550 nmol/liter) at 30–60 min and exceed 25 mg/dl (690 nmol/liter) 6–8 h after initiation of the infusion. Two-Day ACTH Stimulation Test The 2-day ACTH stimulation test is similar to the 8h infusion test, except that 250 mg of ACTH(1–24) is infused for more than 24 h on 2 (or 3) consecutive days. This test may be helpful in distinguishing primary from secondary/tertiary adrenal insufficiency. In primary adrenal insufficiency there is no or a minimal response of plasma or urinary cortisol and urinary

Adrenal Insufficiency

17-OHCS. Increases in these values during the 2 or 3 days of the test are indicative of a secondary/tertiary cause of adrenal insufficiency.

CRH Stimulation Test This test is used to differentiate between secondary and tertiary adrenal insufficiency. In both conditions, cortisol levels are low at baseline and remain low after CRH. In patients with secondary adrenal insufficiency, there is little or no ACTH response, whereas in patients with tertiary disease there is an exaggerated and prolonged response of ACTH to CRH stimulation, which is not followed by an appropriate cortisol response. Additional methods to identify the cause of adrenal insufficiency vary depending on whether the disease is primary, secondary, or tertiary.

TREATMENT Adrenal insufficiency is a potentially life-threatening condition. Treatment should be initiated as soon as the diagnosis is confirmed or sooner if the patient presents in adrenal crisis.

Adrenal Crisis Adrenal crisis is a life-threatening emergency that requires immediate treatment. If the diagnosis is suspected but not known, blood samples should be obtained for measurement of cortisol concentrations. Initial Treatment The aim of initial management in adrenal crisis is to treat hypotension (i.e., to correct hypovolemia) and to reverse the electrolyte abnormalities and cortisol deficiency. Large volumes of normal saline solution should be given intravenously. The glucocorticoid deficiency should be treated by immediate intravenous administration of dexamethasone sodium phosphate or hydrocortisone sodium succinate. Dexamethasone may be preferred because it has a long duration of action and does not interfere with the measurements of serum or urinary steroids during subsequent ACTH stimulation tests. After the initial treatment is provided, the cause of the adrenal crisis should be sought and treated. Subsequent Treatment Once the patient’s condition is stable and the diagnosis has been confirmed, parenteral glucocorticoid

79

Adrenal Insufficiency

therapy should be tapered over 3 or 4 days and converted to an oral maintenance dose. Patients with primary adrenal insufficiency require life-long glucocorticoid and mineralocorticoid replacement therapy.

Chronic Adrenal Insufficiency One of the important aspects of the management of chronic primary adrenal insufficiency is patient and family education. Patients should understand the reason for life-long replacement therapy and the need to increase the dose of glucocorticoid during minor or major stress and to inject hydrocortisone, methylprednisolone, or dexamethasone in emergencies.

Emergency Precautions Patients should wear a medical alert (Medic Alert) bracelet or necklace and carry the Emergency Medical Information Card, which should provide information on the diagnosis, medications and daily doses, and the physician involved in the patient’s management. Patients should also have supplies of dexamethasone sodium phosphate and should be educated about how and when to administer them.

Glucocorticoid Replacement Therapy Patients with adrenal insufficiency should be treated with hydrocortisone, the natural glucocorticoid. The hydrocortisone daily dose is 25–30 mg (10–15 mg/m2 body surface area) and can be given in two or three divided doses. A longer acting synthetic glucocorticoid, such as prednisolone, prednisone, or dexamethasone, may be employed but should be avoided because their longer duration of action may produce manifestations of chronic glucocorticoid excess, such as loss of lean body mass and bone density and gain of visceral fat. The usual oral replacement dosages are 5–7.5 mg of prednisolone or prednisone or 0.25–0.75 mg of dexamethasone once daily.

Glucocorticoid Replacement during Minor Illness or Surgery During minor illness or surgical procedures, the dosage of glucocorticoid can be increased up to three times the usual maintenance dosage for 3 days. Depending on the nature and severity of the illness, additional treatment may be required.

Glucocorticoid Replacement during Major Illness or Surgery During major illness or surgery, high doses of glucocorticoid up to 10 times the daily production rate are required to avoid an adrenal crisis. A continuous infusion of 10 mg of hydrocortisone per hour or the equivalent amount of dexamethasone or prednisolone eliminates the possibility of glucocorticoid deficiency. This dose can be halved on postoperative day 2, and the maintenance dose can be resumed on postoperative day 3.

Mineralocorticoid Replacement Therapy Mineralocorticoid replacement therapy is required to prevent sodium loss, intravascular volume depletion, and hyperkalemia. It is given in the form of fludrocortisone (9a-fluorohydrocortisone) in a dose of 0.1 mg daily. The dose of fludrocortisone is titrated individually based on the findings of clinical examination (mainly body weight and arterial blood pressure) and the levels of plasma renin activity. Patients receiving prednisone or dexamethasone may require higher doses of fludrocortisone to lower their plasma renin activity to the upper normal range, whereas patients receiving hydrocortisone, which has some mineralocorticoid activity, may require lower doses. The mineralocorticoid dose may have to be increased during the summer, particularly if patients are exposed to temperatures higher than 29 8C (85 8F).

Androgen Replacement In women, the adrenal cortex is the primary source of androgen in the form of dehydroepiandrosterone and dehydroepiandrosterone sulfate. Although the physiologic role of these androgens in women has not been fully elucidated, their replacement is being increasingly considered in the treatment of adrenal insufficiency.

Chronic Secondary and Tertiary Adrenal Insufficiency Glucocorticoid replacement in chronic secondary or tertiary adrenal insufficiency is similar to that in primary adrenal insufficiency. However, measurement of plasma ACTH concentration cannot be used to titrate the optimal glucocorticoid dose. Mineralocorticoid replacement is rarely required, whereas replacement of other anterior pituitary deficits may be necessary.

80

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Adrenal Androgens . Adrenal Cortex, Anatomy . Adrenal Cortex, Development . Adrenal Cortex Development, Regulation of . Adrenal Cortex, Physiology . Adrenal Suppression

Further Reading Abdu, T. A. M., Elhadd, T. A., Neary, E., and Clayton, R. N. (1999). Comparison of the low dose short synacthen test (1 mg), the conventional dose short synacthen test (250 mg), and the insulin tolerance test for assessment of the hypothalamic–pituitary– adrenal axis in patients with pituitary disease. J. Clin. Endocrinol. Metab. 84, 838. Addison, T. (1855). “On the Constitutional and Local Effects of Disease of the Supra-renal Capsules.” Highley, London. Crowley, S., Hindmarsh, P. C., Honour, J. W., and Brook, C. G. (1993). Reproducibility of the cortisol response to stimulation with a low dose of ACTH(1–24): The effect of basal cortisol levels and comparison of low-dose with high-dose secretory dynamics. J. Endocrinol. 136, 167. Gold, P. W., Kling, M. A., Khan, I., Calabrese, J. R., Kalogeras, K., Post, R. M., Avgerinos, P. C., Loriaux, D. L., and Chrousos, G. P. (1987). Corticotropin releasing hormone: Relevance to normal

Adrenal Insufficiency

physiology and to the pathophysiology and differential diagnosis of hypercortisolism and adrenal insufficiency. Adv. Biochem. Psychopharmacol. 43, 183–200. Longui, C. A., Vottero, A., Harris, A. G., and Chrousos, G. P. (1998). Plasma cortisol responses after intramuscular corticotropin 1–24 in healthy men. Metabolism 47, 1419. Nye, E. J., Grice, J. E., Hockings, G. I., Strakosch, C. R., Crosbie, G. V., Walters, M. M., and Jackson, R. V. (1999). Comparison of adrenocorticotropin (ACTH) stimulation tests and insulin hypoglycemia in normal humans: Low dose, standard high dose, and 8-hour ACTH-(1–24) infusion tests. J. Clin. Endocrinol. Metab. 84(10), 3648–3655. Oelkers, W. (1996). Adrenal insufficiency. N. Engl. J. Med. 335(16), 1206–1212. Schulte, H. M., Chrousos, G. P., Avgerinos, P., Oldfield, E. H., Gold, P. W., Cutler, G. B., Jr., and Loriaux, D. L. (1984). The corticotropin-releasing hormone stimulation test: A possible aid in the evaluation of patients with adrenal insufficiency. J. Clin. Endocrinol. Metab. 58, 1064. Stewart, P. M. (2003). The adrenal cortex. In “Williams Textbook of Endocrinology” (P. R. Larsen, H. M. Kronenberg, S. Melmed, and K. S. Polonsky, eds.), pp. 525–532. Saunders, Philadelphia. Thaler, L. M., and Blevins, L. S., Jr. (1998). The low dose (1-mg) adrenocorticotropin stimulation test in the evaluation of patients with suspected central adrenal insufficiency. J. Clin. Endocrinol. Metab. 83, 2726.

Adrenal Suppression Maria Alexandra Magiakou Athens University Medical School, Athens, Greece

George P. Chrousos NIH/National Institute of Child Health and Human Development, Bethesda, Maryland, United States

Glossary g9000

g9005

g9010

g9015

ACTH (adrenocorticotropic hormone) A hormone synthesized and secreted by the anterior pituitary that is primarily responsible for the regulation of cortisol and adrenal androgen secretion. adrenal crisis A medical emergency in which the subject is affected by an extreme state of adrenal insufficiency, with symptoms resembling shock or coma; considered to be life-threatening and requiring immediate administration of fluids, electrolytes, glucose, and intravenous glucocorticoids. May occur in situations of extreme stress or after the abrupt cessation of a period of glucocorticoid therapy. Cushing’s syndrome A metabolic disorder relating to the chronic oversecretion of adrenal glucocorticoids (mainly cortisol); may be associated with the long-term administration of pharmacological amounts of synthetic glucocorticoids. A condition first described by U.S. neurosurgeon Harvey Cushing in 1912. HPA axis An anatomical complex consisting of the hypothalamus, the pituitary gland, and the adrenal cortex; ultimately controls the secretion of glucocorticoids from the adrenal cortex and thus plays a key role in various neuroendocrine, behavioral, autonomic, and immune responses to alterations in homeostasis.

G

lucocorticoids are produced by the cortices of the adrenal glands and secreted into the systemic circulation in a circadian fashion and in response to stressful stimuli. These steroid hormones play pivotal roles in the regulation of intermediary metabolism, maintenance of cardiovascular function, stimulation of behavior, and control of the immune inflammatory reaction.

been called hydrocortisone. Cortisone, the 2-keto form of cortisol, was first used therapeutically in the management of rheumatoid arthritis by Hench and coworkers in 1949. Since then, a large number of synthetic compounds with glucocorticoid activity have been developed, and glucocorticoids (administered systemically or in a compartmental fashion) have been used in the therapy of a broad spectrum of nonendocrine and endocrine diseases. One of the adverse effects of long-term glucocorticoid therapy in supraphysiologic doses is suppression of the hypothalamic–pituitary–adrenal (HPA) axis, which can render the adrenal glands unable to generate sufficient cortisol if glucocorticoid treatment is abruptly stopped, and the patient may develop glucocorticoid deficiency manifestations. The true prevalence of clinically significant adrenal insufficiency is not known since physicians usually discontinue high glucocorticoids gradually to allow recovery of the HPA axis. Some of the risk factors for HPA axis suppression are clearly defined, whereas others are less certain. Systemic glucocorticoid therapy is more likely to suppress the HPA axis than compartmentalized use of glucocorticoids, with the possible exception of intra-articular steroids. Systemic glucocorticoid potency is also known to correlate with risk for adrenal insufficiency. Glucocorticoid treatment in endocrine and nonendocrine disorders, the side effects of these medications, their concomitant use and interactions with other drugs, adrenal suppression, and the glucocorticoid withdrawal syndrome are discussed in detail in this article.

PHYSIOLOGY OF THE HPA AXIS INTRODUCTION The major endogenous glucocorticoid in humans is cortisol, the synthetic form of which has traditionally

Encyclopedia of Endocrine Diseases, Volume 1. Published by Elsevier Inc.

The adrenal cortex consists of three anatomic zones: the outer zona glomerulosa, the intermediate zona fasciculata, and the inner zona reticularis. The zona

81

82 glomerulosa is responsible for the production of aldosterone, the zona fasciculata for the production of cortisol, and the zona reticularis for the production of adrenal androgens. Corticotropin (ACTH), synthesized and secreted by the corticotropes of the anterior pituitary, is the primary regulator of cortisol and adrenal androgen secretion. Hypothalamic control of ACTH secretion is exerted primarily by corticotropin-releasing hormone (CRH), a 41-amino acid peptide produced by parvocellular neurons of the paraventricular nucleus and secreted into the hypophyseal portal system. There are several regulatory negative feedback loops that function to constrain the activity of the HPA axis. Prominent negative feedback loops are those exerted by glucocorticoids on CRH and ACTH secretion. The adrenal cortisol secretion rate under basal conditions is 12–15 mg/m2/day. In normal individuals, the highest plasma cortisol levels occur between 6:00 and 8:00 am and the lowest at approximately midnight. Cortisol secretion increases two- to fourfold under stress. Plasma cortisol concentrations are elevated during physical and/or emotional stress, including illness, trauma, surgery, and starvation.

ADRENAL INSUFFICIENCY Adrenal insufficiency results from inadequate adrenocortical function, which may be due to destruction of the adrenal cortex (primary adrenal insufficiency; Addison’s disease), deficient pituitary ACTH secretion (secondary adrenal insufficiency), or deficient hypothalamic secretion of CRH or other ACTH secretagogues (tertiary adrenal insufficiency). Primary and secondary adrenal insufficiency related to natural causes are uncommon, whereas iatrogenic, tertiary adrenal insufficiency caused by suppression of HPA function by glucocorticoid administration is common.

PATHOGENESIS OF GLUCOCORTICOID-INDUCED ADRENAL SUPPRESSION Glucocorticoid treatment may not suppress the HPA axis at all, or it may cause central suppression or complete adrenal gland atrophy. Supraphysiologic glucocorticoid doses inhibit both CRH production in the hypothalamus and ACTH production in the pituitary gland. When this inhibition lasts longer than the duration of the glucocorticoid exposure, it is called adrenal suppression.

Adrenal Suppression

The most severely affected glucocorticoid-treated patients have complete HPA axis suppression, characterized by functional adrenal gland atrophy. Cortisol production seems to depend on intermittent but consistent exposure to circulating ACTH. In states of profound or prolonged ACTH deficiency, the adrenal glands atrophy and are unable to generate cortisol in response to exogenous ACTH. However, in reality, adrenal suppression is a central nervous system problem. The rate-limiting step in HPA axis recovery from suppression is located in the brain. Thus, chronic administration of ACTH does not accelerate recovery, even though it may return cortisol production to normal while administered.

SYNTHETIC GLUCOCORTICOIDS Since the introduction of glucocorticoids in the treatment of rheumatoid arthritis in 1949, intense efforts have been made by science and industry to maximize the beneficial effects and to minimize the side effects of glucocorticoids. Thus, many synthetic compounds with glucocorticoid activity have been manufactured and tested. The pharmacologic differences among these chemicals result from structure alterations of their basic steroid nucleus and its side groups. These changes may affect the bioavailability of these compounds (including their gastrointestinal or parenteral absorption, plasma half-life, and metabolism in the liver, fat, or target tissues) and their abilities to interact with the glucocorticoid receptor and to modulate the transcription of glucocorticoid-responsive genes. In addition, structural modifications diminish the natural cross-reactivity of glucocorticoids with the mineralocorticoid receptor, eliminating their undesirable salt-retaining activity. Other modifications increase glucocorticoids’ water solubility for parenteral administration or decrease their water solubility to enhance topical potency. Thus, prednisolone has the structure of cortisol with an additional double bond between C1 and C2, which increases its glucocorticoid and decreases its mineralocorticoid activity. Introduction of an a-fluoro group at C9, on the other hand, enhances both glucocorticoid and mineralocorticoid activity, whereas addition of a hydroxyl or methyl group at C16 practically eliminates mineralocorticoid activity. Dexamethasone, a potent synthetic glucocorticoid that has a double bond at C1and C2, a fluoro group at C9, and an a-methyl group at C16, has 25–50 times the glucocorticoid potency of cortisol and a minimal mineralocorticoid effect. A double bond between C2

83

Adrenal Suppression

and C3 and methylation at the C2 or C16 positions significantly prolong the plasma half-life of a compound. A keto group at C11 is normally reduced by liver enzymes to an 11b-hydroxyl group, which is necessary for glucocorticoid activity. In contrast, a C11 hydroxyl group is oxidized to a keto group in the kidney, minimizing the access of the compound to the mineralocorticoid receptor, and thus its salt-retaining effect. Most synthetic glucocorticoids (e.g., methylprednisolone and dexamethasone) are minimally bound to cortisol-binding globulin (transcortin) and circulate mostly bound to albumin or in the free form. The percentage of such glucocorticoids bound to plasma proteins is relatively constant, and because binding is concentration independent, the metabolic clearance rate of glucocorticoids remains constant regardless of dose. Table I shows the relative glucocorticoid and mineralocorticoid potencies of commonly used systemic glucocorticoids and their approximate plasma and biologic effect half-lives. Glucocorticoid activity has been mostly defined in rat bioassays and may not always pertain to human responses, especially the growth-suppressing properties of synthetic glucocorticoids, which have been markedly underestimated. Based on the biologic effect half-life of glucocorticoids, they are classified as short, intermediate, or long acting based on the duration of corticotropin suppression after a single dose of the compound.

t0005

SYSTEMIC GLUCOCORTICOID ADMINISTRATION Therapeutic Indications Glucocorticoids may be administered as replacement therapy in patients with primary or secondary adrenal insufficiency, as adrenal suppression therapy in congenital adrenal hyperplasia and glucocorticoid resistance, and as anti-inflammatory or immunosuppressant therapy in a broad range of mostly nonendocrine disorders affecting many different systems. Thus, glucocorticoids are used in endocrine, autoimmune, collagen, renal, gastrointestinal, respiratory, nervous, hematologic, and ophthalmic diseases and are used in the suppression of the host versus graft and graft versus host reaction in cases of organ transplantation. Neoplastic disorders of the lymphoid system, such as leukemia and lymphomas, are also treated with glucocorticoids, along with the appropriate chemotherapy. Acute administration of pharmacologic doses of glucocorticoids is necessary in a small number of nonendocrine diseases, such as malignant hyperthermia, and in patients with craniospinal trauma or brain tumors or in those who are undergoing major neurosurgical operations to decrease the temperature and prevent destruction of neural tissue from the local edema and inflammatory reaction, respectively. In addition, glucocorticoids have been used in the

Table I Glucocorticoid Equivalenciesa Equivalent dose (mg)

Glucocorticoid potency

Mineralocorticoid potency

Plasma half-life (min)

Biologic half-life (h)

Glucocorticoids Short acting Cortisol

20.0

1.0

2

90

8–12

Cortisone

25.0

0.8

2

80–118

8–12

Prednisone

5.0

4.0

1

Prednisone

5.0

4.0

1

Triamcinolone

4.0

5.0

0

30

18–36

Methylprednisolone

4.0

5.0

0

180

18–36

0.5

25–50

0

200

36–54

0.6

25–50

0

300

36–54

Intermediate acting

Long acting Dexamethasone Betamethasone

60 115–200

18–36 18–36

Mineralocorticoids Aldosterone Fluorocortisone Desoxycorticosterone acetate a

From Liapi and Chrousos (1992).

— 2.0 —

0.3

300

15.0

150

200

15–20

0.0

20

70

8–12 18–36 —

84 prevention of the respiratory distress syndrome in the premature neonate, when delivery is anticipated before week 34 of gestation. In this case, treatment of the pregnant woman with 12 mg of betamethasone, followed by 12 mg 18–24 h later, stimulates the production of pulmonary surfactant and the maturation of the fetal lungs. Recently, glucocorticoid therapy has been reintroduced in the treatment of adult acute respiratory distress syndrome. The recommended methylprednisolone doses are moderate and treatment is given continuously during the course of the disease. Significantly improved morbidity and mortality have been observed. Similar glucocorticoid treatment is being studied in patients with systemic inflammation syndrome, septic shock, and multiple organ dysfunction with promising results.

Adrenal Suppression

Table II Effects of Long-Term Glucocorticoid Therapya Endocrine and metabolic Suppression of HPA axis (adrenal suppression) Growth failure in children Carbohydrate intolerance Hyperinsulinemia Insulin resistance Abnormal glucose tolerance test Diabetes mellitus Cushingoid features Moon facies, facial plethora Generalized and truncal obesity Supraclavicular fat collection Posterior cervical fat deposition (buffalo hump) Glucocorticoid-induced acne Thin and fragile skin, violaceous striae Impotence, menstrual disorders Decreased thyroid-stimulating hormone and triiodothyronine

Side Effects Side effects occur only with supraphysiologic doses of glucocorticoids and not with proper replacement, which is equivalent to 12–15 mg of hydrocortisone/ m2 body surface area/day. Major complications are unlikely for short-term treatment (< 2 weeks) with high doses of glucocorticoids, although sleep disturbances and gastric irritation are common complaints, and depression, mania, or psychosis may be infrequently precipitated. On the other hand, many side effects are associated with long-term daily administration of pharmacologic amounts of glucocorticoids (Table II), including the development of varying degrees of Cushing’s syndrome manifestations during therapy and secondary adrenal insufficiency (adrenal suppression) after discontinuation of treatment. Growth retardation is one of the major side effects of long-term daily glucocorticoid therapy in children. The degree of cushingoid features and the severity and length of adrenal suppression depend on the type and dose of the specific compound used, the duration of treatment, the idiosyncrasy, and the stress status of the patient. Most complications of glucocorticoid treatment are totally or partially reversible after discontinuation of glucocorticoid administration, with the exception of posterior subcapsular cataracts and advanced bone necrosis. High doses of glucocorticoids suppress the immune defenses of the organism. Thus, individuals on glucocorticoid therapy are particularly susceptible to viral diseases against which they have not been vaccinated or naturally immunized (e.g., varicella, which can be devastating). Also, such individuals are

Hypokalemia, metabolic alkalosis Gastrointestinal system Gastric irritation, peptic ulcer Acute pancreatitis (rare) Fatty infiltration of liver (hepatomegaly) (rare) Hemopoietic system Leukocytosis Neutrophilia Increased influx from bone marrow and decreased migration from blood vessels Monocytopenia Lymphopenia Migration from blood vessels to lymphoid tissue Eosinopenia Immune system Suppression of delayed hypersensitivity Inhibition of leukocyte and tissue macrophage migration Inhibition of cytokine secretion or action Suppression of the primary antigen response Musculoskeletal system Osteoporosis, spontaneous fractures Aseptic necrosis of femoral and humoral heads and other bones Myopathy Ophthalmic Posterior subcapsular cataracts (more common in children) Elevated intraocular pressure or glaucoma Neuropsychiatric disorders Sleep disturbances, insomnia Euphoria, depression, mania, psychosis Pseudotumor cerebri (benign increase of intracranial pressure) a

From Laue et al. (1989).

t0010

85

Adrenal Suppression

susceptible to contract or to sustain activation of dormant tuberculosis. Individuals treated with massive doses of glucocorticoids can also develop saprophytic, fungal, or protozoan infections such as those seen in patients with severe immunodeficiency. Nonfluorinated glucocorticoids (cortisone, cortisol, prednisone, and prednisolone) cross the placenta poorly. Fluorinated steroids, on the other hand, cross the placenta readily and should be given cautiously to women during pregnancy. The ratio of maternal–fetal plasma concentration gradients is approximately 10:1 for cortisol and prednisolone and approximately 2.5:1 for betamethasone and dexamethasone. Newborns exposed to high doses of synthetic fluorinated corticosteroids in utero should be checked for signs of adrenal insufficiency, and a Cortrosyn stimulation test should be performed to assess the need for glucocorticoid replacement. Special precautions should be taken for premature infants because some glucocorticoid preparations containing benzyl alcohol have been associated with a fatal “gasping” syndrome. To avoid complications, alternate-day administration of intermediate-acting glucocorticoids should be used, when possible, if long-term therapy is necessary. Frequently, such a regimen can control the activity of the disease under therapy without causing Cushing’s syndrome, growth retardation, or adrenal suppression. Termination of long-term daily therapy (> 2 weeks) should be gradual to prevent development of acute adrenal insufficiency and to avoid reactivation of the disease under therapy. Switching to daily hydrocortisone replacement or to alternate-day administration of intermediate-acting glucocorticoids is an acceptable method for weaning patients from glucocorticoid therapy.

COMPARTMENTAL GLUCOCORTICOID ADMINISTRATION Topical Glucocorticoids Glucocorticoids are quite effective when applied topically and are nontoxic to the skin in the short term. The factors that determine local penetration are the structure of the compound employed, the vehicle, the basic additives, occlusion versus open use, normal skin versus diseased skin, and small areas versus large areas of application. Fluorinated steroids (e.g., dexamethasone, triamcinolone acetonide, betamethasone, and beclomethasone) penetrate the skin better than nonfluorinated steroids, such as hydrocortisone. However, fluorinated steroids also cause more local

complications and may be associated with systemic absorption and side effects. The complications of chronic topical skin use of glucocorticoids are mostly local (e.g., epidermal atrophy and hypopigmentation, telangiectasia, or acne and folliculitis) or infrequently systemic, with the classic manifestations of Cushing’s syndrome, growth retardation in children, and adrenal suppression. The frequency of systemic effects by topical corticosteroids is increased in newborns and small children compared to adolescents and adults because glucocorticoids penetrate the skin of newborns and small children more easily and in larger amounts. Systemic effects may also be observed in patients with hepatic disease or idiosyncratically because of decreased drug metabolism. Although most types of dermatitis are generally responsive to topical glucocorticoids, there are rare cases in which intralesional injections might be considered (e.g., hypertrophic scars, acne cysts, or prurigo nodularis).

Ophthalmic Glucocorticoids Patients with autoimmune or idiopathic inflammation of the anterior segment of the eye (e.g., iritis and uveitis) may benefit from local administration of glucocorticoids. Also, patients with postsurgical or traumatic inflammation are given topical glucocorticoids to prevent local destruction from edema. Special care should be taken to avoid treating patients with herpes simplex conjunctivitis or keratitis during the infectious stage of the disease because major spread of the infection may be precipitated.

Inhaled Glucocorticoids Glucocorticoid inhalation therapy is widely used in patients with bronchial asthma and croup. The existing preparations at the recommended doses have a remarkable therapeutic effect without causing manifestations of Cushing’s syndrome, growth retardation, or clinically significant adrenal suppression. Systemic effects may be observed, however, as a result of increased intake of such preparations or altered steroid metabolism. Inhaled glucocorticoids have also been used in ventilator-dependent preterm infants to reduce the severity of respiratory distress syndrome and to facilitate weaning from mechanical ventilation of infants with bronchopulmonary dysplasia. Their effect on the function of the immature HPA axis of these preterm neonates is unclear.

86

Adrenal Suppression

Nasal Glucocorticoids Aerosols containing glucocorticoids are available for the treatment of allergic rhinitis. Frequent and chronic use should be avoided to prevent local and systemic complications.

Intra-articular Glucocorticoids The intra-articular injection of glucocorticoids may be of value in carefully selected patients if strict aseptic techniques are used and repeated and frequent injections are avoided.

CONCOMITANT USE OF GLUCOCORTICOIDS WITH OTHER DRUGS Special attention is required for the concomitant use of glucocorticoids with other drugs because of potential interactions and because some drugs may affect the metabolism of the steroids, which may lead to a decreased or increased glucocorticoid effect on their target tissues. Such interactions and effects are shown in Tables III–V.

PREDICTING GLUCOCORTICOIDINDUCED HPA AXIS SUPPRESSION Several predictors of glucocorticoid-induced HPA axis suppression have been discussed. The following are the most important:

MONITORING OF PATIENTS RECEIVING GLUCOCORTICOID TREATMENT Patients receiving long-term treatment with glucocorticoids should adhere to a high-protein, calorierestricted diet. The diet should also be rich in potassium and calcium and low in sodium. Adequate ambulation or exercise should be recommended to prevent muscular atrophy and osteopenia. Patients should concurrently take antacids or histamine antagonists to prevent gastric irritation or peptic ulcers. Young children should have their growth monitored every 3 months (until age 5), and older children should have their growth monitored every 6 months. For all patients, body weight, length or height, blood pressure, fasting and 2-hpostprandial blood glucose, serum electrolytes, and bone maturation and density should be measured. Because glucocorticoids decrease the organism’s response to infection, care should be taken to determine whether latent infections, such as mycobacterial disease, are present before treatment begins.

1. Kind of steroid used and glucocorticoid potency: The synthetic analogs of glucocorticoids are much better tolerated as anti-inflammatory agents because they cause significantly less sodium retention at supraphysiologic doses than hydrocortisone and cortisone acetate. Glucocorticoid potency (Table I) correlates positively with risk for adrenal insufficiency. Thus, hydrocortisone and cortisone acetate are the least potent and, therefore, least suppressive agents. Prednisone, prednisolone, methylprednisolone, and triamcinolone are moderately suppressive, and dexamethasone suppresses ACTH the longest. 2. Systemic vs compartmental therapy: Systemic glucocorticoid therapy is more likely to suppress the HPA axis than are intra-articular, inhalational, or topical glucocorticoids. 3. Alternate-day therapy: There is evidence that patients are at lower risk for adrenal insufficiency if they can take glucocorticoids on alternate days from the outset or if they can convert to alternate-day therapy before the HPA axis is suppressed.

Table III Interactions of Glucocorticoids with Other Drugsa Drug

Side effect

Comment

Amphotericin B

Hypokalemia

Monitor potassium levels frequently

Digitalis glycosides

Digitalis toxicity

Monitor potassium levels frequently

Growth hormone

Ineffective

—

Potassium-depleting diuretics

Hypokalemia

Monitor potassium levels frequently

Vaccines from live attenuated viruses

Severe generalized infections

—

Hypokalemia

a

From Liapi and Chrousos (1992).

87

Adrenal Suppression

Table IV

Effects of Glucocorticoids on Blood Levels of Other Drugsa

Drug

Drug blood levels

Comments

Aspirin

Decreased

Increased metabolism or clearance; monitor salicylate levels

Coumarin anticoagulants

Decreased

Frequent control of prothrombin levels

Cyclophosphamide

Increased

Inhibition of hepatic metabolism; adjust the dosage of the drug

Cyclosporine

Increased

Inhibition of hepatic metabolism

Insulin

Decreased

Adjust the dosage of the drug

Isoniazid

Decreased

Increased metabolism and clearance

Oral hypoglycemic agents

Decreased

Adjust the dosage of the drug

a

From Liapi and Chrousos (1992).

4. Once-a-day dosing in the morning or mimicking normal diurnal cortisol rhythms: Since evening doses of glucocorticoids tend to suppress the normal early morning surge of ACTH secretion, it is better, whenever possible, to cure patients with a single morning dose. Once-a-day dosing is usually feasible for prednisone, triamcinolone, and dexamethasone. The short-acting hydrocortisone and cortisone acetate are usually given twice a day, at waking and at approximately 5 pm. To mimic normal diurnal cortisol rhythms, the morning dose is two-thirds and the afternoon dose is one-third of the total daily dose. 5. Duration and cumulative dose of glucocorticoid treatment: Although traditionally the duration of glucocorticoid therapy and the cumulative dose of glucocorticoid received have been considered as predictive of the likelihood of HPA axis suppression, several studies suggest that they only roughly predict HPA axis suppression. Adrenal insufficiency is extremely rare in patients treated for 1 week or less. Table V

Perhaps the best predictor of HPA axis suppression is the patient’s current glucocorticoid dosage. A strong correlation has been found between prednisone maintenance doses >5 mg/day and a subnormal ACTH stimulation test result.

WEANING PATIENTS FROM GLUCOCORTICOID THERAPY Termination of long-term daily glucocorticoid therapy (>2 weeks) should be gradual to prevent development of adrenal insufficiency and to avoid reactivation of the disease under therapy. The likelihood of the latter depends on the activity and natural history of the disorder. When there is a chance that the underlying illness may recur, the glucocorticoids should be withdrawn slowly over a period of weeks to months, with frequent reassessment of the patient’s condition. Daily hydrocortisone replacement or double or triple replacement of intermediate-acting glucocorticoids

Effect of Drugs on Plasma Glucocorticoid Concentrationsa

Drug

Glucocorticoid blood levels

Comments

Antacids

Decreased

Possible physical adsorption to antacid

Carbamazepine

Decreased

Increased cytochrome P450 activity

Cholestyramine Colestipol

Decreased Decreased

Decreased gastrointestinal absorption of glucocorticoids Decreased gastrointestinal absorption of glucocorticoids

Cyclosporine

Increased

Inhibition of hepatic metabolism

Ephedrine

Decreased

Probably increased metabolism

Erythromycin

Increased

Impaired elimination

Mitotane

Decreased, with elevated transcortin

Total plasma cortisol unreliable; adjust glucocorticoid levels

Oral contraceptives

Increased

Impaired elimination, increased protein binding

Phenobarbital

Decreased

Increased cytochrome P450 activity; adjust glucocorticoid dosage

Phenytoin Rifampin

Decreased Decreased

Increased cytochrome P450 activity; adjust glucocorticoid dosage Increased cytochrome P450 activity(?); adjust glucocorticoid dosage

Troleandomycin

Increased

Partially resulting from impaired elimination

a

From Liapi and Chrousos (1992).

88 given on alternate days are acceptable methods for weaning patients from glucocorticoid therapy.

ACUTE ADRENAL CRISIS Recovery of the HPA axis can take 12 months or longer. Abrupt cessation of glucocorticoid treatment or quick tapering can precipitate an acute adrenal insufficiency crisis. The main symptoms range from anorexia, fatigue, nausea, vomiting, dyspnea, fever, arthralgia, myalgia, and orthostatic hypotension to dizziness, fainting, and circulatory collapse. Hypoglycemia is occasionally observed in children and very thin adults. The diagnosis is a medical emergency, and treatment should consist of immediate administration of fluids, electrolytes, glucose, and parenteral glucocorticoids.

GLUCOCORTICOID WITHDRAWAL SYNDROME Glucocorticoid withdrawal can present as an acute adrenal crisis or with symptoms of chronic glucocorticoid deficiency. Thus, patients may suffer from anorexia, myalgia, nausea, emesis, lethargy, headache, fever, skin desquamation, arthralgias, weight loss, and postural hypotension. In addition, they may experience exacerbation of a previously present autoimmune disease (e.g., rheumatoid arthritis, atopic dermatitis, and asthma) or develop a new autoimmune disease (e.g., Hashimoto’s thyroiditis and Graves’ disease). Amatruda et al. first defined the steroid withdrawal syndrome as a symptom complex resembling true adrenal insufficiency, with nonspecific symptoms such as weakness, nausea, and arthralgias, occurring in patients who have completed a dosage reduction of glucocorticoid therapy and who respond normally to HPA axis testing. The occurrence of the subjective component of the steroid withdrawal syndrome does not depend on the absence of cortisol from the circulation or an impairment of the HPA axis because these symptoms may occur while the patient is on proper glucocorticoid replacement or when the patient has a normal cortisol response to Cortrosyn. In this instance, the steroid withdrawal syndrome may be a result of difficulties in withdrawing from the high levels of glucocorticoids—a phenomenon that appears to be idiosyncratic. However, when patients become ill after a dosage reduction, the physician should consider a differential diagnosis that includes true adrenal insufficiency, a

Adrenal Suppression

flare of the disease being treated, and steroid withdrawal syndrome. All three conditions resolve after patients are restarted on the glucocorticoid regimen that previously controlled their symptoms.

BIOCHEMICAL DIAGNOSIS OF ADRENAL INSUFFICIENCY As previously mentioned, glucocorticoid treatment may not suppress the HPA axis at all, or it may cause central suppression and adrenal gland atrophy of varying degrees. The insulin tolerance test and the metyrapone test have been employed in the diagnosis of adrenal suppression and are quite sensitive. However, the risks involved with both tests do not justify their use when a rapid ACTH stimulation test can distinguish clinically significant adrenal suppression. To evaluate the adequacy of HPA axis recovery, the rapid Cortrosyn (or high-dose ACTH stimulation test) is most commonly used. An intravenous bolus of 250 mg of corticotropin 1-24 is administered, and cortisol is measured after 30 or 60 min or both. A plasma cortisol concentration > 18–20 mg/dL at these times indicates adequate recovery of the HPA axis. This test can also be done intramuscularly. A modified Cortrosyn test has been recommended in lieu of the standard test. Only 1 mg of corticotropin 1-24 is administered instead of 250 mg. This test is fraught with technical errors as a result of multiple dilutions of the Cortrosyn preparations and adhesion of Cortrosyn to the tubing system. Its purported increased sensitivity may not necessarily signify a greater clinical prediction of adrenal suppression.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Adrenal Cortex, Anatomy . Adrenal Cortex, Development . Adrenal Insufficiency . Corticotropin-Releasing Hormone (CRH) and Inflammation . Glucocorticoids, Overview

Further Reading Ackerman, G. L., and Nolan, C. M. (1968). Adreno cortical responsiveness after alternate corticosteroid therapy. N. Engl. J. Med. 278, 405–409. Axelrod, L. (1976). Glucocorticoid therapy. Medicine 55, 39–65. Boumpas, D. T., Chrousos, G. P., Wilder, R. L., et al. (1993). NIH conference of the combined clinical staff. Glucocorticoid therapy of immune related diseases: Basic and clinical correlates. Ann. Intern. Med. 119, 1198–1208. Christy, N. P. (1992). Pituitary adrenal function during corticosteroid therapy: Learning to live with uncertainty. N. Engl. J. Med. 326, 266–267.

Adrenal Suppression

Chrousos, G. P., and Harris, A. G. (1998). Hypothalamic pituitary adrenal axis suppression and inhaled corticosteroid therapy. Part I. General principles. Neuroimmunomodulation 5, 287. Chrousos, G. P., and Harris, A. G. (1998). Hypothalamic pituitary adrenal axis suppression and inhaled corticosteroid therapy. Part II. Review of the literature. Neuroimmunomodulation 5, 288–308. Gomez, M. T., Magiakou, M. A., Mastorakos, G., and Chrousos, G. P. (1993). The pituitary corticotroph is not the rate limiting step in the postoperative recovery of the hypothalamic pituitary adrenal axis in patients with Cushing’s syndrome. J. Clin. Endocrinol. Metab. 77, 173. Krasner, A. S. (1999). Glucocorticoid induced adrenal insufficiency. J. Am. Med. Assoc. 282, 671. Laue, L., Kawai, S., Udelsman, R., et al. (1989). Glucocorticoid antagonists: Pharmacological attributes of the prototype antiglucocorticoid RU 486. In “Antiinflammatory Steroid Action: Basic and Clinical Aspects” (L. M. Lichtenstein, H. Claman, and A. Oronsky, eds.), pp. 303–329. Academic Press, New York.

89 Liapi, C., and Chrousos, G. P. (1992). Glucocorticoids. In “Pediatric Pharmacology” (S. J. Jaffe, and J. V. Aranda, eds.), 2nd ed., pp. 466–475. Saunders, Philadelphia. Magiakou, M. A., and Chrousos, G. P. (1994). Corticosteroid therapy, nonendocrine disease and corticosteroid withdrawal. In “Current Therapy in Endocrinology and Metabolism” (C. W. Bardin, ed.), 5th ed., pp. 120–124. Mosby Yearbook, St. Louis. Magiakou, M. A., and Chrousos, G. P. (1996). Corticosteroid therapy and withdrawal. In “Endocrinology and Metabolic Diseases, Current Practice of Medicine,” pp. 6.1–6.6. Palatino Helvetica, Philadelphia. Orth, D. N. (1994). Adrenal insufficiency. In “Current Therapy in Endocrinology and Metabolism” (C. W. Bardin, ed.), 5th ed., pp. 124–130. Mosby Yearbook, St. Louis. Rimsza, M. E. (1978). Complications of corticosteroid therapy. Am. J. Dis. Child. 132, 806. Tyrell, J. B., and Baxter, J. D. (1987). Glucocorticoid therapy. In “Endocrinology and Metabolism” (P. Felig, J. B. Baxter, and A. E. Broadus, eds.), 2nd ed., pp. 788–817. McGraw-Hill, New York.

Adrenal Tumors, Molecular Pathogenesis Christian A. Koch NIH/National Institute of Child Health and Human Development and University of Leipzig, Leipzig, Germany

George P. Chrousos NIH/National Institute of Child Health and Human Development , Bethesda, Maryland, United States

Glossary adrenocorticotropic hormone or corticotropin (ACTH) Stimulates the adrenal gland. first hit According to Knudson’s two-hit model of tumorigenesis, in tumor suppressor gene-related tumors, the first hit is the germ-line mutation of the defective tumor suppressor gene. The second hit consists most commonly of a deletion of the wild-type allele of the tumor suppressor gene in question and can cause biallelic inactivation of this gene with subsequent tumor formation in the affected cell and organ. hyperaldosteronism State of aldosterone excess, The most common cause in primary hyperaldosteronism is an aldosterone-oversecreting adrenocortical mass that can lead to secondary hypertension in the affected patient. incidentaloma Mass (e.g., in the adrenal gland) detected by chance during radiologic imaging workup. loss of heterozygosity (LOH) This can occur through deletion of the wild-type allele of a gene that has a germ-line mutation in hereditary tumor syndromes or through deletion of a gene that has a somatic mutation in sporadic tumors. This will move the affected cell into the homozygous state for the respective gene.

A

drenal tumors are found in up to 9% of autopsy studies. Modern imaging modalities facilitate better detection of adrenal masses than in the past but also detect adrenal incidentalomas in patients who are not evaluated for adrenal tumors. Assessing such patients for the presence of subclinical disease caused by adrenal incidentalomas and the potential for malignancy is challenging. Although there are radiological (e.g., intratumor necrosis, irregular margins, and local metastases), histopathological (e.g., Weiss criteria), and genetic (e.g., overexpression of IGF2 and loss of heterozygosity at 17p13) criteria to predict malignancy, no absolute specific marker and feature exist to reliably distinguish benign and malignant adrenal lesions, excluding pure adrenal ganglioneuroma, which by definition is

90

always benign. Although adrenal tumors have been observed in several familial syndromes, the majority of these lesions occur sporadically. If the gene defect in familial syndromes is known, timely identification of family members already affected by or at risk for the development of an adrenal tumor is possible by genetic screening tools including germ-line mutation analysis. Despite these molecular advances and tools, the pathogenesis of adrenal tumors remains widely unknown. The elucidation of adrenal tumorigenesis may be facilitated by studying adrenal tumors of patients with hereditary syndromes since in these tumors at least one hit, the inherited gene defect, is known and presumably represents the “first hit” in tumor evolution. In contrast, sporadic tumors have an unknown first hit, making it more difficult to determine the sequence of genetic events or hits. In this article, we discuss the molecular pathogenesis in hereditary and sporadic adrenal tumors.

ADRENOMEDULLARY TUMORS Pheochromocytoma Von Hippel–Lindau Syndrome Von Hippel–Lindau (VHL) syndrome consists of a variety of masses, including renal carcinomas, hemangioblastomas, and pheochromocytomas. It affects approximately 1 in 36,000 individuals and is caused by mutations in the VHL tumor suppressor gene located at chromosome 3p25–26. Less than 26% of patients with a VHL germ-line mutation develop a pheochromocytoma, and up to one-third of these patients do not suffer from symptoms of catecholamine excess since they may have “silent” (not catecholamineoversecreting) pheochromocytomas. Most VHLassociated pheochromocytomas have loss of function (LOH) at 3p25–26, leading to biallelic inactivation of the VHL gene, thereby following Knudson’s two-hit model of tumorigenesis. A small subset of VHL pheochromocytomas have LOH at 1p or chromosome 11. VHL protein leads to degradation of certain proteins, most in the 26S proteasome complex (Table I).

Encyclopedia of Endocrine Diseases, Volume 1. Published by Elsevier Inc.

p0010

91

Adrenal Tumors, Molecular Pathogenesis

Table I Gene Alterations in Pheochromocytomaa Gene

Chromosomal location

Gene alteration

Prevalence

Tumor suppressor genes VHL

NF1

3p25.5

17q11.2

SDHD

11q23

p53

17p13

LOH of the wild-type allele in VHL tumors

33/46

LOH in MEN2 tumors Somatic mutations in MEN2 tumors

10/18 3/21

Somatic mutations in VHL tumors

0/36

Somatic mutations in sporadic tumors

1/20

LOH in sporadic tumors

35/102 benign, 3/10 malignant

LOH of the wild-type allele

3/7

LOH in sporadic tumors

1/6

Absent neurofibromin expression

6/6

Reduced or absent neurofibromin expression LOH in sporadic tumors

1/4 sporadic, 5/14 MEN-2, 1/2 VHL 13/18

Somatic mutations

1/18

Somatic mutations

6/55

LOH

10/46

MEN1

11q13

LOH of the wild-type allele

2/2

RASSF1A

3p21

Somatic mutations

0/23

Hypermethylation

5/23

9p21

LOH

0/26

10q11.2

Duplication of the mutant RET allele in trisomy 10 or loss of the wild-type allele in MEN2 tumors

7/9

p16 Oncogenes RET

Ras

11p15 (H-ras)

LOH in MEN2 tumors

0/13

Gain at 10q or 10 in sporadic tumors LOH in sporadic tumors

2/10 malignant, 3/42 benign 2/38 benign, 1/10 malignant

Somatic mutations in MEN2 tumors

1/3

Somatic mutations in VHL tumors

0/41

Somatic mutations in sporadic tumors

17/149 benign, 1/29 malignant

Somatic mutations

0/8

12p11 (K-ras)

a

p0015

0/8

GNAS1

20q13.2

Somatic mutations

0/10

EGFR

7p12

Overexpression

4/7

Modified from Koch, Pacak, and Chrousos (2002).

Multiple Endocrine Neoplasia Types 1 and 2 Multiple endocrine neoplasia type 1 (MEN1) is an autosomal dominant tumor syndrome caused by germ-line mutations in the gene menin located at chromosome 11q13. Characteristic endocrine tumors occur in the pituitary, parathyroid gland, and pancreas. Genotype–phenotype correlations in patients with germ-line mutations in the MEN1 gene do not exist. Adrenal nodules occur in patients with MEN1 approximately four times more often than in individuals without MEN1, suggesting a pathogenetic link. Pheochromocytomas in patients with MEN1 germline mutations have rarely been reported. LOH

at 11q13 in these tumors has been observed in pheochromocytomas. Multiple endocrine neoplasia type 2 (MEN2) is caused by germ-line mutations in the RET protooncogene located at chromosome 10q11.2. It is an autosomal dominantly inherited cancer syndrome and affects approximately 1 in 40,000 individuals. MEN2 is characterized by the presence of medullary thyroid carcinoma, pheochromocytoma, and parathyroid hyperplasia. Approximately 50% of patients with germ-line mutations in RET develop a pheochromocytoma during their lifetime. There is a clear genotype–phenotype correlation in patients with MEN2.

92

Adrenal Tumors, Molecular Pathogenesis

C10

Other events

M WT

f0005

C10 Target cell (e.g., chromaffin "cell")

Figure 1 Model for tumorigenesis in MEN2-associated pheochromocytomas. In an individual with MEN2, each cell of the target organs (e.g., adrenal medulla) carries a RET germ-line mutation in the heterozygous state with one intact wild-type RET allele. Selected chromaffin cells undergo a second hit through duplication of the mutant RETallele in trisomy 10 or loss of the wild-type allele, giving these cells overrepresention of mutant RET and a growth advantage with subsequent tumor formation. M, mutant RET allele; WT, wild-type RET allele; C10, chromosome 10. Modified from Koch, Pacak, and Chrousos (2002).

The same germ-line mutation in RET may lead to the development of a pheochromocytoma in one patient but not in another patient. Similarly, one patient with MEN2 may develop medullary thyroid carcinoma in the first month of life, whereas another patient with the same germ-line mutation in RET may develop medullary thyroid carcinoma at his or her 83rd birthday. It is therefore puzzling how germ-line mutations in RET lead to tumor formation in patients with MEN2. Recently, the model of a “second hit” in such patients was introduced. Selected cells in the target organs, such as the chromaffin cells and the C cells of the thyroid gland, may gain a growth advantage by overrepresentation of mutant RET by duplication of mutant RET in trisomy 10, loss of wild-type RET, or tandem amplification (Fig. 1). A subset of MEN2-related pheochromocytomas have somatic VHL gene alterations, possibly leading to an overrepresentation of RET. LOH at 1p has been reported in some MEN2-related pheochromocytomas, highlighting the developmental relationship between pheochromocytoma and neuroblastoma (neuroblastomas frequently have LOH at 1p). Also, reduced neurofibromatosis type 1 (NF1) expression occurs in a subset of MEN2 pheochromocytomas, suggesting a role for the ras pathway in tumor formation. Neurofibromatosis Type 1 Although this inherited tumor syndrome affects 1 in 4000 individuals, pheochromocytomas occur in less than 2% of patients with NF1. The NF1 gene maps

to chromosome 17q11.2 and encodes neurofibromin, which is involved in controlling the ras signaling pathway. Mice heterozygous for one mutant NF1 allele develop pheochromocytoma in 50% of cases. LOH at 17q11 has been observed in a subset of NF1 pheochromocytomas. Because of the rarity of NF1-related pheochromocytomas, there is a lack of larger genetic studies and the molecular pathogenesis of NF1 pheochromocytomas is largely unknown. SDHX Syndromes Pheochromocytomas are also referred to as adrenal paragangliomas (“next to the ganglia”). Paragangliomas can be of sympathetic and parasympathetic origin. Sympathetic paragangliomas are mainly located in the retroperitoneum. Parasympathetic paragangliomas are often located in the neck from the skull base down to the aortic arch. Head and neck paragangliomas usually do not oversecrete catecholamines, whereas those below the neck commonly do. Carotid body paragangliomas develop from the carotid body, a structure that serves as an oxygen-sensing organ. Chronic hypoxia has been shown to be associated with enlargement of the carotid bodies, an observation that drew the attention of investigators to possible genetic defects of the oxygen-sensing and oxygen-signaling pathways. Subsequently, germ-line mutations in the succinate–ubiquinone oxidoreductase subunit D gene (SDHD), a gene belonging to the mitochondrial complex II that is involved in the Krebs cycle and in the aerobic electron transport chain, have been identified in extraadrenal pheochromocytomas. Researchers have also analyzed adrenal pheochromocytomas for mutations in SDHD. Subsequent investigations on genes encoding for the other three subunits of mitochondrial complex II led to the detection of germ-line mutations in SDHB and SDHC in patients with hereditary pheochromocytomas. To date, germ-line mutations in SDHA, the gene encoding for the flavoprotein, have not been found in adrenal pheochromocytomas or in “paragangliomas.” Of the five mitochondrial complexes (I–V ), complex II is the only one with no subunits encoded by the mitochondrial genome. The SDHD gene is located at 11q23 and consists of four exons. Mutation analysis in pheochromocytomas has revealed missense and nonsense mutations. SDHD encodes for the small (cybS) subunit of cytochrome b in the succinate–ubiquinone oxidoreductase (complex II). Functional analysis revealed that inactivation of cybS by a nonsense mutation (R22X) abolishes the enzymatic activity of mitochondrial complex II and activates the hypoxia pathway (e.g., through increased expression of

93

Adrenal Tumors, Molecular Pathogenesis

VEGF). The SDHB gene is located at 1p35–36 and consists of eight exons. Mutation analysis showed inactivating mutations in hereditary pheochromocytomas and paragangliomas. SDHB encodes for the iron sulfur protein in the succinate–ubiquinone oxidoreductase (complex II). The SDHC gene is located at 1q21 and consists of six exons. Mutation analysis revealed a G-to-A transition in exon 1 in one family with paragangliomas. SDHC encodes for the large (cybL) subunit of cytochrome b in the succinate– ubiquinone oxidoreductase (complex II). Sporadic Pheochromocytomas Allele losses at 1p, 3p, 3q, 17p, and 22q are commonly found in pheochromocytomas. Their role in tumorigenesis and tumor progression, however, remains unclear. Less than 10% of sporadic pheochromocytomas have somatic mutations in RET, VHL, SDHD, or SDHB. Among 91 sporadic pheochromocytomas investigated for SDHD mutations, only 1 carried a somatic mutation. Mutation analysis of SDHB in 24 sporadic pheochromocytomas revealed only 1 with a germ-line mutation. Mutations in SDHC have not been identified. Notably, apparently sporadic pheochromocytomas may be part of a familial syndrome with germ-line mutations in VHL, RET, SDHD, or SDHB. In a recent study of 271 patients, Neumann et al. reported up to 24% of such cases. This may justify screening all patients presenting with a pheochromocytoma for germ-line mutations in the previously mentioned genes. Somatic mutations and genetic alterations in other genes, such as p53, NF1, RASSF1A, c-erbB-2, and EGFR, have been observed in sporadic pheochromocytomas, suggesting a role in the pathogenesis of these tumors. No markers reliably distinguish benign and malignant pheochromocytomas. The only criterion is clinical: the presence of metastases at locations at which there is usually no chromaffin tissue (e.g., liver, bone, and lung). Telomerase expression does not predict malignancy. Clonal analyses of adrenomedullary nodules in patients with a RET germ-line mutation suggest that these nodules are monoclonal.

Ganglioneuromas and Ganglioneuroblastomas Ganglioneuromas are neuroectodermal tumors related to neuroblastoma. Rarely, these tumors occur in the adrenal gland and are entirely benign except for mixed or so-called composite tumors, such as ganglioneuroblastomas. Most neuroblastomas show LOH at 1p36

and have a less favorable prognosis. Ganglioneuromas as mature tumors are not expected to reveal LOH at 1p36. A few tumors have low p53 content. The pathogenesis of these adrenal lesions is unclear.

ADRENOCORTICAL TUMORS Li–Fraumeni Syndrome This autosomal dominant familial cancer syndrome is associated with breast cancer, brain tumors, soft tissue sarcomas, leukemia, and adrenocortical carcinoma. Adrenal cancer occurs in approximately 1% of patients with the classic Li–Fraumeni syndrome. Germ-line mutations in p53 at chromosome 17p13 are responsible for this syndrome and can lead to adrenocortical tumors as the sole manifestation. Frequently, adrenal tumors of patients with germline mutations in p53 show LOH at 17p13 and, therefore, evidence of biallelic inactivation of this tumor suppressor gene (Table II).

Beckwith–Wiedemann Syndrome This syndrome shows variable expressivity and occurs sporadically. Affected patients are at increased risk for developing adrenal cancer. The gene locus is at 11p15.5, a chromosomal region that includes the IGF2 and the p57/KIP2genes. p57 is a negative regulator of cell proliferation and inhibits G1 cyclin/cyclindependent kinase (CDK) complexes. Adrenocortical tumors in this syndrome show overexpression of IGF2, probably related to uniparental paternal isodisomy for the IGF2 locus. Duplication of the paternal 11p15 allele containing the IGF2 gene locus and/or loss of the maternal allele are frequently found in adrenal cancer. This is in contrast to adrenocortical lesions that are classified as benign and frequently do not demonstrate overexpression of IGF2.

Carney Complex First described in 1980, this autosomal dominant hereditary syndrome comprises primary pigmented nodular adrenocortical disease, growth hormonesecreting pituitary tumors, spotty skin pigmentations, atrial and peripheral myxomas, and psammomatous melanotic schwannomas. Adrenocortical disease occurs in approximately 26% of patients with this complex. At least two chromosomal loci have been identified: 2p16 and 17q22–24. A subset of patients with Carney complex have germ-line mutations in

94

Adrenal Tumors, Molecular Pathogenesis

Table II Genetic Alterations in Adrenocortical Tumorsa Gene

Chromosomal location

Gene alteration

Prevalence (tumors)

Tumor suppressor genes p53

MEN1

17p13

11q13

LOH at the wild-type allele in Li–Fraumeni syndrome LOH in sporadic tumors Somatic mutations in sporadic tumors Gain at 17 in sporadic tumors LOH of the wild-type allele in MEN1 tumors LOH at 11q13 in sporadic tumors Somatic mutations in sporadic tumors

p57/kip2 and H19

11p15

LOH in sporadic tumors Somatic mutations in sporadic tumors Low expression in sporadic tumors

p21 p16 Oncogenes RET

Overexpression in sporadic tumors LOH in sporadic tumors and low expression

10q11.2

Somatic mutations in sporadic tumors RET/PTC1 rearrangement Somatic mutations in sporadic tumors

22/39 cancers 26/97 16/42 cancers 11/57 6/26 cancers 9/38 1/13 11/13 cancers 28/126 0/18 cancers 2/90 21/38 cancers 35/155 0/61 16/17 cancers 8/19 29/49 cancers 1/7 3/7 cancers

Overexpression of K-ras

1/23 2/23 0/33 0/17 3/24 cancers 10/50 6/18

11p15.5

Overexpression in sporadic tumors with duplication of the paternal allele

51/57 cancers 31/161

EGFR

7p12

Overexpression

EGF

4q25

Expression

2p13

Overexpression

68/69 cancers 10/23 0/5 cancers 0/26 5/5 cancers

20q13.2 3p21 18p11.2

Somatic mutations in sporadic tumors Somatic mutations in sporadic tumors Somatic mutations Deletion Somatic mutations Moderate to high expression

RAS

Growth factors IGF2

TGF-a Signal transduction molecules GNAS1 GNAI2 ACTH-R

a

6p21 9p21

13/17

1p13.2 (N-ras) 12p11.2 (K-ras) 11p15.5 (H-ras)

ATR1

3q21–25

PRKAR1A

17q22–24

Calcium-dependent protein kinase C

17q22–23 (a) 13p21 (d)

Modified from Koch, Pacak, and Chrousos (2002).

LOH in tumors associated with Carney complex Increased activity by cAMP stimulation Normal activity

8/55 3/47 0/16 3/45 0/17 0/1 cancer 21/76 2/2 3/3 17/17

95

Adrenal Tumors, Molecular Pathogenesis

PRKAR1A, a gene that encodes protein kinase A regulatory subunit 1a.

mechanisms unknown.

Multiple Endocrine Neoplasia Type 1

Ectopic G Protein-Coupled Receptors

Approximately 35% of patients with MEN1 have adrenal nodules, most of which are adrenocortical. LOH at 11q13, the menin locus, has been identified in some adrenocortical tumors, but the precise pathogenesis of adrenal lesions in patients with MEN1 remains obscure.

Aberrant expression of ectopic membrane hormone receptors, including gastric inhibitory polypeptide, b-adrenergic agonists, luteinizing hormone, vasopressin, and interleukin-1, has been found in some adrenal adenomas but not in adrenal cancer.

Familial Hyperaldosteronism Hyperaldosteronism can occur in at least two familial syndromes: familial hyperaldosteronism type 1, caused by the formation of a hybrid gene with fusion of the corticotropin-regulated promoter of the 11bhydroxylase gene and the angiotensin II-regulated aldosterone synthase gene at 8q24, and familial hyperaldosteronism type 2, which does not respond to dexamethasone administration by decreasing corticotropin (ACTH) secretion and whose responsible gene is linked to chromosomal subband 7p22 but is still unidentified. The molecular pathogenesis of adrenal nodule development in patients with hyperaldosteronism is unknown.

Congenital Adrenal Hyperplasia Almost all patients with this autosomal recessive disorder have germ-line mutations in the gene coding for 21-hydroxylase at 6p21.3. This genetic defect leads to impaired cortisol secretion with subsequent elevation of ACTH. Almost half of heterozygous carriers have macronodular adrenal disease, the rationale basis for investigators to search for mutations in the 21-hydroxylase gene in patients with apparently sporadic adrenal tumors. How these adrenal nodules develop remains unknown.

McCune–Albright Syndrome This syndrome is a sporadic postzygotic genetic disease that includes growth hormone-secreting pituitary adenomas and nodular adrenocortical disease. Somatic mutations in the a chain of the stimulatory G protein GNAS1 at 20q13.2 are responsible for this syndrome and result in stimulation of cAMP. Mutations in the cAMP inhibitory GNAI2 gene occur in only a few adrenocortical tumors. The precise

of

adrenal

nodule

formation

are

Sporadic Tumors Cell replication errors may lead to numerical changes of chromosomes, chromosomal translocations, amplification and/or loss of genes, somatic sequence alterations in specific genes including DNA repair genes, and other genomic changes. An important step in analyzing adrenal and other tumors is to determine whether they are mono- or polyclonal. The cellular origin of adrenocortical tumors is unknown, although clonal analyses of adrenal cortex lesions have helped better define the nature of these tumors. Usually, the development of cancer and tumors in general is regarded as a multistep process. One tumor-initiating mutation in a single cell may equip the cell with a selective growth advantage, enabling it to become a tumor. This tumor would then be called monoclonal since it derived from a single genetically aberrant cell. In contrast, a polyclonal tumor would develop from a group of aberrant cells arising in parallel. Only one of three studies on the clonal analysis of adrenocortical tumors noted the duration of follow-up for patients with benign adrenal adenomas/hyperplasias, even though this feature is important because differences in the rate of growth may distinguish benign from malignant adrenocortical tumors. It appears that adrenocortical carcinomas are monoclonal, whereas the majority of benign adrenocortical lesions are polyclonal. It can be speculated that tumorigenesis develops from polyclonal adrenocortical cell aggregates, some of which gain a selective growth advantage, giving rise to a monoclonal tumor. However, considering the issues of clonality interpretation, it is unclear whether a clonality assay helps differentiate benign and malignant adrenocortical lesions. Comparative genomic hybridization (CGH) studies and allele typing using microsatellite markers to detect involvement of genes with tumor suppressor or oncogenic function may help in elucidating the pathogenesis of adrenocortical tumors. Studies on adrenocortical tumors must be considered carefully because

96 many investigators do not refer to a specific follow-up time; the natural history is important in classifying an adrenocortical tumor as benign or malignant. Sidhu et al. reported an equal distribution of chromosomal gains and losses in benign and malignant adrenocortical tumors, although the genetic events in both groups were quite different. Limitations of this study are the classification of benign vs malignant tumors, which was not based on the presence of metastases but rather on modified Weiss histologic criteria, and the variable follow-up period for the group of benign tumors, which was up to 41 months. These investigators analyzed 18 benign and 13 malignant adrenocortical tumors. Benign adrenal masses 5 cm) adrenal tumors had gains in chromosomes 5, 12, and 19. Losses in the benign group were limited to chromosome 3q, whereas they occurred at chromosomes 1p, 11, 17p, and 22 in the malignant group. This is in contrast to earlier CGH studies showing gains at chromosome 4 mainly or exclusively in malignant tumors. Sidhu et al. proposed activation of a protooncogene on chromosome 4 as an early event in adrenocortical tumorigenesis and that the presence of four or more CGH alterations in one tumor is suggestive of the malignant phenotype. To elucidate the pathogenesis of adrenocortical tumors, the precise definition of adenoma vs carcinoma is of utmost importance, as is the identification of precursor lesions. Without this critical information, it is difficult to determine the sequence of events from tumor initiation to progression. The assumption that putative oncogenes at regions of chromosomal gain represent the first step in tumor development, and regions of chromosomal loss represent putative tumor suppressor genes allowing subsequent steps of tumor progression, may seem plausible to many investigators but it is speculative. Candidate genes believed to be involved in tumor formation of sporadic adrenal tumors are those that are known to be associated with hereditary tumor syndromes, such as p53, GNAS1, MEN1, and IGF2, and genes and receptors involved in signal transduction systems. Mutations or other genetic alterations in the angiotensin II type 1 receptor and corticotropin receptor gene have rarely been found. Biallelic inactivation of the MEN1 gene in sporadic adrenocortical tumors has not been identified, although either somatic mutation in MEN1 or LOH at its locus 11q13 have been reported. The roles of IGF2, H19, and p57/KIP2, all located at 11p15, in the pathogenesis of adrenocortical tumors were previously mentioned. Overexpression

Adrenal Tumors, Molecular Pathogenesis

of IGF2 is frequently found in malignant sporadic adrenocortical tumors. Genetic alterations in p57/ KIP2 (CDK inhibitor 1C) were sought in adrenocortical tumors because adrenal cancers frequently showed allelic loss at 11p15. However, no somatic mutations were found in this gene in 75 sporadic adrenocortical tumors, but low p57/KIP2 expression was demonstrated in 3 of 10 adrenal adenomas and 6 of 6 carcinomas. Reduced or absent function of this gene seems to lead to enhanced activity of G1 CDK complexes with possibly subsequent promotion of cell proliferation. Consequently, investigators examined adrenocortical tumors for abnormalities in other inhibitors of CDKs. One such protein, P16, whose gene is located at 9p21, is an inhibitor of the CDK 2A gene. One of 7 benign and 3 of 7 malignant adrenocortical tumors showed loss of one p16 allele and absent p16 protein by immunohistochemistry. P21, a CDK inhibitor that can be induced by p53, was overexpressed in 70% (25/38) of adrenocortical cancer samples. Downregulation of p21, however, did not affect the prognosis of patients with adrenal cancer. This suggests a role for these CDK inhibitors in only a small subset of adrenocortical cancers. In addition to IGF2, epidermal growth factor (EGF) is very important. Its receptor, EGFR, has been studied in a small number of adrenal tumors. By immunohistochemistry, EGFR was overexpressed in benign and malignant adrenocortical tumors, whereas EGF was not detected. Instead, transforming growth factor-a (TGF-a) was overexpressed in adrenal cancer. TGF-a is a natural ligand for EGFR. The p53 gene has been studied in many tumors, including adrenocortical tumors, because of its frequent mutation in cancers and its known role in regulating the cell cycle. p53has been classified as a tumor suppressor gene and is mutated in the germ line of most patients with Li–Fraumeni syndrome. Since this syndrome is associated with adrenocortical tumors, it is conceivable that p53 may also play a role in the pathogenesis of sporadic adrenal tumors. Although allelic loss at 17p13, the locus for the p53 gene, and somatic p53 mutations frequently occur in adrenal cancer, they are uncommon in benign adrenocortical tumors, suggesting that genetic alterations in p53 are involved in tumor progression rather than initiation. Only one study from Taiwan reported p53 mutations in adrenocortical adenomas, but the classification of benign vs malignant adrenal tumors, including the follow-up period of the affected Taiwanese patients, remains unclear. In contrast, p53 mutations (predominantly in exons 5–8) were reported in up to 70% of adrenal cancers. However, it is unknown whether

97

Adrenal Tumors, Molecular Pathogenesis

biallelic inactivation of p53 occurred in these tumors since investigators reported either allelic loss at 17p13 or somatic p53 mutations and not whether both events (biallelic inactivation of p53) took place in the respective tumor. Nevertheless, LOH at 17p13 has been suggested as a molecular prognostic marker for sporadic adrenocortical tumors. The ras gene family encodes G proteins, which are involved in signaling pathways through modification of intracellular cAMP concentrations. Overexpression of ras may lead to constitutive signal transduction and/or cell proliferation. ras mutations were first identified in 12.5% of adrenocortical tumors with an equal prevalence in benign and malignant tumors. A subsequent study reported somatic K-ras mutations and an overexpression of K-ras in 33% of benign adrenocortical tumors, whereas somatic H-ras mutations were not detected. However, previous studies did not find this prevalence rate of ras mutations. The RET protooncogene at chromosome 10q11.2 encodes a tyrosine kinase receptor that is involved in the control of cell differentiation and proliferation. Germ-line mutations in RET are found in almost all patients with MEN2. Importantly, RET is only expressed in certain tissues, such as the neural crestderived parafollicular C cells in the thyroid gland and extra- and intraadrenal chromaffin cells. Overrepresentation of mutant RET may initiate tumorigenesis of medullary thyroid carcinoma and pheochromocytoma. Although adrenocortical tumors are not part of MEN2, they may have genetic alterations involving RET. Analysis of 21 sporadic adrenocortical tumors revealed 1 aldosterone-producing tumor with a point mutation in RET and RET/PTC1 rearrangements in 2 tumors—one cortisol-producing and one aldosterone-producing. Telomerase is the ribonucleoprotein enzyme that elongates telomeres (i.e., chromosomal DNA ends). In most normal somatic cells, telomerase is repressed, whereas this enzyme is reactivated in transformed cells. Recently, a considerable amount of data have been obtained on telomerase activity in human cancers, including endocrine tumors. In general, the presence or absence of telomerase activity in adrenocortical tumors does not seem to be a prognostic marker, although some studies suggest that malignant adrenal tumors have increased telomerase activity.

CONCLUSION Most adrenal tumors are sporadic and associated with numerous somatic genetic alterations. The exact

order and time frame of these genetic events are difficult to determine. Therefore, one cannot make precise statements about the pathogenesis of tumor initiation and tumor progression. Identifying reliable genetic prognostic markers is also a challenge and strongly depends on the follow-up period of patients with adrenal tumors that are classified as benign or malignant. Focusing on adrenal tumors that occur in hereditary syndromes may facilitate the determination of their pathogenesis since in these adrenal masses at least one genetic hit, the inherited gene defect, is known. Subsequent genetic alterations may then be easier to place into context. In MEN2-associated pheochromocytomas, a second hit model leading to overrepresentation of mutant RET as a possible tumor-initiating event has been proposed. Somatic mutations in genes that are known to be involved in hereditary syndromes, such as p53, VHL, and RET, are rarely found in sporadic adrenal tumors. Future studies should focus not only on single genes and genetic alterations but also on the interaction of their encoded protein products.

See Also the Following Articles Beckwith-Wiedemann Syndrome (BWS) . McCune-Albright Syndrome . Multiple Endocrine Neoplasia (MEN) Type 2 . Neurofibromatosis . Pheochromocytoma . Von Hippel-Lindau Syndrome

Further Reading Brodeur, G. M., Sawada, T., Tsuchida, Y., and Voute, P. A. (eds.) (2000). “Neuroblastoma.” Elsevier, Amsterdam. Casey, M., Vaughan, C. J., He, J., Hatcher, C. J., Winter, J. M., Weremowicz, S., Montgomery, K., Kucherlapati, R., Morton, C. C., and Basson, C. T. (2000). Mutations in the protein kinase A R1alpha regulatory subunit cause familial cardiac myxomas and Carney complex. J. Clin. Invest. 106, R31–R38. Celik, V., Unal, G., Ozgultekin, R., Goksel, S., Unal, H., and Cercel, A. (1996). Adrenal ganglioneuroma. Br. J. Surg. 83, 263. Eng, C. (1999). RET proto-oncogene in the development of human cancer. J. Clin. Oncol. 17, 380–393. Gicquel, C., Bertagna, X., Gaston, V., Coste, J., Louvel, A., Baudin, E., Bertherat, J., Chapuis, Y., Duclos, J. M., Schlumberger, M., Plouin, P. F., Luton, J. P., and Bouc, Y. L. (2001). Molecular markers and long-term recurrences in a large cohort of patients with sporadic adrenocortical tumors. Cancer Res. 61, 6762–6767. Gimm, O., Koch C. A., Januszewicz, G., Opocher, G., and Neumann, H. P. H. (2003). The genetic basis of pheochromocytoma. Frontiers Horm. Res., in press. Hedeland, H., Ostberg, G., and Hokfelt, B. (1968). On the prevalence of adrenocortical adenomas in an autopsy material in relation to hypertension and diabetes. Acta Med. Scand. 194, 211–214.

98 Huang, S. C., Koch, C. A., Vortmeyer, A. O., Pack, S. D., Lichtenauer, U. D., Mannan, P., Lubensky, I. A., Chrousos, G. P., Gagel, R. F., Pacak, K., and Zhuang, Z. (2000). Duplication of the mutant RET allele in trisomy 10 or loss of the wildtype allele in multiple endocrine neoplasia type 2-associated pheochromocytoma. Cancer Res. 60, 6223–6226. Kirschner, L. S., Carney, J. A., Pack, S. D., Taymans, S. E., Giatzakis, G., Cho, Y. S., Cho-Chung, Y. S., and Stratakis, C. A. (2000). Mutations of the gene encoding the protein kinase A type I-alpha regulatory subunit in patients with the Carney complex. Nature Genet. 26, 89–92. Kloos, R. T., Gross, M. D., Francis, I. R., Korobkin, M., and Shapiro, B. (1995). Incidentally discovered adrenal masses. Endocr. Rev. 16, 460–484. Koch, C. A., Pacak, K., and Chrousos, G. P. (2002). The molecular pathogenesis of hereditary and sporadic adrenocortical and adrenomedullary tumors. J. Clin. Endocrinol. Metab. 87, 5367–5384. Koch, C. A., Vortmeyer, A. O., Zhuang, Z., Brouwers, F. M., and Pacak, K. (2002). New insights into the genetics of familial chromaffin cell tumors. Ann. N. Y. Acad. Sci. 970, 11–28. Koch, C. A., Huang, S. C., Azumi, N., Zhuang, Z., Chrousos, G. P., Vortmeyer, A. O., and Pacak, K. (2002). Somatic VHL gene deletion and mutation in MEN 2A-associated pheochromocytoma. Oncogene 21, 479–482. Koch, C. A., Walther, M. M., and Linehan, W. M. (2002). Von Hippel–Lindau syndrome. In “Adrenal Physiology and Diseases” (G. P. Chrousos, ed.). www.endotext.org/adrenal. Koch, C. A., Brouwers, F. M., Rosenblatt, K., Burman, K. D., Davis, M. M., Vortmeyer, A. O., and Pacak, K. (2003). Adrenal ganglioneuroma in a patient presenting with severe hypertension and diarrhea. Endocr. Rel. Cancer 10, 99–107. Lack, E. E. (1997). “Tumors of the Adrenal and Extraadrenal Paraganglia.” Armed Forces Institute of Pathology, Washington, DC. Lacroix, A., Ndiaye, N., Tremblay, J., and Hamet, P. (2001). Ectopic and abnormal hormone receptors in adrenal Cushing’s syndrome. Endocr. Rev. 22, 75–110. Levy, A. (2001). Monoclonality of endocrine tumors: What does it mean? Trends Endocrinol. Metab. 12, 301–307. Neumann, H. P., Bausch, B., McWhinney, S. R., Bender, B. U., Gimm, O., Franke, G., Schipper, J., Klisch, J., Altehoefer, C.,

Adrenal Tumors, Molecular Pathogenesis

Zerres, K., Januszewicz, A., and Eng, C. (2002). Germ-line mutations in nonsyndromic pheochromocytoma. N. Engl. J. Med. 346, 1459–1466. Schussheim, D. H., Skarulis, M. C., Agarwal, S. K., Simonds, W. F., Burns, A. L., Spiegel, A. M., and Marx, S. J. (2001). Multiple endocrine neoplasia type 1: New clinical and basic findings. Trends Endocrinol. Metab. 12, 173–178. Sidhu, S., Marsh, D. J., Theodosopulos, G., Philips, J., Bambach, C. P., Campbell, P., Magarey, C. J., Russell, C. F. J., Schulte, K. M., Roher, H. D., Delbridge, L., and Robinson, B. G. (2002). Comparative genomic hybridization analysis of adrenocortical tumors. J. Clin. Endocrinol. Metab. 87, 3467–3474. Stratakis, C. A., Carney, J. A., Lin, J. P., Papanicolaou, D. A., Karl, M., Kastner, D. L., Pras, E., and Chrousos, G. P. (1996). Carney complex: A familial multiple neoplasia and lentiginosis syndrome: Analysis of 11 kindreds and linkage to the short arm of chromosome 2. J. Clin. Invest. 97, 699–705. Stratakis, C. A., Jenkins, R. B., Pras, E., Mitsiadis, C. S., Raff, S. B., Stalboerger, P. G., Tsigos, C., Carney, J. A., and Chrousos, G. P. (1996). Cytogenetic and microsatellite alterations in tumors from patients with the syndrome of myxomas, spotty skin pigmentation, and endocrine overactivity (Carney complex). J. Clin. Endocrinol. Metab. 81, 3607–3614. Tischler, A. S. (2000). Divergent differentiation in neuroendocrine tumors of the adrenal gland. Semin. Diagn. Pathol. 17, 120–126. Van der Harst, E., de Krijger, R. R., Dinjents, W. N., Weeks, L. E., Bonjer, H. J., Bruining, H. A., Lamberts, S. W., and Koper, J. W. (1998). Germline mutations in the VHL gene in patients presenting with pheochromocytomas. Int. J. Cancer 77, 337–340. Walther, M. M., Reiter, R., Keiser, H. R., Choyke, P. L., Venzon, D., Hurley, K., Gnarra, J. R., Reynolds, J. C., Glenn, G. M., Zbar, B., and Linehan, W. M. (1999). Clinical and genetic characterization of pheochromocytoma in von Hippel–Lindau families: Comparison with sporadic pheochromocytoma gives insight into natural history of pheochromocytoma. J. Urol. 62, 659–664. Walther, M. M., Herring, J., Enquist, E., Keiser, H. R., and Linehan, W. M. (1999). Von Recklinghausen’s disease and pheochromocytomas. J. Urol. 162, 1582–1586. Weiss, L. M. (1984). Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. Am. J. Surg. Pathol. 8, 163–169.

Adrenarche, Premature Catherine Dacou-Voutetakis, Sarantis Livadas, Antony Voutetakis, and Maria Dracopoulou Athens University Medical School, Aghia Sophia Children’s Hospital, Athens, Greece

Glossary adrenarche Initiation of function of the reticular zone of the adrenal gland. congenital adrenal hyperplasia A disorder resulting from enzymatic defects in the pathway of cortisol synthesis. CYP21 A gene encoding the steroidogenic enzyme 21-hydroxylase. gonadarche Initiation of function of the hypothalamic– pituitary–gonadal axis. g0020

premature pubarche The appearance of pubic hair, with or without axillary hair growth or increased apocrine odor, in the absence of other secondary sexual characteristics, occurring prior to age 8 in girls and age 9 in boys.

P

remature adrenarche refers to the premature activation of the zona reticularis of the adrenal gland, clinically manifested by pubic hair development.

INTRODUCTION Although more than 60 years have elapsed since the endocrinologist F. Albright coined the term adrenarche, the mystery of the mechanisms involved in its development has not yet been unraveled. In studying girls with gonadal dysgenesis, a natural experimental model of agonadism, Albright observed that pubic hair can develop under the influence of adrenal androgens and in the absence of gonadal hormones. He was thus the first to distinguish between gonadal and adrenal puberty (gonadarche and adrenarche, respectively). Subsequent studies of other clinical prototypes, including hypogonadotropic hypogonadism and isolated premature thelarche, confirmed his observation. Dissociation between adrenal androgens and cortisol secretion has also been documented by demonstrating that a rise in dehydroepiandrosterone

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

sulfate (DHEAS) during puberty is not accompanied by a rise in cortisol. Moreover, the rise in adrenocorticotropin hormone (ACTH) and cortisol in Cushing’s syndrome occurs without an analogous increase in DHEAS.

ADRENARCHE Definition Adrenarche refers to the activation of the zona reticularis (ZR) of the adrenal gland, biochemically marked by a rise in dehydroepiandrosterone (DHEA) and DHEAS but not in other androgens. The clinical manifestation of adrenarche is termed pubarche and is characterized by the development of pubic hair with or without axillary hair growth or apocrine odor. The decline in the function of the ZR, later in life, is termed adrenopause.

Phylogenetic Data on Adrenarche Adrenarche and adrenopause appear to be phenomena that are unique to the highest order of primates and, therefore, represent a recent evolutionary development. The serum levels of DHEA, DHEAS and D4adrostenedione (D4A) remain unchanged during sexual maturation in rats, hamsters, guinea pigs, sheep, pigs, goats, horses, and cows. Modest (twofold) changes in DHEA are observed in rabbits and dogs only after their sexual maturation. In primates, the phenomenon of adrenarche is observed only in the chimpanzee and not in other types (e.g., the rhesus monkey).

Ontogenetic Data on Adrenarche During the fifth gestational week, cells from the celomic epithelium migrate to the cephalic part of the mesonephros and form the distinct adrenal primordium. The central part of this primordium consists of large, eosinophilic cells that constitute the so-called

99

100

Adrenarche, Premature

fetal zone, whereas the outer zone consists of densely packed cells that form the definitive zone. By the second trimester of pregnancy, the fetal zone of the human adrenal cortex represents approximately 80– 90% of the fetal adrenal. In the fetal zone, CYP11A and CYP17 [but not 3b-hydroxysteroid dehydrogenase (3b-HSD)] and dehydroepiandrosterone sulfotransferase are expressed. The adult ZR, which also lacks 3b-HSD and expresses dehydroepiandrosterone sulfotransferase activity, is considered the equivalent of the fetal adrenal zone (Fig. 1). During the third trimester of pregnancy, the definitive zone of the fetal adrenal forms two, functionally separate units: the outer unit is the zona glomerulosa and the inner unit comprises the zona fasciculata and the ZR (Fig. 1). It must be underscored that the cortisol- and androgen-producing parts of the adrenal and the adrenal medulla are related ontogenetically, anatomically, and functionally, suggesting a possible link between adrenomedullary function and adrenarche.

Postnatal Development of the Reticular Zone

p0035

During the first postnatal months in humans, the fetal zone regresses and almost disappears. Focal islands of the ZR are noted in the adrenals at the age of 3 years and a continuous ZR starts to develop at the age of 6, an age at which adrenal C19 steroids begin to rise. Adrenarche is initiated at approximately age 6 in girls and approximately age 7 in boys and this is reflected in a gradual rise of the adrenal androgens DHEA and DHEAS. The driving force of these evolutionary events has not been determined. The serum levels of DHEA and DHEAS start to rise approximately 2 years prior to the gonadal

activation (gonadarche), reaching their peak values at 20–25 years. Thereafter, the serum concentrations of the adrenal androgens gradually decline and in elderly people are 10–20% of those encountered in young adults (adrenopause). The DHEAS reduction in the elderly is coupled with a decrease in the width of the ZR without any impact on the size of the rest of the adrenal cortex. The values of cortisol and aldosterone do not show analogous changes. This observation strongly indicates that the events described are unique to the ZR. Some of the mechanisms involved in these alterations of the ZR are outlined herein. A histological study of adrenal samples from individuals between the ages of 4 months and 56 years showed a decrease in the enzymatic activity of 3bHSD in the ZR, beginning at the age of 8–13 months with a further progressive decline up to the age of 25–26 years. Longitudinal clinical data obtained from subjects aged 2.9 to 12.3 years have demonstrated a progressive, age-related increase in the DHEAS values (approximately 22% per year) in parallel with an increased activity in 17,20-lyase and a decreased activity in 3b-HSD (an enzymatic profile encountered in both the fetal and the adult reticular zones). These changes are already evident at the preadrenarcheal stage of development. It can thus be deduced from this study that adrenarche is not the result of a sudden change in the activity of adrenal enzymes, at a particular period of time; it rather reflects a gradual maturational process that begins in early childhood. When controlled for chronological age, no association between weight, body mass index (BMI), and DHEAS (a marker of adrenarche) was evident. Nevertheless, changes in the nutritional status, measurable by changes in BMI, have been suggested as an important physiological regulator of adrenarche. Adult

Fetal Definitive zone Transitional region

Aldosterone

Zona glomerulosa

Cortisol

Zona fasciculata Fetal zone DHEA and DHEA “sulfate”

Zona reticularis Medulla

f0005

Figure 1 The zones in the fetal and adult adrenal gland. Modified from Topical Endocrinology 21, p. 18, with permission of the publisher.

101

Adrenarche, Premature

Origin, Regulation, and Biological Significance of Adrenarche Origin and Regulation The extra- or intra-adrenal factors involved in the development and function of the ZR have not yet been fully elucidated. It has long been postulated that a distinct pituitary factor [adrenal androgenstimulating hormone (AASH)] conducts the development and function of the ZR, but this factor has not been isolated. Studies of natural human models have provided valid data on the putative developmental and/or functional factors that influence the ZR. Among these models are anencephalic fetuses, patients with congenital adrenal hypoplasia (ACTH receptor defect), combined pituitary hormone deficiency, hypogonadotropic hypogonadism, or isolated premature thelarche, and subjects with precocious gonadarche on gonadotropin-releasing hormone analogue (GnRHa) suppression. The study of adrenal androgens in these models suggests that DHEA and DHEAS synthesis and secretion are dependent on an intact corticotropin-releasing hormone (CRH)–ACTH axis, not only later in life but also in the latter part of pregnancy. It is most probable that the CRH–ACTH complex exerts a permissive effect and acts in synergy with a putative extra-adrenal factor to successfully orchestrate the development and function of the ZR. An interesting model in which low DHEAS levels have been detected, despite the normal ACTH– adrenal axis, is the pituitary insufficiency associated with the Prop1 gene defect. The low level of DHEAS in these patients may indicate that the pituitary transcription factor Prop1 is necessary for the normal synthesis of the putative factor (AASH) that initiates adrenarche. Alternatively, the low DHEAS level in patients with the Prop1 gene defect could simply represent an early marker of incipient ACTH insufficiency, which is known to occur later in life in a number of these patients. In addition to CRH–ACTH and the putative AASH, other extra-adrenal factors have been implicated in the development and function of the ZR: prolactin, estrogens, the epidermal growth factor, angiotensin, gonadotropins, proopiomelanocortin-related peptides, growth hormone (GH), insulin growth factor-I (IGF-I), insulin, and possibly adipose tissue factors. None of these factors, however, have been conclusively shown to regulate androgen secretion by the adrenal gland. Age-related alterations in the expression of the adrenal enzymes have also been proposed as a mechanism

for the development and function of the ZR (intraadrenal factors). These changes not only refer to the relative activity of the adrenal enzymes but also to their responsiveness to ACTH. Specifically, the increase in 17-hydroxylase and 17,20-lyase activity occurs along with a decrease in 3b-HSD activity, primarily evident in the developing ZR. It must be stressed, however, that these biochemical changes alone cannot fully explain the initiation of adrenarche. Biological Significance of Adrenarche Since a human model of isolated absence of the ZR has not thus far been identified, the exact biological role of the ZR and the implications of its absence or insufficiency still remain enigmatic. A small, transient increase in growth rate occurring at approximately age 7 (midchildhood growth spurt) has been attributed to the initiation of adrenarche. However, a cause and effect relationship between adrenarche and the midchildhood growth spurt has been disputed. It has also been shown that adrenarche is not a sine qua non for gonadarche since gonadal puberty proceeds normally in clinical entities in which adrenarche is absent. The decline of DHEA coincides with signs of aging and has therefore been interpreted to indicate that aging is, at least in part, a DHEA deficiency syndrome. This observation has prompted studies on the effect of DHEAS replacement in the elderly and in young subjects with DHEAS deficiency of various etiologies, with equivocal results.

PREMATURE ADRENARCHE–PUBARCHE Definitions Premature adrenarche (PA) and premature pubarche (PP) are frequently used interchangeably. Nevertheless, they are not synonymous. PA refers to premature activation of the ZR of the adrenals and is marked by levels of DHEA and DHEAS that are high for the chronological age (CA) but appropriate for the stage of pubic hair development. PP is the term applied to characterize the clinical expression of PA, namely, the appearance of pubic hair, usually at the labiae, with or without axillary hair growth or increased apocrine odor, in the absence of other secondary sexual characteristics, prior to age 8 in girls and 9 in boys. This age cut-off point has been called into question but is still accepted. PP occurs more frequently in girls than in boys, with a male to female ratio of 1/5 to 1/10, for no apparent reason.

102 The term ‘‘exaggerated adrenarche’’ has been coined to describe a form of PP in which androgen levels, basal or post-ACTH stimulation, are above those expected for the stage of pubic hair development.

Etiology of Premature Pubarche

p0100

Premature pubarche may be caused by: (1) premature activation of the ZR without any apparent pathological condition (idiopathic), (2) congenital adrenal hyperplasia (CAH), (3) virilizing adrenal or ovarian tumor, and (4) increased end-organ sensitivity to androgens. In idiopathic PP, increased BMI or a sudden rise in BMI may constitute a trigger factor for its induction. Nevertheless, BMI and leptin levels only partially explain the increased DHEAS values. The GH–IGF-I axis and especially hyperinsulinism, as a consequence of insulin resistance, have been implicated in androgen production by the ZR and generally in the mechanism of PP initiation. The most frequent pathological condition underlying PP, and the one usually creating diagnostic dilemmas, is defective adrenal steroidogenesis and, in particular, nonclassical CAH (NC CAH). This aspect of PP has been quite controversial. Based on the determination of basal and ACTH-stimulated adrenal androgen levels, defective steroidogenesis, indicative of the NC CAH (caused by 21-hydroxylase or 3bHSD deficiency), ranges from 0 to 54% in the various published series. These huge differences are probably due to the variety of criteria used for recruitment and evaluation of the subjects participating and, most important, in the lack of confirmation by molecular analysis. In a study by Dacou-Voutetakis and Dracopoulou, 48 consecutive cases of PP were evaluated by molecular analysis of the CYP21 gene as well as by basal and ACTH-stimulated 17-OH progesterone (17OHP) values. A significantly increased incidence of mutations in the CYP21 gene was detected in PP children in comparison to the general population (heterozygous 37.5% and homozygous 8.3%). The 17OHP values on the ACTH test showed an overlap between carriers and noncarriers. The application of the receiver operating curve (ROC) curve in this study showed that the sum of the basal plus 60 min value of 17OHP was the best indicator of heterozygosity. For children over the cut-off point of 5 ng/ml (15 nmol/ liter), there is a 76.5% certainty of heterozygosity for a CYP21 mutation. By using the nomogram proposed by New et al., the heterozygote’s values fell in the expected area, but this was also the case for the majority of normal values. A value of 17OHP 60 min postACTH stimulation equal to or greater than 10 ng/ml

Adrenarche, Premature

(30 nmol/liter) was indicative of a homozygous mutation and occurred in 8.3% of the cases. The latter finding is accepted by most investigators. An increased incidence of CYP21 heterozygosity in either PP or functional hyperandrogenism in adolescents has also been reported by Witcell et al. and has been postulated by Knorr et al., based on hormonal evaluation. Contrary to these findings, Potau et al. did not a find higher incidence of CYP21 carriers in Spanish subjects with a history of PP. It is not possible to determine whether or not PP CYP21 heterozygote subjects have a higher probability of manifesting hyperandrogenism and the clinical syndrome related to this entity than PP girls who are not CYP21 carriers. Only a long-term follow-up study of such cases will provide a definitive answer to these important questions. Mutations in the 3b-HSD gene have also been detected in girls with PP, though infrequently. The new hormonal criteria for the diagnosis of 3bHSD deficiency in children with PP are as follows: (1) baseline 17OH pregnenolone (17P) and 17P/ cortisol ratio >29 nmol/liter and >103, respectively, and (2) ACTH-stimulated 17P and 17P/cortisol ratio >294 nmol/liter and >363, respectively. Androgen-producing tumors of the adrenals or ovaries are rarely a cause of PP, but they should be considered in the differential diagnosis. Finally, in some children with PP, no androgen excess for either CA or pubertal stage is detected and the pubic hair growth is attributed to increased end-organ sensitivity to androgens.

Long-Term Consequences of Premature Adrenarche–Pubarche Growth and Pubertal Development At presentation, children with PP show an acceleration of linear growth and skeletal maturation. In some children, the difference between their bone age (BA) and CA (DBA
CA) can be up to 2 years, whereas in others the BA is comparable to CA. Long-term follow-up of children with PP has not shown impaired growth potentials: the final height (FH) attained is comparable to target height (TH). It seems that the linear growth pattern in children with PP is modified in that height velocity in the period preceding PP diagnosis is higher than that of controls, peak height velocity occurs at an earlier age, and growth during puberty is compromised. Despite reports indicating good FH prognosis, the physician must individualize the approach for each child with PP and closely watch patients with so-called ‘‘exaggerated adrenarche’’ and/ or those who have BA advancement greater than 1

p0115

103

Adrenarche, Premature

year, for the possibility of an unfavorable effect of PP on growth potentials, namely, a predicted height below the TH. The age of gonadarche and menarche in children with PP is not different than in the general population. Functional Ovarian Hyperandrogenism and Polycystic Ovary Syndrome In a percentage of girls with PP, functional ovarian hyperandrogenism (FOH) (excessive response of androgens to GnRHa), along with an increased incidence of anovulation, hirsutism, and menstrual irregularities [characteristics of polycystic ovary syndrome (PCOS)], has been reported to occur already in late adolescence. It seems that girls with an exaggerated response of 17OHP to ACTH at the time of PP diagnosis are more likely to develop FOH and/or PCOS. Insulin Resistance, Hyperinsulinism, and the Metabolic Syndrome Acanthosis nigricans, decreased insulin sensitivity and consequent hyperinsulinism, and dyslipidemia have been found in a number of girls with PA. Pertinent data in boys are inconsistent. In certain studies, the hyperinsulinemia is already evident at the prepubertal stage, possibly conferring a higher risk for later development of diabetes mellitus type 2 (DM2). Contrary to the above findings, other investigators found no differences in glucose tolerance, insulin resistance indices, or lipid values between PP girls and control girls studied 5 years after menarche. It is quite possible that the stated differences reflect differences in the population (ethnic) groups studied and this must be seriously considered in the study and follow-up of PP girls. The contribution of intrauterine growth retardation to the occurrence of later adverse consequences, and especially FOH in girls with PP, is highly controversial and most authors have not found such an association.

Does PA Represent a Pathologic Condition? p0130

The question of whether or not PP is a benign condition, simply reflecting premature activation of the ZR without long-term adverse consequences, cannot be convincingly answered yet. A number of girls with PP later present insulin resistance and hyperinsulinism, low sex hormone-binding globulin (SHBG), FOH, increased hirsutism, acne, or polycystic ovarylike syndrome with possible later occurrence of DM2 or the metabolic syndrome. Pertinent data in boys are not consistent.

No specific features at PP diagnosis have yet been identified to predict the child at risk of developing these pathological entities. Nevertheless, it seems that some of the late consequences are, to a large extent, restricted to certain population (ethnic) groups. In addition to ethnic group, other factors, such as obesity and a family history of DM2, might contribute to the occurrence of adverse consequences. Exaggerated adrenarche as well as low SHBG (an index of insulin resistance and free androgen levels) and glucose/insulin ratio 1), the main possibilities are as follows: (1) exaggerated adrenarche, (2) defective adrenal steroidogenesis, or (3) virilizing tumor (adrenal or ovarian). In such cases, an intravenous Synachten test (ACTH) is carried out (250 mg, 150 mg/m2, or 10 mg/m2, the last dose being infrequently used). Blood samples, primarily for 17OHP determination, are obtained at 0 and 60 min. A 60 min value of 17OHP >10 ng/ml strongly suggests a homozygous mutation of the CYP21 gene (NC form) and CYP21 gene analysis should be considered (Fig. 3). There is a 75% probability of existence of heterozygosity if the sum value of 0 and 60 min postSynacthen is >5 ng/ml (15 nmol/liter). Obviously, the latter information is of no immediate practical significance and may be of value only for genetic

p0140

p0145

104

Adrenarche, Premature

Hyperinsulinemia (insulin resistance)

• Metabolic syndrome

Possible contributing inputs

• Ethnic group

• Dyslipidemia • Type 2 diabetes

?

Hyperandrogenism

• PP

• Obesity • Family history • IUGR Steroidogenic adrenal defect

Figure 2 Possible pathogenetic mechanisms of premature pubarche (PP). IUGR, intrauterine growth retardation.

counseling. Determination of glucose/insulin value, SHBG, and ovarian sonography might constitute a good baseline record but should not be regarded as necessary diagnostic tools. It must be stated that an increased prevalence of sonographic findings of PCOS has been reported in girls with PP at the prepubertal stage, but the diagnostic and prognostic value of such a finding remains unclear. The dexamethasone suppression test or some form of imaging

study is rarely indicated for the remote possibility of a virilizing tumor. This latter possibility is much higher if pubarche occurs very early and there is an increased level of serum testosterone. Management and Follow-up The therapeutic approach for children with PP will be determined by the pathogenetic mechanism involved and the clinical findings. In most cases, however, drug

Clinical and laboratory data

Diagnostic possibilities

∆BA−CA < 1 year Predicted height within target

High probability of idiopathic PP

DHEAS, D4A, testosterone, 17OHP: Basal values • Within normal limits for CA

Increased end-organ sensitivity

• Within the range of pubertal stage

Idiopathic PP Exaggerated adrenarche

• Above the values for pubertal stage and ∆BA−CA > 1

Defective adrenal steroidogenesis

Synachten test 60 917OH Progesterone > 30 nmol/L 60 917OH Pregnenolone ≥ 294 nmol/L Virilizing tumor f0015

Figure 3 Premature pubarche (pp): diagnostic steps.

NC-CAH (CYP21) NC-CAH (3b-HSD)

105

Adrenarche, Premature

intervention is not required. A follow-up of children with PP must be assured, especially during puberty and the immediate postpubertal years. In general, both the child and the parents should be made aware of the good prognosis of most patients. Instructions concerning dietary habits and exercise should be given in cases with borderline or elevated BMI or a positive family history of DM2. The physician should be aware of the possible long-term consequences of PP and should individualize the care and follow-up, keeping in mind that PP, in certain cases, might not represent an innocent temporal deviation of sexual maturation but, possibly, an early manifestation of a complex metabolic abnormality. Fasting glucose/insulin ratio 6000) with left ventricular dysfunction following myocardial infarction (MI). As contrasted with placebo, the addition of eplerenone to standard therapy with ACE inhibitors or AT1 receptor blocker, beta blocker, aspirin, and diuretics led to a significant reduction in overall mortality and in cardiovascular morbidity and mortality. This article provides a broader perspective of ALDO’s properties that contribute to the pathophysiology of CHF and are mediated by receptor–ligand binding. They include ALDO’s influence on magnesium (Mg2þ) homeostasis and the role of extracellular Mg2þ ([Mg2þ]o) in regulating adrenal ALDO secretion, vascular remodeling and immune cell activation, endothelial cell dysfunction, and the central nervous system.

MG2þ HOMEOSTASIS AND ALDOSTERONE SECRETION p0025

The distribution of Mg2þ within body tissues is as follows: 53% in bone, 27% in skeletal muscle, and 19% in other soft tissues such as the heart. Less than 1% of total body Mg2þ is present in blood, and cytosolicfree Mg2þ ([Mg2þ]i) represents only 0.5 to 5% of total cellular Mg2þ, with approximately 80% bound to adenosine triphosphate (ATP) and other phosphometabolites sequestered within organelles such as mitochondria and endoplasmic reticulum. [Mg2þ]i homeostasis is maintained by exchange with these intracellular stores, whereas [Mg2þ]o is held within narrow limits by Mg2þ efflux from tissue stores.

Epithelial Cells In 1955, Mader and Iseri reported that a patient with adrenal adenoma had experienced spontaneous

Aldosterone in Congestive Heart Failure

episodes of hypomagnesemia together with enhanced Mg2þ excretion in urine and stool. Others would also note the presence of hypomagnesemia in patients with primary aldosteronism, suggesting that Mg2þ deficiency accompanies long-standing, ALDO-induced Mg2þ excretion. These clinical findings implicated ALDO in regulating both Kþ and Mg2þ excretion in classic target tissues. In 1962, Horton and Biglieri addressed urinary Kþ and Mg2þ excretion in five patients with PAL, where each patient had low normal or reduced serum Mg2þ levels. The influence of surgical resection of adrenal adenoma on urinary Kþ and Mg2þ excretion was assessed in two patients (Fig. 1). In both patients, there was an immediate and marked fall in excretion of these monovalent and divalent cations after surgery together with a gradual normalization of plasma Mg2þ. A lower basal excretion of Mg2þ, comparable to values seen for normal controls on the same Mg2þ diet, was seen postoperatively. In another patient, spironolactone was shown to reduce both urinary Kþ and Mg2þ excretion, which returned to previous increased basal levels following discontinuation of spironolactone (Fig. 2). Thus, the importance of ALDO in promoting urinary Mg2þ excretion was evident. In 1963, Conn, who had earlier coined the term PAL to connote autonomous adrenal ALDO production independent of plasma renin activity, concluded that hypomagnesemia, hypokalemia, hypernatremia, hypochloremia, and metabolic alkalosis were cardinal metabolic abnormalities of PAL. The importance of the adrenal cortex in regulating Mg2þ and Kþ excretion was further underscored in adrenal insufficiency or following adrenalectomy. ALDO treatment of dogs or rats with surgically induced bilateral adrenalectomy was shown to increase fecal Kþ and Mg2þ excretion and to normalize their plasma concentrations. Horton and Biglieri treated an adrenalectomized patient with d-ALDO for 8 days (Fig. 3). On day 5 of this regimen, spironolactone cotreatment was initiated. Exogenous ALDO promoted a prompt elevation in urinary Kþ and Mgþ excretion that was abrogated by receptor antagonist. Thus, the body of evidence is compelling that ALDO promotes both Kþ and Mg2þ excretion at classic target tissue sites. Mg2þ excretion is likely increased in patients with CHF, where elevated plasma ALDO levels are expected. In patients with CHF treated with a loop diuretic, there exists an independent stimulus to urinary Mg2þ excretion and the potential for exaggerated Mg2þ loss. However, its serum concentration does not reflect intracellular Mg2þ, and methods to detect

153

Aldosterone in Congestive Heart Failure

SERUM K mEq/L 100

3.3 3.6

3.6 3.6 3.4 2.8 3.3

4.2

4.5 SURGERY INTAKE

URINARY K mEq/24 h

50

0 SURGERY SERUM Mg mEq/L

1.5

1.5

1.8

1

3

5

1.65

1.7

1.9 1.9

10

URINARY Mg mEq/24 h

5

DAYS f0005

0

7

9

11

13

15

1

3

5

7

9

11

13

15

17

Figure 1 Urinary Kþ and Mg2þ excretion in two patients with primary aldosteronism before and after surgical removal

of adrenal adenoma. Serum Kþ and Mg2þ levels are shown for one of these individuals. Reproduced from Horton and Biglieri (1962). Effects of aldosterone on the metabolism of magnesium. J. Clin. Endocrinol. Metab. 22, 1187–1192. ß The Endocrine Society.

biologically active, cytosolic-free [Mg2þ]i are not widely used; therefore, the assessment of this important clinical problem remains to be defined. Intracellular Mg2þ deficiency may contribute to morbid and mortal events such as sudden cardiac death, which occurs in 50% of patients with CHF. The 30% reduction in risk of sudden cardiac death observed in the RALES trial may be related, in part, to spironolactone’s ability to restore and preserve Mg2þ homeostasis.

lymphocytes in exchange for Naþ via receptor–ligand binding. This Naþ-dependent response, measured by mag-fura-2, a fluorescent probe, involves both transcription and protein synthesis, as demonstrated by its respective abrogation by cycloheximide and actinomycin D. Lymphocyte-ionized [Mg2þ]i is reduced in patients with PAL secondary to either adrenal adenoma or hyperplasia. In uninephrectomized rats treated with ALDO by implanted mini-pump, PBMC [Mg2þ]i is significantly reduced and accompanied by immune cell activation (vide infra).

Nonepithelial Cells ALDO regulates Mg2þ exchange by nonepithelial cells, such as peripheral blood mononuclear cells (PBMCs), where it binds to a single class of cytosolic receptors. In patients with either PAL or renin-dependent secondary aldosteronism (SAL), ALDObinding sites are reduced by 50%. Following surgical removal of adenomatous adrenal tissue, ALDO receptor binding is normalized in these cells. ALDO promotes the efflux of Mg2þ from cultured human

Aldosterone Secretion [Mg2þ]o participates in the regulation of adrenal ALDO secretion to create a pathway of reciprocal regulation in Mg2þ homeostasis (Fig. 4). In healthy normotensive men and women, a 3-h intravenous infusion of magnesium sulfate (MgSO4) suppresses plasma ALDO levels. On the other hand, dietaryinduced Mg2þ deficiency with reduced [Mg2þ]o is accompanied by an expanded width to the adrenal

p0050

154

Aldosterone in Congestive Heart Failure

d-ALDOSTERONE 100 mg every 6 h intramuscularly SPIRONOLACTONE 1g/day

SPIRONOLACTONE 1 g/day

100 100

INTAKE URINARY K mEq/24 h

INTAKE

50

URINARY K 50 mEq/24 h

0 0 7 URINARY Mg mEq/24 h

9

5

7 URINARY Mg mEq/24 h

3

3

1 DAYS f0010

5

1 þ

3

5

7

9

2þ

Figure 2 Urinary K and Mg

excretion in a patient with primary aldosteronism before, during, and after oral spironolactone treatment. Reproduced from Horton and Biglieri (1962). Effects of aldosterone on the metabolism of magnesium. J. Clin. Endocrinol. Metab. 22, 1187–1192. ß The Endocrine Society.

zona glomerulosa together with hyperplasia of the renal juxtaglomerulosa cells, increased adrenal ALDO secretion, increased plasma ALDO, and a reduction in urinary Naþ/Kþ ratio. Dietary Mg2þ deficiency combined with a high-Naþ diet attenuates, but does not abrogate, heightened ALDO secretion. In cultured zona glomerulosa cells, superfusate [Mg2þ]o regulates ALDO production; high [Mg2þ]o suppresses, whereas an [Mg2þ]o-free media augments, their elaboration of ALDO. The SAL that accompanies dietary Mg2þ deficiency is associated with a timedependent rise in [Naþ]i and [Ca2þ]i in heart, skeletal muscle, kidney, and bone and is suggestive of an inhibition of Na,K-ATPase, an Mg2þ-dependent pump, and increased Naþ/Ca2þ exchange at these sites.

NEUROHORMONE–IMMUNE INTERFACE A structural remodeling of the cardiovasculature by fibrous tissue accompanies aldosteronism derived from either endogenous or exogenous sources. This fibrogenic phenotype includes intramural arteries of the heart, kidney, pancreas, mesentery, and vaso

1 DAYS

1 þ

3

5

7

9

11

2þ

Figure 3 Urinary K and Mg excretion in a patient with bilateral adrenalectomy before and during aldosterone treatment (without and then with oral spironolactone coadministration). Adapted from Horton and Biglieri (1962). Effects of aldosterone on the metabolism of magnesium. J. Clin. Endocrinol. Metab. 22, 1187–1192. ß The Endocrine Society.

f0015

ADRENAL ALDO SECRETION

Plasma ALDO Na+ dependent ALDO

ALDO

Epithelial cell Mg2+ excretion Nonepithelial cell Mg2+ efflux Na+/Mg2+ exchangers Mg2+/Ca2+ exchangers Na,K-ATPase

Mg2+ Balance [Mg2+]o

Figure 4 Mg2þ balance and reciprocal regulation of adrenal ALDO secretion. Reproduced with permission from Weber, K. T. (2003). Aldosterone revisited: Perspectives on less well recognized actions of aldosterone. J. Lab. Clin. Med. (in press).

f0020

Aldosterone in Congestive Heart Failure

vasorum of the aorta and pulmonary artery. Cotreatment with a receptor antagonist (e.g., spironolactone, eplerenone), in either nondepressor or depressor doses, prevents this remodeling, indicating its independence of elevations in blood pressure. In a substudy of the RALES trial, survival benefit was associated with a reduction in circulating markers of collagen synthesis that presumably reflected an attenuation in ongoing vascular fibrosis. In this connection, urinary excretion of hydroxyproline, a marker of collagen turnover, is increased in adrenalectomized rats treated with ALDO, 1% dietary sodium chloride (NaCl), and cortisone. On the other hand, glucocorticoids reduce urinary hydroxyproline excretion, and their inhibition of collagen formation in bone is associated with osteoporosis. In 1995, Campbell and colleagues found that the perivascular fibrosis of the coronary vasculature that ultimately appears in aldosteronism is preceded by a proinflammatory vascular phenotype that features invading monocytes/macrophages and lymphocytes and adhesion molecule expression. These findings, including their independence of blood pressure, were confirmed by others more recently. An interrogation of molecular responses involved in the invasion of coronary vessels by these inflammatory cells was addressed in animal models of PAL and SAL. In rats receiving ALDO/salt treatment (ALDOST), where plasma renin and angiotensin II both are suppressed, Sun and colleagues more recently tested the hypothesis that oxidative stress was involved in the appearance of the proinflammatory/fibrogenic cardiac phenotype. At week 3 of ALDOST, there was no evidence of cardiac pathology. However, at weeks 4 and 5, inflammatory cells (monocytes/macrophages and lymphocytes) were found to have invaded intramural coronary vessels in both ventricles. By immunohistochemistry, invading PBMCs were found to express the gp91phox subunit of nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, whose activation is a major source of superoxide in leukocytes, and 3-nitrotyrosine, a product of peroxynitrite with stable protein tyrosine residues and where peroxynitrite is formed by the reaction between superoxide and nitric oxide. The RelA subunit of nuclear factor-kB (NF-kB), a redoxsensitive transcription factor integral to inflammatory responses, was likewise activated in these cells. In situ hybridization localized increased mRNA expression of intercellular adhesion molecule-1 (ICAM-1), monocyte chemoattractant protein-1 (MCP-1), and proinflammatory cytokine tumor necrosis factor-a (TNF-a) at vascular sites that involved the normotensive nonhypertrophied right atrium and ventricle and left

155 atrium as well as the hypertensive hypertrophied left ventricle. Cotreatment with either spironolactone or an antioxidant (either pyrrolidine dithiocarbamate or N-acetylcysteine) prevented the appearance of these cells and associated molecular responses as well as the subsequent perivascular fibrosis. Thus, ALDOST induces oxi/nitrosative stress within inflammatory cells invading the intramural coronary vasculature, and it is this proinflammatory vascular phenotype that leads to intramural coronary artery pathology and the subsequent perivascular fibrosis. What is responsible for the induction of oxi/nitrosative stress in PMBCs? Gerling and colleagues identified an ALDO-mediated reduction in [Mg2þ]i that appears in PBMCs at week 1 of ALDOST, long before the appearance of cardiac pathology at week 4. The early activation of these cells is evident in their transcriptome (expressed genes) and proteome (expressed proteins). The interrogation of complex molecular events that account for immune cell activation and subsequent homing of these cells to the heart may explain why cardiac pathology does not appear until week 4 of ALDOST and calls into question the prospect of an autoimmune response similar to that which can follow MI. In rodents treated with a Mg2þ-deficient diet, a putative state of exaggerated aldosteronism, lymphocyte Mg2þ is reduced to an extent comparable to the Mg2þ depletion that appears in skeletal muscle and cardiac tissue. Weglicki and colleagues identified an early (week 1) induction of oxi/nitrosative stress and depletion of antioxidant defenses in PBMCs and endothelial cells. Lymphocyte activation that appears includes their production of proinflammatory cytokines and a neurogenic peptide, substance P, together with the expression of its receptors. Cardiac lesions are first seen in this model at week 3 and can be prevented by a substance P receptor antagonist. The potential for the SAL that accompanies human CHF to likewise be accompanied by reduced [Mg2þ]i and immune cell activation remains unknown. The prospect does exist for an immune cell origin to the ‘‘cytokine storm’’ characteristic of CHF that features elevations in circulating proinflammatory cytokines such as TNF-a and interleukin-6 (IL6). Cells of the monocyte–phagocyte system are a potent source of these cytokines. Another source of cytokine production in heart failure is the central nervous system. Irrespective of their origin, prolonged elevations in these proinflammatory cytokines contribute to the progressive systemic illness that accompanies CHF and features tissue wasting to eventuate in cardiac cachexia.

156

DYSFUNCTION OF THE ENDOTHELIUM p0080

In patients with PAL or renal artery stenosis with SAL, forearm vasomotor reactivity to endothelial celldependent acetylcholine is diminished compared with that in normotensive controls, whereas nonendothelial cell-dependent sodium nitroprusside-induced vasodilatation is preserved. Following surgical removal of adrenal adenoma, the impairment in endothelial celldependent vasodilation is restored. In the SAL that accompanies CHF, diminished forearm vasomotor reactivity to acetylcholine is normalized by spironolactone treatment. Acetylcholine-induced, nitric oxidedependent vasorelaxation of aortic rings is reduced in rats following MI. This vasomotor dysfunction, together with increased superoxide formation by aortic tissue, is normalized by spironolactone alone or in combination with an ACE inhibitor. In cultured aortic endothelial cells, reduced Mg2þ concentration of culture medium is associated with increased oxidant production and reduced intracellular glutathione, an antioxidant reserve consumed in neutralizing oxi/ nitrosative stress. Abnormal Mg2þ homeostasis may account for the underlying pathophysiological basis of endothelial dysfunction seen in either PAL or SAL.

CENTRAL NERVOUS SYSTEM

p0090

ALDO receptors are found at diverse sites in the central nervous system. These include epithelial cells of the choroid plexus. A role for ALDO in the genesis of idiopathic intracranial hypertension (IIH) has been proposed given the reported association between IIH and PAL and SAL as well as the prevalence of headaches among patients with PAL. That this proposition does not apply to all patients with IIH is underscored by its appearance in patients with adrenal insufficiency unless ALDO is produced in situ within the central nervous system, as is now recognized to be the case for the cardiovasculature. The choroid plexus, a site of high-affinity ALDO receptor binding, is involved in the production of cerebrospinal fluid (CSF) and is a target site for ALDO, spironolactone, and ouabain, an endogenous, digitalis-like substance released by the adrenals and the hypothalamic–pituitary axis. ALDO exerts its biological actions on epithelial cells by enhancing the activity and number of Na,K-ATPase pumps in their apical membrane. An ouabain-sensitive Na,K-ATPase is present in the microvilli of the plexus and is involved in the regulation of CSF formation and electrolyte composition (e.g., ouabain reduces CSF

Aldosterone in Congestive Heart Failure

production). ALDO is present in CSF, where its concentration correlates with plasma levels. Either the systemic or intracerebroventricular administration of a mineralocorticoid (ALDO or deoxycorticosterone [DOC]) is accompanied by a fall in CSF Kþ, together with a rise in arterial pressure, without changes in blood volume, cardiac output, plasma catecholamines, or vasopressin. This hypertensive response is abrogated by intracerebroventricular infusion of Kþ or a mineralocorticoid receptor antagonist. Therefore, ALDO has a central action involved in the regulation of blood pressure as well as CSF volume and composition. Produced locally within the brain, ALDO’s paracrine properties may likewise contribute to blood pressure regulation. ALDO’s central actions, which may contribute to the pathophysiology of CHF, were reviewed and expanded by Felder and colleagues. The hypothalamic paraventricular nucleus (PVN), a forebrain site involved in the regulation of extracellular volume and sympathetic nerve activity, is governed by circulating neurohormones and effector signals originating from the brainstem. In rats with MI induced by coronary artery ligation, the activity of the PVN is increased. Systemic or intracerebroventricular administration of spironolactone reduces this activity, improves baroreflex regulation of renal sympathetic nerve activity (albeit in a time-dependent manner), and prevents the increase in Naþ appetite and decline in urinary Naþ and H2O excretion that appear in this model. Plasma levels of TNF-a rise progressively over weeks 1 to 3 following MI, a response abrogated by intracerebroventricular infusion of spironolactone started 24 h after coronary ligation, suggesting that central ALDO receptor activation is involved in regulating the release of this proinflammatory cytokine. However, the cellular source of TNF-a remains uncertain and may include central and/or peripheral tissues.

p0095

FUTURE DIRECTIONS Since its discovery more than 50 years ago, ALDO has had a well-established importance in clinical medicine, including its role in CHF. The past decade or so has witnessed a resurgence of interest in the adrenal’s most potent mineralocorticoid as well as its de novo production through steroidogenesis within the cardiovasculature and brain. An ever-expanding role for this steroid molecule in the metabolism of monovalent and divalent cations by epithelial and nonepithelial cells has warranted an even broader perspective of its portfolio of actions. The many peripheral and central actions of ALDO that can contribute to the

p0100

Aldosterone in Congestive Heart Failure

pathophysiology of CHF syndrome remain to be defined. This is no more evident than in the adverse structural remodeling of the heart and systemic organs that accompanies chronic elevations in plasma ALDO (inappropriate relative to dietary Naþ) and that may be secondary to PBMC activation induced by [Mg2þ]i depletion and transduced by oxi/nitrosative stress. Molecular mechanisms involved in immune cell responses remain to be elucidated. Today’s technologies will permit an assessment of ALDO’s role in altering the molecular phenotype of these immune cells, specifically their transcriptome and proteome. Such insights may provide for the development of serologic biomarkers that address the risk, onset, and progression of vascular injury in CHF and could lead the way toward refined and even newer drug targets.

Acknowledgments The editorial assistance of Richard A. Parkinson is appreciated. This work was supported, in part, by NIH R01-DK62403 and R24-RR-15373 (I.C.G), NHLBI R01-HL67888 (Y.S.), NHLBI R01-HL62229 (K.T.W.), and grants from the Center of Excellence in Connective Tissue Diseases (Y.S. and K.T.W.) at the University of Tennessee Health Science Center.

See Also the Following Articles Aldosterone Receptors . Atrial Natriuretic Factor and Family of Natriuretic Peptides . Hypertension, Endocrine . Immune System, Hormonal Effects on . Mineralocorticoids and Mineralocorticoid Excess Syndromes . Primary Aldosteronism (PAL) . Tissue Renin-Angiotensin-Aldosterone System

Further Reading Campbell, S. E., Janicki, J. S., and Weber, K. T. (1995). Temporal differences in fibroblast proliferation and phenotype expression

157 in response to chronic administration of angiotensin II or aldosterone. J. Mol. Cell. Cardiol. 27, 1545–1560. Conn, J. W. (1963). Aldosteronism in man: Some clinical and climatological aspects. J. Am. Med. Assoc. 183, 871–878. Felder, R. B., Francis, J., Zhang, Z. H., Wei, S. G., Weiss, R. M., and Johnson, A. K. (2003). Heart failure and the brain: New perspectives. Am. J. Physiol. 284, R259–R276. Horton, R., and Biglieri, E. G. (1962). Effect of aldosterone on the metabolism of magnesium. J. Clin. Endocrinol. Metab. 22, 1187–1192. Kagawa, C. M., Cella, J. A., and Van Arman, C. G. (1957). Action of new steroids in blocking effects of aldosterone and deoxycorticosterone on salt. Science 126, 1015–1016. Lo¨sel, R. M., Feuring, M., Falkenstein, E., and Wehling, M. (2002). Nongenomic effects of aldosterone: Cellular aspects and clinical implications. Steroids 67, 493–498. Mader, I. J., and Iseri, L. T. (1955). Spontaneous hypopotassemia, hypomagnesemia, alkalosis, and tetany due to hypersecretion of corticosterone-like mineralocorticoid. Am. J. Med. 19, 976–988. Pitt, B., Remme, W., Zannad, F., Neaton, J., Martinez, F., Roniker, B., Bittman, R., Hurley, S., Kleiman, J., and Gatlin, M. (2003). Eplerenone, a selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction. N. Engl. J. Med. 348, 1309–1321. Pitt, B., Zannad, F., Remme, W. J., Cody, R., Castaigne, A., Perez, A., Palensky, J., and Wittes, W. (1999). The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N. Engl. J. Med. 341, 709–717. Selye, H. (1946). The general adaptation syndrome and the diseases of adaptation. J. Clin. Endocrinol. 6, 117–230. Sun, Y., Zhang, J., Lu, L., Chen, S. S., Quinn, M. T., and Weber, K. T. (2002). Aldosterone-induced inflammation in the rat heart: Role of oxidative stress. Am. J. Pathol. 161, 1773–1781. Tait, J. F., and Tait, S. A. S. (1990). A decade (and even more) of aldosterone and other adrenal steroids. In ‘‘Endocrine Hypertension’’ (E. G. Biglieri and J. C. Melby, eds.), pp. 5–27. Raven Press, New York. Weber, K. T. (2001). Aldosterone in congestive heart failure. N. Engl. J. Med. 345, 1689–1697. Weglicki, W. B., Mak, I. T., Stafford, R. E., Dickens, B. F., Cassidy, M. M., and Phillips, T. M. (1994). Neurogenic peptides and the cardiomyopathy of magnesium-deficiency: Effects of substance P-receptor inhibition. Mol. Cell. Biochem. 130, 103–109.

Aldosterone Receptors Marc Lombe`s and Maria-Christina Zennaro INSERM U 478, Xavier Bichat, Paris, France

Glossary aldosterone A steroid hormone synthesized in the zona glomerulosa of the adrenal cortex; the major mineralocorticoid hormone in humans. mineralocorticoid Corticosteroid hormone secreted by the adrenal gland and exerting its function through the mineralocorticoid receptor; also referred to as the hormone that affects water and electrolyte homeostasis.

A

ldosterone receptors, or mineralocorticoid receptors, are defined as intracellular proteins that are able to bind aldosterone and mediate hormone action within target cells.

INTRODUCTION The aldosterone receptor, also referred to as the mineralocorticoid receptor (MR), is a member of the nuclear receptor superfamily that acts as a liganddependent transcription factor mediating aldosterone effects on a variety of target tissues. These include epithelial cells in the kidney and colon but also nonepithelial cells in the cardiovascular and central nervous systems.

MECHANISM OF ALDOSTERONE ACTION The classical model of aldosterone action is illustrated schematically in Fig. 1. Aldosterone penetrates a target cell, typically a polarized epithelial cell in the distal nephron and presumably by passive diffusion, and specifically binds to the MR. In its unliganded state, the MR resides predominantly in the cytoplasm and is complexed with various receptor-associated proteins, including a dimer of heat shock protein 90 (hsp90), heat shock protein 70 (hsp70), and other proteins such as immunophilins, cyclophilin (Cyp40), and FKBP52, which are known to bind immunosuppressive agents.

158

On aldosterone binding, the MR undergoes conformational change that leads to the dissociation of receptor-associated proteins. Aldosterone receptor complexes then translocate to the nuclear compartment and bind as homodimers to specific DNA sequences that are known as mineralocorticoid response elements (MREs) located in the regulatory regions of aldosterone-sensitive genes. The consensus sequence of MREs generally consists of an inverted hexameric palindrome separated by three nucleotides (AGAACAnnnTGTTCT). The MR is then able to recruit specific coactivators, in a sequential and/or combinatorial manner, that subsequently enhance transcriptional activation through direct interaction with the basal transcription factors and chromatin remodeling involving histone acetylation/methylation. Several aldosterone-regulated genes have been identified, including the serum- and glucocorticoid-inducible kinase (sgk1), a serine threonine kinase that is able to phosphorylate the ubiquitin ligase Nedd4-2, which in turn controls the retrieval of the subunits of the epithelial sodium channel from the apical membrane of the cell. Other aldosterone-induced proteins, such as the small monomeric GTP-binding protein Kirsten Ras (Ki-Ras), the glucocorticoid-induced leucine zipper protein (GILZ), and the N-myc down-regulated gene 2 (NDRG2), also seem to play an important role during the early phase of aldosterone responses in the renal tubule. Collectively, aldosterone stimulates the biosynthesis and activity of sodium channels and pumps, leading to an enhanced vectorial transepithelial sodium transport from the tubular lumen to the basolateral space. Another important issue in the mechanism of aldosterone action is the apparent nonselectivity of the MR. Indeed, aldosterone and glucocorticoids such as cortisol are equally able to bind to the MR with high affinity (Kd in the nanomolar range). However, the enzyme 11b-hydroxysteroid dehydrogenase type 2 (11b-HSD2), by metabolizing 11b-hydroxysteroids into 11 keto-derivatives such as cortisone, which exhibits no affinity for the MR, plays a pivotal role in

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

159

Aldosterone Receptors

Figure 1 Schematic representation of the molecular mechanism of aldosterone action in a polarized epithelial cell. MR, mineralocorticoid receptor or aldosterone receptor; hsp90 and hsp70, heat shock protein 90 and 70, respectively; MREs, mineralocorticoid response elements; ENaC, epithelial sodium channel; Sgk1, serum- and glucocorticoid-inducible kinase; NDRG2, N-myc down-regulated gene 2; Ki-Ras, small monomeric GTP-binding protein Kirsten Ras; GILZ, glucocorticoidinduced leucine zipper protein; Naþ,Kþ-ATPase, sodium/potassium ATPase; 11b-HSD2, 11b-hydroxysteroid dehydrogenase.

mineralocorticoid selectivity in epithelial cells, preventing permanent occupancy of the MR by the more prevalent glucocorticoid hormones.

STRUCTURE OF THE MR The cloning of the human MR cDNA in 1987 by Ron Evans’s laboratory facilitated the deduction of its primary structure and the definition of different functional domains. Like other members of the nuclear receptor superfamily of which it is a member, the MR displays a common modular structure and is composed of three distinct functional domains. A schematic representation of the MR is given in Fig. 2. The amino terminal region, also referred to as the transactivation domain, is 602 amino acids long and so constitutes the longest domain among nuclear receptor family members. Its primary sequence bears little resemblance to that of other family members, sharing less than 15% homology with its closely related receptor, the glucocorticoid receptor. However, among all known mammalian species, more than 85%

of this MR domain is highly conserved, suggesting that it contains specific and important functions. This domain harbors two ligand-independent activation functions: AF1a located in aa 1 to 167 and AF1b spanning aa 445 to 602 and presumably an inhibitory transactivation region between aa 167 and aa 437. The centrally located, highly conserved DNAbinding domain is responsible for the specific interaction with hormone response elements located in the promoters of aldosterone target genes. This domain has a rigid structure, is highly hydrophilic, is very rich in cysteine residues, and is composed of two zinc finger structures. It contains a P box that constitutes the interacting contact with the half-site of the response element and a D box that is responsible for weak dimerization. A nuclear export signal has also been identified between the two zinc fingers, and a weak ligand-independent nuclear localization signal, NSL1, has been shown to be located next to the C-terminal site of the DBD. The ligand-binding domain at the carboxyterminus part of the receptor is 250 amino acids

160

Aldosterone Receptors

Figure 2 Different functional domains of the human mineralocorticoid receptor. Numbering represents the amino acids. NTD, N-terminal domain of the receptor; DBD, DNA-binding domain; LBD, ligand-binding domain; AF, activating function; NLS, nuclear localization signal; NES, nuclear export signal; D Box, dimerization box; hsp90, heat shock protein 90.

long and is separated from the DBD by a hydrophilic proline-rich hinge region. The complex C-terminal domain is responsible for ligand binding and contains a ligand-dependent nuclear localization signal (NLS2), multiple contact sites for hsp90 interaction, and a ligand-dependent activating function AF2 domain. This domain is highly structured and is composed of 12 a-helices (H1–H12) and one b antiparallel sheet on which the steroid hormone lies. Aldosterone binds to the MR with high affinity, but glucocorticoids such as cortisol bind to the receptor with equivalent affinity. Although similar dissociation constants (Kd) have been calculated by Scatchard plot analysis, dissociation rates (k  1) are much faster for glucocorticoids than for aldosterone. This intrinsic property of the MR to discriminate between aldosterone and glucocorticoids constitutes an additional molecular mechanism that ensures the mineralocorticoid selectivity of aldosterone action within target cells from a dynamic point of view. Finally, it has been shown that the aldosterone–MR complex presumably adopts conformation different from that of the glucocorticoid–MR complex, whereby distinct interaction between the N-terminal domain and the LBD occurs. This leads to the recruitment of particular coactivators resulting in a highly specific transcriptional response. The three-dimensional structure of the MR was deduced by using an analogy of the crystal structures of the ligand-binding domain of other steroid hormone receptors. This facilitated the precise definition of the amino acid contacts with the functional groups of the steroid. Thus, the 3-ketone function of aldosterone interacts with the glutamine residue at position 776 and the arginine residue at position 817. On the other hand, the 20-ketone function contacts the cysteine residue at position 942, and the

21-hydroxyl and 18-hydroxyl groups are anchored by the asparagine residue at position 770. On ligand binding, the helix 12 within the AF2 domain rotates tightly against the LBD, and this, together with the changes in helices 3 to 5, facilitates the interaction of the receptor with coactivators of the steroid receptor coactivator (SRC) family.

CELL-SPECIFIC EXPRESSION AND SUBCELLULAR LOCALIZATION OF THE MR The tissue-specific expression of the MR is presented in Table I. The MR is essentially expressed at relatively high levels in polarized cells of sodium-transporting epithelia in the distal parts of the nephron (from the cortical part of the thick ascending limb of Henle’s loop, distal tubule, and connecting tubule to the cortical and medullary collecting tubule), colon, pneumocytes, and salivary and sweat glands. It is now well established that the MR is also present in ‘‘nonclassical’’ aldosterone target tissues, most notably neurons of the hippocampus, cardiomyocytes, adipocytes, vascular smooth muscle cells, and (presumably) other cell types. With respect to intracellular localization, initial immunocytochemical studies demonstrated that the MR is predominantly located in the cytoplasmic compartment in the absence of ligand and is translocated into the nucleus on aldosterone exposure. Experiments using green fluorescent protein–MR chimeras allowed the subcellular localization of the MR and its kinetics in living cells to be examined. As illustrated in Fig. 3, on aldosterone binding, the MR rapidly translocates in the nucleus within minutes and is sequestered in specific areas within this

161

Aldosterone Receptors

Table I Tissue-Specific Expression of Mineralocorticoid Receptor Epithelial cells Distal nephron Colon Salivary glands Sweat glands Pneumocytes Nonepithelial cells Neurons Cardiomyocytes Adipocytes Smooth muscle cells

compartment, where transcriptionally active regions of the chromatin presumably exist.

REGULATION OF MR EXPRESSION p0060

The human MR (hMR) gene is localized to the q31.2 region of chromosome 4 and spans approximately 450 kb. The hMR gene is composed of 10 exons, including 2 untranslated first exons referred to as exon 1a and exon 1b (Fig. 4). Alternative transcription of these two 50 -untranslated exons generates two mRNA isoforms: hMRa and hMRb. Given that the hMR translation initiation site, as defined by the start codon ATG, is located 2 bp downstream from the beginning of exon 2, these two isoforms give rise to the same translation product. The last 8 exons, from 2 to 9, encode the various functional domains of the protein. Exon 2 codes for the N-terminal domain of the receptor. The two small exons 3 and 4 encode each of the two zinc fingers of the DNAbinding domain, whereas the last 5 exons encode the ligand-binding domain of the receptor. However, the existence of other hMR splice variants has been demonstrated, and this seems to play a major role in modulating receptor function. The hMR gene expression is controlled by two different promoters that differ in terms of their basal activity as well as their hormonal regulation. Experiments in transgenic mice have shown distinct tissue-specific use and activity of these two hMR regulatory regions in vivo. The proximal P1 promoter corresponding to the 50 -flanking region of exon 1a is a relatively strong promoter that is transcriptionally active in all aldosterone target tissues, whereas the distal P2 promoter flanking exon 1b is weaker and has a more restricted pattern of expression; thus, it is presumably used during specific developmental stages or physiological situations.

Figure 3 Aldosterone-dependent nuclear translocation of the hMR. RCSV3 rabbit renal cells were transfected with the expression vector encoding for a chimera protein consisting of human mineralocorticoid receptor and enhanced green fluorescent protein EGFP. The intracellular distribution of EGFP–hMR in living cells was observed directly on an inverted IRB microscope. In the absence of aldosterone, the receptor is predominantly in the cytoplasm, whereas 15 min after addition of 108 M aldosterone, the receptor is localized in the nucleus.

PATHOPHYSIOLOGICAL EFFECTS OF ALDOSTERONE Aldosterone is primarily implicated in the maintenance of water and salt homeostasis by regulating sodium reabsorption and potassium excretion across tight epithelia. As such, aldosterone plays a key role in the regulation of blood pressure, and in turn, the dysfunctional regulation of aldosterone secretion and action is implicated in many human diseases such as hypertension and heart failure. In addition to the renal effects of aldosterone, it has become evident that aldosterone exerts direct effects on the cardiovascular system. Aldosterone excess leads to the development of cardiac hypertrophy and fibrosis, which are involved in cardiac remodeling and heart failure. Even though the molecular mechanisms remain obscure, the detrimental effects of aldosterone on cardiovascular function led to major clinical trials aimed at demonstrating beneficial effects of antimineralocorticoid compounds. Initially, the RALES study demonstrated the efficiency of spironolactone treatment by showing that it significantly reduces the morbidity and mortality (by 30%) of patients with severe congestive heart failure. A fairly new selective aldosterone antagonist, eplerenone, has been approved by the Food and Drug Administration for the treatment of high blood pressure. This antimineralocorticoid has reduced progestagenic and antiandrogenic activities more than has spironolactone. The EPHESUS study also clearly demonstrated the beneficial effects of eplerenone in patients with acute myocardial infarction complicated by left ventricular dysfunction and heart failure. It is likely that aldosterone receptor

162

Aldosterone Receptors

Figure 4 Structure of human mineralocorticoid receptor gene, mRNA isoforms, and hMR protein. P1 and P2 are the alternative promoters of the hMR gene. Numbering represents the exons. ATG and TGA are the starting and stop codons, respectively.

blockade preventing renal, cerebral, and vascular injuries will become increasingly important in managing hypertension, heart failure, and atherosclerosis as well as for patients during the postmyocardial infarction period.

GENETIC ALTERATIONS OF THE MR There is now clear evidence that genetic alterations of the MR are associated with human diseases. The first MR mutations were found in patients with autosomal dominant or sporadic pseudohyoaldosteronism type I, an inherited disorder characterized by renal salt wasting during infancy and associated with failure to thrive, hyponatremia, hyperkalemia, and high plasma aldosterone levels. Therefore, these clinical and biological features are consistent with aldosterone resistance. These heterozygous frameshift, nonsense, or missense mutations occur within different functional domains of the MR receptor that, in turn, affect receptor function in different ways, generally resulting in receptor inactivity. Conversely, a gain of function mutation in the MR has been described in a family with severe early-onset hypertension that is exacerbated by pregnancy. The mutation, a substitution of leucine for serine at codon 810, lies within the ligand-binding domain and has

been shown to drastically modify the receptor steroid specificity. Indeed, further experiments demonstrated a constitutive MR activation in the absence of ligand. In addition, progesterone, spironolactone, and even cortisone were able to fully activate the mutant receptor, consistent with the clinical presentation of gestational hypertension. Interestingly, various genetically engineered animals mimicking human disease have proven to be useful in terms of analyzing the in vivo function of the MR. MR gene inactivation achieved by homologous recombination leads to knockout mice developing symptoms of pseudohypoaldosteronism. In contrast, hMR overexpressing transgenic mice exhibit specific alteration in renal and cardiac function. Altogether, these animal models constitute attractive new experimental systems to further explore the widespread and pleiotropic function of aldosterone receptors in vivo and to decipher the molecular and cellular events underlying the aldosterone signaling pathway.

See Also the Following Articles Aldosterone in Congestive Heart Failure . Hypertension, Overview . Mineralocorticoids and Mineralocorticoid Excess Syndromes . Primary Aldosteronism (PAL) . Tissue Renin-Angiotensin-Aldosterone System

Aldosterone Receptors

Further Reading Arriza, J. L., Weinberger, C., Cerelli, G., Glaser, T. M., Handelin, B. L., Housman, D. E., and Evans, R. M. (1987). Cloning of human mineralocorticoid receptor complementary DNA: Structural and functional kinship with the glucocorticoid receptor. Science 237, 268–275. Berger, S., Bleich, M., Schmid, W., Cole, T. J., Peters, J., Watanabe, H., Kriz, W., Warth, R., Greger, R., and Schutz, G. (1998). Mineralocorticoid receptor knockout mice: Pathophysiology of Naþ metabolism. Proc. Natl. Acad. Sci. USA 95, 9424–9429. Fajart, J., Wurtz, J. M., Souque, A., Hellal-Levy, C., Moras, D., and Rafestin-Oblin, M. E. (1998). Antagonism in the human mineralocorticoid receptor. EMBO J. 17, 3317–3325. Farman, N., and Rafestin-Oblin, M. E. (2001). Multiple aspects of mineralocorticoid selectivity. Am. J. Physiol. Renal Physiol. 280, F181–F192. Geller, D. S., Farhi, A., Pinkerton, N., Fradley, M., Moritz, M., Spitzer, A., Meinke, G., Tsai, F. T., Sigler, P. B., and Lifton, R. P. (2000). Activating mineralocorticoid receptor mutation in hypertension exacerbated by pregnancy. Science 289, 119–123. Geller, D. S., Rodriguez-Soriano, J., Vallo Boado, A., Schifter, S., Bayer, M., Chang, S. S., and Lifton, R. P. (1998). Mutations in

163 the mineralocorticoid receptor gene cause autosomal dominant pseudohypoaldosteronism type I. Nat. Genet. 19, 279–281. Le Menuet, D., Isnard, R., Bichara, M., Viengchareun, S., MuffatJoly, M., Walker, F., Zennaro, M-C., and Lombe`s, M. (2001). Alteration of cardiac and renal functions in transgenic mice overexpressing human mineralocorticoid receptor. J. Biol. Chem. 276, 38911–38920. Le Menuet, D., Viengchareun, S., Penfornis, P., Walker, F., Zennaro, M-C., and Lombe`s, M. (2000). Targeted oncogenesis reveals a distinct tissue-specific utilization of alternative promoters of the human mineralocorticoid receptor gene in transgenic mice. J. Biol. Chem. 275, 7878–7886. Pitt, B., Remme, W., Zannad, F., Neaton, J., Martinez, F., Roniker, B., Bittman, R., Hurley, S., Kleiman, J., and Gatlin, M. (2003). Eplerenone, a selective aldosterone blocker, in patients with left ventricular dysfunction after myocardial infarction. N. Engl. J. Med. 348, 1309–1321. Stockland, J. D. (2002). News ideas about aldosterone signaling in epithelia. Am. J. Physiol. 282, F559–F576. Zennaro, M-C., Keightley, M-C., Kotelevtsev, Y., Conway, G., Soubrier, F., and Fuller, P. J. (1995). Human mineralocorticoid receptor genomic structure and identification of expressed isoforms. J. Biol. Chem. 270, 21016–21020.

Alkaline Phosphatase David Goltzman and Dengshun Miao McGill University, Montreal, Quebec, Canada

Glossary alkaline phosphatase A membrane-bound metalloenzyme that consists of a group of isoenzymes, all glycoproteins, encoded by at least four different gene loci: tissue nonspecific, intestinal, placental, and germ cell alkaline phosphatase; catalyzes the hydrolysis of a wide range of phosphomonoesters at alkaline pH. hypophosphatasia An inheritable disorder characterized by defective bone mineralization and a deficiency of tissue nonspecific alkaline phosphatase activity. isoenzymes Enzymes that have the same catalytic activity but differ slightly in amino acid sequence and/or posttranslational modifications.

S

ince its first description by Suzuki and colleagues in 1907, alkaline phosphatase (ALP) has been investigated continuously and extensively. For most of the past century, there has been widespread use of ALP activity in serum as an enzymatic signal for a variety of disease states involving, in particular, the liver and bone. Investigations directed at the molecular properties of the enzyme have been relatively recent. Quantitation of serum ALP activity has been a routine in hospital laboratories since the 1930s, and this test is perhaps the most frequently performed enzyme assay. Bone ALP became the clinically most relevant enzyme in the diagnosis of bone disease. In spite of its broad use as a clinical marker, the physiological function of this protein, which is ubiquitous in nature, is largely unknown. Identification of ALP gene mutations in hypophosphatasia, a rare heritable form of rickets, has confirmed that ALP functions importantly in skeletal mineralization in humans. Still, many questions remain.

GENETICS AND EXPRESSION Alkaline phosphatase (ALP) is a membrane-bound metalloenzyme that consists of a group of isoenzymes. Each isoenzyme is a glycoprotein encoded by different gene loci. At least four loci have been identified: tissue nonspecific, intestinal, placental, and germ cell ALP.

164

It is believed that the evolution of the ALP gene family has involved the duplication of a primordial tissue nonspecific ALP (TN-ALP) gene to create the TN-ALP gene and an intermediate intestinal ALP (IAP) gene, followed by additional duplications of the latter to create intestinal, placental, and germ cell ALP genes. Only humans and great apes have placental ALP; all other mammals have IAP. The gene encoding TN-ALP maps to the short arm of chromosome 1, bands p36.1–p34. It is expressed at its highest levels in liver, bone, and kidney (hence its alternate name L/B/K ALP) as well as at lower levels in various other tissues. Differential processing of the TN-ALP gene product occurs within the cell or during passage of the protein molecule out of its cell of origin. In this way, differential glycosylation of TN-ALP gives rise to tissue-specific isoforms. The gene encoding IAP is a member of the gene family mapping to the long arm of chromosome 2 (q34–q37). IAP is present at high levels in intestinal tissue and at trace levels in the kidney. In contrast to the other ALP isoenzymes, the carbohydrate side chains of IAP are not terminated by sialic acid. Distinct IAPs can be isolated from fetal and adult intestinal tissue, with the fetus forming a sialylated isoenzyme in contrast to the adult. The fetal and adult forms differ not only in the carbohydrate content but also in the protein moiety itself, suggesting that a separate ALP gene locus may exist in humans during fetal development. Fetal IAP is present in amniotic fluid and in the meconium. This fetal/embryonic gene has also been identified in cancer cells and is designated the Kasahara isoenzyme. The human placental ALP gene was also mapped to chromosome 2. It exhibits 87% homology with the IAP gene. There are, however, amino acid differences at their carboxyl-terminals. Placental ALP is a heatstable enzyme present at high levels in the placenta. A trace amount of this isoenzyme can be detected in normal sera. Part of the serum placental-type activity originates from neutrophils. Placental ALP activity has also been detected in normal type I pneumocytes

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

165

Alkaline Phosphatase

and in the human ovary and cervix. The placental ALP gene can be reexpressed by cancer cells as the Regan isoenzyme. Placental ALP is a very polymorphic enzyme, with up to 18 identified allelozymes resulting from point mutations, in contrast to the other ALP isoenzymes. The gene encoding germ cell ALP (GCAP) was also mapped to chromosome 2. It encodes testis/ thymus ALP and can be expressed in the placenta at low levels. GCAP in testis appears to be localized to the cell membrane of immature germ cells and, like the other ALP isoenzymes, is attached to the cell membrane by means of a glycan–phosphatidyl– inositol (GPI) anchor. Like the intestinal and placental ALP genes, it can be reexpressed by cancer cells.

LOCALIZATION OF ALP Localization of ALP in the cell has been performed by fractionating subcellular components by ultracentrifugation and then performing electron microscopic cytochemistry. From these studies, one can conclude that most of the ALP activity in the majority of cell types is located on the plasma membrane. Still, a small amount of enzyme activity has been shown to be present at other intracellular sites such as the nucleus, endoplasmic reticulum, and Golgi apparatus. Histochemical techniques at the level of the electron microscope showed reaction product for ALP on the microvillar membranes, multivesicular bodies, Golgi complex, lysosomes, endoplasmic reticulum, and nuclear membranes of the absorbing epithelial cells of the intestinal mucosa. ALP in the kidneys is localized on the microvilli of the proximal tubular cells. In liver, the reaction product of ALP is present on the microvilli of the bile canaliculi. In bone, ALP activity is present on the outer surface of plasma membranes of maturing and hypertropic chondrocytes and of preosteoblasts and osteoblasts. The reaction product is also present in chondrocyte lacunae, in matrix vesicles in cartilage matrix, as well as among uncalcified collagen fibrils of osteoid in bone. In placenta, ALP activity is found along the plasma membrane lining the microvilli and in vesicles of the syncytiotrophoblast. However, in eosinophil leukocytes, the ALP reaction product was found in only minimal amounts in the plasma membranes but was prominent on the nuclear membranes and outer compartment of mitochondria. The Golgi membranes and endoplasmic reticulum reacted but did so less intensely. The positive cytochemical reactions for ALP are confined to the cytoplasm of the more mature

neutrophilic granulocyte. ALP activity was also localized to fibroblasts, vascular smooth muscle, and endothelial cells in brain, heart, and endocrine glands.

STRUCTURE AND ANCHORING OF ALP TO THE CELL MEMBRANE Human ALP enzymes have not yet been obtained in crystalline forms suitable for X-ray analysis. However, their active sites, and the active site of Escherichia coli ALP, show a high degree of homology, so that key regions of the human isoenzymes can be interpreted with reference to the corresponding regions of the bacterial enzyme. E. coli ALP exists as a dimer of identical subunits, each of which contains 429 amino acids. Crystallographic observations of the molecule from E. coli have revealed the three-dimensional structure of dimeric ALP: a bat-like figure with a metal ion triplet in each active site region. Zinc can bind to all of these sites but binds particularly strongly to four sites per dimer, with magnesium occupying two sites of the dimer. ALP belongs to the large group of proteins attached to the outer surfaces of cells by a C-terminal GPI anchor. As such, it is an ectoenzyme expressed on the outside of the cell. The enzyme is a tetramer when it is membrane bound, but it circulates as a dimer. Phospholipase C or D, which is abundant in plasma, potentially converts the membrane-bound form to a soluble form.

IDENTIFICATION OF ALP ISOENZYMES Total serum ALP activity remains one of the most frequently measured enzyme activities in clinical medicine. Several different substrates have been introduced for its assay. Of these, paranitrophenylphosphate (PNPP) is probably the most widely used. Many methods have been proposed to separate the various ALP isoenzymes, including heat denaturation, chemical inhibition of selective activity, gel electrophoresis, precipitation by wheat germ lectin, and immunoassay. The heat denaturation method is based on the gradation in heat stability at 56 8C of the ALP enzymes found in serum. This heat stability ranges from placental ALP, which is completely heat stable; to liver ALP, which has intermediate stability; to bone ALP, which is very labile. The mean remaining enzyme activities after 15 min at 56 8C are 11, 21, 90, and 87% of the original activity for the bone,

166 liver, intestine, and placenta isoenzymes, respectively. Determining the tissue origin of the enzymes by employing differential heat inactivation demands precise control of the temperature during the assay. Selective chemical inhibitors have also been used to separate ALP isoenzymes. l-phenylalanine and l-tryptophane inhibit intestinal and placental ALP, whereas levamisole and l-homoarginine inhibit TNALP. Each of these inhibitors is stereospecific and noncompetitive. Another common method for distinguishing among ALP isoenzymes is polyacrylamide gel electrophoresis (PAGE). Liver ALP carries the highest net negative charge, followed by the placental, bone, and intestinal forms. Liver and bone ALP can be separated sufficiently to allow visual assessment of their relative proportions, but these methods are quite tedious and there is often overlap between the two isoenzymes, making precise quantification difficult. Wheat germ lectin binds to N-acetylglucosamine and sialic acid residues, and it provides a method by which to separate liver and bone ALP. Based on the differing glycosylation patterns of liver and bone, wheat germ lectin selectively binds the bone form. However, proper standards and lectin concentrations are necessary for accurate resolution. Attempts to produce tissue-specific monoclonal antibodies have resulted in antibodies with preferential recognition of the liver, intestinal, and placental ALP. A two-site immunoradiometric assay that relies on the use of two monoclonal antibodies has been developed for the bone isoform.

FUNCTION OF ALP Biochemical Function Three in vitro functions have been attributed to ALP: (1) phosphohydrolysis of organic phosphomonoesters of low molecular mass, (2) phosphotransferase activity, and (3) protein phosphatase activity. Whether any of these relate to the physiological role of the enzyme is as yet unknown. ALP has little preference for a particular substrate and will hydrolyze all phosphomonoesters (diesters are not substrates). Catalysis includes phosphorylation of a serine residue at the active site, followed by transfer of the phosphoryl group to either water (phosphohydrolysis) or an organic acceptor alcohol (phosphotransferase). However, phosphoester cleavage is faster if the transfer of phosphate is to an acceptor rather than to water, and the hydrophobic nature of plasma membranebound ALP might lead to a preference for organic

Alkaline Phosphatase

acceptors and, thus, principally to a transphosphorylation reaction.

Physiological Role of ALP ALPs are widely distributed in nature; they are present in all species, from bacteria to humans. This is an indication that the enzymes are involved in fundamental biochemical processes. Embryonal Development Studies in amphibian embryos led to predictions concerning the expected distribution, on the embryonic flank mesoderm, of a cell surface molecule involved in guidance information for the embryonic cell migration of pronephric duct cells. An investigation of effects of disruption of the embryonic ALP gene on mouse preimplantation development revealed that the absence of the embryonic ALP gene resulted in fewer blastocysts in vitro, delayed parturition, and reduced litter size in vivo. This observation indicated that the presence of an active ALP gene is beneficial for preimplantation development. Regulation of Lipid Transport and Intestinal and Renal Phosphate Transport The local abundance of ALP on the membranes of cells involved in the transport of many substances (e.g., duodenal cell and small intestinal enterocytes, renal tubular cell, type I pneumocyte in the lung) suggests a role for the enzyme in complex active transport. The evidence has been accumulating that IAP may play a role in lipid transport. Elevation of serum IAP activity has been demonstrated in rats following feeding of a high-fat diet. The magnitude of this response is dependent on fatty acid chain length. The correlation between lipid concentration and IAP activity in human lymph has led to the speculation that IAP might be involved in lipid transport. The views concerning the role of ALP in renal inorganic phosphate (Pi) transport differ considerably. The administration of ALP inhibitors decreases Pi reabsorption in the kidney in vivo. Studies in vitro, however, showed that compounds that inhibit ALP activity did not block the uptake of Pi, suggesting that ALP is not a Pi-transporting enzyme analogous to ATPase in Na+ transport but rather that the role of ALP in Pi transport is indirect. In the kidney, ALP activity is highest in the early proximal tubule, which coincides with the region of the proximal luminal brush border membrane transport of Pi. Positive

167

Alkaline Phosphatase

correlations are found between the brush border membrane transport of Pi and the activity of ALP but not of other brush border membrane enzymes. The function of ALP in the intestinal epithelial transport of Pi (and calcium) is also a controversial subject in that inhibitors of ALP are unable to inhibit Pi transport by the intestinal cell. Bone Formation Another important role that has been assigned to ALP is its role in skeletal mineralization. Robison was the first to recognize the role of ALP in skeletal mineralization. The strongest evidence that TN-ALP functions in skeletal formation come with the delineation of inherited hypophosphatasia. This disorder is characterized by defective bone mineralization. The severity of hypophosphatasia is highly variable, ranging from stillbirth with almost no mineralized bone to pathological fractures first presenting during adulthood. The clinical heterogeneity of hypophosphatasia probably reflects the numerous mutations that have been described in the TN-ALP gene and that may give rise to various degrees of clinical severity. The following roles have been postulated for bone ALP in the mineralization process: (1) hydrolysis of organic phosphate esters, resulting in high local Pi concentration and facilitating precipitation of calcium phosphate; (2) destruction of physiological crystal growth inhibitors such as inorganic pyrophosphate and adenosine triphosphate (ATP) through its hydrolase activity; (3) action as a Pi transporter; and (4) active transport of Ca2+ or Pi via its ATPase activity. In bone matrix vesicles (functionally active shedded plasma membrane fragments of osteoblasts) and the initial site of hydroxyapatite crystal formation, bone ALP activity can be as much as 20 times more than that on the plasma membrane surface of intact osteoblasts. ALP may be involved here in initiation of the calcification process by raising the local concentration of phosphate ions. The extracellular matrix-binding domain of TN-ALP may also be important in directing the migration of matrix vesicles along collagen fibers during the process of bone mineralization. ALP in Liver Although liver ALP generally represents at least half of total serum ALP activity in healthy adults, there is little evidence to support a critical hepatic function for liver ALP. It is generally assumed from its location on the sinusoidal membrane that liver ALP acts as a transport protein. Nevertheless it is doubtful that ALP fulfills a key function in liver given that gross deficiency of the liver isoenzyme in congenital

hypophosphatasia does not appear to give rise to any obvious clinical manifestation. However, one report has suggested that liver ALP may protect liver function from immunological injury by a mechanism involving neutralization of endotoxin. ALP in Microvessels ALP is one of the main enzymes present in brain and heart microvessels. ALP may have a role in the metabolism of pyridoxal 50 -phosphate, a cofactor of enzymes such as glutamate decarboxylase and glutamate transaminase involved in the metabolism of neural tissue, and might be a key enzyme of the blood–brain barrier regulated by insulin. Thus, insulin has been found to significantly inhibit brain ALP activity. ALP activity associated with capillary endothelial cells is also clearly affected by a hypoxic environment. ALP activity was significantly reduced or absent in areas of hypoxic skeletal and cardiac muscle, and a similar reduction was found in brain emboli associated with cardiopulmonary bypass. Focal loss of ALP activity demonstrated by histochemical methods appears to be a useful probe in identifying ischemic or hypoxic loci in these tissues. It has also been proposed that the expression of ALP in endothelia in brain and heart microvessels may contribute to the vascular hardening and calcification observed in humans. This, in turn, could be related to vascular aging, vascular disease, and the resultant weakening and/or rupture of vessel walls.

PATHOLOGICAL SIGNIFICANCE OF ALP The major factors that affect ALP activity are age, sex, and hormonal status (puberty or menopause). From birth to 6 weeks, both bone ALP and intestinal ALP increase. No liver ALP is observed until 6 months of age. In children, a wide range of ALP activity exists and correlates with height and weight, and until puberty the bone isoenzyme represents 77 to 87% of the total. Activity increases in children around the age of puberty, with the maximum being earlier in girls than in boys, and corresponds temporally with growth spurts in both sexes. Bone ALP has been reported to increase during pregnancy; however, a gradual increase in total ALP activity is observed during the first 6 months of pregnancy, followed by a rapid increase during the final trimester. This increase is due primarily to the placental enzyme. In healthy adults, the ratio of bone activity to liver activity is approximately 1:1. After 50 years of age,

168

Alkaline Phosphatase

total ALP again increases. Bone ALP activity is generally found to be higher in postmenopausal women than in premenopausal women. There is a great deal of interindividual variation in adult ALP levels, but for any one individual, values change little with time. ALP is cleared from the blood very slowly; the halflife varies from 40 h for bone to 7 days for placental isoforms. Biological daily variation of total ALP is estimated to be less than 4%.

mineralization of the osteoid matrix synthesized by the osteoblasts, but because the growth plate is closed, disorders of epiphyseal cartilage do not occur and so growth abnormalities are not seen. The result is an accumulation of unmineralized bone, which in turn results in decreased bone density, as shown by X-ray or other techniques. As in children, disorders of vitamin D metabolism in adults are the most common causes of osteomalacia, and bone ALP is significantly increased in these cases.

Bone Disease

Paget’s Disease of Bone Paget’s disease of bone is a common disorder in the elderly where excessive bone turnover occurs, leading to the production of structurally abnormal bone. It is characterized during its initial phase by an increased resorption of bone, followed by an intense osteoblastic response. ALP is a very sensitive biochemical sign manifestation of Paget’s disease of bone, as are indexes of bone resorption. ALP activity in Paget’s disease is higher than in any other bone disease, excluding primary tumors of bone (osteosarcomas) and osteoblastic metastases from extraskeletal tumors (notably prostate cancer). Bone ALP may eventually decrease, possibly as the disease moves into a sclerotic phase. Serum ALP and indexes of bone resorption correlate well with the extent of skeletal involvement and with the response to treatment.

Rickets and Osteomalacia Rickets in growing children is characterized by alterations in chondrocyte differentiation and reduced matrix mineralization in the cartilaginous growth plate, by increased osteoblastic activity, and by defective mineralization of bone matrix. The most common cause is vitamin D deficiency. Primary disorders of phosphate homeostasis and renal tubular disorders also can cause rickets in children. A number of conditions result in vitamin D-deficient rickets, including inadequate exposure of the skin to ultraviolet radiation with inadequate dietary intake or intestinal malabsorption of this vitamin, chronic impaired renal function with insufficient production of 1,25-dihydroxy vitamin D (the metabolically active form of this vitamin), vitamin D receptor abnormalities, and an inherited deficiency of the enzyme 1a–hydroxylase that produces 1,25-dihydroxy vitamin D. Bone ALP in these children is moderately to greatly elevated. Bone ALP activity can be further increased shortly after starting vitamin D treatment when healing begins, and it can decline progressively when therapy is effective. X-linked hypophosphatemia due to a mutation in the enzyme PHEX and autosomal dominant hypophosphatemia due to a mutation in FGF23 are characterized by progressively severe skeletal deformities and dwarfism. Bone ALP activity is also high in these children. Rickets can occur in a variety of disorders with impaired proximal tubular function that produce increased renal clearance of inorganic phosphate and hypophosphatemia. Glomerular filtration can be entirely normal or near normal. Because the serum concentrations of inorganic phosphate and calcium are critical for the formation of hydroxyapatite crystals and mineralization of bone, hypophosphatemia results in defective mineralization. Osteomalacia is the adult equivalent of rickets in children. It is characterized by a defect in the

Congenital Hypophosphatasia Hypophosphatasia is an inherited disorder characterized by defective bone mineralization and a deficiency of TN-ALP activity. Other ALP isoenzyme activity is unaffected. The severe homozygous form is lethal, whereas less severely affected children show generalized bone deformities due to defective mineralization of osteoid. The milder adult form is characterized by pathological fractures and precocious loss of teeth. Three phosphocompounds accumulate endogenously in hypophosphatasia—phosphoethanolamine (PEA), inorganic pyrophosphate (PPi), and pyridoxal 50 phosphate (PLP)—indicating that these are natural substrates for TN-ALP.

Hepatobiliary Diseases Hepatic diseases, such as acute and chronic hepatitis, cirrhosis, carcinoma of the liver, metastatic carcinoma of the liver, and acute and chronic biliary obstruction, are associated with increases in liver ALP activity. Since the first report of elevated ALP activity in human serum in connection with obstructive jaundice,

169

Alkaline Phosphatase

much of the current knowledge concerning ALP in cholestasis has been obtained from experiments with bile duct-ligated rats. Studies in rats have shown that bile duct ligation or exposure of cultured explants of rat liver to bile from bile duct-ligated rats causes a marked increase in hepatic ALP synthesis in the liver, followed by an increase in serum ALP activity in the bile duct-ligated rats. Several bile acids have the capacity to stimulate ALP activity in a dose-dependent fashion in rat liver cell cultures. In patients with liver cirrhosis, the catabolism of ALP may be reduced. Thoracic lymph flow is increased several-fold in cirrhotic patients, and the intestinal isoenzyme enriches the thoracic lymph and the serum, especially after a fatty meal. It is conceivable that these two factors (decreased clearance by the cirrhotic liver and enrichment of thoracic lymph with intestinal isoenzyme) are responsible for the frequent occurrence of the intestinal isoenzyme in the serum of cirrhotic patients.

Tumor Forms of ALP Since the first description of ectopic production of ALP by tumor tissue, three different isoenzymes from independent gene loci have been detected in cancer patients: (1) Regan isoenzyme or term placental ALP; (2) Nagao, testicular, or placental ALP-like; and (3) Kasahara or fetal intestinal ALP. Total serum ALP has been shown to be an effective indicator for metastatic breast cancer, with a sensitivity for metastases of 30 to 40%. Total ALP has also been suggested to be a significant prognostic factor for prostate cancer. Using various statistical methods, sensitivity, specificity, and accuracy for liver metastases in lung cancer were 71, 89, and 86%, respectively. Serum bone ALP was twice as sensitive as total enzyme activity in the diagnosis of the presence of bone metastases. Serum placental ALP has been

described as a potential marker in seminoma patients, but environmental influences have been shown to affect placental ALP levels significantly (e.g., smoking leads to a reexpression of placental ALP). Despite these environmental influences, placental ALP has been used successfully for monitoring testicular and ovarian carcinomas. Therefore, ALP forms appear to play an important role in the diagnosis of a variety of benign and malignant clinical disorders, especially of bone and liver, and can be a useful index for monitoring therapy in these diseases.

Further Reading Bai, X. Y., Miao, D., Goltzman, D., and Karaplis, A. C. (2003). The autosomal dominant hypophosphatemic rickets R176Q mutation in fibroblast growth factor 23 resists proteolytic cleavage and enhances in vivo biological potency. J. Biol. Chem. 278, 9843–9849. Dehghani, H., Narisawa, S., Millan, J. L., and Hahnel, A. C. (2000). Effects of disruption of the embryonic alkaline phosphatase gene on preimplantation development of the mouse. Dev. Dynamics 217, 440–448. Hui, M., and Tenenbaum, H. C. (1998). New face of an old enzyme: Alkaline phosphatase may contribute to human tissue aging by inducing tissue hardening and calcification. Anatomical Rec. 253, 91–94. Miao, D., and Scutt, A. (2002). Histochemical localization of alkaline phosphatase activity in decalcified bone and cartilage. J. Histochem. Cytochem. 50, 333–340. Narisawa, S., Wennberg, C., and Millan, J. L. (2001). Abnormal vitamin B6 metabolism in alkaline phosphatase knock-out mice causes multiple abnormalities, but not the impaired bone mineralization. J. Pathol. 193, 125–133. Van Hoof, V. O., and De Broe, M. E. (1994). Interpretation and clinical significance of alkaline phosphatase isoenzyme patterns. Crit. Rev. Clin. Lab. Sci. 31, 197–293. Whyte, M. P. (1994). Hypophosphatasia and the role of alkaline phosphatase in skeletal mineralization. Endocrinol. Rev. 15, 439–461. Xu, Q., Lu, Z., and Zhang, X. (2002). A novel role of alkaline phosphatase in protection from immunological liver injury in mice. Liver 22, 8–14.

Alternative Promoters Erik Jansen, Torik A. Y. Ayoubi, Wim J. M. Van de Ven, and John W. M. Creemers University of Leuven and Flanders Interuniversity Institute for Biotechnology, Belgium, The Netherlands

Glossary promoter DNA structures containing cis-acting regulatory elements required for efficient initiation of transcription and for controlling expression of a gene. transcription units Templates for RNA polymerases that encode information for the production of a single protein product (simple transcription units) or templates that produce multiple mature mRNAs that can give rise to multiple, albeit related, proteins (complex transcription units).

Alternative promoters can be defined as two or more promoters used to generate the same primary transcript or at least partially overlapping primary transcripts. Alternative promoters will always give rise to differential transcription initiation, resulting in differences in the 50 region of the mRNA isoforms.

INTRODUCTION Probably the biggest surprise coming from the human genome sequence is the much lower than anticipated gene number. Therefore, it might be expected that complexity of the human proteome is achieved by transcriptional, translational, and posttranslational diversity. The use of alternative promoters is a frequently used way in which to generate multiple protein isoforms from a single gene. In addition, alternative promoters play an essential role in the control of transcription in a spatial and temporal fashion necessary for the proper development and differentiation of specialized cell types that define multicellular organisms.

THE NEED FOR ALTERNATIVE PROMOTERS Expression of genes in more than one tissue or developmental stage often requires distinct combinations

170

of transcription factors. One single promoter might not always be sufficient to accommodate all of the required responses. Another role for alternative promoters is to be able to respond in different cell types to the same extracellular signals or to respond in the same cell to different signals. Finally, alternative transcripts can generate diversity by influencing mRNA stability, translation efficiency, and amino terminus of the encoded protein. A different amino terminus, in turn, can lead to alterations in protein levels, functions, or subcellular localization.

STRUCTURAL ORGANIZATION OF ALTERNATIVE PROMOTERS Although there are various patterns of alternative promoter use, the two basic mechanisms are shown schematically in Fig. 1. In the first case (panel A), two tandemly arranged promoters are positioned within the same exon. In the second case (panel B), alternative promoter use will result in alternative first exons. If the start codon is located within the second exon (indicated with a black diamond), there will be no difference between the transcripts at the protein level. However, if start codons are located in the first exons (indicated with arrowheads), different protein isoforms will be generated. Another

A

B

Figure 1 Schematic representation of the organization of genes containing alternative promoters. Exons are depicted as boxes, and intervening sequences are depicted by solid lines. Dotted lines connecting exons indicate splicing patterns. Arrows indicate transcription initiation sites. Diamonds and arrowheads represent start codons, as discussed in the text.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

171

Alternative Promoters

layer of complexity can be added by additional promoters and/or exons upstream of the second promoter.

MODES OF REGULATION OF ALTERNATIVE PROMOTERS Promoters respond to specific stimuli. In most cases, this is a direct result of the presence or absence of specific response elements in a given promoter. Alternative promoters provide the possibility of responding differentially to ubiquitous transcription activators and repressors, tissue- and developmental stage-specific regulators, and maternal or paternal allele-specific (imprinted) regulators. In combination with the generation of different protein isoforms, this will result in the differential expression and presence of, for instance, hormone receptor XA in tissue A and its variant XB in tissue B. This variant could display altered ligand specificity and/or altered intracellular signaling. In addition, feedback mechanisms exist in which the gene product (in)directly up- or down-regulates one or several of its alternative promoters.

s0025

SELECTED EXAMPLES The genes encoding transcription factors CREB (cyclic AMP response element-binding protein) and CREM (cyclic AMP response element-modulatory protein) are an interesting example of regulatory mechanisms. The peculiar aspect resides in the fact that they can encode different isoforms, either activating or inhibiting gene expression, by mechanisms of alternative exon splicing, alternative promoter use, and auto-regulation of promoters. In particular, an internal promoter of the CREM gene directs the expression of a repressor isoform that auto-regulates the alternative promoter, thereby generating a negative feedback loop. The specific regulation of expression of the cytochrome P450 aromatase gene is a good example of alternative promoter use in the context of endocrine disorders. P450 aromatase is expressed in several normal human tissues such as ovary, placenta, testis, brain, adipose, bone, and skin; in some pathological tissues such as breast and endometrial tumors; and tissues from endometriosis and myofibroma of the uterus. The regulation of expression of P450 aromatase is quite different in these tissues; folliclestimulating hormone (FSH) stimulates the expression of P450 aromatase in ovary, phorbol esters and

ligands of retinoic acid receptors stimulate the expression in choriocarcinoma cells, and androgens stimulate the expression in the hypothalamus. It has been shown that the tissue-specific and hormonal milieu-specific expression is achieved mainly by tissue-specific alternative use of different promoters. At least nine alternative promoters and 50 -untranslated first exons have been identified. FSH increases the expression of aromatase in ovary by the use and activation of the proximal promoter II. In placenta, phorbol esters regulate the expression of the most upstream promoter I.1. And in adipose tissue, tumor necrosis factor-a stimulates the use of promoter I.4 through activation of transcription factor AP-1. The potential role of P450 aromatase promoter switching in various physiological and pathological processes is illustrated by observations that distinct transcripts are expressed in fetal versus adult human liver and in healthy versus cancerous breast adipose tissue. A summit of complexity is represented by the GNAS1 gene encoding a stimulatory G protein that is essential for activation of intracellular signaling in response to specific hormone–receptor interactions. This gene has at least four alternative promoters. In addition, it is an imprinted gene that produces different gene products from the maternal and paternal allele through the use of oppositely imprinted alternative promoters. Moreover, abnormal imprinting of the GNAS1 gene promoters can lead to disease states such as pseudo-hypoparathyroidism and somatotrophs adenomas.

CONCLUSION Heterogeneity in the 50 ends of mRNAs generated by alternative promoter use in a tissue- or developmental stage-specific manner is common to a large group of genes. It is obvious that the increasing complexity of the mRNA population is an appropriate means of achieving a differential and spatiotemporal expression of the corresponding gene products and their potential pleiotropic actions. There is increasing evidence for several genes that switching of gene expression from one mRNA variant to another may be a key regulatory mechanism in several physiological and pathological processes.

See Also the Following Articles Alternative Splicing Gene Expression

.

Peptide Hormones, Regulation and

172

Further Reading Ayoubi, T. A. Y., and Van de Ven, W. J. M. (1996). Regulation of gene expression by alternative promoters. FASEB J. 10, 453–460. Bulun, S. E., Noble, L. S., Takayama, K., Michael, M. D., Agarwal, V., Fisher, C., Zhao, Y., Hinshelwood, M. M., Ito, Y., and Simpson, E. R. (2001). Endocrine disorders associated with

Alternative Promoters

inappropriately high aromatase expression. J. Steroid Biochem. Molec. Biol. 61, 133–139. De Cesare, D., and Sassone-Corsi, P. (2000). Transcriptional regulation by cyclic AMP-responsive factors. Prog. Nucleic Acid Res. Mol. Biol. 64, 343–369. Lalande, M. (2001). Imprints of disease at GNAS1. J. Clin. Invest. 107, 793–794.

Alternative Splicing Scott A. Akker, Bernard C. E. Khoo, and Shern L. Chew St. Bartholomew’s Hospital, London, United Kingdom

Glossary exon A nucleotide sequence, separated by introns, that is included in the final mRNA. intron A transcribed nucleotide sequence that is excised in the production of mRNA. splicing The process by which introns are excised and exons united to form mRNA.

A

lternative splicing is the process by which different mRNAs are generated from the same pre-mRNA by the selection of alternative exons.

THE SPLIT GENE, ALTERNATIVE SPLICING, AND PROTEIN DIVERSITY

p0020

“Split” genes were first recognized in 1977. It is now estimated that 95% of human genes are split, with coding exons split by noncoding introns. Splicing is the process by which the introns are excised and the exons ligated to form a translatable message. Thus, DNA is transcribed to form premessenger RNA (pre-mRNA), spliced to form messenger RNA (mRNA), and then exported from the nucleus for translation to protein. Alternative splicing is now recognized as the major contributor to protein diversity in man. Analysis of the human genome suggests that only 1.1% of its 2.91 billion base pairs code for exons. These exons are distributed over an estimated 32,000 genes, and alternative splicing is important in generating diversity and complexity from this limited gene pool. Current estimates suggest that at least 60% of the genes can produce two or more mRNAs by alternative splicing. The potential for generating diversity is demonstrated by the Slo gene, which has the potential to generate more than 500 protein isoforms. The regulation of alternative splicing provides an essential control step in post-transcriptional gene

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

expression, and the endocrine system is likely to be one of the key systems that directs this in a tissueand cell-specific manner. Next, we describe the current understanding of the role, mechanisms, and control of alternative splicing, with particular reference to endocrinology.

THE SPLICING MECHANISM Splicing is performed by a large complex of proteins and small nuclear ribonucleoproteins (snRNPs) called the spliceosome. The spliceosome must be able to recognize the exon/intron boundaries in a precise and reproducible manner. At least part of the exon/intron definition is provided by certain key intronic elements, including a 30 and 50 splice site and a branch site and polypyrimidine tract upstream of each spliced exon (Fig. 1). Splicing occurs in close physical and temporal relationship to pre-mRNA transcription, with the Cterminal domain of RNA polymerase II helping direct splicing factors to the pre-mRNA. The snRNPs are key players in the spliceosome, and each one has a RNA component that is capable of interacting with the pre-mRNA through Watson–Crick base pairing. U1 snRNP binds the 50 splice site of the exon. U2 snRNP then associates with the polypyrimidine tract of the upstream 30 splice site and the U4/U6.U5 trisnRNP is recruited. The spliceosome then catalyzes cleavage of the pre-mRNA at the 50 splice site, and a lariat structure with the branch site is formed (Fig. 1). Cleavage at the 30 splice site precedes exon ligation and lariat release. Splicing is followed by the addition of a poly(A) tail, which is also subject to regulation.

PATTERNS OF ALTERNATIVE SPLICING The most primitive form of alternative splicing is intron retention; in man, however, the most commonly described form of alternative splicing is exon skipping. Several other mechanisms exist, including the use of alternative 30 and/or alternative 50 splice

173

174

Alternative Splicing

5' Splice site

Branch site

gu

a

3' Splice site ag

Exon skipping/inclusion Alternative 3⬘ splice site

U1 E

Intron retention

gu a

ag Alternative 5⬘ splice site

A

U1 ug U2 a

ag U1

B

f0005

U4/U6 ug U2 U5 a ag

Figure 1 Basic mechanisms of splice site selection and spliceosome assembly. The exons are shown as boxes and the intron as a line. The bold line in the intron, between the branch site and the 30 splice site, represents the polypyrimidine tract. The stepwise assembly of the spliceosome complexes, early (E) and subsequent (A and B), is shown. Note that only the U1, U2, and U4/U6.U5 snRNPs are depicted, and the many non-snRNP proteins are omitted for clarity. From Akker et al. (2001), reproduced by permission of the Society for Endocrinology. sites (Fig. 2). Many genes in the endocrine system undergo a combination of different alternative splicing events, employing various mechanisms.

REGULATION OF ALTERNATIVE SPLICING Splice Sites The splice sites that flank exons are necessary but not sufficient to explain exon recognition. Splice sites have a varying degree of similarity to highly degenerate consensus sequences. In fact, splice site-like sequences that also match the consensus occur with great frequency throughout the genome and define a set of pseudosites and pseudoexons that never undergo splicing. Such pseudoexons may outnumber true exons 10 to 1. Thus, there is a spectrum of exons: those that are always spliced (constitutive exons), those that are spliced only under certain conditions (alternatively spliced exons), and those that never undergo splicing (pseudoexons). As a rule, alternatively spliced exons tend to have splice sites with a weaker match to the consensus than constitutive exons, and this allows them to be more amenable to the effects of other sequence elements and factors. Experimentally

Mutually exclusive exons

Figure 2 Common patterns of alternative splicing. The central exons undergo alternative splicing, with one pathway represented by solid lines and the alternative with dashed lines. As seen in IGF-1, the same exon within a gene may undergo more than one form of alternative splicing. enhancing the splice site strength of alternatively spliced exons usually causes them to be included constitutively. The splice site strength of the exons neighboring an alternatively spliced exon is also important. Strengthening the 50 splice site of the upstream exon or the 30 splice site of the downstream exon can lead to an increased degree of skipping of the middle exon.

Other RNA Sequence Elements and Their Interactions Major advances in understanding regulated alternative splicing have occurred with the identification of primary sequence elements involved in promoting exon selection (enhancers) or repressing splicing (silencers). The factors functioning through several such regulatory sequences have been isolated. Many important non-snRNP proteins are involved in spliceosome assembly and function. A number of these also regulate splice site selection. In addition to the elements described previously, examples of elements that may function through secondary structure have been described, such as in the insulin receptor described later. Alternative splicing could be subject to regulation via RNA helicases, which function to unwind the secondary structure to allow access of the spliceosome to splice sites and regulatory elements. Splicing Enhancers and Serine–Arginine-Rich Proteins There are many examples of splicing enhancer elements. These are most commonly sited within exons but may also be intronic. A family of splicing factors has been shown to interact with these elements—the serine–arginine-rich (SR) proteins. SR proteins are characterized by RNA-recognition motifs and domains

175

Alternative Splicing

containing serine and arginine repeats, which are important for protein–protein interactions. SR proteins are required for constitutive splicing and are also crucial mediators of regulated alternative splicing. SR proteins may regulate alternative splicing through other means. The SR protein SF2/ASF has been proposed to increase the stability of certain mRNAs by direct binding. Some proteins act by modulating SR protein function; for example, SRrp86 directly interacts with two other SR proteins to enhance (SRp20) and repress (SC35) their actions in a dose-dependent manner. s0040

p0055

p0065

Splicing Silencers and Heterogeneous Nuclear Ribonucleoproteins Less well documented are examples of splicing silencers. These have also been described in both exonic and intronic contexts. It has been proposed that silencers occur frequently throughout human introns and act to repress pseudo-splice sites. The major family of proteins shown to interact with these elements is the heterogeneous nuclear ribonucleoproteins (hnRNP) family. Only a few have been characterized in terms of their function. Perhaps the best characterized hnRNP is polypyrimidine tract-binding protein (PTB; also known as hnRNP I). PTB exists as several isoforms through alternative splicing and has been associated with several types of exon skipping. PTB binding sites are often found on both upstream and downstream introns, and it has been proposed that an interaction occurs between PTBs, which can form multimers, and together they promote exon skipping. In the case of hnRNP A1, cooperative binding between several molecules occurs, starting from a highaffinity binding site. This results in the “spreading” of hnRNP binding along the pre-mRNA. The overall picture that is emerging is that each alternatively spliced exon lies in a balance of recognition versus nonrecognition. The spliceosome makes its decision based on the splicing signal strength and the degree of signal enhancement or repression present. A major model of alternative splicing holds that the binding of SF2/ASF to a pre-mRNA is in competition with that of hnRNP A1. The binding of SF2/ ASF recruits spliceosomal factors via protein–protein interactions. SF2/ASF may also function to block the propagation of hnRNP A1 binding along the premRNA. Any change in the ratio of enhancing factors to silencing factors can swing the balance toward or away from splicing. The following mechanisms may regulate enhancing/silencing factors: phosphorylation, demonstrated by STREX and insulin action on SRp40; cellular

localization/sequestration (e.g., the stress-induced cytoplasmic sequestration of hnRNP A1); and temporal regulation of factor concentrations in tissues (e.g., insulin up-regulates the expression of SRp40). A combination of all three of these mechanisms is likely to occur in vivo: for example, cytoplasmic sequestration of hnRNP A1 is associated with phosphorylation mediated by the MKK3/6–p38 kinase pathway.

ENDOCRINE EXAMPLES OF ALTERNATIVE SPLICING Table I provides examples of endocrine stimuli that affect alternative splicing. Here, we focus on endocrine examples that demonstrate some principles of alternative splicing regulation.

Table I Examples of Hormonally Regulated Alternative Splicing Eventsa Alternatively spliced mRNA

Stimulus

Insulin receptor

Dexamethasone, glucose, insulin

Calcitonin/CGRP

Dexamethasone

Protein kinase C b

Insulin

IGF-1

Growth hormone

Fibroblast growth factor receptor

Cytokines

Phosphotyrosine-1B

PDGF, EGF, bFGF

hPMCA2

Calcium

CD44

PDGF, IGF-1, via hnRNP A1

Fibronectin EIIIB (rat)

Insulin, via HRS

Fibronectin ED (human)

TGF-b1, vitamin D, retinoic acid

Kv3.1 channel Agrin

bFGF/depolarization Nerve growth factor

MHC-B

Nerve growth factor

SRp20

Serum/cell cycle

Slo BK channel

Hypophysectomy/ACTH

Thyroid hormone receptor-b

Tri-iodothyronine

Tau

Tri-iodothyronine

Activating transcription factor-3

TNF-a

SERCA3

Retinoic acid, “hypertension”

a

Abbreviations used: bFGF, basic fibroblast growth factor; EGF, epidermal growth factor; hPMCA2, human plasma membrane Ca-ATPase; IGF-1, insulin-like growth factor-1; Kv3.1, potassium voltage-gated channel; MHC-B, myosin heavy chain II-B; PDGF, platelet-derived growth factor; SERCA3, sarco/endoplasmic reticulum Ca-ATPase 3; TGF-b1, transforming growth factor-b1; TNF-a, tumor necrosis factor-a.

176

Alternative Splicing

The Calcitonin/a-Calcitonin Gene-Related Peptide Gene: Tissue-Specific Alternative Splicing Altered by a Hormonal Stimulus

p0080

The first described example of a hormonal stimulus altering splicing was in the calcitonin/a-calcitonin gene-related peptide (a-CGRP) gene in 1986. The gene contains six exons and two polyadenylation signals. Alternative splicing leads to two proteins, distinct in both structure and function. In thyroid C cells, exon 4 is recognized and spliced with use of the exon 4 polyadenylation signal to generate calcitonin. In neuronal cells, exon 4 is skipped, and exons 5 and 6 are spliced with use of the exon 6 polyadenylation signal to generate a-CGRP (Fig. 3). Exon 4 is the key because its inclusion leads to cleavage and polyadenylation at its poly(A) site, thus preventing recognition of the downstream exons. The “constitutive” pathway is that of exon 4 inclusion, and it is only skipped in neuronal and cardiac tissue. The important elements in and around exon 4 have been characterized. It is not surprising that exon 4 undergoes skipping because it contains weak splicing signals (non-canonical branch site and polypyrimidine tract) as well as a weak polyadenylation signal. Mutation of these signals to canonical sequences leads to exon 4 inclusion in all circumstances. Given the weak signals, it is perhaps more surprising that exon 4 is generally included. However, no fewer than three exonic splicing enhancers, one intronic splicing enhancer, and five pentanucleotide repeats, which also act as enhancers, have been described. These appear to compensate for the weak splicing signals. The intronic splicing enhancer is particularly interesting because it contains consensus sequences for both 30 and 50 splice sites that interact with U1 snRNP, PTB, and SF2/ASF. The enhancer and the factors binding it regulate both polyadenylation and terminal exon splicing. CGRP mRNA

1

Calcitonin/CGRP pre-mRNA

1

2

3

2

5

3

6 AAAAAA

A 4

In a medullary carcinoma-derived cell line, dexamethasone provides a stimulus that alters alternative splicing, up-regulating calcitonin production over aCGRP production. The intermediaries between dexamethasone and the elements and factors described previously are not known.

Growth Hormone/Insulin-like Growth Factor-1 Alternative Splicing: Generation of Complexity This system provides insight into how levels of complexity in gene expression can be established because growth hormone (GH), insulin-like growth factor-1 (IGF-1), their binding proteins, and their receptors all undergo alternative splicing. GH is coded for by the GH-1 gene, a five-exon gene that produces two major transcripts and at least three minor transcripts. The predominant isoform of GH is the 22-kDa form resulting from inclusion of all exons. Approximately 5–10% of human GH in the pituitary and in lymphocytes is the 20-kDa form, resulting from the use of an alternative splice site within exon 3. Use of this site is dependent on secondary structure elements and leads to a 15-amino acid deletion from the final protein. The role of this isoform is not known; somatotroph adenomas produce proportionally more 20-kDa isoform than 22kDa isoform, and GH is carried by two GH-binding proteins that have different affinities for the 20- and 22-kDa isoforms. A small proportion of GH premRNAs undergo exon 3 or exons 3 and 4 skipping, and these isoforms are believed to have a dominant negative effect, inhibiting the action of normal GH. Exon 3 inclusion has been shown to be dependent on an upstream intronic splicing enhancer. Mutations within this intronic splicing enhancer lead to exon skipping and cause isolated GH deficiency. IGF-1 is derived from a six-exon pre-mRNA. Four of the six exons undergo alternative splicing and lead to the production of different precursor peptides but an identical mature peptide (Fig. 4). The exact roles of these different precursors are not well understood.

A 5

6 A 1

Calcitonin mRNA

1

2

3

2

3

4

5

A 6

4 AAAAAA

Figure 3 Alternative splicing of the calcitonin/CGRP gene. The two potential mRNAs are schematically shown: CGRP mRNA (top) and calcitonin (bottom). A, the polyadenylation signals on the pre-mRNA.

Signal peptides

Mature peptide

E domains

Figure 4 Alternative splicing of the IGF-1 gene. Alternative promoter sites are represented by the arrows of exon 1 and exon 2. Alternative splicing pathways are represented by the sloped lines.

177

Alternative Splicing

Exon 5 splicing is dependent on tissue type and the hormonal milieu, and an exonic splicing enhancer located within exon 5 is responsive to the SR protein SF2/ASF. Increasing the activity/amount of SF2/ASF enhances exon 5 inclusion. The presence of exon 5 is linked to nuclear and nucleolar localization. The role of the precursors at different sites remains to be determined.

Insulin Receptor Alternative Splicing: A Role in Insulin Resistance? The insulin receptor (IR) gene is a 22-exon gene. As in the calcitonin/a-CGRP gene, there are two alternatively spliced isoforms that are generated by the inclusion or skipping of a single exon—exon 11. Exon 11 is only 36 base pairs and codes for 12 amino acids that lie in the C terminus of the a-subunit. The two receptor isoforms have been shown to differ in their tissue distribution, dimerization properties, substrate binding, and function (Table II). In a human hepatoblastoma cell line, dexamethasone increases insulin sensitivity, and this is associated with up-regulation of the IR-B isoform. Glucose also increases IR-B splicing, and insulin levels have been linked with different IR isoform expression, independent of glucose. Closer analysis has revealed some of the elements involved in alternative splicing of exon 11 and that ultimately mediate hormonal regulation. The weak splice sites of exons 10–12 predispose to regulation. These splice sites appear to compete for limiting splicing factors. Changing the splice sites effects the level of splicing of exon 11. Furthermore, four regulatory elements have been identified. Two elements lie in intron 10; a GA-rich region acts as an intronic

splicing enhancer that may enhance the neighboring weak splice sites. Downstream of this element lies an intronic splicing silencer, which is predicted to form a stem-loop structure with exon 11. The final two elements lie in close proximity to exon 11—one acts as an enhancer and the other as a silencer. It is generally accepted that hyperinsulinemia is associated with relatively increased IR-A expression in muscle and that this precedes hyperglycemia. This contributes to insulin resistance, but it is not established whether hyperinsulinemia precedes the change in IR alternative splicing or vice versa. There is an additional level of complexity because both IR isoforms form heterodimers with the homologous IGF-1 receptor. Heterodimers of the IGF-1 receptor with the IR-A isoform bind IGF-1, IGF-2, and insulin, and signaling occurs along the IGF pathway. In contrast, heterodimers of the IGF-1 receptor with the IR-B isoform bind only IGF-1. Therefore, hyperinsulinemia can cause activation of both insulin and IGF pathways via IR-A homodimers and IR-A/IGF-1 receptor heterodimers, respectively. The clinical relevance is clear because heterodimers are found at higher levels in skeletal muscle and adipose tissue in type 2 diabetes.

CaRRE: The First Hormonally Responsive Splicing Regulation Element The Slo gene encodes BK calcium and voltageactivated potassium channels. These channels exhibit functional diversity, partially due to alternative splicing. Two alternatively spliced stress axisregulated exons (STREXs) have been identified that enhance repetitive firing when included. The skipping of these exons in adrenal chromaffin cells is increased

Table II Characteristics of Insulin Receptor Isoforms Insulin receptor isoform

IR-A (exon 11 skipping)

IR-B (exon 11 inclusion)

Tissue distribution

Ubiquitous; predominant during embryonic development

Insulin-sensitive tissues: liver, muscle, adipocytes, kidney

Insulin binding affinity

þþ

þ

Sensitivity to insulin action

þ

þþ

Insulin action in b cells

Increased insulin transcription

Increased glucokinase transcription

IGF-1 binding affinity IGF-2 binding affinity

þþ þ

 

IGF-1 receptor heterodimer

Binds insulin, IGF-1, IGF-2

Binds IGF-1

Expression in type 2 diabetes

Increased

Decreased

Expression in cancer cells

Increased

Decreased

p0120

178

Alternative Splicing

by hypophysectomy in rats. This effect is prevented by ACTH injections. Treatment of these cells with glucocorticoids increases skipping, whereas androgens decrease skipping. It is proposed that stress hormones regulate alternative splicing of STREX in adrenaline-secreting cells. One of the sequence elements involved in this alternative splicing has been identified. Depolarization of GH3 pituitary cells decreases STREX inclusion, and this is dependent on the action of a Ca2þ/ calmodulin-dependent protein kinase (CaMK IV ). CaMK IV in turn acts via an element [CaMK IVresponsive RNA element (CaRRE)] upstream of the STREX. Transfer of this element to another exon confers CaMK IV repressibility on that exon. STREX skipping reduces the BK channel Ca2þ sensitivity, and this provides a feedback pathway by which chronic depolarization will reduce the Ca2þ sensitivity of the cell.

Hormones as Regulators of Alternative Splicing Hormones regulate the alternative splicing of their target genes via multiple mechanisms, and many different second messenger systems mediate their effects (Fig. 5). Many of the downstream splicing factors have been identified, but the intermediaries remain elusive. Many kinases act to regulate alternative splicing, and the phosphorylation of splicing proteins is likely to be a key process. For example, terminal exon inclusion of the protein kinase C (PKC) bII pre-mRNA in skeletal muscle is up-regulated by insulin, which enhances insulin-stimulated glucose uptake. Insulin regulates the alternative splicing via the PI3K signaling pathway by phosphorylation of the SR protein, SRp40. Hormones acting via nuclear receptors may act on the C-terminal domain of RNA polymerase II to regulate the recruitment of splicing factors.

Environ stressors

Insulin

MKK3/6 p38 Phosphorylation of hnRNP A1

hnRNPs

PKC, ras

PI3K

SRPK 1/2 Clk/sty PP2Cg

Phosphorylation of SRp40

SR proteins

(A1, PTB, CUG-BP)

(SF2/ASF, SRp40, SC35)

Alternative splicing

CTD RNA pol II? recruitment of factors

? Vit D RA T3 dex

RNA helicases Unknown splicing factors

(p72)

CamK IV? (CaRRE) ACTH dex androgens depol

Figure 5 Possible pathways involved in the regulation of alternative splicing. Primary stimuli are indicated by the large arrows. dex, dexamethasone; depol, depolarization; environ, environmental; T3, tri-iodothyronine; RA, retinoic acid. Possible effector mechanisms are shown in boxes (examples of molecules). Second messenger molecules: Clk/sty, clk/sty kinase; MKK3/6 p38, p38 mitogen-activated protein kinase kinase; PP2Cg, protein phosphatase 2C gamma; CTD RNA pol II, C-terminal domain of RNA polymerase II; SRPK, SR protein kinases.

p0135

179

Alternative Splicing

SPLICING AND ENDOCRINE DISEASE

THERAPIES

Splicing defects are associated with an increasing array of disease processes and are particularly well represented in inherited endocrinopathies, such as congenital adrenal hyperplasia, multiple endocrine neoplasia, and neurofibromatosis type 1. Mutations that directly affect splicing may be classified into those that disrupt the splice sites and those that change non-splice site sequences. In the former class, mutations at 50 splice sites may cause activation of nearby cryptic 50 splice sites or, more commonly, skipping of the entire adjacent upstream exon. In the latter class, mutations of non-splice site sequences may disrupt regulatory elements for nearby splice sites, as previously described for isolated GH deficiency. An unusual endocrine splicing mutation occurs in the GH receptor gene, in which a deep intronic point mutation directly activates a pseudoexon, resulting in an additional 36 amino acids being included in the GH receptor, which in turn leads to GH insensitivity (Laron’s syndrome). Thus, point mutations may cause disease by disrupting splicing regulatory elements. The best described example of a disease caused by disruption of alternative splicing is myotonic dystrophy (DM1) and insulin resistance. DM1 is caused by CTG trinucleotide repeats at the 30 untranslated end of the DM protein kinase gene. In DM1 skeletal muscle, there are high levels of the hnRNP CUGbinding protein (CUG-BP), causing aberrant regulation of IR alternative splicing and a higher proportion of the IR-A isoform. Overexpression of this protein in normal cells also induces a switch to the IR-A isoform. This has been shown to occur via an intronic CUG-BP binding element upstream of the crucial exon 11. The same binding protein was previously shown to alter alternative splicing of the cardiac troponin T gene in both cardiac and skeletal muscle. Overexpression of CUG repeat RNA in normal cells appears to increase the half-life and steady-state levels of CUG-BP. Thus, a specific pathway has been proposed that leads to insulin resistance and possibly some of the other effects of this multisystem disorder. Interestingly, higher levels of the IR-A isoform have also been demonstrated in poorly differentiated cancer cells. An autocrine loop has been proposed in anaplastic thyroid cancer in which high levels of IGF2 have been associated with high levels of IR-A and IR-A autophosphorylation. Because IGF-2 binds IRA and not IR-B, it may be acting as a mitogen. The presence of heterodimerization with signaling through the IGF system enhances this effect.

Most attempts to alter splicing have focused on the disrupted splicing present in the cancer process, and several novel approaches to chemotherapy have been proposed. Complementary antisense oligonucleotides can switch alternative splicing but have problems with delivery and toxicity. Although suitable for cancer therapy, such approaches are less promising as longterm solutions to genetic disease caused by splicing dysregulation. However, two examples demonstrate the potential for novel therapies as our understanding increases. First, dystrophin synthesis is rescued in a Duchenne muscular dystrophy cell line by chimeric snRNAs with complementarity to exon 51 splice sites. Second, work is focusing on the development of compounds to treat human spinal muscular atrophy (SMA). Mutations in the SMN1 gene cause SMA, but all SMA patients can produce functional SMN protein from the SMN2 gene. This gene is identical to the SMN1 gene except for one nucleotide in exon 7. This single nucleotide disrupts an SF2/ASFdependent exonic splicing enhancer and leads to alternative splicing of exon 7. Exon 7 is therefore excluded from 80% of transcripts derived from the SMN2 gene, and the 20% full SMN protein produced is not sufficient to prevent disease. Several novel splicing compounds have been identified as enhancers of exon 7 inclusion, and the search continues for a way to upregulate exon 7 alternative splicing; however, several candidates have been found from cell culture work.

CONCLUSION Alternative splicing is an essential means of generating diversity from a limited number of genes. Genes coding for hormones, their binding proteins, and their receptors are particularly well represented. Alternative splicing is an integral component of the endocrine system because it both regulates and is regulated by hormones. Recent insights have shown that pre-mRNA has many sequence elements within introns and exons that regulate alternative splicing. Many factors acting directly on such elements have been described. The intermediates between a given hormone and its potential effect on a splicing factor are not known. Neither is it understood how splicing factors, many of which are ubiquitously and generously expressed, are regulated to produce subtle changes in spliceosome action. As further knowledge of the alternative splicing process and its relation to disease processes is obtained, it is hoped that novel therapies can be

p0155

180 designed that divert alternative splicing to the appropriate pathway.

See Also the Following Articles Alternative Promoters . Insulin-like Growth Factors

Further Reading Akker, S. A., Smith, P. J., and Chew, S. L. (2001). Nuclear posttranscriptional control of gene expression. J. Mol. Endocrinol. 27, 123–131. Caceres, J. F., and Kornblihtt, A. R. (2002). Alternative splicing: Multiple control mechanisms and involvement in human disease. Trends Genet. 18, 186–193.

Alternative Splicing

Cartegni, L., Chew, S. L., and Krainer, A. R. (2002). Listening to silence and understanding nonsense: Exonic mutations that affect splicing. Nat. Rev. Genet. 3, 285–298. Krainer, A. R. (1997). “Eukaryotic mRNA Processing.” IRL Press, Oxford. Maniatis, T., and Tasic, B. (2002). Alternative pre-mRNA splicing and proteome expansion in metazoans. Nature 418, 236–243. Stoss, O., Stoilov, P., Daoud, R., Hartmann, A. M., Olbrich, M., and Stamm, S. (2000). Misregulation of pre-mRNA splicing that causes human diseases. Concepts and therapeutic strategies. Gene Ther. Mol. Biol. 5, 9–30. Webster, N. J. G., and Huang, Z. (1999). Hormonal regulation of alternative splicing. In “Post-transcriptional Processing and the Endocrine System” (S. L. Chew, ed.), pp. 1–17. Karger, Basel, Switzerland.

Alzheimer’s Disease and Hormones Sigbritt Rasmuson and Tommy Olsson Umea˚ University Hospital, Umea˚, Sweden

ALZHEIMER’S DISEASE Glossary Alzheimer’s disease The most common cause of cognitive disturbance and dementia; a progressive neurodegenerative disease. apolipoprotein E (apoE) A protein that is important for the metabolism of lipids (e.g., cholesterol); there are three different alleles coding for three isoforms of apoE, and an increased frequency of the apoE e4 allele is seen in Alzheimer’s disease. cognition The process of knowing or perceiving. declarative memory Consists of items that are easy to verbalize and generally accessible to conscious recall; this includes both episodic and semantic memory. entorhinal cortex A structure within the hippocampal region that provides the primary pathway by which sensory information enters and leaves the hippocampal region. hippocampus A region of the brain that plays an important role in memory and learning and that has a high concentration of glucocorticoid receptors. neuroendangerment Exposure that makes neurons less likely to survive neurological insults. plaques and tangles Hallmark lesions in the brain in Alzheimer’s disease; plaques are extracellular deposits consisting of the protein b-amyloid, whereas tangles are abnormalities of the neural cytoskeleton.

A

lzheimer’s disease is the most common form of dementia. The hallmark lesions are neurofibrillary tangles and neuritic plaques. The hippocampal formation, a key area for learning and memory processes, is affected early during the course of the disease. A number of hormones may affect the development and progression of this disease. These include hormones readily entering the brain through the blood–brain barrier as well as hormones locally produced within the brain. Potential prevention of dementia as well as treatment of this devastating disease may be achieved in the future through manipulation of hormone effects on brain cells.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

In 1907, Alois Alzheimer described the first case of the most common cause of dementia. Alzheimer’s disease (AD) accounts for at least 50% of all cases of dementia today. The prevalence increases with age, from approximately 2 to 3% at 70 to 74 years to 12 to 16% at 85 to 90 years. The most common cognitive symptoms in AD are impairments in explicit memory, visuospatial problems, inability to calculate, loss of judgment, and progressive loss of language. Furthermore, there are various noncognitive, behavioral, and psychological symptoms such as depression, anxiety, and aggressiveness. The definitive diagnosis of AD is based on a combination of an appropriate clinical history and a histopathological investigation. There are two hallmark pathological brain lesions: neurofibrillary tangles and neuritic plaques. The hippocampus seems to be the brain region that shows the earliest neuron loss in AD, and the pattern of neurodegeneration results in a functional isolation of the hippocampus from the cortical association areas. There are disturbances in multiple neurotransmitter systems in AD, including a profound loss of cortical and hippocampal cholinergic innervation, reduced central somatostatin levels, and changes in glutamatergic and serotonergic activity. Neurotransmitters and circulating hormones play key roles in modulating the complex neurobiological interactions that occur within the brain. Many hormones are also produced locally in the brain. There are important bidirectional links between the central nervous system and the peripheral endocrine organs. Thus, lesions in the brain of different origins may influence endocrine function.

THE LIMBIC–HYPOTHALAMIC– PITUITARY–ADRENAL AXIS Components of the LHPA Axis The limbic–hypothalamic–pituitary–adrenal (LHPA) axis plays an important role in maintaining homeostasis

181

182 under basal and challenge conditions. The components of this axis function as an integrated feedback system in a hierarchic way (Fig. 1). The hippocampus, a limbic forebrain region, is an important neuroanatomical target for interactions between neurotransmitter systems and circulating hormones. The hippocampus has a high density of glucocorticoid receptors (GRs) and an inhibitory effect on cortisol release and is also important for negative feedback function. Steroid hormones have profound effects on the central nervous system (CNS). These include effects

Figure 1 Schematic diagram of the limbic–hypothalamic–pituitary–adrenal axis. In Alzheimer’s disease, there may be increased glucocorticoid production related to increased corticotropin-releasing hormone (CRH) secretion and disturbances in negative feedback. Circulating adrenal androgen levels are increased during the early phase of the disease, and this may indicate increased androgen production in the zona reticularis of the adrenal cortex or an altered metabolism of androgens. CNS, central nervous system; Ach, acetylcholine; 5-HT, serotonin; NE, norepinephrine; ACTH, adrenocorticotropin.

Alzheimer’s Disease and Hormones

on mood, memory, and learning. These effects are exerted by glucocorticoids as well as by other steroid hormones such as androgens and estrogens.

Experimental Studies Glucocorticoids have potent effects on the CNS, affecting cellular metabolism, receptor and ion channel density, neurotransmission, cell division, maturation, and cell survival. It was originally suggested that prolonged supraphysiological glucocorticoid levels with increasing age and/or stress downregulated central GR expression. Reduced GRs in key brain areas, notably the hippocampus, might decrease the capacity for feedback, leading to further glucocorticoid hypersecretion. This would then lead to a vicious cycle in which hippocampal neuronal cell death might ensue, leading to permanently attenuated feedback and glucocorticoid hypersecretion. There are several caveats to this hypothesis. High glucocorticoid levels have only a moderate influence on GR expression. Instead, neurotransmitter influx, notably by serotonin and noradrenaline, is of major importance. Furthermore, it is not clear whether excessive glucocorticoid exposure in vivo per se might lead to neuron death. It seems more likely that glucocorticoids may lead to neuroendangerment. Thus, glucocorticoids make neurons less likely to survive a brain insult. The aging brain, especially the hippocampus, appears to be more susceptible than other parts of the brain to adverse effects of glucocorticoids. In combination with other insults associated with advanced age such as brain ischemia, various pathological processes might be accelerated by hypercortisolism. On a cellular level, glucocorticoids have a clear impact on neurons. Thus, increased levels of glucocorticoids seem to reduce neurogenesis within the hippocampus in close connection to excitatory amino acid efflux. Second, reversible stress-induced modeling of dendrites in hippocampal neurons is mediated by glucocorticoids along with excitatory amino acids. Finally, excitability of hippocampal neurons is influenced to a major extent. Thus, chronic excess of glucocorticoids might cause or accelerate some aspects of degenerative brain aging, notably hippocampus-associated cognitive dysfunction in rodents and humans.

Glucocorticoids and Cognition In accordance with experimental data, several studies suggest an association between hypercortisolism, on the one hand, and hippocampal function and volume,

183

Alzheimer’s Disease and Hormones

on the other. Healthy adults treated with low/medium doses of synthetic glucocorticoids develop a reversible decrease in both immediate and delayed recall. In patients with Cushing’s syndrome with very high circulating cortisol levels, impaired declarative memory performance and mood changes are common. This is associated with reduced hippocampal volume. Notably, this volume loss seems to be a reversible phenomenon. In long-standing major depression, often associated with a moderate increase in glucocorticoid production, hippocampal volume reduction seems to be common. Interestingly, increasing serum cortisol levels in healthy elderly persons over years has been associated with decreasing hippocampal volume, with concomitant deficits in hippocampus-dependent memory tasks. In individuals with mild cognitive impairment (MCI), often preceding a later development of AD, salivary cortisol levels are inversely correlated with the result in an immediate recall task. From these data, it is not clear whether hypercortisolism is related to reversible hippocampal dysfunction or whether any association exists between hypercortisolism and future neurodegenerative disease.

present in mild to moderate AD, with increased excretion of A-ring-reduced metabolites in urine. This indicates that LHPA activation can be secondary to changes in cortisol clearance in AD. Alternatively, altered cortisol metabolism may be a protective mechanism that, together with down-regulation of GR expression in the periphery, can explain the absence of clinical features of hypercortisolism in peripheral tissues in AD (i.e., no increased prevalence of hypertension, diabetes mellitus, abdominal obesity, etc., as seen in Cushing’s syndrome). Importantly, because of tissue-specific differences in cortisol metabolism and receptor expression, the brain may be excessively exposed to glucocorticoids. Thus, increased cortisol levels in cerebrospinal fluid (CSF) have been found in AD. Adverse effects on the brain through increased glucocorticoid exposure may be worsened via an inability of GRs in the brain to down-regulate their expression when exposed to increased glucocorticoid levels. Thus, hypercortisolism may contribute to neuropsychiatric symptoms and accelerate neuronal damage in AD via tissue-specific alterations in pre-receptor glucocorticoid metabolism and receptor sensitivity/ reactivity.

Cortisol and Alzheimer’s Disease

p0040

Disturbances on several different levels of the LHPA axis are present in AD. Some studies have been performed on patients in advanced stages of the disease, but it seems reasonable to assume that it is important to consider studies on mild to moderate AD, where these disturbances are early events in the disease rather than a late response to extensive CNS damage or medications. There are several indicators of an increased central drive of the LHPA axis in AD, contributing to hypercortisolism. Postmortem studies have demonstrated increased corticotropin-releasing hormone (CRH) mRNA levels in the paraventricular nucleus of the hypothalamus. This putatively increased CRH secretion is associated with an increased cortisol but blunted adrenocorticotropin (ACTH) responsiveness to exogenous CRH in AD patients. An insensitivity to glucocorticoid feedback in AD probably also contributes to this increased activity through an inability to ‘‘shut off’’ temporary increases in LHPA axis activity. In line with this, increased glucocorticoid production has been reported in women with mild to moderate AD. These results show that increased glucocorticoid production is an early event in AD rather than a late response to chronic illness and/or medications. In addition, an altered cortisol metabolism is

ANDROGENS Adrenal androgens include dehydroepiandrosterone (DHEA), its sulfate (DHEAS), and androstenedione. DHEA enhances neuronal and glial survival and enhances memory retention in rodents, and placebocontrolled DHEA supplementation studies in humans have suggested that administration of DHEA may improve physical and physiological well-being. During the more advanced phases of dementia, levels of DHEAS are clearly decreased. The latter finding has been suggested to be of importance for progression of neurodegeneration, partly through loss of a glucocorticoid-antagonistic effect by DHEA, but this is a matter of controversy. In contrast, we have found increased levels of androgens in mild to moderate AD. This indicates an increased androgen production in the zona reticularis of the adrenal cortex or an altered metabolism of androgens. Furthermore, dynamic studies show an enhanced response of androgens, including DHEA, in AD after ACTH stimulation. In one longitudinal study on AD, lower levels of DHEA were associated with superior cognitive performance. Importantly, DHEA has no known receptor; instead, local brain metabolism of this hormone

184 may govern its biological activity. A cytochrome p450 enzyme, CYP7B, converts DHEA to 7a-hydroxyDHEA and may be responsible for putative antiglucocorticoid effects of DHEA. CYP7B mRNA is highly expressed in the human hippocampus, and this expression is significantly decreased in hippocampal neurons in patients with AD. Thus, alterations in androgen metabolism in AD might be important for tissue effects of adrenal androgens, with key interactions with glucocorticoids on a cellular level.

ESTROGEN AND OTHER FEMALE GONADAL HORMONES Estrogens have numerous effects on the brain, including influences on development and adult brain plasticity. Beneficial effects on neuronal plasticity and blockade of neurotoxic effects may prevent or retard the development of neurodegenerative disease, including AD. In favor of this hypothesis, estrogens regulate synaptogenesis in the rat hippocampus, notably in the CA1 subregion, which seems crucial for learning and memory functions. Synaptic spine density is related to circulating estradiol levels, and this is linked to memory function. These effects are associated with an increase in N-methyl-D-aspartate (NMDA) receptors in hippocampal neurons that relates to increased efficacy in long-term potentiation, that is, a proposed neurophysiological correlate to memory. Estrogens may also work in collaboration with neurotropins, mainly nerve growth factor (NGF), to stimulate neuronal development, differentiation, and growth via colocalized receptors on neurons in the rodent forebrain hippocampus and cerebral cortex. Estrogens also seem to promote neurogenesis. In addition, estrogens influence the function in several neurotransmitter systems, including the cholinergic, serotonergic, dopaminergic, and noradrenergic systems. Estrogen-induced enhancement of cholinergic functions via increased activity of choline acetyl transferase (ChAT) in the basal forebrain, hippocampus, and frontal cortex may be of particular relevance for AD. Related to this, the growth hormone response to pyridostigmine is increased in postmenopausal women taking estrogen replacement therapy, indicating an increased central cholinergic tone. As evident from studies of, for example, serotonin receptor expression after manipulation of estradiol and progesterone levels, it is important to notice that estradiol may have very different effects when given

Alzheimer’s Disease and Hormones

alone versus given in combination with progesterone regarding site and type of effects. This is also clear for effects on the noradrenergic system; estradiol given alone inhibits noradrenaline uptake, whereas estradiol followed by progesterone increases reuptake of this neurotransmitter. This influences the interpretation of hormone replacement studies in postmenopausal women that mainly has been done with estrogen in combination with a progestin. In vitro, estrogens have neuroprotective effects, including inhibition of b-amyloid formation from its precursor protein. This may be due partly to reduced accumulation of reactive oxygen and nitrogen species. Importantly, estrogens seem to inhibit the production of proinflammatory cytokines, notably interleukin-6 (IL-6). Glucocorticoid-induced hippocampal neuronal damage also seems to be reduced. In vivo estrogens reduce damage from ischemic stroke insults, and this may be relevant for protection against development and/or progression of AD. Thus, cerebrovascular disease may affect the clinical expression of AD. In contrast, high levels of estradiol might be neurotoxic following injury, and this can be relevant for early stages of AD. Neuroimaging studies in humans have implicated that estrogen influences the pattern of brain activation during memory processing, with regional increments in cerebral blood flow and glucose metabolism and with modulation of activity in specific brain regions affected during the early stages of AD. This may be partly related to direct effects on cerebral blood flow by estrogens. Epidemiological studies have suggested that estrogen is protective against the development and/ or progression of neurodegenerative disorders. Thus, low circulating levels of estrogen have been associated with an increased risk of AD. Some, but not all, case control studies, as well as a couple of prospective studies, have suggested that estrogen treatment may be protective against the development of AD. Importantly, there may be a protective effect of estrogen in long-term users of estrogen and/or when treatment is given during the latent preclinical stage of AD that may extend a decade or more before the onset of diagnosable dementia. Obviously, prospective randomized double-blind studies are needed for verification of these observations. Notably, circulating levels of estradiol have been reported to be slightly increased (and so not decreased) with concomitant elevated adrenal androgen hormone levels (DHEA and androstenedione) in patients with mild to moderate AD. This may reflect increased secretion and/or alterations in metabolism of these hormones during

Alzheimer’s Disease and Hormones

the early phases of AD. These alterations of endogenous gonadal hormone levels during the early stages of AD imply that there may be no need for increasing estradiol levels further by supplementation in mild to moderate dementia. Indeed, increased estradiol levels may even be neurotoxic, and it might be worth exploring beneficial effects on reducing the production of adrenal androgens during the early phases of neurodegeneration/AD, thereby also reducing an excessive increase in estrogens. The finding of a novel estrogen receptor, the estrogen b-receptor (ERb), has intensified the possibility of finding drugs that may act as neuroprotectants specifically through this receptor. The ERb has a clear role in the development of the cerebral cortex and also in survival of hippocampal neurons after exposure to excitatory neurotoxins. In contrast, ERa is the major receptor subtype expressed in basal forebrain cholinergic neurons. Thus, estrogen probably acts via ERa to enhance cognitive functions through the production of acetylcholine. On the other hand, ERb is the only estrogen receptor expressed in the dorsal raphe nucleus suggesting important effects on the serotonin system, indirectly affecting neuronal plasticity. In ERb knockout mice, spatial learning is impaired and treatment with cytotoxins causes marked apoptosis in hippocampus at doses that do not affect wild-type litter mates. Eight exon ERb knockout mice also show an increase in astroglia numbers, with a concomitant decrease in neuron number. This cell loss affects the limbic system, as well as the substantia nigra, to a major extent. Complex interactions among the subtypes of estrogen receptors is important to elucidate further, and it is not surprising that administration of estrogen compounds may induce very complex responses. Development of selective ligands for ERa and ERb may have profound effects on cognitive function and neuronal survival, notably in relationship to b-amyloid-induced neurotoxicity.

THE ROLE OF METABOLIC DYSFUNCTION p0105

A number of recent studies have suggested an association between AD and risk markers for cardiovascular diseases. These include hypertension, diabetes mellitus, lipid abnormalities, and the presence of the apolipoprotein E (apoE) e4 allele. Related to this, antihypertensive treatment and treatment with statins may reduce the incidence of dementia. There are several explanations for these associations, including

185 overlapping pathophysiology and clinical features, notably metabolic dysfunction. In the ‘‘metabolic syndrome,’’ associated with increased risk of type 2 diabetes and cardiovascular disease, hyperinsulinemia is a key element. Insulin receptors are widely distributed in the brain, mainly in the cerebral cortex and hippocampus. Insulin receptors are localized at the synapse, where they regulate neurotransmitter release and receptor recruitment. This indicates a role for insulin in synaptic plasticity. Disruption of cerebral insulin receptor functions leads to progressive cognitive impairments in rodents, and high insulin levels may directly influence the development of neuropathological changes in AD. Thus, insulin may be involved in the formation of neurofibrillatory tangles via its regulatory activity on tau phosphorylation. Insulin also seems to affect amyloid metabolism, inhibiting b-amyloid degradation. Desensitization of the neural insulin receptor reduces transport of glucose, the major nutrient for brain cells, and so might be a crucial link between metabolic dysfunction/hyperinsulinemia and cognitive dysfunction in AD. This could be worsened through secretion of b-amyloid because this protein may influence insulin binding and action through competitive binding to the insulin receptor. Furthermore, glucocorticoid overexposure to the brain, due to either a primary increase in glucocorticoid secretion or local changes in prereceptor metabolism, can decrease insulin sensitivity. Epidemiological studies indeed suggest a link among insulin resistance, diabetes mellitus, and AD. Ethnic factors may be of importance given that longitudinal studies of Caucasians and Japanese Americans have generated conflicting results, with clear associations between these factors found in a population-based study from Rotterdam, Netherlands. Interestingly, antidiabetic drugs may influence AD pathology. Troglitazone, a thiazolidindione antidiabetic agent acting as an ‘‘insulin sensitizer’’ via activation of peroxysome proliferator-activated receptor-g (PPAR-g) receptors, antagonizes an amyloid-stimulated proinflammatory response and neurotoxicity. This indicates a link among an inflammatory component of the metabolic syndrome, development of atherosclerosis, and pathology in AD. Proinflammatory cytokines produced by activated microglia, especially IL-1, seem to trigger enhanced synthesis of amyloid precursor protein and production of b-amyloid. Amyloid deposits per se stimulate further cytokine production by activated microglia, leading to a vicious cycle with continuous production

186 of amyloid precursor protein and b-amyloid. The link between inflammatory mechanisms and AD pathology has been strengthened by the association between head trauma and later development of AD. Epidemiological studies also suggest that nonsteroidal anti-inflammatory drugs (NSAIDs) can retard the development of AD. Notably, the positive epidemiological findings with NSAIDs are reported for drugs that activate PPAR-g receptors, lowering b-amyloid production (Ab42). This strengthens a possible link between insulin resistance and the development of b-amyloid-associated neurodegeneration. Another potential link between insulin resistance and a later development of cardiovascular disease might be the adipocyte-derived hormone leptin. Leptin is an important regulator of satiety and energy expenditure. Based on the fact that overweight is associated with high circulating levels of leptin with associated ‘‘leptin resistance,’’ a basic physiological function for leptin as protective against neuroendocrine consequences of starvation has been proposed. Increased leptin, strongly related to increased fat depot size, has been reported to predict the later development of myocardial infarction and stroke, independent of other risk factors. In patients with early AD, a physiological link between circulating cortisol and leptin levels over 24 h seems to be lost. This suggests that leptin regulation is disturbed during the early phase of neurodegeneration, and this may be linked to abnormalities in cardiovascular risk factors as well as weight loss in patients with AD. In summary, epidemiological and experimental data implicate that metabolic dysfunction precedes and may influence the development, progression, and symptomatology of AD.

GROWTH HORMONE The activity of the growth hormone (GH)–insulin-like growth factor-1 (IGF-1) axis declines during aging. In humans, GH secretion takes place in a pulsatile manner, regulated by stimulatory effects of growth hormone-releasing hormone (GHRH) and inhibitory input from somatostatin. Reduced central somatostatin levels have been found in autopsy studies in AD patients. In the periphery, circulating basal GH levels have been reported to be elevated in a few studies, but the diurnal pattern of GH secretion in AD seems to be unaltered. There also seems to be a high variability in reported responsiveness to a GHRH challenge. Related to these findings, it is not clear whether peripheral IGF-1 levels are altered in AD and, if so,

Alzheimer’s Disease and Hormones

whether this is a trait or a stage-dependent change in GH–IGF-1 axis function.

THYROID FUNCTION In the Rotterdam study of aged individuals, the relative risk of AD at follow-up was increased more than threefold for participants with reduced thyroid-stimulating hormone (TSH) concentrations at baseline. In contrast, no association was found between an increased TSH level and incident AD. Thus, the findings in the Rotterdam study are the first to suggest that subclinical hyperthyroidism among the elderly may increase the risk of AD.

CATECHOLAMINES Noradrenergic axons arising from the locus ceruleus (LC) project to several cortical areas, including the prefrontal and entorhinal cortexes. In the brain, the levels of norepinephrine (NE) are highest in the hypothalamus, and NE plays an important role in attention, arousal, and stress reactions as well as in cognition. The level of NE in plasma is widely used as a marker of the activity of the sympathetic nervous system (SNS), and a decline in NE levels is reported in aging individuals. In AD, studies assessing the SNS are relatively few, with small numbers of patients and varying severities of the disease. Decreases in cortical NE levels are described in various degenerative diseases of the brain, including AD, and a marked LC neuronal loss is considered as a classic postmortem pathological hallmark of AD. Despite these findings, increased concentrations of NE in CSF of AD patients have been described, possibly due to a compensatory activation of remaining LC neurons in this disorder or to increased turnover. Taking into consideration the great variation between studies concerning both severity of the disease and number of patients included, the overall result of the studies on SNS in AD points to a noradrenergic dysfunction that may contribute to cognitive impairment and behavioral symptoms.

See Also the Following Articles Acetylation . ACTH (Adrenocorticotropic Hormone) . Aging and Longevity of Human Populations . Aging, Immunology and . Brain, Effects of Steroid Hormones . Catecholamines . DHEA and the Elderly . Functional Genomics of Aging . Growth Hormone (GH) . Insulinlike Growth Factors . Leptin . Neuroendocrine System

Alzheimer’s Disease and Hormones

and Aging . Somatostatin Analogs . Stress, Aging, and Central Nervous System Interactions

Further Reading Alderson, A. L., and Novack, T. A. (2002). Neurophysiological and clinical aspects of glucocorticoids and memory: A review. J. Clin. Exp. Neuropsychol. 24, 335–355. Ferrari, E., Cravello, R., Muzzoni, B., Casarotti, D., Paltro, M., Solerte, S. B., Fioravanti, M., Cuzzoni, G., Pontiggia, B., and Magri, F. (2001). Age-related changes of the hypothalamic– pituitary–adrenal axis: Pathophysiological correlates. Eur. J. Endocrinol. 144, 319–329. Gasparin, I. L., Netzer, W. J., Greengard, P., and Xu, H. (2002) Does insulin dysfunction play a role in Alzheimer’s disease? Trends Pharmacol. Sci. 23, 288–293. Kalmijn, S., Mehta, K. M., Pols, H. A., Hofman, A., Drexhage, H. A., and Breteler, M. M. (2000). Subclinical hyperthyroidism and the risk of dementia: The Rotterdam study. Clin. Endocrinol. (Oxf.) 53, 733–737. Lupien, S. J., deLeon, M., deSanti, S., Convit, A., Tarshish, C., Nair, N. P. V., Thakur, M., McEwen, B. S., Hauger, R. L., and

187 Meaney, M. J. (1998). Cortisol levels during human aging predict hippocampal atrophy and memory deficits. Nature Neurosci. 1, 69–73. McEwen, B. S., and Sapolsky, R. M. (1995). Stress and cognitive function. Curr. Opin. Neurobiol. 5, 205–216. Rasmuson, S., Andrew, R., Na¨sman, B., Seckl, J. R., Walker, B. R., and Olsson, T. (2001). Increased glucocorticoid production and altered cortisol metabolism in women with mild to moderate Alzheimer’s disease. Biol. Psychiatry 49, 547–552. Rasmuson, S., Nasman, B., Eriksson, S., Carlstrom, K., and Olsson, T. (1998). Adrenal responsivity in normal aging and mild to moderate Alzheimer’s disease. Biol. Psychiatry 43, 401–407. Seckl, J. R., French, K. L., O’Donnell, D., Meaney, M. J., Nair, N. P., Yates, C. M., and Fink, G. (1993). Glucocorticoid receptor gene expression is unaltered in hippocampal neurons in Alzheimer’s disease. Brain Res. Mol. Brain Res. 18, 239–245. Seckl, J. R., and Olsson, T. (1995). Glucocorticoid hypersecretion and the age-impaired hippocampus: Cause or effect? J. Endocrinol. 145, 201–211. Wooley, C. S. (1998). Estrogen-mediated structural and functional synaptic plasticity in the female rat hippocampus. Horm. Behav. 34, 140–148.

Amidation Joseph Bell, Betty A. Eipper, and Richard E. Mains University of Connecticut Health Center, Farmington, Connecticut, United States

large dense core vesicle (LDCV) Also referred to as secretory granule. A membrane-delimited organelle formed from the trans-Golgi network that stores concentrated peptides and proteins for regulated secretion (secretion-on-demand). Standard electron microscopy preparations show an electron-dense core surrounded by a membrane.

other covalent modifications occur in specialized storage organelles called large dense core vesicles (LDCV). More than half of the known biologically active peptides have an a-amide group instead of a free carboxyl group on their COOH terminus (Fig. 1). Amidated fatty acids (e.g., oleamide) and other molecules have also been reported, and the formation of these a-amide groups may follow the same process established for a-amidated peptides. For peptides, the immediate precursor of the aamidated peptide is the corresponding peptide with an additional glycine residue at its COOH terminus. This peptidyl-glycine intermediate is formed from the initial larger precursor by the combined actions of endoproteases and carboxypeptidases; occasionally, the Gly residue is the final residue in the larger precursor. The a-amidation reaction begins with peptidylglycine and finishes with peptidyl-NH2, the bioactive peptide. A single amidating enzyme can produce peptides terminating with all 20 amino acid amides.

A

WHY IS PEPTIDE AMIDATION IMPORTANT?

Glossary a-amidated peptide A peptide with a carboxyterminal amide group (-CONH2) in place of a free -COOH group. a-amidation The enzymatic reaction that results in the presence of a carboxy-terminal a-amide group. bioactive peptide Any peptide with a biological activity. They typically bind to receptors at nanomolar or picomolar concentrations and show very high amino acid sequence specificity. Most bioactive peptides work through G protein-coupled metabotropic receptors with seven transmembrane domains.

midation is the process by which the -CONH2 structure is created. The side chains of asparagine and glutamine contain this moiety, but this article discusses the process by which the carboxy terminus of a peptide acquires an a-amide.

INTRODUCTION Endocrine cells and neurons communicate to other cells in the body (neurons, endocrine cells, muscles, glands, etc.) in part by secreting a variety of small molecules (e.g., glutamate, glycine, norepinephrine, and ATP), steroids, and peptides. Peptides are by far the most numerous class of secreted communication molecules. Bioactive peptides are normally synthesized as part of a larger inactive protein that is cotranslationally inserted into the lumen of the endoplasmic reticulum. The active peptide regions are then liberated by a small group of highly selective endo- and exoproteases. The majority of these cleavages and

188

For most amidated peptides, biological potency decreases two or three orders of magnitude when the a-amide group is absent, with either the intact glycine residue remaining or the free acid exposed after the a-amide group is removed (Fig. 2). The a-amide is important because its presence means that the COOH terminus of the peptide does not change charge as a function of physiological changes in pH. Thus, the uncharged COOH terminus of the peptide can associate closely with its membrane receptor, binding to a site on the receptor at or within the transmembrane domains. This leaves the NH2 terminus of the peptide exposed to the extracellular aqueous solution, enabling it to bind to additional portions of the peptide receptor and thus giving the peptide receptor pair more specificity and tighter binding than usually seen with smaller ligands, such as acetylcholine or glutamate.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

189

Amidation

O O OH H O H OH N COO− PAL Peptidyl NH2 + H COO− N COO− PHM Peptidyl H H Zn(II) 2 Cu(II) O2, 2 Ascorbate Peptidyl-α-hydroxylglycine Amidated peptide Glyoxylate Peptidylglycine H2O, 2 Semidehydroascorbate

Peptidyl

Figure 1

f0005

The two-step a-amidation reaction.

THE a-AMIDATION REACTION The peptide a-amidation reaction is a two-step process (Fig. 1). The first step is performed by peptidylglycine-a-hydroxylating monooxygenase (PHM: E.C. 1.14.17.3; also called peptidylglycine-a-monooxygenase). The second step is performed by peptidyl-a-hydroxyglycine-a-amidating lyase (PAL: E.C. 4.3.2.5; also known as peptidylamidoglycolate lyase). In the a-hydroxylation reaction, ascorbate reduces the two copper ions that are bound to PHM, yielding two semidehydroascorbate molecules that leave the enzyme. Cytosolic reducing equivalents, delivered by cytochrome b561, an integral membrane protein, are used to return the semidehydroascorbate to its fully reduced state. Only after the peptidyl-Gly substrate binds to PHM is molecular oxygen split, yielding peptidyl-a-hydroxy-Gly and water—soluble products that leave the enzyme. The two atoms of molecular oxygen are incorporated into the hydroxylated peptide and water. The a-hydroxylation reaction is closely analogous to the b-hydroxylation reaction performed by dopamine b-monooxygenase, which converts dopamine to norepinephrine and tyramine to octopamine. The most distinguishing features of this

Potency percent

100

10

1

0.1

TRH

CRH Calcitonin Gastrin Oxytocin

SP FMRF-amide

Peptides Free acid (-COOH) f0010

Amidated form (-NH2)

Figure 2 Relative biological potency of a-amided peptides (each normalized to 100%) compared to the corresponding free acid form of each peptide.

reaction are its absolute dependence on copper and its use of ascorbate and molecular oxygen. Although the second, or lyase, reaction can occur spontaneously as the pH is increased above 7.5, in cells it must be catalyzed by a separate enzyme. In LDCVs, with an internal pH of 5.0–5.5, the peptidyla-hydroxy-Gly product created by PHM is quite stable unless the lyase reaction is catalyzed by an enzyme. The lyase reaction, the stereospecificity of which matches that of PHM, requires enzymebound divalent metal, and bound zinc has been detected. It is not clear whether the zinc is catalytic or structural or both. A structural zinc could aid in the folding and stabilization of PAL, just as Ca2þ is thought to stabilize the prohormone convertases involved in propeptide endoproteolysis. A catalytic zinc could play a role like the one it plays in alcohol dehydrogenase, just as Zn2þ plays a crucial catalytic role in the carboxypeptidases.

THE PEPTIDE AMIDATING ENZYMES The two enzymatic activities necessary to perform the peptide a-amidation reaction are initially synthesized as a single bifunctional precursor protein called peptidylglycine-a-amidating monooxygenase (PAM) (Fig. 3). Mammals have a single gene encoding PAM. Alternative splicing generates several forms of PAM that differ in important ways. Elimination of the exon encoding the transmembrane domain yields a soluble, secreted enzyme. Elimination of the flexible linker region between PHM and PAL (exon 16) greatly reduces the ability of cells to separate the two catalytic activities. In LDCVs, the same prohormone convertases that cleave propeptide precursors also cleave the PAM precursor, producing soluble PHM, soluble and membrane-anchored PAL, and soluble bifunctional PAM. In invertebrates such as Drosophila and Cnidarians, PHM and PAL are encoded by separate genes. Elimination of the Drosophila PHM gene is lethal, causing death in the very late embryo and young larvae stages. In snails, both catalytic functions are encoded in the same gene, but there are four copies of the PHM domain, each with a similar

190

Amidation

Secretory granule

P PHM

PAL

CD

Exon16 P PHM

+

PAL

CD

PAL PAL

PHM Lumen Cytosol f0015

Figure 3 Forms of the amidating enzyme as found in mammalian LDCVs.

dependence on copper and ascorbate but a unique peptide substrate specificity. The catalytic core of PHM was defined using controlled protease digestion, and its structure was explored by assigning disulfides, examining site-directed mutants, and employing spectroscopy and X-ray crystallography. The catalytic core of PHM consists of two b-clamshell or sandwich domains. Each approximately 150-amino acid domain contains a single copper binding site. The NH2-terminal domain, with its three disulfide bonds, uses three His residues to bind Cu (the CuA or CuH site). The COOHterminal domain, with its two disulfide bonds, uses two His residues and one Met residue to bind Cu (CuB or CuM). The two domains are held together by a single hydrophilic linker strand, whereas the interiors of the domains are very hydrophobic. All of the histidine and methionine residues involved in coordinating the two catalytic copper ions are conserved in all known PHM sequences. Interestingly, all the spectral and crystallography data indicate that the two copper ions are farther apart than expected for a reaction requiring both copper ions to undergo a reduction–oxidation cycle. The X-ray structure places the two copper ions 11 A˚ apart and separated by a solvent-filled cleft; the Gly-extended peptide substrate binds closer to CuB. Dopamine b-monooxygenase shares many conserved disulfide bonds and contains histidine and methionine residues that may bind to copper in a similar manner. Based on sequence similarity, two additional potential family members, monooxygenase X and dopamine-b-hydroxylase-L, have been identified; their substrate specificity has not been determined.

PAL has not been studied as extensively. However, of the amino acid residues that are conserved among species, site-directed mutagenesis has identified a subset that are candidates for involvement in maintaining the structure of PAL and additional residues that are candidates for involvement in the PAL catalytic mechanism. Unlike the copper ions in the PHM reaction, which cycle from Cu1þ to Cu2þ during each reaction cycle, transferring reducing equivalents to molecular oxygen, the catalytic zinc presumably interacts directly with the a-hydroxylated substrate. PAL remains unique, with no close homologues identified in database screens. The yeast genome lacks enzymes homologous to either PHM or PAL.

FUTURE STUDIES ON PEPTIDE AMIDATION Future work should focus on the peculiar copper/ oxygen chemistry employed by PHM and its relatives and on the unique structure of PAL and the divalent metals it binds. Cell biological studies should focus on how membrane PAM enters and leaves LDCVs and how it traverses the endocytic pathway. The role of copper transporters (e.g., ATP7A and the Menkes protein), cytochrome b561, and ascorbate transporters in providing the right vesicular milieu must be elucidated. The biological consequences of disrupting amidation by interfering with copper metabolism (mottled/brindled mice) or the PAM protein (PAM knockout mice) must be explored.

Further Reading Bell, J., Ash, D. E., Snyder, L. M., Kulathila, R., Blackburn, N. J., and Merkler, D. J. (1997). Structural and functional investigations on the role of zinc in bifunctional rat peptidylglycine a-amidating enzyme. Biochemistry 36, 16239–16246. Eipper, B. A., Stoffers, D. A., and Mains, R. E. (1992). The biosynthesis of neuropeptides: Peptide alpha-amidation. Annu. Rev. Neurosci. 15, 57–85. Eipper, B. A., Milgram, S. L., Husten, E. J., Yun, H.-Y., and Mains, R. E. (1993). Peptidylglycine alpha-amidating monooxygenase: A multifunctional protein with catalytic, processing and routing domains. Protein Sci. 2, 489–497. Feng, J., and May, S. W. (2001). High-performance liquid chromatographic enantiomeric separation of an enzyme inhibitor which possesses both a chiral center and tautomeric moieties. J. Chromatogr. A 905, 103–109. Francisco, W. A., Merkler, D. J., Blackburn, N. J., and Klinman, J. P. (1998). Kinetic mechanism and intrinsic isotope effects for the peptidylglycine alpha-amidating enzyme reaction. Biochemistry 37, 8244–8252. Kolhekar, A. S., Mains, R. E., and Eipper, B. A. (1996). Peptidylglycine alpha-amidating monooxygenase (PAM): An ascorbate requiring enzyme. Methods Enzymol. 279, 35–43.

Amidation

Kulathila, R., Merkler, K. A., and Merkler, D. J. (1999). Enzymatic formation of C-terminal amides. Nat. Prod. Rep. 16, 145–154. Mains, R. E., and Eipper, B. A. (1999). Peptides. In “Basic Neurochemistry” (G. R. Siegel, et al., eds.), pp. 363–382. Lippincott-Raven, Philadelphia. Merkler, D. J. (1994). C-terminal amidated peptides: Production by the in vitro enzymatic amidation of glycine-extended peptides and the importance of the amide to bioactivity. Enzyme Microb. Technol. 16, 450–456.

191 Mounier, C. E., Shi, J., Sirimanne, S. R., Chen, B. H., Moore, A. B., Gill-Woznichak, M. M., Ping, D., and May, S. W. (1997). Pyruvate-extended amino acid derivatives as highly potent inhibitors of carboxyl-terminal peptide amidation. J. Biol. Chem. 272, 5016–5023. Prigge, S. T., Mains, R. E., Eipper, B. A., and Amzel, L. M. (2000). New insights into copper monooxygenases and peptide amidation: Structure, mechanism and function. Cell Mol. Life Sci. 57, 1236–1259.

Amiodarone and Thyroid Wilmar M. Wiersinga Academic Medical Center and University of Amsterdam, Amsterdam, The Netherlands

Glossary action potential Changes in the plasma membrane during depolarization and the subsequent repolarization of cardiomyocytes. The duration of the action potential is reflected by the QT time of the electrocardiogram (interval between the start of the QRS-complex and the end of T-top). Class III anti-arrhythmic drugs (such as amiodarone) act via prolongation of the action potential. colloid Proteinaceous substance in thyroid follicles, containing sizable quantities of iodine, thyroglobulin, and thyroid hormone. deiodination Enzymatic removal of iodine atoms from organic compounds. euthyroidism Normal thyroid function. IC50 value Concentration of a compound at which it inhibits a particular phenomenon by 50%. lysosome A membranous bag of hydrolytic enzymes; this organelle is used for the intracellular digestion of macromolecules. organification The incorporation of iodine atoms in organic compounds. thionamides Class of anti-thyroid drugs (such as carbimazole, methimazole, and propylthiouracil) that inhibit thyroid hormone synthesis by interfering with organification. g0045

thyroid hormone receptor Protein that, after binding of triiodothyronine (T3), binds to specific DNA sequences in the promoter region of T3-dependent genes, thereby modulating the transcription of these genes. Wolff-Chaikoff effect Decreasing yield of organic iodine from increasing doses of inorganic iodide.

A

miodarone is a very potent anti-arrhythmic drug, which is successfully used in the treatment of atrial fibrillation and lifethreatening ventricular arrhythmias. The drug, however, has many side effects. Amiodarone influences thyroid hormone secretion and metabolism in all patients taking the drug. In a

192

subset of patients, this results in amiodarone-induced hypothyroidism or thyrotoxicosis. Amiodarone also acts as a thyroid hormone receptor antagonist.

PHARMACOLOGY The structural features of amiodarone are its high iodine content and its resemblance to thyroxine (Fig. 1). The drug is prescribed as amiodarone hydrochloride (MW 681.82), which contains 37.25% iodine by weight. It was originally introduced in 1962 in clinical medicine for the treatment of angina pectoris, but later was found to be very efficacious in the treatment of cardiac arrhythmias. Amiodarone is classified as a class III anti-arrhythmic agent; it lengthens the duration of the action potential and repolarization time in cardiac tissues. It also has weak anti-adrenergic effects, it causes smooth muscle relaxation resulting in a dilation of coronary arteries and an increase in coronary blood flow, and it induces peripheral arterial vasodilation and a decrease in systemic blood pressure and afterload. Dosage forms of amiodarone (trade name Cordarone) are tablets or injections. Because of the large distribution volume, the onset of the drug’s action is delayed; consequently, a loading dose to saturate the large body stores is frequently required. The highest levels of amiodarone and its major metabolite desethylamiodarone (DEA) are found in adipose tissue, liver, and lung (Table I). The slow turnover of amiodarone from deep compartments such as adipose tissue explains its exceptionally long terminal half-life of
10 days (half-life for DEA is 57
27 days). The major metabolic pathways of amiodarone are N-dealkylation, resulting in its main (and biologically active) metabolite DEA, and deiodination, resulting in monoiodo-, desdiiodo-, and desethyldesdiiodoamiodarone. A daily oral dose of 200 mg amiodarone results in a urinary iodide excretion of approximately 14,000 mg/24 liter, which is approximately 45 times higher than the optimal daily iodine intake of 150–300 mg recommended by the World Health

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

193

Amiodarone and Thyroid

liver, heart, skin, corneal epithelium, and nerve fibers of amiodarone-treated patients.

I O

C2H5

C O

O CH2 CH2 N

C4H9

C2H5

I

Figure 1 Molecular structure of amiodarone (chemical name: 2-butyl-3-[3,5-diiodo-4-(b-diethylamino-ethoxy)-benzoyl] benzofuran).

Organization. Amiodarone medication thus causes chronic iodine excess. Extensive glucuro-conjugation of amiodarone occurs and biliary excretion and fecal elimination account for 65–75% of the ingested drug. Transplacental transfer of amiodarone and DEA varies from 10 to 20%. Plasma amiodarone and DEA concentrations in the newborn are approximately fourfold lower than those in the mother. Amiodarone and DEA concentrations in breast milk are higher than in the plasma of the mother, due to their high lipid solubility. Side effects are numerous, occurring in 80% of patients. Most prevalent are corneal microdeposits (almost 100%), gastrointestinal symptoms such as anorexia and nausea (80%), photosensitivity and unusual blue-gray skin discoloration of exposed areas (55–75%), and neurologic symptoms such as tremor, ataxia, and peripheral neuropathy (48%). Uncommon but severe adverse effects include pulmonary toxicity, liver failure, and proarrhythmias. Most side effects develop slowly and are related to the cumulative dose of amiodarone. The mechanism of amiodarone toxicity is multifactorial, but a drug-induced phospholipidosis with disturbed lysosomal function explains many side effects. Amiodarone as an amphiphilic drug binds strongly to intralysosomal phospholipids, rendering them indigestible by phospholipases. The bound complexes form the intralysosomal multilamellar inclusion bodies, which have been found in lung,

Table I Tissue Distribution of Amiodarone (A) and Desethylamiodarone (DEA) in Human Autopsies Tissue

A (mg/g)

DEA (mg/g)

Adipose tissue

316

76

Liver

391

2354

Lung

A/DEA 4.2 0.12

198

952

0.21

Kidneys

57

262

0.22

Heart Muscle

40 22

169 51

0.24 0.43

Thyroid

14

64

0.22

EFFECTS ON THYROID HORMONE SECRETION AND METABOLISM Amiodarone treatment invariably results in changes in the plasma concentrations of thyrotropin [thyroidstimulating hormone (TSH)] and thyroid hormones. There is an initial rise in plasma TSH starting in the first week of treatment, with a return to normal values after 3 months. This is due to a chronic iodine excess generated during biotransformation of the drug, which transiently inhibits the synthesis and release of thyroid hormones (the so-called Wolff-Chaikoff effect), explaining the rise in TSH. The thyroid usually escapes from these inhibitory effects and plasma TSH returns to normal values. Amiodarone simultaneously affects extrathyroidal thyroid hormone metabolism; it strongly inhibits type I iodothyronine-50 -deiodinase, which catalyzes the deiodination of thyroxine (T4) into triiododothyronine (T3) and that of reverse triiodothyronine (rT3) into 3,30 -diiodothyronine. Consequently, plasma T3 decreases and plasma rT3 increases. Plasma T4 and free thyroxine (FT4) concentrations also increase, predominantly due to a reduced metabolic clearance rate of T4 related to inhibition of T4 uptake in the liver. Inhibition of T4 entry into tissues decreases the availability of the substrate T4 for 50 -deiodination, thereby contributing to decreased production of T3 and decreased clearance of rT3. Chronic administration of amiodarone thus results in elevated plasma T4 and FT4 concentrations in the presence of a normal plasma TSH: a remarkable combination of test results.

AMIDARONE-INDUCED THYROID DISEASES The clinical diagnosis of amiodarone-induced hypothyroidism (AIH) and amiodarone-induced thyrotoxicosis (AIT) can be made very easily if the classical symptoms and signs of thyroid hormone deficiency or excess are present, but this is not always the case. Worsening of cardiac arrhythmia can be an important clue for the diagnosis of AIT. A TSH value within the normal reference range reliably excludes AIH and AIT. An elevated TSH with a low T4 or FT4 in plasma indicates AIH. When plasma TSH is suppressed and plasma T3 is elevated, the diagnosis of AIT is straightforward. However, a normal plasma T3

194

Amiodarone and Thyroid

Table II Incidence of Amiodarone-Induced Hypothyroidism (AIH) and Amiodarone-Induced Thyrotoxicosis (AIT) in Relation to Environmental Iodine Intake Iodine intake High

AIH

AIT

AIH þ AIT

Country

13.2%

1.7%

14.9%

USA, UK

Intermediate

5.7%

7.9%

13.6%

Low

6.4% 11.9%

18.4%

Spain, Australia, The Netherlands Italy, Belgium

does not exclude AIT (in view of the decreased T3 production due to inhibited deiodination of T4 into T3) and AIT may present as T4 toxicosis. AIH or AIT develops in 16% of amiodaronetreated patients (Table II). The incidence of AIH is higher in patients residing in areas with a high iodine intake and AIT occurs more often in regions with a low iodine intake. Amiodarone may also give rise to a small, firm goiter (induced by the iodine excess) in the presence of a normal level of TSH, but this occurs less frequently than AIH and AIT.

Amidarone-Induced Hypothyroidism AIH occurs more often in females than in males and most cases are seen in the first 18 months of treatment. It is caused by a failure of the thyroid gland to escape from the Wolff-Chaikoff effect, resulting in permanent inhibition of organification. This is more likely to occur in subjects with preexisting autoimmune thyroiditis. Consequently, AIH is, to a certain extent, predictable and women with preexisting thyroid peroxidase antibodies are at risk of developing AIH. Thyroidal radioiodine uptake is preserved in AIH despite the increased stable iodide pool; it is explained by the drop in organification, which normally also produces compounds (presumably iodinated lipids) that inhibit iodide uptake.

Discontinuation of amiodarone usually restores euthyroidism in 3–4 months, but permanent hypothyroidism may ensue in patients with preexistent autoimmune thyroiditis. Potassium perchlorate (which acutely inhibits thyroidal iodine uptake via the sodium–iodide symporter) may shorten the time to reach euthyroidism. Thyroxine medication is effective and allows continuation of amiodarone. Fetal hypothyroidism occurs in 11% of patients treated with amiodarone during pregnancy and should be treated at once with thyroxine.

Amidarone-Induced Thyrotoxicosis AIT occurs more often in males than in females and new cases continue to occur throughout the duration of treatment. Two types of AITwith a different pathogenesis have been distinguished (Table III). Type I occurs in patients with preexisting thyroid disease (Graves’ disease, nodular goiter), obviously caused by increased thyroid hormone synthesis due to overrepletion of intrathyroidal iodine stores by the iodine excess. Type II is the result of destructive thyroiditis caused by the cytotoxic effects of amiodarone and DEA on thyrocytes by interference with lysosomes. Disruption of the normal thyroidal architecture allows the release of colloid contents (very rich in thyroid hormone) into the circulation, causing thyrotoxicosis. The stores of thyroid hormone in the colloid are finite, which explains the often self-limited nature of type II. Treatment for AIT depends on its severity, which varies from mild to very severe, and on the cardiac condition, which may or may not allow discontinuation of amiodarone. In AIT type I, discontinuation of amiodarone is recommended, but patients are still thyrotoxic 6–9 months thereafter. Combination therapy using thionamides (which are less effective during iodine excess) with potassium perchlorate shortens

Table III Characteristics of Amiodarone-Induced Thyrotoxicosis Types I and II Type I

Type II

Underlying thyroid abnormality

Yes

No

Pathogenetic mechanism

Excessive thyroid hormone synthesis due to iodine excess

Excessive thyroid hormone release due to destructive thyroiditis

Goiter

Usually diffuse or nodular goiter

Occasionally small diffuse goiter

Thyroid radioiodine uptake

Low, normal, or high

Low

Serum interleukin-6

Normal or slightly elevated

Markedly elevated

Spontaneous remission Preferred drug treatment

Less likely Potassium perchlorate plus thionamides

More likely Glucocorticoids plus thionamides

Subsequent hypothyroidism

Unlikely

Possible

195

Amiodarone and Thyroid

t0020

Table IV

Hypothyroid-like Effects of Amiodarone in Various Tissues

Tissue effect

Hypothyroidism

Amiodarone

Amiodarone þ T3

Heart QT interval

"

"

N

Heart rate

#

#

N

b-adrenoceptor density

#

#

N

Ca2þ ATPase activity of myosin

#

#

# #

# #

N #

(")

"

N

Liver LDL receptor density Triglyceride lipase activity Adipose tissue Lipoprotein lipase activity

Note. ", increase; #, decrease; N, return to normal; ("), increase not significant.

the period until euthyroidism to less than 3 months. 131 I therapy is seldom feasible in view of the low radioiodine uptake. Total thyroidectomy has been performed successfully in resistant cases. In AIT type II, spontaneous recovery to euthyroidism is the rule within 3–5 months after stopping amiodarone. Faster improvement is usually obtained with prednisone given for 7–12 weeks in combination with thionamides. Although most authors still favor discontinuation of amiodarone in AIT type II, a favorable outcome under continuation of amiodarone is certainly possible. Despite all efforts, some patients do not respond to multidrug treatment with thionamides, potassium perchlorate, and steroids. There have been reports of fatal cases in which AIT patients have died of thyroid storm.

AMIDARONE AS A THYROID HORMONE RECEPTOR ANTAGONIST p0075

Amiodarone is prescribed for cardiac arrhythmias and angina pectoris, the rationale being induction of bradycardia, lengthening of the cardiac action potential, and depression of myocardial oxygen consumption. Essentially similar phenomena are observed in hypothyroidism. The hypothesis has thus been put forward that one of the main mechanisms of action of amiodarone is via induction of a local hypothyroidlike condition in extrathyroidal tissues, notably the heart (Table IV ). The hypothesis is quite attractive, in view of the markedly decreased tissue concentrations of T3 induced by amiodarone. However, other drugs, such as iopanoic acid (equally potent in inhibiting type I 50 -deiodination), do not induce hypothyroid-like effects. In support of the hypothesis is the finding that DEA inhibits the binding of T3 to its nuclear receptors, resulting in a dose-dependent

decrease in the expression of several T3-dependent genes. Interestingly, DEA is a competitive inhibitor of T3 binding to thyroid hormone receptor a1 (TRa1) (IC50 value 47 mM ) but a noncompetitive inhibitor of T3 binding to TRb1 (IC50 value 27 mM ). The intracellular concentrations of DEA reached in vivo are high enough (50–500 mM ) for the drug to be able to interfere with T3 binding. Protein–protein binding studies with TRb1 and the coactivator glucocorticoid receptor-interacting protein showed an inhibitory effect of DEA on the T3-dependent binding of the coactivator to TRb1. Further studies have indicated that residues on the outside of the TR ligand-binding domain are involved in the binding of DEA. The available studies provide good evidence that DEA rather than amiodarone itself is responsible for the hypothyroid-like actions. The drug appears to meet the criteria of a thyroid hormone receptor antagonist.

See Also the Following Articles Hypothyroidism, Causes of . Hypothyroidism, Systemic Manifestations of . Hypothyroidism, Treatment of . Iodine . Thyroid Hormone Action . Thyroid Hormone Metabolism . Thyroid Hormone Receptors . Thyrotoxicosis, Overview of Causes . Thyrotoxicosis, Systemic Manifestations . Thyrotoxicosis, Treatment

Further Reading Daniels, G. H. (2001). Amiodarone-induced thyrotoxicosis. J. Clin. Endocrinol. Metab. 86, 3–8. Martino, E., Bartalena, L., Bogazzi, F., and Braverman, L. E. (2001). The effects of amiodarone in the thyroid. Endocr. Rev. 22, 240–254. Newman, C. M., Price, A., Daview, D. W., Gray, T. A., and Weetman, A. P. (1998). Amiodarone and the thyroid: A practical

196 guide to the management of thyroid dysfunction by amiodarone therapy. Br. Med. J. 79, 121–127. Trip, M. D., Du¨ren, D. R., and Wiersinga, W. M. (1994). Two cases of amiodarone-induced thyrotoxicosis successfully treated with a short course of antithyroid drugs while amiodarone was continued. Br. Heart J. 72, 266–268. Trip, M. D., Wiersinga, W. M., and Plomp, T. A. (1991). Incidence, predictability and pathogenesis of amiodarone-induced thyrotoxicosis and hypothyroidism. Am. J. Med. 91, 507–511.

Amiodarone and Thyroid

Van Beeren, H. C., Jong, W. M., Kaptein, E. Visser, T. J., Bakker, O., and Wiersinga, W. M. (2003). Dronedarone acts as a selective inhibitor of 3,5,30 -triiodothyronine binding to thyroid hormone receptor-a1: In vitro and in vivo evidence. Endocrinology 144, 552–558. Wiersinga, W. M. (1997). Pharmacotherapeutics of the thyroid gland. In ‘‘Handbook of Experimental Pharmacology’’ (A. P. Weetman and A. Grossman, eds.), Vol. 128, pp. 225–287. Springer-Verlag, Berlin/Heidelberg, Germany.

Anderson’s Disease (Chylomicron Retention Disease) Nathalie Berriot-Varoqueaux and Marie-Elisabeth Samson-Bouma Xavier Bichat, Paris, France

Lawrence P. Aggerbeck University Pierre and Marie Curie and University of Paris-Sud, Paris, France

lipoproteins Particles in the plasma composed of lipids (phospholipids, free and esterified cholesterol, and triglycerides) and apoproteins (e.g., apoAI, apoB, and apoC) that are responsible for lipid and lipid-soluble vitamin transport. The major lipoprotein classes are very low-density lipoproteins (VLDLs), low-density lipoproteins (LDLs), and high-density lipoproteins (HDLs).

Charlotte Anderson. North American medical scientists used the term chylomicron retention disease to describe a cohort of patients with similar signs and symptoms in whom it appeared that chylomicrons were assembled but not secreted. Both terms have been used subsequently by physicians and scientists in other areas of the world as diagnoses for patients having the clinical and laboratory manifestations described in this article. The diagnosis is made on the basis of the clinical findings (diarrhea, steatorrhea, and lipid malabsorption), endoscopy (the presence of a “white hoary” layer or gele´e blanche on the small intestinal mucosa), the intestinal biopsy (vacuolated enterocytes that stain positively with Oil red), the levels of plasma cholesterol (hypocholesterolemia), the presence of apolipoprotein (apo)B-100 and the absence of apoB-48 in the plasma, and the absence of chylomicrons and apoB-48 in the plasma after a fat feeding. The parents of patients are asymptomatic.

microsomal triglyceride transfer protein (MTP) Protein heterodimer in the endoplasmic reticulum that is necessary for the formation of apo B-containing lipoproteins, such as chylomicrons or VLDLs.

CLINICAL DESCRIPTION

Glossary apolipoproteins plasma lipoproteins, corresponds to the which is produced editing.

(apo) Protein components of such as apoAI and apoB. ApoB-48 amino-terminal 48% of apoB-100, in the intestine following RNA

chylomicrons Triglyceride-rich lipoproteins secreted by the intestine after a meal. hypocholesterolemia (hypocholesterolemic disorders) Syndromes characterized by low levels of plasma cholesterol.

A

nderson’s disease and chylomicron retention disease are terms used for very similar, if not identical, rare hereditary (most likely autosomal recessive) hypocholesterolemic disorders characterized by a lipid malabsorption syndrome with steatorrhea (chronic diarrhea) and growth retardation.

INTRODUCTION The term Anderson’s disease has been used since 1970 by French medical scientists as the diagnosis for patients having manifestations similar to those of a patient described by the Australian physician

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

Three well-described, inherited hypocholesterolemic states, characterized by lipoprotein deficiency, affect apoB-containing lipoproteins: abetalipoproteinemia, familial hypobetalipoproteinemia, and Anderson’s disease or chylomicron retention disease. Abetalipoproteinemia is a rare autosomal recessive disease that manifests in infancy. It is characterized by profound hypocholesterolemia, hypotriglyceridemia, lipid malabsorption, diarrhea, retinitis pigmentosa, acanthocytosis, spinocerebellar degeneration, and the complete absence of apoB-containing lipoproteins. The molecular basis of the disease is a mutation in the gene encoding the large subunit of the microsomal triglyceride transfer protein (MTP), which results in a defect in lipoprotein assembly. Familial hypobetalipoproteinemia is an autosomal codominant

197

198

p0025

disorder. Two major phenotypes have been described. Individuals who are homozygous for null alleles, in which plasma apoB is absent, are phenotypically similar to individuals who have abetalipoproteinemia. Individuals in whom truncated forms of apoB are found in the plasma, whether homozygous or heterozygous, are generally asymptomatic clinically; however, in both cases, there are decreased plasma and low-density lipoprotein (LDL) cholesterol levels. Mutations in the apoB gene form the molecular basis of the disease. Anderson’s disease is clinically distinguishable from abetalipoproteinemia and null alleles of homozygous hypobetalipoproteinemia by the absence of acanthocytosis, retinitis pigmentosa, and severe neurological symptoms and by the presence of apoB100-containing lipoproteins. Anderson’s disease is distinguishable from heterozygous familial hypobetalipoproteinemia and homozygous familial hypobetalipoproteinemia (with truncated apoB) by the presence of diarrhea, malabsorption, and steatorrhea and an autosomal recessive mode of inheritance. At least 35 cases of Anderson’s disease in 26 families have been described. Sixteen cases (10 different families) are of North African origin (Algeria, Morocco, and Tunisia). Ten cases are from 9 Canadian families. The 9 remaining cases are from six different countries (Spain, Pakistan, Turkey, England, Lebanon, and the United States). Consanguinity has been described in 9 families, no consanguinity is present in 12 families, and no information regarding consanguinity is available for the remaining 5 families. Individuals with these disorders exhibit a malabsorption syndrome with steatorrhea and growth retardation under a normolipidemic alimentary regime. The mucosal surface of the small intestine, as observed by endoscopy, is covered with a whitish layer (“a white stippling-like hoar frosting” or gele´e blanche). Bloating of the stomach, osteomalacia, and rickets have been observed in several cases. Hepatic steatosis has been noted in four cases but without evolution to cirrhosis. Neuroretinal manifestations occur later and are less severe than in abetalipoproteinemia. Neurological signs most frequently consist of a loss of deep tendon reflexes. There is occasional alteration of position and vibratory senses, nerve conduction velocities, and evoked auditory and visual potentials. For patients diagnosed as adults, the neurological signs are more severe and may also include areflexia, ataxia, alteration in deep and vibratory senses, myopathy (with lipofuschine deposits on muscle biopsy), and polyneuropathy. Although nystagmus and delayed dark adaptation may occur, there is no visual loss

Anderson’s Disease (Chylomicron Retention Disease)

or retinitis pigmentosa. Acanthocytosis is typically absent. Institution of a low-fat diet supplemented with lipid-soluble vitamins (A and E) and essential fatty acids results in the resumption of normal growth and abatement of the gastrointestinal symptomatology. Departure from the low-fat diet results in recurrence of symptoms.

PLASMA LIPID AND LIPOPROTEIN AND BIOCHEMICAL ANALYSES Plasma cholesterol levels are decreased but remain higher than 50 mg/dl. Fasting plasma triglyceride levels are normal. Postprandially, there is no increase in plasma triglycerides, and chylomicrons are not detected. However, the absorption of luminal fatty acids and their consecutive esterification by epithelial cells appear normal. Although apoB-48-containing lipoproteins are absent from the plasma, lipoproteins containing apoB-100 are present but in decreased amounts. The plasma levels of high-density lipoproteins (HDLs) and apoAI, -AIV, -E, and -C are also decreased, and there are low levels of total lipids, phospholipids, carotenoids, and lipid-soluble vitamins (particularly vitamin E) as well as vitamins A, K, and D. Lipoprotein composition is abnormal in that it has decreased amounts of cholesterol and increased amounts of phospholipid and triglyceride. Very low-density lipoproteins (VLDLs) are increased in size, whereas LDLs and HDLs are decreased in size. Analysis of mRNA synthesis in the human intestine has shown the presence of mRNA for apoAI, -AIV, -E, -CII, and -CIII. ApoB messenger RNA appeared to be correctly edited in the intestine in two cases of chylomicron retention disease in one family. Normal MTP protein and activity were detected in intestinal biopsies of several patients.

IMMUNOCHEMICAL ANALYSES Despite the absence of apoB-48 in the plasma in this disease, this apolipoprotein as well as apoAI, -AIV, -CII, and -CIII have been detected by immunochemical techniques in the enterocytes of patients, along with the lipid components that are normally assembled into triglyceride-rich lipoproteins. Immunoprecipitation with polyclonal antibodies to apoB or apoAIV of the homogenates of organ cultures of intestinal biopsies from patients shows the presence of normal-sized apoB-48 and apoAIV in amounts three- to fivefold more abundant compared to those

199

Anderson’s Disease (Chylomicron Retention Disease)

found in normal individuals. Analysis of homogenates of organ cultures and of the culture media shows that the patients assemble and secrete some normalsized apoB-48 and apoAIV, which are coimmunoprecipitated and which float like chylomicrons. No apoB-100 is detected in the biopsy or culture medium.

ULTRASTRUCTURAL ANALYSES BY LIGHT AND ELECTRON MICROSCOPY Studies of the intestine by light microscopy have shown that villi are present in normal number and length but that the enterocytes are overloaded with fat droplets. The pattern and extent of lipid loading are variable among patients and also among biopsy sections for a given patient. In the regions of the villi that contain lipid-laden enterocytes there are always a few morphologically normal-appearing cells. The cells in the inferior approximately one-third of the villus characteristically show no accumulation of lipids. When examined by electron microscopy, the enterocytes in some regions of biopsies have an intracellular architecture like that found in normal fasted individuals in whom intra- and intercellular lipoprotein-like particles are not readily apparent and the Golgi apparatus is flat and nondistended. In other regions, the enterocytes contain large amounts of lipid particles. Many of these are chylomicron- and VLDL-sized particles (approximately 300 nm in diameter; range, 169–580 nm) in membrane-bound compartments. When clearly identifiable, the Golgi apparatus is frequently distended and empty, although membrane-bound compartments containing particles are in close juxtaposition. These membrane-bound particles resemble lipoprotein particles seen in normal fed individuals that are situated in a membrane-bound compartment and are seen budding from the lateral aspect or located near the Golgi apparatus. The identity of the membrane-bound compartment that contains the lipoprotein-sized particles is not entirely clear, and the composition of the particles has not been carefully defined. Other, larger particles (368–2127 nm mean diameter) appear to be lipid droplets that are free in the cytoplasm. These lipid droplets may derive from the breakdown of membrane-bound compartments that contain lipid particles unassembled with protein (putative second step triglyceride-rich particles). Large non-membrane-bound lipid droplets (presumably not assembled with protein) predominate in abetalipoproteinemia, whereas smaller membrane-bound

particles predominate in Anderson’s disease. Finally, smaller particles (63 nm mean diameter) are rarely found in the intercellular spaces of affected individuals, suggesting that secretion can occur. Even after treatment with a low-fat, lipid-soluble, vitamin-supplemented diet for at least 6 months and abatement of gastrointestinal symptoms, biopsies performed in patients after 12–15 h of fasting remained lipid laden with lipoprotein- and lipid-like particles with densities exceeding those found in the fed normal individual.

LINKAGE ANALYSES Using an autosomal recessive mode of transmission and highly polymorphic microsatellite markers [most frequently the (CA)n type], segregation analyses of four families, excluded as a cause of the disease significant regions of the genome surrounding the genes for apoAI, -CIII, and -AIV (15 cm on chromosome 11); the apoCII gene, which includes the apoCI and apoE genes (24 cm on chromosome 19); and the genes encoding three proteins involved in intracellular lipid transport—MTP (30 cm on chromosome 4) and fatty acid-binding proteins 1 (20 cm on chromosome 2) and 2 (30 cm on chromosome 4). No evidence of linkage was found for a distance of at least 5 cm on either side of the reported location of the apoB gene (results exclude the apoB gene in 13 cases from seven families with Anderson’s disease).

INTRACELLULAR PROCESSING OF APOLIPOPROTEIN B Two basic types of asparagine-linked glycans (Nglycosylation) are found on apoB-48 present in normal enterocytes. One form contains only highmannose oligosaccharides and represents the newly glycosylated protein in the endoplasmic reticulum. The other form contains, in addition to high-mannose glycans, complex oligosaccharides that arise by processing of some of the high-mannose glycans in the trans-Golgi apparatus. Five of six potential asparagine glycosylation sites may be used in apoB-48 based on results obtained with apoB-100, and there are probably one high-mannose chain and four complex-type oligosaccharide chains on the same molecule. The mixed glycosylation pattern (high-mannose and complex oligosaccharides on the same molecule) is apparently due to the masking of some of the oligosaccharide chains on apoB-48 to the action of Golgi

200 glycosyltransferases. The addition and modification of N-linked carbohydrates occur in distinct intracellular compartments. The advancement of a protein along the secretory pathway can thus be assessed by evaluating the differences in the sensitivity of the N-linked oligosaccharides to endoglycosidases. In Anderson’s disease, there is a time-dependent transformation of high-mannose endoglycosidase H-sensitive oligosaccharides of endoplasmic reticulum origin to complex endoglycosidase H-resistant oligosaccharides, added in the Golgi network, as in normal individuals. The apoB-48 containing oligosaccharides only partially sensitive to endo H (high-mannose plus complex oligosaccharides) has reached the trans-Golgi and is certainly already assembled with lipids and is probably destined to be secreted. In contrast, in abetalipoproteinemic patients, there is a single intracellular population of apoB-48 containing only high-mannose and hybrid glycans, indicating that apoB-48 does not reach the Golgi apparatus since the complex glycans that are added in the medial and trans-Golgi network are absent.

CONCLUSION p0085

Biochemical and ultrastructural analyses suggest that triglyceride-rich lipoprotein assembly takes place in enterocytes. ApoB-48 oligosaccharide processing indicates that the defect is not located between the endoplasmic reticulum and the Golgi apparatus but rather is distal to the trans-Golgi apparatus. Given that apo-AIV and apo-AI seem to be well secreted in patients, this implies a cargo-specific defect in transport between the trans-Golgi apparatus and the basolateral surface. Two independent genome-wide linkage analyses have shown that a locus on chromosome 5q31.1 segregates with affected status in several families affected with Anderson’s Disease, chylomicron retention disease, or chylomicron retention disease with the neuromuscular disorder Marinesco-Sjogren syndrome. In 10 affected individuals, coding sequence variants (including two frameshift, one splice site, and five missense mutations) were identified in both alleles of the SARA2 gene, which is located in the region of apparent homozygosity. The SARA2 gene product (sar1b) belongs to the Sar1-ADP ribosylation factor family of small GTPases which are involved in COP-coated vesicle mediated intracellular transport.

Anderson’s Disease (Chylomicron Retention Disease)

Acknowledgments This study was supported by the Centre National de la Recherche Scientifique; the Institut National de la Sante´ et de la Recherche Me´dicale (PROGRES, Grant 4P016D); the Ministe`re de l’Education, de la Recherche et de la Technologie; the Fondation pour la Recherche Me´dicale (N.B.-V.); the Caisse Nationale de l’Assurance Maladie des Travailleurs Salarie´s (Grant 4AIC01); and the Societe´ Nationale Franc,aise de Gastroente´rologie.

See Also the Following Articles Abetalipoproteinemia . Familial Low Syndromes, Hypobetalipoproteinemia

Cholesterol

Further Reading Anderson, C., Townley, R. R. W., Freeman, M., and Johansen, P. (1961). Unusual causes of steatorrhoea in infancy and childhood. Med. J. Aust. 11, 617–622. Berriot-Varoqueaux, N., Dannoura, A. H., Moreau, A., Verthier, N., Sassolas, A., Cadiot, G., Lachaux, A., Munck, A., Schmitz, J., Aggerbeck, L. P., and Samson-Bouma, M. E. (2001). Hereditary malabsorptive and hypocholesterolemic syndromes: Apolipoprotein B48 glycosylation in abetalipoproteinemia and Anderson’s disease. Gastroenterology 121, 1101–1108. Bouma, M. E., Beucler, I., Aggerbeck, L. P., Infante, R., and Schmitz, J. (1986). Hypobetalipoproteinemia with accumulation of an apoprotein B-like protein in intestinal cells. Immunoenzymatic and biochemical characterization of seven cases of Anderson’s disease. J. Clin. Invest. 78, 398–410. Dannoura, A. H., Berriot-Varoqueaux, N., Amati, P., Abadie, V., Verthier, N., Schmitz, J., Wetterau, J. R., Samson-Bouma, M. E., and Aggerbeck, L. P. (1999). Anderson’s disease: Exclusion of apolipoprotein and intracellular lipid transport genes. Arterioscler. Thromb. Vasc. Biol. 19, 2494–2508. Jones, B., Jones, E. L., Bonney, S. A., Patel, H. N., Mesenkamp, A. R., Eichen-Baum-Voline, S., Rudling, M., Myrdal, U., Annesi, G., Naik, S., Meadows, N., Quattrone, A., Islam, S. A., Naoumova, R. P., Angelin, B., Infante, R., Levy, E., Roy, C. C., Freemont, P. S., Scott, J., and Shoulders, C. C. (2003) Mutations in a Sar1 GTPase of COPII vesicles are associated with lipid absorption disorders. Nat. Genet. 34, 29–31. Kane, J. P., and Havel, R. J., (1995). Disorders of the biogenesis and secretion of lipoproteins containing the B apolipoproteins. In “The Metabolic and Molecular Basis of Inherited Disease” (C. R. Scriver, A. L., Beaudet, W. S., Sly, D., Valle, eds.), Vol. 2, pp. 1853–1885. McGraw-Hill, New York. Levy, E., Marcel, Y. L., Deckelbaum, R. J., Milne, R., Lepage, G., Seidman, E., Bendayan, M., and Roy, C. C. (1987). Intestinal apo B synthesis, lipids and lipoproteins in chylomicron retention disease. J. Lipid Res. 28, 1263–1274. Roy, C. C., Levy, E., Green, P. H. R., Sniderman, A., Letarte, J., Buts, J. P., Orquin, J., Brochu, P., Weber, A. M., Morin, C. L., Marcel, Y. L., and Deckelbaum, R. J. (1987). Malabsorption, hypocholesterolemia and fat filled enterocytes with increased intestinal apoprotein B. Chylomicron retention disease. Gastroenterology 92, 390–399.

Androgen Biosynthesis and Gene Defects Mark R. Vanderwel and Richard J. Auchus University of Texas Southwestern Medical Center, Dallas, Texas, United States

Glossary cryptorchidism The condition in which the testes fail to descend into the scrotum and are retained within the abdomen or inguinal canal. follicle-stimulating hormone (FSH) A gonadotropin secreted and released by the anterior pituitary. FSH stimulates the ripening of the follicles in the ovary and formation of sperm in the testes by acting on granulosa and Sertoli cells, respectively. haploinsufficiency A mutation in only one of the two alleles for an autosomal gene, resulting in a reduced gene dosage for the protein encoded on the normal allele. human chorionic gonadotropin (hCG) A hormone similar to the pituitary gonadotropin luteinizing hormone; it is produced by the placenta during pregnancy. hypospadias A congenital abnormality in which the labioscrotal folds have not completely fused, so that the opening of the urethra is on the underside of the penis.

luteinizing hormone (LH) A gonadotropin secreted by the anterior pituitary. LH stimulates androgen synthesis by the Leydig cells of the testis and the theca cells of the ovary; it also stimulates the ‘‘luteinization’’ of ovarian cells after ovulation, forming the corpus luteum, which makes progesterone. male pseudo-hermaphroditism A congenital abnormality in which the genitalia of a 46,XY infant are not completely masculinized, despite the presence of testes. Mu¨llerian structures Structures that develop from the paramesonephric duct in normal females. These include the fallopian tubes, uterus, and upper part of the vagina. Sertoli cells Cells found in the walls of the seminiferous tubules of the testis. They anchor and nourish the developing germ cells.

Leydig cells The cells interspersed between the seminiferous tubules of the testis. They secrete androgens in response to luteinizing hormone.

Wolffian structures Structures that develop from the mesonephric duct in normal males. These include the epididymis and vas deferens.

F

NORMAL ANDROGEN BIOSYNTHESIS

ormation of the male external genitalia in human beings requires the production of dihydrotestosterone during the critical period of sexual differentiation at 8–12 weeks of gestation. At this time, chorionic gonadotropin (hCG) stimulates responsive Leydig cells to convert cholesterol to testosterone, which is transformed to dihydrotestosterone in target tissues. Mutations in the genes encoding the enzymes, cofactor proteins, receptors, and stimulatory hormones may disrupt this process at various steps, leading to congenital defects in androgen production and male pseudo-hermaphroditism. Defects in gonadotropin-releasing hormone will not interfere with fetal androgen production but will prevent the production of luteinizing hormone (LH) and not allow pubertal progression. This article describes the syndromes caused by mutations in these genes, focusing on the clinical and biochemical features of disorders that impair enzymatic steps in dihydrotestosterone production.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

g0045

All steroid hormone production begins with the conversion of cholesterol to pregnenolone. A ‘‘mobile pool’’ of free cholesterol in the outer mitochondrial membrane (OMM) is physically inaccessible to the side chain cleavage enzyme (CYP11A, P450scc), which resides in the inner mitochondrial membrane (IMM). Stimulation of Leydig cells with hCG or LH, both of which bind to the same LH/hCG receptor, increases intracellular cyclic AMP (cAMP). The rise in cAMP induces the expression and activation of the labile steroidogenic acute regulatory (StAR) protein, which allows the cholesterol to flow from the OMM to the IMM, where CYP11A converts the cholesterol to pregnenolone. The microsomal enzyme CYP17

201

g0055

p0010

202 (P450c17) sequentially oxygenates pregnenolone (the 17a-hydroxylase reaction) and cleaves the C17— C20 bond (the 17,20-lyase reaction), yielding dehydroepiandrosterone (DHEA). Although human CYP17 hydroxylates pregnenolone and progesterone with comparable efficiencies, its 17,20-lyase activity is much more efficient for the D5-steroid 17a-hydroxypregnenolone (17Preg) than for its D4-congenor 17a-hydroxyprogesterone (17OHP). Consequently, DHEA is the substrate for 17b-hydroxysteroid dehydrogenase type 3 (17b-HSD3), which reduces DHEA to D5-androstenediol, and 3b-hydroxysteroid dehydrogenase/isomerase type 2 (3b-HSD2) converts this D5 steroid to testosterone (T). The sequence of these latter two reactions (17b-HSD3 and 3b-HSD2) may also proceed in reverse, with androstenedione (AD) as the intermediate. T produced by the testis diffuses into peripheral tissues and those tissues that contain the enzyme 5a-reductase type 2 (SRD5A2) (i.e., prostate, genital skin) metabolize T to the potent androgen dihydrotestosterone (DHT). The importance of all of these steps and the lack of adequate redundancy are demonstrated by the clinical disorders caused by mutations in the genes encoding the key proteins described herein.

DISORDERS OF LEYDIG CELL STIMULATION Hypothalamic Hypogonadism p0015

Because fetal T synthesis during weeks 8–12 of gestation is driven primarily by placental hCG, 46,XY children with defects in gonadotropin-releasing hormone (GnRH) or gonadotropin production are born with relatively normal male external genitalia. However, T production in late gestation is driven by LH, so micropenis is often present, and these individuals fail to experience puberty. Mutations in the genes for the orphan nuclear receptors DAX1 and SF1, the homeodomain transcription factors HESX1, LH3, and PROP1, the b-subunits of LH and follicle-stimulating hormone, the GnRH receptor, or the extracellular matrix protein anosmin-1 have been identified in patients with defects in androgen production and/or infertility due to abnormalities in hypothalamus and/or pituitary development. Most of these disorders are extremely rare and variable in their manifestations, but the more common syndromes whose genetic basis has been at least partially defined are discussed below.

Androgen Biosynthesis and Gene Defects

Kallman’s Syndrome and Variants Kallman’s syndrome refers to the combination of hypogonadotropic hypogonadism and anosmia. Consequently, 46,XY patients are born with normal male genitalia except for micropenis, and they fail to initiate puberty. The diagnosis is made when LH does not rise in response to a bolus of GnRH or GnRH analogue. Magnetic resonance imaging (MRI) with attention to the olfactory bulbs and midline structures is helpful but not essential, and olfactory testing should be performed. At the time of expected puberty, sex steroid replacement appropriate for gender is commenced. Treatment with pulsatile GnRH using a programmable pump will not only induce sexual maturation but may restore fertility in individuals of both sexes. Kallman’s syndrome can be sporadic, autosomal dominant or recessive, or X-linked. The best understood form of Kallman’s syndrome is the X-linked variety, which accounts for nearly half of cases and most often results from mutations in the KAL gene. This gene encodes an extracellular matrix protein called anosmin-1 that guides the migration of both the olfactory and GnRH-producing neurons from the olfactory placode to their proper location in the head and brain. Forms of hypothalamic hypogonadism without anosmia may result from mutations in the gene for the GnRH receptor, but these conditions are rare. In approximately half the cases of Kallman’s syndrome and its variants, the molecular basis is unknown. Septo-optic Dysplasia Developmental defects in midline structures often ablate hypothalamic–pituitary axes, and the growth hormone and GnRH–LH axes are particularly vulnerable. Severe defects, such as holoprosencephaly, characteristically involve large portions of the brain, but milder developmental defects can involve few structures. Septo-optic dysplasia refers to the combination of optic nerve hypoplasia and hypothalamic–pituitary maldevelopment. Evaluation and management are similar to those for Kallman’s syndrome, although vision testing and MRI evaluation are important for guiding follow-up and for prognosis. Mutations in the homeobox gene HESX1 have been identified in patients with septo-optic dysplasia, and the clinical severity roughly correlates with the impairment in DNA binding by the mutant protein.

p0020

p0030

Leydig Cell Agenesis or Hypoplasia (Testicular Unresponsiveness to hCG/LH) When a 46,XY fetus has a mutation in the LH/hCG receptor, Leydig cells fail to develop appropriately,

p0035

203

Androgen Biosynthesis and Gene Defects

p0045

and T production is impaired from conception. This defect leads to varying degrees of genital anomalies at birth, depending on the amount of hCG-independent T production prior to the 10th week of gestation and the severity of the hCG/LH receptor dysfunction. Because secretion of anti-Mu¨llerian hormone by the Sertoli cells is intact, 46,XY children with Leydig cell hypoplasia do not retain Mu¨llerian structures. The testes lack distinct Leydig cells on biopsy, and Sertoli cells may appear at puberty. However, the seminiferous tubules, if present, often show spermatogenic arrest, and the testes degenerate progressively. The 46,XY infants born with completely female genitalia may not be identified until puberty, when they present with failure to develop breasts and to undergo menarche. Milder forms in 46,XY infants cause undervirilization, including hypospadias, micropenis, and cryptorchidism. 46,XX females homozygous for LH receptor defects will have normal female genitalia and may experience some breast development at puberty, but with amenorrhea and infertility. The diagnosis is confirmed by low or absent T, AD, and 17OHP production in response to hCG stimulation testing. Basal and GnRH-stimulated gonadotropin values are elevated in pubertal subjects. Management depends on the age of diagnosis and the degree of virilization. When the defects are severe enough to produce phenotypically female genitalia, assignment of the female gender, with gonadectomy and estrogen replacement therapy at the time of expected puberty, is usually recommended. For less severely affected individuals with undervirilized male genitalia, surgery may be necessary to correct hypospadias, and testosterone therapy is used to stimulate phallic development and to virilize the patient at puberty. Leydig cell hypoplasia is an autosomal recessive condition due to mutations in the hCG/LH receptor. Several genetic defects have been reported, including missense, nonsense, and null mutations. The null mutations, such as Arg554Stop, are associated with the most severe clinical phenotypes.

VARIANTS OF CONGENITAL ADRENAL HYPERPLASIA The most common form of congenital adrenal hyperplasia (CAH) is 21-hydroxylase (CYP21, P450c21) deficiency, but CYP21 is not expressed in the gonads and does not participate in T biosynthesis. Other, less common forms of CAH that involve enzymes or proteins expressed both in the adrenals and the gonads are discussed below.

Lipoid CAH StAR Deficiency Because StAR facilitates the transport of cholesterol from the OMM to the IMM, inactivating mutations in StAR block the production of pregnenolone and thus impair all steroidogenesis, both in the adrenals and in the gonads. Under the stimulation of adrenocorticotropic hormone (ACTH) and LH, cholesterol esters massively accumulate in the adrenal glands and testes, respectively, affording the characteristic enlarged, lipid-laden adrenals from which the name lipoid CAH derives. Secondarily, sterol autooxidation products accumulate in the adrenals and Leydig cells, altering cell structure and ultimately provoking cell destruction. Both 46,XY and 46,XX individuals will have female external genitalia at birth. Affected 46,XY individuals have abdominal, inguinal, or intralabial testes, a blind vaginal pouch, and no uterus or fallopian tubes. Wolffian duct remnants may be preserved in 46,XY individuals secondary to low levels of StAR-independent steroidogenesis. All reported patients are diffusely hyperpigmented from pro-opiomelanocortin excess. Because the theca cells of the ovary do not normally make steroids during fetal and neonatal life, the ovaries of 46,XX subjects do not suffer lipid accumulation and cell death during childhood. Consequently, 46,XX subjects may produce enough estrogens in early puberty to undergo some breast development and may even menstruate until lipid accumulation and auto-oxidation obliterate ovarian function as well. The diagnosis of lipoid CAH is confirmed by low or absent glucocorticoids, mineralocorticoids, gonadal steroids, their precursors, and their metabolites in plasma and/or urine, even after stimulation. In particular, 3b-HSD2 deficiency is excluded by documenting low concentrations of not only the active D4 steroids but of the D5 precursors pregnenolone and 17Preg. On abdominal computed tomography scan or MRI, the lipid-laden adrenals are strikingly enlarged, displacing the kidneys caudad. Treatment requires replacement doses of glucoand mineralocorticoids in the newborn period, which must be continued throughout life. All affected 46,XY males have been reared as females, and orchidectomy is advised. Estrogen replacement therapy for individuals of both genotypes is required at puberty to initiate female secondary sexual characteristics, and low-dose testosterone may be used to elicit a female pattern of sexual hair growth. Lipoid CAH is an autosomal recessive disease with a male/female ratio of approximately 3/1; however, this ratio may be skewed

p0075

204 by ascertainment bias. This condition is rare in the United States and Europe, but is the second most common form of CAH in Japan and Korea. In one series, mutation Gln258Stop accounted for 80% of the affected alleles from Japanese and Korean subjects, suggesting a founder effect that causes the relatively high incidence of lipoid CAH in these countries. Mutation Arg182Leu was found in 78% of affected alleles from Palestinian subjects in the same series.

p0080

Side Chain Cleavage Enzyme (CYP11A or P450scc) Deficiency For many years, it was hypothesized that homozygous CYP11A deficiency caused lipoid CAH, but an absence of CYP11A would preclude placental progesterone synthesis and promote spontaneous abortion after the 8th to 10th week. However, haploinsufficiency of CYP11A has been shown to produce a milder clinical picture of lipoid CAH than in StAR deficiency. Tajima and associates described a 46,XY patient with clitoromegaly, a blind vaginal pouch, hyperpigmentation, and absent Mu¨llerian structures. This patient was raised as a female, and testes were removed from the inguinal region. Adrenal insufficieny with hyperplasia did not occur until the child was 4 years old, and no mutation was found in the gene for StAR. Instead, one allele of the gene for CYP11A had a 6 bp in-frame insertion, adding Gly-Asp between Asp-271 and Val-272. This mutant enzyme had no activity and appeared to impair the function of wild-type CYP11A when expressed in the same cells, suggesting a partial, dominant-negative mode of action, leading to less severe disease in early childhood rather than infancy. Treatment is similar to that for StAR protein deficiency, with gluco- and mineralocorticoid replacement to prevent life-threatening adrenal insufficiency at the time of diagnosis, plus estrogen replacement therapy at the time of puberty.

3b-HSD2 Deficiency The 3b-HSD enzymes catalyze the conversion of the D5 steroids pregnenolone, 17Preg, DHEA, and D5androstenediol to their corresponding D4 steroids progesterone, 17OHP, AD, and T, respectively. One of these conversions is required in the biosynthesis of all active steroid hormones, so severe 3b-HSD deficiency will also result in a form of CAH with impaired androgen production.

Androgen Biosynthesis and Gene Defects

46,XY individuals with 3b-HSD deficiency most frequently exhibit male pseudo-hermaphroditism with a small phallus, hypospadias, partial labioscrotal fusion, and possibly a urogenital sinus with a blind vaginal pouch. Testes usually lie in the lower inguinal region, and Mu¨llerian structures are absent. Paradoxically, 46, XX individuals often have trace clitoral enlargement and progressive masculinization if undertreated. Severe 3b-HSD deficiency can present with salt-wasting crisis from glucocorticoid and mineralocorticoid insufficiency within the first week of life. Less severe forms of 3b-HSD deficiency are usually diagnosed in genetic males because of genital abnormalities, but may be difficult to diagnose in females. Androgen production increases at puberty in both sexes but at a rate that is intermediate for males and females; consequently, girls show signs of androgen excess, but boys often develop gynecomastia. Fertility has been reported in affected individuals of both sexes. The diagnosis of 3b-HSD deficiency hinges on elevated ratios of D5 steroids to their D4 congeners. These ratios, which are already increased at baseline, are accentuated by cosyntropin stimulation and should reach > 12 SD above normal. Adult females with hirsutism often have high circulating DHEA-S concentrations with high ratios of D5 to D4 steroids, so extremely elevated ratios are required to confidently diagnose 3b-HSD deficiency. Therapy includes early gluco- and mineralocorticoid replacement in salt-wasting individuals to prevent life-threatening adrenal insufficiency, and similar replacement is used in non-salt-wasting individuals to limit sexual precocity caused by increased synthesis of adrenal DHEA-S. Females require estrogen replacement, and males require testosterone supplementation to achieve full development of secondary sexual characteristics. Two functional 3b-HSD genes are encoded on chromosome 1p13. The type 2 enzyme is the dominant isoform expressed in the adrenals and gonads, and its gene is mutated in 3b-HSD deficiency. Adult females who present with hirsutism, infertility, and relatively high DHEA-S concentrations do not have mutations in the 3b-HSD genes. The type 1 enzyme is expressed in the placenta but also in liver and skin, and this enzyme accounts for the peripheral conversion of D5 precursors to D4 steroids in 3b-HSD deficiency, leading to the paradoxical androgen excess in females. Mutations in the gene for the type 1 enzyme have not been reported and are probably lethal because, as with homozygous CYP11A deficiency, this disorder would compromise progesterone synthesis

p0095

p0100

p0105

p0110

Androgen Biosynthesis and Gene Defects

by the placenta. Mutations in the type 2 gene have been found from residues Ala-10 to Gly-294, with no mutations dominating large series. Salt-wasting subjects have completely inactive alleles, and non-salt-wasting individuals can harbor mutations that retain partial activity when expressed in heterologous systems.

CYP17 (P450c17) Deficiencies CYP17 catalyzes both the 17a-hydroxylase and 17,20-lyase reactions. Severe mutations will ablate both activities, but milder mutations may either partially impair both activities or preferentially impair 17,20-lyase activity with 17a-hydroxylase activity remaining relatively normal. Furthermore, the 17,20-lyase activity is particularly dependent on the interaction of CYP17 with its redox partners, cytochrome P450 oxidoreductase (CPR) and cytochrome b5. In particular, the presence of cytochrome b5 increases the 17,20-lyase activity 10-fold but minimally influences 17a-hydroxylase activity. Thus, alterations in redox partners and/or their interactions with CYP17 may preferentially impair 17,20-lyase activity.

p0120

p0125

Combined 17a-Hydroxylase/17,20-Lyase Deficiency The classical description of severe, combined, 17ahydroxylase/17,20-lyase deficiency is sexual infantilism and hypokalemic hypertension in both genetic sexes. The presentation in 46,XY individuals varies from completely female external genitalia with a blind vaginal pouch to an undervirilized male genital phenotype with hypospadias and a small phallus. Testes in these subjects may be intra-abdominal, in the inguinal canal, or in labioscrotal folds. Mu¨llerian structures are absent, and Wolffian derivatives are usually hypoplastic. Severely affected patients will fail to develop secondary sexual characteristics, including pubic and axillary hair. With milder disease, males may develop gynecomastia at puberty and females may progress to Tanner stage 5 with ovulatory menses. Based on studies with recombinant enzyme, it has been estimated that at least 25% of normal fetal CYP17 activities are required for masculinization of the external genitalia. Hypertension with hypokalemia occurs in childhood and is often severe, with myopathy and hypertensive complications at a young age. The diagnosis should be entertained not only in male pseudo-hermaphrodites, but also in any individual with hyporeninemic hypertension, hypokalemic alkalosis, and a suppressed aldosterone. This diagnosis may be confirmed by obtaining elevated serum levels

205 of ACTH and the precursors that accumulate proximal to the block in 17-hydroxylation: progesterone, 11-deoxycorticosterone (DOC), corticosterone (B), 18-hydroxy-DOC, and 18-hydroxy-B. Both DOC and B have mineralocorticoid activity, leading to hypertension and hypokalemia. However, signs of adrenal insufficiency rarely develop, because the weak glucocorticoid B is present in abundance. Gonadotropins are elevated at puberty, and serum concentrations of aldosterone, 17OHP, cortisol, and sex steroids are low or absent. Treatment includes replacement of glucocorticoids to suppress DOC and B secretion and thereby to normalize potassium homeostasis and blood pressure. Mineralocorticoid receptor antagonists, such as spironolactone, can be added to reduce the doses of glucocorticoids and to prevent iatrogenic Cushing syndrome. At puberty, sex steroid replacement is indicated, and gonadectomy should be performed in 46,XY patients assigned a female sex of rearing. Mutations in CYP17 have been identified throughout the protein, and some missense mutations retain partial activity when expressed in heterologous systems. Most mutations that yield a completely inactive enzyme also destabilize the enzyme structure and ablate heme binding. A deletion of Phe-53 has been found in several Japanese subjects, and a CATC duplication following Ile-479 has been described in both Dutch Frieslanders and Canadian Mennonites. Recent reports suggest that 17-hydroxylase deficiency is most common in Brazil, where mutations W406R and R362C dominate. Isolated 17,20-Lyase Deficiency When only the 17,20-lyase activity is deficient, as has been suggested in 18 case reports, adrenal gluco- and mineralocorticoid synthesis is normal, but testosterone synthesis is impaired. Therefore, serum potassium and blood pressure are normal, but sexual development may be hampered. In 46,XY patients, the external genitalia are that of an undervirilized male, due to some residual 17,20-lyase activity. Mu¨llerian structures are absent, Wolffian derivatives are either hypoplastic or normal, and testes may be intra-abdominal, inguinal, or in the scrotum. In 46,XX females, this condition is believed to lessen adrenarchal and pubertal development, but no confirmed cases have been studied in detail. Cosyntropin stimulation tests yields normal or elevated 17-hydroxysteroid values, including cortisol and 17OHP, but DHEA and AD do not rise proportionately. Similarly, hCG stimulation testing will produce an increase in 17OHP concentrations, but AD and

p0130

p0135

206 T do not rise normally. The ratio of the rise in the C19 steroids to their C21, 17-hydroxy precursors is the most discriminatory parameter for diagnosing isolated 17,20-lyase deficiency. Only six cases of isolated 17,20-lyase activity have been confirmed with molecular genetic and biochemical studies. These subjects were all 46,XY individuals homozygous for mutations Arg358Gln (one) or Arg347His (three) or heterozygous for a completely inactive allele and one copy of Arg347Cys (two). These residues, Arg-347 and Arg-358, are both located on the redox-partner-binding surface of the CYP17 enzyme. Consistent with the known dependence of the 17,20-lyase activity on the interaction of CYP17 with redox partners CPR and cytochrome b5, these mutations at this binding surface appear to preferentially impair 17,20-lyase activity. Gonadectomy is recommended in 46,XY subjects raised as females and gender-appropriate sex steroid replacement will be necessary at the time of puberty. Optimal management in 46,XX subjects has not been established.

DEFECTS AFFECTING ONLY TESTOSTERONE AND DIHYDROTESTOSTERONE PRODUCTION The final two genetic disorders discussed involve the terminal steps of Tand DHT biosynthesis. These two conditions are unique in that only males (46,XY) experience clinical manifestations that are solely due to androgen deficiency in utero. Furthermore, other enzymes partially compensate for the genetic deficiencies, but only at puberty.

17b-Hydroxysteroid Dehydrogenase Type 3 (17b-HSD3) Deficiency

p0170

The human genome contains several 17b-HSD isoforms, but 17b-HSD3 is the enzyme that is defective in the clinical entity ‘‘17b-HSD deficiency,’’ also known as ‘‘17-ketosteroid reductase deficiency.’’ The 17b-HSD3 enzyme catalyzes the conversion of C19, 17-ketosteroids to 17b-hydroxysteroids using NADPH as cofactor: AD to T, DHEA to D5-androstenediol, 5a-androstanedione to DHT, and 5a-androsterone to 5a-androstanediol. Because 17b-HSD3 is expressed exclusively in the testes, the loss of this enzyme impairs androgen biosynthesis only in males. Most affected 46,XY individuals with 17b-HSD3 deficiency have predominantly female external

Androgen Biosynthesis and Gene Defects

genitalia with a blind vaginal pouch. Surprisingly, Wolffian derivatives, such as the epididymis, vas deferens, seminal vesicles, and ejaculatory duct, are present, suggesting that an alternate pathway in these tissues enables some testosterone production, perhaps mediated by the 17b-HSD type 5 isoform. Testes are usually located in the inguinal canal, and Mu¨llerian structures are absent. Most of these children are raised as females. At puberty, testicular AD production increases and significant extraglandular conversion of this AD to T elicits marked physical changes. The phallus enlarges and can reach lengths of 4 to 8 cm; the voice may deepen, male body hair develops, and muscle mass increases. Several affected individuals have changed gender role from female to male in adolescence because of the prominent physical and psychological masculinization they experience. In contrast, 17b-HSD3 is not expressed in the human ovary, so 46,XX patients with this disorder are asymptomatic. The diagnosis is based on markedly elevated AD concentrations in the face of low T in the neonatal period or in adolescence. The discrepancy in the AD/T ratio is accentuated with hCG stimulation. In the past, affected 46,XY males were frequently raised as females and underwent castration followed by estrogen substitution therapy at puberty, but infants with adequate phallic structures and mild hypospadias may be reared as males and undergo genitoplasty. This approach anticipates the tendency for gender reversal associated with virilization at puberty. However, even within members of a kindred with identical genotypes, affected individuals vary in their decisions about gender reversal at puberty when reared initially as females. If the patient is reared as a male, T replacement at puberty is necessary to achieve full masculinization and to prevent the development of gynecomastia. Spermatogenesis is absent because intratesticular Tsynthesis is blocked, and postpubertal elevations in gonadotropins may increase the risk for testicular neoplasms. Most mutations in the gene for 17b-HSD are located on exon 9 and impair all enzyme functions. One common mutation, identified in both Brazilian and Palestinian subjects, is R80Q, which lies in the Rossman fold area and primarily disrupts the binding of cofactor, but not of steroid.

p0180

5a-Reductase Type 2 (SRD5A2) Deficiency The disorder 5a-reductase deficiency (also known as pseudo-vaginal perineoscrotal hypospadias)

p0190

t0005

Table I

Clinical Manifestations in 46,XY Individuals Leydig cell hypoplasia

Lipoid CAH

3b-HSD2 deficiency

17-OHase deficiency

Isolated 17,20-lyase deficiency

17b-HSD3 deficiency

5a-Reductase 2 deficiency

Possible appearance of genitalia

Female/ambiguous/ hypoplastic male

Female

Ambiguous/ hypospadic male

Female/ambiguous/ hypospadic male

Female/ hypoplastic male
hypospadias

Female/ ambiguous

Ambiguous hypospadias, small phallus

Wolffian duct derivatives

Absent/hypoplastic

Absent/hypoplastic

Normal

Absent/hypoplastic

Hypoplastic/normal

Hypoplastic

Normal

Mu¨llerian duct derivatives Absent Gonads in 46,XY Testes, no a Leydig cells

Absent

Absent

Absent

Absent

Absent

Absent

Testes lipid laden

Testes

Testes

Testes

Testes

Normal testes

Adrenal insufficiency?

No

Severe

Severe–mild

No

No

No

No

Virilization at puberty? Increased hormone concentrations

None to poor LH, FSH

None LH, FSH, renin

Poor to mild LH, FSH, pregnenolone, 17Preg, DHEA

None to poor LH, FSH, DOC, B, Progesterone

Poor LH, FSH

Yes LH, FSH, Estrone, AD

Yes T

Decreased hormone concentrations

T, DHT

All adrenal & gonadal steroids

Progesterone, 17OHP, AD, T, DHT

Renin, cortisol, 17OHP, DHEA, AD, T, DHT

DHEA, AD, T, DHT T, DHT

DHT

Chromosomal location

2p21

StAR: 8p11.2 CYP11A: 15q23–q24

1p13

10q24.3

10q24.3

2p23

a

9q22

Small undescended testes with absent or decreased numbers of Leydig cells to descended testes of normal size with decreased numbers of Leydig cells.

208 provides strong genetic evidence that DHT is required for complete formation of the male genitalia in human beings. The 5a-reductases catalyze the conversion of T to its 5a-reduced metabolite DHT, and the type 2 isoform executes this transformation in the prostate and genital skin. Consequently, 46,XY male infants with a deficiency in the type 2 enzyme are born with hypospadias and a phallic structure that resembles an enlarged clitoris, often bound in chordee. The urogenital sinus with a blind vaginal pouch opens on the perineum, and the scrotum is bifid, with testes located in the inguinal canal or labioscrotal folds. With the abundance of T, Wolffian structures are well differentiated, and Mu¨llerian structures are absent. The ejaculatory ducts terminate in the blind vaginal pouch or onto the perineum next to the urethra, and the prostate is hypoplastic. As is the case in 17b-HSD3 deficiency, masculinizing changes occur at puberty as circulating T concentrations rise into the normal adult male range. The voice deepens, muscle mass increases, the phallus grows to 4 to 8 cm, and the subject may experience erections. The testes enlarge, the scrotal structure becomes rugated and pigmented, and spermatogenesis may occur, but is often impaired from cryptorchidism. Facial hair and body hair are sparse and acne and temporal hair recession do not occur, presumably because DHT production is low. As in 46,XY infants with 17b-HSD3 deficiency, most 46,XY individuals with 5a-reductase deficiency are reared as females but reverse gender role with the masculinizing changes of puberty, and in some cultures where the disorder is endemic, this process has achieved a socially acceptable status. However, unlike 17b-HSD3 deficiency, gynecomastia does not develop. The 46,XX females with 5a-reductase deficiency are phenotypically normal at birth, but at puberty they have decreased body and sexual hair and delayed menarche yet normal fertility. Diagnostic testing includes measurement of serum T and DHT, and a T/DHT ratio >30 confirms the diagnosis. One pitfall of testing is that after puberty, the activity of the type 1 isozyme may provide measurable levels of DHT, emphasizing the importance of the T/DHT ratio. Treatment of 5a-reductase deficiency is DHT therapy, often applied as a cream to the genitalia, to increase phallic length and to facilitate hypospadias repair. Supraphysiologic dosing of T in adults may generate adequate DHT via a partially functional type 2 enzyme and via the type 1 isoform.

Androgen Biosynthesis and Gene Defects

Approximately 40% of children with 5a-reductase deficiency are born to consanguineous parents; uniparental disomy has also been described. The founder mutation Arg246Trp is prevalent in the Dominican Republic, where the disease is common, but other mutations are found in other kindreds there. A 20 kb deletion that includes the SRD5A2 gene is prevalent in Papua New Guinea, and an A insertion into amino acid 251 causes 5a-reductase deficiency in a Turkish cluster. Several mutations are found in Brazil, and Gln126Arg is the most common. The type 1 isozyme is expressed in the liver up to 2 to 3 years of age and in nongenital skin. Mutations in the type 1 isozyme have not been described. Table I compares clinical manifestations in 46,XY children born with these disorders. Hypothalamic hypogonadism is not included, because boys with this disorder generally have normal genitalia, except for micropenis.

See Also the Following Articles Adrenal Androgens . Androgen Biosynthesis Inhibitors and Androgen Receptor Antagonists . Androgens, Gender and Brain Differentiation . Androgen Insensitivity Syndrome . Estrogen and the Male . Pseudohermaphroditism, Male, Due to 5a-Reductase-2 Deficiency . Sexual Function and Androgens . Undescended Testes

Further Reading Achermann, J. C., Weiss, J., Lee, E. J., and Jameson, J. L. (2001). Inherited disorders of the gonadotropin hormones. Mol. Cell. Endocrinol. 179, 89–96. Andersson, S., Geissler, W. M., Wu, L., Davis, D. L., Grumbach, M. M., New, M. I., Schwarz, H. P., Blethen, S. L., Mendonc,a, B. B., Bloise, W., Witchel, S. F., Cutler, G. B., Jr., Griffin, J. E., Wilson, J. D., and Russell, D. W. (1996). Molecular genetics and pathophysiology of 17b-hydroxysteroid dehydrogenase 3 deficiency. J. Clin. Endocrinol. Metab. 81, 130–136. Auchus, R. J. (2001). The genetics, pathophysiology, and management of human deficiencies of P450c17. Endocrinol. Metab. Clin. N. Am. 30, 101–119. Bose, H. S., Sugawara, T., Strauss, J. F., III, and Miller, W. L. (1996). The pathophysiology and genetics of congenital lipoid adrenal hyperplasia. N. Engl. J. Med. 335, 1870–1878. Geller, D. H., Auchus, R. J., Mendonc,a, B. B., and Miller, W. L. (1997). The genetic and functional basis of isolated 17,20-lyase deficiency. Nat. Genet. 17, 201–205. Imperato-McGinley, J., and Zhu, Y. S. (2002). Androgens and male physiology in the syndrome of 5a-reductase-2 deficiency. Mol. Cell. Endocrinol. 198, 51–59. Moisan, A. M., Ricketts, M. L., Tardy, V., Desrochers, M., Mebarki, F., Chaussain, J. L., Cabrol, S., Raux-Demay, M. C., Forest, M. G., Sippell, W. G., Peter, M., Morel, Y., and Simard, J. (1999). New insight into the molecular basis of 3b-hydroxysteroid

p0210

Androgen Biosynthesis and Gene Defects

dehydrogenase deficiency. J. Clin. Endocrinol. Metab. 84, 4410–4425. Richter-Unruh, A., Martens, J. W., Verhoef-Post, M., Wessels, H. T., Kors, W. A., Sinnecker, G. H., Boehmer, A., Drop, S. L., Toledo, S. P., Brunner, H. G., and Themmen, A. P. (2002). Leydig cell hypoplasia: Cases with new mutations, new polymorphisms and cases without mutations in the

209 luteinizing hormone receptor gene. Clin. Endocrinol. 56, 103–112. Tajima, T., Fujieda, K., Kouda, N., Nakae, J., and Miller, W. L. (2001). Heterozygous mutation in the cholesterol side chain cleavage enzyme (p450scc) gene in a patient with 46-XY sex reversal and adrenal insufficiency. J. Clin. Endocrinol. Metab. 86, 3820–3825.

Androgen Biosynthesis Inhibitors and Androgen Receptor Antagonists Stephen E. Borst, David T. Lowenthal, and George Zavros University of Florida and Malcom Randall VA Medical Center, Gainesville, Florida, United States

Glossary androgenetic alopecia Male-pattern baldness, requiring both a genetic predisposition and sufficient action of dihydrotestosterone. clinical flare Temporary, but serious, side effects resulting from administration of gonadotropin-releasing hormone agonists. Includes bone pain, nerve compression, and blockage of the ureters. gynecomastia Excessive development of male mammary glands. hirsutism The presence of excessive facial and body hair, especially in women. paraphilias Sexual practices that are socially prohibited.

receptor. Both testosterone and DHT bind to the androgen receptor in the cell, leading to the transcription of certain genes. The receptor’s affinity is four times greater for DHT than for testosterone.

Flutamide Flutamide is a nonsteroidal androgen receptor antagonist used in the management of metastatic prostate cancer and in the treatment of hirsutism in women. It is a pure androgen antagonist and produces no androgenic or other steroidal effects. It is metabolized in the liver to hydroxyflutamide, which is the active antiandrogen (see Fig. 1). It has a relatively short halflife of approximately 5.2 h. Side effects of flutamide treatment are gynecomastia, abnormal liver function, diarrhea, and gastrointestinal complaints. It is contraindicated in patients with severe hepatic impairment.

T

he androgens testosterone and dihydrotestosterone (DHT) play a central role in a number of disease states, including the progression of prostate cancer, benign prostatic hyperplasia, male pattern baldness, hirsutism, acne, and virilizing syndromes in women. In adults, most of the undesirable effects of androgens are mediated specifically by DHT. This article discusses the use of drugs that produce anti-androgenic effects by the following mechanisms of action: (1) androgen receptor antagonism, (2) inhibition of the conversion of testosterone to DHT, and (3) inhibition of testosterone synthesis.

ANDROGEN RECEPTOR ANTAGONISTS The androgen receptor, the gene of which is located near the centromere on the long arm of the X chromosome, was first described in 1969 and was cloned in 1988. It is present in highest concentrations in androgen target tissues such as the accessory organs of male reproduction. Tissues such as skeletal muscle, heart, and placenta have smaller amounts of androgen

210

Nilutamide Nilutamide is structurally related to flutamide and binds to the androgen receptor with an affinity similar to that of hydroxyflutamide. It is used in combination with surgical castration for the treatment of metastatic prostate cancer. Nilutamide has a longer half-life than flutamide, approximately 40 h, allowing once a day oral administration. The major side effects of nilutamide are diarrhea, impaired adaptation to darkness, alcohol intolerance, and the occasional serious side effect of acute interstitial pneumonitis. It is contraindicated in patients with severe respiratory insufficiency or severe hepatic impairment.

Casodex Casodex is another nonsteroidal androgen receptor inhibitor used in the management of prostate cancer. Casodex is indicated for use in combination therapy

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

211

Androgen Biosynthesis Inhibitors and Androgen Receptor Antagonists

CH3 NHCOCH

NO2 CF3

NHCOCOH

NO2 CF3

CH3

Flutamide

O

CH3 CH3

CH3

CH3 O

Hydroxylated active form of flutamide

Figure 1 Chemical structure of flutamide and its hydroxylated

O

O

active form.

SCCH3

Spironolactone

with a luteinizing hormone-releasing hormone analogue for the treatment of stage D2 metastatic carcinoma of the prostate. It is generally well tolerated and the most common side effects include gynecomastia and breast pain. It should be used with caution in patients with moderate to severe hepatic impairment as hepatotoxicity has been reported during the first 3 to 4 months of treatment.

Cimetidine Cimetidine was the first histamine-2 blocker introduced for general clinical use in the treatment of duodenal ulcers and other gastric hypersecretory conditions. It exerts anti-androgenic properties by binding to the androgen receptor, causing loss of libido, impotence, and gynecomastia. It has been used as an orphan drug for the treatment of androgenetic alopecia, hirsutism, and warts.

Cyproterone Cyproterone acetate, a potent androgen antagonist, has been used for the treatment of acne, male pattern baldness, and hirsutism in men and for virilizing syndromes in women as well as in the treatment of the paraphilias (see Fig. 2). Its usefulness is limited by the fact that it can cross the placenta and cause male pseudo-hermaphroditism in male embryos and it has been associated with severe liver damage. It is the most widely used anti-androgen internationally. It has orphan drug status in the United States.

CH3 C=O CH3 OAc CH2

O

CH3

Cl

Cyproterone acetate

Figure 2 Chemical structure of cyproterone acetate.

Figure 3 Chemical structure of spironolactone.

Spironolactone Spironolactone is a steroidal androgen receptor blocker that has been shown to be beneficial in the management of acne, hirsutism, and androgenetic alopecia (see Fig. 3). Spironolactone is a weak antiandrogen in blocking the androgen receptor but it also acts by inhibiting androgen biosynthesis. It is best known as an aldosterone receptor antagonist, it has been used traditionally as a potassium-sparing diuretic, and hyperkalemia is one of its side effects. Other side effects include menstrual irregularities, breast tenderness, fatigue, and headache.

Organophosphates Organophosphate insecticides are widely used in agricultural and residential settings. The organophosphate insecticide fenestration has structural similarity to flutamide, a known ant androgen, and has been identified as an androgen receptor antagonist.

ANDROGEN BIOSYNTHESIS INHIBITORS Androgen Biosynthesis Gonadotropin-releasing hormone (GnRH) is released from the hypothalamus in a pulsatile (on–off) manner, stimulating the anterior pituitary to produce luteinizing hormone (LH) and follicle-stimulating hormone (FSH). LH stimulates the Leydig cells of the testes to secrete testosterone. FSH stimulates the Sertoli cells of the testes, increasing spermatogenesis. Testosterone is the major circulating androgen and is converted to DHT in tissues expressing the 5a-reductase enzyme. Muscle and bone express very low levels of 5a-reductase and the anabolic actions of testosterone in these tissues are the direct result of binding of testosterone to androgen receptors. The prostate highly expresses 5a-reductase and the effects of testosterone in promoting prostate enlargement and

212

Androgen Biosynthesis Inhibitors and Androgen Receptor Antagonists

cancer are both mediated by DHT. Testosterone and DHT bind to the same androgen receptors, but DHT has greater potency. In addition, binding of DHT dimerizes the androgen receptor, causing additional effects. The following tissues express 5areductase and exhibit androgenic effects that are dependent on DHT: hair follicles (chest, scalp, and beard), liver, kidney, adrenals, seminal vesicles, prostate, testes, foreskin, and scrotum. Virilization of the external genitalia is dependent on the conversion of testosterone to DHT and a deficiency in 5a-reductase results in an incomplete form of male pseudohermaphroditism. However, in adults the effects of DHT are generally considered to be undesirable. The latter include increased body hair, acne, male-pattern baldness, and prostate enlargement. In contrast, the direct effects of testosterone are generally considered to be desirable. The latter include increased muscle and bone mass, deepening of the voice, increased libido and sense of well-being, spermatogenesis, penile and scrotal enlargement, and increased hematocrit (desirable in the absence of polycythemia). Blocking the effects of DHT, while preserving the direct effects of testosterone, presents an attractive therapeutic target.

Finasteride Finasteride is a selective inhibitor of 5a-reductase type 1 (see Fig. 4). Finasteride is effective for treating male-pattern baldness and is marketed for this

purpose as Propecia (1 mg/day). Finasteride is also effective for the treatment of benign prostatic hyperplasia (BPH) and is marketed for this purpose as Proscar (5 mg/day). Finasteride causes an approximate 20% reduction in prostate size and brings symptomatic relief of BPH within several months. In contrast, a-adrenergic receptor antagonists bring more rapid relief by decreasing urinary hesitation and spasm in the bladder neck. However, a-antagonists are less effective on a long-term basis, because they do not reduce or prevent an increase in the size of the prostate. Finasteride is less effective than other treatments for treating prostate cancer and is not indicated for this purpose. The ineffectiveness of finasteride in prostate cancer may be due to incomplete suppression of DHTor to elevation of testosterone within the prostate. Whereas 5a-reductase type 1 predominates in the prostate, 5a-reductase type 2 is also present and finasteride is a weak inhibitor of the latter enzyme. Finasteride reduces prostate tissue levels of DHT by approximately 70%, but the remaining DHT may be sufficient to produce a permissive effect in prostate cancer. Several compounds that are dual inhibitors of both type 1 and type 2 5areductase are under evaluation. These compounds cause a near-total suppression of prostate DHT levels. Whether they have greater efficacy than finasteride in treating BPH and prostate cancer remains to be determined. By blocking the conversion of testosterone to DHT, finasteride causes an increase in prostate testosterone levels. Although testosterone promotes

NHC(CH3)3 CH3 C

O

CH3 H O

OH

N H H

Finasteride

5 α-Reductase O

OH

Inhibition

O H

Testosterone

Dihydrotestosterone

Figure 4 Finasteride inhibits 5a-reductase and thus blocks the conversion of testosterone to DHT.

213

Androgen Biosynthesis Inhibitors and Androgen Receptor Antagonists

5-O-Pro-His-Trp-Ser-Tyr-Gly-Leu-Arg-Gly-Pro-Gly-NH2 Endogenous GnRH

5-O-Pro-His-Trp-Ser-Tyr-Gly-D-Leu-Arg-Gly-Pro-NHCH2CH3 Leuprolide (Leupron)

and producing a hypogonadal state in both men and women. The GnRH agonist naferelin (Synarel) is administered as a nasal spray for the treatment of endometriosis. Suppression of ovarian steroidogenesis and clinical effectiveness are similar to those of danazol.

Figure 5 Peptide sequence of endogenous GnRH and that of a GnRH analogue, leuprolide (Leupron), with a d-amino acid substituted at the 6 position and with ethylamide substituted at the 10 position.

prostate cancer to a much lesser degree than does DHT, 5a-reductase inhibitors may elevate prostate testosterone levels sufficiently to promote or to maintain a permissive effect on prostate cancer.

GnRH Agonists GnRH is a decapeptide secreted by the hypothalamus in a pulsatile manner, causing the anterior pituitary to release LH and FSH. Endogenous GnRH has a short half-life. Analogues with a d-amino acid substituted at the 6 position and with ethylamide substituted at the 10 position are GnRH agonists with greater potency and longer duration of action than endogenous GnRH (see Fig. 5). Leuprolide (Leupron) and goserelin (Zoladex) are GnRH agonists that have been used successfully to treat prostate cancer. Leupron is formulated in microspheres and Zoladex is formulated in a polymer matrix. In both cases, the slow release of the drug from an intramuscular injection site suppresses androgens for several months. GnRH agonists continuously stimulate pituitary GnRH receptors, as opposed to endogenous GnRH, which stimulates them intermittently. Continuous stimulation of GnRH receptors produces a transient ‘‘clinical flare reaction,’’ which is caused by a surge of androgen production and which lasts 5 to 12 days. Clinical flare occurs in approximately 10% of patients with prostate cancer who are treated with GnRH agonists. The condition is dangerous and is associated with bone pain, nerve compression, and blockage of the ureters. These symptoms may be controlled with androgen receptor antagonists, such as flutamide. However, androgen antagonists also produce serious side effects. Eventually, continuous stimulation causes GnRH receptors to be down-regulated, resulting in a marked and long-lasting inhibition of LH and FSH secretion

GnRH Antagonists In an effort to avoid the transient androgen surge associated with GnRH agonists, a number of GnRH antagonists have been developed. Abarelix is the first to reach clinical trials. One study has shown that aberelix causes a near-complete suppression of serum testosterone, without a transient increase. Evaluation of the effectiveness of aberelix in treating prostate cancer is under way.

See Also the Following Articles Androgen Biosynthesis and Gene Defects . Androgen Insensitivity Syndrome . Gonadotropin-Releasing Hormone (GnRH) Actions . Gynecomastia . Sexual Function and Androgens

Further Reading Bartsch, G., Rittmaster, R. S., and Klocker, D. (2000). Dihydrotestosterone and the concept of 5a-reductase inhibition in human benign hyperplasia. Eur. Urol. 37, 367–380. Chrousos, G. P., Zoumakis, E., and Gravanis, A. (2001). In ‘‘Basic and Clinical Pharmacology,’’ (Katzung, B. G., ed.), 8th ed., pp. 634–636 and 704–706. Lange Medical Books/McGraw-Hill, New York. Cook, T., and Sheridan, W. P. (2000). Development of GnRH antagonists for prostate cancer: New approaches to treatment. Oncologist 5, 162–168. Curtis, L. C. (2001). Organophosphate antagonism of the androgen receptor. Toxicol. Sci. 60, 1–2. Grumbach, M. M., and Conte, F. A. (1998). In ‘‘Williams’ Textbook of Endocrinology,’’ (Wilson, J. D. and Foster, D. W., eds.), 9th ed., p. 859. W. B. Saunders, Philadelphia, PA. Mandell, J. (1998). Sexual differentiation: Normal and abnormal. In ‘‘Campbell’s Urology,’’ (Walsh, P. C., et al., eds.), 7th ed., p. 2147. W. B. Saunders, Philadelphia, PA. McLeod, D. G. (1993). Antiandrogenic drugs. Cancer 71, 1046–1049. Miller, J. W. (2000). In ‘‘Melmon and Morrelli’s Clinical Pharmacology,’’ (Carruthers, S. G., Hoffman, B. B., Melmon, K. L., Nierenberg, D. W., eds.), 4th ed., pp. 630–640. McGraw-Hill, New York. Taplin, M. E., and Ho, S. M. (2001). The endocrinology of prostate cancer. J. Clin. Endocrinol. Metab. 86, 3467–3477.

Androgen Insensitivity Syndrome Yuan-Shan Zhu and Julianne Imperato-McGinley Weill Medical College of Cornell University, New York, New York, United States

Glossary androgen receptor An intracellular protein that specifically binds to androgens and mediates androgen action. g0010

coregulator A protein that interacts with nuclear receptors to enhance (coactivators) or reduce transactivation (corepressors) of target genes but does not significantly affect the basal transcription rate. dihydrotestosterone A derivative of testosterone formed by the conversion of testosterone in the action of 5a-reductase isozymes. testosterone A male hormone that is produced by the testes and induces and maintains male secondary sexual characteristics.

A

ndrogen receptor (AR), a member of the nuclear receptor superfamily and a ligand-dependent nuclear transcription factor, mediates the action of androgens. AR mutations that inactivate its function result in androgen insensitivity, classified as either complete androgen insensitivity syndrome (CAIS) or partial androgen insensitivity syndrome (PAIS). Affected 46,XY individuals with CAIS have female external genitalia, normal female breast development, absent or scanty axillary and public hair, and absent female internal genitalia. Testes are present, with levels of plasma testosterone, estradiol, and luteinizing hormone that are high-normal or elevated relative to those of normal males. The plasma dihydrotestosterone (DHT) level is normal or low-normal, which can sometimes result in an elevated testosterone:DHT ratio. Individuals with PAIS have a wide spectrum of phenotypic features, ranging from decreased body hair, infertility, and/or gynecomastia to severe ambiguity of the genitalia. Studies of androgen insensitivity highlight the importance of androgen–AR function in male sexual differentiation and physiology.

MALE SEXUAL DETERMINATION AND DIFFERENTIATION Male sexual determination and differentiation is a complex process involving multiple steps, including

214

testes formation (sex determination) under the control of the SRY (sex-determining region of the Y chromosome) gene on the short arm of the Y chromosome. Other autosomal and X-chromosomal genes are also known to be involved in gonadal or testes development. Before 6 weeks of gestation, human embryos with either a 46,XY (genetic male) or 46,XX (genetic female) karyotype develop identically, and an undifferentiated gonad is present in both genetic male and female fetuses. At approximately 6 or 7 weeks of gestation, testicular cords evolve from the primary sex cords of the indifferent gonad. The Sertoli cells within the cords enlarge, become contiguous, and engulf the germ cells. The seminiferous cords interconnect to form a network of solid cords, which connect with the mesonephric tubules and ultimately to the ductuli efferentes. Leydig cells are apparent by 8 weeks of fetal life and completely fill the interstitial spaces of the developing testes at 3 months of gestation. Two major hormones, testosterone and antiMu¨llerian hormone, are synthesized in the testes and involved in the translation of gonadal sex to phenotypic sex. Anti-Mu¨llerian hormone (also called Mu¨llerianinhibiting substance or Mu¨llerian-inhibiting factor), a glycoprotein and a member of the transforming growth factor-b family, secreted from the Sertoli cells of the testes, promotes regression of the Mu¨llerian ducts, resulting in lack of development of female internal structures (uterus, fallopian tubes, and upper vagina). The Mu¨llerian ducts, which appear at 40–48 days of gestation, regress at approximately 81⁄2 weeks of gestation in the male fetus. Testosterone secreted by the Leydig cells of the testes beginning at 8 weeks of gestation mediates the differentiation of the Wolffian ducts to the epididymides, vasa deferentia, and seminal vesicles, or the internal masculinization. The process of Wolffian duct differentiation is mediated by testosterone, probably via a paracrine action, and completed by 12 weeks of gestation.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

215

Androgen Insensitivity Syndrome

Male external genital differentiation begins soon after Wolffian ductal differentiation. The development of the prostate is dependent on the local conversion of testosterone to the more potent androgen dihydrotestosterone (DHT) by the 5a-reductase-2 isozyme present in these tissues. The urogenital tubercle becomes the glans penis, the urogenital folds become the shaft of the penis, and the urogenital swellings become the scrotum. The urogenital sinus forms the prostate, bulbourethral glands, and the prostatic and membranous portion of the urethra. The entire process of male external sexual differentiation is completed by 14–16 weeks of gestation. Descent of the testes and growth of genitalia occur during the last two trimesters of pregnancy. The actions of both testosterone and DHT on internal and external genital masculinization are mediated via androgen receptor (AR). Thus, a functional AR is required for normal sexual differentiation. Any defect in the production or action of androgens during these critical periods can result in disorders in male sexual differentiation.

s9000

p0050

COMPLETE ANDROGEN INSENSITIVITY SYNDROME Clinical Syndrome Androgen unresponsiveness in utero causes a 46,XY fetus with testes and normal androgen secretion to be born with female genitalia and an absent or severely hypoplastic Wolffian ductal system. The labia, especially the labia minora, may be underdeveloped. The clitoris is normal or small. The vagina ends blindly. Due to normal secretion of anti-Mu¨llerian hormone by the testes in utero, Mu¨llerian-derived structures are absent or rudimentary, and thus the uterus and cervix are absent or rudimentary. During puberty, there is normal or augmented breast development due to the unopposed estrogenic action by androgens. Pubic and axillary hair is scant or absent. The testes of patients with complete androgen insensitivity syndrome (CAIS) are usually located in the abdomen or inguinal canal. They cannot be distinguished histologically from those of normal males before puberty. However, postpubertal histologic studies reveal immature tubular development with Sertoli cells, spermatogenia, and no spermatogenesis. There is frequent clumping of tubules with formation of tubular adenomas. Leydig cells are hyperplastic and electron microscopy reveals ample smooth endoplasmic reticulum and mitochondria with tubular cristae.

This correlates well with the usually elevated plasma testosterone levels, although in some respects Leydig cells have been reported to resemble fetal Leydig cells with absent crystals of Reinke.

Biochemical Characteristics In postpubertal individuals with CAIS, plasma luteinizing hormone (LH) is frequently increased with high-normal or elevated testosterone and correlates well with the histological findings of Leydig cell hyperplasia. The elevation of LH apparently results from androgen unresponsiveness in the hypothalamus and/or pituitary. However, LH levels are not in the castrate range due to the negative feedback effect of estrogen on the hypothalamus and/or pituitary. Follicle-stimulating hormone is normal or elevated. Although the plasma testosterone level is in highnormal or elevated, plasma DHT levels may be lownormal, resulting in an elevated testosterone:DHT ratio. This may be due to a secondary deficiency of 5a-reductase activity since DHT exhibits positive feedforward control of 5a-reductase activity. Plasma and urinary estrogens range from high male to low female levels. The estrogens originate mainly from the testes and, to a lesser extent, peripheral aromatization of androstenedione and testosterone by aromatase. An unopposed estrogen effect, due to increased estrogen and androgen unresponsiveness, is the likely explanation for breast development during puberty. The sex hormone binding globulin (SHBG) levels in individuals with CAIS are higher than those of normal males and similar to those of normal females. Castrated CAIS patients not receiving estrogen have SHBG levels similar to those found in normal males.

Molecular Biology of the Androgen Receptor and Genetic Basis of Androgen Insensitivity Syndrome The AR, a member of the nuclear steroid receptor superfamily and a ligand-dependent nuclear transcription factor, was cloned in 1988. It has approximately 910–919 amino acids encoded by the AR gene located on Xq11–12 (Fig. 1). The AR gene is a single-copy X-chromosomal gene that spans approximately 90 kilobases of genomic DNA. The encoding region of the AR gene comprises eight exons separated by seven introns. Like other steroid receptors, the AR is a single polypeptide composed of relatively distinct domains: an amino-terminal domain, a DNA-binding

216

Androgen Insensitivity Syndrome

Figure 1 AR gene location, gene structure, and protein domains. (Top) The location of the AR gene at the q11–12 of the X chromosome. (Middle) The AR gene consists of eight exons (boxes) and seven introns (dashed line). (Bottom) The AR complementary DNA and AR protein, for which the AR domains and the exons encoding each domain are illustrated. Relative positions of glutamine (Gln), proline (Pro), and glycine (Gly) repeats within the N-terminal domain are shown. The transactivation functions AF-1 and AF-2 are located within the N-terminal domain and ligand-binding domain, respectively.

domain, a hinge region, and a steroid-binding domain in the carboxyl terminal. The large amino-terminal domain that comprises approximately half of the AR molecule is encoded by exon 1. It is involved in the transcriptional activation of target genes and contains a transactivation domain, activation function 1 (AF-1). This domain plays a role in AR functions by intramolecular and/or intermolecular interaction with other factors. There are three highly polymorphic direct repeats of amino acid residues: one each containing glutamine, proline, and glycine residues. The increase in size of the glutamine homopolymeric segment is related to the pathogenesis of the spinal and bulbar muscular atrophy (Kennedy’s disease). The DNA-binding domain encoded by exons 2 and 3 contains two ‘‘zinc finger’’ motifs that are hallmarks of all nuclear steroid receptors, and it is the most highly conserved region among steroid receptors. The formation of the two zinc fingers involves eight cysteine residues. These two zinc-coordinated stem-loop structures are responsible for the specific interaction with the cognate DNA of target genes by interacting with the major groove of the DNA duplex. The first zinc finger (proximal to the N-terminal domain) is associated with the determination of the sequence specificity of DNA binding, whereas the second finger helps stabilize the DNA–receptor complex. Despite their exquisite functional specificity in the physiological context, receptors for androgens, glucocorticoids, progesterone, and mineralocorticoids can recognize the same DNA response element both in an in vitro binding assay and in functional analysis

using transiently transfection analysis. This paradox remains to be solved. The carboxyl terminal of the AR is the ligand-binding domain, encoded by the 30 portion of exon 4 and exons 5–8, and is responsible for the specific highaffinity ligand binding. Studies indicate that androgens interact with the ligand-binding domain mostly through hydrophobic interaction as well as some hydrogen bonding. The carboxyl terminal also contains subdomains involved in dimerization and transcriptional activation. The second transactivation function (AF-2) domain of AR resides within the ligand-binding domain. Upon ligand binding, the AF-2 domain can interact with coregulators, such as coactivators, to affect AR function. Between the DNAbinding domain and the steroid-binding domain is the hinge region, which contains the nuclear translocation signal. Both testosterone and dihydrotestosterone, potent natural androgens, bind to the same AR at the ligandbinding domain to regulate androgen target gene expression. The binding of androgen on the AR results in an AR conformational change that promotes the dissociation of chaperone proteins and facilitates receptor dimerization, nuclear transportation, phosphorylation, and DNA binding. Upon the recruitment of coregulators and general transcription factors, the transcription of a target gene is either induced or inhibited and ultimately leads to a change in androgen target proteins and cellular or biological structures and functions (Fig. 2). A number of coregulators have been identified that can interact with AR to either enhance or reduce androgen–AR action on target

217

Androgen Insensitivity Syndrome

Cytoplasm Testosterone hsp

Phosphorylation P AR

hsp

hsp

AR

Nucleus P AR

hsp

TFs T DHT 5α-Reductase

Male sexual differentiation Male secondary sexual features Prostate differentiation and development Prostate hypertrophy and cancer f0010

P

CoR

AR

RNA polymerase P

ARA

AR

ARE

GTFs TATA

Genes

Transcription

New proteins

mRNAs

Figure 2 The molecular events of androgen–AR action in a target cell. When testosterone enters the cell, it can be converted to DHT by the action of 5a-reductase. Both testosterone (T) and DHT bind to the same AR. The androgen–AR complex translocates to the nucleus, interacting with other factors and target genes to regulate gene transcription. The changes in androgen–target proteins in the cell eventually affect cellular structure and function related to male sexual differentiation, physiology, and pathophysiology. ARE, androgen response element; GTFs, general transcription factors; ARA, androgen receptor-associated proteins; CoR, coregulator; TFs, transcription factors; hsp, heat shock protein. gene transcription. Although significant progress has been made in the past decade in understanding androgen–AR action, the detailed process from androgen binding on AR to the alteration of target gene transcription remains to be elucidated. The mutations of the AR that cause loss of function result in androgen insensitivity, which is inherited as an X-linked recessive condition, with genetic males expressing the condition (Fig. 3). To date, more than 490 different mutations in all eight exons in the AR gene have been reported, including the mutation for the largest pedigree of CAIS (see the AR gene mutation database at www.mcgill.ca/androgendb). These mutations range from a single point mutation to an entire gene deletion and can result in various AR dysfunction, including impaired androgen binding, impaired DNA binding, impaired cofactor interaction, blocked formation of a functional receptor (deletions, nonsense mutations, splice-junction alterations), decreased AR expression, and an unstable androgen–receptor complex. Depending on the type of dysfunction, various degrees of functional impairment of the AR occur, resulting in a wide spectrum of symptoms, ranging from a total female phenotype to

normal male phenotype with decreased secondary sexual hair, infertility, or gynecomastia. Although various individual mutations have been characterized, no correlation between the severity of the syndrome and a particular gene defect has been observed. A variety of genetic defects impair the normal functioning of the AR. Generally, mutations that delete the entire AR gene or interrupt the AR open reading frame, blocking the formation of a functional receptor resulting from premature termination, aberrant splicing, or deletion of partial or complete exon segments, are associated with a phenotype of complete androgen insensitivity. This is due to the fact that AR DNA- and hormone-binding domains, critical for AR function, are located at the carboxyl terminus of the AR protein. As a consequence, defects that truncate the receptor protein at any point during its synthesis result in the removal of a portion of one or both of these important functional domains and will lead to a complete loss of AR function. Mutations in the ligand-binding domain represent approximately 67% of AR mutations related to androgen insensitivity. The majority of these mutations are single nucleotide substitutions and cause defects in

218

Androgen Insensitivity Syndrome

Figure 3 Pedigree of a kindred with complete androgen insensitivity indicates the maternal transmission of this disease. s, normal females; N, normal males; Z, androgen insensitive subjects.

androgen binding that have been categorized as receptor-negative (the absence of detectable ligand binding), receptor-positive (quantitatively normal but qualitatively abnormal ligand binding), and receptor-reduced or receptor-deficient (reduced capacity or reduced affinity of ligand binding) defects, as demonstrated in patient genital skin fibroblasts. Substitution mutations in the AR ligand-binding domain have been identified in patients with the entire spectrum of androgen-insensitive phenotypes. Approximately 15% of AR mutations are located in the DNA-binding domain, resulting in either complete or partial androgen insensitivity. Studies of these mutations have shown that the mutant receptors bind androgens with normal or near-normal affinity but fail to bind normally to the target DNA sequences by in vitro DNA-binding assays, resulting in a defect in androgen–AR function as demonstrated by in vitro transfection analyses. It has been reported that some AR mutations result in decreased AR protein levels and decreased androgenbinding capacity in the genital skin fibroblasts of affected individuals. These mutant ARs may have subtle differences in function on selected target genes when analyzed by in vitro transfection. However, the levels of AR in these individuals are significantly decreased due to the mutations that alter AR gene transcription, translation, or posttranslational processes. AR mutations that alter the stability of the androgen–receptor complex can also cause androgen insensitivity. It has been shown that replacement of arginine 774 of AR with cysteine residue leads to androgen resistance and undetectable androgen binding in genital skin fibroblasts. However, substitution of

the same residue by histidine leads to normal levels of androgen binding in fibroblasts that display a marked thermal liability in in vitro assays, suggesting that the stability of the androgen–receptor complex is important for normal AR activity. A new type of defect has been reported at the postreceptor level in an individual with CAIS as diagnosed based on clinical and biochemical features. The AR gene in this patient is normal. However, the transmission of the activation signal from the AF-1 domain of the AR is disrupted, probably due to a defect in an AR coregulator or coactivator, indicating that defects in AR signal transmission to target gene expression can also result in androgen insensitivity despite the absence of AR mutation.

Diagnosis and Management Since the phenotypic appearance of CAIS is totally feminine, the condition is usually diagnosed following breast development at puberty when patients seek medical advice for primary amenorrhea. Complete androgen insensitivity is the most likely diagnosis in a phenotypic female who presents with primary amenorrhea, breast development, scanty or absent pubic and axillary hair, a short vagina, and absent cervix and uterus. These subjects also have a clear, smooth, and acne-free complexion. Occasionally, individuals with CAIS are diagnosed before or soon after birth. This diagnostic evaluation results from the discrepancy between the findings of a 46,XY karyotype on amniocentesis and the presence of a female phenotype on prenatal ultrasound examination or at birth. Some individuals with CAIS are

219

Androgen Insensitivity Syndrome

p0165

diagnosed during infancy when they develop inguinal herniae, often bilaterally. Occasionally, a 46,XY individual with 17a-hydroxylase deficiency can also have the same phenotypic presentation at puberty. 5a-Reductase-deficient individuals are most frequently born with severe ambiguity of the external genitalia (pseudovaginal and perineoscrotal hypospadias) and can occasionally present with female genitalia. However, puberty is significantly different from that in androgen-insensitive patients because distinct virilization without gynecomastia occurs. Urinary 5b/5a C21 steroid metabolite ratios are markedly elevated in 5a-reductase-2-deficient patients, whereas these are only slightly changed or normal in patients with complete androgen insensitivity. In summary, the diagnosis of CAIS can usually be made based on the clinical presentation and the biochemical features. Genetic analysis with identification of an AR mutation and functional demonstration of the mutation can assist and confirm the diagnosis. Once diagnosed, the management of patients with CAIS depends on whether the individuals are prepubertal or postpubertal. CAIS subjects are phenotypic and psychosexual females. It has been reported that up to 25% of patients with CAIS develop testicular neoplasms, although seldom before 25 years of age. If a diagnosis is made before puberty, the patient is usually followed until puberty is completed and breasts have developed naturally before being referred to a surgeon for gonadectomy. However, the optimal time for gonadectomy with regard to psychological adjustment is still a matter of debate. It should be realized that waiting for the postpubertal period to perform the gonadectomy means that it will often be done during the tumultuous years of adolescence or soon thereafter. Gonadectomy performed at this time, however, has the advantage of involving adolescents or young adults in decision making. It can be argued that gonadectomy should be performed in childhood, thus avoiding the difficulties of yet another psychologic adjustment during adolescence. However, some adult patients with this condition, who were gonadectomized in childhood, believe that gonadectomy should have been delayed until early adulthood, when they could have been more actively involved in decision making; thus making the psychologic adjustment easier. If a child has an inguinal hernia and is brought to a physician for this reason and the diagnosis is made, consideration should be given to perform gonadectomy at the time of hernia repair, thereby avoiding a second operation. It is important that the patient learn about the condition from his or her doctor in a careful and

sensitive way. This will avoid severe psychologic problems. Therefore, it is critical that the doctor develop a good relationship with the patient and the parents. The timing of the information from initiation of the topic to a detailed discussion must be individualized. Psychologic counseling is needed for the patient and the family. The postpubescent patient should be gonadectomized due to the probability of testicular neoplasms. Cyclic estrogen replacement therapy should be prescribed at the appropriate time to prevent osteoporosis and to maintain breast development, and patients should be followed regularly. For most patients, the vagina is of sufficient size to allow normal coital function. In a patient whose vagina is too small, vaginal dilation is the usual corrective measure, and vaginal surgery is rarely required.

PARTIAL (INCOMPLETE) ANDROGEN INSENSITIVITY SYNDROME In partial or incomplete forms of androgen insensitivity, a spectrum of clinical phenotypes is present and includes gynecomastia and severely ambiguous genitalia, mild hypospadias, gynecomastia only, normal male genitalia with infertility, and decreased body hair in adulthood. A range of phenotypic abnormalities have been reported in 11 members of one family, indicating that a single mutant gene, variably expressed, may be a factor in the variant phenotypic forms of partial androgen insensitivity syndrome (PAIS). The endocrine profile, analogous to that in patients with CAIS, shows elevated LH and testosterone levels. Total 17b-estradiol produced and secreted by the testes has been reported to be greater than in patients with CAIS. However, the degree of feminization during puberty is not as prominent. This may indicate a less severe androgen and estrogen imbalance at the cellular level due to some androgen responsiveness. Genetic analysis has revealed a variety of AR mutations responsible for PAIS. These mutations occur throughout the entire encoding region and result in various AR dysfunctions. There is no correlation between a specific AR gene mutation and phenotypic expression, although numerous mutations have been identified and characterized. Treatment of patients with PAIS is dependent on the degree of virilization. In most cases, patients with profound genital ambiguity should be raised as females since deficient masculinzation and gynecomastia will undoubtedly occur at puberty. Surgical correction of external genitalia and gonadectomy are indicated and female sex hormone therapy should be prescribed at

220 puberty. Surgical correction of mild hypospadias is necessary in patients raised as males, and surgical correction of gynecomastia may be necessary. Supplemental high-dose androgen therapy to increase virilization is controversial due to its possible deleterious effect on the cardiovascular system. However, due to the fact that some AR mutations display conditional androgen-binding defects, the use or development of specific androgen analogs that can overcome the androgen-binding defects manifest in natural androgens is a therapeutic strategy in the management of PAIS that remains to be investigated.

SUMMARY Androgen insensitivity syndrome due to AR mutations is a natural model for elucidating androgen actions in male sexual development, physiology, and pathophysiology. The identification and characterization of various AR mutations provide important information for understanding AR structure and function. The elucidation of AR–coregulator interaction in androgen action and the identification of an AR–coregulator defect in androgen insensitivity further indicate the complexity of androgen–AR action and open the door to a new therapeutic strategy.

Acknowledgments This study was supported in part by National Institutes of Health Grant M01-RR-00047 (General Clinical Research Center) and a Merck Foundation fellowship.

See Also the Following Articles Adrenal Androgens . Androgen Biosynthesis and Gene Defects . Androgen Biosynthesis Inhibitors and Androgen

Androgen Insensitivity Syndrome

Receptor Antagonists . Androgens, Gender and Brain Differentiation . Anti-Mu¨llerian Hormone . Endocrine Disrupters and Male Sexual Differentiation . Gender Assignment and Psychosocial Management . Genes and Gene Defects Affecting Gonadal Development and Sex Determination . Gynecomastia . Testes, Embryology of

Further Reading Adachi, M., Takayanagi, R., Tomura, A., Imasaki, K., Kato, S., Goto, K., Yanase, T., Ikuyama, S., and Nawata, H. (2000). Androgen-insensitivity syndrome as a possible coactivator disease. N. Engl. J. Med. 343(12), 856–862. Imperato-McGinley, J., and Canovatchel, W. J. (1992). Complete androgen insensitivity—Pathophysiology, diagnosis and management. Trends Endocrinol. Metab. 3(3), 75–81. Imperato-McGinley, J., and Zhu, Y. S. (2002). Gender and behavior in subjects with genetic defects in male sexual differentiation. In ‘‘Hormones, Brain and Behavior’’ (D. W. Pfaff et al., eds.), Vol. 5, pp. 303–345. Academic Press, San Diego. Imperato-McGinley, J., Peterson, R. E., Gautier, T., Cooper, G., Danner, R., Arthur, A., Morris, P. L., Sweeney, W. J., and Shackleton, C. H. L. (1982). Hormonal evaluation of a large kindred with complete androgen insensitivity: Evidence for secondary 5a-reductase deficiency. J. Clin. Endocrinol. Metab. 54(5), 931–941. Patterson, M. N., McPhaul, M. J., and Hughes, I. A. (1994). Androgen insensitivity syndrome. Baillieres Clin. Endocrinol. Metab. 8(2), 379–404. Quigley, C. A., DeBellis, A., Marschke, K. B., El-Awady, M. K., Wilson, E. M., and French, F. S. (1995). Androgen receptor defects: Historical, clinical, and molecular perspectives. Endocrine Rev. 16(3), 271–321. Sultan, C., Lumbroso, S., Paris, F., Jeandel, C., Terouanne, B., Belon, C., Audran, F., Poujol, N., Georget, V., Gobinet, J., Jalaguier, S., Auzou, G., and Nicolas, J. C. (2002). Disorders of androgen action. Semin. Reprod. Med. 20(3), 217–228. Zhu, Y. S., Cai, L. Q., Cordero, J. J., Canovatchel, W. J., Katz, M. D., and Imperato-McGinley, J. (1999). A novel mutation in the CAG triplet region of exon 1 of androgen receptor gene causes complete androgen insensitivity syndrome in a large kindred. J. Clin. Endocrinol. Metab. 84, 1590–1594.

Androgens, Gender and Brain Differentiation Julianne Imperato-McGinley and Yuan-Shan Zhu Weill Medical College of Cornell University, New York, New York, United States

Glossary androgen receptor An intracellular protein that binds to androgens and mediates androgen action. dihydrotestosterone A derivative of testosterone formed by the action of 5a-reductase isozymes that is involved in male sexual differentiation and development. 5a-reductase isozymes Microsomal nicotinamide adenine dinucleotide phosphate (NADPH)-dependent proteins that reduce the double bond at the 4–5 position of a variety of C19 and C21 steroids such as testosterone. gender identity The sense of being male or female; the self-awareness of knowing one’s sex. gender role The expression of one’s gender identity to the public; manifested by one’s behavior or actions. male pseudohemaphrodite A 46,XY individual who has the testicular tissue but whose external genitalia and other sexual characteristics are so ambiguously developed that the sex of the individual is uncertain. testosterone A male hormone that is produced by the testes and that induces and maintains male secondary sexual characteristics.

G

enetic and hormonal influences on phenotypic sex determination in humans have been shown to be similar to those in other mammals. Many animal studies have also demonstrated that hormones are essential for sexual differentiation of the brain during development and for the maintenance of sexually dimorphic behavior throughout life. However, the effect of hormonal influences on sexual dimorphic differences in the human nervous system and on gender identity and sex differences in human behavior is still an emerging field. This article discusses the importance of androgens in (1) determination of male gender identity and (2) cognitive function by reviewing specific inherited genetic defects in androgen biosynthesis and action. The complex interaction of nature versus nurture is also addressed.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

ROLE OF ANDROGENS IN BRAIN MORPHOLOGY Studies of brain morphology of various animals demonstrate the presence of sexual dimorphism. Androgens administered perinatally to female rats result in structural changes in the sexually dimorphic nucleus of the preoptic area of the brain, making its similar to that of male rats. In canaries (Serinus canarius) and zebra finches (Poephila guttata), areas in the brain that control the vocal cords are noticeably larger in males than in females. The area X of the lobus parolfactorius is well developed in males of both species but is hardly identifiable in females. These size differences correlate with differences in singing ability; males produce a complex song, whereas females do not normally sing at all (zebra finches) or sing an infrequent simple song (canaries). The influence of testosterone at a critical period in development induces enhancement of these song areas of the brain in female canaries and zebra finches. The action of testosterone on the song areas of the brain may be mediated by conversion to estradiol given that it has been shown that masculine patterns of song area development in genetic females can be induced with estrogen. However, feminine patterns cannot be reproduced in males with antiestrogens or inhibitors of estrogen synthesis. An unresolved issue involves studies of genetic females that have functional testicular tissue and virtually no ovarian tissue but that still have feminine song circuitry. Morphological sex differences that can be induced by androgens are also present in the auditory system of the tree frog (Eleuthero dactylus coqui), the spinal cord of the cat, the brain of the juvenile macague monkey, and the neurons innervating the bulbocavernosus and levator ani muscles as well as the neurons innervating the ischiocavernosus and levator ani muscles (dorsolateral motor nucleus [DLN]) of the rat. Studies of the human brain have found that the volume of the nuclei in the preoptic area is 2.5
0.6

221

p0010

222 times larger in men, containing approximately twice as many cells. One area (INH-3) is larger by a factor of three in the male compared with the female. Another cell group is twice as large in the male brain, varies in women according to hormone levels, and is 3.7-fold larger in women of childbearing age. However, it is still not clear whether these nuclei are subdivisions of the paraventricular nucleus or are separate anatomical entities. Sex differences in areas such as the left cortical language regions, as well as in the shape and surface area of the human callosum, have been reported in some studies but not in others. Because the callosum consists of myelinated connecting fibers, larger callosal volumes in women are interpreted as providing better interhemispheric communication, resulting in less functional specialization of the two hemispheres.

ROLE OF ANDROGENS IN MALE BEHAVIOR In animal studies, androgens administered early in females have been shown to organize neural systems in such a way as to stimulate or induce male sexual response during adulthood and to inhibit female sexual response. The critical period for this central nervous system effect of androgens might not coincide with the critical period for external genital differentiation. In some animals, pre- or perinatal androgen stimulation also affects nonsexual masculine patterns of behavior that are independent of hormonal exposure during adulthood. Comparable effects have been seen in animal models such as the guinea pig, rabbit, hamster, rat, dog, and rhesus monkey. The critical period of exposure (pre- and/or perinatal) differs among the various species. In offspring of pregnant rhesus monkeys, treatment with androgens during a specific period of gestation increases mounting and ‘‘rough and tumble’’ play in females. Thus, sexually dimorphic behavior in animals is secondary to sex steroid-induced differentiation and activation of the brain at critical periods.

GENDER IDENTITY DEVELOPMENT The relative influence of hormonal versus environmental factors in the determination of gender identity has been an ongoing debate for many years.

Environmental Influence In 1955, Money proposed that human sexuality was undifferentiated at birth and becomes differentiated as

Androgens, Gender and Brain Differentiation

a consequence of various experiences. This theory was revised later to acknowledge that male sexually dimorphic behavior is expressed at birth but that it can be incorporated into either a male or female gender identity pattern. This theory was tested by matching individuals with ambiguous genitalia that were ‘‘chromosomally, gonadally, and otherwise diagnostically the same.’’ The ‘‘matched pairs’’ were reported as differing only in their sex assignment and sex of rearing. The results demonstrated that gender identity correlated with sex of rearing, and not with chromosomal or gonadal sex, leading to the conclusion that sex of rearing was predominant in establishing gender identity. However, at the time of the studies, adequate hormone evaluation was unavailable and individuals were not assessed for their hormonal environment. Therefore, the degree of androgen exposure was assumed to be similar for the phenotypically matched pairs but was not objectively known to be so. Consequently, the issue of nature (androgen imprinting) versus nurture (sex of rearing) was not resolved due to incomplete hormonal characterization of the individuals. Individuals with inadequate testosterone production or action are not suitable models to assess the relative importance of environment (sex of rearing) versus nature (androgen imprinting) in determining male gender identity because deficient androgen exposure would result in a decrease in the androgen effect on the brain. In addition, when castration and sex hormone therapy are initiated in an individual born with ambiguous genitalia to coincide with the determined sex of rearing, the natural hormonal sequence of events in the evolution of gender is interrupted. As a consequence, no valid conclusions can be drawn about the relative importance of these factors.

Hormonal Influence Gender Identity in Research Subjects with 5a-Reductase-2 Deficiency In research subjects with 5a-reductase-2 deficiency, the unique biochemical defect affects male external genital development without altering testosterone response. In these individuals, the biosynthesis of testosterone and its peripheral action are normal; thus, prenatal and neonatal testosterone exposure of the brain proceeds as in the normal male. However, deficient 5a-reductase-2 enzyme activity impairs conversion of testosterone to dihydrotestosterone (DHT), and the deficiency in utero results in genital ambiguity. As a result, many affected individuals with

Androgens, Gender and Brain Differentiation

p0055

p0060

this condition are believed to be female at birth and are raised as girls. However, at puberty, significant virilization occurs under the influence of normal to elevated plasma testosterone levels and a gender change occurs. Thus, 5a-reductase-2-deficient individuals provide a unique opportunity to evaluate the relative influences of nature (testosterone) versus environment (sex of rearing) in the determination of gender identity in humans. In the affected subjects from the Dominican kindred, those who were raised as females began to realize that they were different from other girls in the village because they did not develop breasts and/or they felt masses in either the inguinal canal or the ‘‘labia.’’ They passed through a number of stages that included no longer feeling like girls, feeling like men, and finally identifying as men. When they became convinced of their maleness, a change in gender role occurred either during puberty or during the postpubertal period. In some individuals, the gender role change to male was delayed until they were confident of their ability to defend themselves. The average age of the gender role change was 16 years. With one exception, those who were raised as girls changed to a male gender role and perform male work in a society where there is a definite division of labor; women perform household activities, whereas men work as farmers, miners, or woodsmen. Females are the affected males’ choice for sexual intimacy. The adequacy of sexual intercourse depends on the severity of the chordee and the size of the phallus. With one exception, all of the subjects who changed to a male gender role either live or have lived with women in common law marriages. Some chose women with children from previous unions. One man lives alone in the hills working as a farmer since adopting a male gender role at 20 years of age. The social and psychosexual development of New Guinean research subjects of the Sambian tribe with 5a-reductase-2 deficiency has been recorded in field studies of Carlton Gajdusek over decades with similar observations. The Sambian tribal culture of the eastern highlands is rigidly gender segregated. Women are caretakers, whereas men are hunters and warriors. Male initiation rites during puberty include ritualized fellatio and other rituals for men of premarital age. Gender segregation during pubertal initiation was allegedly strictly enforced and included killing any female who accidentally witnessed these events. In the past, New Guinean 5a-reductase-2-deficient individuals who were raised as girls until puberty, when they made the transition to boys, were reported. But today, as in the Dominican kindred, the condition is usually

223 recognized at birth by experienced midwives or is recognized during childhood. The female-to-male gender change produced the term ‘‘turnim man’’ from the Melanesian pidgin and was incorporated into the Sambian lexicon, connoting that these individuals are innately and biologically driven to change gender. Some believe that New Guinean research subjects with 5a-reductase-2 deficiency, as well as others with this condition, are regarded as a ‘‘third sex’’ and change to a male gender role to adapt to their ‘‘male-admiring’’ society. It has been our experience, as well as Gadjusek’s experience, that these individuals clearly regard themselves as male. Three affected New Guinean 5a-reductase-2 patients specifically requested and obtained genital correction during adulthood so that they could be, in their words, ‘‘complete men.’’ In addition, gender change in 5a-reductase-2 deficiency has been noted in affected individuals from many other countries such as Brazil, Turkey, Mexico, Cyprus, Algeria, Italy, Lebanon, Pakistan, Saudi Arabia, and Sweden as well as in Dominican individuals not part of the large kindred described previously. When puberty is allowed to occur without medical intervention, the majority of individuals identify as male. It can be theorized that in normal males, the sex of rearing and androgen imprinting of the brain act in concert to determine male gender expression. Studies of 5a-reductase-2-deficient individuals have shown that when the sex of rearing (female) is discordant with the testosterone-mediated biological sex, the biological sex will prevail when puberty occurs in a nonintervening environment. Thus, under the influence of testosterone-mediated puberty, ‘‘masculinization’’ of the brain theoretically occurs and a male gender identity develops, overriding the female sex of rearing. Therefore, androgens appear to act as inducers and activators in the evolution of male gender identity in humans. It is unclear whether it is testosterone or DHT that directly mediates this process given that both type 1 and type 2 5a-reductase isozymes are present in the brain, and that the type 1 isozyme is normal in individuals with 5a-reductase-2 deficiency. Gender identity has been postulated to become fixed at around the time of language development, between 18 months and 4 years of age. Studies of individuals with 5a-reductase-2 deficiency appear to demonstrate that the development of gender identity is not fixed during early life but rather evolves throughout childhood and becomes fixed during or after puberty.

p0075

224

Androgens, Gender and Brain Differentiation

Gender identity

In utero androgen Postnatal androgen Appearance of external genitalia Self-awareness (testes−− appearance of external genitalia Environmental influences a. Sex of rearing b. Society Male puberty Male

Female

Figure 1 Schematic illustration of the critical factors involved in the evolution of a male gender identity in humans. In summary, environmental and/or sociocultural factors are not the sole factors responsible for the formation of male gender identity; androgens make a strong and definite contribution (Fig. 1). Analogous to the induction of a male phenotype from an inherent female phenotype, the formation of male gender identity in humans appears to be at least partially induced by androgens from an undifferentiated and/or inherently female nervous system.

p0090

p0095

Gender Identity in Research Subjects with 17b-Hydroxysteroid Dehydrogenase-3 Deficiency Genetic male research subjects with 17b-hydroxysteroid dehydrogenase-3 (17b-HSD-3) deficiency are born with severe ambiguity of the genitalia due to deficiency in conversion of androstenedione to testosterone. At puberty, they develop male secondary sexual characteristics (phallic enlargement and abundant body and facial hair), consequent to the peripheral conversion of androstenedione to testosterone by other 17b-HSD isozymes. Gender change from female to male has been reported in subjects with this condition, including those from a large Arab kindred in the Gaza Strip extending over eight generations. From published cases with 17b-HSD-3 deficiency, at least half of the research subjects raised as female change their gender from female to male spontaneously or in consultation with a physician and/or psychiatrist at various ages. In the Arab kindred, some individuals raised as girls exhibited aggressive behavior, leading to their dismissal from school. Those who changed gender

did so on their own initiative, and some did so without parental consent or psychiatric help. The affected individuals are capable of having erections with ejaculations. Three individuals were castrated after diagnosis at the decision of the physicians, and they were raised as women. It is interesting that no individuals from this kindred living as females are known to have married. Some have stated that they would rather have been raised as males, revealing a questionable female identity. One elderly individual was found to have a female gender role but apparently identifies as male. This individual was a farm laborer and was proud to be stronger and more productive than male colleagues. Conversion of androstenedione to testosterone is possible in the human brain. Except for the 17bHSD-3 isozyme that is deficient in these research subjects, other 17b-HSD isozymes are normally expressed in the brains of both humans and animals. Thus, alternate pathways for testosterone and/or DHT formation via other 17b-HSD isozymes are in the brain as well as in extragonadal tissues of patients with 17b-HSD-3 deficiency and can theoretically cause androgen imprinting of the brain. These studies suggest that hormones play a significant role in gender identity formation and that both environmental and hormonal factors influence male gender identity formation in humans.

SEX DIMORPHISM IN COGNITIVE FUNCTION AND LATERALITY Meta-analyses of sex differences in spatial abilities have shown that spatial abilities of males are consistently greater than those of females on certain tasks. A number of tests showed significant male advantages that were stable across age and have not decreased during recent years. Tests included mental rotation tasks, the primary mental abilities (PMA) spatial relations subtest, and the rod and frame test. Male superiority in mathematics performance on the SAT has remained constant, and on a computerized version of a task of spatial ability, males outperform females across a variety of measures. On functional magnetic resonance imaging (fMRI) using a three-dimensional maze, sexually dimorphic differences in areas of brain activation have also been noted. Women, on average, have slightly better verbal skills than do men. When all language measures are combined, women excel in tests of speech production and verbal fluency. Using echo-planer fMRI, a sex difference in brain organization was found during letter recognition, rhyme, and lexical–semantic tasks.

225

Androgens, Gender and Brain Differentiation

There is a trend toward greater right ear superiority in men. Auditory testing of laterality produces the most robust effect, with dichotic consonant–vowel (CV) syllable pairs tasks being the most reliable approach to testing of this type. From the past emerging data, it seems reasonable to suggest that the sex differences described previously are linked to differences in functional organization of the brain.

COGNITIVE ABILITIES IN ANDROGEN-INSENSITIVE INDIVIDUALS p0125

p0130

Individuals with complete androgen insensitivity are phenotypic and psychosexual females due to the lack of androgen response consequent to androgen receptor gene mutations. Individuals with complete androgen insensitivity were studied using the Spanish version of the Wechsler Intelligence Scale (EIWA) to evaluate the relationship between androgen unresponsiveness and cognitive abilities, with particular attention to subtests of visual–spatial ability. In this study, affected individuals, as well as control males and females, all are members of a large kindred with this condition. Subjects with androgen insensitivity demonstrated significantly lower scores on subtests of spatial ability than did either control males or females from the kindred. Because androgen-insensitive subjects are raised as females having a totally female psychosexual orientation, it should be considered that their cognitive performance may reflect the influence of their sex roles as a factor of socialization. However, this consideration does not explain their significantly lower overall performance on the perceptual organization factor and subtests of spatial ability when compared with control females from the same kindred. This exaggerated female pattern of performance suggests an effect of androgen unresponsiveness. Studies of individuals with complete androgen insensitivity by other investigators demonstrate a modest but consistent and significant tendency toward superiority of verbal abilities over space–form abilities using the Wechsler Adult Intelligence Scale. Performance–perceptual scores are poorer in affected individuals than in both male and female controls.

CONCLUSION Over the past two decades, studies of natural human genetic models with deficiency in androgen production or action have demonstrated the importance of

androgens in male sexual differentiation, development of male gender identity/role, and male-patterned behavior and cognitive function. These studies have also revealed that androgen, mainly testosterone, plays a vital ‘‘imprinting’’ role during a critical period of development. These actions appear to be mediated via androgen interacting with the androgen receptor, not via conversion of androgen to estrogen interacting with estrogen receptors as demonstrated in animal studies. This is further supported by the fact that 46,XY individuals with an estrogen receptor mutation or aromatase mutation, resulting in deficiency in estrogen production and action, develop a male gender identity and role. The fact that the majority of research subjects with 5a-reductase-2 deficiency or 17b-HSD-3 deficiency, who were raised as females, changed their gender identity and role to male during or after puberty suggests that the development of gender identity evolves throughout childhood and is fluid until the events of puberty. It is important to bear in mind that androgen acts together with social or environmental factors within the endocrine milieu in the development of the male gender identity/role (Fig. 1). The critical period of androgen exposure of the brain for this development remains uncertain.

See Also the Following Articles Adrenal Androgens . Androgen Biosynthesis and Gene Defects . Androgen Biosynthesis Inhibitors and Androgen Receptor Antagonists . Androgen Insensitivity Syndrome . Endocrine Disrupters and Male Sexual Differentiation . Estrogen and the Male . Germ Cell Differentiation Signaling Events, Male . Hyperandrogenism, Gestational . Mu¨l. lerian Inhibiting Substance: New Insights Pseudohermaphroditism, Male, Due to 5a-Reductase-2 Deficiency . Sexual Function and Androgens

Further Reading Arnold, A. P. (1997). Sexual differentiation of the zebra finch song system: Positive evidence, negative evidence, null hypotheses, and a paradigm shift. J. Neurobiol. 33, 572–584. Goy, R. W., and McEwen, B. S. (1980). ‘‘Sexual Differentiation of the Brain.’’ MIT Press, Cambridge, MA. Imperato-McGinley, J., Peterson, R. E., Gautier, T., and Sturla, E. (1979). Androgens and the evolution of male-gender identity among male pseudohermaphrodites with 5a-reductase deficiency. N. Engl. J. Med. 300, 1233–1237. Imperato-McGinley, J., Peterson, R. E., Stoller, R., and Goodwin, W. E. (1979). Male pseudohermaphroditism secondary to 17bhydroxysteroid dehydrogenase deficiency: Gender role change with puberty. J. Clin. Endocrinol. Metab. 49, 391–395. Imperato-McGinley, J., Pichardo, M., Gautier, T., Voyer, D., and Bryden, M. P. (1991). Cognitive abilities in androgen-insensitive

226 subjects: Comparison with control males and females from the same kindred. Clin. Endocrinol. (Oxf.) 34, 341–347. Imperato-McGinley, J., and Zhu, Y. S. (2002). Gender and behavior in subjects with genetic defects in male sexual differentiation. In ‘‘Hormones, Brain, and Behavior’’ (D. W. Pfaff, A. P. Arnold, A. M. Etgen, S. E. Fahrbach, and R. T. Rubin, eds.), vol. 5, pp. 303–345. Academic Press, Orlando, FL. Poletti, A., Coscarella, A., Negri-Cesi, P., Colciago, A., Celotti, F., and Martini, L. (1998). 5a-reductase isozymes in the central nervous system. Steroids 63, 246–251.

ANF see Atrial Natriuretic Factor

Androgens, Gender and Brain Differentiation

Rosler, A., and Kohn, G. (1983). Male pseudohermaphroditism due to 17b-hydroxysteroid dehydrogenase deficiency: Studies on the natural history of the defect and effect of androgens on gender role. J. Steroid Biochem. Mol. Biol. 19, 663–674. Shaywitz, B. A., Shaywitz, S. E., Pugh, K. R., et al. (1995). Sex differences in the functional organization of the brain for language. Nature 373, 607–609. Voyer, D., Voyer, S., and Bryden, M. P. (1995). Magnitude of sex differences in spatial abilities: A meta-analysis and consideration of critical variables. Psychol. Bull. 117, 250–270.

Angiogenesis Andreas Bikfalvi University of Bordeaux I, Bordeaux, France

Glossary angiogenesis In a strict sense, the formation of vessels from preexisting vessels. Generally, the development of new vessels. angiopoietins Factors implicated in the recruitment of accessory cells to the vasculature and vessel remodeling. arteriogenesis Formation of larger blood vessels including arterioles.

mural cells Cells lining the vessel wall in close contact with the subendothelial matrix and endothelial basement membranes. Pericytes are among the principal mural cells and are found in capillaries. platelet-derived growth factors Growth factors implicated in the recruitment of accessory cells (mainly pericytes) to the vasculature.

endothelial cells Major cell type that is targeted by angiogenesis factors or inhibitors. Endothelial cells are in close contact with the blood and organize into tubular structures.

proteolytic enzymes Enzymes that degrade proteins such as extracellular matrix molecules. The major proteolytic systems for the degradation of matrix molecules are the plasmin/plasminogen activator–inhibitor system and matrix metalloproteinases.

fibroblast growth factors Pleiotropic growth factors that also induce angiogenesis.

pruning Process that yields a mature remodeled vascular network.

hypoxia-inducible transcription factors (HIFs) There are two HIFs, HIF-1a and HIF-2a, that regulate genes involved in angiogenesis. integrins Heterodimeric cell surface receptors for matrix molecules. lymphangiogenesis Formation of lymphatic vessels.

T

he formation of vascular channels, angiogenesis, is a fundamental process that takes place during the embryonic life and also plays a crucial role in the adult organism.

INTRODUCTION Generally, angiogenesis refers to two different mechanisms of vasoformation and types of vessels. Angiogenesis, in a strict sense, describes the formation of new vessels from preexisting functional vessels. On the other hand, vasculogenesis involves the differentiation of endothelial progenitor cells that are incorporated subsequently into vessels. Until recently, it was believed that angiogenesis occurs in the embryo and the adult organism, whereas vasculogenesis occurs

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004, Elsevier Inc. All rights reserved.

RIP-Tag mouse model Transgenic mouse model with targeted expression of large T antigen in pancreatic b cells. A model for multistage carcinogenesis. vascular endothelial growth factors (VEGFs) The major regulators of angiogenesis. vasculogenesis Formation of blood vessels from progenitor cells.

only during embryonic development. This view has been challenged by the identification of circulating endothelial progenitor cells in the adult that are thought to arise in the bone marrow. These circulating progenitors are able to actively participate in angiogenesis processes, such as tissue ischemia, tumor angiogenesis, or ocular neoangiogenesis. In the past, angiogenesis researchers mainly studied the mechanisms of formation of blood vessels. However, lymphatic vessels have become a focus of intense research. The study of arteriogenesis has also broadened the angiogenesis field. Arteriogenesis describes the formation of larger functional blood vessels—a process that is particularly important in the formation of collaterals that limit tissue ischemia.

227

228

Angiogenesis

This article provides a brief overview of the mechanisms of angiogenesis and the impact of angiogenesis research on the pathophysiology and therapy of disease.

BASIC MECHANISMS OF VASOFORMATION Vasoformation is dependent on molecular regulations in both healthy and pathological tissue. This process is dependent on paracrine angiogenesis signals that induce proliferation and migration of vascular cells and their assembly into functional vessels. Vessel stabilization and remodeling are very important events that occur at a later stage. In microvessels, two major cell types, endothelial cells and pericytes, participate in these processes. In larger vessels, smooth muscle cells are involved instead of pericytes. A number of soluble factors, receptors, and extracellular matrix molecules play a role in vascular morphogenesis. Expression of these factors is under the control of a molecular switch inside the cells. Hypoxia is a driving force for angiogenesis in tumors or ischemic tissue (Fig. 1). Hypoxia regulates angiogenesis via an increase in hypoxia-inducible transcription factor-1a (HIF-1a) that initiates a program of survival and adaptive gene expression. In angiogenesis, the major factor regulated through the HIF-1a system is vascular endothelial growth factor (VEGF). In the presence of oxygen, the enzyme prolyl 4-hydroxylase (PHD) binds molecular oxygen and hydroxylates proline residues in HIF-1a. Hydroxylated HIF-1a associates with the von Hippel– Lindau (VHL) gene product, passes to the proteasome, Hypoxia Pro564 HIF1α

Normoxia O2

Fe2+ PHD

Cofactors

Pro564OH HIF1α

VHL

Pro564OH Co-factor HIF1β

HIF1α

HIF1α

HRE Proteasome

Degradation f0005

Figure 1 The oxygen-sensing pathway. HIF1a, hypoxiainducible factor-1a; HIF1b, hypoxia-inducible factor-1b; HRE, hypoxia response element; VHL, von Hippel–Lindau.

and is rapidly degraded. Under hypoxic conditions, hydroxylation is inhibited and HIF-1a levels are stabilized. There are three sequence-related PHDs. Probably only one of these forms, PHD2, which resides in the cytoplasm, is involved in HIF-1a regulation during angiogenesis. PHD2 is also transcriptionally induced by HIF-1a in a low-oxygen environment. This provides an autoregulatory feedback loop. In vascular cells, another form of HIF, HIF-2a, is responsible for regulating gene expression. For example, an important receptor that binds VEGF, VEGFR2, is up-regulated through a HIF-2adependent mechanism. The molecular angiogenesis switch is dependent on hypoxia, oncogenic transformation, which occurs in tumor cells, or on autocrine growth factor loops. For example, activation of the ras gene product induces the expression of VEGF, a stimulator of angiogenesis, and down-regulates inhibitors such as thrombospondins in tumor cells. Similar effects are also observed when autocrine growth factor loops are present in tumor cells.

KEY FACTORS REGULATING VASCULAR DEVELOPMENT Among the most important regulators of angiogenesis are the VEGFs. This family is composed of VEGFA,-B,-C, and-D and the related placental growth factors (PLGFs). VEGFs are essential in embryonic and postnatal vascular development. They also play an important role in ischemia-driven or tumor angiogenesis. In fact, one of the VEGF prototypes, VEGF-A, is a permissive factor for multistage carcinogenesis, as evidenced in the RIP-Tag mouse model. VEGFs bind three types of tyrosine-kinase receptors: VEGFR1 (flt1), VEGFR2 (flk1 or KDR), and VEGFR3 (flt3). VEGF-A binds VEGFR2 and VEGFR1. In contrast, VEGF-B binds only VEGFR1. VEGF-C and VEGF-D both preferentially bind VEGFR3 but also interact with VEGFR2. Finally, PLGFs bind only VEGFR1. VEGFR2 is the critical receptor for angiogenesis in blood vessels and for vascular permeability. VEGFR1 may synergize with VEGFR2 in postnatal and pathological angiogenesis. VEGFR1 is also found on hematopoietic cells that accumulate at angiogenic sites. However, only VEGFR2 seems to be necessary for embryonic vascular development. Nevertheless, both receptors are needed for repair-associated, tumor, or retinal neoangiogenesis in the adult.

229

Angiogenesis

p0060

p0065

p0070

p0075

p0080

VEGFR3/flt4 is required for the growth and the maintenance of the lymphatic vessels and is also mutated in primary human lymphedema (Milroy’s disease), which illustrates the importance of this receptor in the lymphatic tissue. Targeted inactivation of VEGFR3 in mice results in abnormal lymphatic vessel growth. VEGF prototypes also use coreceptors for binding to target cells. Among the coreceptors, neuropillins seem to be the most important. For instance, neuropillin-1 is a coreceptor of VEGFR2. This interaction is critical for VEGF-A binding to VEGFR2. VEGF-A is the main VEGF form that regulates angiogenesis in blood vessels, whereas lymphangiogenesis is primarily dependent on VEGF-C. Several splice variants of VEGF-A (VEGF-165, VEGF-121, and VEGF-189) have been identified that exhibit variable affinities for heparan sulfates. This interaction is important for VEGF’s biodisponibility. VEGF expression is regulated by a number of factors, including hypoxia, oxidative stress, reactive oxygen species, tumor suppressor genes or growth factors, or cytokines that activate the MAP kinase pathway. VEGF alone appears to be unable to direct blood vessel organization or maturation. A number of other factors act not only on vascular endothelial cells but also on other vascular cells, such as pericytes. They may also contribute to vessel stabilization or pruning. Among these factors are fibroblast growth factors (FGFs), platelet-derived growth factors (PDGFs), angiopoietins, developmental gene products such as Notch/Delta, or molecules classically involved in axonal guidance. Particularly noteworthy among the latter are ephrins and neuropillins. The major system involved in vessel stabilization is the angiopoietin system. Angiopoietin-1 (Ang-1) is thought to stabilize vessels by rendering them less sensitive to VEGF. The stabilizing effect of Ang-1 is disrupted by angiopoietin-2 (Ang-2) binding to the Ang-1 receptor Tie-2. Ang-1 stimulates endothelial cell migration by activating tie-2 and inducing the recruitment of Dok-R, NcK, and Pack. Ang-1, but not Ang-2, has been shown to be in the extracellular matrix. Ang-1 induces platelet-derived factors (PDGF), which is released and stimulates migration of pericytes, smooth muscle cells, and other accessory cells and, thus, favors mural cell coverage. Ang-1 is able to trigger vessel remodeling, as demonstrated in the retina. Angiogenesis is also under the control of proteolytic enzymes and inhibitors, including the plasminogen activator system and matrix metalloproteinase and their inhibitors. The basic principle of the activity of

these enzymes is that they must be present at a critical concentration at the cell surface to promote invasion of vascular tubes. Furthermore, inhibitors such as tissue inhibitor of metalloproteinase-2 for matrix metalloproteinase (MMPs) may be required to localize the proteolytic activity at the cell surface, thus promoting activation of MMPs. This may account for the paradoxical stimulatory effects observed for these inhibitors in some circumstances. Cell adhesion molecule receptors such as integrin avb3 (and, to a lesser extent, avb5) also play a critical role in angiogenesis. For example, the integrin avb3 is highly expressed in proliferating endothelial cells, and both a monoclonal antibody to avb3 and a lowmolecular-weight antagonist have been shown to inhibit angiogenesis in in vivo models. This indicates that integrin avb3 has a promoting role in angiogenesis and may constitute a potential interesting therapeutic target. However, recent observations of avb3 knockout mice have challenged this view. In particular, it has been reported that mice lacking b3 integrin show enhanced pathological angiogenesis. The reason for these apparent differences is not clear. Cheresh and collaborators attempted to explain the differences using the concept of ligated and unligated integrins. Unligated integrins provide dead signals to endothelial cells that are blocked by ligation. Knocking out integrins affects both ligated and unligated integrins and may artificially increase the invasiveness of endothelial cells by increasing endothelial cell survival, assuming that more unligated than ligated integrins are present.

ANGIOGENESIS AND PATHOLOGY Angiogenesis is a driving force for a number of pathologies, such as cancer, ocular neovascular disease, ischemic disease, and chronic inflammatory diseases. Cancer cells express a complex molecular repertoire that critically impacts the surrounding vascular stroma (Fig. 2). It is clearly recognized that tumor cells produce both negative and positive regulators of vasoformation, and that the net effect on the vasculature is the outcome of the balance between these two types of regulators. That the growth of tumors is dependent on the vasculature was recognized by Algire in 1945 and later formulated as a paradigm by Judah Folkman. In general, tumors less than 2 or 3 mm are avascular and grow slowly. Tumor cells are then activated intracellularly by a mechanism called angiogenic switch and then start to favor the positive

230

Figure 2 Regulation of tumor angiogenesis. Tumors produce both positive (activators) and negative (inhibitors) regulators and regulate both angiogenesis and lymphangiogenesis. Inhibitors of lymphangiogenesis have not been identified. VEGF, vascular endothelial growth factor; FGFs, fibroblast growth factors; TSP, thrombospondins.

regulators over the negative regulators. First insight into the molecular mechanisms of the angiogenic switch was gained through the analysis of tumor progression in the RIP-Tag mouse model. Overexpression of the T antigen under the control of the insulin promoter induces an insulinoma that progresses to highly malignant and vascularized lesions. Tumor nodules from the early stage do not induce angiogenesis when cocultured with endothelial cells in vitro. At the later stage, however, angiogenesis is potently induced. A number of molecules are involved in the angiogenic switch. VEGF is required since mice with conditional inactivation of VEGF do not undergo the angiogenic switch. MMP3 and MMP9 are also required. These proteases allow the translocation of VEGF from the tumor matrix to the blood vessels. FGF prototypes also seem to be needed in this model. Tumors are not identical and may be more or less dependent on one or the other angiogenesis factor or inhibitor. For example, prostate carcinoma and melanoma are more dependent on FGF and VEGF, whereas other tumors, such as ovarian carcinoma, are primarily dependent on VEGF. Furthermore, tumors may modify their angiogenesis repertoire during their lifetime and switch from one to the other growth factor/growth factor receptor system. This may constitute an escape mechanism for tumors due to environmental constraints.

Angiogenesis

Tumors may also express angiogenesis inhibitors, such as thrombospondin-1 (TSP-1) and TSP-2. Involvement of TSPs in tumor angiogenesis has been particularly well studied in mouse models of skin carcinogenesis. TSP-1 and TSP-2 both seem to inhibit tumor angiogenesis and tumor growth in models of skin carcinoma in mice. On the other hand, TSP-2, but not TSP-1, increases with tumor progression, indicating that TSP-2, but probably not TSP-1, is part of a host antitumor defense mechanism. In addition to blood vessels, lymphatic vessels are also required for dissemination of tumor cells. Factors critically involved in the development of lymphatics, such as VEGF-C and FGFs, may also play a role in lymphangiogenesis in tumors. Blocking lymphangiogenesis in a highly metastatic human lung cancer cell line by inhibiting VEGF-C suppresses lymph node metastasis. Furthermore, crossing RIP-Tag mice with mice expressing VEGF-C under the control of the insulin promoter yields bigenic mice that develop pancreatic tumors with metastatic spread through the lymphatic system. Most important, Dadras et al. reported that intratumoral lymphangiogenesis in melanoma patients is associated with poor survival. Ocular neovascular disease is another pathological condition in which abnormal angiogenesis plays a leading role. Diseases include diabetic retinopathy and age-related macular dystrophy (ARMD). VEGF seems to be one of the principal factors in the pathophysiology of these diseases. In the case of ARMD, the fas/fas ligand system may also have a critical role. Indeed, fas/fas ligands are expressed in the retinal pigmented epithelium and the choroid vessels. Knockout mice for fas/fas ligand exhibit aberrant ocular neovascularization resembling ARMD. Angiogenesis is also implicated in ischemic disease in two ways. First, capillary growth within the walls of large arteries may contribute to the establishment of a proliferative lesion and invasion into the intima. It has been reported that inhibition of plaque neovascularization reduces the accumulation of macrophages and progression to advanced atherosclerotic lesions. Second, the growth of the collateral circulation that limits the extent of the ischemic lesion is also controlled by angiogenesis factors such as VEGF or FGFs. This has offered novel therapeutic opportunities to salvage the ischemic area and to improve morbidity and mortality in patients with coronary or peripheral artery disease. Finally, chronic inflammatory disorders such as chronic polyarthritis are also angiogenesis dependent. In the early phase, this disease is characterized by a

p0115

p0120

231

Angiogenesis

proliferative lesion of synoviocytes in the synovia. Within the inflamed synovia, the number and quality of microvessels are also altered. VEGF and integrin avb3 seem to play an essential role since blocking their activity in animal models results in disease improvement.

ANGIOGENESIS AND THERAPY

p0135

A number of molecules are under preclinical or clinical evaluation for angiogenic or antiangiogenic therapy. Clinical trials using VEGF or FGFs to stimulate the collateral circulation in patients with coronary disease or peripheral arterial ischemic disease are under way. On the other hand, inhibition of angiogenesis is therapeutically relevant in cancer, retinal neovascular disease, or chronic inflammatory disease. The major disease for which antiangiogenesis strategies have been investigated is cancer. Targeting the vasculature for angiogenesis inhibition may be achieved through extracellular or intracellular mechanisms (Fig. 3). Extracellularly, inhibitors may block binding of angiogenesis factors to receptors, interfere with coreceptor/ligand/receptor interactions, activate inhibitor receptors, or modulate integrin–extracellular matrix interactions or protease activity. Intracellularly, inhibitors may directly inhibit tyrosine kinase receptors or signaling modules. Among inhibitors, molecules interfering with VEGF activity are a major focus of research. VEGF activity may be

blocked with anti-VEGF or VEGF receptor antibodies or specific VEGF receptor kinase inhibitors. Ferrara and collaborators were the first to show that antibodies to VEGF slowed tumor growth. Human forms of the antibody (Avastin) are now in phase 3 clinical trial for the treatment of solid tumors. Recently, remarkable results were announced in a phase 3 clinical trial with Avastin in combination with chemotherapy in colon carcinoma. Low-molecular-weight inhibitors of the tyrosine kinase activity of the VEGF receptor (VEGFR-KIs) are another venue for antiangiogenesis tumor therapies. Encouraging results have been obtained with VEGF-KIs in kidney and colon carcinoma. Another target for development of angiogenesis inhibitors is molecules encrypted within larger regulatory proteins, including endostatin, angiostatin, thrombospondin, platelet factor-4 (PF-4), and endorepellin. For example, a peptidomimetic of thrombospondin has been synthesized that exhibits high antiangiogenesis activity and is currently in clinical trials. A C-terminal fragment of PF-4 has also been developed that exhibits increased antiangiogenic and antitumor activity in glioblastoma. Furthermore, antiangiogenesis properties of molecules already known to exhibit inhibitory activities on other cell types have recently been discovered. Particularly noteworthy is the fact that chemotherapy, when given at low dose and metronomically, has mainly antiangiogenic effects and greatly inhibits tumor development in mice.

Figure 3 Angiogenesis inhibition versus vascular targeting. For angiogenesis inhibition, inhibitors may block the interaction of angiogenesis factors with their receptor by disrupting specific binding to receptors or coreceptors. Furthermore, inhibitors may activate inhibitor receptors such as CD36 for thrombospondin-1. Inhibitors may also disrupt the interaction between integrins and matrix molecules. Targeting agents bind surface molecules that are specifically present on the surface of angiogenic endothelial cells or in the extracellular matrix around blood vessels.

232

p0155

Combinatory approaches involving several molecules or treatment regiments have been developed to increase therapeutic efficacy. Antiangiogenic molecules may be combined with chemotherapy or radiotherapy. These approaches have been validated in the mouse, and phase 3 clinical trials are under way involving, for example, the association of chemotherapy with Avastin in colon or kidney carcinoma. Another approach is to combine several antiangiogenesis drugs. For example, the combination of endostatin, angiostatin, and TNP470 has been shown to be more effective than any single agent alone. Bergers and colleagues associated a VEGFR-KIs with Gleevec, a drug used for the treatment of chronic myelogenous leukemia. In addition to inhibiting the BCR-abl tyrosine kinase activity, Gleevec also inhibits c-kit and PDGFR tyrosine kinase. Since PDGFs are critically involved in pericyte recruitment and vessel stabilization, inhibition of this receptor tyrosine kinase may be important to resensitize vessels to VEGF inhibition. This is indeed the case, as demonstrated by Bergers and colleagues using the RIP-Tag mouse model. Treatment of established tumors leads to regression when the two inhibitors are combined. Vascular targeting is different from the approaches described previously (Fig. 3). It involves selective targeting of tumor blood vessels and their subsequent destruction. This approach is based on the assumption that the tumor vasculature is different from that in normal tissues. Much evidence supports this claim. Tumor endothelial cells are often hypoxic, lack nutrients, and exposed to a number of cytokines and stress factors. This modifies the gene expression profile, making them angiogenic. The proof of concept that targeting is a valuable approach for cancer treatment was provided in 1993 when Burrows and Thorpe used an anti-MHC class II antibody coupled to ricin toxin to destroy tumor vasculature in experimental animal models. Other compounds that target the vasculature include the combretastatins, which are currently in phase 1 trials. They are selectively toxic to tumor vasculature by disrupting the tubulin cytoskeleton. Recently, impressive results were reported by Halin and collaborators using a vascular-specific antibody (L19) that was isolated from an antibody phage library and that recognized the ED-B domain of fibronectin. This antibody is selectively targeted to the tumor endothelium. Chimeras of the antibody with interleukin-12 or tissue factor caused regression of established tumors when injected in experimental mouse tumor models. This oxygen-sensing pathway also offers opportunities for therapeutic intervention. PHD inhibitors should be proangiogenic, whereas molecules that

Angiogenesis

abrogate HIF-1a binding to coactivators and/or the hypoxia response element reduce angiogenesis and tumor growth. Modulation of the iron content of the cell may also affect the activity of the PHDs because these enzymes are sensitive to the local Fe2þ concentration.

CONCLUSION Angiogenesis is a fundamental mechanism in embryonic and postnatal development that also plays a major role in pathologies such as cancer, ocular neovascular disease, ischemic disease, and chronic inflammatory disease. Many molecules, receptors, and intracellular signaling modules have been implicated in vascular morphogenesis. These discoveries have offered novel opportunities for therapeutic intervention. There is an increasing repertoire of drugs with which to manipulate angiogenesis and new endothelial-specific genes with which to target the vasculature. Thus, angiogenesis research is currently at an exciting stage.

See Also the Following Articles Fibroblast Growth Factor (FGF) . Platelet-Derived Growth Factor (PDGF) . Prostate Cancer

Further Reading Algire, G. H., Chalkely, H. W., Legallais, F. Y., and Park, H. (1945). Vascular reactions of normal and malignant tumors in vivo: I. Vascular reactions of mice to wounds and to normal and neoplastic transplants. J. Natl. Cancer Inst. 6, 73–85. Alitalo, K., and Carmeliet, P. (2002). Molecular mechanisms of lymphangiogenesis in health and disease. Cancer Cell 1, 219–227. Bergers, G., Song, S., Meyer-Morse, N., Bergsland, E., and Hanahan, D. (2003). Benefits of targeting both pericytes and endothelial cells in the tumor vasculature with kinase inhibitors. J. Clin. Invest. 111, 1287–1295. Bikfalvi, A., and Bicknell, R. (2002). Recent advances in angiogenesis, anti-angiogenesis and vascular targeting. Trends Pharmacol. Sci. 23, 576–582. Carmeliet, P. (2000). Mechanisms of angiogenesis and arteriogenesis. Nature Med. 6, 389–395. Cheresh, D. A., and Stupack, D. G. (2002). Integrin-mediated death: An explanation of the integrin knockout phenotype? Nature Med. 8, 193–194. Dadras, S. S., Paul, T., Bertoncini, J., Brown, L. F., Muzikansky, A., Jackson, D. G., Ellwanger, U., Garbe, C., Mihm, M. C., and Detmar, M. (2003). Tumor lymphangiogenesis: A novel prognostic indicator for cutaneous melanoma metastasis and survival. Am. J. Pathol. 162, 1951–1960. Ferrara, N. (2003). Role of vascular endothelial growth factor in physiologic and pathologic angiogenesis: Therapeutic implications. Semin. Oncol. 29(6 Suppl. 16), 10–14. Folkman, J. (1971). Tumor angiogenesis: Therapeutic implications. N. Engl. J. Med. 285, 1182–1186.

Angiogenesis

Hagedorn, M., and Bikfalvi, A. (2000). Target molecules for antiangiogenic therapy: From basic research to clinical trials. Crit. Rev. Oncol. Hematol. 34, 89–110. Halin, C., Rondini, S., Nilsson, F., Berndt, A., Kosmehl, H., Zardi, L., and Neri, D. (2002). Enhancement of the antitumor activity of interleukin-12 by targeted delivery to neovasculature. Nat. Biotechnol. 20, 264–269. Javerzat, S., Auguste, P., and Bikfalvi, A. (2002). The role of fibroblast growth factors in vascular development. Trends Mol. Med. 8, 483–489. Kerbel, R., and Folkman, J. (2002). Clinical translation of angiogenesis inhibitors. Nature Cancer Rev. 2, 727–739. Oliver, G., and Detmar, M. (2002). The rediscovery of the lymphatic system: Old and new insights into the development and

233 biological function of the lymphatic vasculature. Genes Dev. 6, 773–783. Rak, J., Yu, J. L., Kerbel, R. S., and Coomber, B. L. (2002). What do oncogenic mutations have to do with angiogenesis/vascular dependence of tumors? Cancer Res. 62, 1931–1934. Reyes, M., et al. (2002). Origin of endothelial progenitors in human postnatal bone marrow. J. Clin. Invest. 109, 337–346. Ruoslahti, E. (2002). Specialization of tumour vasculature. Nature Cancer Rev. 2, 83–90. Semenza, G. (2000). HIF-1: Mediator of physiological and pathological responses to hypoxia. J. Appl. Physiol. 88, 1474–1480. Vale, P. R., Isner, J. M., and Rosenfield, K. (2001). Therapeutic angiogenesis in critical limb and myocardial ischemia. J. Interv. Cardiol. 14, 511–528.

Angiotensin, Evolution of Werner Kloas Leibniz Institute of Freshwater Ecology and Inland Fisheries, Berlin, Germany

Glossary angiotensin receptors All angiotensin receptors identified in several classes of vertebrates share a high degree of homology and belong to a superfamily of seven-transmembrane G protein-coupled receptors. angiotensins Bioactive peptide hormones of varying length between 6 and 10 amino acids. The main functions of angiotensins are associated with osmomineral and water balance as well as maintenance of blood pressure in vertebrates. evolution of the renin–angiotensin system (RAS) The functioning systems to produce angiotensin are present in all classes of vertebrates and some invertebrate groups, indicating an early appearance of endocrine function of angiotensin during evolution.

Angiotensin (Ang) is a classical peptide hormone important in osmomineral regulation and blood pressure maintenance of vertebrates. The Ang-producing renin–angiotensin system (RAS) and corresponding receptors for Ang are present in all classes of vertebrates. Comparative studies indicate that homologous Ang and RAS are present in several groups of invertebrates, implicating an important biological role of Ang during early evolution.

INTRODUCTION The presence of the peptide hormone Ang produced by the renin–angiotensin system (RAS) and its corresponding receptors is reported in all classes of vertebrates. Ang plays an important role in the control of blood pressure and osmomineral regulation directly by vasoconstrictory actions in the circulatory system or indirectly by regulating osmominerals, affecting hormones such as corticosteroids, catecholamines, or nonapeptides, and by influencing drinking behavior as a dipsogenic agent. However, studies indicate that the evolution of Ang occurred at an earlier evolutionary period in several groups of invertebrates.

234

COMPONENTS OF THE RENIN–ANGIOTENSIN SYSTEM In vertebrates, the major pathway of systemic Ang is the generation of the decapetide Ang I (angiotensins 1–10) by the enzyme renin from liver-produced angiotensinogen. Ang I is subsequently converted to the octapetide Ang II (angiotensins 1–8) by Ang-converting enzyme (ACE) and then metabolized to smaller peptides exhibiting biological activities, such as Ang III (angiotensins 2–8), Ang IV (angiotensins 3–8), and Ang V (1–7). In addition, the tissue-specific existence of Ang-producing RAS has been demonstrated within the brain, pituitary, gonads, and adrenal, where Ang is processed not only via renin and ACE but also by various other enzymes.

Angiotensins Amino acid sequences of Angs are highly preserved throughout evolution and differ only in positions 1, 3, 5, and 9, where exchanges of amino acids may occur (Table I). However, species-specific differences in the structure of Angs can markedly modify their biological activities depending on the corresponding Ang receptors of the respective species. It is noteworthy that the structure of Ang I in humans and leeches is identical.

Renins and Angiotensin-Converting Enzymes Renins and renin-like enzymes cleaving Ang I by proteolytic cleavage from angiotensinogen are present in all classes of vertebrates and also found in leeches and insects. The presence of ACEs generating the biologically active octapetide Ang II from Angs I is established for all vertebrates, and functioning ACEs have been demonstrated in invertebrates such as insects, leeches, and mollusks.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

235

Angiotensin, Evolution of

Table I Angiotensin I Amino Acid Sequences from Vertebrates and Annelidsa Amino acid sequence

Common structure

1

2

3

4

5

6

7

8

—

Arg

—

Tyr

—

His

Pro

Phe

9

10

—

Leu

Species variation Mammals (human)

Asp

Val

Ile

His

Birds (chicken) Reptiles (alligator)

Asp Asp

Val Val

Val Val

Ser Ala

Amphibians (bullfrog)

Asp

Val

Val

Asn

Bony fishes (eel)

Asn

Val

Val

Gly

Cartilaginous fishes (dogfish)

Asn

Pro

Ile

Gln

Cyclostomes (hagfish)

Asn

Val

Val

Thr

Annelids (leech)

Asp

Val

Ile

His

a

Native angiotensin I from all vertebrate classes and annelids. Among vertebrate classes additional species-specific differences in amino acid sequences in positions 1, 3, 5, or 9 are present or may be found.

ANGIOTENSIN RECEPTORS AND CELL SIGNALING PATHWAYS All Ang receptors identified and cloned in vertebrates belong to the superfamily of seven-transmembrane G protein-coupled receptors, which have several second messenger systems, including phosphatidylinositol-4,5-bisphosphate, intracellular Ca2+, endothelium-derived relaxing factor, nitric oxide, and cGMP. In invertebrates, evidence of functioning Ang receptors in leeches was obtained by binding studies and physiological experiments.

BIOLOGICAL ACTIONS The classical functions of Ang in vertebrates are the maintenance of water and osmomineral contents and blood pressure. In higher vertebrates, Ang decreases the glomerular filtration rate in the kidney and stimulates the adrenal to secrete mineralocorticoid aldosterone by the cortex and to release the catecholamines adrenaline and noradrenaline by the medulla. The effects on the central nervous system include dipsogenic behavior and increased hypophyseal secretion of nonapeptides (vasopressin or vasotocin), which act positively on water conservation and blood pressure. In addition, Ang acts on memory processes, reproduction, and immune response modulation of mammals. The endocrine counterpart on nearly all target organs for Ang and its biological actions is the family of natriuretic peptides, which counteract the

maintenance of water and osmomineral homeostasis in all classes of vertebrates. However, main target organs for Ang may differ in vertebrates during early evolution. In amphibians and fish, stimulatory effects of Ang on the adrenal homologues producing catecholamines or corticosteroids are found in some species but absent in others. Additional Ang targets in fish are osmoregulatory organs, such as gill, intestine, and chondrichthyean rectal gland, all of which are very important for osmoregulation. Angiotensin-induced effects on dipsogenic behavior and renal functions also vary widely among lower vertebrates, whereas vasopressor activities are always prominent. In invertebrates, biological actions of Ang are reported for leeches, in which Ang has a diuretic effect by influencing Cl transport, demonstrating its physiological significance for maintenance of osmoregulation after a blood meal in these annelids. Angiotensin is one of the classical vertebrate hormones involved in osmomineral and blood pressure regulation. The development of more complex endocrine effects of Ang on several target organs during vertebrate evolution is associated with changes from aquatic to semiterrestrial and terrestrial lifestyles, for which salt and water conservation as well as maintenance of blood pressure are essential. However, comparative studies indicate that Ang originated much earlier during evolution in annelids, in which it is also associated with water and osmomineral regulation.

236

See Also the Following Articles Hypertension, Renin and . Renin

Further Reading Anderson, W. G., Cerra, M. C., Wells, A., Tierney, M. L., Tota, B., Takei, Y., and Hazon, N. (2001). Angiotensin and angiotensin receptors in cartilaginous fishes. Comp. Biochem. Physiol. A 128, 31–40. Kobayashi, H., and Takei, Y. (1996). The renin–angiotensin system. Comparative aspects. Zoophysiology 35, 77–92.

Anorchia see Agonadism

Angiotensin, Evolution of

Nishimura, H. (2001). Angiotensin receptors—Evolutionary overview and perspectives. Comp. Biochem. Physiol. A 128, 11–30. Russell, M. J., Klemmer, A. M., and Olson, K. R. (2001). Angiotensin signaling and receptor types in teleost fish. Comp. Biochem. Physiol. A 128, 41–51. Salzet, M., Deloffre, L., Breton, C., Vieau, D., and Schoofs, L. (2001). The angiotensin system in elements in invertebrates. Brain Res. Rev. 36, 35–45. Sandberg, K., and Ji, H. (2001). Comparative analysis of amphibian and mammalian angiotensin receptors. Comp. Biochem. Physiol. A 128, 53–75.

Anorexia Nervosa Jesu´s Argente University Auto´noma and Hospital Universitario Infantil Nin˜o Jesu´s, Madrid, Spain

Marı´a Teresa Mun˜oz Hospital Infantil Universitario Nin˜o Jesu´s, Madrid, Spain

PREVALENCE Glossary amenorrhea The absence of menstruation during a minimum period of 3–6 months in women who have previously experienced menstruation (secondary amenorrhea) or the lack of menarche at 16 years of age (primary amenorrhea). bone mass Total quantity of bone tissue, including the total volume of bone tissue and the total quantity of mineralized extracellular matrix. malnutrition A pathological condition occurring when energy and nutrient requirements are not met by dietary intake. There is a wide spectrum of clinical forms, depending on the severity and duration of the deficit, the age of the subject, and the cause of the condition. osteopenia Significant diminution of bone mass per unit volume in relation to what is considered normal for the age, pubertal stage, and sex of the subject. osteoporosis A disease characterized by low bone mass and microarchitectural deterioration of bone tissue, leading to enhanced bone fragility and a consequent increase in fracture risk.

A

norexia nervosa (AN) is a childhood psychiatric disorder characterized by patient-induced and maintained weight loss that leads to progressive malnutrition and specific pathophysiological signs (disturbance of body image and fear of obesity). Diagnostic criteria were proposed for the first time in 1972 by Feighner et al. and were later modified by the American Academy of Psychiatry in the third edition of the Diagnostic and Statistical Manual of Mental Disorders. These criteria allowed uniform classification of patients with these characteristics and also suggested the existence of less severe forms, called nonspecified eating disorders. These forms can also have a dangerous evolution and therefore require therapy equal to that of the clinical forms. The diagnostic criteria for AN were updated in 1994 in the fourth edition of the DSM.

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

The prevalence of anorexia nervosa (AN) among adolescents and young adults is approximately 0.5–1%, with a bimodal age distribution including peak ages of 14 and 18 years. However, a significant number of cases in girls during the initial stages of puberty, as well as before the onset of puberty, have been observed. The female/male ratio ranges from 5:1 to 20:1. A study of a Spanish female population between 12 and 21 years of age showed a prevalence of 0.3% for AN, 0.8% for bulimia, and 3.1% for nonspecified eating disorders; thus, a total of 4.1% of the population suffers from some type of eating disorder. However, 50–67% of adolescent females are dissatisfied with their weight and body shape, and most adolescents have been on a diet. Many of these teens use unhealthy weight control methods, including fasting, diet pills, and vomiting. Adolescents with chronic diseases, especially females, have a higher incidence of eating disorders and special attention should be given to these patients to ensure that they do not develop clinically significant cases of anorexia or bulimia. Indeed, a cross-sectional study indicates that the prevalence is approximately twice that seen in age-matched controls.

ETIOPATHOGENESIS The etiology of AN is multifactorial, including genetic, biological, psychological, and cultural factors. The coexistence of various risk factors increases the possibility of developing this disease (Table I).

Individual Factors Patients with AN frequently have personality disorders, including low self-esteem and high anxiety. They are introverted, obsessive, and perfectionists, and feel as if they are little effective, although the

237

238

Anorexia Nervosa

Table I Diagnostic Criteria for Anorexia Nervosaa Refusal to maintain body weight at or above a minimally normal weight for age and height (e.g., weight loss leading to maintenance of body weight less than 85% of that expected; or failure to make expected weight gain during period of growth, leading to body weight less than 85% of that expected) An intense fear of gaining weight or becoming fat, even though underweight A disturbance in the way in which one’s body weight or shape is experienced, undue influence of body weight or shape on self-evaluation, or denial of the seriousness of the current low body weight In postmenarcheal females, amenorrhea (the absence of at least three consecutive menstrual cycles) Two subtypes are described: restricting subtype and bingeing/purging subtype a

From the American Psychiatric Association (1994).

reverse is usually the case. In addition, those patients that use purgative methods have a tendency to steal and often have problems with alcohol and drugs. It is significant that in these patients the loss of self-esteem and autocontrol culminates in them going on a diet. To this culturally accepted activity, they often add other useful activities, such as studying many hours and intense physical exercise, which often result in a brief improvement in their condition. The malnutrition provokes a series of alterations, both physical and mental, and the capacity to relate socially is altered, producing a new decrease in self-esteem and autocontrol. The response is a more strict diet, which can then perpetuate a dangerous spiral of events. Some high-level athletes develop AN, but others present incomplete or subclinical forms with a more difficult diagnosis. A form called athletic anorexia has been described that includes high-level female athletes who have an intense fear of gaining weight or becoming fat, even though they are below their ideal weight, with low-caloric intake and often excessive exercise.

Genetic Factors No specific marker has been identified that suggests an increased possibility of AN. Investigation has concentrated on analysis of polymorphisms of genes related to weight control and the serotoninergic and dopaminergic pathways. One of the most frequently studied genes is that of the serotonin receptor. Serotonin [5-hydroxytryptamine (5-HT)] is involved in a broad range of biological, physiological, and behavioral functions and has been implicated in the development of eating disorders. Polymorphisms within different 5-HT receptor genes and the tryptophan hydroxylase gene have also been analyzed. Some studies have described an association between the-1438 A allele of the -1438 G/A polymorphism within the promoter region of the 5-HT2A receptor

gene and AN. These studies indicate that patients with AN have enhanced 5-HT2A receptor binding and provide further evidence for a serotoninergic dysfunction in eating disorders, although further studies are needed. The hypothalamus is also an important organ in the control of energy metabolism because it is responsible for the sensation of hunger and satiety and hence energy intake. In addition, via modulation of the sympathetic nervous system, it is involved in thermogenesis and, as a result, energy expenditure. Various neuropeptides control these functions. Regulation of acute eating behavior incorporates a system of satiety signals originating in the food. Cholecystokinin, bombesin, gastrin-releasing peptide, and others are involved in this signaling system and reach the brain via peripheral innervation or the circulation to activate their receptors. Long-term energy balance is regulated via a system composed of different hormones secreted in proportion to corporal adiposity, such as leptin and insulin, that act at the level of the central nervous system. These respond to changes in body fat by activating anabolic or catabolic pathways, the first through production of neuropeptide Y (NPY), which stimulates food intake, and the second via the hypothalamic melanocortin system, which reduces food intake and stimulates weight loss (Fig. 1). Leptin, a hormone synthesized by adipose tissue, plays an important role in the regulation of food intake and energy expenditure. Its mechanisms of action are unknown, although its primary target seems to be the hypothalamus. Leptin acts at the level of the hypothalamus to decrease appetite, resulting in weight loss. Plasma leptin levels and secretory pattern vary during the night and day and are influenced by food intake. Molecular analysis of the coding region and part of the promoter region of the leptin gene in patients with AN has yielded negative results. Hence, involvement of the leptin gene in the etiology of AN seems unlikely.

239

Anorexia Nervosa

Energy balance Negative

Positive

Leptin

Leptin

+

− α-MSH CRH

NPY

−

+ α-MSH CRH

(a-MSH), which has a high affinity for the MC receptors, especially MC3 and MC4. Mutation screening of the coding region of MC4-R in patients with AN and bulimia nervosa revealed two common polymorphisms in both groups. Allele and genotype frequencies did not differ between these groups and probands of different weight extremes.

NPY

Familial Factors

Food intake and weight gain

Food intake and weight gain

Figure 1 Hormonal response to malnutrition. It has been demonstrated that a gastrointestinal peptide hormone, ghrelin, stimulates growth hormone secretion in rats and humans. Ghrelin also plays an important role in the regulation of energy balance. Plasma ghrelin levels are regulated by acute and chronic changes in energy balance (e.g., fasting increases, whereas feeding decreases, circulating ghrelin concentrations). The actions of NPY, incrementing ingestion and decreasing thermogenesis, are opposite those of leptin. Neuropeptide YY5 and YY1 receptors in rats and humans are assumed to play a major role in NPYinduced food intake. The neuropeptide YY5 receptor gene (NPYY5R) is expressed in brain regions involved in the central regulation of feeding behavior, including the lateral hypothalamus, the paraventricular nucleus, and the arcuate nucleus. Systematic mutation screening within the coding region of the NPYY5R revealed a rare Glu-4-Ala variant in a single patient with AN. This allele was transmitted from the mother, who had no antecedents of any eating disorder. Association and transmission disequilibrium studies pertaining to variations and polymorphisms within NPYY1R and NPYY5R and AN were negative. Glucocorticoids are also implicated in energy regulation. Via their effects on NPY, they act as endogenous antagonists of leptin and insulin. Other neuropeptides that stimulate food intake and energy storage are melatonin-concentrating hormone and orexin A and B, which increase in response to fasting and stimulate appetite. The melanocortins (MCs) are peptides derived from the precursor proopiomelanocortin (POMC) and act on specific receptors. The endogenous MC most implicated in food intake and body weight is a-melanocyte-stimulating hormone

The families of patients with AN have certain characteristics in common. They are often overprotective, strict, and have a relative incapacity to solve their own conflicts. The mother figure is described as the boss of the family and the father figure as distant. In many families, the patient is often recognized as an individual only after the onset of the illness; consequently, the patient continues with the illness in order to remain the center of the family’s attention.

Sociocultural Factors These adolescents are often very vulnerable, receiving a large quantity of information that they cannot assimilate, which creates tension regarding problems normal for their age, including sexuality, competitiveness, individuality, and independence within the family. It is well-known that every historic period determines the prototypes for fashion and beauty. During the past two decades, thinness for women and strong physic for men have been fashionable. These stereotypes have been perpetuated mainly via publicity. The adolescent is constantly bombarded with information regarding the ideal weight and figure, how to have the perfect body, what type of exercise one must practice to achieve this perfect body, and miracle diets.

EVALUATION Multiple organ system complications are seen, including those involving the cardiovascular and peripheral vascular systems and gastrointestinal, hematological, renal, skeletal, endocrine, and metabolic disorders. These alterations are related not only to the state of malnutrition but also to the conduct of these patients to control their weight. A number of endocrine and metabolic disturbances described in patients with AN indicate hypothalamic dysfunction, including amenorrhea–oligomenorrhea, delayed puberty, hypothyroidism, hypercortisolism, interferon growth

240

Anorexia Nervosa

factor-1 (IGF-1) deficiency, electrolyte abnormalities, hypoglycemia, and hypophosphatemia.

Medical Complications The clinical manifestations of AN are broad, affecting all systems of the organism, and depend largely on whether the form is restrictive or purgative. Some anorexic patients (10–20%) have bulimic tendencies, which mainly include provocation of vomiting, the use of laxatives, and a compulsive increase in physical activity. Cardiovascular problems occur in up to 80% of patients, including bradycardia and hypotension, due to autonomic nervous system imbalances. Electrocardiographic abnormalities may include atrial and ventricular arrhythmias and QTc abnormalities. In addition, changes in myocardial function have been reported with decreases in myocardial tissue mass, mitral valve prolapse, and pericardial effusions. Gastrointestinal complications are also common. AN can cause a decrease in gastrointestinal motility, resulting in chronic constipation. Laxative abuse can lead to cathartic colon syndrome and chronic constipation that is sometimes refractory to treatment. Cases of acute gastric dilatation have been described during the phase of realimentation of extremely affected anorexia patients since gastric emptying of solids is retarded, and that of liquids is retarded in some patients. Esophageal problems include severe esophagitis and even ruptures of the esophagus associated with induced vomiting. Neurological consequences result from severe malnutrition (Fig. 2). Computed tomography and magnetic resonance imaging have demonstrated cortical atrophy and ventricular dilatation. Malnourished patients have greater cerebrospinal fluid volumes and reduced white and gray matter volumes. Abnormalities on computed tomography scans are reversible with refeeding and nutritional recovery.

Biochemical Abnormalities Hematological findings include anemia, leucopenia (relative neutropenia and lymphocytosis), thrombocytopenia, low erythrocyte sedimentation rate, and decreased fibrinogen levels in plasma. The anemia and occasional pancytopenia appear to be due to hypoplasia of the bone marrow, which is filled with a gelatinous mucopolysaccaride. Vomiting results in the loss of sodium, hydrogen, and potassium, causing metabolic alkalosis. The use of

Figure 2 Patient with anorexia nervosa and severe malnutrition.

laxatives provokes the loss of potassium and bicarbonate, which can result in metabolic acidosis. Finally, the use of diuretics can cause increased loss of sodium, potassium, and calcium in the urine, depending on the dose and drug used. Renal complications are present in 7% of these patients and can include a decrease in glomerular filtration, an increase in plasma urea and creatinine levels, electrolyte alterations, edema, and hypokalemic nephropathy. Renal concentration ability is impaired and polyuria may occur. An abnormal response of arginine vasopressin to an osmotic stimulus may be seen. Cellular immune functions are also altered as a consequence of poor nutrition. These include modifications in some immunoglobulin fractions (IgG and IgA and complement factors C3 and C4), low response in the cutaneous delayed hypersensitivity test, and alterations in the lymphocyte subpopulations CD3, CD4, and CD57. However, infections are infrequent in these patients. Plasma protein levels are usually normal, although in some cases hypoalbuminemia is present. Elevated amylase has been observed in the absence of clinical signals of pancreatitis. Predisposition factors are duodenal and jujenal dilation, and these are more frequent in patients with bulimia. There is an increase in

241

Anorexia Nervosa

ß-carotenes and vitamin A, although the reason for this is unclear. In contrast, blood levels of copper and zinc are reduced and iron and ceruloplasmin levels are normal. Mild hypercholesterolemia is frequent in AN, with the elevation occurring in the low-density lipoprotein fraction, whereas both high-density lipoprotein and very low-density lipoprotein levels are normal. Plasma triglyceride levels are normal despite low values for hepatic lipase and lipoprotein lipase activities. The cause of the hypercholesterolemia is not known.

DIFFERENTIAL DIAGNOSIS A differential diagnosis must be made in situations characterized by loss of weight in young persons, such as brain tumors and lymphomas, and in cases with gastrointestinal symptoms, such as chronic inflammatory disease. In addition, different endocrine pathologies, such as Addison’s disease, hyperthyroidism, and diabetes mellitus, must be taken into consideration. Depending on the predominant psychopathological symptoms, depression and obsessive–compulsive alterations, social phobia, and schizophrenia must be ruled out.

neurotransmitters are primary or secondary to malnutrition remains to be elucidated. Malnutrition may be responsible for the delayed puberty and reduction in growth seen in these patients. This phenomenon has been interpreted as a mechanism of adaptation to the reduction in nutrients. Delayed puberty is present when symptoms appear in prepubertal patients; in contrast, if the disease begins after development has begun, puberty is detained and the growth spurt delayed and smaller. Finally, if symptoms appear after puberty, secondary amenorrhea is present. One of the indications that the process of adaptation to malnutrition has begun is hypoinsulinemia, present as a consequence of low glucose and amino acid levels. Furthermore, growth hormone (GH) abnormalities and low IGF-1 levels contribute to poor growth in prepubertal patients,

Table II

Endocrine Changes in Anorexia Nervosaa

Hypothalamic–pituitary axis LH + FSH + GH +* IGF-1 + IGFBP-1 * IGFBP-3 +

COMPLICATIONS

GHBP +

Endocrine and Neurotransmitter Disturbances

Corticotropin (+ response to CRF) Prolactin , +

Hypothalamic–Pituitary–Ovarian Axis Patients with AN exhibit isolated hypogonadotropic– hypogonadism of hypothalamic origin. The etiology is uncertain, although multiple factors may play a role, including hypothalamic dysfunction, reduction of weight, sex steroids and neurotransmitters alterations, as well as physical exercise (Table II). Adolescents with AN exhibit low basal levels of both luteinizing hormone (LH) and follicle-stimulating hormone (FSH) as well as low estradiol levels, indicating the abnormal function of the hypothalamic–pituitary–gonadal axis. In addition, spontaneous secretion of LH during a 24-h period is decreased in both frequency and amplitude of the secretory bursts. With weight gain, serum levels of both LH and FSH are increased, suggesting that malnutrition may play a role in the regulation of gonadotropin secretion. Disturbances in neurotransmitters have been described, including the dopaminergic system and endogenous opioids, both peptides related to GnRH regulation. Whether these alterations in

Thyrotropin (delayed response to TRH)

ADH: abnormal regulation Thyroid gland T4 + T3 + Reverse T3 * Adrenal gland Cortisol , * Urinary free cortisol * Abnormal dexamethasone suppression Ovary Estradiol + Estrone + Progesterone + Testis Testosterone + a

Abbreviations used: normal, ,; decreased, +; increased, *; LH, luteinizing hormone; FSH, follicle-stimulating hormone; LHRH, luteinizing hormone-releasing hormone; IGF-1, insulin-like growth factor-1; IGFBP-1, insulin-like growth factor-binding protein type 1; IGFBP-3, insulin-like growth factor-binding protein type 3; GHBP, growth hormone-binding protein; TRH, thyrotropinreleasing hormone.

242 leading to a reduction in their final height. Reliable data are not available to establish the percentage of growth lost in relation to the target height. However, one study estimated that some patients were 3 cm shorter than their target height. Whether leptin is a permissive factor or plays a central role in the initiation of puberty is unknown. In subjects with reduced fat stores, problems with reproductive system functioning are frequent, including a reduction in serum sex steroid levels. A similar phenomenon, the shutdown of the hypothalamic– pituitary–gonadal axis, occurs in patients with anorexia nervosa after the loss of fat stores. In both cases, it is speculated that the problems with gonadal function could be related to the decreased serum leptin levels as a result of the loss of fat tissue. We have observed that in patients with AN during partial weight recovery, there is no significant increase in leptin levels or in the recovery of gonadal function. Of course, it is possible that a longer weight recovery period is necessary for this to occur, which would result in loss of the linear correlation with the body mass index (BMI). Studies suggest that BMI is the most important control factor for the secretion of leptin in situations of modified food intake and that there is a loss in circadian rhythm in patients with AN. Of great interest is whether leptin is necessary for the recovery of menstruation in these patients. However, more studies are necessary to answer this question. There is evidence of an association between melatonin levels and gonadal function in humans, with woman with hypothalamic amenorrhea having elevated nocturnal levels of melatonin. Patients with AN are also reported to have elevated nocturnal levels of 6-sulfatoximelatonin (the principal metabolite excreted in the urine) both at diagnosis and after weight recovery if amenorrhea exists. The percentage of total body fat can be evaluated by using dual-energy X-ray absorptiometry. In patients with AN and moderate malnutrition, the percentage of total body fat is a better indicator of the nutritional state than BMI. There is a significant correlation between leptin levels and the percentage of total body fat that is not found between BMI and leptin. It has been reported that if patients with AN recuperate weight to obtain at least 90% of the weight adequate for their height, their menstrual cycles will return within 6 months. It therefore follows that one of the decisive factors for the normalization of gonadal function is the recuperation of the nutritional state.

Anorexia Nervosa

Hypothalamic–Pituitary–GH Axis Most studies indicate that a large percentage of patients with AN have elevated basal and GH-releasing hormone (GHRH)-stimulated GH levels. There are alterations in the GH response to different stimuli (decreased response after hypoglycemia, clonidine, and hexarelin) as well as paradoxical hormone responses [elevated GH after thyrotropin-releasing hormone (TRH) thyrotropin or after intravenous glucose], although these responses are heterogeneous. Few studies have analyzed spontaneous GH secretion (SGHS) in AN patients. We studied SGHS in a group of anorexic patients at diagnosis and at two different times during weight recovery and found that at diagnosis SGHS is heterogeneous. In 40% of subjects, mean 24-h GH secretion was >3 ng/ml (lower limit of normality), and the remaining 60% had levels below the normal range. The difference between these groups and the controls was due to modification in the amplitude of the GH peaks and not to pulse frequency. In both groups, recovery of at least 10% of initial weight resulted in the normalization of the parameters of SGHS. These observations suggest that alterations in GH secretion in these patients are due to modifications in its neuroendocrine control, with an increase in GHRH release and decreased somatostatin tone. The GH pattern in conjunction with the negative correlation between basal and pulsatile GH secretion and BMI suggests that the observed alterations in GH secretion are directly related to malnutrition. One possible mechanism may involve the reduced IGF-1 levels caused by malnutrition. This would effect a reduction in the negative feedback action that IGF-1 exerts on GH secretion at the level of the hypothalamus and pituitary. Another variable that may be involved is the hypoestrogenism that accompanies amenorrhea. It has been suggested that malnutrition underlies the increase in the amount of GH secreted in each pulse and that hypoestrogenism is responsible for the increased pulse frequency. Definite conclusions, however, can not be drawn from these studies. In patients with AN, plasma ghrelin levels are significantly increased and rapidly normalize after partial weight recovery. Whereas decreased leptin levels in AN patients might simply reflect reduced body fat mass, increased gastric ghrelin secretion in AN might reflect a physiological effort to compensate for the lack of nutritional intake and stored energy. Serum GH-binding protein (GHBP) levels in patients with AN are dramatically reduced and tend to normalize with weight recovery. The reduction in GH

Anorexia Nervosa

receptors is most likely one of the principal causes of GH resistance. In malnutrition, the low GHBP levels may be related to hypoinsulinemia, alterations in thyroid function, or hypoestrogenism. On the other hand, many studies have demonstrated a correlation between serum GHBP levels and BMI or the percentage of body fat or, more specifically, visceral fat. Given that it has not been demonstrated that circulating GHBP is uniquely or even preferentially derived from liver GH receptors, it is possible that other tissues, such as adipose tissue, may contribute to plasma GHBP levels. If this is the case, the extreme reduction in adipose tissue in patients with AN may cause the observed decrease in serum GHBP levels. Patients with AN have extremely reduced serum IGF-1 levels that tend to normalize with weight recovery; however, as observed for other forms of malnutrition, the time necessary for this to occur may be prolonged. Circulating IGF-1 is largely dependent on GH, but it is also very sensitive to nutritional changes. In AN patients, serum IGF-1 levels do not correlate with GH secretion, which suggests that the decrease in IGF-1 is independent of GH and probably directly due to the state of malnutrition. The coexistence of reduced IGF-1 levels and normal or elevated GH secretion suggests that these patients exhibit resistance to GH action. Data on serum levels of free IGF-1 are limited and contradictory. Some authors have found normal levels, whereas others report a decrease; in both cases, weight recovery tended to increase free IGF-1 levels. Likewise, some authors have found serum IGF-2 levels to be normal at the time of diagnosis and to increase with weight recovery. In contrast, others report IGF-2 levels to be decreased, although not significantly, and to normalize with weight recovery. Serum levels of free IGF-2 are reported to be decreased in patients with AN and to increase with weight recuperation. Patients with AN have elevated serum IGFBP-1 and IGFBP-2 levels that tend to normalize with weight recovery. Both are reported to be GH independent and very sensitive to nutritional regulation. The increase in IGFBP-1 in these patients is most likely related to hypoinsulinism, although other metabolic or hormonal factors, such as increased glucagon and glucocorticoids levels, as well as decreased intracellular glucose or other specific substrates may be involved. In AN, as in other forms of malnutrition, the increase in IGFBP-2 most likely depends on the combined influence of caloric–protein restriction, hypoinsulinism, and GH resistance.

243 Serum IGFBP-3 levels are decreased in AN patients as a consequence of GH resistance and tend to normalize after weight recovery. Indeed, all components of the trimolecular complex formed by the union between IGFBP-3, IGF, and the acid-labile subunit (ALS) are decreased. Given that these proteins are all GH dependent and regulated by the nutritional state, this is not unexpected. IGFBP-3 decreases significantly with caloric restriction, but in adults it decreases only with protein restriction. In contrast to other catabolic situations, in AN increased proteolysis of IGFBP-3 has not been observed. Although present in serum in low concentrations, IGFBP-4 and IGFBP-5 are very important in bone formation: At the cellular level, they regulate the actions of the IGFs. In AN, serum levels of both of these IGFBPs are dramatically reduced and do not normalize with partial weight recovery. The biological significance of the changes in IGFBPs that occur in AN or that are due to malnutrition is difficult to explain, in part because the physiological roles of the binding proteins are not totally understood. The decrease in serum IGFBP-3, and therefore the trimolecular complex IGFBP-3–IGF– ALS, impedes the retention of IGF in the vascular space, favoring the decrease in plasma levels. On the other hand, the increase in IGFBP-1 and IGFBP-2, two proteins with low molecular weights that can cross the vascular barrier, favors even more the movement of the IGFs to the tissues on which they can act. Therefore, modification of the serum and tissue levels of the IGFBPs could be one of the mechanisms by which malnutrition regulates the concentration and actions of the IGFs and the somatotrope axis. Hypothalamic–Pituitary–Thyroid Axis Thyroid function is affected by malnutrition in AN. Clinically, patients appear to be in a relative hypothyroid state, sometimes called euthyroid sick syndrome. Clinical manifestations include hair loss, dry skin, hypothermia, and bradycardia. All these findings are reversible with appropriate refeeding and successful treatment. Laboratory findings include low-normal levels of thyroxine (T4) and thyroid-stimulating hormone (TSH), below normal levels of triiodothyronine (T3), and elevated levels of reverse T3. All these findings are the result of malnutrition and weight loss. In fact, the low T3 level correlates with the amount of weight loss. The extremely reduced T3 levels in these patients are due to altered peripheral deiodination that preferentially transforms T4 into the inactive metabolite, reverse T3. Finally, blunting of the TSH response to

244 TRH is indicative of hypothalamic–pituitary dysfunction. Ultrasonographic methods in patients with AN demonstrate that the thyroid is markedly reduced in comparison to that of age- and sex-matched controls. This glandular atrophy is not due to low TSH levels because TSH levels are usually normal in AN. However, thyroid size is influenced by IGF-1, and the low IGF-1 levels in these patients may contribute to thyroid atrophy. These alterations normalize with weight recovery and thyroid hormone replacement is usually not indicated. Hypothalamic–Pituitary–Adrenal Axis Adrenocorticotropin (ACTH) and the endogenous opioids are derived from the same precursor, POMC. Increased ACTH secretion is preceded by activation of the POMC system. The opioid system has both direct and indirect influences over food intake and the level of physical activity. In laboratory animals, opioid administration stimulates appetite via receptors in the paraventricular nucleus, and opioid antagonists such as naloxone reduce appetite. Corticotropin-releasing hormone (CRH), synthesized in neurons of the paraventricular nucleus, is regulated, at least in part, by leptin and insulin. Its administration intracerebroventricularly induces a reduction in food intake and weight loss. In AN, plasma cortisol levels are often elevated while circadian rhythm is conserved. Dexamethasone can partially suppress this hypercortisolemia, which is similar to what is observed in patients with depression and Cushing’s disease. In acute situations of AN, the dexamethasone test has no medical significance; however, in patients who are gaining weight, it may have prognostic value. Refeeding studies of anorexic patients have shown that, irrespective of the initial weight, weight gains as low as 10% are associated with the normalization of cortisol secretion. The mean plasma half-life of cortisol is prolonged and ACTH levels are within the normal range, but the ACTH response to CRH is inferior to that of control patients. This hypercortisolemia may reflect a defect at the level of the hypothalamus, or even higher, that results in hypersecretion of CRF. Taken together, these observations suggest CRH hypersecretion more than cortisol resistance.

Osteoporosis or Osteopenia Bone Mineral Density Osteopenia is a frequent and often persistent complication in patients with AN, leading to clinical

Anorexia Nervosa

fractures and increased fracture risk through life. According to the international literature, more than 50% of patients with AN present osteopenia at diagnosis. The pathogenesis of osteopenia and osteoporosis is not completely known; however, numerous studies have highlighted a number of factors, including severe malnutrition, poor calcium intake, excessive exercise patterns despite malnutrition, hypoestrogenism, increased serum cortisol levels, and hormonal imbalances, such as decreased progesterone levels and decreased IGF-1 levels. During infancy and puberty, obtaining adequate bone mineralization requires normal nutrition and metabolic and endocrine functions and the absence of chronic pathologies. Hormones and growth factors that play a major role in regulating bone metabolism include T3, sex steroids, vitamin D, paratohormone, and GH, together with the IGF–IGFBP system and an intricate local system of growth and transcriptional factors. The degree of osteopenia possibly depends on the age at which the amenorrhea began as well as its duration. Indeed, patients with primary amenorrhea have a more severe osteopenia than those that present with secondary amenorrhea. There are significant concerns about the lasting impact and irreversibility of osteoporosis. Evaluation of bone density is recommended in patients who have been amenorrheic for 6 months to 1 year. Investigation of the changes in bone mineral density (BMD) showed that nonrecovered AN patients with the binge eating/purging type have a significantly reduced BMD compared to patients with the restricting type. These results suggest that patients with the binge eating/purging type are at high risk for osteoporosis and may need additional therapy to prevent bone loss. Karllson et al. reported that a substantial proportion of bone mass deficit in anorexic patients was due to smaller bone size. Recovery from illness was associated with near normal bone size and volumetric BMD. However, incomplete recovery of lean and fat mass may account for part of the remaining deficit in bone size but not volumetric BMD. Administration of estrogens and gestagens to adolescents with reduced bone mass and amenorrhea for at least 1 year indicated that osteopenia cannot be reversed. In contrast, in patients who recovered menstruation spontaneously, a 20% increase in bone mass compared to that at diagnosis was seen. The effects of estrogens on bone metabolism have been described as inhibitory for the resorption

Anorexia Nervosa

process, although direct effects on osteoblastic activity have been described. It has been reported that AN occurring during adolescence impairs both mineral accrual and bone size. Although reduced volumetric BMD may be related to estrogen deficiency, there was no reduced bone size after adjusting for fat and lean mass. Weight, but not estrogen use, is a significant predictor of BMD in anorexic women at all skeletal sites. The reason why estrogens are incapable of increasing bone mass in adolescents with AN and amenorrhea is unknown. In may be due to the failure to administer estrogen therapy at diagnosis, poor compliance, or perhaps the short duration of recovery. In addition, decreases in other nutritional and hormonal factors are also involved in the pathogenesis of bone mass loss; hence, estrogen replacement alone may not be sufficient for BMD recovery. IGF-1 and Leptin IGF-1 is one of the most important regulators of bone metabolism. Circulating serum levels of IGF-1 correlate with BMD in the normal population. IGF-1 exerts a double effect on bone metabolism by stimulating osteoblastic activity and the resorption process. Deficiency in growth factors, especially IGF-1, most likely due to the state of malnutrition, as well as the slow recuperation of their plasma levels with weight gain, is known to occur in these patients. However, it is not known whether these patients would benefit from the administration of biosynthetic GH or recombinant IGF-1. Several trials have analyzed the effect of recombinant human IGF-1 (rhIGF-1) on bone formation in AN patients. The administration of rhIGF-1 at a dose of 100 mg/kg subcutaneously (sc) twice a day for 6 days increased metabolic markers of bone formation as well as resorption, whereas the injection of rhIGF-1 at a dose of 30 mg/kg sc per day stimulated only bone production formation marker. Improvement in nutritional status in AN patients via intravenous hyperalimentation therapy results in a rapid increase in serum IGF-1 levels, followed by a progressive increase in osteocalcin. This indicates that bone formation begins immediately. Nevertheless, increased bone resorption appears to continue for at least 5 weeks. The effect of leptin administration compared to estrogen therapy in ovariectomy-induced bone loss in rats has also been reported. Leptin was effective in reducing trabecular bone loss, trabecular architectural changes, and periosteal bone formation. These findings suggest that leptin may regulate bone remodeling, and this effect may be, at least in part, mediated

245 by the osteoprotegerin/RANK (receptor for activation of nuclear factor kappa B) ligand pathway. RANK and RANK ligand (RANKL) are members of the tumor necrosis factor (TNF) and TNF receptor superfamilies, which are essential for osteoclast differentiation. In the bone microenvironment, the stimulatory effects of RANKL are neutralized by the secreted decoy receptor, osteoprotegerin (OPG). It follows that the balance between OPG and RANKL secretion by stroma cells is critical for the regulation of osteoclast formation. The dramatic decline in leptin levels observed in AN may be one of the major hormonal factors involved in the pathogenesis of the associated bone fragility through diminishing cortical bone formation rates and skeletal growth. Leptin may play an important protective role in bone metabolism by inhibitory bone resorption. Bone Markers The available data are few and contradictory regarding markers of bone formation. The bone isoenzyme of alkaline phosphatase (bAP) and the amino-terminal pro-peptide of procollagen I (PNIP) have the greatest diagnostic sensitivity in detecting anomalies in bone remodeling, at least in osteoporetic women. Among the markers of bone resorption, the telopeptide carboxy terminal of the 1 chain of type 1 collagen (CTX) has demonstrated great specificity and sensitivity in the investigation of bone metabolism. Osteoporosis in adolescents with AN is related to decreased bone formation and actual bone resorption. It is clear that markers of bone formation are decreased and markers of bone mineral resorption are increased. Trabecular bone seems to be more vulnerable than cortical bone. A significant positive correlation of BMI, IGF-1, and IGFBP-3 with osteocalcin as a bone formation marker was demonstrated, whereas a negative correlation for the bone resorption marker CrossLaps was found only with BMI. Patients with AN exhibit a high degree of osteopenia that correlates with their bAP levels. Furthermore, the urinary fragments of CTX in patients with AN are derived primarily from old bone (b-CTX), whereas in young adolescents they are primarily from new bone (a-CTX). Therefore, a-CTX is more adequate for measuring bone resorption in controls, whereas b-CTX is more adequate in anorexic patients. One study demonstrated that during undernutrition and amenorrhea, with low IGF-1 and extremely low circulating estradiol, biochemical markers of bone metabolism indicated a shift toward

246 bone resorption because bone formation markers (osteocalcin, bAP, and PNIP) were normal and bone resorption markers (CTX) elevated. However, in fully recovered patients, bone metabolism markers indicate accelerated bone turnover with an increase in both formation and resorption. Bone formation markers correlated positively with IGF-1 and bone resorption markers negatively with estradiol, indicating that IGF-1 is a major bone formation stimulator, whereas estradiol action predominately inhibits bone resorption. In patients with AN, the mechanism of bone loss does not appear to be due to an increase in absorption over formation. It is possible that the observed increase in bone remodeling is a mechanism developed in an attempt to restore bone mass. However, the large deficit of calcium in these patients (the loss of exogenous sources due to the deficit in alimentation produces the liberation of bone calcium to maintain the homeostasis of the extracellular fluid) and the deficit in amino acids as a result of fasting make it very difficult to restore bone mass. The best predictors for osteopenia are BMI and the duration of amenorrhea, followed by the duration of regular menses before amenorrhea. Treatment protocols emphasize the importance of nutritional rehabilitation and recovery. The addition of 1500 mg/day of calcium with vitamin D is recommended and lifestyle counseling with an emphasis on smoking cessation is relevant. The issue of exercise is controversial; for most women, including teens, weight-bearing exercise is an important factor contributing to bone mineral accretion and is more important than calcium intake in adolescent bone health. Hormone replacement therapy with oral contraceptives has been studied as a treatment for osteopenia/osteoporosis; results have been mixed. Some preliminary studies have demonstrated a minor protective effect on bone density of the lumbar spine, whereas others have not.

TREATMENT An integrated treatment program should be instituted and carried out by a multidisciplinary team, including a pediatrician, endocrinologist, psychiatrist, psychologist, nurse, and possibly others. It is very important that the correct diagnosis is made and that the patient and family are made aware of the importance of the disease and the various aspects of treatment. They should be aware that treatment will necessarily last for a period not less than 5 years.

Anorexia Nervosa

The role of the physician in the control of this process and the establishment of an adequate relationship with the patient and family is fundamental for successful treatment. Treatment should be on an outpatient basis if the diagnosis was made early, the degree of undernutrition is not too severe, and the mental disturbance is not incapacitating. If this is not the case, the patient must be hospitalized. Treatment objectives should have a strict priority: prevent the death of the patient, prevent the disease from becoming chronic, and start physical and mental recuperation.

Nutritional Treatment Refeeding To successfully begin refeeding, it is fundamental that a therapeutic alliance with the patient be established in which the patient understands and accepts that he or she has a disease. This can be accomplished by asking the patient if he or she has the various signs and symptoms of AN so that, little by little, the patient can identify with this condition. The physician should use height and weight graphs to explain what percentile the patient is currently in and what the patient’s weight should be for his or her age and height. A target weight, acceptable to both the patient and the physician, should be agreed upon. The patient’s understanding of caloric requirements to maintain a normal weight should be explored because very few possess knowledge of the appropriate requirements for their height, age, and sex. Patients must understand that their growth and physical activity depend on the adequate intake of calories, including proteins, fats, carbohydrates, and vitamins and minerals. Obtaining and Maintaining an Adequate Weight After the target weight has been achieved, a maintenance diet must be prescribed, and intake of foods not on the diet and eating between scheduled meals must be prohibited. This helps the patient overcome the fear of losing control or gaining weight.

Psychological Treatment Psychotherapeautic help can consist of individual, familial, or group treatment. In AN, the initial phases of psychotherapy consist of helping the patient begin refeeding by alleviating the feelings of guilt of the patient or family. The next step is to treat the overall

247

Anorexia Nervosa

psychological well-being of the patient such that eventually the medical problems become secondary. It is fundamental to work with the family using familial therapy or counseling.

Pharmacological Treatment Psychotropic medication should not be used as the only or principal treatment but, rather, as a prevention of relapse in patients who have gained weight or for treatment of symptoms associated with AN, such as depression or obsessive–compulsive behavior. A variety of substances have been used to treat AN because it is basically unresponsive to pharmacological treatment.

PROGNOSIS AND EVOLUTION The normal evolution of AN consists of cycles of recuperation and relapse even in the best controlled cases, and the evolution rarely lasts less than 2 or 3 years. Approximately 50% of AN patients achieve total recuperation, whereas 20% have residual problems and 30% become chronic. The mortality rate is between 0.5 and 1% per year of observation. The most frequent causes of death are severe undernutrition, gastrointestinal complications, infections, and suicide. With prolonged follow-up, the mortality rate increases slightly (up to 20% for patients older than 20 years of age).

CONCLUSION The Academy Pediatric Association recommends a multidisciplinary treatment. The team should include a medical physician, dietician, and psychiatrist/psychologist. Evidence suggests that there is a hypothalamic dysfunction in patients with AN, and in general this normalizes with weight recovery. Disturbances in neurotransmitter, neuropeptide, and neuroendocrine systems have been reported in acutely ill patients and in patients during follow-up. Due to scarce macro- and micronutrients, intact nonvital processes that increase energy output and are not necessary for survival, such as growth, pubertal development, and reproduction, are shut down until the nutritional situation improves. Early implementation of appropriate psychological and nutritional therapy and the best treatment for osteopenia/ osteoporosis in these patients is necessary. The subtypes of AN and BMI at follow-up are the best

predictors of bone mineral density, and leptin may play an important protective role in bone metabolism. Analysis of the genetic mechanisms underlying weight regulation is progressing rapidly. The genetic analysis of AN may help to define new drug targets and therefore lead to new treatment strategies. The prognosis for adolescents with eating disorders is favorable. With early and aggressive treatment, most adolescents with AN recover, as demonstrated by a number of long-term follow-up studies at adolescent medicine centers.

See Also the Following Articles Body Proportions . Constitutional Delay of Growth and Puberty (CDGP) . Eating Disorders and the Reproductive Axis . Obesity, Childhood and Adolescence . Obesity Regulation . Osteoporosis, Overview . Puberty: Physical Activity and Growth

Further Reading American Psychiatric Association. (1994) “Diagnostic and Statistical Manual of Mental Disorders.” American Psychiatric Association Washington, DC. Argente, J., Barrios, V., Chowen, J. A., Sinha, M. K., and Considine, R. V. (1997). Leptin plasma levels in healthy Spanish children and adolescents, children with obesity and adolescents with anorexia nervosa and bulimia nervosa. J. Pediatr. 131, 833–838. Argente, J., Caballo, N., Barrios, V., Mun˜oz, M. T., Pozo, J., Chowen, J. A., Morande´, G., and Herna´ndez, M. (1997). Multiple endocrine abnormalities of the growth hormone and insulin-like growth factor axis in patients with anorexia nervosa: Effect of short- and long-term weight recuperation. J. Clin. Endocrinol. Metab. 82, 2084–2092. Audı´, L., Vargas, D. M., Gussinye´, M., Teste, D., Martı´, G., and Carrascosa, A. (2002). Clinical and biochemical determinants of bone metabolism and bone mass in adolescent female patients with anorexia nervosa. Pediatr. Res. 51, 497–504. De la Piedra, C., Calero, J. A., Traba, M. L., Asensio, M. D., Argente, J., and Mun˜oz, M. T. (1999). Urinary and C-telopeptides of collagen I: Clinical implications in bone remodeling in patients with anorexia nervosa. Osteop. Int. 10, 480–486. Ducy, P., Ambling, M., Takeda, S., Priemel, M., Schilling, A. F., Beil, F. T., et al. (2000). Leptin inhibits bone formation through a hypothalamic relay: A central control of bone mass. Cell 100, 197–207. Gorwood, P., Ades, J., Bellodi, L., Cellini, E., Collier, D. A., DiBella, D., et al. (2002). The 5-HT (2A)-1438 G/A polymorphism in anorexia nervosa: A combined analysis of 316 trios from six European centres. Mol. Psychiatr. 7, 90–94. Karlsson, M. K., Weigall, S. J., Duan, Y., and Seeman, E. (2000). Bone size and volumetric density in women with anorexia nervosa receiving estrogen replacement therapy and in women recovered from anorexia nervosa. J. Clin. Endocrinol. Metab. 85, 3177–3182.

248 Mun˜oz, M. T., and Argente, J. (2002). Anorexia nervosa: Hypogonadotrophic hypogonadism and bone mineral density. Hormone Res. 57(Suppl.), 57–62. Mun˜oz, M. T., Morande´, G., Garcı´s, F., Pozo, J., and Argente, J. (2002). The effects of estrogen administration on bone mineral density in adolescents with anorexia nervosa. Eur. J. Endocrinol. 46, 45–50. Otto, B., Cuntz, U., Fruehauf, E., Wawarta, R., Folwaczny, C., Riepl, R. L., et al. (2001). Weight gain decreases elevated plasma

Anorexia Nervosa

ghrelin concentrations of patients with anorexia nervosa. Eur. J. Endocrinol. 145, R5–R9. Pozo, J., and Argente, J. (2002). Delayed puberty in chronic illness. Best Pract. Res. Clin. Endocrinol. Metab. 16(1), 73–90. Støving, R. K., Flyvbjerg, A., Frystyc, J., Fisker, S., Hangaard, J., Hansen-nord, M., and Hagen, C. (1999). Low serum levels of free and total insulin-like growth factor I (IFG-I) in patients with anorexia nervosa are not associated with increased IGF-binding protein-3 proteolysis. J. Clin. Endocrinol. Metab. 84, 1346–1350.

Anti-Mu¨llerian Hormone Nathalie Josso and Nathalie di Clemente Unite INSERM 693, Clamart, France

Glossary activin signaling factor-b signaling genetic (AMH)].

receptor-like kinases (ALKs) ALKs act as (type I) receptors for the transforming growth (TGF-b) family. They may be involved in the pathway of several ligands [e.g., bone morphoproteins (BPMs) and anti-Mu¨llerian hormone

bone morphogenetic proteins (BMPs) Members of the TGF-b family; they are involved in many morphogenetic processes, namely cartilage and gonadal differentiation. GATA-4 Members of the GATA family of proteins (GATA1–6) recognize a consensus sequence on DNA due to conserved zinc fingers in their DNA-binding domain. They play important roles in erythroid cell differentiation. GATA-4 is abundantly expressed in the Sertoli cells of the differentiating testis and binds to the AMH promoter through a conserved proximal element. high-mobility group (HMG) A 79-amino acid DNAbinding domain shared by SOX proteins and SRY. g0025

Mu¨llerian ducts Precursors of the uterus, Fallopian tubes, and upper part of the vagina; the main target organ of AMH.

A

nti-Mu¨llerian hormone, also called Mu¨llerian inhibiting substance, is a member of the transforming growth factor-b family, whose main effect is the initiation of regression of the Mu¨llerian ducts in the male fetus.

INTRODUCTION Anti-Mu¨llerian hormone (AMH), also known as Mu¨llerian inhibiting substance, is a ‘‘sexually specialized’’ member of the transforming growth factorb (TGF-b) family in the sense that it is expressed exclusively by gonadal somatic cells and that its effects are targeted to the gonads and reproductive tract. Its main biological effect, the inhibition of Mu¨llerian duct development in male fetuses, represents the first

Encyclopedia of Endocrine Diseases, Volume 1. ß 2004 Elsevier Inc. All rights reserved.

Smad proteins Smad proteins are a family of intracellular signal mediators for members of the TGF-ß family. They are classified into three broad categories. Receptor-specific Smads bind specifically either to type I receptors of TGF-b or activin (Smads 2 and 3) or to members of the BMP family (Smads 1, 5, and 8). After activation by their specific type I receptor, they bind to a co-Smad (Smad-4) and enter the nucleus to modulate gene transcription. Inhibitory Smads, Smads 6 and 7, prevent the activation of receptor-specific Smads. SOX-9 Members of the SOX gene family are critical transcription factors for many developmental processes. SOX-9 is a key regulator of cartilage and testis development; mutations in humans lead to a malformative syndrome known as campomelic dysplasia and to gonadal aplasia in XY subjects. SOX-9 is a powerful activator of AMH transcription. steroidogenic factor-1 (SF-1) Also known as Ad4BP, SF-1 is an orphan nuclear receptor that activates AMH transcription. It also activates P450 steroid hydroxylases and aromatase. In its absence, gonads, adrenals, the pituitary, and the hypothalamus do not develop normally.

step of phenotypic male sex differentiation. Mu¨llerian ducts are sensitive to AMH only during a short developmental window, before 8 weeks in the human fetus. The existence of a specific factor dedicated to Mu¨llerian regression was predicted in 1953 by Alfred Jost, who proposed that the fetal testis has a dual hormonal secretion: testosterone, responsible for the virilization of Wolffian ducts and external genitalia, and a second hormone, responsible for Mu¨llerian regression (Fig. 1), that was identified and purified only many years later. AMH inhibits meiotic prophase maturation and the growth of the fetal ovary, in which it sometimes induces the differentiation of structures resembling seminiferous tubules. AMH also has effects postnatally. In the male, it inhibits the differentiation and function of Leydig cells; in

255

256

Anti-Mu¨llerian Hormone

Figure 1 (Right) The role of AMH in fetal sex differentiation. AMH, secreted by fetal Sertoli cells, binds to an AMH receptor located on the membrane of target organs and induces the regression of Mu¨llerian ducts, which would otherwise give rise to the uterus. (Left) The effect of testosterone secreted by Leydig cells. After binding to the androgen receptor, testosterone induces the differentiation of male excretory ducts and masculinizes the penis and urogenital sinus after reduction to DHT.

the female, it inhibits follicular maturation and breast cancer cell growth. However, the postnatal effects of AMH are not crucial because testicular or ovarian functions are not disrupted by mutations of AMH or its receptors.

BIOCHEMISTRY AND MOLECULAR BIOLOGY OF AMH AND ITS RECEPTORS AMH is a 560-amino acid glycoprotein formed by two 70-kDa monomers linked by disulfide bonds. The hormone is cleaved at a proteolytic site 109 amino acids upstream of the C terminus, yielding a short bioactive C-terminal domain with homology to other members of the TGF-ß family and a long N-terminal prodomain with no bioactivity of its own but which enhances the bioactivity of the C terminus. The human AMH gene was cloned in 1986. It is located on the tip of the short arm of chromosome 19, band p13.3. Only 2.75 kilobase pairs (kbp) in length, it contains five exons; the 30 -end of the fifth one is particularly GC rich and codes for the bioactive Cterminal domain. The 200-bp minimal promoter is flanked by a household gene, SAP-62, coding for a spliceosome, and contains binding sites for various transcription factors (Fig. 2). In keeping with its status as a TGF-b family member, AMH signals through two membrane bound receptors, both with serine/threonine kinase activity. Paradoxically, receptor type II is the primary

receptor; it binds to AMH and then recruits the signaling, so-called type I, receptor. The gene for the type II receptor, AMHR-II, cloned in 1994, is AMH specific; it is divided into 11 exons, has a molecular mass of 82 kDa, and maps to chromosome 12q13. Based on evidence obtained from different tissues and cell lines, it is believed that three type I receptors—ALK2,-3, and-6—involved in the signaling pathway of the bone morphogenetic protein (BMP) family also function as AMH type I receptors. As expected, all three signal through the BMP receptor-specific Smad molecules 1, 5, and 8 (Fig. 3).

Figure 2 The AMH gene. The 30 -end of exon 5 (hatched) codes the 109-amino acid biologically active C terminus. The 200-kbp promoter contains binding sites for transcription factors and is adjacent to a housekeeping gene, SAP62. AMH maps to chromosome 19 p13.3.

257

Anti-Mu¨llerian Hormone

Figure 3 AMH signal transduction. AMH binds to autophosphorylating receptor type II (AMHR-II), which then recruits a type I receptor, ALK 2, 3, or 6. The complex transiently binds and phosphorylates a receptor-specific Smad (1, 5, or 8), which then binds to a common Smad (co-Smad), Smad 4. The receptor-specific and co-Smad then enter the nucleus and bind to the promoter of target genes with the help of cofactors and transcription factors.

In the male, AMH is produced in large amounts by Sertoli cells just prior to the differentiation of fetal seminiferous tubules until puberty (Fig. 4). In sexually mature individuals, Sertoli cells continue to synthesize small amounts of AMH; after puberty, however, the hormone is preferentially secreted toward the lumen of the seminiferous tubule. Because of the formation of the blood–testis barrier, AMH concentration is higher in the seminal plasma than in serum. AMHR-II is expressed in the mesenchymal cells surrounding the fetal Mu¨llerian duct, in Sertoli cells, and, to a lesser degree, in Leydig cells. In the female, low amounts of AMH and AMHR-II are coexpressed by granulosa cells of preantral and small antral follicles. Expression decreases slightly as follicle maturation progresses. When follicles disappear from the ovary at the end of reproductive life, AMH is no longer detectable in serum. The chronological expression of AMH is tightly regulated. In the male, it is essential that the hormone be produced during the short window of Mu¨llerian responsiveness. Several transcription factors—SOX-9 (a member of the high-mobility group gene family), steroidogenic factor-1, and GATA-4, all of which are expressed in the developing testis but not the ovary —bind to response elements on the AMH promoter (Fig. 2) and cooperate to trigger the early initiation of AMH transcription in the testis. After birth, AMH is regulated positively by follicle-stimulating hormone (FSH) and negatively by androgens. During puberty, when Leydig cells begin to produce testosterone, the

expression of AMH by Sertoli cells is repressed, provided normal androgen receptors are present. In their absence, the stimulatory influence of FSH becomes apparent. In the female, initiation of AMH expression begins as soon as follicle maturation reaches the preantral stage, either after birth or in the last weeks of pregnancy. Little is known, however, about the mechanism of regulation.

800 Male

Female

600 (pmol/liter)

EXPRESSION AND REGULATION

400

200

0

o. yr yr yr m 1 4 8 1 o.− 1− 4− − 0 m 1

ND

I

II

III −V dult IV A

Puberty

al

lt e rt du us be A opa u ep en Pr t-m s Po

Figure 4 Serum AMH in postnatal life measured by ELISA. Serum AMH can be measured by ELISA in both sexes. In the human male, serum AMH is high until puberty, except in the first month of life. It decreases slowly until puberty, when its serum concentration is inversely correlated to pubertal maturation. In the adult male, serum AMH is very low or undetectable. In the female, serum AMH is very low and becomes undetectable after menopause.

258

Anti-Mu¨llerian Hormone

AMH IS A MARKER OF GONADAL FUNCTION AMH is produced postnatally by the gonad of both sexes; ascertainment of its serum level is useful to assess gonadal function.

Testicular Function Seminiferous Tubule Function in Prepubertal Subjects AMH is expressed constitutively by prepubertal Sertoli cells. It follows that its circulating level reflects the presence and function of seminiferous tubules, with no need for prior stimulation by chorionic gonadotropin (Fig. 4). Pediatric endocrinologists assay serum AMH in children with impalpable testes to discriminate between anorchia and bilateral cryptorchidism and in patients with ambiguous genitalia to distinguish testicular dysgenesis from disorders affecting testosterone synthesis or sensitivity. Androgen Action Normally, AMH is down-regulated by testosterone produced by Leydig cells during puberty. A significant reduction of serum AMH is observed when testosterone concentration reaches 7 nmol/ng/liter. If testosterone is not synthesized or if the androgen receptor is abnormal, not only does serum AMH fail to decrease but also it increases higher than normal childhood levels because FSH is now produced and stimulates AMH transcription in Sertoli cells. Serum AMH may also be used as a sensitive marker for the effectiveness of medication inactivating the androgen receptor, such as cyproterone acetate. Azoospermia In oligospermic adult males, the level of AMH in seminal fluid is correlated with sperm concentration. In patients with nonobstructive azoospermia, it is higher if spermatozoa are detectable in a testicular biopsy. Thus, it may aid in predicting the success of an intracytoplasmic sperm injection procedure.

Adult Ovarian Function In women, serum AMH reflects the function and number of granulosa cells. Between birth and menopause, serum AMH levels are less than 75 pmol/liter or may be undetectable. This is always the case after natural menopause or ovariectomy (Fig. 5).

Figure 5 Serum AMH in women with or without granulosa cell tumors. In normal women, serum AMH is low ( 80 nm). Histologically, they appear similar with few mitoses and uniform nuclei. Frequently, they synthesize multiple peptides/amines, which can be detected immunocytochemically but may not be secreted. The presence or absence of clinical syndrome or type cannot be predicted by immunocytochemical studies. Histological classifications do not predict biologic behavior. Only invasion or metastases establish malignancy. Similarities of biologic behavior They are generally slow growing, but a proportion are aggressive. They secrete biologically active peptides/amines, which can cause clinical symptoms. They generally have high densities of somatostatin receptors, which are used for both localization and treatment.

p0030

distinguish between benign and malignant NETs unless metastases or invasion are present. PETs are classified according to the specific clinical syndrome they cause or, if no clinical syndrome occurs, as nonfunctional (Table II). Nonfunctional PET in the strict sense is a misnomer because these tumors frequently secrete multiple peptides (pancreatic polypeptide, CgA, and neurotensin); however, they do not cause a specific clinical syndrome. The symptoms caused by nonfunctional PETs are due to the tumor per se. Carcinoid tumors are frequently classified by location of origin (foregut, midgut, or hindgut) because they share histochemical, functional, and biologic characteristics within these areas (Table III). Foregut carcinoids generally have low serotonin (5-HT) content, are argentaffin negative on silver staining, synthesize several hormones, and occasionally secrete ACTH or 5-hydroxytryptophane (5-HTP) causing an atypical carcinoid syndrome. Midgut carcinoids are generally argentaffin positive, have high serotinin content, frequently multihormonal, rarely secrete 5HTP or ACTH, and most frequently cause carcinoid syndrome when they metastasize (Table III). Hindgut carcinoids (transverse colon to rectum) are argentaffin negative; rarely contain serotonin, 5-HTP, and ACTH or cause carcinoid syndrome; contain numerous peptides; and frequently metastasize to bone. Most carcinoids (70%) occur in one of four sites—bronchus, jejunum/ileum, rectum, or appendix—although they can occur in any tissue (Table III). Overall, the GI tract is the most common site for carcinoids (74% of

cases), and the respiratory tract the second most common site (25% of cases). The frequency of NETs varies according to whether they are symptomatic. The incidence of clinically significant PETs is 10 cases/1 million population/year. Their relative incidence is insulinoma, gastrinoma, and nonfunctional (0.5–2.5 cases/1 million/year) > VIPomas (2–8 times less common) > glucagonomas (17–30 times less common) > somatostatinomas. In autopsy studies, 0.5–1.5% of all cases have PETs, with less than 1 in 1000 symptomatic. Clinically significant carcinoids occur in 7–13 cases/1 million/year and malignant cases at autopsy in 21–84 cases/1 million/year. Both PETs and carcinoids can show malignant behavior. With PETs, 50–100% are malignant, except for insulinomas, in which < 10% are malignant (Table II). With carcinoid tumors the percentage that are malignant varies with different locations (Table III). A number of factors are predictive of tumor aggressiveness and/or survival for NETs (Table IV). The presence of liver metastases is the most important prognostic factor. Primary tumor size is also an important predictor. The factors in Table IV need to be considered when determining the aggressiveness of treatment of NETs. Until recently, the molecular pathogenesis of NETs was largely unknown because common oncogenes (ras, fos, etc.) and common tumor suppressor genes (retinoblastoma gene, p53) were usually not altered. Recent studies provide evidence that alteration in the HER2/neu oncogene, MEN-1 gene, and p16INK4a

p0035

t0010

Table II

Gastrointestinal Neuroendocrine Tumor Syndromes Biologically active peptide(s) secreted

Syndrome

Incidence (new cases/ 106 population/year)

Tumor location

Malignant (%)

Main symptoms/signs

Established specific functional syndrome Pancreatic endocrine tumor Zollinger–Ellison syndrome (gastrinoma)

Gastrin

0.5–1.5

Insulinoma VIPoma (Verner– Morrison syndrome, pancreatic cholera, WDHA) Glucagonoma

Insulin Vasoactive intestinal peptide

1–2 0.05–0.2

Glucagon

0.01–0.1

Somatostatinoma

Somatostatin

Rare

GRFoma

Growth hormonereleasing hormone

Unknown

ACTHoma

ACTH

Rare

Serotonin ? Tachykinins PTHrP Others unknown

Rare (43 cases)

Serotonin, possibly tackykinins, motilin, prostaglandins

0.5–2

Rare

a

PET causing carcinoid syndrome PET causing hypercalcemia Carcinoid tumor Carcinoid syndrome

Possible specific functional PET syndrome PET secreting calcitonin Calcitonin

Rare

Duodenum (70%) Pancreas (25%) Other sites (5%) Pancreas (> 99%) Pancreas (90%; adult) Other (10%; neural, adrenal, periganglionic) Pancreas (100%)

> 60

> 95

Cushing’s syndrome (100%)

60–80

Same as carcinoid syndrome above

84

Abdominal pain due to hepatic metastases

Midgut (75–87%) Foregut (2–33%) Hindgut (1–8%) Unknown (2–15%)

95–100

Diarrhea (32–84%) Flushing (63–75%) Pain (10–34%) Asthma (4–18%) Heart disease (11–41%)

> 80%

Diarrhea (50%)

Unknown

Hypertension

> 60

Weight loss (30–90%) Abdominal mass (10–30%) Pain (30–95%)

Pancreas (55%) Duodenum/jejunum (44%) Pancreas (30%) Lung (54%) Jejunum (7%) Other (13%) Pancreas (4–16% all ectopic Cushing’s) Pancreas (< 1% all carcinoids) Pancreas (rare cause of hypercalcemia)

PET secreting renin No functional syndrome

Renin

Rare

PPoma/nonfunctional

None

1–2

Pancreas (100%)

PET, pancreatic endocrine tumor.

< 10 40–70

50–80

Pain (79–100%) Diarrhea (30–75%) Esophageal symptoms (31–56%) Hypoglycemic symptoms (100%) Diarrhea (90–100%) Hypokalemic (80-100%) Dehydration (83%) Rash (67–90%) Glucose intolerance (38–87%) Weight loss (66–96%) Diabetes mellitus (63–90%) Cholelithiases (65–90%) Diarrhea (35–90%) Acromegaly (100%)

Pancreas (rare cause of hypercalcitonemia) Pancreas

a

60–90

> 70

114

t0015

Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS)

Table III Carcinoid Tumor Location, Frequency of Metastases, and Association with Carcinoid Syndrome Location a (% of total)

Incidence of a metastases

Incidence of carcinoid b syndrome

Table IV Molecular Abnormalities and Prognostic Factors in NETsa Molecular abnormalities PETs HER2/neu (erbB-2) expressed in 100% GAS MEN1 gene LOH, 46%; mutation, 42% GAS/10% INS

Foregut Esophagus Stomach

5 liters/ day of fluid and > 350 mm/day of potassium when the diarrhea is not controlled. Octreotide will control the diarrhea in 85–90% of VIPoma patients. In nonresponders or patients who have become refractory, combinations of octreotide and glucocorticoids may be helpful. Other drugs that may help in individual patients include clonidine, indomethacin, prednisone, lithium, phenothiazine, loperamide, lidamidine, propanolol, and metoclopramide. In 40–70% of adult patients with VIPoma, diffuse metastatic disease in the liver is present initially; therefore, surgical cure is not possible. For these patients, long-acting somatostatin analogues such as octreotide or lanreotide (Fig. 1) are the drugs of choice for treatment. If these fail or tumor growth continues, treatment with chemoembolization, hepatic embolization, or chemotherapy or radiolabeled somatostatin analogues (Fig. 1) may be helpful. If the majority of the metastatic disease in the liver can be safely resected, cytoreductive surgery may be of value in helping to control the symptoms of the hormone-excess state.

Glucagonomas Glucagonomas are neuroendocrine tumors of the pancreas that ectopically secrete excessive amounts of glucagon and cause a syndrome characterized by a dermatitis (migratory necrolytic erythema), glucose intolerance or diabetes, and weight loss (Table VIII). Clinical Features Glucagonomas usually occur in middle-aged and elderly populations. Migratory necrolytic erythema usually starts as an erythematous area typically at perifacial or intertriginous areas, such as the groin, buttocks, or thighs. It subsequently becomes raised and bullae form and break, resulting in eroded areas. The skin lesions may wax and wane and frequently precede the diagnosis of glucagonoma (mean, 6–8 years). Hypoaminoacedemia is a characteristic feature of glucagonomas, with plasma amino acid levels decreasing to < 25% of normal, especially glycogenic amino acids such as alanine and glycine. Weight loss is another prominent feature of this syndrome and is not seen early in other PETs unless malabsorption is present, suggesting that it is an intrinsic feature of this syndrome. Experimental studies support

p0125

118

Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS)

Table VIII Frequency of Clinical Symptoms and Laboratory Findings in Patients with Glucagonoma Frequency (%) Clinical symptoms Dermatitis

64–90

Diabetes/glucose intolerance

38–96

Weight loss

56–96

Glossitis/stomatitis/cheilitis Diarrhea

29–40 14–15

Abdominal pain

12

Thromboembolic disease

12–35

Venous thrombosis

24

Pulmonary emboli

11

Psychiatric disturbance

0–17

Laboratory abnormality Anemia Hypoaminoacidemia

33–90 26–100

Hypocholesterolemia

80

or stomatitis, glucose intolerance or diabetes that may precede the diagnosis by years, and anemia (Table VIII).

Figure 1 Structure of somatostatin and synthetic analogues used for diagnostic (tumor localization) or therapeutic indications in patients with neuroendocrine tumors. The indium (111In), yttrium (90Y), or lutetium (177Lu) compounds are used for therapeutic purposes. DOTA-1,4,7,10-tetra-azacylododecane-N,N0 ,N00 , N000 -tetra-acetic acid.

the conclusion that a novel anorectic substance, independent of glucagon, is released by these tumors and is responsible for the weight loss. Other prominent symptoms/clinical findings include glossitis

Pathology/Pathophysiology Glucagonomas are characteristically diagnosed late in their course and are usually large (average, 5–10 cm), and 50–80% have evidence of metastatic spread at diagnosis, usually to the liver (43–80%). Glucagonomas are usually single tumors and 50–80% occur in the pancreatic tail. Glucagon is a naturally occurring, 29-amino acid peptide characteristically released by the pancreatic A cells. Most of the findings of the syndrome are compatible with the known actions of glucagon stimulating glucogenolysis and gluconeogenesis and affecting gut secretion and motility. The exact pathogenesis of the migratory necrolytic erythema remains unclear and has been attributed by some to hypoaminoacidemia or to nutritional deficiencies such as zinc. Diagnosis/Differential Diagnosis The diagnosis is established by demonstrating an elevated plasma glucagon level (normal level is usually < 150 ng/liter). Plasma glucagon levels usually exceed 1000 ng/liter in glucagonoma patients (90%), and in patients with symptoms/laboratory findings of glucagonoma (Table VIII), a level > 1000 ng/liter is diagnostic. Other conditions can also cause elevated plasma glucagon levels, including pancreatitis, hepatic failure disease, renal failure, Cushing’s syndrome, prolonged fast, or familial hyperglucagonemia. Except

119

Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS)

for cirrhosis, these disorders usually do not cause increases in plasma glucagon levels > 500 ng/liter. Treatment Diffuse hepatic metastases are usually present at diagnosis (in up to 80% of patients), so curative surgical resection is usually not possible. Surgical debulking or other antitumor treatments may be of palliative benefit. The drugs of choice are the long-acting somatostatin analogues (octreotide and lanreotide) (Fig. 1), which improve the rash in 75–80% of patients and may improve the weight loss, pain, and diarrhea but usually do not improve the diabetes/glucose intolerance.

Somatostatinoma Syndrome The somatostatinoma syndrome is caused by a neuroendocrine tumor of the GI tract that ectopically secretes excess amounts of somatostatin (Fig. 1), which frequently causes diabetes mellitus, gallbladder disease, diarrhea, and steatorrhea (Table IX ). There is no general agreement on what constitutes a somatostatinoma, and there is no distinction in the literature regarding the presence of a somatostatinoma and/or the somatostatinoma syndrome. In the literature, the term somatostatinoma is generally used to mean a GI neuroendocrine tumor possessing somatostatin-like immunoreactivity, and it may (11–45%) or may not (55–89%) be associated with clinical symptoms due to ectopic release of somatostatin (Table IX). Because of this confusion, the term somatostatinoma syndrome is used here to indicate a PET-releasing

t0045

Table IX Clinical and Laboratory Findings in Patients with Somatostatinomas with or without Somatostatinoma Syndrome a

Somatostatinoma

Overall frequency in Pancreatic Intestinal somatostatinoma (%) (%) a syndrome (%) Clinical finding Diabetes mellitus

95

21

95

Gallbladder disease

94

43

68

Diarrhea

66–97

11–36

37

Weight loss

32–90

20–44

68

Laboratory finding Steatorrhea

83

12

47

Hypochlorhydria

86

17

26

a

Somatostatinoma is the occurrence of a pancreatic endocrine tumor containing somatostatin by immunocytochemistry that can occur with (11%) or without (89%) the somatostatinoma syndrome, which is due to ectopically released somatostatin.

somatostatinoma, which causes clinical symptoms, and the term somatostatinoma is used to indicate a PET containing somatostatin immunoreactivity.

Clinical Features The mean age of onset is 51–53 years. In one large review (N ¼ 173), only 11% of all somatostatinomas in the literature were associated with the somatostatinoma syndrome. The principal clinical features of the somatostatinoma syndrome are gallbladder disease, diabetes mellitus, diarrhea, weight loss, and steatorrhea. The frequency of these symptoms depends on the location of the somatostatinoma (Table IX ). Each symptom characteristic of the syndrome is reported more frequently with pancreatic than intestinal somatostatinomas.

Pathology/Pathophysiology Somatostatin is a naturally occurring tetradecapeptide (Fig. 1) found widely in the central nervous system and GI tract, where it functions as a neurotransmitter or has paracrine or autocrine actions. In general, it is a potent inhibitor of processes, including the release of numerous hormones, gastric and intestinal pancreatic secretion, and absorption. The hypochlorhydria, diabetes, steatorrhea, and gallbladder diseases that occur in somatostatinoma syndrome are caused by the known inhibitor effects of somatostatin. Somatostatinomas occur in the pancreas in 56– 74% of cases and are mainly found in the pancreatic head. In most of the remaining cases, they are found in the intestine, with 90% in the duodenum, particularly in the periampullary area. They are usually solitary, large tumors (mean, 3.6–4.9 cm), and 53–84% have metastatic spread at diagnosis (usually to the liver).

Diagnosis/Differential Diagnosis In most cases in the literature, somatostatinomas have been found incidentally either at the time of cholecystectomy or during endoscopy. The presence of psammoma bodies in a duodenal neuroendocrine tumor should raise this possibility. Duodenal somatostatinomas are associated with von Recklinghausen’s disease. Most of these patients do not develop the somatostatinoma syndrome and have normal plasma somatostatin levels. The diagnosis of the somatostatinoma syndrome requires the demonstration of an elevated plasma somatostatin level.

120

s0140

p0195

Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS)

Treatment

NONFUNCTIONAL PETs

Symptoms of the somatostatinoma syndrome are improved by administration of long-acting synthetic somatostatin analogues such as octreotide (Fig. 1). Pancreatic tumors (70–92%) and intestinal tumors (30–69%) are associated with liver metastases at presentation and antitumor treatment is frequently needed.

Nonfunctional PETs are neuroendocrine tumors of the pancreas that secrete no products or the secreted products do not cause a distinct functional syndrome. The symptoms from these tumors are therefore due entirely to the tumor per se (i.e., pain, jaundice, weight loss, etc.). Nonfunctional PETs frequently secrete chromogranin A (CgA) (80–100%), CgB (90–100%), pancreatic polypeptide (PP) (58%), and the a-subunit of hCG (40%), and many secrete other hormones such as neurotensin. These tumors characteristically present late in the disease course. They are generally large (72% > 5 cm) and invasive, and hepatic metastases are usually present (64–92%). The most common symptoms are abdominal pain (30–50%), jaundice (20– 35%), weight loss, fatigue, and bleeding. In 10–15% of cases, they are found incidentally. The diagnosis is established by histology with appropriate neuroendocrine tumor immunohistochemistry and by assessing plasma hormone levels/clinical symptoms. Plasma PP is increased in 22–71% of patients, CgA levels are increased in 80–100%, and in patients with a pancreatic mass without a functional syndrome, this finding suggests that a nonfunctional PET is present. Curative surgical resection is rarely possible, and treatment needs to be directed against the malignant PET.

GRFomas AND OTHER RARER FUNCTIONAL PET SYNDROMES GRFomas are neuroendocrine tumors that ectopically secrete growth hormone-releasing factor (GRF). GRF is a 44-amino acid peptide that stimulates growth hormone release. GRFomas occur in lung (47–54%), PETs (29–30%), small intestinal carcinoids (8–10%), and other sites (12%). The symptoms are those of acromegaly and the mean age of patients is 38 years. The acromegaly is indistinguishable from classical acromegaly due to a pituitary adenoma. Pancreatic GRFomas are usually large (mean, > 6 cm) and liver metastases are present in 39% of cases. They should be suspected in a patient with acromegaly with an abdominal tumor, MEN-1, or hyperprolactinemia, which occurs in 70% of GRFomas. The diagnosis is established by performing plasma assays for GRF and growth hormone. Surgery is the treatment of choice if possible. Symptoms can be controlled in > 75% of patients by long-acting somatostatin analogues, such as octreotide or lanreotide (Fig. 1). Cushing’s syndrome due to a PEToccurs in 4–16% of all patients with ectopic Cushing’s syndrome. It occurs in 5% of patients with sporadic gastrinomas and is associated with hepatic metastases and a poor prognosis. Paraneoplastic hypercalcemia due to a PET releasing PTH-related peptide is rare. These tumors are usually large and malignant. PETs secreting calcitonin may cause a specific syndrome associated with diarrhea (Table II). PETs also cause the carcinoid syndrome (Table II). These are characteristically large and malignant (68–88%) and may cause an atypical carcinoid syndrome because they lack DOPA decarboxylase. A renin-producing PET was described in a patient presenting with hypertension (Table II). Ghrelin is a 28-amino acid peptide with a number of metabolic functions. Although expression has been demonstrated in most PETs, in a one study only 1 of 24 patients (4%) with a PET had an elevated plasma ghrelin level. This patient was asymptomatic, suggesting that no specific syndrome is associated with ectopic release of ghrelin by a PET.

FUNCTIONAL SYNDROMES DUE TO CARCINOID TUMORS Carcinoid tumors can cause a specific functional syndrome, the carcinoid syndrome, or occasionally can release biologically active peptides that cause the specific PET syndromes discussed previously. Because carcinoids are also malignant (Table III), specific treatments need to be directed at both the carcinoid tumor and the functional syndrome it produces.

Carcinoid Syndrome The carcinoid syndrome is caused by a neuroendocrine tumor, usually present in the gastrointestinal tract, ectopically secreting bioactive amines/peptides, which results in a clinical syndrome characterized by diarrhea, flushing, asthma/wheezing, and, occasionally, heart disease (Table X ). Clinical Features The mean age at presentation is 57 years, but it occurs over a wide age range (9–91 years) (Table X). The principal symptoms are diarrhea and flushing, which occur in up to 73% of cases initially and 84% during

Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS)

Table X Clinical Features in Patients with Carcinoid Syndrome At presentation

During course of disease

Symptoms/signs Diarrhea (%)

32–73

68–84

Flushing (%)

23–65

63–74

Pain (%)

10

34

Asthma/wheezing (%)

4–8

3–18

Pellagra (%) None (%)

2 12

5 22

Carcinoid heart disease (%)

11

14–41

Demographics Male (%)

46–59

46–61

Mean age, years (range)

57 (25–79)

53 (9–91)

Tumor location Foregut (%) Midgut (%)

5–9

2–33

78–87

60–87

Hindgut (%)

1–5

1–8

Unknown (%)

2–11

2–15

the course of the disease. The flush is usually of sudden onset, associated with a deep red to violaceous erythema of the upper body, and often associated with a feeling of warmth and occasionally with lacrimination, pruritis, or diarrhea. It may be precipitated by food, exercise, alcohol, or drugs, particularly serotonin reuptake inhibitors. Flushing is usually caused by metastatic midgut tumors. Flushing with midgut tumors and bronchial or gastric carcinoids may differ in duration, associated symptoms (salivation and lacrimination), and skin color. Diarrhea usually occurs with flushing (85% of cases) and is usually watery and of small volume (60% < 1 liter/day). Steatorrhea is present in 67%, and in 50% of patients it is > 15 g/day of fecal fat. Cardiac manifestations occur in 11% initially and 14–41% during the disease course (Table X), and they are due to endocardial fibrosis, primarily on the right side (tricuspid > pulmonary), but they can also occur on the left side. Fibrosis results in valve constriction and up to 80% of patients develop heart failure. Pellagra-like symptoms (2–25%) and symptoms due to increased fibrotic tissue (i.e., retroperitoneal fibrosis, Peyronie’s disease, and intraabdominal fibrosis) are unusual features of this disease. A life-threatening complication of the carcinoid syndrome is the development of a carcinoid crisis. This is most frequently seen in patients with high 5hydroxyindoleacetic acid (5-HIAA) levels and may be provoked by anesthesia, endoscopy, stress, surgery, a

121 radiological procedure, or a biopsy. Patients develop intense flushing, diarrhea, abdominal pain, hypotension, and cardiac abnormalities. If not adequately treated, it can be fatal. Pathology/Pathobiology Carcinoid symptoms occurred in 8% of 8876 patients with carcinoid tumors. Carcinoid syndrome occurs only when tumor-secreted products reach the systemic circulation in sufficient concentrations. In 91% of cases, this occurs due to liver metastases, and in the remainder it occurs due to retroperitoneal invasion by gut or pancreatic tumors or due to lung or ovary carcinoids with direct access to the systemic circulation. Midgut carcinoids account for 60–67% of cases of the carcinoid syndrome, foregut tumors account for 2–33%, hindgut accounts for 1–8%, and an unknown primary accounts for 2–15%. One of the main secretory products of carcinoid tumors is serotonin (5-HT), and overproduction occurs in 90–100% of patients with carcinoid syndrome. Serotonin is thought to primarily mediate diarrhea by its effects on gut motility and intestinal secretion. However, prostaglandins and tachykinins (substance P, K; neuropeptide K) may also be important in causing diarrhea in some patients. Flushing is not relieved by serotonin antagonists, and in patients with gastric carcinoids it can be due to histamine secretion. Tachykinins are released during flushing and may be important in its mediation. Both histamine and serotonin may be responsible for the wheezing/asthma as well as the fibrotic reactions characteristic of this disease. The heart disease is likely due to the actions of serotonin because it is similar to that seen with appetite-suppressant drugs (e.g., fenfluramine) that have high affinity for serotonin receptors. Patients may develop a typical or atypical carcinoid syndrome. The typical syndrome is due to a midgut tumor oversynthesizing serotonin from tryptophan. The atypical syndrome occurs when there is a deficiency in the enzyme required to convert 5-HTP to 5-HT and there is overproduction of 5-HTP. The atypical syndrome is more likely to occur with foregut carcinoids. Diagnosis/Differential Diagnosis The diagnosis of carcinoid syndrome requires measurement of urinary or plasma serotonin or its metabolites. The measurement of urinary 5-HIAA is most frequently performed. False positives may occur if a patient eats serotonin-rich foods (bananas, walnuts, pecans, and pineapple) or takes certain medications (l-Dopa, cough syrups with guaniforesin, and

p0230

122 salicylates). If an atypical carcinoid syndrome is suspected, urinary 5-HIAA may be only slightly elevated, and other metabolites of tryptophan such as 5-HTP should be considered. Serum cgA levels are elevated in 50–100% of patients with carcinoid tumors, and the level correlates with tumor bulk. However, serum CgA levels are not specific for carcinoids because they can occur with other NETs, including PETs. Flushing can be caused by other conditions, including systemic mastocytosis, reactions to alcohol or glutamate, effects of drugs, and menopause. These need to be excluded. Treatment Treatment consists of avoiding conditions that precipitate attacks and treatment of wheezing with bronchodilators, treatment of heart failure with diuretics, and treatment of diarrhea with antidiarrheal agents, such as loperamide or diphenoxylate. If symptoms present, somatostatin analogues (octreotide and lanreotide) (Fig. 1) or serotonin receptor antagonists are the drugs of choice. Somatostatin analogues control symptoms in 80% of patients with flushing or diarrhea, and 70% show a > 50% decrease in urinary 5HIAA. Approximately 40% of patients show some resistance after 4–6 months and doses may have to be increased. Sustained-release preparations, such as octreotide-LAR (long-acting release) and lanreotidePR (prolonged release), are widely used because they can be given less frequently. Octreotide-LAR is usually given monthly and lanreotide-PR every 10–14 days as opposed to the usual preparation of octreotide/lanreotide that is given subcutaneously every 6–12 h. Somatostatin analogues can be given to both treat and prevent carcinoid crises. Prior to a possible precipitating event, such as surgery, anesthesia, or stress, it is recommended that octreotide be administered. Side effects from the somatostatin analogues occur in 40–60% of patients with subcutaneous analogues. Pain at the injection site and GI side effects (discomfort, nausea, diarrhea, and cramping) are the most common. They are usually short-lived and do not interrupt treatment. Long-term complications include an increased incidence of gallstones/biliary sludge (52%), steatorrhea, and worsening glucose tolerance. Interferon-a may also control symptoms in some patients. Surgery should be performed if possible; however, almost all patients with carcinoid syndrome have metastatic disease in the liver and curative resection is not possible. Treatment directed against the tumor may be needed.

Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS)

NET TUMOR LOCALIZATION Tumor localization is needed for all management phases of both PETs and carcinoid tumors. Both the localization of the primary tumor and the determination of the location and extent of metastatic disease are required to appropriately manage these patients. Conventional imaging studies [CT, magnetic resonance imaging (MRI), ultrasound, and angiography] and SRS are widely used. For PETs, endoscopic ultrasound is also widely used. Bronchial carcinoids are usually detected by chest X-ray and assessed by CT. Rectal, duodenal, colonic, and gastric carcinoids are usually detected by GI endoscopy. Both PETs and carcinoids frequently (90–100%) overexpress somatostatin receptors, which have a high affinity for radiolabeled somatostatin analogues (Figs. 1–3) that can be used to localize them. Because of its greater sensitivity compared to that of conventional imaging studies and its ability to localize a number of tumors throughout the body simultaneously, SRS is the imaging modality of choice for localizing all primary and metastatic NET tumors except insulinomas. Insulinomas are usually small and have a low density of somatostatin receptors, with the result that SRS is positive in only 12–50% of patients with insulinomas. In contrast, SRS is positive in 73–89% of patients with carcinoids and 60– 100% of patients with PETs other than insulinomas. Figures 2 and 3 show two examples of the ability of SRS to image a primary PETand metastatic disease in the liver when conventional imaging studies were negative. For PETs localized in the pancreas, endoscopic ultrasound is highly sensitive, localizing 73– 100% of insulinomas, which occur almost exclusively within the pancreas. SRS occasionally gives false positives (12% in one study) because normal and abnormal cells can have increased somatostatin receptors such as granulomas, thyroid disease, and activated lymphocytes (abscess and infection). Furthermore, SRS does not provide information on tumor size or the exact location of metastases, and a CT scan or MRI are frequently used to provide this information.

TREATMENT OF ADVANCED DIFFUSE METASTATIC DISEASE IN PATIENTS WITH MALIGNANT NETs Of the numerous prognostic factors identified for NETs (Table IV), the presence and the extent of the hepatic metastases are the most important in almost every study. For patients with gastrinomas, 5-year

Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS)

f0010

123

Figure 2 Computed tomography (CT) scan (top) and somatostatin receptor scintigraphy (SRS) (bottom) of a patient with a pancreatic endocrine tumor (gastrinoma). This patient previously had a duodenal tumor removed. The CT scan was negative, whereas the SRS showed a tumor medial and anterior to the right kidney. At surgery, a PET was found in a metastatic lymph node in the periduodenal area. This case illustrates the greater sensitivity of SRS compared to conventional imaging studies.

Figure 3 Computed tomography (CT) scan and somatostatin

survival is 98–100% without liver metastases, 78% with limited metastases in one lobe, and 16% with diffuse metastases. For carcinoid tumors without liver metastases, 5-year survival is 80–90%, and with diffuse metastases it is 50%. A number of treatments are reported to be effective, including embolization, chemoembolization, chemotherapy, cytoreductive surgery (removal of all visible tumor), somatostatin analogues, a-interferon, radiotherapy using radiolabeled somatostatin analogues to target the tumor (Fig. 1), and liver transplantation.

support this conclusion. However, studies suggest that it may increase survival; therefore, cytoreductive surgery is recommended, if possible. Chemotherapy for metastatic carcinoids is generally disappointing, with response rates of 0–40% with various two–three drug combinations. With PETs, chemotherapy has been more successful, with response rates of 30–70%. The regimen of choice is streptozotocin and doxorubicin. Unfortunately, chemotherapy is almost never curative and has substantial toxicity. Therefore, it is generally reserved for patients with rapidly growing PETs who fail other treatments. Long-acting somatostatin analogues (Fig. 1) and a-interferon rarely decrease tumor size (i.e., 0–15%); however, these drugs have prominent tumoristatic effects, stopping growth in 50–90% of patients with NETs. Studies suggest that they are particularly effective in slower growing tumors. It has not been proven that these agents extend survival; however, because of

Specific Antitumor Treatments Unfortunately, cytoreductive surgery is only possible in 10–20% of patients who have limited hepatic metastases, allowing surgical removal of at least 90% of visible tumor. Although it is reported to provide palliative treatment, there are no control studies to

receptor scintigraphy (SRS) of a patient with a metastatic PET to the liver. The CT scan showed no liver lesions, whereas the SRS showed two hepatic metastases (one in the right lobe and one in the left lobe). This case illustrates the increased sensitivity of SRS compared to conventional imaging studies for localizing hepatic metastases.

124 the availability of long-acting formulations (e.g., one injection of somatostatin/month), their low toxicity, and their long-term effectiveness in some patients, they are the antitumor agents of choice. Hepatic embolization or chemoembolization (i.e., chemotherapy with embolization) can decrease tumor bulk and help control the symptoms of the hormoneexcess state. This approach is usually reserved for patients who have disease largely confined to the liver, who have a patent portal vein, and who fail treatment with other modalities. Somatostatin receptor-directed cytotoxicity using radiolabeled somatostatin analogues that are internalized by somatostatin receptors (Fig. 1) overexpressed on the NET are being widely investigated. The following are being evaluated for treatment: 111indium (111Ln)-labeled compounds, which emit gamma rays, internal conversion, and Auger electrons; 90yttriumcoupled analogues, which emit high-energy b particles; and 177lutetium (177Lu)-coupled analogues, which emit both. The 177Lu and 111In compounds have respectively been shown to stabilize disease in 41 and 40% and decrease tumor size in 38 and 30% of patients with advanced metastatic NETs. Liver transplantation, although largely abandoned for most metastatic tumors, is still a consideration for patients with metastatic NETs because of their slower growth. In 103 cases of malignant NETs (43 carcinoid, 48 PETs), liver transplantation achieved 2- and 5-year survival rates of 60 and 47%, respectively. Liver transplantation has been suggested for younger patients with metastatic NETs limited to the liver.

See Also the Following Articles Gastrinomas . Ghrelin . GI Hormones in Cancer . GI Tract, General Anatomy (Cells) . GI Tract, General Pathology of Endocrine Growths . Substance P

Gastrointestinal Hormone see GI Hormone entries

Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS)

Further Reading Alexander, R. A., and Jensen, R. T. (2001). Pancreatic endocrine tumors. In ‘‘Cancer: Principles and Practice of Oncology’’ (V. T. DeVita, S. Hellman, and S. A. Rosenberg, eds.), 6th ed., pp. 1788–1813. Lippincott Williams & Wilkins, Philadelphia. Arnold, R., Simon, B., and Wied, M. (2000). Treatment of neuroendocrine GEP tumours with somatostatin analogues: A review. Digestion 62, 84–91. Caplin, M. E., Buscombe, J. R., Hilson, A. J., Jones, A. L., Watkinson, A. F., and Burroughs, A. K. (1998). Carcinoid tumour. Lancet 352, 799–805. Corleto, V. D., Delle Fave, G., and Jensen, R. T. (2002). Molecular insights into gastrointestinal neuroendocrine tumors: Importance and recent advances. Digestive Liver Dis. 34, 668–680. Eriksson, B. (2001). Systemic therapy for neuroendocrine tumors of the pancreas. In ‘‘Surgical Endocrinology’’ (G. M. Doherty and B. Skogseid, eds.), pp. 393–403. Lippincott Williams & Wilkins, Philadelphia. Gibril, F., and Jensen, R. T. (2003). Diagnostic use of radiolabeled somatostatin analogues in patients with gastroenteropancreatic endocrine tumors. Digestive Liver Dis. (in press). Jensen, R. T., and Doherty, G. M. (2001). Carcinoid tumors and the carcinoid syndrome. In ‘‘Cancer: Principles and Practice of Oncology’’ (V. T. DeVita, Jr., S. Hellman, and S. A. Rosenberg, eds.), 6th ed., pp. 1813–1833. Lippincott Williams & Wilkins, Philadelphia. Jensen, R. T., and Norton, J. A. (2002). Pancreatic endocrine tumors. In ‘‘Sleisenger and Fordtran’s Gastrointestinal and Liver Disease: Pathophysiology, Diagnosis, and Management’’ (M. Feldman, L. S. Friedman, and M. H. Sleisenger, eds.), 7th ed., pp. 988–1016. Saunders, Philadelphia. Kulke, M. H., and Mayer, R. J. (1999). Carcinoid tumors. N. Engl. J. Med. 340, 858–868. Kwekkeboom, D., Krenning, E. P., and deJong, M. (2000). Peptide receptor imaging and therapy. J. Nucl. Med. 41, 1704–1713. Lehnert, T. (1998). Liver transplantation for metastatic neuroendocrine carcinoma. Transplantation 66, 1307–1312. Metz, D. C., and Jensen, R. T. (2003). Endocrine tumors of the gastrointestinal tract and pancreas. In ‘‘Gastrointestinal Cancers’’ (A. K. Rustgi, ed.), pp. 681–720. Saunders, New York. Mignon, M. and Jensen, R. T. (eds.) (1995). ‘‘Endocrine Tumors of the Pancreas: Recent Advances in Research and Management,’’ Vol. 23. Karger, Basel, Switzerland. O’shea, D., and Bloom, S. R. (guest eds.) (1996). Gastrointestinal endocrine. Balliere’s Clin. Gastroenterol. 10, 555–766.

126

Gender Assignment and Psychosocial Management

of the twentieth century. It was generally assumed that, if gender assignment had been carried out correctly, that is, in agreement with the true sex, the final outcome would be a heterosexual man or a heterosexual woman, in agreement with the assigned gender. There were problems with the true-sex policy, however. The presumed definitive criterion of the true sex did not always agree with a person’s somatic physical appearance as a man or woman, and also did not always agree with the person’s gender identity, and could therefore lead to dramatically adverse social and legal consequences for the intersexed individual.

development. A third consequence was the instruction to keep the intersex status of the child secret from all people who do not belong to the core family and to educate the child gradually, in line with his or her cognitive development. These guidelines were intended to prevent the parents from developing chronic doubts about the sex of their child and to protect the child from stigmatization by other people, but have frequently led to parental attempts of preventing the disclosure of details of the medical history to the intersex children themselves, sometimes even after they attained adulthood.

The Optimal-Gender Policy

The True-Brain-Sex Policy

On the basis of a critical examination of the true-sex policy and studies of the overall psychosocial outcome and quality of life of intersex patients, John Money and the Hampsons at Johns Hopkins Hospital in Baltimore, Maryland, formulated an optimal-gender policy during the 1950s. This policy had a number of underlying assumptions. (1) There is no single biological criterion that determines the development of psychological gender; rather, there is a cascade of biological processes that culminate in the development of gender identity. (2) Socialization factors have the decisive role in psychological gender development. (3) If one minimizes the barriers to socialization in the assigned gender, the outcome will be a heterosexual man or a heterosexual woman, in agreement with the assigned gender. (4) The condition of the body and particularly the genitalia limits the psychosocial and especially psychosexual functioning of the intersexed person in later years. Therefore, the newborn with intersexuality should be assigned to that gender that permits the optimal psychosexual and psychosocial functioning when all available medical treatment options are taken into account. So, in contrast to the question of the true-sex policy—‘‘Is this a boy or a girl’’?—the optimal-gender policy asked ‘‘Will this newborn have a better function later in life as a male or a female?’’ From the central role of socialization followed the concept of the birth of an intersex newborn as a psychosocial emergency, with the implication that the decision time for gender assignment should be kept as short as possible. Another consequence was the recommendation of early feminizing or masculinizing surgery of the external genitalia so that their appearance would be as similar to the gender norm as possible and would therefore not interfere with gender-typical rearing conditions and body-image

Increasingly over the past decade, the optimal-gender policy has come under criticism. One reason is that also under this policy some individuals will turn out dissatisfied with their assigned gender and may seek gender reassignment, which is made more difficult if the external genitalia have been operated on to be more compatible with the originally assigned gender. A second argument is that, even without later gender change, genital surgery carries a risk of damage to sexual functioning in adolescence and adulthood. Intersex activists, therefore, demand that psychosocially indicated genital surgery be performed only with the informed consent of the mature individual. Some critics advocate a moratorium on genital surgery, unless necessary for strictly medical indications, and claim—without data—that psychological counseling can take care of all attending psychosocial problems that children with marked genital ambiguity and their families may face. Third, the optimal-gender policy is blamed (although unjustly; see above) for keeping the medical facts secret, especially from the patient him- or herself, which contributes to the maintenance of the social stigma of intersexuality and thereby to the negative self-image (especially shame) of the intersexed individual. Many intersex activists, therefore, argue for early comprehensive disclosure of their medical history to intersex patients. A fourth criticism of the optimal-gender policy derives from the extensive data, especially data generated by animal research, regarding the influence of sex hormones on the developing brain and long-term sex-dimorphic behavior. Because of these data, some biological determinists suggest that the decisive factor for gender identity formation is the prenatal androgenization of the brain and that psychosocial factors have only a secondary role. If such a ‘‘true-brain-sex policy’’ is valid, gender assignment decisions should be based on the

127

Gender Assignment and Psychosocial Management

degree of androgenization/masculinization of the brain and minimization of barriers to socialization is unimportant. However, because brain imaging techniques are not yet capable of rendering such assessments, these theorists must fall back on the status of the genitalia as a—rather uncertain—indicator of brain androgenization/masculinization.

Policy Effects on Gender Assignment The three major policies outlined above may lead to quite varied decisions on gender assignment. Two cases will illustrate this: (1) a chromosomally female newborn with extreme genital masculinization (Prader stage V) due to prenatal androgen excess associated with classical congenital adrenal hyperplasia (CAH) and (2) a chromosomally male newborn with penile agenesis. According to the true-sex policy, case (1) should be raised female because of the clearly female histology of the gonads (ovaries) and female chromosomes (see Table I). By contrast, a 46,XY newborn with penile agenesis would be assigned to the male gender, because of the normal-male histology of the testicular gonads and the clearly male karyotype. Under the optimal-gender policy, case (1) would be assigned to the female gender, because the external genitalia can be feminized so as to permit coitus and to thereby retain the option of conception and pregnancy, because the internal female reproductive organs are intact. The 46,XY infant with penile agenesis would also be assigned to the female gender, because the surgical de novo construction of a functioning penis is not yet possible and therefore sexual functioning as a male would be compromised. Advocates of the true-brain-sex policy would recommend assigning the CAH case to the male gender because of the putative effects of prenatal androgens on the brain and recommend the same for the case of penile agenesis. Some activists recommend assigning the intersex child to the gender that seems to offer the more promising outcome, but to do so provisionally and to consider from the outset the possibility of later gender change and therefore not to operate on the genitals before the age of consent, unless medically necessary. t0005

Table I Gender Assignment as a Function of Policy Disorder

True-sex policy

Optimal-gender policy

True-brain-sex policy

Penile agenesis

Male

Female

Male

CAH/Prader stage V

Female

Female

Male

The Status of the Evidence The major objections to the optimal-gender policy are important to consider, but a consensus about a new management policy has yet to emerge, especially for 46,XY individuals with intermediate degrees of genital ambiguity. The major problem is insufficient data concerning the long-term psychological outcome. Patient-initiated later gender change, for instance, can be observed in intersex individuals in both directions, from male to female and from female to male (and, in some cases, from the assigned gender to ‘‘intersex’’). This is not surprising, because such gender change is also well known—as transsexualism—in non-intersex individuals, but the relative frequency of later gender change is increased in intersex patients. Given the psychological, social, and medical problems that are associated with later gender change, clinicians usually favor a policy that minimizes its occurrence. Thus, long-term follow-up data are needed that will allow clinicians to state the relative frequency of gender change of patients with a given syndrome who were managed under a defined policy and in a particular cultural context. Because gender change can take place as late as in midlife, follow-up studies need to reach at least into that age range. Given the relative rarity of intersex patients, it is not surprising, therefore, that the data thus far available are clearly insufficient. As an extreme example, consider the question of gender reassignment in chromosomal males with traumatic loss of the penis in infancy. One such case, in which reassignment to female at 17–21 months of age was followed by patient-initiated re-reassignment to male in adolescence, has drawn enormous attention among care providers as well as in the media, once his unfortunate story was investigated and published, when he was in his mid-thirties. However, only one other case with a history of traumatic loss of the penis in infancy and reassignment to female has been followed into adulthood and published and this person continues to live as a woman without gender dysphoria. Two cases with strikingly different outcomes are not sufficient for evidence-based policy decisions. Even for intersex syndromes that are more prevalent and less shrouded in secrecy than the cases of traumatic loss of the penis, long-term outcome data are hard to come by and those that have become available all suffer from problems that limit their validity: loss to follow-up of patients who have moved and cannot be found, who have died (especially if for other than medical reasons), or who have refused to participate in research (especially if for reasons of disappointment with their medical treatment).

128 Outcome data are also insufficient with regard to the question of impairment of sexual functioning after gender-confirming genital surgery. From a neuroanatomic standpoint, it is plausible that excision of the clitorophallus— widely practiced as a feminizing technique in the 1950s and 1960s—would lead to a diminution of erotic sensitivity and orgasmic capacity, although the limited outcome data available show considerable variability. But surgical techniques have undergone many changes since then and reduction of the clitoropenis rather than excision has become the norm. To evaluate sexual functioning in patients who were operated on with these newer surgical techniques, researchers need to wait at least until the patients are sexually active, sufficiently sexually experienced, and able to talk about these intimate details of their lives, and this in a population that is known to be relatively delayed in attaining psychosexual milestones. The same applies to patients with a similarly ambiguous genital status at birth who for one reason or another have not undergone genital surgery and could therefore serve as comparisons. Most surgical outcome reports are limited to anatomical data and the few that include data on sexual functioning leave much to be desired in terms of sample representativeness, assessment methods, and study details provided, quite apart from the fact that patients who have undergone surgeries with the newer techniques are not yet old enough for sexual functioning studies. The true-brain-sex policy appears very plausible at first glance, but becomes more problematic on closer scrutiny. The assumption that the effects of prenatal hormones on the brain are the decisive biological factor in the development of gender and therefore the best criterion for the assignment of gender is primarily derived from research on the sexual differentiation of brain and behavior in nonhuman mammals, especially rodents, which has yielded several neuroendocrine models of the sexual differentiation of brain and behavior. In humans, most pertinent studies have been limited to androgens. In general, the resulting data allow the preliminary conclusion that effective prenatal androgens are indeed associated with the masculinization of gender-role behavior. However, the core gender identity, for instance, as a girl or as a woman, is compatible with large variability in genderrole behavior, as is vividly demonstrated in girls and women with CAH, thus making a close linkage to prenatal hormones less likely. It also must be kept in mind that in the limited literature available, the association between prenatal androgens and later gender change takes place predominantly in the context of further postnatal and pubertal cross-gender

Gender Assignment and Psychosocial Management

hormonalization and the associated increased somatic sexual ambiguity. Therefore, in all likelihood, it is not the prenatal hormonal milieu by itself that determines such later gender change.

Inferring Brain Masculinization from Genital Masculinization Adherents of the true-brain-sex policy recommend a male assignment for a newborn with intersexuality if the child’s brain was prenatally strongly masculinized, independent of the status of the genitalia. However, how can one determine in a newborn whether and to what extent the brain was prenatally masculinized and/or defeminized? Appropriate brain imaging techniques are not yet available, and even if they were available, it is rather uncertain that the brain is, at the time of birth, already sufficiently sexually dimorphic to show such clear-cut gender dimorphic structures. Some advocates of the true-brain-sex policy therefore suggest using the genital status at birth (staged according to Prader for excess masculinization in chromosomal females and according to Quigley for undermasculinization in chromosomal males) as an indicator of brain androgenization/masculinization. There are several problems with this approach. (1) Genital staging at birth is not suitable for all syndromes of interest. For instance, Prader staging is problematic for many prenatally dexamethasone-treated cases of 46,XX CAH and Quigley staging is problematic for 46,XY newborns with 5a-reductase deficiency, and Quigley staging is not at all suitable for other conditions of interest, such as 46,XY penile agenesis and 46,XY cloacal exstrophy of the bladder. (2) The growth of the penis does not depend exclusively on sex hormones. (3) The hormone model of genital differentiation is not identical with the hormone models of brain differentiation. As far as is known, the prenatal development of the genitalia is determined by testosterone, dihydrotestosterone, and anti-Mu¨llerian hormone, whereas the development of the brain depends on androgens and, possibly, estradiol. (4) Attempts to correlate in the 46,XX CAH syndrome the Prader stage with gender-role behavior have had only very limited success. (5) Available (scanty) data on patient-initiated gender change in adults with a given intersex syndrome show little relationship to genital status at birth. (6) It also must be kept in mind that gender identity incongruence occurs in nonintersexed persons as well, although prior to surgery the genitals of transsexuals are compatible with their original gender assignment. (7) Research has made it

Gender Assignment and Psychosocial Management

likely that contributions to the sexual differentiation of brain and behavior by hormone-independent genetic factors and brain-tissue-specific hormonal factors need to be taken into account. (8) There may be early effects of the social environment on central nervous system structures, as is already known from animal research. If any of these mechanisms participate in the development of persons with gender identity incongruence without somatic intersexuality, they are likely to also contribute to the variability and development of intersex patients. It can be concluded from all of these considerations that the genital status at the time of birth cannot be interpreted as a clear-cut indicator of brain masculinization.

Gender Assignment by Diagnostic Category Intersex children and their families cannot be left in gender limbo. Despite controversies and uncertainties, clinicians must continue making decisions in this area. The degree of uncertainty varies with the syndrome. The situation is relatively clear-cut for the most frequent syndrome, CAH in 46,XX newborns, for which there is a general consensus that the assignment should be to the female gender, because the internal reproductive structures (ovaries, fallopian tubes, distal vagina) permit reproduction, provided that external genitalia and vaginal introitus are surgically corrected where necessary. Although many girls with CAH show variable degrees of behavioral masculinization, gender identity typically is female. Some genitally highly masculinized 46,XX children with CAH have been inadvertently raised as males and maintain a male gender identity when correctly diagnosed in adolescence, and a few CAH females change to male in adolescence or adulthood, apparently mostly in cases of protracted marked genital ambiguity and inconsistent hormonal control. However, the evidence is not sufficient to support the routine assignment of the most masculinized (Prader stage V) newborns to the male gender (as has been recommended). Also, among 46,XY intersex syndromes not all gender assignment decisions are problematic. There is a consensus that 46,XY newborns with complete androgen insensitivity should be raised as female, because they will never respond to male sex hormones, neither somatically nor psychologically, and, given that it is (mostly) prenatal rather than postnatal androgens that affect gender-related behavior, the argument can be extended to the rare syndrome of complete gonadal dysgenesis. Even in the syndromes of partial androgen

129 insensitivity and/or testosterone biosynthesis defect, there is a consensus that the least undermasculinized cases should be raised male and the most undermasculinized cases should be raised female. However, given the scarcity of long-term follow-up reports on such syndromes, the best cutoff point on the Quigley scale for assignment to the male or female gender cannot be arrived at on an empirical basis and what is known about the variability of gender outcome in individuals with a given molecular genotype and endocrine or genital phenotype lets one expect that any cutoff point on the Quigley scale will be associated with occasional later patient-initiated gender reassignment, however rare. Similar uncertainties also affect the decision-making in cases of micropenis [i.e., a fully differentiated penis with the urethral meatus at the tip, but a (stretched) length 2½ standard deviations or more below the norm]. Individual cases of satisfactory long-term outcome have been shown among female-assigned as well as male-assigned individuals, but unsatisfactory outcomes have also been documented in both types of assignments, and evidence-based decision-making requires substantially larger samples and more comprehensive assessments to weigh the advantages and disadvantages of either assignment. Also problematic are the relatively rare pubertal change syndromes such as 5a-reductase deficiency and 17b-hydroxysteroid-dehydrogenase deficiency in 46,XY. At birth, such children appear more female than male and they are typically assigned to the female gender, unless diagnosed correctly. There are a number of clinical–medical reports from resource-poor countries showing that, in the absence of medical intervention, the majority of such individuals start virilizing dramatically during spontaneous puberty and later change to the male gender, but very few data are available on the long-term outcome of such cases when medical intervention is introduced at newborn age or at the beginning of puberty. Similarly, the long-term outcome of those few cases who reportedly have had early hormone replacement therapy to masculinize the genitalia is not yet known. Most controversial is the gender assignment of 46,XY infants who presumably had a normal-male sex-hormonal environment in utero so that the brain was normally masculinized, but who, for nonhormonal reasons, had genital abnormalities, for instance, in cases of penile agenesis, cloacal exstrophy of the bladder, or traumatic loss of the penis due to a circumcision accident. Although there are some wellknown individual cases of later patient-initiated gender reassignment to male, there are other cases where this has not occurred, and in general, the

130 long-term outcome data are much too limited for an evidence-based decision. In the syndrome of true hermaphroditism, the gender assignment decision is based on considerations of the preponderance of masculine or feminine structures in both the external and internal genitalia and the level of anti-Mu¨llerian hormone, but again there is uncertainty regarding the best cutoff point for male versus female assignment, and more long-term followup data are needed for the formulation of a consensus policy. In the case of ‘‘XX males’’ secondary to a translocation of the SRY locus onto an X chromosome and consequent differentiation of a male gonad, there is no doubt that such infants should be raised as boys. Although the rapidly expanding knowledge on multiple genes participating in the sexual differentiation of the gonads and the brain complicates the picture and leads to the discovery of additional genotypes associated with intersex syndromes, it can be expected that over the next decade these discoveries will help refine both the diagnostic picture and the prognostic picture, especially when complemented by data about sexual brain dimorphism from brain imaging.

Conclusion Overall, the available data from both earlier and newer studies can be tentatively summarized in the statement that, with the exception of the pubertal gender-change syndromes, the majority of intersex patients develop a gender identity commensurate with the assigned gender. Later initiation of gender change by an intersex person appears to be the more likely the more the prenatal and postnatal biological factors and the postnatal psychosocial factors push in the same cross-gender direction. It follows that the gender assignment of the newborn should be based on the best prognosis for the future psychosocial and psychosexual functioning of the patient in the given cultural context, taking into account everything that is known about the likely steroid effects on the brain in the syndrome in question and what is known from follow-up investigations.

PSYCHOSOCIAL MANAGEMENT Introduction The issues of gender assignment and genital surgery and the potential implications for the development of gender, sexuality, and parenthood—three highly salient, emotional, moral, and frequently controversial social issues—make intersex conditions a challenge not only for patients and their families, but also for

Gender Assignment and Psychosocial Management

physicians and other professionals involved in their clinical care and psychosocial management. Moreover, the psychosocial management itself is in a period of flux, since the optimal-gender policy of the second half of the twentieth century has come under severe criticism, whereas the evidence on which to base systematic improvements where warranted leaves much to be desired and no systematic randomized trials of psychosocial management procedures for intersex patients have been conducted. In these circumstances, recommendations for specific treatment options need to be more tentative, and not only the providers, but also the parents and, when old enough, the patients themselves will be burdened by more uncertainties in their decision-making than desirable. The professionals involved in the medical and psychosocial management of intersex patients, especially when gender assignment is problematic, may involve multiple disciplines and subspecialties, e.g., specialists in neonatology, pediatric endocrinology, pediatric urology, gynecology, genetics, genetic counseling, and mental health. To minimize the burden on intersex patients and their families, these professionals should form a smoothly operating team and follow a common policy, also with regard to criteria for decisions on gender. Parents of intersex newborns are typically in a state of high stress while the gender of their child remains uncertain, and most are not in a position to decide between conflicting professional opinions. Those team members—both medical and mental health staff—who have the most influence on decisions regarding gender assignment and genital surgery and those who most likely will be involved with the parents and patients over the coming years must keep up with medical advances, emerging new treatment options, the growing research on psychosocial outcomes, and the prognostic alternatives for both health and psychological development in the intersex field. To be fully effective, they need to acquire up-to-date information on all psychosocially relevant aspects of the patient’s syndrome and specific symptoms. For instance, some intersex syndromes may be associated with short stature or neuropsychological impairments, both of which have psychosocial consequences of their own.

Gender Assignment and Reassignment Gender Assignment at Birth The identification of ambiguous genitalia in a newborn only occasionally constitutes a medical emergency. Yet, having a newborn of undetermined

Gender Assignment and Psychosocial Management

p0110

gender is highly stressful for most parents and thus presents a psychosocial emergency and speedy processing of necessary medical tests and rapid decision-making with regard to gender assignment are important. Prolonged periods of nondecision are thought to run the risk of chronically ambiguous or inconsistent sex typing by the family, or of rejection of the child altogether. Given the diverse factors that influence gender identity development in intersex patients, the decision on gender assignment required in a patient with highly ambiguous genitalia is not only based on medical criteria, but also needs to take into account the differential prognosis for psychological (including psychosexual and reproductive) functioning in either gender role as well as the prognosis for how well the family will be able to cope with either decision. The prognostic considerations take into account the various options of genital surgery and sex-hormone treatment and their respective implications for future functioning, especially in terms of gender-role fit, sexuality, reproduction, and overall quality of life. As genderrelated values differ in various societies, cultural factors have significant impact on gender assignment decisions. In many Asian countries, for example, gender assignment to male is more strongly favored, and infertility in the female more stigmatized, than in the West. Intersex counseling at this stage should involve both parents and the inclusion of other family members should be considered. Legally, gender assignment is the parents’ decision, but they are likely to rely on expert medical advice. Along with medical education about the nature and origin of the problem and the medical tests involved, the parents must be adequately informed by their clinicians about the diversity of long-range outcomes and not given overly optimistic assurances. Commensurate with their cognitive capacity, the parents need to be made aware of significant gaps in scientists’ knowledge about intersex disorders or major controversies about management decisions. Such parent counseling requires a careful balance between providing gross oversimplification on the one hand and too much detail on the other, with a resulting paralysis of decision-making. If either parent is not convinced that the gender decision was correct, appropriate child rearing may be in jeopardy. Monitoring Behavioral Development The psychosocial management needs of intersex patients vary with developmental stage. Once the diagnosis is established and the decision on gender assignment made, an approximate timetable for future

131 medical procedures and preventive psychosocial measures can be projected and visits planned accordingly. Again, a team approach in which medical procedures and psychosocial management are closely integrated is highly recommended. Active outreach by the mental health professional may facilitate prevention of psychosocial problems and adherence to medical treatment. As for other congenital disorders, regular monitoring of behavioral development is also recommended for intersex children. This includes monitoring for medical problems and treatment compliance, as needed, and monitoring for cognitive and behavioral problems, with a special emphasis on atypical genderrole behavior. Some parents (and even some children) need help with handling atypical gender-role behavior. If symptoms of gender dysphoria emerge, parents and the child may need some help in dealing with gender issues and in persistent extreme cases the option of gender reassignment needs to be considered. Gender Reassignment Sometimes, intersex children are initially misdiagnosed, especially when born outside of intersexexperienced medical settings, and receive their correct diagnosis later from medical specialists. The definitive diagnosis may alter the prognostic considerations and make a gender reassignment desirable. Such reassignment, if well justified, is usually not a problem during infancy, provided the parents are adequately counseled. After infancy, during the toddler and preschool years, many intersex children develop some degree of gender-atypical behavior, but these rarely indicate a problem of gender dysphoria. Parents who are anxious about their child’s gender atypicality may need reassurance and counseling. Gender reassignment decisions after infancy, which are quite uncommon, should never be based on purely medical considerations, but require careful psychological evaluation (over a prolonged period of time) of the child’s overall behavioral development, with particular attention to the child’s gender-role behavior and to any symptoms of gender dysphoria, as well as to the behavioral niche the child has occupied in the family system. Gender dysphoria and wishes to change one’s gender may also emerge later in adolescence or adult life, usually as the result of a long and gradual process. Even if the problem is of long standing, through counseling the patient can consider various options, and gender change and the attendant hormone treatment and surgical procedures are not the invariable outcome.

132

Genital Surgery

p0145

Some children with intersexuality require genital surgery for medical reasons. In the majority of cases, however, genital surgery has been performed for psychosocial reasons in order to confirm the assigned gender by genital appearance and to thereby facilitate gender-appropriate rearing, help develop a gendertypical body image, and avoid social stigmatization and also to facilitate later peno-vaginal intercourse. Because genital surgery is associated with some risk to sexual tissues and erotic innervation and thereby sexual functioning, a vigorous debate has ensued as to whether such nonmedically necessitated surgery should be delayed until the intersex child is old enough to give informed consent. Surgeons argue that modern surgical techniques are much improved, but it may still require years until the patients so treated are old enough to provide data on sexual functioning. Experienced clinicians are concerned that obviously ambiguous genitalia put an unoperated intersex child at risk of undue attention and stigmatization. Moreover, in the absence of sufficient sexual experience, even an older adolescent or young adult may not be capable of giving appropriately informed consent, and the very fact of genital ambiguity may contribute to the delay of sexual initiation and some patients may never reach that developmental stage. Thus, there is as yet no consensus regarding the issue of early genital surgery, except to say that milder cases of genital ambiguity are less likely to be operated on than they would have been one or two decades ago and that there is a growing consensus that genital surgery should be confined to intersex-experienced centers of excellence. If the decision for genital surgery has been made, both medical and psychological considerations determine the choice of time. From a psychological perspective, genital surgery is performed more easily in infancy, when counseling for the child is not an issue, and in adolescence, when cognitive maturation facilitates counseling and the patient has achieved a degree of autonomy, than in early and middle childhood. The older the child, the more he or she should be empowered to have the decisive vote in the decision for or against genital surgery and in choosing the time when it should take place. Before and after genital surgery, genital examinations of intersex children and adolescents are frequent. Given the low prevalence rates of the various syndromes of intersexuality, it has been common medical practice to have many different physicians, especially medical students and residents, perform genital examinations on the same patient. However,

Gender Assignment and Psychosocial Management

considerable evidence has accumulated showing that this practice has negative and often severe psychological aftereffects. In medical training, this practice should be replaced by the use of photographs, videotapes, and physical models. Also, older children and teenagers should have a say in whether someone, and, if yes, who, should accompany them during the examination. In adolescence and adulthood, psychosocial monitoring needs to include concerns about romantic relationships and sexual functioning, because of the increased problems associated with genital ambiguity and related surgery.

Sex-Hormone Treatment During infancy, androgen treatment is sometimes used for chromosomally male children with underdeveloped male genitals to test for an androgen-receptor defect or, if the assignment decision was for male, to enhance penile size. Occasionally, such children show behavioral change that normalizes with the end of the treatment and the parents may need reassurance. Intersex children who have no gonads or underfunctioning gonads will need sex-hormone replacement therapy to initiate or support pubertal maturation. This is best done at the age when their classmates experience endogenous puberty. Undue delay of pubertal development may add to difficulties in the later development of romantic and sexual relationships.

Information Management The medical and sexual education of intersex children and their parents and related counseling (including information about how to deal with relatives and friends) is a crucially important part of psychosocial management. The most difficult issue is the disclosure of their medical status and history to the intersex patients themselves, especially to those assigned to a gender discrepant with sex chromosomes and/or gonadal structure. Professionals generally support the right of patients to have access to their medical information and many experienced professionals agree that by the time of graduation from high school the patient should be fully informed, commensurate with the level of cognitive capacity. However, many parents are anxious and would like to postpone full disclosure, sometimes permanently. Appropriate timing of disclosure is important. Early disclosure runs into the problem of cognitive limitations. Late disclosure implies cover-up and deception and may lead to the

133

Gender Assignment and Psychosocial Management

patient’s long-term distrust of and anger at medical professionals and even endanger appropriate future utilization of medical services. Disclosure is best conceptualized as a long-term process starting in late preschool age, with many installments, preferably in conjunction with meaningful events. Substantively, there are the issues of past genital surgery and their tell-tale signs, namely, scars on the abdomen, possibly persisting abnormalities in genital appearance, and plans for future surgery; the need for the induction of puberty by sex-hormone treatment and the maintenance of secondary sex characteristics by its lifelong continuation; the issue of infertility or, better, the potential realization of parenthood by means other than pregnancy or insemination; the question of variability in gender-role behavior among individuals of the same core gender identity; and the potential for romance, sexuality, and sexual pair-bonding—regardless of sexual orientation. All of these issues of functional potential are more important to the average developing intersex person than the details of the molecular structure of genes, the variants of hormone synthesis, receptor defects, enzyme abnormalities, and the organization and activation of neural networks—although the occasional adolescent or adult patient may have specialized interests or even choose a career path in the respective biological sciences and may seek as complete an understanding of his or her biological condition as is possible. For the developing child and adolescent, the medical information must be carefully tailored to the cognitive maturity level. The fact that the meaning and connotations of terms differ vastly between medical personnel and lay persons makes a deliberate and cautious choice of terminology necessary. The intended medical ‘‘truth’’ may be dramatically at variance with what the patient or parents perceive. Visual aids are often very helpful. For both the patients and the parents, discussing the medical aspects of intersexuality tends to be highly emotional, so adequate retention is problematic. In the author’s unit, the ideal procedure for the medical education of a newly referred family is that the physician in charge of the overall coordination of care, usually a pediatric endocrinologist or urologist, gives a carefully worded summary of the medical information to the parents, in the presence of the mental health professional, who then takes over. Typically, the parents are asked to recount in detail how they understand the medical condition, its origin, and its prognosis. Then, the instruction is repeated, the

parents’ misunderstandings are corrected, and their particular concerns are addressed. At subsequent visits, the procedure is repeated as necessary, keeping in mind that new misunderstandings and different concerns develop over time. Similar procedures are used with intersex children, once they are old enough, except that, in addition, the clinician lets them explain to their parents (in the clinician’s presence) what they have learned from the instruction, to open the channels of communication.

Support Groups One of the recurrent problems of persons suffering from rare disorders and their families is the feeling of isolation and having no one to talk to who really understands what they must cope with. Many support groups have sprung up, including groups specific for intersex persons and/or their parents, and the internet has greatly aided in this development. Generally, such groups have been a great help to many patients and their families. However, the quality of the groups is highly variable and the medical and other information that comes from them is sometimes quite misleading. Thus, frequent exchanges between support groups and professionals are recommended and group participants should be strongly encouraged to seek second opinions from their physicians and other care providers regarding the information they receive.

See Also the Following Articles Androgen Insensitivity Syndrome . Congenital Adrenal Hyperplasia, Prenatal Diagnosis and Therapy . Endocrine Disrupters and Male Sexual Differentiation . Genes and Gene Defects Affecting Gonadal Development and Sex Determination . Sexual Maturation, Female . Sexual Maturation, Male

Further Reading Cohen-Kettenis, P. T., and Pfa¨fflin, F. (2003). Atypical sexual differentiation. In ‘‘Transgenderism and Intersexuality in Childhood and Adolescence: Making Choices’’ (P. T. Cohen-Kettenis and F. Pfa¨fflin, eds.), Chap. 3, pp. 23–49. SAGE Publications, Thousand Oaks, CA. Cohen-Kettenis, P. T., and Pfa¨fflin, F. (2003). Clinical management of intersex conditions. In ‘‘Transgenderism and Intersexuality in Childhood and Adolescence: Making Choices’’ (P. T. CohenKettenis and F. Pfa¨fflin, eds.), Chap. 5, pp. 85–104. SAGE Publications, Thousand Oaks, CA. Daaboul, J., and Frader, J. (2001). Ethics and the management of the patient with intersex: A middle way. J. Pediatr. Endocrinol. Metab. 14, 1575–1583.

134 Meyer-Bahlburg, H. F. L. (1999). Variants of gender differentiation. In ‘‘Risks and Outcomes in Developmental Psychopathology’’ (H.-C. Steinhausen and F. C. Verhulst, eds.), pp. 298–313. Oxford University Press, Oxford, UK. Meyer-Bahlburg, H. F. L. (2002). Gender assignment and reassignment in intersexuality: Controversies, data, and guidelines for research. In ‘‘Pediatric Gender Assignment: A Critical Reappraisal’’ (S. A. Zderic, D. A. Canning, M. C. Carr, and H. M. Snyder, eds.), pp. 199–223. Kluwer/Plenum, New York.

Gender Assignment and Psychosocial Management

Money, J. (1994). ‘‘Sex Errors of the Body and Related Syndromes: A Guide to Counseling Children, Adolescents, and Their Families,’’ 2nd ed. Paul H. Brookes, Baltimore, MD. Ruble, D. N., and Martin, C. L. (1998). Gender development. In ‘‘Handbook of Child Psychology’’ ( W. Damon, ed.) 5th ed., Vol. 3, pp. 933–1016. Wiley, New York. Wilson, J. D. (1999). The role of androgens in male gender role behavior. Endocr. Rev. 20, 726–737. Zucker, K. J. (1999). Intersexuality and gender identity differentiation. Annu. Rev. Sex Res. 10, 1–69.

136

t0005

Genes and Gene Defects Affecting Gonadal Development and Sex Determination

Table I Genes Involved in Mammalian Gonadal Development and Sex Determination Chromosomal localization Gene

Human

Mouse

Gene function

a

Mutation phenotype

WT1

11p13

2

Transcription factor

XY male-to-female sex reversal Denys-Drash and Frasier syndromes

SF1/AD4BP/NR5A1

9q33

2

Transcription factor

LIM1/LHX1

11p13

11

Transcription factor

XY male-to-female sex reversal Gonadal dysgenesis (mouse KO) Gonadal dysgenesis (mouse KO)

LHX9

1q31

1

Transcription factor

Gonadal dysgenesis (mouse KO)

EMX2

10q26

19

Transcription factor

Gonadal dysgenesis (mouse KO)

SRY

Yp11

Y

Transcription factor

XY gonadal dysgenesis

SOX9

17q24.3–25.1

11

Transcription factor

XY male-to-female sex reversal Campomelic dysplasia

DMRT1

9p24

19

Transcription factor

GATA4

8p23.1–22

14

Transcription factor

XY male-to-female sex reversal Multigene deletion in chromosome 9 encompassing DMRT1 gene XY gonadal dysgenesis (mouse mutation)

FOG2

8q23

15

Transcription factor

XY gonadal dysgenesis (mouse KO)

DAX1

Xp21.3–p21.2

X

Transcription factor

Gene duplication: XY gonadal dysgenesis Abnormal testis cord formation (mouse KO)

M33

17q25

19

Transcription factor

FGF9

13q11–q12

14

Signaling molecule

Delay in genital ridge formation and XY male-to-female sex reversal (mouse KO) XY male-to-female sex reversal (mouse KO)

WNT4

1p35

4

Signaling molecule

AMH/MIS

19p13.3–p13.2

10

Signaling molecule/ hormone

a

XY gonadal dysgenesis in individuals carrying a duplication of a portion of chromosome 1 encompassing the WNT4 gene XX masculinization (mouse KO) Persistent Mu¨llerian duct syndrome in XY individuals

Human data unless indicated otherwise.

Phenotypic sex Chromosomal sex

Gonadal sex

SRY WT1 GATA4 FOG2 SF1 SOX9 M33 DAX1 DMRT1 FGF9

XX or XY

SF1 WT1 SOX9 GATA4 Sertoli cell Testis Leydig cell

Testosterone

Wolffian duct development 5α-reductase

DHT

Virilization of external genitalia

WT1 SF1 LIM1 LHX9 EMX2

Urogenital ridge

Mullerian duct regression

MIS

Bipotential WNT4 gonad

Ovary

f0005

Figure 1 Molecular determinants of mammalian gonadal development and sex determination. DHT, dihydrotestosterone.

137

Genes and Gene Defects Affecting Gonadal Development and Sex Determination

GENES INVOLVED IN EARLY GONADAL DEVELOPMENT A number of genes have been show to be crucial for the earliest steps of gonadal morphogenesis. These include WT1, SF1, LIM1/LHX1, LHX9, and EMX2. Since these genes act prior to sex determination, they are required for early gonadal development in both sexes. Mice deficient in these genes either fail to develop genital ridges or have an early block in genital ridge development. Human mutations involving the LIM1, LHX9, and EMX2 genes have not yet been linked to cases of human gonadal agenesis or dysgenesis. The WT1 and SF1 genes merit further discussion since these genes are known to be required for both early gonadal development and sex differentiation.

WT1 The W T1 gene, encoding a DNA-binding protein containing four zinc-fingers, was first identified as a tumor suppressor gene via genetic studies of Wilms’ tumor patients in humans. Mutations in the W T1 gene in humans were shown to be responsible for both Denys-Drash syndrome and Frasier syndrome, characterized by anomalies in kidney development and gonad formation including sex reversal. Mice homozygous for a null mutation in the Wt1 gene show a phenotype of renal and gonadal agenesis and adrenal dysgenesis, demonstrating that Wt1 is required for genital ridge formation prior to the events of sex determination. It is evident that the W T1 gene can result in multiple protein isoforms via posttranscriptional processes including alternative splicing; these isoforms can function not only as transcription factors but also as RNA-processing factors. Important among these isoforms is the presence of a 3-amino-acid insertion (þKTS) between the third and fourth zinc-fingers or the absence of the insertion (KTS). The WT1 (KTS) isoform functions as a transcription factor that can activate the AMH and DAX1 promoters as well as the SRY promoter in vitro. The WT1 (þKTS) isoform shows little transactivation activity but has been implicated in RNA-processing events. Isoformspecific null allele mouse studies have been reported and they show that the Wt1 (þKTS) null allele results in a phenotype of XY sex reversal. This indicates that WT1 (þKTS) isoforms as well as RNA processing are important contributors to the molecular mechanisms leading to sex determination.

SF1 Steroidogenic factor 1 (SF1; also known as Ad4BP and NR5A1) is a transcription factor belonging to the nuclear receptor superfamily. SF1 was first identified as an essential regulator of the P450 hydroxylases in the adrenals and gonads. Later studies showed that SF1 is also present in the pituitary and ventromedial nucleus of the hypothalamus (VMH). In agreement with its expression pattern, Sf1 knockout mice lack adrenal glands and gonads, exhibit VMH and pituitary gonadotrope abnormalities, and display female internal genitalia. At the transcriptional level, SF1 regulates the expression of many genes involved in reproduction, steroidogenesis, and sexual differentiation. An important SF1 target for sexual differentiation is the anti-Mu¨llerian hormone gene (AMH/ MIS) involved in Mu¨llerian duct regression in the developing male embryo. In humans, the role of SF1 in male sex differentiation is supported by the identification of SF1 gene mutations in a subpopulation of 46,XY sex-reversed patients: a dominant de novo heterozygous G35E mutation, a recessive homozygous R92Q mutation, and a heterozygous deletion of eight nucleotides causing a frameshift mutation and C-terminal truncation of the SF1 protein. For two of these SF1 mutations (G35E and R92Q), the presence of normal Mu¨llerian structures and the retention of a uterus in the affected individuals is indicative of insufficient AMH production. This suggests that the mutated SF1 proteins might interfere with the expression of a Sf1 target gene(s) involved in the male sexual differentiation pathway, including AMH.

GENES INVOLVED IN SEX DETERMINATION AND DIFFERENTIATION The era of molecular genetics of mammalian sex determination was launched more than a decade ago with the identification and cloning of the elusive Y chromosome-linked testis determining gene, termed SRY. Since then, a multitude of additional genes have been described as being involved in testis differentiation; these include SOX9, DMRT1, GATA4, FOG2, DAX1, and FGF9.

SRY The first major advance in the molecular characterization of mammalian sex determination was reported

138 in 1990 with the positional cloning of the SRY gene, a candidate for the testis-determining factor (TDF) located on the Y chromosome. The equivalence of SRY and TDF was subsequently demonstrated via gain-offunction studies in the transgenic mouse, whereby a transgene coding for Sry caused an XX animal to develop with a male phenotype. Biochemical studies showed that SRY can bind to DNA via a structural motif termed an HMG-box and mutations in the HMG-box were shown to be of clinical relevance in cases of XY females in the human population. Structural comparisons of the SRY gene between mammalian species revealed a notable lack of sequence conservation. Although this is unremarkable for a gene located on the Y chromosome, it was not anticipated for such a key player in the mammalian sex determination process. Also surprising is the fact that after 13 years of study, no direct target genes of SRY have yet been identified, and the mechanisms by which SRY causes the activation of male-specific gene expression, cell proliferation, and cell migration within the male genital ridge remain obscure. However, some progress has been made in deciphering SRY promoter function, with in vitro evidence that WT1, SF1, SOX9, and GATA4 can activate SRY gene transcription. Genetic evidence suggests that DAX1 competes with SRY activity and additional in vitro evidence suggests that DAX1 inhibits the function of proteins that activate SRY activity.

SOX9 p0040

The description of the HMG-box of SRY allowed the cloning of a family of related HMG-box-containing genes, termed SOX (SRY-related HMG-box) genes. One of these new SOX family members, SOX9, was associated via positional cloning with the musculoskeletal disease campomelic dysplasia and also with autosomal sex reversal in humans. Demonstration of SOX9 gene expression within the male genital ridge at approximately the time of sex determination in mammals as well as in birds suggested a conserved role for SOX9 in sex determination and/or sex differentiation. Transgenic expression of SOX9 within the genital ridge in the absence of SRY resulted in a male phenotype within XX animals, suggesting that SOX9 expression is sufficient for sex determination and can account for downstream functions of SRY expression. Like SRY, SOX9 contains a DNA-binding HMG-box domain, but unlike SRY, it also contains a recognizable transactivation domain. A key event in sex differentiation is the activation of the AMH gene, which codes for a secreted protein of the TGF-b family.

Genes and Gene Defects Affecting Gonadal Development and Sex Determination

Sox9 has been shown by both in vitro and in vivo studies to transactivate the Amh promoter. In vitro evidence suggests that at least in some species, SOX9 can also transactivate the SRY promoter, placing it both upstream and downstream of sex determination. Whether SOX9 is necessary for sex determination per se, or can be compensated for by other SOX genes known to be expressed within the genital ridge, awaits further experimentation.

DMRT1 DMRT1 was initially identified and cloned via a conserved DNA-binding domain sequence (DM domain) identified within the Drosophila melanogaster Doublesex gene, the Caenorhabditis elegans Mab-3 sex regulator gene, and human expressed sequences derived from a testis cDNA library. The chromosomal location of human DMRT1 (9p) makes it a candidate gene responsible for XY gonadal dysgenesis seen with 9p deletions in humans. This region is syntenic with the chicken Z chromosome, making chicken DMRT1 a candidate for a dosage-sensitive sex determination system in birds. Unlike most other genes involved in sex determination and differentiation that show fairly wide tissue expression patterns, DMRT1 expression is restricted to the gonads. Furthermore, DMRT1 expression is up-regulated concurrently with testis development, not just in humans and mice but also in other vertebrates. Male mice that are deficient for the Dmrt1 gene display problems with cell differentiation and cell survival in the postnatal testis, but these animals undergo normal sex determination. These results could indicate that DMRT1 is simply not involved in sex determination or that functional redundancy can compensate for the absence of DMRT1; there are seven DMRT genes in mouse and humans. A DMRT family member (Dmrt1bY) was proposed as the sex-determining factor in the Medaka fish; however, this gene was found to be absent in closely related fish species. Thus, although DMRT1 sequences are widely distributed within the animal kingdom and are unquestionably implicated in gonadal development, a universal involvement of DMRT1 sequences in sex determination does not appear likely. Where mammalian DMRT1 integrates with other genes known to be involved in mammalian sex determination and differentiation pathway also remains to be determined.

GATA4 GATA4 is one of six proteins belonging to the GATA family of transcriptions factors. GATA factors are

139

Genes and Gene Defects Affecting Gonadal Development and Sex Determination

named for the nucleotide sequence (WGATAR) that they bind to in the promoter regions of target genes. They were originally identified as crucial regulators of cardiac development and hematopoietic cell differentiation. GATA expression, however, is not limited to these two systems. In the mouse, the Gata4 gene is abundantly expressed in the somatic cell population of the developing genital ridge prior to and during the time of Sry expression. Thus, based on its expression pattern, GATA4 was proposed to play a central role in early gonadal development and sex determination. This hypothesis has been confirmed in the mouse, where in vivo disruption of GATA4 function via a mutation of the Gata4 gene leads to a block in testis development and a marked down-regulation of Sry expression. Thus, GATA4 appears to function as a direct upstream regulator of SRY expression in the developing testis. Although the latter has yet to be conclusively demonstrated, the presence of multiple GATA regulatory motifs in the mouse, human, and pig SRY promoters strongly supports this notion. After gonadal differentiation, GATA4 has been shown to be recruited as an important regulator of the AMH gene involved in Mu¨llerian duct regression and, consequently, male sexual differentiation. GATA4 has been shown to regulate both the mouse and human AMH promoters through direct transcriptional cooperation with SF1. Although no human GATA4 gene mutations have yet been linked to gonadal defects, evidence suggests that disruption of GATA4/SF1 synergism may account for some cases of human male-to-female sex reversal involving insufficient AMH expression.

FOG2 GATA factors regulate the expression of target genes through cooperative interactions with other transcription factors. These include the Friend of GATA proteins (FOG1 and FOG2), which were cloned as GATA-specific cofactors. Although the FOG proteins do not directly bind to DNA, they have been shown to act as either enhancers or repressors of GATA transcriptional activity depending on the cell and promoter context being studied. In this regard, it has been suggested that the FOG proteins act as bridging molecules that link GATA proteins with other factors involved in either activation or repression. In the mouse, Fog2 is coexpressed with Gata4 in the developing genital ridge. Like Gata4, Fog2 is likely expressed upstream of Sry in the sex determination pathway since Fog2 knockout mice have markedly diminished Sry expression and do not develop a normal testis.

DAX1 The X chromosome-linked DAX1 [dosage-sensitive sex reversal-adrenal hypoplasia congenital (AHC) critical region on the X chromosome, gene 1; NR0B1] gene encodes an atypical member of the nuclear receptor superfamily. DAX1 is expressed in several endocrine tissues, including the gonads, where it is present in the Sertoli and Leydig cells of the testis and the granulosa and theca cells of the ovary. DAX1 plays a critical role in adrenal development and gonadal function; it was originally identified as the causative gene for AHC associated with hypogonadotropic hypogonadism (HHG), a recessive disorder affecting males. Indeed, several different groups have characterized numerous DAX1 mutations (nucleotide substitutions, insertions, and deletions) in individuals affected with AHC and HHG. Females carrying heterozygous mutations in the DAX1 gene are normal and there has been no report of homozygous females since males needed to transmit the nonfunctional allele are infertile. Since Dax1 expression is down-regulated in the developing male gonad after overt testis differentiation in the mouse, the Dax1 gene was initially believed to be an important ovarian determinant. However, female Dax1 knockout mice are phenotypically normal. In contrast, male Dax1 knockout mice are infertile due to a loss of germ cells. Thus, Dax1 is apparently not required for ovarian differentiation but rather is essential for the maintenance of the integrity of the seminiferous epithelium of the testis. Closer examination of Dax1-deficient mice has revealed that Dax1 is also essential for testis cord formation and, consequently, testis determination. In addition to DAX1 mutations, another human disorder has been associated with the DAX1 gene. In this case, duplication of a small region on the X chromosome that encompasses the DAX1 gene leads to dosagesensitive sex reversal in XY males. Overexpression of Dax1 in transgenic mice has also been shown to induce sex reversal in males carrying a weakened Sry allele. Thus, adequate Dax1 levels are required for normal testis development and function, but too much Dax1 appears to have an ‘‘anti-testis’’ effect. In the female developing gonad, Dax1 transcription appears to be up-regulated by Wnt4, a member of the Wnt family of signaling molecules. Wnt4 signaling has been shown to function via activation of b-catenin and its subsequent interaction with Sf1 on the Dax1 promoter. In the developing male gonad, the action of Wnt4 has been suggested to be inhibited by Sry.

140

AMH In mammals, male sexual differentiation is regulated by two hormones produced by the fetal testis: testosterone, secreted by Leydig cells, and AMH, produced by Sertoli cells. In the male embryo, AMH induces regression of the Mu¨llerian ducts (the anlagen of the internal female reproductive tract). AMH is the earliest marker of testis formation; it is found in Sertoli cells of the developing human testis starting at approximately 8 weeks of gestation. In the mouse, Amh is first detected in Sertoli cells on embryonic day 12.5; its expression remains high throughout fetal life before declining markedly after birth. Consistent with its role in male sexual differentiation, the absence of AMH expression in humans causes persistent Mu¨llerian duct syndrome, a form of pseudo-hermaphroditism characterized by the retention of Mu¨llerian duct structures. The transcriptional regulation of the AMH gene has been intensely studied. Interestingly, several transcription factors involved in primary sex determination (SF1, WT1, SOX9, and GATA4) are later recruited as important regulators of AMH expression. Although many of these factors have been shown to activate AMH transcription on their own by directly binding to the AMH promoter, functional cooperation between these factors appears to be of paramount importance in directing the proper spatiotemporal expression of the AMH gene.

OTHER GENES INVOLVED IN GONADAL DEVELOPMENT AND SEX DETERMINATION Ever since the discovery of SRY, the majority of newly identified genes involved in gonadal development and sex determination have been transcription factors. Signaling molecules such as FGF9 and WNT4 have also been implicated in these processes. In the mouse, Fgf9 was shown to be involved in the induction of mesonephric cell migration into the XY gonad, an essential step required for seminiferous cord formation and, hence, proper testis organization. Wnt4 expression is detected in the genital ridge of both sexes but its expression becomes largely restricted to the ovary after overt gonadal differentiation. Wnt4-deficient female mice are masculinized due to the expression of genes involved in testosterone biosynthesis that are normally not expressed in the fetal ovary. Much like DAX1, overexpression of WNT4 in humans has been reported to cause male-to-female sex reversal; this

Genes and Gene Defects Affecting Gonadal Development and Sex Determination

anti-testis effect of too much WNT4 is thought to occur via up-regulation of DAX1.

PERSPECTIVES Although significant progress has been made in elucidating how the different genes in the gonadal development and sex determination pathway come together to form a complex regulatory network, it remains evident that many pieces of the genetic puzzle are still missing. Indeed, the majority of the cases of human XY sex reversal and approximately one-quarter of the cases of XX maleness have not yet been defined at the genetic level. Technologies for large-scale gene identification based on differential expression have the potential for making important inroads into finding new genes in this fundamental developmental process. For example, expression-based strategies such as gene microarray technology have begun to be successfully applied to the discovery of genes involved in testis and ovary development. The vanin 1 gene, which encodes a glycosylphosphatidylinositol-anchored cell surface protein, is an example of a gene identified through the sex-specific screening process. Clearly, other genes are likely to be identified in coming years. This is especially true for genes that might control ovarian development, an area that has been long overlooked by the drive to study testis differentiation.

See Also the Following Articles Agonadism, Male and Female . Androgens, Gender and Brain Differentiation . Anti-Mu¨llerian Hormone . Delayed Puberty and Hypogonadism, Male . Endocrine Disrupters and Male Sexual Differentiation . Sexual Maturation, Female . Sexual Maturation, Male

Further Reading Cotinot, C., Pailhoux, E., Jaubert, F., and Fellous, M. (2002). Molecular genetics of sex determination. Semin. Reprod. Med. 20, 157–167. Jameson, J. L., Achermann, J. C., Ozisik, G., and Meeks, J. J. (2003). Battle of the sexes: New insights into genetic pathways of gonadal development. Trans. Am. Clin. Climatol. Assoc. 114, 51–63. Novartis Foundation Symposium (2002). ‘‘The Genetics and Biology of Sex Determination’’ (D. Chadwick and J. Goode, eds.). Wiley, West Sussex, UK. Parker, K. L., and Schimmer, B. P. (2002). Genes essential for early events in gonadal development. Ann. Med. 34, 171–178. Tilmann, C., and Capel, B. (2002). Cellular and molecular pathways regulating mammalian sex determination. Rec. Prog. Horm. Res. 57, 1–18.

142

p0020

p0030

unclear because autosomal dominant disorders usually occur as a result of either haploinsufficiency or secondary to dominant negative activity. IGHD-3 may be associated with hypogammaglobulinemia and is inherited as an X-linked disorder. No mutations in GH1 or other candidate genes have been identified for this type. Mutations causing MPHD have been identified in the genes encoding pituitary transcription factors that direct the embryonic development of the anterior pituitary gland, including POU1F1 (formerly referred to as Pit-1), PROP1, HESX1, and LHX3 and LHX4. Autosomal recessive and dominant mutations of POU1F1 (POU domain, class 1, transcription factor 1), which is located on chromosome 3p11, result in deficiency of GH, prolactin, and thyroid-stimulating hormone (TSH); adrenocorticotropin hormone (ACTH) and the gonadotropins are spared. Autosomal recessive mutations of PROP1 (Prophet of Pit-1), located on chromosome 5, cause deficiency of GH, prolactin, and TSH, as well as the gonadotropins. ACTH is spared initially, but there is a tendency toward the development of deficiency with increasing age. Mutations of HESX1 (also referred to as RPX ), a pituitary transcription factor that also plays a role in the development of the optic nerves, have been implicated in septo-optic dysplasia. This is a heterogeneous disorder characterized by midline neurological abnormalities associated with pituitary hypoplasia and optic nerve hypoplasia. Mutations of LHX3 have been identified in humans with abnormal neck and cervical spine development in whom there is a deficiency of all anterior pituitary hormones except ACTH. Mutations have been identified in LHX4 in a family with GH, TSH, and ACTH deficiency in combination with cerebellar and skull-base malformations. Potentially, inactivating mutations of SST, or one of the five specific G protein-coupled receptor subtypes (SSTR1–5) to which somatostatin binds, could lead to loss of inhibition and subsequent excess GH production. Although a mutation of the sst5 gene has been demonstrated in a patient with a GH-secreting pituitary adenoma, the simultaneous presence of a gsp mutation makes the role of the SSTR5 mutation unclear. Activating Mutations GHRHR signals through a Gsa-containing G protein. Activating somatic mutations of the gene encoding the Gsa subunit give rise to the gsp oncogene. These gsp mutations have been identified as the molecular cause of 30–40% of GH-producing pituitary tumors. They have also been found in GH-producing tumors, which occur in patients with McCune–Albright

Genetic Testing for Pituitary Disease

syndrome. The mutations result in continuous and markedly elevated levels of intracellular cAMP, with subsequent increases in the secretion of GH and the proliferation of somatotrophs.

Thyroid-Stimulating Hormone TSH is a glycoprotein hormone that consists of a and b subunits. The a subunit is common to TSH, luteinizing hormone (LH), follicle-stimulating hormone (FSH), and human chorionic gonadotropin, whereas the b subunit in each of these hormones is unique and confers specificity of action. Production of TSH is stimulated by thyrotropin-releasing hormone (TRH) and inhibited by triiodothyronine.

p0035

Inactivating Mutations Central hypothyroidism, which is a rare disorder characterized by insufficient production of TSH by the anterior pituitary, occurs most commonly as part of MPHD. Rare cases of isolated TSH deficiency have been described, typically resulting from missense, nonsense, or frameshift mutations of the TSH-b gene on chromosome 1; a nonsense mutation of the TRHR gene has also been identified in a patient with isolated central hypothyroidism. Central hypothyroidism secondary to the production of a mutant TSH molecule has been demonstrated in two families in which the affected members demonstrated thyroid hormone levels consistent with hypothyroidism but TSH levels that were inappropriately in the normal range. Activating Mutations TSH-secreting pituitary adenomas are a rare cause of central hyperthyroidism. Potential molecular causes that have been investigated, such as activating mutations of the Gsa subunit, TRHR and POU1F1, have not been identified.

p0045

Adrenocorticotropin Hormone ACTH is a peptide hormone derived from a precursor polypeptide, pro-opiomelanocortin (POMC). The POMC gene encodes several peptides, including an N-terminal peptide, b-lipotropin, and ACTH. Synthesis of ACTH is primarily stimulated by hypothalamic corticotropin-releasing factor (CRF) through its interaction with the CRF receptor, another G protein-coupled receptor. Inactivating Mutations Although rare cases of isolated ACTH deficiency have been described, deficiency of this hormone usually

p0055

143

Genetic Testing for Pituitary Disease

occurs as part of MPHD. ACTH deficiency can be simulated by isolated glucocorticoid deficiency because there is resistance to the action of ACTH at the level of the adrenal cortex. In approximately half of cases identified with this disorder, there is a demonstrable mutation in the ACTH receptor (also known as the melanocortin-2 receptor). This condition, referred to as GCCD1, is inherited in an autosomal recessive manner. Subjects are biochemically characterized by low serum cortisol and high ACTH levels, with normal mineralocorticoid activity. Activating Mutations The prevalence of gsp mutations in ACTH-secreting pituitary adenomas has been demonstrated to be less than 10%.

Luteinizing Hormone and Follicle-Stimulating Hormone The pituitary gonadotropins LH and FSH, which belong to the family of glycoprotein hormones, are produced by the anterior pituitary in response to stimulation by the hypothalamic peptide gonadotropin-releasing hormone (GnRH). The a subunit, common to both LH and FSH, is encoded by a gene on chromosome 6q12.21. The hormone-specific LH b subunit is encoded by the LHb gene on chromosome 19q13.32, whereas the FSH b subunit is encoded by the FSHb gene on chromosome 11p13.

p0070

Inactivating Mutations Five subjects with isolated FSH deficiency have been reported with inactivating mutations of FSHb. Two different mutations (a 2-base pair deletion and a missense mutation) were described in three women who presented with primary amenorrhea, poorly developed secondary sexual characteristics, and infertility. Mutations of FSHb have also been reported in two men, one of whom presented with azoospermia but normal puberty, and the other with azoospermia and delayed puberty. These mutations consisted of the same 2-base pair deletion detected in the females plus a new missense mutation. Only one homozygous missense mutation causing inactivation of LHb has been reported in a male, who presented with delayed puberty. No mutations of the human a subunit gene have been described in the literature. Hypogonadotropic hypogonadism (HH), which results from deficiency of both gonadotropins, can occur in isolation, in association with anosmia such

as in Kallmann’s syndrome (KS), or with adrenal insufficiency such as in adrenal hypoplasia congenita. Inactivating mutations of the gene encoding the GnRH receptor (GnRHR), a member of the G protein-coupled receptor family located on chromosome 4q21.2, have been reported in isolated HH. The defects in KS and adrenal hypoplasia congenita both occur at the hypothalamic rather than pituitary level. Briefly, KS is a form of HH that is associated with anosmia/hyposmia. X-linked recessive KS results from mutations of the KAL1 gene that is located in the pseudo-autosomal region of Xp. This gene encodes the KAL protein (also known as anosmin), which plays a central role in the migration of both GnRH and olfactory neurons during embryonic development. However, the majority of KS is not X-linked and has been shown to have either autosomal recessive or dominant patterns of inheritance in familial cases. Adrenal hypoplasia congenita (AHC) is a rare X-linked disorder characterized by primary adrenal insufficiency and HH. It is caused by missense, nonsense, and frameshift mutations of the DAX-1 gene (dosage-sensitive sex reversal, AHC-critical region of the X chromosome, gene 1), located on the short arm of the X chromosome Xp21. More than 60 different mutations of DAX-1, which is expressed in the hypothalamus, pituitary, adrenals, and gonads, have been reported in X-linked AHC.

p0080

Activating Mutations These mutations have not been identified in the LHb, FSHb, or GnRHR genes.

Prolactin Prolactin (PRL) is a polypeptide hormone encoded by the PRL gene on chromosome 6. Its release is stimulated by a number of prolactin-releasing factors, including vasoactive intestinal peptide, TRH, and PRL-releasing peptide. It is inhibited by prolactininhibiting factors, predominantly dopamine but possibly also by the 56-amino acid portion of the precursor to GnRH, known as GnRH associated peptide. Inactivating Mutations Isolated PRL deficiency is extremely rare, usually occurring as a component of MPHD. Activating Mutations No such mutations of genes encoding receptors for the hypothalamic factors or of PRL have been identified.

p0090

144

Pituitary Disease Associated with Other Disorders Multiple Endocrine Neoplasia This disorder is characterized by the presence of tumors involving two or more endocrine glands in a single patient. The two major forms, multiple endocrine neoplasia 1 (MEN1) and 2 (MEN2, which comprises 2A and 2B), are inherited as autosomal dominant conditions, with penetrance that increases with age. MEN1 is characterized by tumor development in the parathyroid glands, pancreatic islet cells, and the anterior pituitary. Biochemically, approximately 60% of MEN1-associated pituitary tumors secrete prolactin, less than 25% secrete GH, and 5% secrete ACTH. The remainder appear to be nonfunctioning tumors. There is variability in the phenotype, although some manifestations tend to be more common. This variability also occurs within family members of an affected kindred. Since penetrance is incomplete (especially early in life), MEN should always be considered in the diagnosis of ‘‘isolated’’ endocrine tumors, especially when familial. MEN1 tumors are caused by inactivation of a tumor suppressor gene, MEN1. This gene is located on chromosome 11q13 and consists of 10 exons that encode a 610-amino acid protein, referred to as menin. Approximately 140 germ-line mutations of MEN1 have been identified: approximately 25% are nonsense mutations, 45% deletions, 15% small insertions, 10% missense mutations, and less than 5% donor splice mutations.

Carney Complex Carney complex is an autosomal dominant condition that may be considered a multiple endocrine neoplasia syndrome. Endocrine tumors in Carney complex may involve the adrenal gland, the thyroid, the gonads, and the anterior pituitary gland. The majority of the pituitary tumors secrete GH and prolactin. The disorder is caused by mutations in one of two genes—the tumor suppressor gene PRKAR1 (the type 1a regulatory subunit of cAMP-dependent protein kinase A), located on chromosome 17q, or an unknown gene on chromosome 2p16.

POSTERIOR PITUITARY Arginine Vasopressin Arginine vasopressin (AVP) is synthesized in the supraoptic nuclei of the hypothalamus and transported within axons to the posterior lobe of the

Genetic Testing for Pituitary Disease

pituitary for storage. The gene encoding AVP is arginine vasopressin–neurophysin II (AVP–NPII), located on chromosome 20p13. It encodes the precursor protein of AVP, consisting of AVP (i.e., antidiuretic hormone) and neurophysin II, the carrier protein for AVP. Diabetes insipidus may result from a mutation of the AVP portion or the neurophysin portion of the gene. Deficiency of AVP results in central or neurogenic diabetes insipidus. There are a number of genetic forms, such as familial neurohypophysial diabetes insipidus (FNDI) and Wolfram’s syndrome. FNDI is a rare autosomal dominant syndrome in which more than 35 different germ-line mutations of AVP–NPII have been reported. Wolfram’s syndrome, also known as DIDMOAD (diabetes insipidus, diabetes mellitus, optic atrophy, and deafness), is an autosomal disorder that is caused by inactivating mutations of the gene encoding wolframin (WFS1) on chromosome 4p16.1.

Acknowledgement We thank Brian Betensky for his editorial assistance in the preparation of the manuscript.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . FSH (FollicleStimulating Hormone) . Gigantism: Excess of Growth Hormone . Growth Hormone Deficiency, Genetic . Growth Hormone (GH) . LH (Luteinizing Hormone) . Multiple Endocrine Neoplasia (MEN) Type 2 . Pituitary Gland Anatomy and Embryology . Pituitary Tumors, Molecular Pathogenesis . Prolactin (PRL) . TSH (Thyroid-Stimulating Hormone; Thyrotropin)

Further Reading Achermann, J. C., and Jameson, J. L. (2001). Advances in the molecular genetics of hypogonadotropic hypogonadism. J. Pediatr. Endocrinol. Metab. 14(1), 3–15. Achermann, J. C., Ozisik, G., Meeks, J. J., and Jameson, J. L. (2002). Genetic causes of human reproductive disease. J. Clin. Endocrinol. Metab. 87(6), 2447–2454. Brandi, M. L., Gagel, R. F., Angeli, A., Bilezikian, J. P., BeckPeccoz, P., Bordi, C., Conte-Devolx, B., Falchetti, A., Gheri, R. G., Libroia, A., Lips, C. J., Lombardi, G., Mannelli, M., Pacini, F., Ponder, B. A., Raue, F., Skogseid, B., Tamburrano, G., Thakker, R. V., Thompson, N. W., Tomassetti, P., Tonelli, F., Wells, S. A., Jr., and Marx, S. J. (2001). Guidelines for diagnosis and therapy of MEN type 1 and type 2. J. Clin. Endocrinol. Metab. 86(12), 5658–5671. Cogan, J. D., and Phillips, J. A. (1998). Growth disorders caused by genetic defects in the growth hormone pathway. Adv. Pediatr. 45, 337–361. Dattani, M., Martinez-Barbera, J., Thomas, P., Brickman, J., Gupta, R., Martensson, I., Toresson, H., Fox, M., Wales, J., Hindmarsh,

Genetic Testing for Pituitary Disease

P., Krauss, S., Beddington, R., and Robinson, I. (1998). Mutations in the homeobox gene HESX1/Hesx1 associated with septo-optic dysplasia in human and mouse. Nature Genet. 19, 125–133. Gertner, J. M., Wajnrajch, M. P., and Leibel, R. L. (1998). Genetic defects in the control of growth hormone secretion. Horm. Res. 49(Suppl. 1), 9–14. Layman, L. C. (1999). Genetics of human hypogonadotropic hypogonadism. Am. J. Med. Genet. 89(4), 240–248. Machinis, K., Pantel, J., Netchine, I., Leger, J., Camand, O., Sobrier, M., Dastot-Le, M. F., Duquesnoy, P., Abitbol, M., Czernichow, P., and Amselem, S. (2001). Syndromic short stature in patients with a germline mutation in the LIM homeobox LHX4. Am. J. Hum. Genet. 69, 961–968. Marx, S., Spiegel, A. M., Skarulis, M. C., Doppman, J. L., Collins, F. S., and Liotta, L. A. (1998). Multiple endocrine neoplasia type 1: Clinical and genetic topics. Ann. Internal Med. 129(6), 484–494. Netchine, I., Sobrier, M., Krude, H., Schnabel, D., Maghnie, M., Marcos, E., Duriez, B., Cacheux, V., Moers, A., Goossens, M., Gruters, A., and Amselem, S. (2000). Mutations in LHX3 result in a new syndrome revealed by combined pituitary hormone deficiency. Nature Genet. 25, 182–186. Parks, J. S., Brown, M. R., Hurley, D. L., Phelps, C. J., and Wajnrajch, M. P. (1999). Heritable disorders of pituitary development. J. Clin. Endocrinol. Metab. 84(12), 4362–4370. Rittig, S., Robertson, G. L., Siggaard, C., Kovacs, L., Gregersen, N., Nyborg, J., and Pedersen, E. B. (1996). Identification of 13 new mutations in the vasopressin-neurophysin II gene in 17

145 kindreds with familial autosomal dominant neurohypophyseal diabetes insipidus. Am. J. Hum. Genet. 58, 107–117. Robinson, A. G., and Verbalis, J. G. (2003). Posterior pituitary. In ‘‘Larsen Williams Textbook of Endocrinology’’ ( J. Verbalis, ed.), 10th ed., pp. 291–292. Saunders, Philadelphia. Sandrini, F., and Stratakis, C. (2003). Clinical and molecular genetics of Carney complex. Mol. Genet. Metab. 78(2), 83–92. Spiegel, A. M. (1997). Inborn errors of signal transduction: Mutations in G proteins and G protein-coupled receptors as a cause of disease. J. Inherited Metab. Dis. 20(2), 113–121. Spiegel, A. M. (1997). The molecular basis of disorders caused by defects in G proteins. Horm. Res. 47(3), 89–96. Spiegel, A. M. (2000). G protein defects in signal transduction. Horm. Res. 53(Suppl. 3), 17–22. Stratakis, C. A. (2001). Clinical genetics of multiple endocrine neoplasias, Carney complex and related syndromes. J. Endocrinol. Invest. 24(5), 370–383. Stratakis, C. A., Kirschner, L. S., and Carney, J. A. (2001). Clinical and molecular features of the Carney complex: Diagnostic criteria and recommendations for patient evaluation. J. Clin. Endocrinol. Metab. 86(9), 4041–4046. Thakker, R. V. (2001). Multiple endocrine neoplasia. Horm. Res. 56(Suppl. 1), 67–72. Themmen, A. P. N., and Huhtaniemi, I. T. (2000). Mutations of gonadotropins and gonadotropin receptors: Elucidating the physiology and pathophysiology of pituitary-gonadal function. Endocr. Rev. 21(5), 551–583. Tsigos, C. (1999). Isolated glucocorticoid deficiency and ACTH receptor mutations. Arch. Med. Res. 30(6), 475–480.

147

Germ Cell Differentiation Signaling Events, Male

CELL SIGNALING DURING SPERMATOGENESIS Differentiation of male germ cells within the seminiferous epithelium proceeds in a complex series of stages, including the multiplication of spermatogonia, the successive phases of meiosis, the extensive remodeling of the nucleus and cellular structures during the haploid phase, and the final release of sperm cells. The whole process depends on a coordinated network of endocrine, paracrine, and cell-to-cell communication involving a variety of somatic partners. The kinds of extracellular molecules that are involved in signaling to the diverse types of spermatogenic cells are thought to be numerous, and not all have been identified. The most common are the gonadotropins luteininzing hormone (LH) and follicle-stimulating hormone (FSH), which have their target cells in the testicular somatic Leydig cells and Sertoli cells, respectively. Under hormonal signaling, Leydig cells secrete the male hormone testosterone, whereas Sertoli cells, under stimulation by FSH and testosterone, produce factors that induce germ cell proliferation, meiosis, and spermiogenesis. Among the well-established Sertoli factors are inhibins, activins, and follistatin. In addition to the endocrine action of these peptides, which are generally considered endocrine regulators of FSH secretion, experimental evidence shows that activins, inhibins, and follistatin act locally within the testis as paracrine and autocrine factors involved in the regulation of spermatogenesis. Some studies suggest that inhibin can inhibit spermatogonial proliferation, whereas activin A seems to stimulate spermatogonial proliferation and follistatin may neutralize the action of activin. Other Sertoli factors reported to have a possible role in male germ cell differentiation are specific neurotropins and cytokines and also putative ‘‘contacting’’ molecules—that is, the membrane components with selective expression and exposure on the Sertoli cell’s surface that exert a specific stage- or age-related action on spermatogenesis. How is the message carried by so many different types of signaling molecules transduced in male germ cells? The answer is not known because of the inability to culture male germ cells for long periods of time. In fact, in contrast to cultured cells, no male germ cell lines are available for transfection experiments or to assay the signaling pathway evocated by a growth factor, an activator or inhibitor, a mutagen, and so on. Only recently have successful conditions for culturing primary spermatocytes and spermatogonia been reported; these may provide the groundwork

for future physiological or transfection studies. However, no cell lines are able to differentiate—that is, to undergo the spermiogenetic process to yield spermatozoa. Therefore, the only way to study spermatogenesis physiologically is to test in vivo-specific constructs by creating individual transgenic mice or mutant mice. The question is, what kind of mutant animal? To find an answer to this question, one has to ask the following: What are the major mechanisms regulating cell signaling?

PROTEIN PHOSPHORYLATION Protein phosphorylation is generally accepted as the universal tool that cells have developed to switch on or off dynamic processes. Phosphorylation can regulate the activity of transcription factors at multiple levels, including nuclear transport, dimerization, DNA binding and transcriptional activation, as well as a series of enzyme-catalyzed reactions, including enzyme-linked membrane receptor signaling, G protein-linked receptor signaling, and nutritional functions such as glycogen breakdown. It can also regulate complex phenomena, such as learning and memory. Consequently, protein phosphorylation may play a key role in spermatogenesis; this has been recognized for at least 40 years. How can the hierarchy of phosphorylation events that constitute a signaling pathway be identified?

c-kit/SCF Signaling Under cAMP and FSH stimulation, Sertoli cells express a growth factor that exists in both a soluble form and a transmembrane form and that is known as steel factor or stem cell factor (SCF). The SCF receptor is the c-kit transmembrane tyrosine kinase that is expressed by spermatogonia in the testis. In the whole organism, the c-kit/SCF system is required for normal hematopoiesis, melanogenesis, and gametogenesis; regarding spermatogenesis, c-kit/SCF signaling has been shown to be essential for proliferation and survival of the only proliferating germ cells, the spermatogonia. Which transduction pathway elicited by c-kit/SCF is able to support these effects? Among the c-kit/SCF-mediated pathways, there is direct activation of phosphatidylinositol (PI) 30 -kinase, which in turn leads to activation of Akt, a serine–treonin kinase with a key role in the growth factor-dependent proliferative and survival responses. Since PI 30 -kinase is active when it binds to a specific tyrosine phosphorylated residue of c-kit (Tyr719 in the mouse),

p0040

148 Blume-Jensen and coworkers generated homozygous mutant mice carrying a Phe719 instead of Tyr719 in c-kit (Y719F/Y719F mutant mice) to search for c-kit/ SCF signaling via PI 30 -kinase. No hematopoietic or pigmentation defects were found in homozygous mutant mice, but males were sterile due to a block in spermatogenesis, with a decrease in proliferation and then extensive apoptosis of spermatogonia. In contrast, female homozygotes were fully fertile. With such an approach, the researchers demonstrated the role of an individual signaling pathway downstream of c-kit/SCF in intact animals. Also, they demonstrated the cell line specificity of cell signaling because melanogenesis and hematopoiesis were not impaired by the inhibition of the transduction pathway acting through PI 30 -kinase.

UBIQUITIN SYSTEM The importance of post-translational protein modification for cell signaling within the cell is well recognized. In addition to protein phosphorylation, post-translational glycosylation also plays an essential role in the control of biological activities as the ligand receptor recognizing crosstalk; however, its dynamism and reversibility of action are not comparable to those of protein phosphorylation/dephosphorylation. Less canonical in comparison to protein phosphorylation and glycosylation is another highly dynamic mechanism, protein ubiquitination, involved in regulation of cell signaling. The ubiquitin system determines the half-life, stabilization, refolding, and translocation of proteins crucial for cell physiology. Ubiquitination is crucial for the downregulation of plasma membrane receptors, steroid hormone receptors, plasma membrane transporters, and ion channels. The ubiquitin machinery is complex; in fact, although the signaling is in the ubiquitin moiety covalently ligated to a well specified protein, other indispensable components are the ubiquitin-activating enzymes, the ubiquitinconjugating enzymes, and the ubiquitin protein ligases that, through the cooperation of members of molecular chaperones, deliver the ubiquitinated protein to the proteasome or endocytotic vacuolar compartment. In addition to the enzymes involved in linking ubiquitin to proteins, there are a large number of deubiquitinating enzymes, which remove ubiquitin from the ubiquitinated proteins, allowing the ubiquitin moiety to recycle. Therefore, as phosphatases switch out the signal elicited by protein kinases, deubiquitinases counterbalance the signaling by ubiquitin ligases. The ubiquitin pathway is undoubtedly important for spermatogenesis. Using mouse transgene and knockout

Germ Cell Differentiation Signaling Events, Male

models, it has been shown that certain components of the ubiquitin system are required for successful spermatogenesis. Different phases of mammalian spermatogenesis require different specialized activities of the ubiquitin machinery.

HR23B Knockout The HR23B gene encodes a mammalian homolog of Saccharomyces cerevisiae RAD23, an ubiquitin-like fusion protein involved in nucleotide excision repair (NER). To study its biological relevance, Ng and coworkers generated HR23B/ mice. Unexpectedly, HR23B deficiency does not result in a NER defect, but HR23B/ mice show impaired embryonic development and a high rate of intrauterine or neonatal death. When they survive, animals display abnormalities such as retarded growth and male sterility. Considering that HR23B is expressed in all mouse tissues and organs, why does male sterility occur? Because disruption of HR23B causes a failure of spermatogenesis, yielding a phenotype like that of the Sertoli cell-only syndrome. This finding indicates that HR23B, and consequently the protein stability via the ubiquitin/proteasome pathway, may be required for the postnatal initiation phase of spermatogenesis.

Siah1a Knockout Experimental evidence indicates a strict link among DNA repair, the ubiquitin system, the heat shock proteins/chaperones machinery, and the control of the meiotic cycle in spermatocytes. Dickins and coworkers used gene targeting to analyze the function of Siah1a, a component of E3 ubiquitin ligase complexes, during mammalian development. Mutant mice have normal weights at term but are postnatally growth retarded, despite normal levels of pituitary growth hormone. Also, serum gonadotropin levels are normal in mutant animals; however, females are subfertile, whereas males are sterile. Male sterility is due to a block in spermatogenesis. Although spermatocytes display normal meiosis I spindle formation, they accumulate at metaphase and then undergo apoptosis. Consequently, Siah1a, as a component of a male germ cell-specific ubiquitin ligase complex, is necessary for normal progression of meiosis.

HR6B Knockout It is during the postmeiotic phase, spermiogenesis, that ubiquitin signaling plays more roles. This cytodifferentiative process implies a massive remodeling

149

Germ Cell Differentiation Signaling Events, Male

of cell structure and loss of most of the cytoplasm and a large fraction of cellular proteins. Some of this reorganization occurs due to condensation of cytoplasm into the residual body, phagocytosed by Sertoli cells while the residue is eliminated as cytoplasmic droplets by epididymal spermatozoa. However, this is not sufficient. In 1996, the first mammalian model of a defect in an enzyme involved in the ubiquitin pathway indicated the essential role of protein ubiquitination during spermatogenesis. To acquire its stream-like structure and simultaneously protect the genetic material during the long trip to the egg, the spermatozoon has evolved a very peculiar chromatin organization. Nucleosomes are disassembled, somatic and testis-specific histones are replaced by transition proteins, which in turn are replaced by protamines, and a highly condensed and inaccessible organization of chromatin is obtained. Roest and coworkers developed HR6B-deficient mice by gene targeting. The mouse HR6B gene is an autosomal homolog of the S. cerevisiae Rad6 gene that codifies an ubiquitin-conjugating enzyme. This gene is highly conserved from yeast to mammals. The phenotype of HR6B knockout mice is remarkable and, at first glance, surprising: Although the HR6B gene is expressed throughout the body, the only pronounced defect of HR6B/ mice is male infertility. This defect is due to impairment of the process of postmeiotic chromatin remodeling, indicating the involvement of the ubiquitin pathway in chromatin dynamism. The mouse HR6B knockout does not cause a complete and uniform block of spermatogenesis at a given point, but spermatozoa of knockout mice are morphologically abnormal and not able to complete fertilization. This finding advanced the study of the ubiquitin system and spermatogenesis, and since 1996 numerous laboratories have been investigating the signaling supported by the ubiquitin system during male germ cell differentiation. Not surprisingly, it has been found that the ubiquitin machinery during spermiogenesis is involved not only in chromatin remodeling but also in the targeted destruction of certain proteins. It appears that whereas the removal of residual cytoplasm as residual body displays the features of a rough phenomenon, the signal-mediated degradation/elimination or rescue of specific proteins by the ubiquitin/proteasome system plays a key role in yielding a fully fertile spermatozoon.

CONCLUSION Cell signaling is a complex network of various interacting objects. Paradoxically, in some cases, data

obtained in the past two decades have complicated our understanding because it has been found that most molecular signal transducers operate in the context of networks in which there is a considerable overlap of functions, some with a synergizing and some with an antagonizing effect. On the other hand, male germ cell signaling is the Cinderella of cell signaling. Historically, these cells have been the model for morphological, caryological, and comparative studies, but because of the lack of a cell culture technology they have attracted the interest of very few researchers engaged in signal transduction. In this article, I have tried to provide a glimpse of the underexplored and fascinating isle of male germ cell signaling. Of course, due to space limitations and my intention to focus onto two fundamental mechanisms of control of cell signaling (i.e., the traditional protein phosphorylation and the emerging protein ubiquitination), I have not provided a comprehensive review of the topic. However, I hope to communicate to the scientific community that male germ cell signaling is a matter worthy of attention and study. A better understanding of the mechanisms governing spermatogenesis is highly desirable.

See Also the Following Articles FSH (Follicle-Stimulating Hormone) . LH (Luteinizing Hormone) . Spermatogenesis, Endocrine Control of

Further Reading Blume-Jensen, P., Jiang, G., Hyman, R., Lee, K.-F., O’Gorman, S., and Hunter, T. (2000). Kit/SCF receptor-induced activation of PI 30 -kinase is essential for male fertility. Nature Genet. 24, 157–162. de Krester, D. M., Loveland, K. L., Meehan, T., O’Bryan, M. K., Phillips, D. J., and Wreford, N. G. (2001). Inhibins, activins and follistatin: Actions on the testis. Mol. Cell. Endocrinol. 180, 87–92. Dickins, R. A., Frew, I., House, C. M., O’Bryan, M. K., Holloway, A. J., Haviv, I., Traficante, N., de Krester, D. M., and Bowtell, D. D. (2002). The ubiquitin ligase component Siah1a is required for completion of meiosis I in male mice. Mol. Cell. Biol. 22, 2294–2303. Hicke, L. (1999). Getting down with ubiquitin: Turning off cellsurface receptors, transporters and channels. Trends Cell. Biol. 9, 107–112. Ng, J. M., et al. (2002). Developmental defects and male sterility in mice lacking the ubiquitin-like DNA repair gene mHR23B. Mol. Cell. Biol. 22, 1233–1245. Roest, H. P., et al. (1996). Inactivation of the HR6B ubiquitinconjugating DNA repair enzyme in mice causes male sterility associated with chromatin modification. Cell 86, 799–810. Sette, C., Dolci, S., Geremia, R., and Rossi, P. (2000). The role of stem cell factor and of alternative c-kit gene products in the establishment, maintenance and function of germ cells. Int. J. Dev. Biol. 44, 599–608.

151

Ghrelin

extent, thereby depriving it of its ability to bind and activate the ghrelin receptor, GHS-R1a. It is curious, however, that a peptide encoded obviously without a fatty acid side chain, in which form it also circulates in large quantities, would have no biological activity, and some findings suggest a possible biological role for unacylated ghrelin, i.e., in cell proliferation processes. Future research will shed light on this apparent paradox.

DETERMINANTS OF GHRELIN ACTION Whereas an enormous amount of data on ghrelin biology and physiology is emerging, the possibility that additional ghrelin receptors (in addition to GHSR1a/b), as well as other endogenous ligands, might exist and could play a relevant role should not be overlooked. The vast majority of data either focus on GHS-R1a or extrapolate from changes in the overall concentration of circulating (bioactive and biologically inactive) ghrelin peptide. Whereas the detection of relative differences in total circulating ghrelin levels (predominantly representing bio-inactive peptide) between disease states or in response to physiological challenges can be regarded as useful, it is important that these data not be overinterpreted. When more sophisticated methods become available for monitoring the plasma concentrations of active ghrelin and the expression and activation levels of specific ghrelin receptor subtypes, substantial parts of the view on ghrelin physiology might have to be readjusted. The extent and magnitude of ghrelin action most likely involve multiple regulatory levels that may sometimes be independent of one another. Relevant mechanisms include the following: the regulation of transcription and translation of the ghrelin gene, the level of enzymatic activity of the putative acyltransferase that is responsible for the posttranslational octanoylation of the ghrelin molecule, secretion rates of the bioactive ghrelin molecule, putative enzymatic processes that deactivate circulating ghrelin, the possible influence of ghrelin-binding proteins on the hormone’s bioactivity (e.g., binding of HDL), the variable accessibility of target tissue (i.e., blood–brain barrier transport), the clearance or degradation of ghrelin by passage through the kidney or liver, the circulating concentration of additional endogenous ligands or other possibly cross-reacting hormones, the level of expression of the ghrelin receptor(s) in target tissues, and the sensitivity of the target tissues at the level of intracellular signaling mechanisms.

GASTRIC AND HYPOTHALAMIC GHRELIN Ghrelin is primarily expressed in the stomach and upper intestinal tract. However, studies using polymerase chain reaction (PCR) amplification techniques have revealed the potential localization of ghrelin mRNA in several other tissues, such as the kidneys, immunocompetent blood cells, placenta, testes, ovaries, pancreas, pituitary, and hypothalamus. With respect to the putative role of ghrelin in the regulation of energy homeostasis, as well as other homeostatic mechanisms, localization in the hypothalamus seems to be particularly interesting (Fig. 1). In a study using antibodies as well as reverse transcription-PCR (RTPCR), a uniquely distributed hypothalamic group of mostly bipolar neurons has been shown to produce small amounts of ghrelin. These neurons are not

Figure 1 Schematic illustration of the relationship between ghrelin and hypothalamic peptidergic circuits. Ghrelin may reach the hypothalamus via the circulation and affect neuronal activity within the arcuate nucleus (ARC), where the NPY/AgRP- and POMCproducing neurons are located. Ghrelin is also produced in a subset of hypothalamic neurons located in the periventricular area. Ghrelin axon terminals innervate arcuate and paraventricular neurons (PVN), including those producing CRH. It is also likely that the central ghrelin system affects the lateral hypothalamic orexin (ORX) and MCH neuronal systems. Note that the hypothalamus operates with many more peptidergic and classical neuromodulators, such as g-aminobutyric acid and glutamate, and this part of the brain is only one component in the central regulation of homeostasis. Therefore, it is reasonable to suggest that ghrelin’s effect in the central nervous system will involve a variety of neurotransmitters and regions not depicted in this figure.

152 colocalized with any known centrally expressed hormone or neuropeptide, but, intriguingly, they do project directly to several previously identified hypothalamic appetite control centers. Ghrelin receptor expression and binding can furthermore be localized in multiple hypothalamic areas that directly neighbor cells containing neuropeptide Y (NPY), agouti-related protein (AgRP), proopiomelanocortin (POMC), orexin, and other neuropeptides and neurotransmitters that are substantially involved in appetite control. These neuroanatomical findings, complemented by electrophysiological studies, provide evidence for the existence of a central circuit that involves ghrelin as a key modulator. The full understanding of the role of hypothalamic ghrelin in endocrine and autonomic regulation remains elusive. The main source of circulating endogenous ghrelin is the stomach. Circulating ghrelin derived from the gastrointestinal tract could be comodulating central networks regulating energy balance together with ghrelin derived from the hypothalamus after crossing the blood–brain barrier or by neural projection from areas that are not protected by the blood–brain barrier (circumventricular organs, e.g., the median eminence). Gastrointestinally derived circulating ghrelin could be responsible for the peripheral effects of ghrelin, including direct effects on the endocrine axes at the pituitary level and cardiovascular, anti-proliferative, or adipocyte-specific effects, whereas centrally derived ghrelin modulates mainly energy balance control circuits. Gastrointestinally derived ghrelin comodulates the central regulation of energy balance at the gastric level via the vagal nervous system and the brainstem. There is no reason that ghrelin could not act concurrently via paracrine mechanisms in the brain, via endocrine mechanisms by way of the circumventricular organs or after crossing the blood–brain barrier, and via the parasympathetic nervous system at the gastrointestinal level to regulate energy balance. Ghrelin administration induces adiposity, raises the respiratory quotient (reflecting reduced fat utilization), and suppresses spontaneous locomotor activity in rodents. Neutralization studies with polyclonal ghrelin antibodies yielded encouraging results, showing decreased food intake in rodents after intracerebroventricular injection. These data were confirmed by a transgenic rat model that overexpresses antisense oligonucleotides against the ghrelin receptor, GHSR1a, and decreased food intake and lower body fat were observed as a consequence. There is clear evidence that, despite its relatively short half-life, administration of ghrelin in physiologically relevant doses

Ghrelin

induces a positive energy balance. Ghrelin affects body weight and food intake more than 1000-fold more potently following central administration, strongly supporting the hypothesis that ghrelin influences energy homeostasis predominantly by the modulation of central mechanisms. Therefore, whatever the derivation of the decisive endogenous amount of ghrelin that regulates energy balance, neutralizing or blocking its endogenous actions acutely as well as chronically will, at the very least, provide a valuable physiology lesson, yet has the potential to pave the way for the development of a new drug.

REGULATION OF GASTRIC GHRELIN EXPRESSION AND SECRETION Ghrelin is secreted mainly by gastric A/X-like cells within the oxcyntic glands; the half life of ghrelin is relatively short (5–15 min) and less than 20% of the circulating immunoreactive ghrelin appears to be octanoylated and therefore bioactive. Gastrointestinal X/A-like cells represent approximately one-quarter of all endocrine cells in the oxyntic mucosa, and other cells within these glands, such as histamine-rich enterochromaffin-like cells (ca. 70%) and d-(somatostatin) cells (10%), are ghrelin-negative. From the stomach to the colon, ghrelin is found with caudally decreasing expression. Most ghrelin-containing enteroendocrine cells have no continuity with the lumen, probably respond to physical and/or chemical stimuli from the basolateral side, and are closely associated with the capillary network running through the lamina propria. Ghrelin-secreting cells occur as open- and closed-type cells (open or closed toward the stomach lumen), with the number of open-type cells gradually increasing in the direction from the stomach to the lower gastrointestinal tract. A closer look at the structural and functional relationship between ghrelin and its receptor, as well as the structure of motilin and its receptor, suggests that a larger family of peptide hormones is comodulating gastrointestinal motility, appetite, secretion of pituitary hormones, and other physiological processes. This peripheral endocrine network most likely also includes gastrointestinal hormones such as cholecystokinin, peptide YY (PYY1-36, PYY3-36), glucagon-like peptide 1, and gastric inhibitory peptide. Both ghrelin peptide secretion and ghrelin mRNA expression are regulated according to metabolic challenges. Acute and chronic periods of food deprivation increasing from mild to severe (e.g., fasting) raise ghrelin peptide levels and ghrelin mRNA concentrations, whereas refeeding reduces both.

153

Ghrelin

OTHER SOURCES OF GHRELIN It is evident that ghrelin plays a classical endocrine role as a peptide hormone that is secreted into a capillary network, but local, paracrine activities of ghrelin might play an additional role. Removal of the stomach or of the acid-producing part of the stomach in rodents reduces serum ghrelin concentration by approximately 80%, further supporting the notion that the stomach is the main source of this endogenous GHS receptor ligand. However, in another study, plasma levels of ghrelin after total gastrectomy gradually increased again, suggesting that other tissues can compensate for the loss of ghrelin production after gastrectomy. The fact that total plasma ghrelin is barely detectable following gastric bypass surgery was interpreted by Cummings and co-workers as a ‘‘shutdown’’ of gastric ghrelin secretion due to a complete lack of contact with ingested nutrients. Ghrelin mRNA and ghrelin peptide have also been detected in rat and human placenta. Here, ghrelin is expressed predominantly in cytotrophoblast cells and very sporadically in syncytiotrophoblast cells. A pregnancy related time-course, represented by an early increase in ghrelin expression in the third week, decreasing levels in the latest stages of gestation, and still detectable amounts of ghrelin at term, was found in rats. In human placenta, ghrelin is expressed mainly in the first half of pregnancy and is not detectable at term, whereas a putative involvement of ghrelin in fetal–maternal interactions via autocrine, paracrine, or endocrine mechanisms still remains to be shown. Small concentrations of ghrelin are found in the pancreas, where ghrelin immunoreactivity has been localized in a subgroup of endocrine cells that are also immunopositive for pancreostatin. Ongoing, partially contradictory studies suggest that ghrelinpositive cells must be either the pancreatic alpha or beta cells. In normal pituitary cells as well as in pituitary tumors, ghrelin mRNA expression and ghrelin-immunopositive cells were detected. This is in addition to the known presence of GHS receptors in pituitary cells. This suggests a possible autocrine or paracrine role for hypophyseal ghrelin, although only approximately 5% of the detected ghrelin peptide derived from the pituitary has been found to be octanoylated. Using real-time PCR methodology, small amounts of ghrelin were detected in the adrenal glands, esophagus, adipocytes, gallbladder, muscle, myocardium, ovary, prostate, skin, spleen, thyroid, blood vessels, and liver. Preproghrelin production was shown in rat mesangial cells and mouse podocytes,

indicating the production of ghrelin in kidney, glomerulus, and renal cells and suggesting possible paracrine roles for ghrelin in the kidney. Human ghrelin expression and GHS receptor mRNA expression were shown by real-time PCR and confirmed by DNA sequencing in human T lymphocytes, B lymphocytes, and neutrophils from venous blood of healthy volunteers. Cell type and the maturity of the cells did not seem to have an influence on ghrelin production in immune cells. Interestingly, it has been shown that the small molecule GHS has a considerable immuneenhancing effect. In summary, ghrelin is predominantly expressed by the stomach and is expressed at decreasing levels as one moves caudally through the gastrointestinal tract. Although its physiological significance as a paracrine factor in extra-gastrointestinal tissue is the subject of ongoing studies, a classical endocrine role for extra-gastrointestinal ghrelin appears to be unlikely since ghrelin expression levels in other organs are relatively low in comparison. Thus, published studies on the regulation of ghrelin expression primarily focused on gastric ghrelin. However, studies on ghrelin expression or secretion in rodents are not necessarily relevant to the physiological regulation of ghrelin in humans.

REGULATION OF CIRCULATING GHRELIN LEVELS A very intriguing series of clinical studies by Cummings et al. indicates that each daily meal is followed by decreases in circulating ghrelin levels, most likely reflecting acutely reduced ghrelin secretion from the gastrointestinal tract. The authors speculate further that an observed premeal rise in circulating human ghrelin levels might reveal a role for ghrelin in meal initiation, which fits well with the observation that ghrelin administration in healthy volunteers causes hunger sensations. Ghrelin levels might also reflect the acute state of energy balance, signaling to the central nervous system in times of food deprivation that increased energy intake and an energy-preserving metabolic state are desirable. Only a few determinants of circulating ghrelin concentration have been identified; these include insulin, glucose, somatostatin, and possibly growth hormone, leptin, melatonin, and the parasympathetic nervous system tone. In several species (e.g., mice, rats, cows, and humans), ghrelin mRNA expression levels or circulating ghrelin levels have been shown to be increased by food deprivation and to be decreased postprandially. This phenomenon further supports

154 the concept of ghrelin as an endogenous regulator of energy homeostasis that has apparently been preserved throughout evolution in all species. Rat ghrelin expression can also be stimulated by insulin-induced hypoglycemia, leptin administration, and central leptin gene therapy. Ingestion of sugar suppresses ghrelin secretion in rats in vivo, indicating a direct inhibitory effect of glucose/caloric intake on ghrelin-containing X/A-like cells in the oxyntic mucosa of the rat stomach rather than an exclusively insulin-mediated effect. The fact that insulin is an independent determinant of the circulating ghrelin concentration has been shown by several research groups using hyperinsulinemic– euglycemic clamp studies in humans. These findings add further evidence that ghrelin provides the link between mechanisms governing energy balance and the regulation of glucose homeostasis. However, it remains to be shown whether postprandially occurring insulin peaks are sufficient to decrease circulating ghrelin levels, since hyperinsulinemic–euglycemic clamp studies causing decreased ghrelin secretion involve either supraphysiological or markedly prolonged (e.g., > 120 min) periods of hyperinsulinemia. Further insight into the complex mechanisms regulating ghrelin secretion is based on studies showing an increase in circulating ghrelin levels in rats following surgical interventions such as vagotomy and hypophysectomy. Human growth hormone (GH) deficiency, however, does not seem to lead to increased plasma ghrelin levels. On the other hand, administration of synthetic GH in rats decreases circulating ghrelin levels and therapeutic intervention causing normalization of GH levels in patients with acromegaly increases endogenous ghrelin levels. Somewhat contradictory observations could be caused by species-specific differences between rodents and humans or could indicate that an acute, but not a chronic, change in GH levels modulates ghrelin concentrations. A previously neglected pathophysiological factor that might increase circulating ghrelin levels is the production of ghrelin by tumors of the stomach and the intestine, such as carcinoids. In summary, ghrelin expression and ghrelin secretion are predominantly influenced by changes in energy balance and glucose homeostasis and influenced to a somewhat lesser degree by alterations in the endocrine axes (e.g., increasing GH concentrations). Based on the available data, ghrelin seems to represent a molecular regulatory interface between energy homeostasis, glucose metabolism, and physiological processes regulated by the classical endocrine axes, such as growth and reproduction. One particular biological purpose of these multiple roles of ghrelin

Ghrelin

might be to ensure the provision of calories that GH requires for growth and repair.

GHRELIN AND OBESITY In contrast to earlier models that expected the endogenous ligand of the growth hormone secretagogue receptor to exclusively govern growth hormone secretion, ghrelin is believed to play its main physiological role in the regulation of energy balance. As the only peripherally circulating orexigenic agent known, ghrelin triggers appetite and nutrient intake. Ghrelin might even represent the first known ‘‘meal initiation factor.’’ However, conclusive evidence that mealrelated circadian changes in plasma ghrelin concentrations are responsible for the initiation of nutrient intake rather than representing an epiphenomenon of trained meal patterns is still missing. Based on clinical investigations of meal-related changes in plasma ghrelin levels and data generated by insulin- and glucoseclamp studies, plasma insulin and blood glucose levels are very likely to be involved in the general regulation of ghrelin secretion. Although hyperinsulinemic– euglycemic clamps have been repeatedly shown to decrease circulating ghrelin levels, it remains unclear whether experimental conditions during clamp studies are comparable with the lower maximum peaks and the shorter duration of postprandial insulin levels. The possibility cannot be excluded that additional blood-derived factors may be responsible for mealrelated changes in ghrelin concentrations or that gastrointestinal nutrient sensors may modulate ghrelin expression and secretion rates. Although counterintuitive, the finding that circulating ghrelin concentrations are low in obesity not only mirrors earlier observations of hyperleptinemia in obesity, but may be explained by compensatory mechanisms aiming to communicate to the central regulatory centers that energy stores are full. Though it is unclear which signal communicates increased adipocyte size to ghrelin-secreting cells (leptin, interleukin-6, and adiponectin would be candidates), other phenomena and symptoms that commonly occur during obesity (such as a frequently filled stomach or insulin resistance) should be carefully investigated to determine their contribution to hypoghrelinemia. Ghrelin gene polymorphisms have been described by several groups; linkage analysis studies, however, failed to prove a solid association between ghrelin and obesity. Although diet-induced human obesity and polygenic (e.g., Pima Indians) or monogenic (e.g., MC4-R defect) causes of human obesity all present with low plasma ghrelin levels, there is one group

155

Ghrelin

of severely obese patients in whom markedly increased plasma ghrelin concentrations have been observed. Prader-Willi syndrome, an impressive hunger syndrome in which patients exhibit morbid obesity and numerous other symptoms, is caused by a defect in the short arm of chromosome 15 and is accompanied by circulating ghrelin levels that are three- to fivefold higher than in healthy controls. Although the overlap between symptoms of Prader-Willi syndrome and the effects of ghrelin administration is impressive, only treatment with a potent, but safe ghrelin antagonist compound will reveal whether ghrelin is part of the pathogenesis in Prader-Willi syndrome. Discussion is ongoing as to whether increased plasma ghrelin is only a consequence of the severe caloric restrictions that are a central part of the treatment strategies for patients with Prader-Willi syndrome in an attempt to control their energy balance. The only other population in which comparably high ghrelin levels have been reported are patients with cachexia or anorexia nervosa, in whom high ghrelin levels are believed to reflect a physiological compensation effort in response to either a chronically empty stomach or a markedly decreased fat mass. Though circulating ghrelin levels are significantly lower in obese individuals, these levels are still very substantial when compared to nearly undetectable ghrelin concentrations in patients after gastric bypass surgery. The superior effectiveness of this bariatric procedure is considered to be partially due to a ‘‘knock-down’’ of endogenous ghrelin secretion caused by the lack of stimulation of gastric cells by incoming nutrients. On the other hand, an increase in endogenous ghrelin in response to diet-induced weight loss could contribute to the very high likelihood of recurrence of obesity. Carefully conducted clinical studies are imperative to discover the answer to this important question.

SUMMARY In summary, ghrelin represents a gastric hormone that induces hunger and increases fat deposition via central and possibly peripheral mechanisms in response to a negative energy balance. Ghrelin is one of the most potent orexigenic agents and the only known peripherally circulating orexigenic agent; it may also represent the first meal initiation factor. Although, normally, plasma ghrelin concentrations are negatively correlated with fat mass, substantial levels are still secreted in the vast majority of obese individuals, whereas weight loss occurs along with the loss of circulating ghrelin in patients with a gastric bypass.

Apart from the possible effectiveness of a ghrelin antagonist for the general prophylaxis and treatment of adiposity, the blockade of ghrelin could be the first specific pharmacotherapeutic approach to successfully treat patients with Prader-Willi syndrome. However, although the various additional effects of ghrelin on physiological processes and organ systems suggest other possible therapeutic uses, they also make unwanted cardiovascular, gastrointestinal, or proliferative effects caused by the blockade of ghrelin action a likely occurrence.

See Also the Following Articles CCK (Cholecystokinin) . GI Hormone Development (Families and Phylogeny) . GI Hormones Outside the Gut: Central and Peripheral Nervous System . GIP (Gastric Inhibitory Polypeptide) . Hunger and Satiation . Leptin . Motilin . Natural and Synthetic Growth Hormone Secretagogues . Neuropeptide Y . Obesity and Diabetes, Regulation of Food Intake . Peptide YY (PYY)

Further Reading Ariyasu, H., Takaya, K., Tagami, T., Ogawa, Y., Hosoda, K., Akamizu, T., Suda, M., Koh, T., Natsui, K., Toyooka, S., Shirakami, G., Usui, T., Shimatsu, A., Doi, K., Hosoda, H., Kojima, M., Kangawa, K., and Nakao, K. (2001). Stomach is a major source of circulating ghrelin, and feeding state determines plasma ghrelin-like immunoreactivity levels in humans. J. Clin. Endocrinol. Metab. 86, 4753–4758. Baldanzi, G., Filigheddu, N., Cutrupi, S., Catapano, F., Bonissoni, S., Fubini, A., Malan, D., Baj, G., Granata, R., Broglio, F., Papotti, M., Surico, N., Bussolino, F., Isgaard, J., Deghenghi, R., Sinigaglia, F., Prat, M., Muccioli, G., Ghigo, E., and Graziani, A. (2002). Ghrelin and des-acyl ghrelin inhibit cell death in cardiomyocytes and endothelial cells through ERK1/2 and PI 3-kinase/AKT. J. Cell Biol. 159, 1029–1037. Beaumont, N. J., Skinner, V. O., Tan, T. M., Ramesh, B. S., Byrne, D. J., MacColl, G. S., Keen, J. N., Bouloux, P. M., Mikhailidis, D. P., Bruckdorfer, K. R., Vanderpump, M. P., and Srai, K. S. (2003). Ghrelin can bind to a species of high density lipoprotein associated with paraoxonase. J. Biol. Chem. 278, 8877–8880. Bednarek, M. A., Feighner, S. D., Pong, S. S., McKee, K. K., Hreniuk, D. L., Silva, M. V., Warren, V. A., Howard, A. D., Van der Ploeg, L. H., and Heck, J. V. (2000). Structure–function studies on the new growth hormone-releasing peptide, ghrelin: Minimal sequence of ghrelin necessary for activation of growth hormone secretagogue receptor 1a. J. Med. Chem. 43, 4370–4376. Chanoine, J. P., Yeung, L. P., Wong, A. C., and Birmingham, C. L. (2002). Immunoreactive ghrelin in human cord blood: Relation to anthropometry, leptin, and growth hormone. J. Pediatr. Gastroenterol. Nutr. 35, 282–286. Date, Y., Murakami, N., Toshinai, K., Matsukura, S., Niijima, A., Matsuo, H., Kangawa, K., and Nakazato, M. (2002). The role of the gastric afferent vagal nerve in ghrelin-induced feeding and growth hormone secretion in rats. Gastroenterology 123, 1120–1128.

156 Havel, P. J. (2001). Peripheral signals conveying metabolic information to the brain: Short-term and long-term regulation of food intake and energy homeostasis. Exp. Biol. Med. 226, 963–977. Horvath, T. L., Castaneda, T., Tang-Christensen, M., Pagotto, U., and Tschop, M. H. (2003). Ghrelin as a potential anti-obesity target. Curr. Pharm. Des. 9, 1383–1395. Horvath, T. L., Diano, S., Sotonyi, P., Heiman, M., and Tschop, M. (2001). Minireview: Ghrelin and the regulation of energy balance—A hypothalamic perspective. Endocrinology 142, 4163–4169. Hosoda, H., Kojima, M., Mizushima, T., Shimizu, S., and Kangawa, K. (2003). Structural divergence of human ghrelin: Identification of multiple ghrelin-derived molecules produced by post-translational processing. J. Biol. Chem. 278, 64–70.

Ghrelin

Kojima, M., Hosoda, H., Date, Y., Nakazato, M., Matsuo, H., and Kangawa, K. (1999). Ghrelin is a growth-hormone-releasing acylated peptide from stomach. Nature 402, 656–660. Muccioli, G., Tschop, M., Papotti, M., Deghenghi, R., Heiman, M., and Ghigo, E. (2002). Neuroendocrine and peripheral activities of ghrelin: Implications in metabolism and obesity. Eur. J. Pharmacol. 440, 235–254. Tschop, M., Smiley, D. L., and Heiman, M. L. (2000). Ghrelin induces adiposity in rodents. Nature 407, 908–913. Van der Lely, A. J., Tscho¨p, M., Heiman, M. L., and Ghigo, E. Biological physiological pathophysiological and pharmacological aspects of ghrelin. Endocr. Rev., in press. Zhang, Y., Proenca, R., Maffei, M., Barone, M., Leopold, L., and Friedman, J. M. (1994). Positional cloning of the mouse obese gene and its human homologue. Nature 372, 425–432.

169

GI Hormone Development (Families and Phylogeny)

Figure 1 Evolutionary relationship of the chordate classes. Apart from the tunicates (or urochordates), the remaining classes are vertebrate animals. The lengths of the horizontal lines indicate the period when the respective classes are assumed to have segregated.

Thus, observations in one species based on antibodies raised against peptides from another species must be evaluated with caution, since both false-positive and false-negative results can occur. During evolution, duplication of genes and exons followed by substitutions, insertions, and deletions has been an important mechanism for the creation of new peptides with new functions. Two major rounds of duplications occurred during the evolution of vertebrates, one at the beginning of the ‘‘Cambrian explosion’’ (approximately 500 million years ago) and one in the early Devonian (approximately 400 million years ago), to produce up to four copies of the original genome. In fish, a third duplication appears to have occurred in the late Devonian. Many of the duplicated gene products developed into new proteins with new functions, thus giving rise to distinct families of proteins (e.g., regulatory peptides), with family members having related or even very distant functions. Many of the gastrointestinal hormones can clearly be classified as belonging to families originating from a common ancestor. Others appear as singular peptides although they can be traced to most vertebrate classes. The following discussion of the relationships between specific peptides and peptide families considers mainly solid structural information and to some extent function and anatomical origin. Also, only hormones from the Chordata (including tunicates) are considered.

HORMONE FAMILIES The PACAP/Glucagon Superfamily The PACAP/glucagon superfamily includes nine structurally related hormones in human, namely, PACAP, GHRH (also known as growth hormonereleasing factor), glucose-dependent insulinotropic polypeptide (GIP), peptide histidine–methionine (PHM) or peptide histidine–isoleucine (PHI), secretin, vasoactive intestinal peptide (VIP), glucagon, and glucagon-like peptide-1 and-2 (GLP-1 and-2). In mammals, these peptides are encoded by six genes: VIP and PHM are encoded by the same gene, glucagon, GLP-1, and GLP-2 are encoded by a single gene, and the remaining peptides come from separate genes. PACAP/GHRH The PACAP/GHRH superfamily can be traced back to tunicates, in which PACAP and GHRH have been identified; actually, there are two different precursors in tunicates, each coding for both peptides. Throughout evolution, PACAP and GHRH have remained together on the same precursor until the emergence of mammals, in which they appear on separate genes. This could be due to a late gene duplication in a common ancestor of mammals. Alternatively, it has been suggested that a second

170 gene remains undetected in nonmammalian vertebrates and that one has evolved into the mammalian PACAP preceded by a related flanking peptide and the other has evolved into the mammalian GHRH followed by another related flanking peptide. Interestingly, PACAP has been extremely well conserved from tunicate to human, whereas the structure of GHRH is highly variable and even between different mammals there is a low degree of similarity. The Glucagon Family Glucagon, GLP-1, and GLP-2 have been detected in all vertebrate classes, but not (yet) in tunicates. The three peptides are derived from a common precursor and this organization has been conserved from jawless fish to mammals. Glucagon is highly conserved in all vertebrates but both GLP-1 and GLP-2 appear to be considerably more variable. GIP, PHM/PHI, Secretin, and VIP GIP has been identified only in mammals and secretin and PHM/PHI have been identified only in mammals and birds. Whether they are represented in the other vertebrate classes remains an open question. In contrast, VIP has been detected in all vertebrates except jawless fish. Furthermore, the structure of VIP is highly conserved with only minor variations allowed at nine positions of the 28-residue peptide.

The PP-fold Family The PP-fold family of peptides derives its name from a structural characteristic of all the members: a hairpin-like fold with an extended proline helix containing three proline residues, a turn, and an a-helix with two tyrosine residues that fit in between the three proline residues. The family consists of neuropeptide Y (NPY) and peptide YY (PYY), present in all vertebrate classes, pancreatic polypeptide (PP), found only in tetrapods, and a pancreatic peptide Y (PY), found in some fish. NPY and PYY most likely originate from duplication of a common ancestor before the development of the jawless fish. NPY displays a high degree of sequence conservation, whereas PYY is more varied. The occurrence of PP only in tetrapods suggests that it is the result of yet another gene duplication that occurred before the emergence of the amphibians. The sequence variability of PP is quite high, with only 50% identity between mammals, birds, and amphibians. PY constitutes a special case and probably arose from a duplication of the PYY gene unrelated to the duplication event that generated PP.

GI Hormone Development (Families and Phylogeny)

The Cholecystokinin/Gastrin Family Cholecystokinin (CCK) and gastrin constitute a small family comprising only these two members. They are characterized by a common amidated C-terminal tetrapeptide sequence, which also constitutes the minimal structure necessary for biological activity of both peptides. Hence, it appears most likely that CCK and gastrin have evolved from a common ancestor. Cionin, isolated from Ciona intestinalis, a representative of the tunicates, which occupy a key position at the transition to vertebrates, also contains the characteristic tetrapeptide sequence and thus represents the oldest genuine member of the CCK/gastrin family thus far known, dating the emergence of these peptides back to at least 500 million years ago. The CCK/ gastrin family is represented throughout the entire chordate phylum, including cartilaginous and bony fish, amphibians, reptiles, birds, and mammals. A duplication of the ancestral gene appears to have occurred at the level of cartilaginous fish, giving rise to two peptides most likely homologous to CCK and gastrin. At the amphibian level, the two separate peptide systems have been shown to exert distinct physiological gastrin and CCK actions. Interestingly, though CCK is well conserved in all vertebrate species, the gastrins vary more and the mammalian gastrins at first sight appear as a distinct group with little similarity to the nonmammalian gastrins outside the invariant C-terminal tetrapeptide. However, closer examination reveals that even if a major structural change was introduced at the transition to mammals, there exists a clear evolutionary relationship between mammalian and nonmammalian gastrins.

The Tachykinin Family The tachykinins share a C-terminally amidated pentapeptide, -Phe-Xaa-Gly-Leu-Met-amide, where Xaa can vary among a few possible amino acids. In mammals, the family comprises four peptides encoded by two genes: preprotachykinin A, encoding substance P (SP), neurokinin A (NKA), and neuropeptide g (NPg, which is an N-terminally extended form of neurokinin A); and preprotachykinin B, encoding neurokinin B. SP, NKA, and NPg have been identified in most vertebrate classes, whereas NKB has been identified only in mammals and a frog (Amphibia). NKA (a decapeptide) has been highly conserved and SP (an undecapeptide) somewhat less so. Interestingly, SP from amphibians shows less identity to the mammalian form than SP from the phylogenetically more distant fish. However, the peptides are rather short

171

GI Hormone Development (Families and Phylogeny)

and even if many tachykinins from nonmammalian vertebrates have been identified, almost no genes for these tachykinins have been discovered. Thus, solid conclusions regarding the evolution of the tachykinins are difficult to make, but there is no doubt that the tachykinins constitute a family, that they must have arisen from a common ancestor, and that their diversity must have been brought about by exon duplication, gene duplication, and point mutations.

The Somatostatin Family In mammals, the somatostatin family contains two members, somatostatin (known since 1973) and cortistatin, which shows a high degree of similarity to somatostatin. The bioactive forms of somatostatin are somatostatin-14 and the N-terminally extended somatostatin-28. Somatostatin-14 is highly conserved, being identical in all vertebrates from jawless fish to mammals. In addition to genuine somatostatin, many fish and amphibian species express a second gene very similar to preprosomatostatin-I. It has been suggested that the duplication of the original somatostatin gene occurred early in evolution, predating or concomitant with the development of the chordates. Then, a second duplication may have occurred in the bony fish after divergence from the line leading to tetrapods.

The Gastrin-Releasing-Peptide/Bombesin/ Neuromedin B Family Gastrin-releasing peptide (GRP) has been identified in all major vertebrate classes except jawless fish. It consists of 22 (goldfish) to 29 (rat) amino acid residues and has an invariant C-terminal octapeptide sequence in all species. Neuromedin B, which shows similarity to GRP, has been identified only in mammals and a toad (Amphibia) and is encoded by a separate gene. The bombesins constitute an interesting subgroup, all being isolated from the skin of amphibians, where they may serve protective purposes, and also arising from independent genes. Since only GRP is known to exist in many vertebrate classes, it is not possible to further elaborate on the phylogeny of these peptides.

HORMONES WITH NO OBVIOUS FAMILY RELATIONSHIPS Galanin Galanin is a widespread neuropeptide and is also common in the gut. Since its discovery in 1983,

galanin has been identified in a number of species representing all major vertebrate classes. It comprises 29 (or 30, in human) amino acid residues, of which the 14 N-terminal residues are identical in all species investigated. Despite repeated efforts to discover related peptides, none were found until 1999, when a new peptide, galanin-like peptide (GALP), was isolated from hypothalamus based on its stimulation of galanin type 2 receptors. Thirteen residues (positions 9–21) of GALP are identical to the N terminus of galanin. GALP has been identified only in mammals (human, pig, and rat). Thus, it still remains to be seen whether a galanin family can be identified.

Motilin and Ghrelin Motilin is produced in endocrine cells of the upper intestine. Since the identification of porcine motilin in 1972, motilin has been identified in a number of mammals but in only one nonmammalian species, the chicken. The mature form of motilin consists of 22 amino acid residues and between human and chicken there are seven substitutions, although most are conservative. Thus, little can be said about the evolution of motilin. In 1999–2000, a new GHRH was identified independently by two groups and was designated ghrelin and prepromotilin-related peptide, respectively. The latter name refers to a limited structural similarity with motilin. Identification of the two hormones from more vertebrate classes will be needed before possible relationships can be evaluated.

See Also the Following Articles ACTH, a-MSH, and POMC, Evolution of . Angiotensin, Evolution of . Insulin and Insulin-like Growth Factors, Evolution of . Natriuretic Peptide System, Evolution of . Neuropeptide Y, Evolution of . Pituitary Adenylate Cyclase-Activating Polypeptide (PACAP)/Glucagon Superfamily . Prolactin, Evolution of . Somatostatin, Evolution of . Steroid Receptors, Evolution of

Further Reading Cerda-Reverter, J. M., and Larhammar, D. (2000). Neuropeptide Y family of peptides: Structure, anatomical expression, function, and molecular evolution. Biochem. Cell. Biol. 78, 371–392. Feng, D.-F., Cho, G., and Doolittle, R. F. (1997). Determining divergence times with a protein clock: Update and reevaluation. Proc. Natl. Acad. Sci. USA 94, 13028–13033. Holmgren, S., and Jensen, J. (2001). Evolution of vertebrate neuropeptides. Brain Res. Bull. 55, 723–735. Hoyle, C. H. (1998). Neuropeptide families: Evolutionary perspectives. Regul. Pept. 73, 1–33. Irwin, D. M. (2001). Molecular evolution of proglucagon. Regul. Pept. 98, 1–12.

172 Johnsen, A. H. (1998). Phylogeny of the cholecystokinin/gastrin family. Front. Neuroendocrinol. 19, 73–99. Lin, X., Otto, C. J., Cardenas, R., and Peter, R. E. (2000). Somatostatin family of peptides and its receptors in fish. Can J. Physiol. Pharmacol. 78, 1053–1066. Montero, M., Yon, L., Kikuyama, S., Dufour, S., and Vaudry, H. (2000). Molecular evolution of the growth hormone-releasing

GI Hormone Development (Families and Phylogeny)

hormone/pituitary adenylate cyclase-activating polypeptide gene family: Functional implication in the regulation of growth hormone secretion. J. Mol. Endocrinol. 25, 157–168. Sherwood, N. M., Krueckl, S. L., and McRory, J. E. (2000). The origin and function of the pituitary adenylate cyclase-activating polypeptide (PACAP)/glucagon superfamily. Endocr. Rev. 21, 619–670.

174 GI hormone cholecystokinin (CCK) is expressed in the postnatal rat pancreas. Transgenic overexpression of gastrin and transforming growth factor-a has been found to result in islet hyperplasia, perhaps reflecting a trophic role of gastrin during early islet development.

GI Hormones and Endocrine Pancreas: Expressional Regulation

Glucagon The glucagon gene is expressed by both pancreatic A cells and intestinal L cells. Differences in proteolytic processing account for the production of classical glucagon in A cells and of glucagon-like peptide-1 and -2 and other fragments in L cells.

Secretin Secretin is the main product of endocrine S cells of the duodenum and jejunum. S cells occur on villi, but never in crypts, documenting their identity with postmitotic cells. Secretin is transiently expressed in insulin cells of the developing pancreatic islets.

Preprotachykinin A Expression of substance P was first noted in endocrine, enterochromaffin (EC) cells of the intestinal mucosa. Subsequently, preprotachykinin A (PPTA) was found to be transiently expressed in both insulinand non-insulin-producing endocrine cells of the fetal and neonatal pancreas. Postnatally, the number of PPTA-positive cells declines and no such cells can be detected in adults. It has been suggested that substance P and neurokinin A, which both are encoded by the PPTA gene, could serve as growth factors during islet development.

Vasoactive Intestinal Polypeptide VIP is normally expressed by nerve cells of the GI tract and pancreas but is also expressed by pancreatic tumors causing the WDHA (watery diarrhea, hypokalemia, achlorhydria) syndrome, showing that neoplastic pancreatic endocrine cells have the capacity for VIP production.

Serotonin Serotonin (5-hydroxytryptamine) is produced by endocrine EC cells of the GI tract and pancreas and is also detected in pancreatic insulin cells of certain species, such as the guinea pig. Early maturing GI endocrine cells coexpress serotonin with other hormones.

Insulin Transient insulin expression occurs in a few neuronlike cells present at the juxtaduodenal myenteric plexus at embryonic days 18–19 in the rat.

Somatostatin Somatostatin is produced by D cells present in the GI tract and pancreas and is expressed by both the adult and the developing gut and pancreas. It is believed to act primarily as a local paracrine regulator of hormone release. In some species, both pancreatic and GI somatostatin cells emit long neuron-like processes with which they contact their target cells.

PP Family The PP family includes three structurally related peptides: pancreatic polypeptide (PP), neuropeptide Y (NPY), and peptide tyrosine tyrosine (PYY). PP was originally localized to the fourth (PP) cell type of the pancreas. Cat and dog PP (F) cells are characterized by large granules. In other species, such as human and rat, PP cells are small granulated cells, distinct from the fifth (D1) islet cell type, believed to produce an as yet unidentified hormone. Reverse transcription-polymerase chain reaction data show that pancreatic PP expression occurs before PP cells can be immunocytochemically identified and transgenic data indicate that insulin cell progenitors express PP. Only limited GI expression of PP occurs. PYY is expressed by both gut and pancreatic endocrine cells (including antropyloric gastrin cells) and may be a marker for the earliest appearing endocrine pancreatic cells. NPY is expressed in a subpopulation of insulin cells in rat pancreas and by nerves in the GI tract.

Islet Amyloid Polypeptide (Amylin) Amylin was originally discovered as the chief constituent of amyloid present in insulinoma tissue and in islets of patients with non-insulin-dependent diabetes. Amylin occurs in islet insulin and somatostatin cells as well as in a majority population of somatostatin cells and a minority population of gastrin cells of the antropyloric mucosa of the stomach.

p0060

175

GI Hormones and Endocrine Pancreas: Expressional Regulation

HORMONE COEXPRESSION IN THE GASTRODUODENOPANCREATIC REGION Many studies have shown that mature endocrine cell types frequently coexpress multiple hormones. Although these observations challenged well-guarded concepts, such as ‘‘the one cell–one hormone principle,’’ they have stood the test of time better than the original concepts. It is widely accepted that many endocrine cells secrete a host of molecules that act in concert to elicit coordinated biological responses. Also, developing or maturing endocrine cells frequently express multiple hormones. Observations of such coexpression led to the finding that the three main antropyloric endocrine cell types (gastrin, serotonin, and somatostatin cells) develop from common multihormone-expressing precursors in the isthmus of gastric glands. In addition, duodenal endocrine cells develop from multihormonal precursors present in the crypts of Lieberku¨hns. Cells coexpressing insulin and glucagon have been detected in the developing pancreas and it has been suggested that they represent islet cell precursors. However, transgenic studies have documented that insulin and glucagon cells are derived from cell lineages that do not coexpress these two hormones. Instead, mature insulin cells may develop from cells transiently expressing PP. Moreover, the large increases in insulin cell numbers that occur during rat pancreatic development suggest that most of these cells are derived from insulin-negative precursors.

TRANSCRIPTION FACTORS AND ENDOCRINE DIFFERENTIATION Gene knockout studies have demonstrated that a number of transactivating factors are important for the differentiation of both pancreatic and gastroduodenal endocrine cells. This is not surprising since the distal stomach, duodenum, and pancreas develop from closely adjacent parts of the primitive foregut. Furthermore, overlaps in the distribution of these factors may explain the transient and permanent overlaps in the expression of gut hormones in the pancreas and of pancreatic hormones in the gut. Such factors include pancreatic(-antropyloric-) duodenal homeobox-1 (Pdx-1), b2/NeuroD, Pax4, and Pax6. Additionally, indirect data suggest that factors such as islet-1 (Isl-1) and Nkx6.1 may share this dual importance.

Table I Effects of Gene Knockouts of Pdx-1, Pax4, Pax6, and b2/NeuroD on Hormone Expression in the Gastroduodenopancreatic Region Deficiency Pdx-1

Pancreas No pancreas

Duodenum

Stomach

Neurotensin #

Gastrin #

GIP #

Serotonin "

Secretin # Pax4

Insulin #

Serotonin #

Serotonin #

Somatostatin #

CCK #

Somatostatin #

Glucagon "

GIP # PYY # Secretin #

Pax6

Glucagon #

GIP #

Gastrin # Somatostatin #

b2/NeuroD

Insulin # Glucagon #

CCK # Secretin #

Somatostatin #

Pdx-1 was identified as a factor that interacts with the insulin and somatostatin promoters. In Pdx-1-deficient mice, the pancreas fails to develop, the rostral duodenum shows anomalies, and in the antropyloric mucosa, gastrin cells fail to develop, whereas somatostatin cells occur in normal numbers and serotonin cells occur in increased numbers (Table I). These data show that Pdx-1 is not required for gastric somatostatin expression, but is needed for the differentiation of gastrin cells, probably from a common gastrin–somatostatin precursor cell. Additionally, Pdx-1
/
mice show normal gastric expression of amylin, probably reflecting normal expression in somatostatin cells and reduced expression of PYY, reflecting the decrease in gastrin cells. The pancreas of Pax4
/
mice is deficient in insulin and somatostatin cells, but contains increased numbers of glucagon cells. In the antropyloric mucosa, Pax4 deficiency eliminates somatostatin cells and severely reduces the number of serotonin cells, while leaving gastrin cells unaffected (Table I). The duodenum of Pax4
/
mice shows near total elimination of serotonin, secretin, CCK, GIP (gastric inhibitory polypeptide), and PYY cells. Mice deficient in Pax6 show abnormal pancreatic development with no glucagon cells. In the antropyloric mucosa, Pax6
/
mice are deficient in gastrin and somatostatin cells, but show normal numbers of serotonin cells, whereas in the duodenum severe reductions in CCK and GIP, but not in serotonin, secretin, or PYY cells, are observed. These data show that the two Pax genes are needed for the differentiation of virtually all endocrine cells of the gastroduodenopancreatic region. Pax4 is transiently

176 expressed during pancreatic development and may constitute a transcriptional repressor. A fourth factor of proven importance is b2/NeuroD. Mice deficient in this factor show reductions in the number of secretin and CCK cells, while expressing apparently normal numbers of PYY, GIP, and somatostatin cells. In the pancreas, b2/NeuroD
/
mice show a 75% reduction in insulin cells, a 40% reduction in glucagon cells, and a 20% reduction in somatostatin cells. A number of additional signaling or transactivating factors have been shown to be important for cell-specific hormone expression in the pancreas but have not yet been tested for their effects on gut endocrine development. Primary candidates include transforming growth factor-b family cytokines, sonic hedgehog, neurogenin-3, Nkx6.1, and Isl-1.

GI Hormones and Endocrine Pancreas: Expressional Regulation

document that GI and pancreatic endocrine cells are of endodermal origin. It appears that different combinations of factors, acting in a strict spatial and temporal hierarchy as permanent or transient transcriptional activators and repressors, drive cell-specific hormone gene expression in the gastroduodenopancreatic region. The quest for new factors and/or combinations that may be important for directing insulin expression and tumor-specific hormone expression is ongoing.

Acknowledgments Grant support for this work was provided by the Danish Medical Research Council and Cancer Society.

See Also the Following Articles CONCLUSIONS The distal stomach, duodenum, and pancreas develop from closely adjoining areas of the primitive foregut that show very similar hormone and transcription factor expression profiles. Targeted deletion of genes shows that several factors, such as Pdx-1, Pax4, Pax6, and b2/NeuroD, are important in the development of both pancreatic and gastroduodenal endocrine cell types. As documented by gene knockout experiments, these factors are required for the expression of multiple hormones (Table I). Thus, they probably act together in different combinations to direct the cellspecific expression of different hormones. This helps to explain why certain gut hormones are transiently expressed in the pancreas and vice versa. Moreover, these observations may help to explain why certain hormones, such as gastrin, are expressed in pancreatic endocrine tumors. Interestingly, many of the factors identified are also important in the development of specific subsets of neurons. This adds to the long list of factors that are shared by endocrine cells and neurons, but does not indicate a common embryological origin. Thus, many observations solidly

CCK (Cholecystokinin) . Gastrin . GI Hormones and Endocrine Pancreas: Growth . GI Hormones Outside the Gut: Central and Peripheral Nervous System . GI Hormones Outside the Gut: Other Tissues . Pancreatic Cancer

Further Reading Alpert, S., Hanahan, D., and Teitelman, G. (1988). Hybrid insulin genes reveal a developmental lineage for pancreatic endocrine cells and imply a relationship with neurons. Cell 53, 295–308. Grappin-Botton, A., Majithia, A. R., and Melton, D. A. (2001). Key events of pancreas formation are triggered in gut endoderm by ectopic expression of pancreatic regulatory genes. Genes Dev. 15, 444–454. Herrera, P. L. (2000). Adult insulin- and glucagon-producing cells differentiate from two independent cell lineages. Development 127, 2317–2322. Jonsson, J., Carlsson, L., Edlund, T., and Edlund, H. (1994). Insulin-promoter-factor 1 is required for pancreas development in mice. Nature 371, 606–609. Larsson, L.-I. (1988). Regulatory peptides and amines during ontogeny and in non-endocrine cancers: Occurrence and possible functional significance. Prog. Histochem. Cytochem. 17, 1–222. Larsson, L.-I. (2000). Developmental biology of gastrin and somatostatin in the antropyloric mucosa of the stomach. Microsc. Res. Tech. 48, 272–281.

178 demonstrated the expression of heparin-binding epidermal growth factor-like growth factor (HB-EGF) during pancreas development. This factor, expressed abundantly in ductal cells and immature endocrine cells, can be potentially regulated by PDX-1 as both molecules colocalize during development. It is tempting to suggest that HB-EGF could mediate the effects of PDX-1 along with betacellulin and TGF-b2 via the EGF receptor expressed in the endocrine and exocrine cells of the fetal pancreas. It would seem that EGF itself is not involved since it is not expressed in these cells during development. Furthermore, Sarvetnick clearly demonstrated significant immunoreactivity of heregulin isoforms in fetal pancreatic ducts; these factors bind to the Erbb receptors, of which Erbb1 is the EGF receptor. It is strongly suggested that this receptor family operates through receptor heterodimerization between family members, with Erbb2 being the preferred partner. The association of these receptor–ligands with endocrine cells and islet formation strongly supports their involvement in mediating islet growth during pancreatic development. Hayek’s group established that hepatocyte growth factor (HGF), an expressed fetal mesenchyme-derived factor, can regulate beta cell growth and differentiation as the HGF receptor is preferentially expressed in developing beta cells. When used alone in monolayer culture, this factor was associated with a tremendous increase in the number of fetal pancreatic epithelial cells and marked down-regulation of insulin and glucagon gene expression. However, when cells were combined into three-dimensional aggregates, hormone gene expression increased, suggesting the importance of cell–cell contact for the complete biological effect of HGF. The implication of vascular endothelial growth factor (VEGF) in endocrine pancreas development was suggested when coculture of isolated endoderm with tissue fragments containing vascular endothelium induced the expression of both PDX-1 and insulin in the endodermal cells; however, tissues without vascular endothelium did not support pancreatic differentiation. Furthermore, mice expressing VEGF under the control of the PDX-1 promoter to target the factor to the developing pancreas exhibited hypervascular islets with islet hyperplasia. Although VEGF increased mitogenesis, insulin content, and insulin release in response to glucose from fetal islets, it did not stimulate direct differentiation of ductal epithelial cells into endocrine islet cells. Instead, VEGF increased the number of blood vessels or the production of an unknown factor from the vascular endothelial cells. These few examples

GI Hormones and Endocrine Pancreas: Growth

clearly indicate the importance of specific growth factors in the physiological processes leading to the development and differentiation of the fetal pancreatic islets.

Hormones Experimental data that clearly show the involvement of hormones and more specifically GI hormones during fetal pancreas development are rather scattered. Studies by Swenne indicate that growth hormone (GH) can stimulate the in vitro replication of fetal rat beta cells, an effect mimicked by prolactin and placental lactogen. However, because of the doses used, the physiological significance of the observed effects was called into question. Later, Swenne proposed that part of GH’s mitogenecity in beta cells could be attributed to insulin-like growth factor-I (IGF-I). Rhodes described IGF-I and GH signal transduction pathways and suggested that each molecule operates via its own specific route to activate different mitogenic signals. Although cholecystokinin (CCK), a gastrointestinal peptide hormone, can stimulate proliferation of the exocrine pancreas after birth, it remains unlikely that it would be involved during fetal development since its expression in fetal life is negligible. However, by immunochemistry and immunofluorescence methods, Sarvetnick detected CCK in pancreatic cells located in the acinar region of the pancreas on embryonic day 16. These data suggest that CCK signaling could be established early in development but its exact function remains unknown. However, treatments throughout pregnancy with caerulein, a CCK analogue, led Morisset to show that it induced pancreas aplasia in the mature fetus; unfortunately, the endocrine pancreas was not examined in this study. Another GI hormone that could influence the growth of the endocrine pancreas is gastrin, as it is transiently expressed in the islets, its major source in the fetus. Although the role of pancreatic gastrin in islet development remains undefined, it may influence islet growth since fetal pancreas has high levels of CCKB/gastrin receptor mRNA transcripts and high concentrations of gastrin. Although once again endocrine pancreas was not investigated, Morisset clearly demonstrated that pentagastrin treatment throughout pregnancy resulted in fetal pancreas hypertrophy, whereas treatment with L365,260, a specific CCKB receptor antagonist, caused pancreas atrophy. It may be hypothesized that the endocrine pancreas could also have been affected.

179

GI Hormones and Endocrine Pancreas: Growth

POSTNATAL ENDOCRINE PANCREAS GROWTH Normal Growth Discovery of the mechanisms of islet cell development and regeneration remains an important medical issue because of the steady increase in the incidence of diabetes among populations. For this reason, research has mostly focused on beta cell proliferation. As indicated by Bonner-Weir, an increase in islet mass in the adult can occur by neogenesis (islet formation from ductal precursor cells), replication of existing islet cells, or beta cell hypertrophy. Although the adult beta cell population has a limited proliferative potential, a considerable number of endocrine cells can regenerate. Such a process involves an orchestration of hormones and growth factor stimuli. Response to Growth Factors Kim showed that a mutation in the activin receptors led to hypoplastic pancreatic islets with a normal exocrine pancreas. Along this line, Korc established the presence of TGF-b isoforms in the adult human pancreas. TGF-b2 and TGF-b3 in particular are present and coexpressed with insulin. These observations led Korc to suggest that these factors may participate in the regulation of the biological functions of the endocrine pancreas. In perinatal rats fed a low-protein diet, reduced expression of IGF-I led to hypoplastic islets. With the HGF receptor preferentially expressed in beta cells, adult beta cells proliferate in response to HGF stimulation. These few experimental data indicate the importance of certain growth factors in controlling the normal functions of the endocrine pancreas, including turnover of their representative cell components. Response to Hormones and GI Hormones GH, prolactin, and placental lactogen are implicated in regulating marked beta cell hyperplasia and increased islet proliferation during gestation in mice and humans. From studies performed on the pancreatic beta cell line INS-1, it seems that the effect of GH on DNA synthesis would be direct and dependent on glucose concentration. Among the gastrointestinal hormones, pituitary adenylate cyclase-activating protein (PACAP), a member of the secretin/glucagon/ VIP family, was first shown to be a potent insulin secretagogue and its distribution in pancreatic nervous fibers surrounding the islets suggests a paracrine effect. However, studies by Petruzzo raised the possibility of PACAP being a growth factor. Indeed,

incubation of the peptide with purified islets resulted in the appearance of PACAP in nuclei of 80% of the islet cells; such an observation may suggest a genomic action of the internalized PACAP. Intestinal glucagonlike peptide-1, known to modulate insulin, glucagon, and somatostatin release, has been shown to stimulate PDX-1 expression and to increase islet size in mouse pancreas. Even though it has been reported that in the rat pancreas, free CCKA receptors lost their ability to secrete insulin and glucagon, the fact remains that chronic CCK treatment alone or with secretin in normal rats did not affect total insulin content while inducing exocrine pancreas growth. Similarly, as shown by Morisset, caerulein, a CCK analogue, did not modify the labeling index of the endocrine cells after 4 days of treatment, whereas all the other cell populations had their labeling index significantly increased, by even as much as 26% in acinar cells. Finally, data presented by Brand convincingly indicated that gastrin does not stimulate islet cell growth. Indeed, the pancreas of transgenic mice overexpressing gastrin had normal histology with an islet mass comparable to that of the controls.

Endocrine Pancreas Regeneration The aims of most studies related to endocrine pancreas regeneration have been to restore the beta cell mass to maintain euglycemia. After the neonatal period, the replication rate of the beta cells is low although it is known that its mass continues to grow well into adulthood. In order to study islet regeneration, different experimental animal models were used, including partial pancreatectomy, duct ligation, cellophane wrapping, and overexpression of specific genes. Pancreatectomy In 40% pancreatectomized rats, Leahy observed regrowth of much of the excised islet mass after 3 weeks. Similarly, in 60% pancreatectomized rats, beta cell regeneration was still observed but remained incomplete, with the islet non-beta-cell mass remaining unchanged. In large animals, such as the dog, there is a critical threshold of resection at approximately 88 to 92% that causes the immediate onset of diabetes. Development of diabetes is also a major setback for regeneration; indeed, after 74–92% pancreatectomy, the regeneration rate, evaluated as changes in pancreas size, was more than 40% in the nondiabetic dogs but only 15 to 23% in the diabetic animals. Treatment with insulin enhanced DNA synthesis in the remnant

p0060

180 pancreas, thus promoting regeneration. Unfortunately, the factors other than insulin that are involved in islet regeneration are still unknown. Pancreatic Duct Ligation Previous studies indicated that pancreatic duct ligation induced islet cell neogenesis from duct cells in the adult rat pancreas. Klo¨ppel found that, in response to this procedure, gastrin mRNA was strongly expressed in the newly developed duct-like cell structures along with gastrin. Expression of three isoforms of the transforming growth factor-a (TGF-a) protein was also observed. The appearance of these two growth factors preceded the peak of DNA synthesis in beta cells. These data strongly suggest that gastrin and TGF-a act as growth factors during islet neogenesis. Using a similar duct ligation model in rats, Bouwens found that infusion of gastrin for 3 days increased the beta cell mass. However, there was no mitogenic effect of gastrin on the beta cells, which is not surprising since they do not express CCKB receptors. It was then suggested that gastrin extends the process of neogenesis already under way by the duct ligation procedure with an up-regulation of gastrin and the CCKB receptors in the ductal complexes of the ligated part of the pancreas. Under these conditions, gastrin would have a paracrine effect on the responsive ductal cells through its newly expressed CCKB receptor. Cellophane Wrapping A new factor was discovered in response to cellophane wrapping of the head of the pancreas in Syrian golden hamsters. This surgical procedure led to the induction of new islet formation from ductal cells within 2 weeks through a paracrine or autocrine mechanism responsible for islet neogenesis. The responsible factor was cloned and designated INGAP (islet neogenesisassociated protein) and is a member of the Reg family. Its administration resulted in a two- to threefold increase in DNA synthesis in ductal epithelial cells, the potential precursor cells; the peptide had no effect on existing beta cells. The INGAP site of origin would be the acinar cells and its action on ductal cells would likely be paracrine, leading to islet neogenesis, as proposed by Vinik. The phenomenon was also observed in monkeys, with the same wrapping procedure being used. Overexpression of Growth Factors From studies by Brand, it was shown that pancreatic coexpression of gastrin and TGF-a resulted in significant increases in islet mass in mice with both

GI Hormones and Endocrine Pancreas: Growth

transgenes. These studies indicate that both factors can act in synergy to stimulate islet growth, whereas each peptide alone failed to do so. In transgenic mice expressing interferon-g in their islets, pancreatic duct cell proliferation and new islet formation were observed. These specific modifications were associated with the expression in the islets of the Erb-b receptors and their specific agonists, the heregulin isoforms, suggesting their involvement in islet neogenesis. Finally, the overexpression of insulin-like growth factor-II in mouse pancreas led to oversized islets with no change in the number of islets per area. In pancreas, the area occupied by insulin cells was decreased, the area occupied by glucagon cells was increased, and the area occupied by somatostatin cells was unchanged. However, given the islet cell hyperplasia, the total number of beta cells per islet was increased in these transgenic animals. Overexpression of this growth factor also resulted in increased cell replication and reduced apoptosis.

AN APPROACH TO ENDOCRINE PANCREAS GROWTH CONTROL BY GI HORMONES From data presented in this article, it is clear that the role played by the GI hormones in the development and regeneration of the pancreatic islets has remained underinvestigated over the years. Although the involvement of the various GI hormones in secretion of the different islets’ hormones has been well investigated, results on this specific subject remain controversial. As an example, ambiguity can come from the location of CCK in the pancreas. Indeed, the hormone colocalized either with insulin in rat or with glucagon in mice. Similar divergent data regarding the location of the CCKA and CCKB receptor subtypes exist. The research groups of Fourmy and Bouwens established the presence of the CCKB receptor on glucagon cells in human and rat pancreas, respectively, and Amselgruber identified the CCKA receptor on glucagon cells in the pig pancreas. Neither Fourmy nor Bouwens was able to localize the CCKB receptor on pancreatic somatostatin cells; however, Morisset and co-workers clearly showed their colocalization in human, rat, mouse, pig, dog, and calf pancreas. The authors’ approach to evaluating the growth potential of the GI hormones on pancreatic endocrine cells is systematic. Indeed, before any experimental assays on islet cell proliferation are performed, the rat purified islets were adopted as a model and the GI

GI Hormones and Endocrine Pancreas: Growth

f0005

Figure 1 CCKA and CCKB receptor mRNA (A) and protein expression (B) in total rat pancreas and purified islet. (A) RT-PCR was performed using the TITANIUM One-Step RT-PCR Kit (Clontech Laboratories, Palo Alto, CA) and 1 mg of total RNA from purified islets. Amplification of CCKAR, CCKBR, and somatostatin (SS) mRNA is observed. The amplified product sizes are 812, 669, and 231 bp for CCKAR, CCKBR, and SS, respectively. (B) Western blots of total pancreas homogenate (H), purified membranes (M), purified islets (I), and acini (A) represent 25 mg protein fractionated on a 12% sodium dodecyl sulfate–acrylamide gel transferred to a polyvinylidene difluoride membrane and probed with the polyclonal rabbit anti-CCKAR and antiCCKBR described in Table I. Specificity was established by preincubation of each antibody with its specific peptide. Bands were visualized by enhanced chemiluminescence. Localization of CCKAR, CCKBR, insulin (INS), glucagon (GLUC), and somatostatin by confocal microscopy of purified rat islets (C) and specificity of the reactions (D). The primary and secondary antibodies were incubated in 1
phosphate-buffered saline, 0.2% Triton X-100, and 1.4% normal donkey serum overnight at 4 8C and 1 h at room temperature. For specificity (D), the peptide antigens and hormones were preincubated at 40 mg/ml with their respective antibodies.

181

182

GI Hormones and Endocrine Pancreas: Growth

Table I Origin and Characteristics of the Antibodies Primary antibody

Dilution and source

Secondary antibody

Mouse anti-insulin

1:50 (Santa Cruz Biotechnologies)

Rhodamine-conjugated donkey anti-mouse IgG (Santa Cruz Biotechnologies)

Mouse anti-glucagon

1:200 (Sigma Chemical Co.)

Rhodamine-conjugated donkey anti-mouse IgG (Santa Cruz Biotechnologies)

Goat anti-somatostatin Rabbit anti-CCKAR 1122

1:50 (generous gift from P. Brazeau) 1:1500 (generous gift from M. L. Kruze)

Alexa Fluor 546 donkey anti-goat (Molecular Probes) Alexa Fluor 488 donkey anti-rabbit (Molecular Probes)

Rabbit anti-CCKBR 9262

1:200 (generous gift from J. Walsh)

Alexa Fluor 488 donkey anti-rabbit (Molecular Probes)

hormones CCK and gastrin were chosen because their receptors are well characterized and because specific antibodies, raised against parts of these molecules, are also readily available. On these islets, the authors have thus far evaluated (1) the expression of the CCKA and CCKB receptor mRNA, (2) the presence of these two receptor proteins by Western blot, and (3) the specific localization of each receptor protein subtype among the islet cells by confocal microscopy. As shown in Fig. 1, the data clearly show that rat purified islets express both CCKA and CCKB receptor subtypes as indicated by the presence of their respective mRNA (Fig. 1A) and protein (Fig. 1B). Note that the specificity of the two CCK receptor antibodies used was established by preincubation with their specific antigens (Fig. 1B, Peptide; see also Table I). Furthermore, by confocal microscopy, the CCKA receptors were specifically localized on insulin and glucagon cells, whereas the CCKB receptors appeared exclusively on the somatostatin cells (Fig. 1C). As shown in Fig. 1D, preabsorption of each antibody with its corresponding peptide or hormone confirms the specificity of the data. With this CCK receptor distribution in mind, it becomes possible to design more accurate protocols (1) to study the anti-diabetogenic action of CCK8 in type 2 diabetics as shown by Ahren, (2) to explain the increased pancreatic contents of somatostatin in diabetic animals and its potential role in the development of type 2 diabetes, and (3) to determine the growth potential of CCK and gastrin on specific islet cells using hormone concentrations corresponding to the affinity of each CCK receptor subtype. It will also become easier to test and eventually use specific CCK receptor antagonists to target and treat pathologies of

the endocrine pancreas. It is sincerely believed that characterization of the receptors and determination of their specific cellular location are critical data needed to study the role of GI hormones in any physiological response attributed to these hormones.

See Also the Following Articles CCK (Cholecystokinin) . Gastrin . GI Hormones and Endocrine Pancreas: Expressional Regulation . GI Hormones as Growth Factors . Growth Hormone (GH) . Pancreatic Cancer . Pancreatic Polypeptide (PP)

Further Reading Bonner-Weir, S. (2000). Islet growth and development in the adult. J. Mol. Endocrinol. 24, 297–302. Morisset, J., and Grondin, G. (1989). Dynamics of pancreatic tissue cells in the rat exposed to long-term caerulein treatment. 2. Comparative analysis of the various cell types and their growth. Biol. Cell 66, 279–290. Morisset, J., Laine´, J., and Mimeau-Worthington, T. (1999). Hormonal control of rat fetal pancreas development. Biol. Neonate 75, 327–336. Rhodes, C. J. (2000). IGF-I and GH post-receptor signaling mechanisms for pancreatic b-cell replication. J. Mol. Endocrinol. 24, 303–311. Rooman, I., Lardon, J., and Bouwens, L. (2002). Gastrin stimulates b-cell neogenesis and increases islet mass from transdifferentiated but not from normal exocrine pancreas tissue. Diabetes 51, 686–690. Sumi, S., and Tamura, K. (2000). Frontiers in pancreas regeneration. J. Hepatobiliary Pancreatic Surg. 7, 286–294. Wang, R. N., Rehfeld, J. F., Nielsen, F. C., and Kloppel, G. (1997). Expression of gastrin and transforming growth factor a during duct to islet cell differentiation in the pancreas of duct-ligated adult rats. Diabetologia 40, 887–893.

184 number of gastric inhibitory peptide-, substance P-, somatostatin-, and serotonin-producing enteroendocrine cells was also observed in these mice, suggesting that several small intestinal endocrine cell lineages are developmentally related. Finally, expression of a dominant negative Pax6 allele in SEYNeu mice results in reduced levels of proglucagon mRNA transcripts in the small and large intestine and an almost complete absence of enteroendocrine cells containing glucagon-like peptide-1 (GLP-1) or glucagon-like peptide-2 (GLP-2) immunoreactivity.

Enteric Nervous System

p0025

The ENS consists of autonomic ganglia located in the submucosal and myenteric plexus and associated connecting neural structures in the bowel wall. Afferent sensory neurons detect both chemical and mechanical stimuli arising from the intestinal wall and lumen and efferent motor neurons interact with effector cells including smooth muscle cells, pacemaker cells, mucosal glands, blood vessels, epithelia, and cells that regulate both immune response and intestinal endocrine function. The neurons and glial cells of the ENS are derived from the neural crest. During early fetal development, a wave of migrating neural crest (NC) cells (somites 1–5 in the mouse) enters the foregut mesenchyme and colonizes the entire length of the gastrointestinal tract in a rostral-to-caudal manner. Truncal-derived NC cells (somites 6 and 7 in the mouse) also contribute to the development of the ENS in the foregut (esophagus and stomach) and sacral-derived NC cells contribute to the development of the hindgut. As NC cells colonize the gut, these precursor cells undergo cellular proliferation and differentiation into glial cells and other cell types that constitute the ENS. Little is known about the various factors that regulate cellular differentiation; however, genetic loss-of-function experiments in rodents have identified several genes that are essential for regulating the spatial development of the ENS. For example, the transcription factor MASH1 is required for the development of the ENS in the esophagus; glial-derived neurotrophic factor and its receptors RETand GDNF family receptor a-1 as well as the transcription factor Phox2b are essential for the development of the ENS distal to the stomach; whereas endothelin 3 and its receptor Ednrb/edn3/ece1 as well as the transcription factors sox10 and Hoxa-4 are essential for the development of the ENS in the hindgut. Several members of the hedgehog gene family are also important for developmental formation of the ENS.

GI Hormones as Growth Factors

ENTEROENDOCRINE CELL-DERIVED PEPTIDES Gastrin In the gastrointestinal tract, gastrin is produced predominantly in G cells located in the gastric antrum and duodenal bulb. The 101-amino-acid preprogastrin precursor polypeptide is posttranslationally processed into several biologically active molecular forms, including G-34, G-17, and G-14, which can be either amidated or glycine-extended. The amidated forms of gastrin are essential for regulating gastric acid secretion and the progastrin-derived peptides exhibit multiple trophic effects on the stomach, pancreas, and colon. Oncofetal Relationships Stomach Gastrin content increases gradually from birth and reaches a peak at 5 weeks of age, after which levels decrease in adult rats. Although glycine-extended gastrin (G-Gly) precursors increase in the first day of life, the process of weaning results in a further increase in gastric gastrin content and enhanced gastrin amidation. The amidated forms of gastrin (gastrin-17 and -34) stimulate proliferation of gastric stem cells and enterochromaffin-like (ECL) cells in the oxyntic mucosa, resulting in an increased parietal and ECL mass and ensuring adequate gastric acid production in the rodent stomach during the weaning process. Mice with a null mutation in the gastrin gene exhibit decreased numbers of parietal and ECL cells and reduced gastric acid production. Gastric carcinoid tumors occur in approximately 30% of patients with hypergastrinemia due to gastrin-producing tumors as part of the multiple endocrine neoplasia type 1 syndrome, whereas approximately 5% of patients with pernicious anemia and hypergastrinemia secondary to reduced parietal cell mass will develop ECL cell carcinoid tumors, which may resolve following surgical resection of the antrum. Long-term treatment of animals with proton pump inhibitors is also associated with hypergastrinemia, ECL hyperplasia, and the development of carcinoid tumors. Pancreas The biological effects of the progastrin-derived peptides on the endocrine pancreas have yet to be clearly defined. The fetal pancreas of rodents contains a large amount of amidated gastrin that is first detected by E15 and persists until just after weaning. This time period is associated with considerable fetal islet

185

GI Hormones as Growth Factors

growth and differentiation and suggests a possible role for gastrin in pancreatic endocrine cell development. Transgenic mice expressing amidated gastrin in the islet adult beta cell exhibit a minimal phenotype; however, a doubling of islet cell mass is detected when these mice are crossed with transgenic mice expressing transforming growth factor-a (TGF-a). Gastrin is not essential for islet development, as gastrin-deficient mice do not demonstrate abnormalities in pancreatic islet morphology. Colon Progastrin and the progastrin-derived peptides can be detected in the rat colon during fetal development. Following birth, amidated gastrins are no longer detected in the colon and the levels of glycineextended gastrins gradually decrease such that only progastrin can be detected in the adult rat colon. The mechanisms underlying the trophic effects of progastrin and G-Gly on the colonic epithelium remain incompletely defined. Transgenic mice expressing progastrin or G-Gly exhibit increased colonocyte proliferation and mucosal thickness and are more prone to form aberrant crypt foci following the administration of azoxymethane. Consistent with these observations, targeted inactivation of the gastrin gene results in reduced rates of colonic epithelial proliferation. Gastrin stimulates the proliferation of several colon cancer cell lines expressing the gastrin receptor (CCK2); however, the majority of colon cancer cell lines and normal colonic epithelium do not normally express CCK2. A truncated gastrinbinding receptor and a constitutively active CCK2 receptor mutant have been identified in some human colorectal cancers. These findings form the basis for the potential use of gastrin-neutralizing antisera for the treatment of subsets of patients with colon cancer.

Cholecystokinin CCK was initially described as a factor that stimulates gallbladder contraction. In the gastrointestinal tract, CCK is expressed in ‘‘open-type’’ enteroendocrine I cells located in the proximal small intestine and in nerve fibers branching to the gastric and colonic myenteric plexus and submucosal plexus, where it acts as a neurotransmitter. The CCK gene encodes a 94-amino-acid prohormone that is posttranslationally processed in a tissue-specific fashion into CCK83, CCK58, CCK39, CCK33, CCK22, CCK8, and CCK5, all sharing a common C terminus. The major active form of CCK is an octapeptide containing a sulfated

tyrosine residue and an amidated C-terminal phenylalanine residue. Oncofetal Relationships CCK cannot be detected in fetal and neonatal pancreatic islet cells but is expressed in islets during and after weaning, suggesting that CCK is probably not important for development of the endocrine pancreas. CCK stimulates pancreatic enzyme secretion and exhibits trophic effects on pancreatic acini. Chronically elevated levels of CCK are associated with pancreatic hyperplasia and enhanced tumor formation. Rats fed a soybean trypsin inhibitor that enhances the release of CCK develop preneoplastic pancreatic lesions that eventually progress to carcinoma (Table I). Similarly, long-term pancreatobiliary diversion in the rat causes elevated levels of plasma CCK, pancreatic growth, and premalignant changes. Elevated circulating levels of CCK also enhance the development of preneoplastic acinar lesions induced by azaserine, a potent pancreatic carcinogen in rats. In contrast, loss of CCK signaling produces variable effects on the pancreas. The Otsuka Long-Evans Tokushima Fatty rat fails to express the CCK-A receptor and exhibits reduced pancreatic size, whereas the CCK-A receptor knockout mouse exhibits normal pancreatic morphology. CCK may also play a role in stimulating the growth and invasiveness of human pancreatic cancer cell lines.

GLP-1 and GLP-2 The proglucagon gene is expressed in pancreatic A cells, in enteroendocrine L cells, and in the brainstem and hypothalamus. In mammals, both glucagon and the glucagon-like peptides are encoded within a single proglucagon precursor prohormone. Posttranslational processing of proglucagon occurs in a tissue-specific manner, resulting in the liberation of the proglucagon-derived peptides, which possess remarkably diverse biological actions. In the pancreatic A cell, posttranslational processing liberates glucagon and the major proglucagon fragment, whereas in enteroendocrine L cells and brain, posttranslational processing liberates glicentin, oxyntomodulin, GLP-1, GLP-2, and several spacer or intervening peptides. Oncofetal Relationships Proglucagon-immunoreactive cells and proglucagon mRNA transcripts are first detected in the rat intestine by E14; however, an intestinal profile of

186

t0005

GI Hormones as Growth Factors

Table I Intestinal-Derived Factors and Their Growth-Promoting Effects Growth factor

Site of intestinal expression

Gastrin

G cells

Stomach oxyntic mucosa, colonic epithelium, pancreas (?)

ECL cellular hyperplasia, carcinoid tumors, colonic adenocarcinoma

CCK

I cells and ENS

Pancreatic acini

Pancreatic hyperplasia/carcinoma, tumor invasion (?)

GLP-1

L cells

Pancreatic beta cell differentiation and neogenesis

?

GLP-2

L cells

Stomach, small intestine, colonic epithelium

?

NT

N cells and ENS

Small intestinal and colonic epithelium, pancreas (?)

Colonic and pancreatic adenocarcinoma

PYY

L cells

Small intestinal and colonic epithelium

?

TRH

G cells and ENS

Pancreas

?

VIP

ENS

Stimulates the release of neurotrophic factors important in development and functioning of ENS

Promotes growth of non-small-cell lung cancer and pancreatic adenocarcinoma but inhibits growth of gastric adenocarcinoma

Bombesin and related peptides

ENS

Pancreas

Pancreatic adenocarcinoma, gastrinoma, colonic adenocarcinoma, invasion and metastasis of colon and prostate cancer (?)

Trophic effects

Oncogenic effects

Note. ECL, enterochromaffin-like; ENS, enteric nervous system; CCK, cholecystokinin; GLP-1, glucagon-like peptide-1; GLP-2, glucagonlike peptide-2; NT, neurotensin; PYY, peptide YY; TRH, thyrotropin-releasing hormone; VIP, vasoactive intestinal peptide.

glucagon-like immunoreactive peptides cannot be detected until E17–19 when the molecular machinery associated with posttranslational processing of proglucagon has developed. GLP-1 GLP-1 inhibits food intake, gastric emptying, and glucagon secretion and stimulates insulin secretion. Although the GLP-1 receptor is expressed in multiple tissues during murine development, a role for GLP-1 during embryonic development has not been established. GLP-1 stimulates beta cell proliferation, beta cell differentiation, and beta cell neogenesis; however, GLP-1 receptor
/
mice develop normally and GLP-1R
/
islets exhibit only minor abnormalities in islet cell formation, precluding a major role for GLP-1 in islet cell development. Administration of GLP-1 to aging rats leads to an increase in beta cell mass, whereas activation of GLP-1R signaling enhances islet proliferation in rats and mice with experimental diabetes. GLP-2 GLP-2 is produced in the gut during late gestation, following which levels increase progressively in neonatal rats and then subsequently decrease to adult levels. Similarly, GLP-2 receptor expression can be detected along the entire length of the gastrointestinal

tract during late fetal development and during the neonatal period. Exogenously administered GLP-2 induces growth of the stomach, small intestine, and colon of neonatal rats. In contrast, administration of GLP-2 to fetal pigs does not produce an intestinotrophic response. Furthermore, although Pax6 mutant SEYNeu mice exhibit a marked reduction in the number of GLP-2-producing L cells, intestinal development appears relatively normal in these mice, strongly suggesting that GLP-2 is not essential for the development of the fetal murine intestine. GLP-2 exerts trophic effects on the small intestine and colonic mucosa by stimulating crypt cell proliferation and inhibiting apoptosis within the crypt and villus compartments. GLP-2 also exerts multiple actions independent of intestinal growth including enhancement of intestinal epithelial barrier function and stimulation of intestinal hexose transport. The therapeutic utility of GLP-2 has been demonstrated in pilot studies of human subjects with short bowel syndrome and in rodent models of intestinal disease including major small bowel resection, total parenteral nutrition (TPN)-induced intestinal hypoplasia, small and large intestinal inflammation, chemotherapy-induced mucositis, and intestinal ischemia–reperfusion injury. Intriguingly, the GLP-2 receptor has been localized to human endocrine cells and to murine submucosal and myenteric neurons, highlighting the

p0085

187

GI Hormones as Growth Factors

interaction between the enteric endocrine and nervous systems in controlling intestinal epithelial growth.

Neurotensin Neurotensin (NT) is a 13-amino-acid peptide originally discovered in bovine hypothalamus. Peptides that share structural homology with NT include neuromedin N, xenin, and xenopsin. In the gastrointestinal tract, NT is synthesized by N cells located within the small intestinal mucosa, predominantly the ileum, and can also be detected within enteric neurons. Oncofetal Relationships In the human fetal colon, NT is transiently expressed at midgestation between 16 and 18 weeks and the extent of NT expression in the adult human colon is unclear. In the adult rat, NT administration augments the adaptive response to small bowel resection in the intestinal remnant and stimulates growth of the colonic epithelium in vivo. NT receptor expression has been detected in human colonic adenocarcinomas and pancreatic cancer cell lines. The importance of NT for pancreatic growth remains unclear as NT binding is detectable by autoradiography in pancreatic cancer cells but not in normal human pancreas.

Peptide YY PYY, neuropeptide Y, and pancreatic polypeptide are members of the pancreatic polypeptide superfamily. All three peptides contain 36 amino acids, share considerable amino acid homology, demonstrate amidated C-terminal ends, and possess several tyrosine residues (single-letter abbreviation Y) including tyrosines at both the amino and carboxy termini. Only PYY has been implicated as a putative intestinal epithelial growth factor. Oncofetal Relationships PYY is one of the first hormones to be expressed in the developing fetal gastrointestinal tract and colocalizes with GLP-1 in a subset of enteroendocrine L cells present in the ileum, colon, and rectum. PYY coexpression is detected in the majority of colonic enteroendocrine cells throughout fetal and postnatal development, with the exception of serotoninproducing cells, suggesting a possible common PYYþ progenitor cell for enteroendocrine cells in the intestine. Immunoreactive PYY has also been detected

in the developing endocrine pancreas and in a subpopulation of glucagon-producing A cells in mature islets, although a role for PYY in the endocrine pancreas has yet to be elucidated. In the intestine, PYY regulates small and large bowel motility and inhibits both water and chloride secretion. PYY expression is first detected in the rat colon at E17 and levels remain relatively elevated until after weaning (postnatal day 21) when PYY mRNA transcripts decline to adult levels. Administration of PYY to nursing rats increased the weight and DNA content of the duodenum, suggesting a potential trophic role for PYY during intestinal development and dietary adaptation. In the adult mouse, administration of PYY increased the weight and DNA content of the duodenum, ileum, and colon, whereas co-infusion of PYY in TPN-fed rats increased jejunal mass and protein content. In the pancreas, PYY inhibits pancreatic exocrine secretion but does not exhibit direct trophic effects on either the exocrine or the endocrine pancreas.

Thyrotropin-Releasing Hormone Thyrotropin-releasing hormone (TRH) is expressed throughout the gastrointestinal tract in gastric G cells, in pancreatic islet beta cells, and in neurons constituting the myenteric plexus of the esophagus, stomach, and intestine. In the stomach, TRH modulates pentagastrin-stimulated gastric acid secretion and may attenuate acid secretion in subjects with acid secretory disorders. In the pancreas, TRH is expressed during perinatal development and TRH administration in the adult rodent induces pancreatic hyperplasia and inhibits amylase release. TRH also reduces CCK-induced gallbladder smooth muscle contraction and inhibits cholesterol synthesis within the intestinal mucosa but is not trophic to the intestinal mucosa.

ENS-DERIVED PEPTIDES Vasoactive Intestinal Peptide Vasoactive intestinal peptide (VIP) is a 28-amino-acid neuropeptide originally isolated from porcine small intestine. VIP belongs to a family of neuropeptides that includes pituitary adenylate cyclase-activating peptide, peptide histidine isoleucine, and peptide histidine methionine, which function as neurotransmitters and neuromodulators of the enteric nervous system.

188 Oncofetal Relationships In the human digestive tract, immunoreactive VIP can be detected as early as 10 weeks gestation and in the pancreas as early as 9 weeks gestation. VIP stimulates the release of neurotrophic factors from glial cells, promotes the growth of astrocytes and several nonsmall-cell lung cancers, stimulates the growth of VIP-1 receptor-bearing pancreatic adenocarcinomaderived cells in vitro, and inhibits the growth of colon adenocarcinoma cells in vitro.

Bombesin and Related Peptides The bombesin family of peptides was originally isolated from frog skin extracts and includes bombesin, gastrin-releasing peptide (GRP; the mammalian homologue of bombesin), neuromedin B, and neuromedin C. Bombesin is a tetradecapeptide, both neuromedin B and neuromedin C are decapeptides, and GRP is a 27-amino-acid peptide. Oncofetal Relationships Administration of bombesin to 7-day-old rats resulted in increased stomach, intestinal, and pancreatic weights. GRP is also expressed in human cancer cells, including neuroendocrine tumors of the lung (carcinoids and small-cell lung cancer), breast cancer, neoplastic human thyroid C cells, pancreatic cancer, stomach cancer, and colon cancer. Bombesin stimulates the growth of several human small-cell lung cancers and pancreatic cancer cell lines and the detection of GRP and GRP RNA in neoplastic cells suggests that GRP may potentially function as an autocrine growth factor. Bombesin stimulates enhanced growth of xenografted tumors in vivo and GRP receptor mRNA has been detected in colonic tumors but not in normal colonic epithelium.

SUMMARY p0150

Trophic peptides exert effects on the intestinal epithelium via stimulation of crypt cell proliferation and/or inhibition of apoptosis in the crypt and villus compartments. These biological actions may be important for intestinal development and for ensuring adequate epithelial regeneration following intestinal injury. Although the enteric endocrine and nervous systems secrete peptides important for gut growth, intestine-derived growth factors, such as epidermal growth factor, TGF-a, hepatocyte growth factor,

GI Hormones as Growth Factors

and keratinocyte growth factor, along with growth hormone, insulin-like growth factor-I, interleukin-11, and interleukin-15, also exhibit trophic effects on the gastrointestinal epithelium. These factors function in a coordinated interdependent manner to regulate fetal organ maturation and mucosal healing and repair following epithelial injury. Understanding how these peptides interact to exert their cytoprotective and regenerative effects may be important for realizing the clinical potential of one or more gut peptides and growth factors for the treatment of human intestinal injury.

See Also the Following Articles CCK (Cholecystokinin) . Gastrin . Gastrin-Releasing Peptide . GI Hormone Development (Families and Phylogeny) . GI Hormones and Endocrine Pancreas: Growth . GI Hormones in Cancer . Glucagon-like Peptide 2 (GLP-2) . Neurotensin . Peptide YY (PYY) . Thyrotropin-Releasing Hormone (TRH)

Further Reading Boushey, R. P., Yusta, B., and Drucker, D. J. (2001). Glucagon-like peptide (GLP)-2 reduces chemotherapy-associated mortality and enhances cell survival in cells expressing a transfected GLP-2 receptor. Cancer Res. 61, 687–693. Chave, H. S., Gough, A. C., Palmer, K., Preston, S. R., and Primrose, J. N. (2000). Bombesin family receptor and ligand gene expression in human colorectal cancer and normal mucosa. Br. J. Cancer 82, 124–130. Drucker, D. J., Ehrlich, P., Asa, S. L., and Brubaker, P. L. (1996). Induction of intestinal epithelial proliferation by glucagon-like peptide 2. Proc. Natl. Acad. Sci. USA 93, 7911–7916. Hellmich, M. R., Rui, X. L., Hellmich, H. L., Fleming, R. Y., Evers, B. M., and Townsend, C. M., Jr. (2000). Human colorectal cancers express a constitutively active cholecystokinin-B/gastrin receptor that stimulates cell growth. J. Biol. Chem. 275, 32122–32128. Jensen, J., Pedersen, E. E., Galante, P., Hald, J., Heller, R. S., Ishibashi, M., Kageyama, R., Guillemot, F., Serup, P., and Madsen, O. D. (2000). Control of endodermal endocrine development by Hes-1. Nat. Genet. 24, 36–44. Naya, F. J., Huang, H., Qiu, Y., Mutoh, H., DeMayo, F., Leiter, A. B., and Tsai, M.-J. (1997). Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in b2/NeuroD-deficient mice. Genes Dev. 11, 2323–2334. Wang, T. C., Koh, T. J., Varro, A., Cahill, R. J., Dangler, C. A., Fox, J. G., and Dockray, G. J. (1996). Processing and proliferative effects of human progastrin in transgenic mice. J. Clin. Invest. 98, 1918–1929. Yusta, B., Boushey, R. P., and Drucker, D. J. (2000). The glucagonlike peptide-2 receptor mediates direct inhibition of cellular apoptosis via a cAMP-dependent protein kinase-independent pathway. J. Biol. Chem. 275, 35345–35352.

190 Elevation of serum gastrin levels is a common occurrence via a number of pathological, physiological, and pharmaceutical mechanisms. First, it occurs as a result of pernicious anemia, in which autoimmune antibodies directed against parietal cells result in reduced acid secretion. Second, it can occur in patients with Zollinger-Ellison syndrome, caused by islet cell tumors secreting gastrin. Finally, Helicobacter pylori infection and administration of proton pump inhibitors lead to an increase in mainly amidated, mature forms of gastrin. Colonic proliferation has been shown to be altered in these conditions of hypergastrinemia, with an increase in normal crypt cell proliferation. Preclinical models evaluating the effect of hypergastrinemia, induced by proton pump inhibitors (PPIs), on the development of carcinogeninduced colon cancer, show no deleterious effect of the hormone. However, in the APCMin mouse model of familial adenomatous polyposis and in a human colonic adenoma xenograft grown in nude mice, PPI-induced hypergastrinemia resulted in increased adenoma proliferation and reduced survival. A carefully controlled clinical study has consolidated these preclinical findings by showing that elevated gastrin levels (>90 pg/ml) were associated with a three- to fourfold increased risk of colorectal cancer development. A number of studies have demonstrated cholecystokinin-2 (CCK2)/gastrin receptors on GI tumor cells and exogenously administered gastrin peptides were found to have proliferative effects both in vitro and in vivo. In addition to the endocrine mechanism of gastrin’s action, activation of the gastrin gene within GI adenocarcinoma cells results in the production of immature gastrin peptides, mediating proliferation in an autocrine/paracrine manner. Specific gastrin peptides implicated in the promotion of GI tumor growth include amidated G17, GlyG17, glycine-extended G34, and progastrin. Gastrin mediates its physiological effects through the G protein-coupled seven-transmembrane domain CCK2 receptor. The receptors mediating the effects of trophic precursor forms of gastrin still require clarification and universal agreement. However, a number of CCK2 receptor isoforms that result from alternative splicing have been described, including an amino-terminal truncation and retention of introns 2 and 4. Certain members, in particular those with retained introns, may be tumor-specific and have been implicated in binding precursor gastrin species. The gastrin gene is a downstream target of b-catenin-mediated transcription, suggesting that

GI Hormones in Cancer

the gastrin autocrine pathway is operational early in the adenoma–carcinoma sequence and may play a role in CCK2 receptor up-regulation and tumor progression. Gastrin circumvents apoptosis in a number of experimental systems. Gastrin stimulated the phosphorylation and subsequent activation of the survivalinducing protein kinase B/Akt, activated in many cancers, promoting proliferation and resistance to chemotherapy and radiation. Furthermore, autocrine gastrin produced by the colorectal cancer cell line HCT116 reduced the activation of caspase-3 and up-regulated cytochrome c oxidase Vb. This resulted in a decreased sensitivity of the cells to pro-apoptotic stimuli by retaining cytochrome c within the mitochondria. Camptothecin activation of caspases 3 and 9 was also reduced by progastrin in the intestinal cell lines IEC18 and IEC6. The incidence of mutated p53 was elevated in Mastomysnatelensis carcinoid model mice with serum hypergastrinemia. In vitro gastrin stimulation of the AGS gastric cancer cell line increased the expression of p53. Such an effect on mutant p53 may also increase resistance to apoptotic stimuli. The role of gastrin as a transcriptional activator has expanded to include many key targets involved in the establishment and maintenance of malignancy. Gastrin activates the transcription of heparin-binding epidermal growth factor, a potent angiogenic factor, and amphiregulin. This leads to a concomitant upregulation of the EGF receptor. Gastrin peptides increase activated matrix metalloproteinase 2 secretion in the human colon cancer cell line LoVo, resulting in enhanced invasion, as well as up-regulated cyclooxygenase 2 (COX-2) gene and protein expression. The Reg family of genes, which are activated during regeneration of intestinal mucosa and which are positively correlated with colorectal cancer recurrence, have been shown to be up-regulated by gastrin. Longterm administration of the PPI lansoprazole strongly increased the expression of the Reg gene, which was reversed by inclusion of a CCK2 receptor antagonist.

OPTIONS FOR BLOCKADE OF GASTRIN-MEDIATED GROWTH PATHWAYS Gastrin Receptor Blockade Higher affinity CCK2 receptor antagonists have been developed, including the benzodiazepam derivative L-365,260, which, in preclinical studies, reverses gastrin-stimulated GI tumor growth. Another CCK2

p0045

p0050

p0055

191

GI Hormones in Cancer

receptor antagonist, Gastrazole ( JB5008), has been shown to be highly potent in preclinical studies and has been used in an open-labeled pilot clinical trial in patients with pancreatic cancer. When compared to historical controls, a significant survival advantage was suggested. A number of CCK2 receptor antagonists block the effect of Gly-gastrin peptides, thus circumventing potential autocrine gastrin pathways. YM022 was shown to reverse both G17- and GlyG17-stimulated proliferation of the human colon cancer cell line LoVo, in addition to inhibiting basal growth. JMV1155 reduced the GlyG17-stimulated in vivo growth of the human colon tumor DLD1. CCK2 receptor antagonists that block endocrine and potential autocrine pathways exist. Studies delineating the importance of splice variants in terms of tumor expression and ligand specificity will direct development of the next generation of CCK2 receptor antagonists.

Gastrin Immunoneutralization p0075

Anti-gastrin antibodies have been shown to be effective at reducing the basal growth of a gastrin-secreting colon cancer cell line in vivo and in vitro. A continuous infusion of anti-gastrin antibodies is difficult to administer, may be unstable, may cause anaphylaxis, and may fail to achieve consistent therapeutic serum levels sufficient to neutralize postprandial gastrin surges. Such limitations can be overcome by the use of active immunization against gastrin species. G17DT is a gastrin immunogen, composed of a nonapeptide derived from the amino-terminal of human gastrin-17 linked to diphtheria toxoid which acts as an immunogenic carrier. The G17 epitope is monovalent, thereby preventing complement fixation. Antibodies raised do not cross-react with other hormones, such as G34, smaller C-terminal fragments, or CCK. G17DT antibodies raised in rabbits have a high affinity for G17 and GlyG17 and functionality was confirmed by their ability to competitively displace gastrin ligands (125I-G17) from the CCK2 receptor expressed by a tumor cell line even at dilutions of 1:100 of the original serum titer.

Preclinical Efficacy of G17DT Antibodies G17DT has been shown to exert therapeutic effects following passive immunization in nude mouse models of gastric, pancreatic, and colorectal cancer.

Clinical Studies of G17DT G17DT has been assessed in phase I/II trials in advanced colorectal cancer patients; increased survival of G17DT patients compared to a well-matched placebo group was observed. A dose-finding phase II study of G17DT in 22 patients with pancreatic carcinoma demonstrated greater survival in patients who mounted an adequate antibody response than in nonresponders. G17DT is being assessed in phase III trials for gastric, pancreatic, and colorectal cancer.

SOMATOSTATIN Somatostatin exerts its actions through interaction with specific heptahelical G protein-coupled plasma membrane receptors. Five different somatostatin receptor subtypes have been cloned in humans. Different receptor subtypes are coupled to different intracellular transmission cascades in a cell typedependent manner. Somatostatin can also exert cytostatic (G1-phase cell arrest) or cytotoxic (apoptosis induction) effects, depending on the receptor subtype expressed on the target cell. In gastroenteropancreatic neuroendocrine tumors, a predominance of sst1 and sst2 with a lesser extent of sst3 and sst5 subtype receptors has been demonstrated. Since the short half-life of somatostatin makes continuous intravenous infusion mandatory, several long-acting analogues have been synthesized. Of these, octreotide (which binds mainly to somatostatin receptor subtypes sst2 and sst5) has been the most extensively investigated. These synthetic analogues have specific decreasing affinity for sst2 > sst5 > sst3 receptor subtypes and have been used as antiproliferative drugs in the treatment of gastroenteropancreatic tumors. Octreotide and lanreotide treatment resulted in a modest growth-inhibition activity, in functioning or nonfunctioning tumors. Longer-lasting formulations of somatostatin analogues have been developed to provide patients with the convenience of monthly administration and to ensure stable drug serum concentrations between injections. Side effects of these agents consist mainly of gastrointestinal complaints, cholelithiasis, and effects on glucose metabolism. Inoperable liver tumors have an unfavorable natural course despite various therapeutic modalities. Octreotide, a somatostatin analogue, has shown considerable antitumoral activity in animal models of various hepatic tumors and in isolated cell culture lines. A randomized controlled trial of octreotide in the treatment of advanced hepatocellular carcinoma

p0095

192

p0110

has shown a significant survival benefit in the treated patients. Literature reports indicate a stimulatory effect of octreotide on Kupffer cells as a possible antitumor mechanism. Octreotide administration is the best available treatment for advanced inoperable hepatocellular carcinoma and better patient selection, based on receptor subtypes, might further improve the results in the future. Somatostatin analogues have been used in numerous preclinical studies providing contradictory evidence on the growth inhibition of ductal pancreatic adenocarcinoma. Monotherapy did not result in a prolongation of survival; however, in 15–20% of patients, the progression of the process was halted for several months accompanied by a significant improvement of the clinical condition without notable side effects. Somatostatin analogues have also been used in combination with tamoxifen in which patients with unresectable and resected ductal adenocarcinoma of the pancreas had an apparently increased survival when compared to historical controls. Octreotide has been used to treat hypersecretion in Zollinger-Ellison syndrome in a small study. It had an antitrophic effect on parietal cell mass. Octreotide is extremely useful for palliative care. It has analgesic properties when administered by the spinal and intraventricular routes. Its actions in reducing gut motility and secretions make it a valuable adjunct in the management of inoperable bowel obstruction, fistulae, and intractable diarrhea. Octreotide, for example, has been extensively used to ameliorate the gut motor dysfunction that characterizes carcinoid diarrhea.

SOMATOSTATIN AS A DIAGNOSTIC TOOL Radiolabeled somatostatin analogues have been employed for the localization of primary and metastatic tumors expressing somatostatin receptors. The so-called ‘‘somatostatin receptor scintigraphy’’ is an important clinical diagnostic investigation for patients with suspected neuroendocrine tumors.

SOMATOSTATIN TARGETED THERAPY p0130

An extension of somatostatin scintigraphy is somatostatin receptor-targeted chemotherapy and radiotherapy (with conjugates of somatostatin peptides and cytotoxic drugs) and gene therapy (e.g., transferring the sst2 gene into neoplastic cells). Having been successfully tested in experimental studies, these therapies are being evaluated in clinical trials.

GI Hormones in Cancer

GASTRIN-RELEASING PEPTIDE Gastrin-releasing peptide (GRP) and its seventransmembrane-domain G protein-coupled receptor (GRPR) are frequently expressed by cancers of the gastrointestinal tract, breast, lung, and prostate. Most studies have found that GRP acts to increase tumor cell proliferation, leading to the hypothesis that it is an important mitogen for the growth of these cancers. Gastrin-releasing peptide receptors are normally expressed on intestinal smooth muscle cells rather than epithelial cells of the GI tract. However, aberrant receptor expression has been shown on GI adenocarcinomas. Early studies revealed that 24–40% of human colon cancers overexpressed GRP receptors. However, a later study showed that 62% of the 50 colonic cancers examined aberrantly overexpressed both GRPR and the GRP hormone, unlike the normal adjuvant epithelium. The receptors were expressed equally across the different grades of tumor and appeared to be down-regulated in metastases, being expressed in only 1 of 37 metastases. Poorly differentiated tumors were less likely to coexpress GRP and GRPR than well-differentiated tumors and it was concluded that the proteins act as morphogens rather than as mitogens. GRPR mutations have also been described and result in the lack of production of functional receptor protein, indicating that studies exclusively examining gene expression of the GRPR may not reflect their biological significance in malignant progression. Such mutated GRPRs have also been shown in nonantral gastric adenocarcinomas. Further evidence for the role of GRPR as a morphogen was provided in both wild-type and GRPRdeficient C57BL/J6 mice treated with the carcinogen, axoxymethane. Tumors that were induced in wild-type mice expressed GRPR and were well differentiated, whereas tumors in GRPR-deficient mice were poorly differentiated mucinous adenocarcinomas. It was concluded that aberrant expression of GRPR and GRP did not result in larger tumors but due to GRP increasing focal adhesion kinase, the receptors promoted a well-differentiated phenotype. The lack of a proliferative effect of GRP via GRPR was reinforced in a study of colorectal tumor specimens, where it was shown that 93% of tumor specimens expressed GRPR mRNA, which was expressed at higher levels in tumors with lymphatic vessel invasion but there was no relationship with p53 expression or proliferation index. Activation of GRPR results in stimulation of activation protein 1 expression, which has been shown to impact on

p0165

193

GI Hormones in Cancer

COX-2 expression. In a rat intestinal cell line overexpressing GRPR, bombesin was shown to stimulate COX-2 mRNA and protein expression in addition to prostaglandin E2 production. These data suggested that a COX-2-dependent pathway may be responsible for the effect of GRP on GRPR expressed by colon cancer cells. A number of bombesin receptor antagonists have been described by Schally’s group. The bombesin/ GRP antagonist RC-3095 inhibited the growth of the human gastric MKN45 xenograft model and the HT29 human colon xenograft model in nude mice.

FUTURE THERAPEUTIC APPLICATIONS Radiolabeled agonists may be used for imaging and therapy, as they appear to be internalized, yielding a higher target:background ratio.

See Also the Following Articles CCK (Cholecystokinin) . Childhood Cancer, Endocrine Effects of . EGF and Related Growth Factors . Gastrin . Gastrin-Releasing Peptide . Gastrointestinal Neuroendocrine Tumor Syndromes (GI NETS) . GI Hormones as Growth Factors . GI Tract, General Pathology of Endocrine Growths . Pancreatic Cancer . Pancreatic Islet Cell

Tumors . Parathyroid Cancer . Prostate Cancer . Somatostatin Analogs . Thyroid Carcinoma

Further Reading Carroll, R. E., Matkowskyj, K. A., Chakrabarti, S., McDonald, T. J., and Benya, R. V. (1999). Aberrant expression of gastrin-releasing peptide and its receptor by well-differentiated colon cancers in humans. Am. J. Physiol. 276, G655–G665. Carroll, R. E., Ostrovskiy, D., Lee, S., Danilkovich, A., and Benya, R. V. (2000). Characterization of gastrin-releasing peptide and its receptor aberrantly expressed by human colon cancer cell lines. Mol. Pharmacol. 58, 601–607. Hocker, M., and Wiedenmann, B. (1999). Therapeutic and diagnostic implications of the somatostatin system in gastroenteropancreatic neuroendocrine tumour disease. Ital. J. Gastroenterol. Hepatol. 31, S139–S142. Jensen, J. A., Carroll, R. E., and Benya, R. V. (2001). The case for gastrin-releasing peptide acting as a morphogen when it and its receptor are aberrantly expressed in cancer. Peptides 22, 689–699. O’Byrne, K. J., Schally, A. V., Thomas, A., Carney, D. N., and Steward, W. P. (2001). Somatostatin, its receptors and analogs, in lung cancer. Chemotherapy 47, 150–161. Scarpignato, C., and Pelosini, I. (2001). Somatostatin analogs for cancer treatment and diagnosis: An overview. Chemotherapy 47, 1–29. Vainas, I. G. (2001). Octreotide in the management of hormonerefractory prostate cancer. Chemotherapy 47, 62–77. Watson, S. A., and Gilliam, A. D. (2001). G17DT—A new weapon in the therapeutic armoury for gastrointestinal malignancy. Expert Opin. Biol. Ther. 1, 309–317.

GI Hormones Outside the Gut: Central and Peripheral Nervous Systems

195

(and tumors). The gene encoding a specific regulatory peptide may also be expressed by multiple cell types. These patterns may themselves change during evolution, giving rise to species differences in cellular expression.

endoplasmic reticulum (Fig. 1). The product (prepropeptide) is usually biologically inactive and is rapidly converted to a precursor peptide (propeptide) that progresses through the Golgi complex and is sequestered in secretory vesicles that bud from the trans-Golgi network. During transit through the Golgi complex, or in secretory vesicles, the propeptide is subject to any of a variety of different posttranslational modifications. These may include glycosylation, Tyr sulfation, and Ser phosphorylation (all of which occur in the Golgi complex), cleavage of the peptide chain, or COOH-terminal a-amino amidation (which occurs in secretory vesicles). The patterns of posttranslational modification for a specific peptide may differ in neurons and endocrine cells. The Golgi-derived secretory vesicles often possess electron-dense cores and their morphology is characteristic of particular cell types. These vesicles are distinct from the small, clear, synaptic vesicles that recycle at nerve terminals and contain classical transmitters such as acetylcholine, monoamines, g-aminobutyric acid (GABA), and glutamate. Expression of the genes encoding gut hormones is physiologically regulated, but different mechanisms may regulate the expression of these genes in PNS or CNS neurons.

Receptors Hormonal and neurotransmitter peptides typically act at G protein-coupled receptors (GPCRs). GPCRs constitute a large family of structurally related proteins that are thought to have evolved by the duplication and divergence of ancestral genes. Multiple GPCRs may exhibit differential affinity for the various members of regulatory peptide families. The receptors may themselves be expressed by multiple cell types including CNS or PNS neurons and nonneuronal cells, e.g., secretory cells or smooth muscle cells.

CRITERIA FOR THE IDENTIFICATION OF GUT HORMONES IN PNS OR CNS NEURONS Evidence for the presence of gut hormones in PNS and CNS neurons should be based on the following lines of evidence: (1) expression of the gene identified by techniques such as reverse transcription-polymerase chain reaction, Northern blot analysis, and in situ hybridization and confirmation by cloning and sequencing of the relevant cDNA; (2) the presence of the relevant peptide determined by radioimmunoassay, preferably using antibodies to several epitopes, coupled with high-performance liquid chromatography and confirmation by bioassay, chemical isolation, and amino acid sequencing; and (3) demonstration of cellular origins by immunocytochemistry, preferably using multiple antibodies to several epitopes. Localization of neuropeptides may be a useful way of distinguishing subsets of neurons already identified by the presence of classical transmitters and for this purpose double-labeling immunocytochemistry is useful. There may be problems with the specificity of particular antibodies. For these reasons, the interpretation of studies based on the use of a single experimental method, e.g., immunohistochemistry, using a single antibody should be approached with caution.

CELLULAR ASPECTS Regulatory peptides produced in gut endocrine cells, CNS neurons, or PNS neurons are, in each case, generated by mRNA translation at the rough

COTRANSMISSION Increased intracellular Ca2þ leads to exocytosis of both Golgi-derived (neuropeptide-containing) and synaptic vesicle populations (Fig. 2). The actions of neuropeptides are exerted at GPCRs, which are not exclusively localized to postsynaptic membranes or even to neurons. As a consequence, peptides may act over longer distances and for longer times than classical neurotransmitters. Where both neuropeptide and classical transmitters act on the same postsynaptic cell, there may be interactions between them. For example, increased or decreased sensitivity to the classical transmitter may be mediated by neuropeptides.

OVERVIEW OF GUT HORMONES EXPRESSED BY PNS NEURONS Representatives of the gut hormone families are commonly expressed in the major divisions of the autonomic nervous system. In postganglionic parasympathetic neurons, the secretin-like peptides VIP, peptide histidine isoleucine amide (PHI), and pituitary adenylate cyclase-activating peptide (PACAP) are

196

GI Hormones Outside the Gut: Central and Peripheral Nervous Systems

Figure 1 Schematic representation of the intracellular progression of hormonal peptides produced by gut endocrine cells (A) and by central or peripheral neurons (B). In both cases, mRNA is translated at the endoplasmic reticulum (ER) and precursor peptides progress through the Golgi complex; mature peptides are stored in Golgi-derived secretory vesicles (GD-SV) and released by calcium-dependent exocytosis. Peptide released at the basolateral membrane of endocrine cells is conveyed to target cells via the circulation. In neurons, secretory vesicles progress along the axon to nerve terminals for release; in this case, the secreted peptide typically diffuses to its target cells, although peptides released into the hypothalamo-hypophyseal portal vessels are an exception.

commonly expressed, whereas in postganglionic sympathetic neurons, neuropeptide tyrosine (NPY) is commonly expressed. The enteric nervous system, which is sometimes considered to be a third division of the autonomic nervous system, is an abundant source of neuropeptides, including VIP, CCK, and NPY. Many primary afferent neurons express neuropeptides; some of those belonging to the main families of gut hormones (VIP, NPY) are up-regulated after nerve damage. Gut hormonal peptides are not generally found in somatic efferent neurons.

regions, including cerebral cortex, hippocampus, brainstem, and striatum.

MAJOR GUT HORMONE FAMILIES AND THEIR EXPRESSION IN CNS AND PNS NEURONS The following brief sketches are grouped on the basis of peptide families that include at least one member that is produced and released by gut endocrine cells and at least one that is produced and released by CNS or PNS neurons.

OVERVIEW OF GUT HORMONES EXPRESSED BY CNS NEURONS

The Cholecystokinin/Gastrin Family

The expression of neuropeptides is a property of very many CNS neurons. Gut hormones or related peptides, e.g., somatostatin, CCK, and NPY, are expressed in hypothalamic neurons, where they may act in local circuits or in the control of pituitary hormone release after secretion into hypothalamohypophyseal portal vessels. In addition, these and other gut hormones are found in many other CNS

The CCK gene is expressed in intestinal I-type endocrine cells and in many CNS neurons and some PNS neurons. There are differences in posttranslational processing that account for the reported variation between neurons and endocrine cells in the major form of CCK present. In cerebral cortical neurons, CCK may occur together with GABA and in some nigrostriatal neurons it occurs with dopamine.

197

GI Hormones Outside the Gut: Central and Peripheral Nervous Systems

Figure 2 Schematic representation of some events involving neurotransmission by classical (A) and neuropeptide (B) transmitters. Classical transmitters are stored in small, clear synaptic vesicles that are recycled at the nerve terminal. There are reuptake mechanisms for the neurotransmitter, which is resequestered in these vesicles. Receptors for classical transmitters are concentrated in the postsynaptic membrane and are often ion channels. In contrast, neuropeptides are located in Golgi-derived dense-cored vesicles and after secretion they diffuse away from the release site. Neuropeptides typically act at G protein-coupled receptors (GPCR) to trigger second-messenger responses; the receptors may be diffusely distributed on postsynaptic cells and may be some distance from the release site. (Z) Putative transmitter.

Gastrin originates from pyloric antral G cells and may be found in small amounts in hypothalamus. The CCKA (also called CCK1) receptor has high affinity for CCK and low affinity for gastrin; it is expressed in pancreas, gallbladder, and some CNS and PNS neurons. The gastrin–CCKB (also called CCK2) receptor has high affinity for both gastrin and CCK and is found on parietal and enterochromaffin-like cells and many CNS neurons. The main function that has been ascribed to CCK in the brain is inhibition of food intake. In addition, CCK is thought by some to be associated with anxiety or panic attacks and it may modulate noxious sensations.

The Secretin/Glucagon/VIP/PHI/PACAP/ GIP Superfamily The superfamily of structurally related peptides that includes the classical gut hormones secretin and glucose-dependent insulinotropic peptide (GIP) and the pancreatic hormone glucagon also includes major neuropeptide transmitters such as VIP, PHI, and PACAP. In the case of glucagon, there is good evidence that a single precursor molecule may yield several different products through alternative posttranslational processing. In the pancreatic islet alpha cells, glucagon itself is a primary product; in intestinal L cells, glucagon-like peptide-1 (GLP-1) and GLP-2

are primary products (but not glucagon). CNS neurons resemble intestine more closely than pancreas in the processing of the glucagon precursor; one possible role for the CNS GLPs is in the control of food intake. Two members of the family that seem to be predominantly expressed in neurons, and not endocrine cells, are VIP (the precursor of which also gives rise to the related peptide PHI) and the closely related peptide PACAP.

Somatostatin Somatostatin-producing endocrine (D) cells are found throughout the gastrointestinal tract and in the islets of Langerhans. In these systems, somatostatin acts locally as a paracrine inhibitor of secretion from nearby cells, e.g., G cells in the pyloric antrum and beta cells in the pancreas. In addition, somatostatin is produced in hypothalamic neurons and functions as an inhibitor of growth hormone secretion from the anterior pituitary. It is also found in enteric neurons, sympathetic neurons, and some somatic afferent neurons.

Neurotensin Neurotensin was first isolated from the hypothalamus and then discovered to be produced in N-type

p0065

198 endocrine cells of the ileum and colon, where it is a putative mediator of the ileal brake in the gut. Neurotensin has been identified in multiple CNS neurons; interactions with dopamine have been identified and possible actions as an anti-psychotic agent reported. The Peptide YY/NPY/Pancreatic Polypeptide Family Peptide YY is named for the presence of tyrosine residues (Y in the single-letter notation) at both the COOH- and NH2-terminal positions. It is normally produced in endocrine cells of the ileum and colon. Putative hormonal actions include the inhibition of pancreatic secretion, intestinal transit, and food intake. There is little evidence for neurotransmitter functions. But the closely related NPY is widely expressed in CNS and peripheral neurons. One of its most striking actions is the stimulation of food intake on injection into the hypothalamus. In addition, it is present in sympathetic postganglionic neurons and may modulate responses to noradrenaline. Another member of the family is pancreatic polypeptide, produced in the islets of Langerhans; its functions as either a hormone or a neurotransmitter remain uncertain.

Ghrelin and Motilin The related peptides ghrelin and motilin are produced in gastric and small intestinal endocrine cells,

GI Hormones Outside the Gut: Central and Peripheral Nervous Systems

respectively. They are distinct from other gut hormones in that they are released from these cells during fasting. Ghrelin is a naturally occurring ligand for an orphan receptor originally identified as a putative mediator of growth hormone secretion. Small amounts are thought to be produced in the hypothalamus. It is a powerful stimulant of food intake and it is thought that the peptide released from the stomach may stimulate food intake by delivery to the hypothalamus through the circulation.

See Also the Following Articles CCK (Cholecystokinin) . Gastrin . Ghrelin . GI Hormones Outside the Gut: Other Tissues . Motilin . Neuropeptide Y . Neurotensin . Peptide YY (PYY) . Pituitary Adenylate Cyclase-Activating Polypeptide (PACAP)/Glucagon Superfamily . Somatostatin Analogs

Further Reading Dockray, G. J. (1999). The brain gut axis. In ‘‘Textbook of Gastroenterology’’ (T. Yamada, D. H. Alpers, L. Laine, C. Owyang, and D. W. Powell, eds.), 3rd ed. Vol. 1, pp. 67–81. Raven Press, New York. Furness, J. B. (2000). Types of neurons in the enteric nervous system. J. Auton. Nervous Syst. 81, 87–96. Hokfelt, T., Broberger, C., Xu, Z. Q., Sergeyev, V., Ubink, R., and Diez, M. (2000). Neuropeptides—An overview. Neuropharmacology 39, 1337–1356. Walsh, J. H., and Dockray, G. J. (1994). ‘‘Gut Peptides: Biochemistry and Physiology.’’ Raven Press, New York.

200

GI Hormones Outside the Gut: Other Tissues

Secretin Family

Gastrin-Releasing Peptide/Bombesin

Several members of the secretin family [secretin, glucagon, vasoactive intestinal polypeptide (VIP), pituitary adenylate cyclase-activating peptide (PACAP), and growth hormone-releasing factor (GRF)] have been localized to Leydig cells, germ cells, or both. Additionally, PACAP, VIP, and GRF have been localized to the ovaries and GRF has also been localized to the placenta. Secretin expression has been detected in the pituitary and VIP expression has been noted in adrenal chromaffin cells, in mast cells, and in leukocytes. VIP suppresses T cell proliferation and production of interleukin-2 (IL-2), IL-4, and IL-10 and acts as an anti-inflammatory mediator. However, VIP may also enhance certain aspects of lymphocyte function and these opposing actions may be mediated by different subtypes of VIP receptors. The effects of VIP on lymphocytes also reflect the rather abundant innervation of lymphoid tissues by VIP nerves.

Gastrin-releasing peptide (GRP) has been localized to somatotrophs in the anterior lobe and to melanotrophs of the intermediate lobe of the pituitary and it exerts growth hormone-stimulating activity. GRP also occurs in NEBs of the bronchopulmonary tree, particularly during fetal and neonatal life, and has been localized to endocrine-like cells of the prostate. It is produced by a wide selection of human cancer cells of lung (SCLC), mammary, and prostate origin in which it may act as an autocrine growth factor.

Neurotensin Neurotensin has been detected in adrenal chromaffin cells and in leukocytes. In addition, neurotensin is produced by SCLC cells.

Somatostatin Somatostatin shows a widespread distribution and has been localized to neuroendocrine neurons ending on hypophyseal portal blood vessels in the median eminence, thyroid parafollicular C cells, parathyroid cells of some species, adrenal chromaffin cells, Langhans cells of the epidermis, endocrine-like cells of the prostate and toad urinary tract, and leukocytes. In all of these locations, it seems certain that somatostatin acts as a local paracrine mediator or as a short-range (portal system-delivered) hormone. In general, somatostatin acts as an inhibitor of secretory activities and also reduces cell proliferation, possibly through actions on specific phosphatases. Somatostatin has also been localized to several types of tumors, including breast cancers. It inhibits the proliferation of several normal and neoplastic cell types. Interestingly, antisense oligonucleotides blocking somatostatin expression stimulate the proliferation of rat splenocytes. This effect could be inhibited by the addition of exogenous somatostatin. Additionally, somatostatin inhibits the release of mediators (histamine, leukotriene D4) from human and leukemic basophils.

CONCLUSIONS The above data clearly demonstrate that many regulatory peptides simultaneously are produced by cells derived from all three germ layers and that their designation as either gastrointestinal or hormonal is contextual rather than absolute. Many of the peptides may play roles in normal organs, in reproductive functions, and in the immune and cellular defense system of the body (Fig. 1). Additionally, some factors are transiently expressed during development and may serve as tumor growth factors. Indeed, an impressive array of peptides are expressed by lung, gastrointestinal, pancreatic, breast, and prostate cancers and preclinical data suggest that they are potential targets for therapy.

Genitourinary System GRF PACAP VIP

Immune System

Gastrin Secretin CCK GRP

Somatostatin

Neurotensin Pituitary Thyroid Parathyroid Adrenal Medulla NEBs

Figure 1 Scheme summarizing some of the overlaps in the distribution of GI hormones and regulatory peptides.

GI Hormones Outside the Gut: Other Tissues

Acknowledgments Grant support for this work was provided by the Danish Medical Research Council and Cancer Society.

See Also the Following Articles CCK (Cholecystokinin) . Gastrin . GI Hormones and Endocrine Pancreas: Expressional Regulation . GI Hormones Outside the Gut: Central and Peripheral Nervous System

Further Reading Aguila, M. C., Rodriguez, A. M., Aguilla-Mansilla, H. N., and Lee, W. T. (1996). Somatostatin antisense oligodeoxynucleotidemediated stimulation of lymphocyte proliferation in culture. Endocrinology 137, 1585–1590. Bajo, A. M., Schally, A. V., Krupa, M., Hebert, F., Groot, K., and Szepeshazi, K. (2002). Bombesin antagonists inhibit growth of MDA-MB-435 estrogen-independent breast cancers and

201 decrease the expression of the ErbB-2/HER-2 oncoprotein and c-jun and c-fos oncogenes. Proc. Natl. Acad. Sci. USA 99, 3836–3841. Heasley, L. E. (2001). Autocrine and paracrine signaling through neuropeptide receptors in human cancer. Oncogene 20, 1563–1569. Kimoto, Y. (1998). A single human cell expresses all messenger ribonucleic acids: The arrow of time in a cell. Mol. Gen. Genet. 258, 233–239. Lambrecht, B. N. (2001). Immunologists getting nervous: Neuropeptides, dendritic cells and T cell activation. Respir. Res. 2, 133–138. Larsson, L.-I. (1988). Regulatory peptides and amines during ontogeny and in non-endocrine cancers: Occurrence and possible functional significance. Prog. Histochem. Cytochem. 17, 1–222. Schalling, M., Persson, H., Pelto-Huikko, M., Ødum, L., Ekman, P., Gottlieb, C., Høkfelt, T., and Rehfeld, J. F. (1990). Expression and localization of gastrin mRNA and peptide in spermatogenic cells. J. Clin. Invest. 86, 660–669. Sherwood, N. M., Krueckl, S. L., and McRory, J. E. (2000). The origin and function of the pituitary adenylate cyclase-activating polypeptide (PACAP)/glucagon superfamily. Endocr. Rev. 21, 619–670.

209

GI Tract, General Anatomy (Cells)

t0005

EC cells (displaying a yellow fluorescence due to their high endogenous content of 5-HT) from other endocrine cells (displaying green fluorescence due to the newly formed dopamine after L-DOPA administration). These amine-handling properties of gut endocrine cells, as well as some other histochemical features, resemble those of neurons and it was long thought that the embryonic origin of the gut endocrine cells was neuroectodermal. However, it is generally agreed that the endocrine cells, like other epithelial cells lining the GI tract, are of endodermal origin. During the past few decades, immunocytochemistry, with the use of antibodies against individual hormones or other specific cell markers, has become the most important technique for distinguishing the individual endocrine cell populations with respect to their cellular morphology, their regional and topographic distribution, and—when hormone antibodies are used—their function. However, there are still histochemically or ultrastructurally defined cell populations to which a hormone has not yet been linked. Conversely, there are cell populations that differ in their hormone content, but nevertheless share both light microscopic features and secretory granule appearance.

Table I History of Gut Endocrine Cells Heidenhein (1870) Nussbaum (1879),

Chromaffin cells described Gritzner and Menzel (1879) Osmiophilic cells described

Kultschitzky (1897) Ciaccio (1907)

Basigranular acidophil cells described Term enterochromaffin cells coined

Kull (1913)

Nonargentaffin endocrine-like cells described

Masson (1914)

Argentaffin cells described

Vialli and Erspamer (1937)

Enteramine in EC cells described

Feyrter (1938)

Clear cells described

Erspamer (1939)

Argyrophilic cells (interpreted as EC cell progenitors) described

Dawson (1948)

Argentophilic (argyrophil) cells in the stomach described

Erspamer and Asero (1952) Enteramine determined to be identical to 5-HT (serotonin)

formaldehyde vapors under certain conditions (the Falck-Hillarp technique). Using this technique, it was found that the endocrine cells in the GI tract had the capacity to form fluorogenic monoamines, such as dopamine and 5-HT, on administration of the corresponding amine precursor (L-DOPA and 5-hydroxytryptophan, respectively) and to store them in large amounts in their secretory granules. These properties of the cells have sometimes been referred to as APUD (amine precursor uptake and decarboxylation) and became an important means of collectively demonstrating gut endocrine cells, at the same time providing the possibility of discriminating the

t0010

REGIONAL DISTRIBUTION The endocrine cells have as a rule a regional distribution along the GI tract that is characteristic for each

Table II Endocrine Cells in the Mammalian Gastrointestinal Tract Stomach Cell type

Oxyntic

Small intestine

Pyloric

a

Upper

Lower

Large intestine

Hormone

A

þ









Glucagon

A-like/X

þ

þ

þ





Ghrelin

D

þ

þ

þ

þ

þ

Somatostatin

D1 EC

þ þ

þ þ

þ þ

þ þ

þ þ

Unknown (ghrelin in human) Serotonin þ various peptides

ECL

þ









Histamine

G



þ

þ





Gastrin

I





þ

þ



Cholecystokinin

K





þ

þ



GIP

L





þ

þ

þ

Gut glucagon/GLP-1 and -2 þ PYY

MO





þ

þ



Motilin

N P

 þ

 þ

 þ

þ 

 

Neurotensin Unknown

S





þ

þ



Secretin

a

In certain species, e.g., carnivores.

210

GI Tract, General Anatomy (Cells)

Proximal duodenum

EC

D

L

K

I

G

S

MO

N

"All endocrine cells"

Distal duodenum Jejunum Distal ileum Proximal colon Distal colon Rectum 20 cells/unit length of mucosa

Figure 1 Schematic diagram illustrating the distribution and relative frequency (number of cells/unit length of mucosa) of several different populations of endocrine/paracrine cells in the human intestine. Data are based on immunocytochemical observations using hormone antibodies. (Left) Black dots on the outline of the intestine indicate tissue sampling sites. (Right) A compilation of the data. Note that some cell types are distributed throughout the intestine, whereas other cell types have a more restricted distribution. Note also that the rectum is almost as rich in endocrine/paracrine cells as the duodenum. cell type (Table II, Fig. 1). Thus, certain cell types are restricted to, or greatly predominate, in the stomach. Others are restricted to the upper small intestine and still others predominate in the distal small intestine and/or the large intestine. A few cell types are distributed all along the GI tract. Interestingly, there are several indications that such widely distributed cells may serve to influence cells in their vicinity rather than to deliver their product(s) to the circulation. Such widely distributed cells are therefore often described as paracrine cells. However, since the possibility cannot be excluded that they may, in fact, have dual paracrine and endocrine functions, and since they may share several major morphological features with cells known to be genuinely endocrine, they are sometimes referred to as endocrine/paracrine. With respect to genuinely endocrine cells, it has been argued that each such population should have a restricted distribution along the GI tract in order for the target cells to be able to respond to a circulating messenger in a meaningful way. Nevertheless, an established endocrine function of a cell does not preclude an additional paracrine function.

Stomach Based on various histochemical techniques, including silver stains, the APUD properties, and immunohistochemistry, and, in particular electron microscopy, with special emphasis on the morphological features of the secretory granules (exemplified in Fig. 2), at least seven endocrine cell types have been described in the stomach. They comprise G cells, D cells, EC cells, enterochromaffin-like (ECL) cells, A-like cells (also referred to as X cells), and two types of cells with small

granules and designated P cells and D1 cells (Table II). Additionally, in certain species, notably carnivores such as dogs and cats, ‘‘true’’ glucagon-producing A cells, being well-known constituents of the pancreatic islets, are also present in substantial numbers in the stomach. The ECL cells constitute one—and possibly the only—gastric endocrine cell type that is confined to the acid-producing, upper part (corpus fundus or body) of the stomach (Table III). This seems to be the case in all vertebrate species examined, from fish to human. Moreover, the ECL cells are the largest endocrine cell population in this area of the stomach. The ECL cells have a very characteristic secretory granule ultrastructure, which at least in part seems to be linked to the histamine production of these cells. Also, the A-like cells, and, when present, the A cells, are primarily found in this part of the stomach, but they may also occur in smaller numbers in the distal non-acid-producing part (antrum, pylorus) of the stomach. The G cells are exclusively found in the antrum, where they are located within a zone at the upper part of the glands in most species. However, in rodents they predominate at the basal portion of the glands. The EC cells greatly predominate in the antrum in certain mammals, such as rodents, whereas in the stomach of human and many larger mammals they also occur in the acid-producing part of the stomach in quite high numbers. The D cells are diffusely distributed all over the stomach in all species. These cells very often have narrow cytoplasmic processes that can have a beaded, nerve fiber-like appearance, ending with a knob-like swelling filled with granules (Fig. 3). The small granule cells (P and D1 cells), like the D cells, seem to occur all over the stomach. The

211

GI Tract, General Anatomy (Cells)

Figure 2 Electron micrographs of secretory granules of endocrine/paracrine cells in the rat stomach. (A) A-like cell. The granules are round and highly electron-dense. (B) ECL cell. The granules are of the vesicular type with a flocculent electrondense core surrounded by a wide electron-lucent halo. (C) G cell. Note the varying electron density of the granules. (D) EC cell. Note granule pleomorphism.

Table III Relative Frequency of Different Endocrine Cells in Gastric Oxyntic Mucosa of Rat and Human Cell type

Rat (%)

Human (%)

ECL cells

65

35

A-like (X) þ P þ D1 cells

24

14

Somatostatin cells

10

26

0

25

EC cells

D cells have been found to often make direct contact with G cells in the antrum and with parietal cells in the acid-producing part of the stomach. This contact is usually established via the cytoplasmic processes of the D cells. These features, together with the wide distribution of the cells both within the GI tract and outside of it (the pancreatic islets in particular), have contributed to the designation of the D cells as paracrine cells.

212

f0015

GI Tract, General Anatomy (Cells)

Figure 3 (A) Schematic morphological distinction between endocrine and paracrine cells. (B and C) D cells of rat gastric mucosa (somatostatin immunofluorescence). (B) The cells issue long, slender cytoplasmic processes, sometimes ending in a knob-like swelling. (C) D cell process ending on a G cell (stained black by gastrin immunoperoxidase). (D) Electron micrograph showing a D cell process filled with secretory granules. The process runs along the base of the gland epithelium near the top of the figure.

A general morphological characteristic of the endocrine cells in the upper, acid-producing part of the stomach (i.e., the corpus or fundus) is that they are ‘‘closed’’ in that the apical end of the cell does not reach the gland lumen. In the antrum, on the other hand, the vast majority of cells do reach the lumen via a narrow neck (Fig. 4). One possible explanation for this morphological difference is that open cells respond to lumenal stimuli (e.g., nutrients, pH

changes), whereas closed cells are sensitive to other types of stimuli (e.g., distension, temperature changes, neuronal and hormonal messengers).

Small Intestine Endocrine cells are distributed as single cells within the epithelium of both crypts and villi, but collectively they predominate in the crypts. They are

213

GI Tract, General Anatomy (Cells)

characteristically flask-shaped with a narrow, often quite elongated neck reaching the lumen and thus of open type. The secretory granules accumulate at the base of the cell, but may sometimes be more diffusely distributed, although they are always confined to the cytoplasm and are never found in the nucleus. The classification and naming of the intestinal endocrine cells were for a long time based primarily on the size and morphology of the secretory granules. Accordingly, the cells were categorized as, e.g., S (small granules), I (intermediate-size granules), and L (large granules) cells (Table II, Fig. 1). Subsequent immunohistochemical observations about the cellular localization of various hormones complicated matters and resulted in attempts to retain the letter classification, the inevitable renaming of some cells, and the addition of hormonally classified cells that did not fit into previous classifications. Thus, the classification of endocrine cells in the small intestine is often a mixture of letter naming and naming based on the hormone content of the cells. At least 11 different cell types have been identified (Table I). The EC cells are distributed throughout the small intestine. The D cells also occur throughout the small intestine, although in gradually smaller numbers distally. Some cell types are more restricted in their regional distribution. Thus, S cells, I [cholecystokinin (CCK)] cells, K cells, and MO (motilin) cells predominate in the duodenum and upper jejunum and are only rarely found in the ileum. The opposite is true for L cells and N cells, which predominate in the ileum, but are scarce in the duodenum. According to ultrastructural criteria, P cells are also present in the small intestine, with a distribution mainly in the upper part. One cell type with very large granules (VL cells) has been reported to occur throughout the human small intestine.

Large Intestine

Figure 4 Electron micrographs illustrating (A) an endocrine cell of closed type (A-like cell) in the acid-producing mucosa and (B) an open endocrine cell in the antral mucosa (G cell) of the rat stomach. Note numerous microvilli at the lumenal surface.

The large intestine also harbors endocrine cells of several different kinds, although it seems that the number of different populations is somewhat smaller than in the small intestine. Interestingly, the endocrine cells are collectively more numerous in the rectum than in the colon, as calculated in the human gut (Fig. 1). At least five different endocrine cell populations can be distinguished in the large intestine. It is notable that there is a marked overlap in the distribution of certain cell populations between the distal small intestine and the large intestine (Table II). Thus, EC cells and L cells are quite numerous also in the large intestine. N cells, on the

214

GI Tract, General Anatomy (Cells)

other hand, are only rarely found in the large intestine and D cells are also clearly fewer in number here than in the small intestine. As to morphology, many of the endocrine cells have a very characteristic appearance in that each cell at its base issues a long, tapering, sometimes beaded process running along the base of the epithelium. Furthermore, the process often seems to be directed toward the base of the crypt. This gives the cells a paracrine-like appearance. However, since cells thought to play a role as hormone producers, such as the L cells, also have this morphology, it is still open to speculation why the cells are equipped with such long processes. It is not inconceivable that the cells serve both endocrine and paracrine functions.

CONCLUDING REMARKS As is obvious from the data presented here, there is some confusion and remaining uncertainties as to the classification and naming of endocrine cells in the GI tract. This is due to the fact that endocrine cells sharing staining characteristics with cells occurring outside the GI tract, notably, the pancreatic islets, which are known by the ‘‘islet’’ name. This applies to islet A cells and D cells. Another problem arose when an ultrastructurally defined cell type, e.g., the L cell, was later found to comprise at least two functionally distinct cell populations, one producing proglucagonderived peptides (retaining the L cell designation) and another producing neurotensin (renamed N cells).

Figure 5 Endocrine cell plasticity illustrated by the marked hyperplasia of rat ECL cells (B) and G cells (D) on pharmacological blockade of acid secretion for 8 weeks. Acid blockade causes G cell activation and hyperplasia with hypergastrinemia, which in turn brings about activation of the ECL cells with increased histamine secretion and, with time, proliferation. ECL cells are visualized with histidine decarboxylase immunofluorescence and G cells are visualized with gastrin immunofluorescence.

215

GI Tract, General Anatomy (Cells)

Another problem is related to the I cell. Obviously, several functionally distinct endocrine cell populations fall into this somewhat vague category. Thus, CCK cells, K cells, and MO cells are all ‘‘I’’ cells, but with the current possibilities of a more detailed functional classification, the hormone name, or its short name, is gradually replacing the original letter name. Furthermore, the largest population of endocrine cells, the EC cells, are probably not one single cell type, but rather several functionally distinct populations, unified merely because they produce large amounts of serotonin. This view is favored by available ultrastructural data showing different types of granules in EC cells in different regions of the gut, and sometimes even in different EC cells within the same region, together with the fact that certain regulatory peptides are restricted to subpopulations of EC cells only. Finally, it ought to be mentioned that there is a remarkable potential for plasticity in gut endocrine cell systems. This is perhaps most readily observed in the stomach, where each of the endocrine cell populations has a restricted regional distribution and the cells are more frequent per unit area. Thus, marked hyperplasia of G cells and ECL cells (within their normal regional boundaries) evolves within some time (days to a few weeks) after, e.g., profound pharmacological blockade of acid secretion (Fig. 5) and ECL cell hypoplasia after, e.g., removal of the G cells (antrectomy).

See Also the Following Articles GI Hormone Development (Families and Phylogeny) . GI Hormones Outside the Gut: Central and Peripheral Nervous System . GI Hormones Outside the Gut: Other Tissues . GI Tract, General Pathology of Endocrine Growths

Further Reading Bordi, C., and D’Adda, T. (1991). Ultrastructural morphometry of gastric endocrine cells. In ‘‘The Stomach as an Endocrine Organ’’ (R. H a˚kanson and F. Sundler, eds.), pp. 53–69. Elsevier, Amsterdam, The Netherlands. Buchan, A. M. J. (1999). Structure and function of gastrointestinal endocrine cells. In ‘‘Gastrointestinal Endocrinology’’ (G. H. Greeley, ed.), pp. 1–30. Humana Press, Totowa, NJ. Capella, C., and Solcia, E. (1972). The endocrine cells of the pig gastrointestinal mucosa and pancreas. Arch. Histol. Jpn. 35, 1–29. Capella, C., Finzi, G., Cornaggia, M., Usellini, L., Luinetti, O., Buffa, R., and Solcia, E. (1991). Ultrastructural typing of gastric endocrine cells. In ‘‘The Stomach as an Endocrine Organ’’ (R. Ha˚kanson and F. Sundler, eds.), pp. 27–51. Elsevier, Amsterdam, The Netherlands. Dayal, Y. (1999). Neuroendocrine cells of the gastrointestinal tract: Introduction and historical perspective. In ‘‘Endocrine Pathology of the Gut and Pancreas’’ (Y. Dayal, ed.), pp. 1–31. CRC Press, Boca Raton, FL. ¨ ber diffuse endokrine epitheliale Organe.’’ Feyrter, F. (1938). ‘‘U J. A. Barth, Leipzig, Germany. Sjo¨lund, K., Sande´n, G., Ha˚kanson, R., and Sundler, F. (1983). Endocrine cells in human intestine: An immunocytochemical study. Gastroenterology 85, 1120–1130. Solcia, E., Capella, C., Buffa, R., and Frigerio, B. (1976). Histochemical and ultrastructural studies on the argentaffin and argyrophil cells of the gut. In ‘‘Chromaffin, Enterochromaffin and Related Cells’’ (R. E. Coupland and T. Fujita, eds.), pp. 209–225. Elsevier, Amsterdam, The Netherlands. Solcia, E., Capella, C., Vassallo, G., and Buffa, R. (1975). Endocrine cells of the gastric mucosa. Int. Rev. Cytol. 42, 223–286. Sundler, F., and Ha˚kanson, R. (1988). Peptide hormone-producing endocrine/paracrine cells in the gastro-entero-pancreatic region. In ‘‘Handbook of Chemical Neuroanatomy’’ (A. Bjo¨rklund, T. H o¨kfeldt, and Ch. Owman, eds.), Vol. 6, pp. 219–295. Elsevier, Amsterdam, The Netherlands. Sundler, F., and Ha˚kanson, R. (1991). Gastric endocrine cell typing at the light microscopic level. In ‘‘The Stomach as an Endocrine Organ’’ (R. Ha˚kanson and F. Sundler, eds.), pp. 9–26. Elsevier, Amsterdam, The Netherlands. Sundler, F., Ha˚kanson, R., Lore´n, I., and Lundquist, I. (1980). Amine storage and function in peptide hormone-producing cells. Invest. Cell Pathol. 3, 87–103.

217

GI Tract, General Pathology of Endocrine Growths

manipulation in mice. Proliferating or transformed gut endocrine cells may variably express the same antigens as their normal counterpart depending on their differentiation status. Expression of the SSR2 in endocrine tumor cells is important for both diagnostic and therapeutic applications.

NONNEOPLASTIC GROWTHS OF GUT ENDOCRINE CELLS Nonneoplastic growths of gut endocrine cells have been reported only in the stomach and occasionally in the small intestine.

Stomach Nonneoplastic growths are restricted to proliferation of histamine-producing enterochromaffin-like (ECL) cells in the corpus/fundus and of gastrin-producing (G) and somatostatin-producing (D) cells in the antrum/pylorus. Hypergastrinemia-promoted ECL cell hyperplasia is categorized as diffuse (an increase in endocrine cells of more than twice the standard deviation compared to age- and sex-matched controls), linear (sequences of five or more cells inside the basement membrane of the gastric gland), micronodular (clusters of five or more endocrine cells up to 150 mm in diameter), and adenomatoid (collection of five or more micronodules adherent to one another but with interposition of basement membranes and thin strands of lamina propria). Dysplasia is characterized by 150 to 500 mm lesions formed by moderately atypical endocrine cells and defined as

t0010

enlarged micronodules (nodules of 150 mm), adenomatous micronodules (collections of at least five micronodules), fused micronodules (disappearance of the intervening basal membrane between adjacent micronodules), and microinfiltrative lesions (microinfiltration of the lamina propria by endocrine cells filling the space in between glands). G cell hyperplasia is defined as an increase in gastrin cell number (above 140) when counted per linear millimeter of mucosa in 5 mm thick histological sections. It may be associated with reduced somatostatin (D) cell count, resulting in an elevated G/D cell ratio. Long-standing hyperchlorhydria associated with duodenal G cell tumor may result in increased D cells of the antrum and a reduced G/D cell ratio.

Intestine

s0020

Increased numbers of somatostatin D cells are observed in the small intestine of patients with celiac disease. Hyperplasia of unspecified argyrophil endocrine cells was reported in chronic inflammatory bowel disease.

TUMORS General Endocrine tumors of the gut are found at any level of the gastrointestinal tract (Table II). According to epidemiological data, in Western countries endocrine tumors more commonly display an age-standardized rate of approximately 1/100,000 and more frequently

Table II Tumor Type, Distribution, and Cell Features of the Endocrine Tumors of the Gastrointestinal Tract Tumor type Well-differentiated endocrine/tumor/carcinoma

Preferred site

Main cell type

Hormonal products

Stomach, body/fundus

ECL

Histamine 5-HT/5-HTP

Duodenum, antrum, jejunum

G

Gastrin

Duodenum

D

Somatostatin

Appendix, ileum, jejunum, cecum

EC

Serotonina

Rectum, colon

L

Glicentin, PP, PYY

Gangliocytic paraganglioma

Duodenum

PP, D

PP, somatostatin

Poorly differentiated endocrine carcinoma

Stomach intestine

Proto-endocrine

None

Note. ECL, enterochromaffin-like cell; EC, enterochromaffin; 5-HT, 5-hydroxytryptamine (serotonin); 5-HTP, 5-hydroxytryptophan; PP, pancreatic polypeptide; PYY, peptide tyrosine tyrosine. a Substance P, tachykinins, and other peptides.

p0040

218 occur in the colorectum, followed by the small intestine and stomach, occur more frequently in females, and increase in incidence with age. The diagnosis of endocrine tumors is based on the identification of their characteristic morphology and of their antigenic asset, as defined by the expression of markers of endocrine differentiation (as described above) and hormones. According to the new World Health Organization classification, gut endocrine tumors are classified as follows: (1) well-differentiated endocrine tumors, including tumors with benign behavior and tumors with indefinite behavior at diagnosis; (2) well-differentiated endocrine carcinomas, low grade; and (3) poorly differentiated endocrine carcinomas, high grade.

Well-Differentiated Neoplasms

p0060

Well-differentiated tumors are characterized by bland histological and cytological features, low cytological atypia, low mitotic index, and diffuse expression of general markers of endocrine differentiation. Cellspecific markers are in general observed in tumor cell subpopulations. The definition of carcinoma is restricted to welldifferentiated endocrine neoplasms with overt malignancy at diagnosis, i.e., synchronous metastasis, and/ or deep wall invasion. In the absence of the above features, the behavior of well-differentiated tumors may be unpredictable, though independent of tumor cell typing and related to various clinicopathological variables. Predictors of poor outcome are as follows: tumor size (larger tumors usually are more aggressive); invasion of nearby tissue (appendix) or wall invasion beyond the submucosa; angioinvasion and invasion of perineural spaces; solid, nonorganoid structure; necrosis; overt cell atypia; more than two mitoses in 10 microscopic high-power fields (HPF); Ki67 proliferation index higher than 2% or of >100 in 10 HPF; reduction in or loss of chromogranin A immunoreactivity, argyrophilia, and/or nuclear p53 protein accumulation. The predictive value of most of these variables still needs confirmation from investigations on large series of endocrine tumors of the small and large intestine. Stomach ECL Cell Tumors By far the most frequent endocrine tumor of the stomach, ECL cell tumors display strong argyrophilia by silver impregnation techniques, immunoreactivity for chromogranin A and vesicular monoamine

GI Tract, General Pathology of Endocrine Growths

transporter 2, and no or focal immunoreactivity for ghrelin, serotonin, gastrin, and somatostatin. Three clinicopathologic subtypes are defined as follows: type I, associated with diffuse chronic atrophic gastritis of autoimmune or A type; type II, associated with hypertrophic gastropathy, usually in conjunction with MEN1 and Zollinger-Ellison syndrome (ZES); and type III or sporadic, not associated with a specific gastric pathology. Of the three subtypes, type I tumors are the most frequent (70–80% of published series), occur in elderly female patients, and, though multiple and multicentric, are small, limited to the mucosa and submucosa and demonstrate an excellent prognosis for survival. Type II tumors are rare (6% of published series), multiple, and multicentric; they usually display a good prognosis, though metastases may be present and rare cases with an aggressive course have been described. In contrast, the solitary type III tumors are less frequent (15% of published series), occur more frequently in male patients, are larger, and, in general, display a more aggressive behavior with frequent metastases and malignant course. Of note, type III tumors may associate with the so-called ‘‘atypical carcinoid syndrome’’ due to histamine and/or 5-hydroxytryptophan hypersecretion. Other Tumor Types Exceedingly rare gastrin G cell tumors have been reported in the antrum, similar to rare D cell tumors of both the oxyntic and antral mucosa. Such tumors usually are not associated with any hyperfunctional syndrome. Intestine Gastrin G Cell Tumors Gastrin G cell tumors occur in the duodenum or upper jejunum of male patients. When ‘‘functioning,’’ they can be defined as gastrinomas, are the cause of the ulcerogenic ZES, and frequently are associated with MEN1. Though small in size, G cell tumors, especially when functioning, frequently metastasize, especially to local lymph nodes. D Cell Tumors Somatostatin D cell tumors are rare tumors of the duodenum (preferentially at the ampulla of Vater) and upper jejunum in patients in the fifth decade of life. Often associating with type 1 neurofibromatosis and only rarely with diabetes and/or gallstones, they may be defined as functioning, though the complete somatostatinoma syndrome (diabetes mellitus, diarrhea, steatorrhea, hypo- or achlorhydria, anemia, and

p0070

219

GI Tract, General Pathology of Endocrine Growths

gallstones) has been described for pancreatic D cell tumors only. Intestinal D cell tumors are often malignant and may metastasize to both local lymph nodes and the liver. Ganglioneuromatous Paraganglioma Rare tumors composed of endocrine cells, mature ganglion cells, and Schwann-like spindle cells develop in the submucosa of the periampullary duodenal region in middle-aged patients. In general, ganglioneuromatous paragangliomas display benign behavior, though local lymph node metastases of the endocrine component have been reported in occasional cases.

p0095

Serotonin EC Cell Tumors Most serotonin-producing enterochromaffin (EC) cell tumors develop in the intestine, with decreasing frequency in the ileum, cecum, appendix, jejunum, duodenum, distal colon, and rectum. Intestinal EC cell tumors may display relatively aggressive behavior since often they are multiple (small intestine), are of large size (colon), and deeply infiltrate the intestinal wall, with frequent lymph node metastases. A typical carcinoid syndrome due to the unregulated release of serotonin and other active substances by tumor EC cells may occur in association with such tumors, depending on the establishment of liver metastases. In contrast, appendiceal EC cell tumors, despite the deep wall invasion often observed, in general run a benign course. Glicentin/PYY L Cell Tumors L cell tumors most frequently develop in the colon and rectum and rarely at other sites of the intestine and appendix. The majority of L cell tumors are small, do not associate with any hyperfunctional syndrome, and run a benign course.

Poorly Differentiated Endocrine Carcinomas Poorly differentiated endocrine carcinomas are aggressive carcinomas reported to occur at any site

in the gut, except the appendix, of patients in the seventh decade of life. Poorly differentiated endocrine carcinomas display a solid structure, abundant necrosis, small cell cytology, severe cellular atypia, a high mitotic index, no or faint chromogranin A expression, and diffuse expression of NSE and synaptophysin. Additional features include a high Ki67 index and p53 accumulation as observed with immunohistochemistry, whereas cell-specific endocrine markers are absent. Similar to undifferentiated carcinomas, poorly differentiated endocrine carcinomas run a very aggressive course.

See Also the Following Articles GI Hormones in Cancer (Cells)

.

GI Tract, General Anatomy

Further Reading Capella, C., Riva, C., Rindi, G., Sessa, F., Usellini, L., Chiaravalli, A., Carnevali, L., and Solcia, E. (1991). Histopathology, hormone products and clinico-pathologic profile of endocrine tumors of the upper small intestine: A study of 44 cases. Endocr. Pathol. 2, 92–110. Fiocca, R., Capella, C., Buffa, R., Fontana, R., Solcia, E., Hage, E., Chance, R. E., and Moody, A. J. (1980). Glucagon-, glicentin-, and pancreatic polypeptide-like immunoreactivities in rectal carcinoids and related colorectal cells. Am. J. Pathol. 100, 81–92. Moertel, C. G., Sauer, G., Dockerty, M. B., and Baggenstoss, A. H. (1961). Life history of the carcinoid tumor of the small intestine. Cancer 14, 901–912. Rindi, G., Azzoni, C., La Rosa, S., Klersy, C., Paolotti, D., Rappel, S., Stolte, M., Capella, C., Bordi, C., and Solcia, E. (1999). ECL cell tumor and poorly differentiated endocrine carcinoma of the stomach: Prognostic evaluation by pathological analysis. Gastroenterology 116, 532–542. Rindi, G., Luinetti, O., Cornaggia, M., Capella, C., and Solcia, E. (1993). Three subtypes of gastric argyrophil carcinoid and the gastric neuroendocrine carcinoma: A clinico-pathologic study. Gastroenterology 104, 994–1006. Solcia, E., Capella, C., Fiocca, R., Sessa, F., LaRosa, S., and Rindi, G. (1998). Disorders of the endocrine system. In ‘‘Pathology of the Gastrointestinal Tract’’ (S. C. Ming and H. Goldman, eds.), 2nd ed., pp. 295–322. Williams & Wilkins, Philadelphia, PA. Solcia, E., Klo¨ppel, G., and Sobin, L. H. (2000). ‘‘Histological Typing of Endocrine Tumors.’’ World Health Organization International Histological Classification of Endocrine Tumors, Springer-Verlag, New York.

158

GI-CGRP (Calcitonin Gene-Related Peptide)

1

10

20

30

37

Rat CGRP-α

H2N

S C N T A T C V T H R L A G L L S R S G G V V K D N F V P T N V G S E A F

CONH2

Rat CGRP-β

H2N

S C N T A T C V T H R L A N L L S R S G G V V K D N F V P T N V G S K A F

CONH2

Human CGRP-α

H2N

A C D T A T C V T H R L A G L L S R S G G V V K N N F V P T N V G S K A F

CONH2

Human CGRP-β

H2N

A C N T A T C V T H R L A G L L S R S G G M V K S N F V P T N V G S K A F

CONH2

Human amylin

H2N

K C N T A T C A T Q R L A N F L V H S S N N F G A I

CONH2

Human adrenomedullin

G C R F G T C T V Q K L A H Q I F S R L G Q F N N M S Q R Y NH2

L S S T N V G S N T Y

Y Q F T D K D K D N V A P R S K I

S P Q G Y CONH2

Figure 1 Comparison of the amino acid sequences of human and rat CGRP-a and CGRP-b, human amylin, and human adrenomedullin.

CGRP RECEPTORS The biological effects of CGRP are brought about by interaction with specific membrane receptors, two subtypes of which, the CGRP1 and CGRP2 receptors, have been proposed to exist on the basis of bioassay observations. Characteristic of the CGRP1 receptors is their sensitivity to the antagonistic effect of the human CGRP8–37 fragment. The molecular identification of functional CGRP receptors turned out to be very difficult because they are assembled from three different proteins (Fig. 2): the calcitonin receptor-like receptor (CRLR), the receptor-associated membrane protein 1 (RAMP-1), and the receptor component protein (RCP). Although CRLR is the CGRP-recognizing protein, the CGRP receptor becomes functional only if CRLR is associated with RAMP-1. This chaperone protein is important for the intracellular translocation of CRLR and its insertion into the plasma membrane and for conferring a CGRP1 receptor-like binding profile on CRLR. RCP seems to be required for efficient coupling of the receptor to the G protein– adenylate cyclase signaling machinery (Fig. 2). CRLR can associate not only with RAMP-1 and RCP to produce a CGRP1 receptor, but also with RAMP-2 or RAMP-3, two other chaperone proteins sharing approximately 30% sequence homology with RAMP-1. These RAMPs dictate the pharmacological profile of the receptor complex inasmuch as CRLR associated with RAMP-2 behaves as an adrenomedullin receptor and human calcitonin receptor isotype 2 associated with RAMP-1 or RAMP-3 behaves as an amylin receptor.

EXPRESSION AND RELEASE OF CGRP IN THE GI TRACT The principal sources of CGRP in the digestive system are extrinsic primary afferent nerve fibers and

intrinsic enteric neurons. Although the CGRP that is expressed in primary afferents of the rat is CGRP-a, the only form of CGRP in enteric neurons is CGRPb. The relative contribution of extrinsic afferent and intrinsic neurons to the overall CGRP content of the GI tract varies among different gut regions and different mammalian species. Most of the CGRP found in the esophagus and stomach of small rodents is derived from primary afferents, whereas the small intestine and large intestine contain a sizable number of CGRP-positive enteric neurons that issue specific projections to other enteric ganglia, muscle, and mucosa. The majority of the CGRP-expressing extrinsic afferent neurons in the rodent gut originate from cell bodies in the dorsal root ganglia and reach the gut via sympathetic and sacral parasympathetic nerves. Within the wall of the GI tract, they innervate primarily arteries and arterioles but also project to the mucosa, enteric nerve plexuses, and muscle layers. CGRP-positive vagal afferents originating from the nodose ganglion supply the esophagus and proximal part of the stomach but make a relatively small contribution to the content of CGRP in the gastric corpus, antrum, and intestine. As is expected for substances with a vesicular localization, CGRP is released from GI neurons in a calcium-dependent manner when these cells are stimulated. This is, for instance, the case when the mucosa of the stomach and duodenum is exposed to excess acid, which releases CGRP from primary afferent nerve fibers. Peptide release from spinal afferents in the gut can also be elicited by the vanilloid capsaicin because these nerve cells, but not enteric neurons, express vanilloid receptors of type 1. With the use of capsaicin, it has furthermore been found that CGRP in the general circulation represents primarily an overspill of peptide released from peri- and paravascular afferent nerve fibers.

159

GI-CGRP (Calcitonin Gene-Related Peptide)

Figure 2 Diagram of the CGRP receptor complex consisting of the calcitonin receptor-like receptor (CRLR), the receptorassociated membrane protein 1 (RAMP-1), and the receptor component protein (RCP). Association of CRLR with RAMP-1 forms a high-affinity CGRP1 receptor, which, through RCP, is coupled to the G protein Gs and thereby can activate adenylate cyclase (AC) to produce cAMP.

Apart from neurons, CGRP is also found in endocrine cells of the human GI mucosa and rat pancreas and in blood-derived or resident immune cells within the lamina propria of the rat gastric mucosa. However, the quantitative and functional significance of these sources is still not very well understood.

EFFECTS, PHYSIOLOGICAL ROLES, AND PATHOLOGICAL IMPLICATIONS OF CGRP IN THE GI TRACT Motor Activity The most prominent motor action of CGRP in the active gut is muscle relaxation via CGRP1 receptors, which leads to retardation of gastric emptying and attenuation of motor activity throughout the digestive tract. However, CGRP is also able to excite cholinergic motor pathways, which is in keeping with the peptide’s ability to depolarize intrinsic sensory neurons of the myenteric plexus, to enhance the release of acetylcholine from enteric neurons, and to cause contraction of the resting muscle. There is still scarce information as to whether CGRP released from intrinsic or extrinsic neurons of the gut plays a physiological role in the neural control of GI motility. The claim that CGRP released from sensory neurons contributes to distension-induced peristalsis is not universally accepted. There is, however, evidence that CGRP released from extrinsic afferent nerve fibers contributes to the pathological shutdown of GI motility in postoperative and peritonitis-induced ileus. This is consistent with the

observation that CGRP acting via CGRP1 receptors contributes to the inhibition of intestinal peristalsis that ensues after sensory neuron stimulation.

Secretory Processes Electrolyte and fluid secretion in the small intestine of the dog and in the colon of the guinea pig and rat is stimulated by CGRP. Whereas the secretory effect of CGRP in the rat colon arises from a direct action on enterocytes, as is the case with human epithelial cell lines, CGRP’s secretory action in the guinea pig colon is mediated by enteric neurons. Although a physiological role for CGRP in the control of intestinal secretory activity has not yet been determined, it seems as if the peptide contributes to the pathological fluid secretion that in the rat ileum is evoked by Clostridium difficile toxin A. The secretagogue-evoked secretion of enzyme, bicarbonate, and fluid from the pancreas of the dog and rat in vivo is blunted by CGRP through an action involving somatostatin, whereas amylase secretion from isolated acini of the rat and guinea pig pancreas is enhanced by the peptide. There is good evidence that, in the stomach, CGRP contributes to the homeostatic regulation of endocrine and exocrine secretory processes. Thus, CGRP potently depresses basal and secretagogueevoked output of acid and pepsin in the stomach of human, dog, rabbit, and rat, an action that is brought about by CGRP1 receptors, depends on somatostatin as an essential mediator, and goes along with attenuation of the release of acetylcholine, gastrin, and

160

GI-CGRP (Calcitonin Gene-Related Peptide)

Afferent nerve fiber Mucosa

D G CGRP

CGRP1-R

ECL

H+

Parietal cell Reduced H+ secretion

ACh

H+

Low pH

High pH

H+

Lumen

Figure 3 Diagram illustrating the role of CGRP-releasing afferent nerve fibers in the feedback control of gastric acid secretion in the rat stomach. When the acidity in the lumen rises, afferent nerve fibers release CGRP, which, via activation of CGRP1 receptors (CGRP1-R), stimulates the release of somatostatin and inhibits the release of gastrin, histamine, and acetylcholine. D, endocrine D cells releasing somatostatin; G, endocrine G cells releasing gastrin; ECL, enterochromaffin-like cells releasing histamine; ACh, neurons releasing acetylcholine. histamine (Fig. 3). These effects are physiologically relevant, given that the CGRP1 receptor antagonist CGRP8–37 augments basal and stimulated acid secretion and acid accumulation in the gastric lumen releases CGRP from sensory nerve fibers. Through its effects on the release of somatostatin, gastrin, histamine, and acetylcholine, CGRP halts further secretion of acid and thus mediates feedback inhibition of gastric acid output (Fig. 3).

Vascular Functions The arteries and submucosal arterioles of the GI tract receive the densest innervation by extrinsic afferents containing CGRP. This finding, the expression of CGRP1 receptors on the endothelium and smooth muscle of arteries and arterioles, and the peptide’s vasodilator activity point to a vasoregulatory function of CGRP. Indeed, nonadrenergic noncholinergic dilation of the rat superior mesenteric artery is mediated by capsaicin-sensitive afferent nerve fibers releasing CGRP. In contrast, the physiological significance of CGRP in the microcirculation of the small and large intestine is little known. Although CGRP dilates submucosal arterioles in the guinea pig ileum, it fails to alter mucosal blood flow in the rat small and large intestine. The situation is different in the rat stomach, where CGRP is highly potent in causing a CGRP1

receptor-mediated dilation of submucosal arterioles, but not venules. Vasodilation induced by low doses of CGRP is mediated through a mechanism that involves nitric oxide, whereas high doses of the peptide increase blood flow independent of nitric oxide. Although CGRP does not seem to regulate gastric blood flow in physiological circumstances, there is evidence that CGRP comes into play under pathological conditions. This role is best exemplified by the hyperemic response that ensues when the gastric mucosal barrier is disrupted by ethanol or bile salts so that acid can enter the tissue and damage the gastric mucosa. The CGRP-mediated rise in blood flow in response to acid influx serves a protective role in the gastric mucosa as it helps to neutralize and wash away intruding acid and delivers bicarbonate and other factors to defend and repair the mucosa. Another example relates to the C. difficile toxin Aevoked inflammation in the rat ileum, which involves CGRP as a pro-inflammatory mediator.

Mucosal Homeostasis There are several lines of evidence to indicate that GI mucosal integrity and repair are under the control of extrinsic afferent neurons releasing CGRP. First, CGRP protects the mucosa in a number of experimental models of gastric injury and colonic inflammation. The action of CGRP to reduce

161

GI-CGRP (Calcitonin Gene-Related Peptide)

ethanol-induced damage in the gastric mucosa is mediated by CGRP1 receptors and involves nitric oxide. Second, CGRP mediates the gastroprotective effect of a number of factors and drugs that stimulate capsaicin-sensitive afferent nerve fibers. Thus, blockade of CGRP1 receptors with CGRP8–37 prevents the ability of intragastric capsaicin to attenuate experimentally imposed injury as does immunoneutralization of CGRP with polyclonal and monoclonal antibodies to the peptide. Third, CGRP8–37 and active immunization of rats against CGRP exacerbate experimental injury in the stomach. In summary, CGRP released from sensory nerve fibers strengthens gastric mucosal defense and facilitates repair of the wounded mucosa via vasodilation, hyperemia-dependent processes, such as appropriate delivery of bicarbonate, and hyperemia-independent mechanisms, such as secretion of mucus. Complementary evidence for such a homeostatic role for CGRP-releasing nerve fibers in the GI mucosa comes from the observation that sensory neuropathies weaken the resistance of the tissue to injury. This applies not only to the stomach, but also to the esophagus, small intestine, and colon, where experimentally induced inflammation and damage are aggravated.

Gastrointestinal Sensitivity and Nociception Since CGRP is a transmitter of nociceptive afferent neurons innervating the gut, it does not come as a surprise that CGRP can mediate GI pain and inflammatory hyperalgesia. Intraperitoneal administration of exogenous CGRP or acetic acid-induced release of endogenous CGRP in the rat peritoneum triggers abdominal muscle contractions, a reaction that is indicative of pain. Of particular importance is the finding that CGRP8–37 prevents inflammationinduced hypersensitivity to colonic distension. Since, in this respect, intrathecal CGRP8–37 is more potent than intravenous CGRP8–37, the site of CGRP-mediated hyperalgesia is primarily in the spinal cord.

SUMMARY CGRP is a transmitter candidate of intrinsic enteric neurons and extrinsic afferent nerve fibers in the GI tract. As such, this neuropeptide seems to be involved in the neural regulation of GI functions. Its major

actions include depression of GI motility, rise of gastric blood flow, inhibition of gastric acid secretion, enforcement of GI mucosal resistance, and mediation of inflammatory hyperalgesia. Furthermore, there is evidence that a disturbance of the GI CGRP system may contribute to a number of GI disorders and that a correction of these perturbations may be of therapeutic potential.

Acknowledgment Evelin Painsipp is greatly appreciated for drawing the figures.

See Also the Following Articles Calcitonin, Overview . CCK (Cholecystokinin) . Gastrin . Gastrin-Releasing Peptide . GIP (Gastric Inhibitory Polypeptide)

Further Reading Foord, S. M., and Marshall, F. H. (1999). RAMPs: Accessory proteins for seven transmembrane domain receptors. Trends Pharmacol. Sci. 20, 184–187. Gschossmann, J. M., Coutinho, S. V., Miller, J. C., Huebel, K., Naliboff, B., Wong, H. C., Walsh, J. H., and Mayer, E. A. (2001). Involvement of spinal calcitonin gene-related peptide in the development of acute visceral hyperalgesia in the rat. Neurogastroenterol. Motil. 13, 229–236. Holzer, P. (1994). Calcitonin gene-related peptide. In ‘‘Gut Peptides: Biochemistry and Physiology’’ ( J. H. Walsh and G. J. Dockray, eds.), pp. 493–523. Raven Press, New York. Holzer, P. (1998). Neural emergency system in the stomach. Gastroenterology 114, 823–839. Holzer, P. (2000). Calcitonin gene-related peptide in gastrointestinal homeostasis. In ‘‘The CGRP Family: Calcitonin Gene-Related Peptide (CGRP), Amylin, and Adrenomedullin’’ (D. Poyner, I. Marshall, and S. D. Brain, eds.), pp. 125–139. Landes Bioscience, Georgetown, TX. Ichikawa, T., Ishihara, K., Kusakabe, T., Hiruma, H., Kawakami, T., and Hotta, K. (2000). CGRP modulates mucin synthesis in surface mucus cells of rat gastric oxyntic mucosa. Am. J. Physiol. 279, G82–G89. Juaneda, C., Dumont, Y., and Quirion, R. (2000). The molecular pharmacology of CGRP and related peptide receptor subtypes. Trends Neurosci. 21, 432–438. Rosenfeld, M. G., Emeson, R. B., Yeakley, J. M., Merillat, N., Hedjran, F., Lenz, J., and Delsert, C. (1992). Calcitonin generelated peptide: A neuropeptide generated as a consequence of tissue-specific, developmentally regulated alternative RNA processing events. Ann. N. Y. Acad. Sci. 657, 1–17. Tache´, Y. (1992). Inhibition of gastric acid secretion and ulcers by calcitonin gene-related peptide. Ann. N. Y. Acad. Sci. 657, 240–247. Wimalawansa, S. J. (1996). Calcitonin gene-related peptide and its receptors: Molecular genetics, physiology, pathophysiology, and therapeutic potentials. Endocr. Rev. 17, 533–585.

163

Gigantism: Excess of Growth Hormone

Table I Causes of Gigantism

Table II

Pituitary origin Gsa gene mutation, with or without McCune-Albright syndrome

Somatotroph adenomas Sparsely granulated

Pituitary adenoma due to other molecular causes Excess GHRH Eutopic—Hypothalamic dysfunction and tumors Ectopic Somatostatin deficiency

Histological Classification of Adenoma

Densely granulated Plurihormonal tumors of the acidophil cell line Mixed somatotroph–lactotroph adenomas Mammosomatotroph adenomas Acidophil stem cell adenomas Miscellaneous tumors Plurihormonal adenomas

Pituitary GH Excess Caused by Activation of the Gsa Subunit of the G Protein Receptor with or without McCune-Albright Syndrome p0030

The G protein-coupled receptors are the largest family of membrane receptors. Peptides (vasopressin, glycoproteins, adrenocorticotrophic hormone [ACTH], GHRH, melanocyte-stimulating hormone [MSH]), neurotransmitters, and prostaglandins can bind to this receptor family. When a stimulatory ligand binds to the receptor, the a subunit of the G protein (Gsa) is activated by replacement of GDP by GTP. This subsequently leads to activation of adenylate cyclase and cyclic AMP formation. It finally results in phosphorylation of substrates that control metabolism, gene transcription, secretion, and cell proliferation. Normally, intrinsic GTPase stops activation of Gsa. A mutation in the gene encoding the a subunit of the stimulatory G protein receptor results in constitutive activation of the protein, by inhibiting the intrinsic GTPase activity. This has a tumorigenic effect. Thirty to 50% of pituitary adenomas in gigantism are caused by this mechanism. McCune-Albright syndrome is a rare disorder, classically defined by fibrous dysplasia, cafe´ au lait spots, precocious puberty, and other hyperfunctional endocrinopathies. In all patients with the McCune-Albright syndrome, a mutation in the gene for the a subunit is found. It is assumed that the timing of the mutation in development determines which tissues are affected and in this way contributes to the heterogeneity of the clinical presentation. Approximately 20% of the cases of gigantism are associated with McCune-Albright syndrome. GH-producing tumors in these patients are the result of a mutation in the gene encoding the a subunit of the G protein receptor of somatotroph cells, leading to constitutive activation of the Gsa protein, resulting in hyperplasia and adenoma with GH overproduction.

Pituitary GH Excess Caused by an Adenoma by Other Molecular Causes Pituitary adenomas are benign tumors consisting of adenohypophysial cells. They are often slowly growing, expansive tumors confined to the sella turcica, which are usually detected because of the clinical effects of the oversecreted hormones. Adenomas can be classified according to their size, microadenomas (10 mm), but a morphological classification (Table II) is more useful. In gigantism, mammosomatotroph adenomas are the most frequent histological type. The mammosomatotroph cells, producing prolactin and GH, observed in these adenomas resemble the mammosomatotrophs identified in the fetal pituitary at 9.5 weeks gestation. It is thought that pituitary adenomas are derived from an intrinsic pituitary cell defect, leading to clonal expansion of a single transformed cell. The exact mechanisms by which these cells transform into tumorous cells is unknown. Inactivating mutations of tumor suppressor genes are associated with pituitary tumorigenesis, as shown in multiple endocrine neoplasia (MEN-I) syndrome, which is characterized by hyperfunction or tumor formation in the parathyroids, anterior pituitary, and pancreatic cells and rarely in the thyroid and adrenal cells. Pei and Melmed cloned a novel pituitary tumor-transforming gene. High-level expression of this gene in GH-secreting adenomas suggests that it plays a role in pituitary tumorigenesis.

Excess of GHRH Eutopic GHRH Excess Secondary GH overproduction due to hypothalamic GHRH excess is an important cause of gigantism. GHRH excess usually results in mammosomatotroph hyperplasia and rarely in adenoma formation.

p0045

164 In a case report of congenital gigantism due to GH and prolactin-producing cell hyperplasia, high serum levels of GHRH were found. There was no evidence of a hypothalamic GHRH-producing tumor, suggesting that a congenital hypothalamic regulatory defect resulting in early GHRH exposure leads to mammosomatotroph hyperplasia. Ectopic GHRH Excess Ectopic GHRH-producing tumors are a well-known cause of acromegaly, but very rarely cause gigantism. Thus far, only two cases have been described. In one patient, a 15-year-old girl, the GHRH-producing primary tumor was located in the jejunum and resulted, together with GHRH-producing liver metastases, in severe GH hypersecretion causing gigantism. The other case is an 18-year-old boy with GHRH-like activity in a metastatic carcinoid tumor in the foregut. Another cause of intracranial GHRH excess occurs in the setting of neural tumors, such as gangliocytoma or neurocytoma, within or in close proximity to the sella.

GH Excess by Somatostatin Deficiency A few cases have been documented in which GH excess is found in children with neurofibromatosis and optic gliomas or astrocytomas. It has been suggested that infiltration of the glioma in the medial temporal lobe causes a reduction in somatostatin, leading to increased GH levels and interference with the normal pulsatility of GH release.

DIAGNOSIS In all patients with an increased growth velocity and progressive upward deviation from the population reference curves, GH excess should be considered. In young children, body proportions are normal and usually the increased growth is the only sign of gigantism. In adolescents, the typical features of acromegaly, such as thickening of the skin, enlargement of the lower jaw, hands, and feet, coarsening of facial features, and excessive body sweating, may be present. Later, a eunochoid habitus can develop because of incomplete or absent puberty. Depending on the type and expansion of the adenoma, secretion of other pituitary hormones can be affected. As adenomas often secrete GH as well as prolactin, hyperprolactinemia may occur, manifesting as galactorrhea. Gonadotropin deficiency leads to hypogonadotropic hypogonadism and incomplete or absent puberty.

Gigantism: Excess of Growth Hormone

Less often, ACTH or TSH deficiency occurs. Because of compression of the optic chiasm, impaired vision and visual field abnormalities can be presenting symptoms. Serum GH levels are elevated in gigantism. Whereas in overnight GH profiles of healthy individuals, baseline levels do not exceed 0.5 ng/ml, in giants these levels are seldom less than 5 ng/ml. Serum levels of IGF-I and IGF-binding protein 3 (BP3) are elevated compared with age and gender references. However, interpretation of IGF-I levels is difficult during puberty because of the physiological IGF-I increase at that time. Normally, oral glucose ingestion (1.75 g per kilogram of body weight) suppresses GH levels to less than 1 ng/ml, whereas in patients with GH excess, GH levels are not suppressed and may even show a paradoxical increase. However, in children and adolescents, the oral glucose tolerance test is less useful for diagnosing GH excess, as absent suppression is also common in healthy tall children and adolescents. Similarly, the presence of paradoxical GH response to thyrotropin-releasing hormone and luteinizing hormone-releasing hormone is of little use in this age group. In the diagnostic workup, the other pituitary hormones should obviously also be assessed. Prolactin is elevated in approximately half the cases and may be associated with galactorrhea and contribute to hypogonadism. Gonadotropin secretion is often decreased because of compression of gonadotropin-secreting cells by the hyperplastic or adenomatous somatotroph cells. The thyroid-stimulating hormone and ACTH axes should be investigated, although these are mostly not affected. GHRH measurement can be helpful in differentiating GHRH excess from primary GH hypersecretion. As insulin resistance can occur, the insulin response to glucose must be measured. Bone age should be assessed, but is usually not advanced, or only little advanced, during childhood. A skull X ray may reveal enlargement of the sella turcica. When GH excess is confirmed, neuroradiological studies including magnetic resonance imaging for tumor localization and extension are essential. For illustrative purposes, two cases with gigantism are shown in Figs. 1 and 2 . Case 1 presented with tall stature at 8.5 years of age. He had serum GH levels fluctuating between 100 and 150 ng/ml, very elevated serum IGF-I levels, and a GH-producing adenoma with suprasellar extension. After transsphenoidal resection of the adenoma, a small tumor remnant remained. Bromocriptine treatment was sufficient for

p0075

p0095

165

Gigantism: Excess of Growth Hormone

DIFFERENTIAL DIAGNOSIS

Figure 1 Growth curve of a patient with a growth hormoneproducing adenoma.

a normalization of plasma GH levels and final height was limited by supraphysiologic testosterone treatment. His growth pattern is shown in Fig 1. Case 2 presented with pseudo-precocious puberty, a pigmented macula with irregular margins, resembling the coast of Maine, typical for McCune-Albright syndrome, and polyostotic fibrous dysplasia. Initially, he was treated with cyproterone acetate and ketoconazole for his pseudo-precocious puberty, but soon it became clear that he also had GH excess, with mean GH levels between 10 and 15 ng/ml. From 12.4 years of age, he was subsequently treated with bromocriptine, octreotide, and cabergoline. Bisphosphonates were prescribed to restrict the damage of the fibrodysplastic lesions. His growth curve is shown in Fig 2.

Figure 2 Growth curve of a patient with McCune-Albright syndrome.

Tall stature in childhood and adolescence can be classified in various ways. The authors prefer the classification summarized in Table III, in which three main groups are distinguished: constitutional tall stature, primary growth disorders (with a defect presumably in the growth plate), and secondary growth disorders (abnormalities in the milieu interieur). By far the most frequent cause of tall stature is constitutional (familial) tall stature, which is a reflection of the fact that 50–90% of height variation is accounted for by genetic factors. In most cases, height standard deviation scores (SDS) lie within a range of 1.3 SDS above or below the target height SDS (the sex- and secular trend-corrected midparent height). Measurements of parental height and calculating target height are therefore of critical importance. Birth length is usually more than þ0.7 SDS (75th percentile). In the first 3–4 years, height velocity is accelerated, followed by a growth pattern parallel to, but above, the þ1.9 SDS (97th percentile). No abnormalities are present on physical examination. Primary growth disorders include a number of genetic syndromes. There are a number of sex chromosome-related disorders, of which Klinefelter syndrome (47, XXY) is most frequent. Usually, boys with this syndrome are not exceptionally tall in childhood, but adolescents and adults can be tall and often have eunuchoid body proportions. It is assumed that the additional chromosomal material and inadequate pubertal development account for the tall stature in this syndrome. Sotos syndrome, Weaver syndrome, and MarshallSmith syndrome are characterized by increased birth length and head circumference, increased growth velocity during the first year of life, tall stature during childhood and adolescence, advanced bone age, and moderately increased final height. All these syndromes have psychomotor retardation and their own set of typical craniofacial features. The phenotype of fragile X syndrome is very similar to that of Sotos syndrome. In Beckwith-Wiedemann syndrome, birth weight and length are increased and infants have a large protruding tongue. Final height is often within the normal range due to advanced bone maturation. Marfan syndrome is an autosomal dominant inherited disorder, in which tall stature is the most prominent feature. Associated symptoms are arachnodactyly, joint hyperlaxity, cardiovascular anomalies, and lens subluxation. Homocystinuria is an autosomal recessive disorder caused by an enzyme deficiency. Its

166 Table III Causes of Tall Stature Constitutional (familial) Primary growth disorders Sex chromosome-related disorders: Klinefelter syndrome (XXY), XXX syndrome, XYY syndrome Fragile X syndrome Syndromes characterized by advanced growth and maturation: Sotos syndrome, Weaver Syndrome, Marshall-Smith syndrome Marfan syndrome

Gigantism: Excess of Growth Hormone

cardiovascular, neuromuscular, and pulmonary problems as in acromegaly. The risk of colorectal tumors and other malignancies seems to be correlated with the degree of GH control. The diabetogenic effects of GH cause insulin resistance in acromegaly. In young children with GH excess, diabetes is rare because of their increased islet reserve. Periods of stress, however, can exhaust these reserves and result in temporary diabetes and even diabetic ketoacidosis.

Marfan-like phenotype: MEN type IIB, homocystinuria Secondary growth disorders GH excess Hyperthyroidism Precocious puberty True (central) precocious puberty Pseudo-precocious puberty: Leydig cell hyperplasia (testotoxicosis), McCune-Albright syndrome, congenital adrenal hyperplasia, gonadal tumors Impaired estrogen effect (estrogen receptor defect or aromatase deficiency)

clinical features are similar to those of Marfan syndrome, with the exception of mental retardation, which is always present in homocystinuria. MENIIB syndrome mimics the signs of Marfan syndrome, but has additional signs, such as nodules at the tongue and lips. Of the secondary growth disorders, GH excess is most important. Hyperthyroidism can cause an increased growth velocity and bone maturation, but in childhood tall stature is rare and final height is normal. In infants, hyperinsulinism, due to nesidioblastosis or as a result of diabetes mellitus of the mother, can result in overgrowth in terms of weight and length, but this usually normalizes shortly afterward. Pseudo-precocious puberty temporarily leads to increased growth and thus tall stature in childhood, but due to premature closure of the epiphyses final height is usually in the lower normal range. It has been observed that an absence of estrogen effect (in one case occurring due to a mutation in the estrogen receptor and in two cases occurring due to aromatase deficiency) results in normal growth in childhood, but brings about an absence of epiphyseal closure and thus failure to stop growing, leading to very tall stature in adulthood.

COMPLICATIONS In addition to the psychological and social problems that patients with extremely tall stature encounter, complications of gigantism include increased risk of

TREATMENT If an adenoma is present, transsphenoidal surgery is the treatment of choice. Irradiation can cause permanent damage to the pituitary and the central nervous system and is therefore not desirable in children. Primary or adjuvant pharmacotherapy is used pre- and postoperatively and in cases of pituitary hyperplasia. Dopamine agonists (bromocriptine and cabergoline) release GH in normal individuals but paradoxically suppress GH hypersecretion in some patients with gigantism. The reaction to this therapy depends on the expression of dopamine receptor type D2 on the tumor cells. Often dopamine agonists alone fail to achieve complete biochemical control. In cases with pituitary hyperplasia or in cases with an insufficient response to transsphenoidal surgery, somatostatin analogues (octreotide) are often used. Somatostatin receptor subtypes 2 and 5 mediate the inhibitory effect of somatostatin. Complete biochemical control was reported in a case with an acidophilic stem cell adenoma using a combination of a long-acting dopamine agonist and a long-acting somatostatin analogue. A GH receptor antagonist has shown to be effective in acromegaly. It is a mutated human GH molecule that binds to the GH receptor, but does not lead to signal transduction. Thus far, it has not been used in gigantism, but theoretically it should be as effective as in acromegaly. A competitive GHRH antagonist, blocking the GHRH receptor, has proven to be effective in the ectopic GHRH syndrome.

Acknowledgments The authors are grateful to Dr. M. Jansen and Dr. A. M. Pereira Arias for providing information on both patients.

See Also the Following Articles Acromegaly, Diagnosis of . Genetic Testing for Pituitary Disease . Growth Hormone (GH) . McCune-Albright Syndrome . Pituitary Tumors, Molecular Pathogenesis . Postnatal Non-Endocrine Overgrowth . Postnatal Normal

Gigantism: Excess of Growth Hormone

Growth and Its Endocrine Regulation . Somatostatin, Evolution of

Further Reading Daughaday, W. H. (1992). Pituitary gigantism. Endocrinol. Metab. Clin. N. Am. 21, 633–647. Drake, A. J., Lowis, S. P., Bouffet, E., and Crowne, E. C. (2000). Growth hormone hypersecretion in a girl with neurofibromatosis type 1 and an optic nerve glioma: Resolution following chemotherapy. Horm. Res. 53, 305–308. Eugster, E. A., and Pescovitz, O. H. (1999). Gigantism. J. Clin. Endocrinol. Metab. 84, 4379–4384. Holl, R. W., Bucher, P., Sorgo, W., Heinze, E., Homoki, U., and Debatin, K. M. (1999). Suppression of growth hormone by oral glucose in the evaluation of tall stature. Horm. Res. 51, 20–24. Kahn, C. R., Smith, R. J., and Chin, W. W. (1998). Mechanism of action of hormones that act at the cell surface. In ‘‘Williams’ Textbook of Endocrinology’’ ( J. D. Wilson, D. W. Foster,

167 H. M. Kronenberg, and P. R. Larsen, eds.), 9th ed., pp. 95–143. Saunders, Philadelphia, PA. Maheshwari, H. G., Prezant, T. R., Herman-Bonert, V., Shahinian, H., Kovacs, K., and Melmed, S. (2000). Long-acting peptidomimergic control of gigantism caused by pituitary acidophilic stem cell adenoma. J. Clin. Endocrinol. Metab. 85, 3409–3416. Pei, L., and Melmed, S. (1997). Isolation and characterization of a pituitary tumor-specific transforming gene. Mol. Endocrinol. 11, 433–441. Sano, T., Asa, S. L., and Kovacs, K. (1988). Growth hormonereleasing hormone-producing tumors: Clinical, biochemical, and morphological manifestations. Endocr. Rev. 9, 357–373. Shimon, I., and Melmed, S. (1997). Pituitary tumor pathogenesis. J. Clin. Endocrinol. Metab. 82, 1675–1681. Zimmerman, D., Young, W. F., Ebersold, M. J., Scheithauer, B. W., Kovacs, K., Horvath, E., Whitaker, M. D., Eberhardt, N. L., Downs, T. R., and Frohman, L. A. (1993). Congenital gigantism due to growth hormone-releasing hormone excess and pituitary hyperplasia with adenomatous transformation. J. Clin. Endocrinol. Metab. 76, 216–222.

203

GIP (Gastric Inhibitory Polypeptide)

Glucose stimulation of GIP release appears to be direct as it occurs in vitro from a GIP-secreting cell line. The GIP response to fat is slower to develop, greater in magnitude, and more prolonged. The delayed GIP response to fat may be in part due to delayed gastric emptying caused by fat. It should be noted that GIP released by fat would not be insulinotropic unless circulating glucose values were also elevated. In addition to carbohydrates and fat, there are reports of GIP release by ingested amino acids as well as peptone and protein hydrolysate.

GIP GENE STRUCTURE AND STRUCTURE–ACTIVITY RELATIONSHIPS The secreted form of GIP is a 42-amino-acid peptide derived by proteolytic processing of precursors that are 144 and 153 amino acids in length in rat and human, respectively. The gene coding for the human GIP precursor spans approximately 10 kb, whereas that for the rat spans only 8.2 kb. Both consist of six exons, with exons 3 and 4 coding for GIP1–42 (Fig. 1A). The promoter region of the human GIP gene contains potential binding sites for a number of transcription factors including Sp1, activator protein-1 (AP-1), and AP-2. The human GIP gene has been assigned to chromosome 17q21.3–q22. A number of studies have been directed at determining the regions of the GIP molecule that are important for biological activity. Experimental evidence indicates that there are four dissociable domains in GIP1–42. Residues 6–30 in the GIP molecule constitutes a high-affinity binding domain. Truncation of 12 amino acids from the carboxyl terminus of GIP1–42 was shown to result in a peptide with intact insulinotropic activity, but without the ability to stimulate gastric somatostatin secretion, indicating the existence of two bioactive domains in the native molecule. The insulinotropic domain of GIP has been further localized to residues 19–30 and Hinke and co-workers presented evidence suggesting that a third bioactive domain of GIP resides in residues 1–14.

Figure 1 (A) Diagrammatic representation of the GIP gene, propeptide, and the peptides produced by processing. (B) Primary structure of human GIP showing the cleavage site for dipeptidyl peptidase IV.

cloned and it has been shown that the amino acid sequences of the rat, human, and hamster receptors share 40–47% identity with the glucagon-like peptide-1 (GLP-1) and glucagon receptors. Splice variants of the GIP receptor exist in some tissues and it is possible that there is variability in hormone–receptor interaction. Through the study of mutant and chimeric receptors expressed in different cell lines, it has been shown that the N-terminal domain and first extracellular loop are important for GIP binding and that intracellular (IC) loops are involved in G protein coupling, whereas threonine and serine residues in IC3 and the C-terminal tail are important for receptor internalization. Localization studies have confirmed the broad distribution of the GIP receptor in pancreatic islets, fat, stomach, brain, heart, lung, endothelium, adrenal cortex, and bone, in agreement with GIP’s broad pleiotropic actions.

THE GIP RECEPTOR The pancreatic islet GIP receptor is a glycoprotein with an estimated molecular weight of 59 kDa that belongs to the class B family of heptahelical G proteincoupled receptors. GIP receptor cDNAs have been

The GIP Receptor Gene The human receptor gene contains 14 exons and 12 introns, with a protein coding region of 12.5 kb, whereas that of the rat receptor gene spans 10.2 kb

204

GIP (Gastric Inhibitory Polypeptide)

and contains 13 exons. The promoter region of the human gene has not yet been characterized, but 50 -flanking sequences of the rat gene contain several transcription factor-binding motifs, including a cyclic AMP (cAMP)-response element, an octamerbinding site, three Sp1 sites, and an initiator element. One potential Sp1-binding site has been shown to be important for transcriptional activity and distal negative regulatory sequences have been suggested to control cell-specific expression. Missense mutations in the receptor gene have been identified in Japanese and Danish populations. Cells expressing one of the mutated receptors (Gly198/Cys) exhibited elevated EC50 values for cAMP production. However, allelic frequency studies have shown no association of the mutations with type 2 diabetes, although in the Danish group those homozygous for a Glu354/ Gln variant had decreased fasting serum C-peptide levels, suggesting that GIP normally regulates beta cell secretion even in the fasting state. Lynn and co-workers showed that GIP receptor mRNA and protein levels are dramatically reduced in the Vancouver Diabetic Zucker fatty rat, thus partially explaining the resistance to GIP found in these animals.

GIP-Activated Signal Transduction Pathways The endocrine pancreas is the only target organ for which there is information on the mode of action of GIP. Glucose is the main nutrient stimulator of insulin secretion (Fig. 2). Glucose enters the pancreatic beta cell and its metabolism results in an increase in the ratio of cytosolic ATP:ADP. The resulting closure of ATP-dependent Kþ channels leads to membrane depolarization and opening of voltage-dependent Ca2þ channels. The increase in cytosolic Ca2þ is the major stimulus for insulin secretion. Agents such as GIP, GLP-1, and PACAP (pituitary adenylate cyclase-activating polypeptide) act as both competence factors for glucose-induced secretion and potentiators of the glucose-induced responses. GIP has been shown to stimulate adenylyl cyclase (AC) in pancreatic beta cells and the major pathway by which GIP potentiates insulin secretion is thought to be at the level of Ca2þ-induced exocytosis via activation of protein kinase A (PKA). Although the pathways involved have not been completely identified, GIP is capable of increasing Ca2þ influx through both L-type Ca2þ channels and nonselective ion channels and releasing

K+

ATP ADP

Ca2+

Glucose Insulin

Ca2+ Ca2+ PKA

Metabolites G? R GIP

AA cAMP

AA Gs

PLA 2

AC R GIP

Figure 2 Diagram of a pancreatic islet beta cell showing K -dependent ATP channel-associated glucose stimulation of insulin secretion and proposed mechanisms by which GIP activation of adenylyl cyclase and phospholipase A2 can potentiate glucosemediated responses. þ

p0050

205

GIP (Gastric Inhibitory Polypeptide)

Ca2þ from intracellular stores. Ehses and co-workers have demonstrated that GIP can also activate a Ca2þindependent phospholipase A2 (PLA2) in beta cells, resulting in the generation of arachidonic acid. Interestingly, in different cell types, GIP may stimulate PLA2 via receptor coupling to Gbg dimers or via coupling of Ga and adenylyl cyclase.

GIP, Insulin Gene Transcription, and Beta Cell Proliferation In addition to its effect on insulin secretion, glucose is the most important stimulator of proinsulin gene transcription. Its action involves stimulation of cytosolic/nuclear transport of the transcription factor PDX-1 and increased PDX-1 gene transcription. However, both GIP and GLP-1 stimulate the expression of several genes in the beta cell, including those of proinsulin, glucose transporters, and hexokinases. Additionally, these peptides appear to be involved in regulating the differentiation of beta cells from ductal progenitor cells and stimulating beta cell proliferation. Incretins are therefore important regulators of beta cell replication. Although it is not known exactly how GIP and GLP-1 exert their mitogenic actions, it is known that they can induce c-fos activity and activate mitogen-activated protein kinase pathways.

METABOLISM Through the use of assays of immunoreactive GIP levels, it was determined that the kidney is the major site of GIP clearance. By measuring the rate of disappearance of infused peptide, the circulating half-life in humans was calculated to be 20 min. These measurements can be called into question as it is known that postsecretory degradation of incretins results in products that may be immunoreactive but biologically inactive. Thus, RIAs may overestimate their biological half-lives. Dipeptidyl peptidase IV (DP IV) is a ubiquitously distributed enzyme with substrate specificity for small peptides with proline or alanine residues in the penultimate position from the N terminus. Both GIP and GLP-1 serve as substrates for this enzyme (penultimate alanine residues), yielding N-terminally truncated, biologically inactive products (GIP3–42,GLP-19–36). It has been determined that DP IV degradation is the major route of GIP inactivation in the circulation and that the biological half-life of the hormone is approximately 2 min. The fact that most RIAs employ antibodies

directed against the C-terminal region of GIP calls much of the previously published data on circulating GIP levels into question since biologically inactive GIP contributes to immunoreactive GIP levels measured with these RIAs. The physiological role played by DP IV in the regulation of incretin activity, and thus the regulation of blood glucose, has been exploited by Pederson and co-workers in strategies used to enhance insulin secretion and improve glucose tolerance in type 2 diabetes. These strategies have taken the form of administration of DP IV inhibitors that increase the circulating half-life of incretins and the design and synthesis of DP IV-resistant GIP and GLP-1 analogues. An extensive body of evidence indicates that altering plasma incretin concentrations by inhibition of DP IV is a valid therapeutic approach to type 2 diabetes.

BIOLOGICAL ACTIONS Acid Secretion—The Enterogastrone Concept As discussed earlier, GIP was isolated on the basis of its acid inhibitory properties. GIP was pursued as a candidate for the inhibitor of gastric acid secretion: enterogastrone. This proposal gained considerable support when, in addition to its inhibitory actions on the stomach, it was determined that ingestion of fat was the most potent stimulus for the release of GIP into the circulation. Initial studies were carried out in the denervated canine stomach and subsequent studies in the intact innervated stomach of humans and rats demonstrated weaker inhibitory effects. In later studies, GIP was shown to be a potent stimulus for somatostatin release from the stomach and this release could be inhibited by the vagus or cholinomimetic agents. These studies suggested an indirect pathway for the action of GIP on the parietal cell, i.e., by GIPinduced somatostatin release with neuronal modulation by the parasympathetic nervous system. GIP is undoubtedly not the only enterogastrone and it is likely that it acts in concert with other peptides and/ or nervous inhibitory mechanisms that would be initiated during the digestive and absorptive process in the upper small bowel.

GIP Stimulation of Islet Hormones—The Incretin Concept As outlined earlier, both enterogastrone and incretin actions of GIP were hypothesized as a result of studies

206 on impure (GIP-containing) preparations of CCK that demonstrated their ability to inhibit acid secretion and stimulate insulin release. Dupre and coworkers showed that concomitant infusion of porcine GIP and glucose in humans produced a pronounced increase in circulating insulin levels and improved glucose tolerance compared to glucose infusion alone. The insulin response was sustained for the duration of the GIP infusion and was not observed in fasted (euglycemic) individuals. This glucose dependency was subsequently verified in the isolated perfused rat pancreas and a threshold of 5.5 mM glucose for the insulinotropic action of GIP was established. Elegant studies by Anderson et al. established conclusively that endogenous GIP released by oral glucose could account for a major part of the enteric contribution to the insulin response to oral glucose in human. The glucose-dependent insulinotropic action of GIP has been demonstrated in many mammalian species as well as a variety of tumorderived insulin secreting beta cell lines. The action of other insulin secretagogues, such as CCK, acetylcholine, GLP-1, and arginine, have also been demonstrated to be glucose-dependent. A common mechanism whereby glucose metabolism in the beta cell permits stimulation by nonglucose secretagogues has been proposed. GIP has been shown to be glucagonotropic in both the isolated rat pancreas and pancreatic islets. Several studies support a direct effect of GIP on glucagon release. Receptors for GIP are found on rat A cells and GIP increased whole-cell Ca2þ currents and potentiated exocytosis in purified rat pancreatic A cells.

GIP and Fat Metabolism In addition to its effect on insulin secretion, GIP has been implicated in the regulation of fat metabolism, an action that could also be of paramount importance in the normal regulation of insulin secretion. Ingestion of triglycerides increases circulating GIP by a pathway that involves both metabolism and absorption of fatty acids. Several lines of evidence support a role for GIP in the subsequent disposal of circulating triglycerides. Administration of GIP promotes chylomicron-associated triglyceride clearance from blood in dogs and reduces plasma triglyceride levels during intraduodenal fat infusion in rats. GIP has also been shown to stimulate lipoprotein lipase activity in cultured preadipocytes. Regarding the effects of GIP on adipocyte metabolism, it has been reported that GIP enhances the incorporation of

GIP (Gastric Inhibitory Polypeptide)

both fatty acids and glucose into lipids and inhibits glucagon-stimulated lipolysis and cAMP production. This seemed paradoxical, however, in that GIP signals via cAMP, which is known to stimulate hormonesensitive lipase and thus would be expected to exert a lipolytic effect. It was hypothesized by McIntosh and co-workers that the response of the adipocyte to GIP may depend on the prevailing circulating insulin concentration and evidence that GIP exerts a lipolytic effect on adipocytes that is inhibited by insulin was provided. One can postulate a physiological role for GIP in fat metabolism in light of the demonstration that critical levels of circulating free fatty acids (FFAs) are required for optimal glucose stimulation of insulin secretion during fasting. GIP could be capable of stimulating lipolysis under conditions in which insulin levels are of insufficient magnitude to inhibit its action and this may ensure that levels of circulating FFAs are optimal for glucose- and GIP-stimulated insulin secretion.

See Also the Following Articles CCK (Cholecystokinin) . Gastrin . Gastrin-Releasing Peptide . GI-CGRP (Calcitonin Gene-Related Peptide) . GI Hormone Development (Families and Phylogeny) . Pancreatic Islet Cell Tumors

Further Reading Andersen, D. K., Elahi, D., Brown, J. C., Tobin, J. D., and Andres, R. (1978). Oral glucose augmentation of insulin secretion: Interactions of gastric inhibitory polypeptide with ambient glucose and insulin levels. J. Clin. Invest. 62, 152–161. Brown, J. C., Pederson, R. A., Jorpes, E., and Mutt, V. (1969). Preparation of highly active enterogastrone. Can. J. Physiol. Pharmacol. 47, 113–114. Dupre, J., Ross, S. A., Watson, D., and Brown, J. C. (1973). Stimulation of insulin secretion by gastric inhibitory polypeptide in man. J. Clin. Endocrinol. Metab. 37, 826–828. Ehses, J. A., Lee, S. T., Pederson, R. A., and McIntosh, C. H. S. (2001). A new pathway for glucose-dependent insulinotropic polypeptide (GIP) receptor signaling. J. Biol. Chem. 276, 23667–23673. Hinke, S. A., Manhart, S., Pamir, N., Demuth, H.-U., Gelling, R. W., Pederson, R. A., and McIntosh, C. H. S. (2001). Identification of a bioactive domain in the amino-terminus of glucose-dependent insulinotropic polypeptide (GIP). Biochim. Biophys. Acta 364, 1–13. Kosaka, T., and Lim, R. K. S. (1930). Demonstration of the humoral agent in fat inhibition of gastric secretion. Proc. Soc. Exp. Biol. Med. 27, 890–891. Lynn, F. C., Pamir, N., Ng, E. H. C., McIntosh, C. H. S., and Pederson, R. A. (2001). Defective glucose-dependent insulinotropic polypeptide receptor expression in diabetic fatty Zucker rats. Diabetes 50, 1004–1011.

GIP (Gastric Inhibitory Polypeptide)

McIntosh, C. H. S., Pederson, R. A., Koop, H., and Brown, J. C. (1981). Gastric inhibitory polypeptide stimulated secretion of somatostatin-like immunoreactivity from the stomach: Inhibition by acetylcholine or vagal stimulation. Can. J. Physiol. Pharmacol. 59, 468–472.

Gitelman’s Syndrome see Bartter’s Syndrome

207 Pederson, R. A., White, H., Schleznig, D., Schmidt, J., McIntosh, C. H. S., and Demuth, H. U. (1998). Improved glucose tolerance in Zucker fatty rats by oral administration of the dipeptidyl peptidase IV (DP IV) inhibitor, isoleucine-thiazolidide. Diabetes 47, 1253–1258.

221

Glucagon and Glucagon-like Peptides

Figure 1 Homology of amino acid sequences of glucagon and the glucagon-like peptides. of the gut GLP-1 and GLP-2 are the main gene products. The steps in the expression of the preproglucagon gene are outlined in Fig. 2. Several other potentially bioactive peptides are generated from proglucagon such as IP2 (intervening peptide 2) (as shown in Fig. 2) but this article focuses on glucagon and the glucagon-like peptides.

HORMONE RECEPTORS AND CELL SIGNALING The functional effects of glucagon and the GLPs are mediated by specific receptors located within the plasma membrane of target cells. Like the peptide hormones, the hormone-specific receptors share extensive amino acid sequence homology and form part of a closely related subfamily of receptors as part of the larger superfamily of G protein-coupled receptors. This superfamily is distinguished by the tertiary structure of its members, comprising seven transmembrane domains, and the coupling to trimeric G proteins. The coupling to the stimulatory G protein Gs, the activation of adenylyl cyclase, and the subsequent increase in the intracellular production of cyclic AMP (cAMP) constitute the most established functional responses elicited by all the members of the

glucagon receptor subfamily. This uniform response in intracellular function suggests that the divergent physiological effects of glucagon, GLP-1, and GLP2 are due primarily to tissue-specific expression of the receptors, whereby cAMP imparts distinct cellular functions due to the equally diverse array of gene expression patterns in different cell types. Furthermore, temporal and tissue-specific patterns of secretion of the hormones themselves play a critical role in initiating the appropriate functional response. Notably, in the pancreatic beta cell, in which GLP-1 and glucagon receptors are expressed, both glucagon and GLP-1 stimulate insulin secretion, mediated at least in part by cAMP. Paradoxically, however, insulin gene expression is inhibited by glucagon and stimulated by GLP-1. This contradiction suggests that more subtle distinctions in receptor signaling exist, leading to differential intracellular responses to the two hormones, glucagon and GLP-1.

PHYSIOLOGY Glucagon The glucagon receptor is widely expressed and can be found in the liver, adipose tissue, heart, kidney,

Figure 2 Tissue-specific processing of the preproglucagon gene.

222

Glucagon and Glucagon-like Peptides

pancreatic islets, stomach, small intestine, thyroid, and skeletal muscle. Although contradictions in the specific expression pattern exist, likely due to the sensitivity of detection techniques between laboratories, the less ambiguous targets with high receptor levels correspond to the most well-established physiological actions of glucagon. The hepatocyte is a primary target cell of glucagon to which it is exposed when the hormone is released in the portal vein following secretion from the pancreatic alpha cells. Glucagon increases glucose output from the liver, an effect that results from inhibition of glycogen synthesis and stimulation of both glycogenolysis and gluconeogenesis (Table I). Evidence suggests that these effects of glucagon are mediated by cAMP, although the glycogenolytic effect of glucagon may occur via a cAMP-independent mechanism. It is not surprising, therefore, that glucagon helps to maintain hepatic glucose output and thereby to sustain blood glucose levels. There exists a general consensus that the hyperglycemic activity of glucagon represents the first line of defense against hypoglycemia. Accordingly, stimulants of glucagon secretion from alpha cells include the presence and actions of factors at times when the body requires glucose for sustenance (during fasting) or fuel (during exercise). The interplay between insulin and glucagon provides a tightly controlled equilibrium in blood glucose concentration. Insulin stimulates glucose uptake into fat and muscle cells and also inhibits glucagon expression and secretion from alpha cells. Thus, during periods when plasma insulin levels are low, such as during periods of fasting and exercising, plasma glucagon levels are elevated. The anatomical structure of the pancreatic islets of Langerhans further reflects this mechanistic action of insulin on glucagon secretion. Morphological studies have shown that in a given islet, the microcirculation goes from the core to the mantle. This circumstance suggests that in vivo, peripheral alpha cells are exposed to high concentrations of insulin released from the more centrally located beta cells in the islets. Thus, insulin and glucagon function inversely to regulate blood glucose levels.

t0005

Table I Actions of Glucagon "

Hyperglycemia

"

Glycogenolysis

"

Gluconeogenesis

"

Glycerol and free fatty acid release

"

Lypolysis

#

Glycogen synthesis

Dysregulation of glucagon by a breakdown in the intra-islet insulin–glucagon relationship appears to be a substantial component in the pathogenesis of diabetes. As a result of insufficient insulin secretion, plasma glucagon levels are elevated in type 1 (insulin-dependent) diabetes and also to a lesser extent in type 2 (non-insulin-dependent) diabetes, thereby exacerbating the hyperglycemic state. This effect is further heightened when plasma glucagon levels fail to be suppressed following a meal and postprandial hyperglycemia ensues. It is compelling to speculate, therefore, that agents that are able to antagonize glucagon action or secretion may be of value in the treatment of patients with diabetes. Paradoxically, a major feature of long-term diabetes is the impaired response to insulin-induced hypoglycemia, a condition of impairment that is almost universally present after 5 years of duration of type 2 diabetes. Studies have shown that hyperinsulinemia, which may be caused by insulin therapy, and insulin resistance can induce defective counterregulation in the manifestation of hypoglycemia, via defective glucagon release. These examples serve to emphasize the finely balanced hormonal equilibrium that characterizes the normal physiological state of glucose homeostasis. Consistent with the role of glucagon in providing the body with usable energy, additional established targets of glucagon action are the adipocyte and skeletal myocytes (Table I). Although studies of the effects of glucagon on adipose tissue and other tissues are difficult to evaluate owing to the rapid breakdown of the peptide by proteolytic activity associated with these cells, substantial evidence demonstrates that glucagon induces lipolysis, stimulates the release of glycerol and free fatty acids in adipocytes, and stimulates glycogenolysis in myocytes. Therefore, glucagon could qualify as the ‘‘hormone of fuel need’’ rather than just a ‘‘hormone of glucose need.’’

GLP-1 The major secreted and bioactive form of GLP-1 from enteroendocrine L cells is a truncated isopeptide cleaved from the full 37-amino-acid peptide. Cleavage of the six N-terminal amino acids of GLP-1(1–37) to form GLP-1 (7–37) and removal of the glycine residue at position 37 followed by amidation of the exposed arginine at position 36 generate GLP-1(7– 36)amide. The most significant physiological function of GLP-1 is the stimulation of the synthesis and secretion of insulin from pancreatic islet beta cells

223

Glucagon and Glucagon-like Peptides

t0010

Table II Actions of GLP-1 "

Insulin biosynthesis and secretion

"

Islet proliferation and neogenesis

"

Islet cell differentiation

"

Glucose uptake

#

Gastric emptying

# #

Acid secretion Food intake (appetite)

#

Glucagon secretion

(Table II). The mechanism of action of GLP-1 is a potentiation of glucose-induced insulin secretion. In the absence of glucose, GLP-1 fails to stimulate insulin secretion, whereas at elevated glucose concentrations, GLP-1 stimulates secretion at a rate that exceeds that induced by glucose alone and thus potentiates secretion in a glucose-dependent manner. Notably, the secretion of GLP-1 from L cells is stimulated by glucose (and other nutrients) following oral ingestion. Thus, the temporal release of GLP-1, coincident with enhanced blood glucose levels, cooperates synergistically to provide a highly sensitive insulin-releasing mechanism. GLP-1 is thus regarded as a ‘‘glucose competence factor.’’ As described above, insulin suppresses plasma glucagon levels and activates glucose uptake and metabolism, thus returning blood glucose levels to normal following a meal and maintaining glucose homeostasis in general. In this regard, insulin and GLP-1 play an important role in preventing hyperglycemia. Evidence for additional mechanisms of action of GLP-1 in lowering blood glucose levels have emerged; GLP-1 has been shown to augment insulin-mediated glucose uptake in patients with diabetes. Extrapancreatic targets of GLP-1 actions, particularly those of adipose and skeletal muscle tissues, are under investigation. It seems possible that a novel GLP-1 receptor mediates these effects of GLP-1 on adipose tissue and skeletal muscle, as well as liver. Additional functions of GLP-1 include the inhibition of gastric emptying (by inhibiting intestinal motor activity in response to nutrients in the distal gut) and of meal-stimulated gastric acid secretion. A role for GLP-1 in feeding behavior has also been identified. Centrally administered GLP-1 inhibited both feeding and drinking activity in rats, suggesting that GLP-1 may be a central neurotransmitter that modulates visceral functions. Finally, in addition to the detection of pluripotential stem cells within the pancreas, some evidence suggests that GLP-1 stimulates the proliferation and differentiation of these

stem/progenitor cells into beta cells and thus GLP-1 may play a developmental role in the endocrine pancreas. As the characterization of GLP-1 action continues and a clearer picture of its multifunctional profile appears, it is evident that GLP-1 offers considerable potential as a therapeutic treatment for diabetes mellitus. By enhancing both the means to secrete insulin (by incretin action and differentiation of new beta cells) and the functional ability of insulin to induce glucose disposal, it may be possible to better control circulating blood sugar levels in diabetic patients.

GLP-2 GLP-2 is synthesized in the enteroendocrine L cells (along with GLP-1) and in the brainstem and hypothalamus of the central nervous system. GLP-2 may act on the hypothalamus to curtail appetite and food intake. However, the regulation and function of GLP2 in the intestinal system are better understood than those of GLP-2 in the central nervous system. Initially, several lines of evidence suggested the importance of GLP-2 in maintaining the functional integrity of the gastrointestinal epithelium. Examples included the coincidence of intestinal injury and inflammatory bowel disease with elevated levels of circulating GLP-2 and patients with glucagon-secreting tumors who presented with small bowel villus hyperplasia (Table III). It was subsequently shown that administration of GLP-2 to mice leads to induction of intestinal epithelial proliferation. Furthermore, following treatment of mice and rats with exogenous GLP-2 for 7–10 days, a marked increase in villus height and small bowel mass and a smaller increment in small bowel length are observed. GLP-2 also appears to slow the ingestion and transit of food through the gastrointestinal tract, while increasing the absorption of nutrients from the small intestine. As with glucagon and GLP-1, the tissue-specific effects of GLP-2 action are closely reflected by the expression profile of the GLP-2 receptor, which is predominantly found in the stomach, jejunum, ileum, and colon. Although GLP-2 presumably exerts direct effects on cells expressing its putative receptor, some physiological effects of GLP-2 in the intestine are thought to be exerted indirectly in an autocrine, paracrine, or endocrine manner by stimulating the release of as yet unknown factors from GLP-2 receptor-positive cells. This hypothesis arose due to conflicting data regarding the exact tissue distribution of the GLP-2 receptor. Whereas one group observed

224

t0015

p0090

Glucagon and Glucagon-like Peptides

Table III Actions of GLP-2 "

Nutrient absorption

"

Crypt cell proliferation

"

Small and large bowel growth

"

Villus height

#

Food intake (appetite)

# #

Gastric motility Enterocyte and crypt apoptosis

that binding of labeled GLP-2 localized diffusely along the villous epithelium, others have reported the absence of the receptor among several intestinal epithelial cell lines. Thus, because the localization profile of the GLP-2 receptor is not known with certainty, the mechanism of the pleiotropic biological actions of GLP-2 in the gut, such as modulation of gastric motility, small bowel permeability, and both crypt cell proliferation and apoptosis, is confined to speculation. However, some evidence suggests that GLP-2 can directly mediate proliferative and antiapoptotic effects through its receptor in cultured cells. Apart from its intestinal functions, GLP-2 is also expressed and received in the brain. Receptors were located in the hypothalamus and in certain cell groups in the cerebellum, medulla, amygdala, and hippocampus. Although studies of GLP-2 action in the brain are just beginning, a functional role of GLP-2 appears to be in the regulation of food intake. Pharmacological and behavioral studies indicate that GLP-2 acts as a specific transmitter that may inhibit feeding behavior and has potential long-term effects on body weight homeostasis.

INACTIVATION OF GLPs The active half-life of both GLP-1 and GLP-2 is regulated by dipeptidyl peptidase IV (DPP IV/ CD26), an extracellular soluble or membraneanchored protease, which cleaves the two N-terminal amino acids and subsequently inactivates the hormones, with respect to their known functions. This potent inactivation restricts the circulating bioactive

half-lives to 2–3 min in vivo. Particularly high levels of expression of DPP IV are found in the kidney, lung, liver, and jejunum. Expression in endothelial cells of the blood vessels and the presence of small amounts of DPP IV as a soluble enzyme in the blood result in close contact with circulating substrate and added rapidity of peptide inactivation. Consequently, specific DPP IV inhibitors are of special interest for physiological investigations and for potential clinical applications, such as treatment of type 2 diabetes and improvement of mucosal regeneration. Studies have shown that prolonged activation of GLP-1 by DPP IV inhibitors results in a potentiation of the insulinotropic effect of GLP-1 in vivo and administration of GLP-2 to DPP IV-deficient rats markedly increased the bioactivity of GLP-2, resulting in a significant increase in small bowel weight. A second approach to enhance the bioactive half-life of the GLPs is the synthesis of DPP IV-resistant analogues of the GLPs for pharmacological administration. These approaches are in development.

See Also the Following Articles GI Hormone Development (Families and Phylogeny) . Insulin Secretion: Functional and Biochemical Aspects . Pituitary Adenylate Cyclase-Activating Polypeptide (PACAP)/ Glucagon Superfamily

Further Reading Ahren, B. (1998). Glucagon-like peptide-1 (GLP-1): A gut hormone of potential interest in the treatment of diabetes. BioEssays 20, 642–651. Drucker, D. J. (1998). Glucagon-like peptides. Diabetes 47, 159–169. Habener, J. F. (2001). Glucagonlike peptide-1 agonist stimulation of b-cell growth and differentiation. Curr. Opin. Endocrinol. Diabetes 8, 74–81. Kieffer, T. J., and Habener, J. F. (1999). The glucagon-like peptides. Endocr. Rev. 20, 876–913. Lefgbvre, P. J. (1995). Glucagon and its family revisited. Diabetes Care 18, 715–730. Lovshin, J., and Drucker, D. J. (2000). New frontiers in the biology of GLP-2. Regul. Pept. 90, 27–32. Mentlein, R. (1999). Dipeptidyl-peptidase IV (CD26)-role in the inactivation of regulatory peptides. Regul. Pept. 85, 9–24.

226

Glucagon-like Peptide 2 (GLP-2)

Figure 1 Structure of proglucagon and biological actions of GLP-2. GLP-2 administration reduces epithelial permeability in the small intestine, effects noted within hours of peptide administration. The capacity of the GLP2-treated bowel for nutrient absorption appears to be normal to enhanced, as assessed by nutrient challenge studies in normal mice following chronic GLP-2 administration.

GLP-2 ACTIONS IN THE BRAIN p0030

GLP-2 is synthesized in the brain, predominantly in the nucleus of the solitary tract in the brainstem and to a lesser extent in the hypothalamus. GLP-1 and GLP-2 are transported from the brainstem to distant central nervous system (CNS) nuclei via axonal transport. Within the hypothalamus, GLP-2 is distributed in a pattern similar to that of GLP-1; however, GLP-2 administration specifically activates neurons in the dorsomedial nucleus of the hypothalamus, where GLP-2 receptors have been localized by in situ hybridization. The GLP-2 receptor is also expressed in thalamic, cortical, cerebellar, hippocampal, and amygdaloid nuclei. Although the physiological role of GLP-2 in the CNS remains unclear, intracerebroventricular administration of pharmacological amounts of GLP-2 transiently inhibits food intake in mice and rats.

GLP-2 AND GUT DEVELOPMENT p0035

GLP-2 and GLP-2 receptors are expressed in the fetal and neonatal gut and GLP-2 stimulates intestinal epithelial proliferation in neonatal pigs and rats. The trophic effects of GLP-2 in the gastrointestinal tract may be restricted to postnatal life, as exogenous GLP-2 does not stimulate gut growth in late-gestational fetal pigs despite the presence of immunoreactive circulating GLP-2 in the fetal pig.

THE GLP-2 RECEPTOR GLP-2 exerts its actions through a distinct 550amino-acid G protein-coupled GLP-2R (GLP-2 receptor). The cDNA encoding the GLP-2R was initially cloned from human and rat hypothalamic and intestinal cDNA libraries. Consistent with the known actions of GLP-2, GLP-2R expression is predominantly confined to the stomach, duodenum, jejunum, ileum, and colon. GLP-2R transcripts have also been detected in the brain and lung, although the biological significance of GLP-2R expression in lung remains unclear. Within the gastrointestinal tract, GLP-2 receptor-like immunopositivity has been localized to a subset of human endocrine cells using GLP-2R-specific antisera and GLP-2R expression was demonstrated in the murine enteric nervous system by in situ hybridization. Although cell lines expressing an endogenous GLP-2 receptor have not yet been identified, GLP-2 increases intracellular cyclic AMP (cAMP) and activates protein kinase A but does not increase intracellular Ca2þ release in cells expressing a transfected human or rat GLP-2 receptor. GLP-2 directly activates immediate-early gene expression and cellular growth in serum-starved fibroblasts in vitro; however, the growth-promoting effects of GLP-2 in the gut are largely indirect in vivo. GLP-2 also prevents apoptosis following administration of cycloheximide (an inhibitor of protein synthesis) or chemotherapeutic agents to cells expressing a GLP-2R through cAMP-dependent mechanisms; hence, GLP-2R activation directly and indirectly signals intracellular pathways regulating both mitogenic and anti-apoptotic actions.

THERAPEUTIC POTENTIAL OF GLP-2 GLP-2 infusion completely prevented the mucosal atrophy associated with withdrawal of enteral

227

Glucagon-like Peptide 2 (GLP-2)

nutrients in the small intestine but not the large intestine of the rat. Similarly, GLP-2 administration following small bowel resection in rats increased intestinal remnant adaptation and improved absorptive function. Reparative and protective effects of GLP-2 have also been observed in mice following ischemic injury induced by occlusion of the superior mesenteric artery. GLP-2 exerts reparative and cytoprotective effects in rodents in the setting of intestinal inflammation. The protease-resistant GLP-2 analogue h[Gly2]GLP-2 significantly reduced the severity of enteritis, bacterial sepsis, and mortality in mice following administration of the nonsteroidal anti-inflammatory agent indomethacin. Similarly, mice treated with h[Gly2]-GLP-2 exhibit reduced large bowel injury and significantly less weight loss in the setting of dextran sulfate colitis. The anti-apoptotic actions of GLP-2 are also detectable in vivo following chemotherapy administration. Administration of h[Gly2]-GLP-2 prior to irinotecan or 5-fluorouracil attenuated epithelial injury in the gut, decreased intestinal epithelial apoptosis, and reduced mortality in mice. Although GLP-2 promotes intestinal epithelial proliferation in the gastrointestinal tract, co-infusion of GLP-2 and parenteral nutrition does not affect tumor progression or tumor growth in tumor-bearing rats. The reparative and proabsorptive effects of GLP-2 have been examined in a pilot trial in human subjects with short bowel syndrome. Eight patients with stable chronic intestinal insufficiency without a colon in continuity were treated with human GLP-2, 400 mg twice daily by subcutaneous injection for 35 days. GLP-2-treated subjects exhibited significantly increased energy, wet weight, and nitrogen absorption, as measured in metabolic balance studies. Body weight and lean body mass increased, whereas fat mass decreased and the time to 50% gastric emptying was increased in GLP-2-treated subjects. Analysis of intestinal histology revealed increases in crypt depth and villus height following GLP-2 administration.

SUMMARY GLP-2 has been established as the elusive proglucagon-derived peptide with potent effects on gut growth and nutrient absorption. The essential physiological role(s) of GLP-2 has not yet been elucidated due to the lack of effective specific antagonists or immunoneutralizing antisera. The cytoprotective, proabsorptive, and regenerative properties of GLP-2 in vivo

appear to be indirect, mediated by a GLP-2 receptor expressed on endocrine cells and enteric neurons. The downstream effectors released following GLP-2R activation are unknown. Although intracerebroventricular administration of GLP-2 inhibits food intake in rodents, it seems likely that GLP-2 exerts additional as yet unknown actions in the central nervous system. The rapid degradation and clearance of GLP-2 suggest that protease-resistant analogues may exhibit advantages for long-term therapeutic administration in vivo. The diverse actions of GLP-2 in the gut are focused on maintaining the integrity of the epithelial mucosa and optimizing nutrient absorption, consistent with the physiological importance of proglucagon-derived peptides in the integrated control of energy homeostasis.

See Also the Following Articles GI Hormone Development (Families and Phylogeny) Hormones as Growth Factors

.

GI

Further Reading Bjerknes, M., and Cheng, H. (2001). Modulation of specific intestinal epithelial progenitors by enteric neurons. Proc. Natl. Acad. Sci. USA 98, 12497–12502. Boushey, R. P., Yusta, B., and Drucker, D. J. (1999). Glucagon-like peptide 2 decreases mortality and reduces the severity of indomethacin-induced murine enteritis. Am. J. Physiol. 277, E937–E947. Boushey, R. P., Yusta, B., and Drucker, D. J. (2001). Glucagon-like peptide (GLP)-2 reduces chemotherapy-associated mortality and enhances cell survival in cells expressing a transfected GLP-2 receptor. Cancer Res. 61, 687–693. Brubaker, P. L., Crivici, A., Izzo, A., Ehrlich, P., Tsai, C.-H., and Drucker, D. J. (1997). Circulating and tissue forms of the intestinal growth factor, glucagon-like peptide 2. Endocrinology 138, 4837–4843. Burrin, D. G., Stoll, B., Jiang, R., Petersen, Y., Elnif, J., Buddington, R. K., Schmidt, M., Holst, J. J., Hartmann, B., and Sangild, P. T. (2000). GLP-2 stimulates intestinal growth in premature TPN-fed pigs by suppressing proteolysis and apoptosis. Am. J. Physiol. Gastrointest. Liver Physiol. 279, G1249–G1256. Drucker, D. J. (2001). Glucagon-like peptide 2. J. Clin. Endocrinol. Metab. 86, 1759–1764. Drucker, D. J., Ehrlich, P., Asa, S. L., and Brubaker, P. L. (1996). Induction of intestinal epithelial proliferation by glucagon-like peptide 2. Proc. Natl. Acad. Sci. USA 93, 7911–7916. Drucker, D. J., Shi, Q., Crivici, A., Sumner-Smith, M., Tavares, W., Hill, M., Deforest, L., Cooper, S., and Brubaker, P. L. (1997). Regulation of the biological activity of glucagon-like peptide 2 by dipeptidyl peptidase IV. Nat. Biotechnol. 15, 673–677. Lovshin, J., Estall, J., Yusta, B., Brown, T. J., and Drucker, D. J. (2001). Glucagon-like peptide-2 action in the murine central nervous system is enhanced by elimination of GLP-1 receptor signaling. J. Biol. Chem. 276, 21489–21499.

228 Yusta, B., Boushey, R. P., and Drucker, D. J. (2000). The glucagonlike peptide-2 receptor mediates direct inhibition of cellular apoptosis via a cAMP-dependent protein kinase-independent pathway. J. Biol. Chem. 275, 35345–35352. Xiao, Q., Boushey, R. P., Drucker, D. J., and Brubaker, P. L. (1999). Secretion of the intestinotropic hormone glucagon-like peptide

Glucagon-like Peptide 2 (GLP-2)

2 is differentially regulated by nutrients in humans. Gastroenterology 117, 99–105. Yusta, B., Huang, L., Munroe, D., Wolff, G., Fantaske, R., Sharma, S., Demchyshyn, L., Asa, S. L., and Drucker, D. J. (2000). Enteroendocrine localization of GLP-2 receptor expression. Gastroenterology 119, 744–755.

230

f0005

Glucocorticoid Receptor

Figure 1 (A) Genomic, complementary DNA, and protein structures of the human glucocorticoid receptor (hGR). The hGR gene consists of 10 exons. Exon 1 is an untranslated region, exon 2 codes for the amino-terminal domain, exons 3 and 4 code for the DNA-binding domain, and exons 5–9 code for the hinge region and the ligand-binding domain. The glucocorticoid receptor gene contains two terminal exons 9 (exon 9a and 9b), alternatively spliced to produce the hGRa and hGRb isoforms. (B) Functional domains of the glucocorticoid receptor. The functional domains and subdomains are indicated beneath the linearized protein structures. AF, activation function; DBD, DNA-binding domain; LBD, ligand-binding domain; NLS, nuclear localization signal.

MOLECULAR MECHANISMS OF GR ACTION Nucleocytoplasmic Shuttling of GR p0025

In the absence of ligand, hGRa resides primarily in the cytoplasm of cells as part of a large multiprotein complex, which consists of the receptor polypeptide, two molecules of hsp90, and several other proteins. The hsp90 molecules are thought to sequester hGRa in the cytoplasm of cells by maintaining the receptor in a conformation that masks or inactivates its nuclear localization signals (NLSs). Upon hormone binding, the receptor undergoes an allosteric change, which results in dissociation from hsp90 and other proteins,

unmasking of the NLSs, and exposure of the ligandbinding pocket. In its new conformation, the activated, ligand-bound hGRa translocates into the nucleus, where it binds as homodimer to glucocorticoidresponse elements (GREs) located in the promoter region of target genes. HGRa then communicates with the basal transcription machinery and regulates the expression of glucocorticoid-responsive genes positively or negatively, depending on the GRE sequence and promoter context (Fig. 2). The receptor can also modulate gene expression, as a monomer, independently of GRE binding, by physically interacting with other transcription factors, such as activating protein-1 (AP-1) and nuclear factor kB (NF-kB).

231

Glucocorticoid Receptor

f0010

Figure 2 Nucleocytoplasmic shuttling of the glucocorticoid receptor. GR, glucocorticoid receptor; GREs, glucocorticoidresponse elements; HSPs, heat shock proteins; TF, transcription factor; TFREs, transcription factor-responsive elements.

Mechanisms of Transcriptional Activation by GR p0030

p0035

Following binding to GREs, the liganded, activated hGRa enhances the expression of glucocorticoidresponsive genes by regulating the assembly of a transcriptional preinitiation complex at the promoter region of target genes. This is achieved by interaction of the liganded receptor with the basal transcription factors, a group of proteins composed of RNA polymerase II, TATA-binding protein (TBP), and a host of TBP-associated proteins (TAFIIs). The interaction between the activated receptor and the basal transcription factors is mediated by the coactivators, which are nucleoproteins with chromatin-remodeling activity and other enzymatic activities (Fig. 3). Like other transcriptional activators, hGRa uses its transcriptional activation domains AF-1 and AF-2 as surfaces to recruit chromatin remodeling factors and to interact with the coactivators that link enhancer-bound transcription factors to general transcription factors, thereby initiating transcription. At least two regions of hGRa possess intrinsic transcriptional activation functions: AF-2, which maps to the carboxyl terminus, is glucocorticoid-dependent, with ligand binding promoting the formation of a surface that permits protein–protein contacts between AF-2 and additional regulatory factors. In contrast, AF-1 is located at the amino terminus of the receptor, is glucocorticoidindependent, and can recruit both positive and negative regulatory factors that differentially regulate hGRa transcriptional enhancement.

Several families of nuclear hormone receptor coactivators have been described, including the p160 coactivators, such as the steroid receptor coactivator 1 and the glucocorticoid receptor-interacting protein 1, the p300 and cyclic AMP-responsive elementbinding protein (CBP) cointegrators, and the p300/ CBP-associated protein (p/CAF). The p160 coactivators are the first to be tethered to the promoter region of steroid target genes, thus playing a pivotal role in hGRa-mediated transactivation. These coactivators interact directly with both the AF-1 of hGRa through their carboxyl-terminal domain and the AF-2 of hGRa through multiple amphipathic LXXLL signature motifs, which are located in their nuclear receptor-binding domain (NRB). They also contain an additional binding site for hGRa, called auxiliary nuclear receptor-interacting domain, which is located between the NRB and the carboxyl-terminal AF-1-binding site. The p160 proteins have intrinsic histone acetyl-transferase activity, which disrupts the DNA nucleosomal interactions at these promoters, allowing the initiation of transcription. Other coactivators that interact with hGRa include the switching/ sucrose nonfermenting complex and the newly described chromatin remodeling complex, vitamin D receptor-interacting protein (DRIP)/thyroid hormone-associated protein (TRAP) complex. The DRIP/TRAP complex interacts with both the AF-2 and AF-1 domains of hGRa via its components DRIP205 and DRIP150, respectively. Through coordinated interactions with AF-1 and AF-2, the

p0040

232

f0015

Glucocorticoid Receptor

Figure 3 Schematic representation of the interaction of the glucocorticoid receptor with coactivators and other chromatin modulators. AF, activation function; p/CAF, p300/CBP-associated factor; CBP, cAMP-responsive element-binding protein (CREB)-binding protein; DRIP, vitamin D receptor-interacting protein; SNF, sucrose nonfermenting; SWI, mating-type switching; TRAP, thyroid hormone receptor-associated protein.

coactivators enhance the transmission of signals from the DNA-bound hGRa to the transcriptional initiation machinery, loosen chromatin structure, and facilitate access and/or binding of other transcription factors and transcription initiation components to DNA, leading to full transcriptional activity of ligand-bound hGRa.

Interaction of GR with Other Transcription Factors In addition to binding to GREs, glucocorticoids may regulate transcription by physically interacting with other transcription factors. Protein–protein interactions between hGR and NF-kB and between AP-1 and signal transducers and activators of transcription result in negative or positive regulation of their responsive genes, mediating many of the anti-inflammatory and immunosuppressive effects of glucocorticoids. hGR may also interact with transcription factor Nur77 to regulate the expression of the proopiomelanocortin gene, p53, which functions as a tumor suppressor gene, as well as others, such as HNF6, Oct-1 and-2, and GATA-1.

FACTORS THAT MODULATE GR ACTIVITY Phosphorylation of GR Following ligand binding, hGR is phosphorylated at several sites. The mitogen-activated protein kinase

(MAPK) and the cyclin-dependent kinases (CDKs) phosphorylate hGR and modulate its transcriptional activity. MAPK suppresses the activity of hGR in yeast, whereas CDKs stimulate it. JNK, another mitogen-activated kinase, also phosphorylates hGR and suppresses its transcriptional activity. All these kinases phosphorylate hGR at different sites, indicating that the function of hGR may be also regulated by other signal transduction pathways through phosphorylation.

Chaperones and Cochaperones hGR forms heterocomplexes with several heat shock proteins, including hsp90, hsp70, hsp56, and possibly hsp23, which may affect its transcriptional activity. hsp90 and receptor-associating protein 46 regulate the transcriptional activity of hGR negatively.

Chemicals and Compounds Several chemical compounds may modulate the transcriptional activity of hGR. 2,3,7,8-Tetrachlorodibenzo-p-dioxin, a widespread environmental contaminant that produces adverse biological effects such as carcinogenesis, reproductive toxicity, immune dysfunction, hepatotoxicity, and teratogenesis, suppresses the activity of hGR by decreasing binding to glucocorticoids. Geldanamycin, a benzoquinone ansamycin, suppresses hGR function by inhibiting its translocation into the nucleus. On the other hand, thioredoxin, a compound that is accumulated during

p0055

233

Glucocorticoid Receptor

Figure 4 Location of the known mutations of the human glucocorticoid receptor in the genomic structure and its linearized molecule.

oxidative stress, and ursodeoxycholic acid, one of the hydrophilic bile acids, enhance hGR transactivation.

Natural GR Mutations p0065

Mutations in the hGRa gene may impair one or more of the molecular mechanisms of hGR function, resulting in alterations in tissue sensitivity to glucocorticoids and the clinical phenotype of glucocorticoid resistance. Glucocorticoid resistance is a rare, familial or sporadic condition characterized by generalized, partial end-organ insensitivity to physiologic glucocorticoid concentrations. Patients have compensatory elevations in circulating cortisol and adrenocorticotropic hormone concentrations and resistance of the hypothalamic–pituitary–adrenal axis to dexamethasone suppression, but no clinical evidence of hypo- or hypercortisolism. More than 10 kindreds and sporadic cases with the condition have been reported. Abnormalities of several hGRa characteristics, including cell concentrations, affinity for glucocorticoids, stability, and translocation into the nucleus, have been associated with this condition (Fig. 4).

has not been fully elucidated, but may involve competition between hGRa and hGRb for binding to GREs, formation of hGRa–hGRb heterodimers that are transcriptionally inactive, and/or titration or squelching of coactivators needed by hGRa for full transcriptional activity. The ability of hGRb to antagonize the function of hGRa suggests that hGRb may play a critical role in regulating target tissue sensitivity to glucocorticoids. Increased expression of hGRb has been documented in generalized and tissue-specific glucocorticoid resistance and leads to a reduction in the ability of hGRa to bind to GREs. Therefore, an imbalance in hGRa and hGRb expression may underlie the pathogenesis of several clinical conditions associated with glucocorticoid resistance, such as rheumatoid arthritis, systemic lupus erythematosus, and ulcerative colitis.

See Also the Following Articles Glucocorticoid Resistance Syndromes and States . Glucocorticoids and Immunity . Glucocorticoids in Aging: Relevance to Cognition . Glucocorticoids, Overview . Growth and Glucocorticoids . Nuclear Factor-kB and Glucocorticoid Receptors

Glucocorticoid Receptor-b Isoform p0070

hGRb functions as a dominant-negative inhibitor of hGRa activity and inhibits hGRa-mediated transactivation of many target genes in a dose-dependent manner. The mechanism(s) underlying this inhibition

Further Reading Auboeuf, D., Honig, A., Berget, S. M., and O’Malley, B. W. (2002). Coordinate regulation of transcription and splicing by steroid receptor coregulators. Science 298, 416–419.

234 Bamberger, C. M., Schulte, H. M., and Chrousos, G. P. (1996). Molecular determinants of glucocorticoid receptor function and tissue sensitivity to glucocorticoids. Endocr. Rev. 17, 245–261. Carson-Jurica, M. A., Schrader, W. T., and O’Malley, B. W. (1990). Steroid receptor family: Structure and functions. Endocr. Rev. 11, 201–220. Chrousos, G. P. (1995). The hypothalamic–pituitary–adrenal axis and immune-mediated inflammation. N. Engl. J. Med. 332, 1351–1362. Clark, J. K., Schrader, W. T., and O’Malley, B. W. (1992). Mechanism of steroid hormones. In ‘‘Williams Textbook of Endocrinology’’ ( J. D. Wilson and D. W. Foster, eds.), pp. 35–90. W. B. Saunders, Philadelphia, PA. Dalman, F. C., Scherrer, L. C., Taylor, L. P., Akil, H., and Pratt, W. B. (1991). Localization of the 90-kDa heat shock proteinbinding site within the hormone-binding domain of the glucocorticoid receptor by peptide competition. J.Biol. Chem. 266, 3482–3490. Hollenberg, S. M., Weinberger, C., Ong, E. S., Cerelli, G., Oro, A., Lebo, R., Thompson, E. B., Rosenfeld, M. G., and Evans, R. M. (1985). Primary structure and expression of a functional human glucocorticoid receptor cDNA. Nature 318, 635–641.

Glucocorticoid Receptor

Kino, T., Vottero, A., Charmandari, E., and Chrousos, G. P. (2002). Familial/sporadic glucocorticoid resistance syndrome and hypertension. Ann. N. Y. Acad. Sci. 970, 101–111. McKenna, N. J., and O’Malley, B. W. (2002). Combinatorial control of gene expression by nuclear receptors and coregulators. Cell 108, 465–474. McKenna, N. J., Xu, J., Nawaz, Z., Tsai, S. Y., Tsai, M. J., and O’Malley, B. W. (1999). Nuclear receptor coactivators: Multiple enzymes, multiple complexes, multiple functions. J. Steroid Biochem. Mol. Biol. 69, 3–12. Miura, T., Ouchida, R., Yoshikawa, N., Okamoto, K., Makino, Y., Nakamura, T., Morimoto, C., Makino, I., and Tanaka, H. (2001). Functional modulation of the glucocorticoid receptor and suppression of NF-kB-dependent transcription by ursodeoxycholic acid. J. Biol. Chem. 276, 47371–47378. Rogatsky, I., Logan, S. K., and Garabedian, M. J. (1998). Antagonism of glucocorticoid receptor transcriptional activation by the c-Jun N-terminal kinase. Proc. Natl. Acad. Sci. USA 95, 2050–2055. Williams, S. P., and Sigler, P. B. (1998). Atomic structure of progesterone complexed with its receptor. Nature 393, 392–396.

Glucocorticoid Resistance Syndromes and States Alessandra Vottero University of Parma, Italy, and NIH/National Institute of Child Health and Human Development, Bethesda, Maryland, United States

George P. Chrousos NIH/National Institute of Child Health and Human Development, Bethesda, Maryland, United States

Glossary glucocorticoids Any of several hormones, including cortisol, corticosterone, and cortisone, released by the cortex of the adrenal gland. receptor Molecule on the cell surface or inside the cell that recognizes a specific antigen or hormone. resistance Degree of nonresponsiveness to a particular hormone or disease in the body. Also called insensitivity.

G

lucocorticoids (GCs) are major effectors of basal and stress-related homeostasis, influencing cardiovascular function, regulating carbohydrate, protein, and fat metabolism, and modifying the immune/inflammatory response.

INTRODUCTION GCs activate the central nervous system and participate in the development and basal functions of several organs and systems. Glucocorticoids bring about these effects by binding to specific intracellular receptor proteins, called glucocorticoid receptors (GRs). Abnormalities of steroid sensitivity can be divided into two major groups: resistance and hypersensitivity. Resistance syndromes have been described for all steroids; they can be transient or permanent, partial or complete, and compensated or noncompensated. Complete glucocorticoid resistance is incompatible with life because of severe neonatal respiratory distress syndrome, as demonstrated in an in vivo study in which GR
/
knockout mice died within a few hours after birth. On the other hand, when partial steroid resistance occurs, the hypothalamic–pituitary–adrenal (HPA) axis is reset at higher circulating levels of adrenocorticotropic hormone (ACTH) and cortisol.

Encyclopedia of Endocrine Diseases, Volume 2. Published by Elsevier Inc.

The increased ACTH levels compensate for the insensitivity of tissues to glucocorticoids, but also result in increased secretion of glucocorticoid precursors with mineralocorticoid activity [deoxycorticosterone (DOC) and corticosterone] and adrenal androgens. Congenital glucocorticoid resistance has been described in over 10 kindreds and in a few sporadic cases and the molecular mechanisms of resistance in some of these patients have been analyzed.

p0020

MOLECULAR ACTIONS OF GLUCOCORTICOIDS At the cellular level, GCs carry out their actions through an 94 kDa intracellular receptor protein, the GR. This receptor belongs to the nuclear receptor superfamily (Fig. 1) of steroid/thyroid/retinoic acid receptor proteins, which function as ligand-dependent transcription factors. Since 1985, when the sequence of GR cDNA was first published, two alternative splicing products of the same gene located on chromosome 5 have been described (Fig. 2): the classic, active receptor, GR-a, and the dominant-negative isoform, GR-b. The two receptors are highly homologous, differing in just the last 50 and 15 amino acids, respectively. Therefore, the first eight exons of the GR gene containing the 50 noncoding and coding sequences are common to both receptor isoform cDNAs, whereas exons 9a and 9b containing the coding and 30 noncoding sequences are specific for GR-a and GR-b, respectively. In the absence of the ligand, GR-a resides mostly in the cytoplasm of cells in a multiprotein complex consisting of the receptor polypeptide, two molecules of heat shock protein 90 (Hsp90), and several other proteins. The Hsp90s are thought to allow proper folding and stabilization of the receptor, maintaining the latter in a high-affinity, hormone-binding friendly

235

p0025

236

Glucocorticoid Resistance Syndromes and States

Figure 1 Members of the nuclear receptor superfamily and their major homologies.

state and preventing its interaction with proteins of the importin system, which are involved in the cytoplasm-to-nucleus translocation of many proteins. Experimental evidence supports a model of constant bidirectional shuttling of the complex in the unliganded state between the cytosol and the nucleus. Once the hormone binds, the receptor–ligand complex releases the Hsps and homodimerizes with another activated GR-a. This complex interacts with the importin system and translocates via the nuclear pore into the nucleus, where it positively or negatively regulates gene expression by binding as a homodimer with GC-response elements (GREs) in the presence of coregulators [CBP/p300 and SRC-1 (p160) families of ‘‘coactivators’’; NcoR/RIP13 and SMRT/TRAC families of ‘‘corepressors’’] or as a monomer via protein–protein interactions with other transcription factors (Fig. 3A). Through those transcription factors, including activator protein-1 (AP-1), nuclear factor kB (NF-kB), and Stat-5, the GR modulates the transcription rates of non-GRE-containing genes regulated by these factors (Fig. 3B). In contrast to GR-a, GR-b does not bind to glucocorticoids and functions as weak dominant inhibitor of GR-a. This action is mediated by GRE binding, since no major protein–protein interaction with other transcription factors has been described. The intracellular localization of GR-b is uncertain; green fluorescent protein–GR-b fusion hybrids have been found in the nucleus of transfected cells, confirming the potential presence of GR-b in the nucleus, at a site other than the cytoplasm, under certain conditions even in the absence of ligand. The export of GR-b from the nucleus is much slower than that of GR-a, which could explain its accumulation in the nucleus.

MOLECULAR MECHANISMS OF GLUCOCORTICOID RESISTANCE The molecular basis of glucocorticoid resistance has been ascribed to mutations in the GR-a gene that impair one or more of the molecular mechanisms of GR function, thus altering tissue sensitivity to GCs. Inactivating mutations within the DNA- and ligandbinding domains, as well as a 4 bp mutation at the 30 boundary of the GR-a gene, have been described in five kindreds and three sporadic cases (Table I). In 1976, Vingerhoeds et al. described the first two patients, a father and a son, with long-term ‘‘hypercortisolism’’ not associated with clinical manifestations of Cushing’s syndrome. In 1982, Chrousos et al. demonstrated that these patients had abnormal glucocorticoid receptor properties and suggested that they suffered from glucocorticoid resistance. In 1991, this group described the molecular mechanism underlying the disease in this family: the propositus was a homozygote for a single nonconservative point mutation, replacing aspartate at amino acid position 641 with valine, in the hormone-binding domain of the GR. This mutation reduced glucocorticoid receptorbinding affinity for dexamethasone by threefold and caused loss of transactivation activity on the mouse mammary tumor virus (MMTV) promoter. The second family was described in 1993 by the Chrousos group as well; the proposita of this family was a young woman with manifestations of hyperandrogenism. Molecular analysis showed a 4 bp deletion at the 30 boundary of exon 6, removing a donor splice site. This resulted in complete ablation of expression of one of the GR alleles associated with a 50% decrease in GR protein in the affected members of the family.

Glucocorticoid Resistance Syndromes and States

f0010

237

Figure 2 Structure of the hGR gene and gene products. Alternative splicing events, No. 1 (default splicing pathway) and No. 2 (alternative splicing pathway), generate two different hGR messages, which differ in size due to the use of alternative polyadenylation signals. Translation of the messages produces two isoforms of the glucocorticoid receptor, hGR-a and hGR-b, the structures of which are identical through amino acid 727 but then diverge. hGR-a, 777 amino acids long, has a molecular weight of approximately 98 kDa, whereas hGR-b, 742 amino acids long, has a molecular weight of approximately 94 kDa. Functional domains and their putative sites for these functions are indicated below the linearized GR protein. Boxes and lines represent exons and introns, respectively.

The propositus of the third kindred, a boy with precocious puberty, had a single homozygotic point mutation at amino acid 729 (valine to isoleucine) in the hormone-binding domain, which reduced both the affinity and the transactivation activity of the GR. In 1996, another interesting, sporadic case of a man with a history of infertility, hypertension, and 5- to 10-fold elevation of urinary free-cortisol levels was described by the Chrousos group. This patient had a de novo, germ-line heterozygotic mutation at amino acid 559 (isoleucine to asparagine) in the hormonebinding domain, at the hinge region of the GR. This receptor had a 2- to 3-fold more potent dominantnegative activity than GR-b on the wild-type receptor

and was expressed at a 1:1 ratio with the normal GR-a in the patient’s cells. Interestingly, this mutant receptor prevented translocation of the normal receptor into the nucleus, an effect that would be overcome at very high dexamethasone concentrations. Later this patient developed severe Cushing’s disease due to an ACTH-secreting pituitary adenoma. A fifth case/kindred with glucocorticoid resistance was studied; the proband was a young woman with signs of hyperandrogenism but no other complaint. Molecular analysis detected a heterozygotic T-to-G substitution at nucleotide 2373 of exon 9a of the GR, in the ligand-binding domain at amino acid position 747, replacing isoleucine with methionine. This

238

Glucocorticoid Resistance Syndromes and States

Figure 3 (A) Putative mechanisms of action of hGR-a and hGR-b. In the unliganded state, the classic glucocorticoid receptor-a resides predominantly in the cytoplasm as part of a heteromeric complex including at least five molecules of heat shock proteins. After binding to the hormone, the GR-a molecule changes its conformation and is released by the heat shock proteins. It thus homodimerizes with another hormone-activated receptor molecule and translocates to the cell nucleus, where it can regulate the transcriptional activity of target genes. hGR-b is unable to bind glucocorticoids and is transcriptionally inactive, but may have a cell-specific dominant-negative effect on GR-a, primarily on GRE-mediated actions. The GRE consensus sequence (15 bp) is shown. (B) Nuclear actions of GR-a and GR-b. After entering the nucleus, ligand-bound GR-a influences the transcription of target genes by different mechanisms: as a homodimer, it may activate transcription by binding to GRE; as a heterodimer with GR-b, it may have a diminished ability to transactivate a GRE-containing gene and, hence, may act as a dominant-negative inhibitor. Binding of a GR-a homodimer to a negative GRE (nGRE) may also lead to repression; a GR-a–GR-b heterodimer may lose the ability to repress a nGRE. Monomeric GR-a interacts with other transcription factors, such as AP-1 and NF-kB, and prevents them from carrying out their activities, as their target genes contain AP-1 and NF-kB sites, respectively. In a protein– protein interaction similar to that occurring between GR-a and Stat-5, GR-b acts synergistically with Stat-5 in the activation of Stat5responsive genes. There is no evidence that GR-b interacts with or influences the activity of AP-1 or NF-kB.

239

Glucocorticoid Resistance Syndromes and States

t0005

Table I Glucocorticoid Resistance: Molecular Findings cDNA position 2054 (exon 7) 2317 (exon 9)

Amino acid position

Mutation

Domain

A-to-T mutation (nonconservative) (new site for HaeII)

Hormone binding

Hurley et al. (1991)

D!V 729

G-to-A mutation

Hormone binding

Malchoff et al. (1993)

A-to-G mutation (conservative)

Amino-terminal

Karl et al. (1993)

4 bp deletion (exon 6–intron 6)

Deletion ¼ no splice in 1 allele (segregating)

Hormone binding

Karl et al. (1993)

559

T-to-A mutation

Hormone binding

Karl et al. (1996)

641

Reference

V!I 1220 (exon 2)

363 A!S (conservative change)

1808 (exon 5)

I!N 1430 (exon 4)

477

G-to-A mutation

DNA binding

Ruiz et al. (2001)

2035 (exon 8)

R!H 679

G-to-A mutation

Hormone binding

Ruiz et al. (2001)

C-to-T mutation

Hormone binding

Mendonca et al. (2002)

T-to-G mutation

Hormone binding

Vottero et al. (2002)

G!S 1844 (exon 5)

571 V!A

2373 (exon 9)

747 I!M

mutation was located close to helix 12 at the Cterminus of the ligand-binding domain, which plays a pivotal role in the formation of activation function2, a subdomain that interacts with p160 coactivators. The mutant receptor had an approximately 2-fold reduced affinity for dexamethasone and its transcriptional activity on the glucocorticoid-responsive MMTV promoter was compromised by 20- to 30fold; interestingly, it also had dominant-negative activity on the wild-type receptor, probably secondary to the defective interaction of the mutant receptor with p160 coactivators. Two novel mutations in the GR gene were found in two unrelated patients with primary cortisol resistance as defined by a pathologic dexamethasone suppression test: R477H in the DNA-binding domain, which is the first reported mutation in that region of the human GR gene, and G679S in the ligand-binding domain. The R477H mutation showed no transactivating capacity, whereas the G679S mutation had reduced transactivation capacity and 50% binding affinity compared to the wild-type GR. A new phenotype, female pseudo-hermaphroditism and severe hypokalemia, caused by a novel inactivating mutation of the GR gene has been described. A homozygous T-to-C substitution at nucleotide 1844 in exon 5 of the GR gene, which caused a valine to alanine substitution at amino acid 571 in the ligand-binding domain of the receptor, was identified

in the patient. This phenotype indicated that both pre- and postnatal virilization can occur in females with the glucocorticoid resistance syndrome. The molecular mechanisms of action of some of the natural glucocorticoid receptor mutants causing familial glucocorticoid resistance have been studied. In particular, each of the mutations analyzed imparted different functional defects on the GR signal transduction pathway, which might partly explain the variable clinical phenotype of generalized glucocorticoid resistance. This variability includes patients with no abnormalities in the coding sequences and intron–exon boundaries of their GR genes. The importance of GR-b under physiologic conditions is controversial, but it has been proposed that in pathologic situations, such as glucocorticoid resistance, it might play a pathophysiologic role. Taking into consideration the finding that increasing amounts of GR-b produce a dose-dependent decrease in wildtype GR-a transcriptional activity, an imbalance in the expression of these two isoforms might determine an altered sensitivity to glucocorticoids. Supporting a possible role of GR-b in glucocorticoid sensitivity, a genetically determined imbalance in the expression of the glucocorticoid receptor isoforms was observed in cultured lymphocytes from a patient with congenital generalized glucocorticoid resistance and chronic leukemia; in this patient, a low GR-a to GR-b ratio was found compared to a group of normal controls,

240 possibly explaining the glucocorticoid resistance since no abnormalities in the sequence of the entire cDNA or in individual exons of this patient’s gene were found. In keeping with these findings, a significantly higher number of GR-b immunoreactive cells were found in peripheral blood and bronchial lavage cells from glucocorticoid-resistant asthma type 1 patients than from glucocorticoid-sensitive asthmatic patients and normal controls. Animal models of systemic glucocorticoid resistance, such as New World primates, including squirrel monkeys, marmosets, and owl monkeys, have been described. These animals have total plasma cortisol levels that are 7–20 times higher than in humans or other Old World primates, whereas the concentration, affinity, and predicted amino acid sequence of their GRs are similar to those of the human receptor. Interestingly, these animals exhibit resistance to a variety of other steroid/sterol hormones, including estrogens, progesterone, androgens, aldosterone, and vitamin D. Immunoreactivity of both isoforms of the GR has been found in Epstein-Barr virus-transformed B lymphocytes from marmosets, with the b-isoform being 10 times overexpressed compared to the corresponding human cells. An altered splicing pattern of the GR pre-mRNA or differential rates of mRNA translation, mRNA degradation, and/or GR protein degradation might contribute to the steroid resistance of these animals. Alternatively, these animals may have decreased coactivator activity and/or increased corepressor activity, leading to their ‘‘pansteroid’’ resistance. Two sisters with manifestations of glucocorticoid resistance were described by New et al. Their evaluation revealed resistance not only to glucocorticoids, but also to mineralocorticoids and androgens; however, they displayed no resistance to vitamin D or thyroid hormones. The diagnosis of these patients was multiple, partial steroid resistance. The New World primate physiologic and biochemical syndrome and the two pathologic human multiple steroid resistance syndrome cases are the first conditions in which a defective steroid receptor coregulator has been suggested to be responsible for an altered clinical and/or biochemical picture.

CLINICAL CHARACTERISTICS OF GLUCOCORTICOID RESISTANCE Patients with glucocorticoid resistance syndrome have compensatory elevations in circulating cortisol and ACTH concentrations, which maintain circadian

Glucocorticoid Resistance Syndromes and States

rhythmicity and appropriate responsiveness to stressors, albeit at higher hormone concentrations, and resistance of the HPA axis to dexamethasone suppression, but no clinical evidence of hypo- or hypercortisolism. The excess of ACTH results in increased production of adrenal steroids with mineralocorticoid and/or androgenic activity. The clinical spectrum of this disease is quite broad, ranging from completely asymptomatic to mild to severe symptomatic conditions. A large number of subjects may be asymptomatic or display biochemical alterations only. Clinical manifestations due to the excess of mineralocorticoids (including cortisol itself), acting on the intact mineralocorticoid receptor of the patients, are hypertension with or without hypokalemic alkalosis. The increased amounts of androgens, on the other hand, lead to acne, hirsutism, male pattern baldness, menstrual irregularities (oligoamenorrhea), oligo-anovulation, and infertility in women. In children, early and excessive prepubertal adrenal androgen secretion has been associated with ambiguous genitalia and precocious puberty. In adult men, oligospermia and infertility have been observed, possibly as the result of interference with folliclestimulating hormone feedback regulation by the excessive adrenal androgens or by the ACTH-induced intratesticular growth of adrenal rests, which may occur as they do in classic and ‘‘late-onset’’ congenital adrenal hyperplasia. Because of the excessive secretion of adrenal androgens, bone mass density is usually high-normal to elevated in patients with glucocorticoid resistance, in contrast to patients with Cushing’s syndrome, in whom osteoporosis is observed.

DIAGNOSIS The hallmark of the diagnostic evaluation of glucocorticoid resistance is increased serum cortisol and urinary free-cortisol levels without Cushing’s syndrome clinical stigmata. Despite cortisol excess, plasma ACTH concentration is normal or high. The circadian rhythm of cortisol and its responsiveness to stress are intact in patients with glucocorticoid resistance who are also resistant to single or multiple doses of dexamethasone. Differential diagnosis includes (1) early or mild forms of Cushing’s syndrome; (2) pseudo-Cushing states, such as generalized anxiety disorder, melancholic depression, and/or chronic active alcoholism, conditions associated with obesity and increased cortisol secretion; (3) other causes of mineralocorticoid-induced hypertension; and (4) hirsutism (including idiopathic hirsutism), polycystic

Glucocorticoid Resistance Syndromes and States

ovary syndrome, and late-onset congenital adrenal hyperplasia.

THERAPY OF GLUCOCORTICOID RESISTANCE SYNDROME Asymptomatic, normotensive subjects with primary glucocorticoid resistance do not require any treatment. In contrast, patients with symptomatic generalized glucocorticoid resistance are treated with high, individualized doses of oral dexamethasone (0.5–1.0 mg two or three times daily), a synthetic, potent glucocorticoid with minimal intrinsic mineralocorticoid activity. The goal is to suppress ACTH and, therefore, endogenous cortisol, DOC, corticosterone, and adrenal androgen secretion, correcting the states of mineralocorticoid and androgen excess in these patients. Hypertensive patients should receive the smallest dose that lowers the serum concentration of cortisol and other mineralocorticoids and corrects electrolyte abnormalities. Hirsute patients should be treated with doses able to reduce the androgen excess. Untreated patients have no risk of adrenal insufficiency and do not need extra doses of dexamethasone in particularly stressful situations, such as surgery and illness. In contrast, patients undergoing chronic treatment should receive the appropriate glucocorticoid coverage.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Glucocorticoid Receptor . Glucocorticoids and Immunity . Glucocorticoids in Aging: Relevance to Cognition . Glucocorticoids, Overview . Growth and Glucocorticoids . Hypercorticolism and Cushing’s Syndrome

Further Reading Arai, K., and Chrousos, G. P. (1995). Syndromes of glucocorticoid and mineralocorticoid resistance. Steroids 60, 173–179. Bronnegard, M., Werner, S., and Gustafsson, J. A. (1986). Primary cortisol resistance associated with a thermolabile glucocorticoid receptor in a patient with fatigue as the only symptom. J. Clin. Invest. 78, 1270–1278. de Castro, M., and Chrousos, G. P. (1997). Glucocorticoid resistance. Curr. Ther. Endocrinol. Metab. 6, 188–189.

241 Hurley, D. M., Accili, D., Stratakis, C. A., Karl, M., Vamvakopoulos, N., Rorer, E., Constantine, K., Taylor, S. I., and Chrousos, G. P. (1991). Point mutation causing a single amino acid substitution in the hormone binding domain of the glucocorticoid receptor in familial glucocorticoid resistance. J. Clin. Invest. 87, 680–686. Iida, S., Gomi, M., Moriwaki, K., Itoh, Y., Hirobe, K., Matsuzawa, Y., Katagiri, S., Yonezawa, T., and Tarui, S. (1985). Primary cortisol resistance accompanied by a reduction in glucocorticoid receptors in two members of the same family. J. Clin. Endocrinol. Metab. 60, 967–971. Jenster, G. (1998). Coactivators and corepressors as mediators of nuclear receptor function: An update. Mol. Cell. Endocrinol. 143, 1–7. Karl, M., Lamberts, S. W., Detera-Wadleigh, S. D., Encio, I. J., Stratakis, C. A., Hurley, D. M., Accili, D., and Chrousos, G. P. (1993). Familial glucocorticoid resistance caused by a splice site deletion in the human glucocorticoid receptor gene. J. Clin. Endocrinol. Metab. 76, 683–689. Karl, M., Lamberts, S. W., Koper, J. W., Katz, D. A., Huizenga, N. E., Kino, T., Haddad, B. R., Hughes, M. R., and Chrousos, G. P. (1996). Cushing’s disease preceded by generalized glucocorticoid resistance: Clinical consequences of a novel, dominant-negative glucocorticoid receptor mutation. Proc. Assoc. Am. Physicians 108, 296–307. Kino, T., Stauber, R. H., Resau, J. H., Pavlakis, G. N., and Chrousos, G. P. (2001). Pathologic human GR mutant has a transdominant negative effect on the wild-type GR by inhibiting its translocation into the nucleus: Importance of the ligand-binding domain for intracellular GR trafficking. J. Clin. Endocrinol. Metab. 86, 5600–5608. Lamberts, S. W., Koper, J. W., Biemond, P., den Holder, F. H., and de Jong, F. H. (1992). Cortisol receptor resistance: The variability of its clinical presentation and response to treatment. J. Clin. Endocrinol. Metab. 74, 313–321. Malchoff, C. D., Javier, E. C., Malchoff, D. M., Martin, T., Rogol, A., Brandon, D., Loriaux, D. L., and Reardon, G. E. (1990). Primary cortisol resistance presenting as isosexual precocity. J. Clin. Endocrinol. Metab. 70, 503–507. Malchoff, D. M., Brufsky, A., Reardon, G., McDermott, P., Javier, E. C., Bergh, C. H., Rowe, D., and Malchoff, C. D. (1993). A mutation of the glucocorticoid receptor in primary cortisol resistance. J. Clin. Invest. 91, 1918–1925. Mendonca, B. B., Leite, M. V., De Castro, M., Kino, T., Elias, L. L. K., Bachega, T. A. S., Arnhold, I. J. P., Chrousos, G. P., Lewicka, S., and Latronico, A. C. (2002). Female pseudohermaphroditism caused by a novel homozygous missense mutation of the GR gene. J. Clin. Endocrinol. Metab. 87, 1805–1809. Nawata, H., Sekiya, K., Higuchi, K., Kato, K., and Ibayashi, H. (1987). Decreased deoxyribonucleic acid binding of glucocorticoid-receptor complex in cultured skin fibroblasts from a patient with the glucocorticoid resistance syndrome. J. Clin. Endocrinol. Metab. 65, 219–226. Ruiz, M., Lind, U., Gafvels, M., Eggertsen, G., Carlstedt-Duke, J., Nillson, L., Holtmann, M., Stierna, P., Wikstrom, A. C., and Werner, S. (2001). Characterization of two novel mutations in the glucocorticoid receptor gene in patients with primary cortisol resistance. Clin. Endocrinol. 55, 363–371.

Glucocorticoids and Immunity

administration. Also, glucocorticoids decrease in a timely manner the expression of l-selectin on bone marrow progenitors and on differentiating neutrophils that, then, reach the bloodstream with low l-selectin levels. Chemokines released from the inflamed tissue are important for activating neutrophils and initiating the program of cell transmigration. The inhibition of the release of chemokines, such as interleukin-8 (IL-8) and other cysteine-X-cysteine (CXC) chemokines, also participates in the inhibition of neutrophil transmigration by glucocorticoids. Eosinophils, basophils, and T helper 2 (Th2) cells are recruited by chemokines, which are released by immune and nonimmune cells, such as epithelial cells or airway smooth muscle cells. These b-chemokines, or cysteine-cysteine (CC) chemokines, enhance adhesion molecule expression on endothelial cells to allow eosinophils and basophils to roll through the vessel wall. In fact, glucocorticoids prevent tissue invasion of eosinophils by inhibiting the release of CC chemokines, such as eotaxin, eotaxin 2, MCP4 (macrophage inflammatory protein 4), and RANTES (regulated on activated normal T-cell expressed and secreted), released by bronchial epithelial cells. Simultaneously, whereas glucocorticoids decrease the expression of some chemokine receptors, they enhance the expression of others on eosinophils, further complicating the overall picture of eosinophil trafficking. The flow and movement of monocytes seem tightly regulated by glucocorticoids via similar mechanisms, namely, through the repression of monocytes and endothelial adhesion molecules and the regulation of chemokines and their receptors.

Innate Immune Response Glucocorticoids act on the immune system by both suppressing and stimulating a large number of proinflammatory or anti-inflammatory mediators. In many ways, glucocorticoids lead to the termination of inflammation by enhancing opsonization and the activity of scavenger systems and by stimulating macrophage phagocytotic ability and antigen uptake. Glucocorticoids stimulate the expression of the mannose receptor or the scavenger receptor CD163, promoting the clearance of microorganisms, dead cell bodies, and antigens. Moreover, they potentiate interferon-g (IFN-g)-induced FcgRI. At the same time, they prevent inflammation from overshooting, by suppressing the synthesis of many inflammatory mediators, such as several cytokines and chemokines, prostaglandins, leukotrienes, proteolytic enzymes, free oxygen radicals, and nitric oxide.

243 A large number of cytokines [including IL-1b, tumor necrosis factor a (TNFa), IL-6, IL-8, IL-12, and IL-18] are broadly down-regulated by glucocorticoids. Similarly, the secretion of many CC and CXC chemokines is strongly suppressed. Interestingly, antiinflammatory cytokines such as IL-10 and transforming growth factor-b (and to some extent, although it is a matter of some debate, IL-1RA, the IL1 receptor antagonist) are up-regulated by glucocorticoids. Although the negative regulation of pro-inflammatory cytokines by glucocorticoids has been clearly demonstrated both in vitro and in vivo, glucocorticoids enhance the receptor of some of these pro-inflammatory cytokines and chemokines. It is tempting to speculate that they increase receptor expression to improve the sensitivity of target immune cells to these mediators and help these cells to resolve inflammation. Soluble or decoy receptors inhibiting or further enhancing the inflammatory process are also regulated by glucocorticoids. For example, the decoy receptor IL-1RII, which binds IL-1 without driving its signaling, is enhanced by glucocorticoids. This represents an anti-inflammatory mechanism of action of glucocorticoids. Inversely, the soluble IL-6R is increased by glucocorticoids and further amplifies the pro- or anti-inflammatory response to interleukin-6. An intriguing inflammatory mediator is the macrophage inhibitory factor (MIF), whose secretion is triggered by glucocorticoids. Synthesized by anterior pituitary cells and macrophages in response to endotoxin challenge and pro-inflammatory cytokines such as TNFa, MIF exerts a major pro-inflammatory effect on macrophages and T cells and overrides the antiinflammatory and immunosuppressive actions of glucocorticoids. In fact, MIF increases lethality, whereas genetic deletion or therapeutic neutralization confers protection against endotoxemia, acute distress respiratory syndrome, and septic shock. The finding that Toll-like receptor (TLR) 4 expression is increased by MIF underscores the essential role of MIF in the macrophage response to endotoxins and gram-negative bacteria. This indeed may explain why MIF-deficient mice were hyporesponsive to lipopolysaccharide (LPS). Thus, the notion that glucocorticoids enhance the secretion of a major pro-inflammatory cytokine that counteracts their effects represents a yin/yang mechanism of control of the acute-phase response or septic shock. By inhibiting phospholipase A2, inducible cyclooxygenase 2 (COX2), and inducible prostaglandin (PG) synthase 2 (PGS2), glucocorticoids block the release of arachidonic acid and the synthesis of prostaglandins (PGE2, PGH2, and PGD2) and platelet-activating factor; both constitutive COX1 and PGS1 remain

p0030

p0040

244 unaffected by glucocorticoids. Several lines of evidence suggest that glucocorticoids prevent COX2 expression through both posttranscriptional and posttranslational mechanisms. Strikingly, glucocorticoids could enhance 5-lipooxygenase and 5-lipooxygenaseactivating protein expression in monocytes and eosinophils. This is rather surprising given the overall reduced secretion of leukotrienes after glucocorticoid exposure. It seems that the decreased production of leukotrienes is related to the inhibition of phospholipase A2 expression and activity by glucocorticoids. Glucocorticoids may also suppress the reduced form of nicotinamide adenine dinucleotide phosphate oxidase and superoxide dismutase expression and, hence, the production of free oxygen radicals. Nitric oxide appears to be an important intra- and intercellular signaling molecule in shaping the innate and adaptive immune response with both detrimental and protective effects. Glucocorticoids suppress inducible nitric oxide synthetase expression, which results in a decrease in nitric oxide release by endothelial cells. This inhibition also seems to be mediated by the glucocorticoid second-messenger lipocortin 1 and would prevent early endothelial cell-mediated inflammatory reaction.

ADAPTIVE IMMUNITY Antigen Presentation and Adaptive Immune Response

p0050

Dendritic cells (DCs) represent the crucial interplay between innate and adaptive immunity. On encountering microorganisms or antigens, tissue-resident DCs rapidly differentiate and migrate to secondary lymphoid organs. These immature DCs demonstrate increased antigen uptake ability and specialized antigen-processing machinery. Glucocorticoids potentiate this ‘‘immature’’ phenotype. They improve opsonization and the activity of scavenger systems and stimulate macrophage phagocytosis, pinocytotic ability, and antigen uptake. TLRs mediate patterns of microorganism antigen recognition in macrophages, immature DCs, and mature DCs. Interestingly, glucocorticoids modulate the expression of TLR4, the LPS receptor, emphasizing the crucial role of glucocorticoids early in the innate immune response. During their migration, DCs mature and express major histocompatibility complex (MHC) class II and costimulatory molecules to efficiently present the antigen, as professional APCs (antigen-presenting cells), to naive or memory T cells. Crosstalk between

Glucocorticoids and Immunity

T cells and dendritic cells through TCR (T-cell receptor)/MHC II-bound antigen, costimulatory molecules, and cytokines allows the development of a T cell immune response and T cell expansion or deletion. When exposed to glucocorticoids during maturation, these DCs have a decreased ability to present antigens and elicit a Tcell response, primarily because glucocorticoids prevent MHC class II up-regulation and the expression of the costimulatory molecules, such as B7.2 (CD86) and to some extent B7.1 (CD80), CD40, and the ICAM-1/LFA-1 (intercellular adhesion molecule-1/lymphocyte function associated antigen-1) complex. It is noteworthy that terminally differentiated DCs continue to express these molecules due to their relative resistance to glucocorticoids. The timing of exposure to glucocorticoids thus appears to be essential during dendritic cell maturation. Once a T cell immune response has flared up, glucocorticoids may modulate and interfere with the type of T cells involved. The differentiation of CD4þ T cells into T helper 1 (Th1) lymphocytes, which drive cellular immunity, or into Th2 lymphocytes, which drive humoral immunity, depends on the type of antigen encountered and the type of cytokines produced during antigen presentation. Indeed, glucocorticoids block IL-12 secretion by monocytes and DCs. Interleukin-12 is the link between innate and cellular immunity and is crucial for the development of the Th1-directed cellular immune response. On the other hand, glucocorticoids promote Th2 development by enhancing IL-10 secretion by macrophages and immature DCs. In fact, several studies clearly showed, in both mice and humans, that the presence of glucocorticoids during the primary immune response, associated with defective IL-12 production by macrophages or DCs, enhanced Th2 cytokine secretion and decreased Th1 cytokine secretion by CD4þ lymphocytes on secondary stimulation. This effect seemed IL-12-dependent since it was reversed by the addition of exogenous IL-12 to glucocorticoidtreated APCs during the primary stimulation. Moreover, the poor glucocorticoid regulation of costimulatory molecules in mature DCs ruled out any role for costimulation in this process. Terminally differentiated DCs have been classified into DC1 and DC2, depending on their cytokine secretion profile and their aptitude to force the differentiation of naive T cells into Th1 and Th2 mature lymphocytes. Thus, DC phenotypes seem to be twisted by glucocorticoids into a DC2 program that will ultimately generate a Th2 immune response and the secretion of Th2 cytokines.

Glucocorticoids and Immunity

Cellular Immune Response Interleukin-12 is required for Th1 lymphocyte differentiation and secretion of Th1 cytokines, such as IFN-g and TNFa. Alterations in IL-12 signaling, such as those observed in IL-12/ and IL-12R-deficient mice, are associated with a defective Th1 immune response. In fact, glucocorticoids not only block IL-12 secretion by monocytes/macrophages and DCs, they also suppress IL-12Rb1 and IL12Rb2 expression on T cells. They also interfere with IL-12 signaling and prevent IL-12-induced signal transducer and activator of transcription 4 (Stat4) phosphorylation and Stat4-dependent gene expression such as that for interferon regulatory factor-1. Blocking Stat4 activation mimics the situation in Stat4-deficient mice, which are unable to elicit a Th1 immune response. Glucocorticoids also profoundly suppress secretion of the Th1 cytokines IFN-g and TNFa, lessening NK and T cytotoxic effector functions. Such massive inhibition of the Th1 immune response by glucocorticoids led to severe cellular immunodeficiency and impaired defense against intracellular and opportunistic infections.

Humoral Immune Response Whereas glucocorticoids profoundly decrease the secretion of Th2 cytokines during primary stimulation, they promote Th2 differentiation during secondary stimulation. Independent of monocytes/macrophages and DCs, glucocorticoids prime naive T cells to Th2 commitment during secondary immune response. Glucocorticoids that are present during primary stimulation promote IL-10 secretion during secondary stimulation in both naive and memory T cells. Sequential contact with glucocorticoids during the primary and/or secondary immune response may thus influence the pattern of Th2 cytokines. This context-dependent action of glucocorticoids may explain some discrepancies reported in the literature on the regulation of Th2 cytokines by glucocorticoids. Very interestingly, by up-regulating IL-10 secretion by macrophage/DCs and Th2 cells, glucocorticoids may participate in the emergence of regulatory T cells, high-IL-10-producing T cells with major in vivo immunoregulatory properties as shown in experimental allergic encephalomyelitis. Following exposure to endogenous or exogenous glucocorticoids, a progressive shift takes place from a cellular Th1 immune response to a humoral Th2 immune response. A major question remains regarding

245 how glucocorticoids induce Th2 lymphocyte differentiation and the humoral immune response. First, it is well recognized that the lack of commitment in developing a Th1 immune response, namely, the absence of IL-12, is associated with Th2 development. Therefore, the profound suppression of Th1 differentiation by glucocorticoids may undoubtedly participate in Th2 immune response development. Second, glucocorticoids also have differential action on IL-12 and IL-4 signaling. Interleukin-12 and IL-4 activate Stat4 and Stat 6, respectively. Deletion of Stat4 or Stat6 in mice is associated with poor Th1 and Th2 immune responses, respectively. In fact, glucocorticoids block IL-12-mediated Stat4 activation without altering IL-4-induced Stat6 phosphorylation and, hence, help a Th2 immune response to develop. Finally, a growing number of transcription factors have been shown to play a critical role in the hierarchical control of Th1 and Th2 lymphocyte differentiation. Thus, by acting directly on either lymphocytes or DCs, endogenous glucocorticoid hypersecretion or administration of excessive amounts causes a progressive shift from a Th1-directed cellular immune response to a Th2-driven humoral immune response. It is undeniable that glucocorticoids favor a humoral immune response and antibody production. In vivo administration of glucocorticoids raises immunoglobulin E (IgE) serum levels in asthma or atopic patients. Despite the paucity of information, glucocorticoids modulate B cell development. They restrain B cell proliferation and early steps in B cell development but promote the generation of antibodysecreting plasma cells and the secretion of IgE and IgG4. Like DCs and effector T cells, B cells become resistant to the inhibitory actions of glucocorcticoids as they proceed through the different stages of differentiation and maturation. Surprisingly, the question of how these hormones influence B cell receptor signaling compared to TCR signaling has never been explored, underscoring the lack of interest in glucocorticoid-mediated actions on B cells. Interleukin-4 is the critical cytokine that induces Th2 differentiation and promotes B cell differentiation and IgE isotype switching. Interestingly, glucocorticoids act in synergy with IL-4 in B cell differentiation and isotype switching, leading to IgE-secreting B cells. Such IgE isotype switching is dependent on the CD40/CD40 ligand (CD40L) complex since it is not observed in Xlinked hyper-IgM (CD40L-deficient) patients. In fact, glucocorticoids up-regulate CD40L expression on B cells and this finding may explain their synergistic actions on IgE isotype switching.

246

Glucocorticoids and Immunity

Allergy-Mediated Immune Response and Inflammation

T CELL DEVELOPMENT AND HOMEOSTASIS

The fact that glucocorticoids are used in the treatment of atopy and asthma and promote humoral Th2 immune response and IgE secretion represents a disturbing and challenging paradox. In truth, glucocorticoids may enhance IgE secretion but they strongly suppress allergic inflammation and chemokine-driven tissue infiltration of eosinophils. Furthermore, it has been suggested that glucocorticoids would inhibit antigen-specific IgE production while raising total IgE levels. This may elucidate why IgE serum levels remained high in asthma patients in clinical remission during steroid treatment. Although they prime DCs and T cells for Th2 development on secondary stimulation, glucocorticoids inhibit Th2 cytokine secretion during primary antigen exposure. Despite an ongoing Th2 immune response, glucocorticoids can still prevent the potentially deleterious IgE-induced allergic immune response by mast cells. Indeed, they interfere with IgE receptor-mediated release of inflammatory mediators and deplete bronchial mucosa from resident mast cells. Suffice it to say, mucosal mast cells are the predominant effector cells orchestrating allergic inflammation. In vivo studies suggest that glucocorticoids deplete tissue mast cells by inhibiting essential survival factors such as IL-4 and fibroblast-derived stem cell factor. Cross-linking of their FceRI with specific IgE results in the degranulation of preformed inflammatory mediators such as histamine, proteases, cytokines, and lipid mediators responsible for the early phase of type I hypersensitivity manifestations. Interestingly, glucocorticoids block IgE-triggered degranulation of human mast cells. They also prevent the late phase of type 1 hypersensitivity characterized by mast cell activation with de novo synthesis of inflammatory mediators and secondary infiltration of Th2 cells, basophils, and esosinophils. Glucocorticoid-induced down-regulation of high-affinity FceRI and lowaffinity FceRI (CD23) expression may account for the poor response of IgE-triggered mast cell activation only in the late phase. Glucocorticoids also interfere with FceRI signaling through the disruption of raf-1/ heat shock protein 90 and the subsequent mitogenactivated phosphokinase activation and phospholipase A2 responsible for the de novo synthesis of arachidonic acid-derived metabolites. Finally, glucocorticoids inhibit cytokine synthesis of pro-inflammatory cytokines and chemokines by mast cells necessary for their own survival and the chemotaxis and expansion of eosinophils and basophils.

Though supra-pharmacologic doses of glucocorticoids induce T cell apoptosis, adrenal- or thymusderived glucocorticoids or physiologic doses of glucocorticoids can induce Tcell survival or apoptosis, depending on the cell type and differentiation stage. Moreover, both the degree of Tcell activation and the timing of glucocorticoid exposure (before, during, or after activation) render T cells sensitive or resistant to glucocorticoid-induced apoptosis. Several studies have shown that concomitant TCR signaling and glucocorticoid receptor (GR) signaling promote Tcell survival, whereas either TCR signaling alone or GR signaling alone induces Tcell apoptosis. Indeed, thymus-derived glucocorticoids appear to play a role in early thymocyte expansion and in central positive selection. Yet, GR/ knockout mice demonstrate a normal thymus, suggesting that glucocorticoid actions on TCR signaling could occur through nongenomic actions that are independent of GR transcriptional activity. Similarly, there is some evidence suggesting that glucocorticoids could influence peripheral T cell development and selection by simultaneously preventing TCR-induced T cell deletion and enhancing T cell survival. Interestingly, glucocorticoids enhance a key cytokine receptor for T cell development, IL-7Ra, whose deletion in mice and humans is associated with a lack of T cells. Moreover, IL-7 potently enhances thymic-independent peripheral expansion and restores immunity in athymic T cell-depleted hosts in mice. The positive regulation of IL-7Ra expression by glucocorticoids suggests their strong influence in the maintenance of peripheral Tcell pool homeostasis.

CONCLUDING REMARKS The description of the actions of glucocorticoids on the immune response elucidates their positive and negative effects on several components of the innate and adaptive immune responses. Glucocorticoidinduced immunomodulation requires both immunoenhancing and immunosuppressive actions at the same time and these should be integrated in a dynamic, ongoing process. Indeed, pro-inflammatory mediators participate, to the same extent, with antiinflammatory mediators in the so-called immunosuppressive actions of glucocorticoids. The great advantage of their clinical use in Th1-inflammatory and Th1-autoimmune diseases is obvious because they restrain the inflammatory reaction, prevent tissue

247

Glucocorticoids and Immunity

destruction, and block Th1-driven immune responses. Also, the fact that there are glucocorticoids that favor innate immunity and antibody responses may explain why they exhibit clinical benefit in bacterial infections such as in bacterial meningitis. Similarly, the fact that glucocorticoids enhance the phagocytotic ability of neutrophils and macrophages, boosting scavenging and opsonization systems, underscores the clinical benefit and importance of early and continuous administration of glucocorticoids in patients in septic shock. However, these beneficial effects coexist with deleterious adverse effects. Because of their suppressive actions on Th1 immunity, they increase susceptibility to intracellular and opportunistic infections such as tuberculosis or viral infections. Finally, glucocorticoids may worsen disease activity due to their selective immunoenhancing actions. This was hieratically reported in steroid-resistant asthma and ulcerative colitis. The overall knowledge of the actions of glucocorticoids on adaptive immunity clearly revealed that by decreasing the ability of DCs to elicit a Tcell immune response, by promoting Th2 differentiation and helping the development of regulatory T cells, glucocorticoids may induce tolerance. Therefore, glucocorticoids could be extremely helpful in cell therapy through their ex vivo action not only on DCs but also potentially on regulatory T cells. This ex vivo therapeutic use of glucocorticoids would be very beneficial, thanks to their immune modifier effect, and avoid their long-term deleterious adverse effects. Glucocorticoids could thus represent an adjuvant treatment for cell therapies in autoimmune diseases and organ transplantations.

See Also the Following Articles Aging, Immunology and . Corticotropin-Releasing Hormone (CRH) and Inflammation . Cytokine Actions, Cellular Mechanism of . Cytokine Receptors . Glucocorticoid Receptor . Glucocorticoids, Overview . Immune System, Hormonal Effects on

Further Reading Ashwell, J. D., Lu, F. W., and Vacchio, M. S. (2000). Glucocorticoids in T cell development and function. Annu. Rev. Immunol. 18, 309–345. Barnes, P. J. (2001). Corticosteroids, IgE, and atopy. J. Clin. Invest. 107, 265–266. Blotta, M. H., DeKruyff, R. H., and Umetsu, D. T. (1997). Corticosteroids inhibit IL-12 production in human monocytes and enhance their capacity to induce IL-4 synthesis in CD4þ lymphocytes. J. Immunol. 158, 5589–5595. Calandra, T., Bernhagen, J., Metz, C. N., Spiegel, L. A., Bacher, M., Donnelly, T., Cerami, A., and Bucala, R. (1995). MIF as a glucocorticoid-induced modulator of cytokine production. Nature 377, 68–71. Franchimont, D., Galon, J., Vacchio, M. S., Fan, S., Visconti, R., Frucht, D. M., Geenen, V., Chrousos, G. P., Ashwell, J. D., and O’Shea, J. J. (2002). Positive effects of glucocorticoids on T cell function by up-regulation of IL-7 receptor a. J. Immunol. 168, 2212–2218. Hogger, P., Dreier, J., Droste, A., Buck, F., and Sorg, C. (1998). Identification of the integral membrane protein RM3/1 on human monocytes as a glucocorticoid-inducible member of the scavenger receptor cysteine-rich family (CD163). J. Immunol. 161, 1883–1890. Jabara, H. H., Brodeur, S. R., and Geha, R. S. (2001). Glucocorticoids upregulate CD40 ligand expression and induce CD40Ldependent immunoglobulin isotype switching. J. Clin. Invest. 107, 371–378. Nakagawa, M., Bondy, G. P., Waisman, D., Minshall, D., Hogg, J. C., and van Eeden, S. F. (1999). The effect of glucocorticoids on the expression of L-selectin on polymorphonuclear leukocyte. Blood 93, 2730–2737. Piemonti, L., Monti, P., Allavena, P., Sironi, M., Soldini, L., Leone, B. E., Socci, C., and Di Carlo, V. (1999). Glucocorticoids affect human dendritic cell differentiation and maturation. J. Immunol. 162, 6473–6481. Strausbaugh, H. J., and Rosen, S. D. (2000). A potential role for annexin 1 as a physiologic mediator of glucocorticoid-induced l-selectin shedding from myeloid cells. J. Immunol. 166, 6294–6300. Van Laethem, F., Baus, E., Smyth, L. A., Andris, F., Bex, F., Urbain, J., Kioussis, D., and Leo, O. (2001). Glucocorticoids attenuate T cell receptor signaling. J. Exp. Med. 193, 803–814. Wu, C. Y., Sarfati, M., Heusser, C., Fournier, S., Rubio-Trujillo, M., Peleman, R., and Delespesse, G. (1991). Glucocorticoids increase the synthesis of immunoglobulin E by interleukin 4-stimulated human lymphocytes. J. Clin. Invest. 87, 870–877.

249

Glucocorticoids in Aging: Relevance to Cognition

REGULATION OF THE HYPOTHALAMIC–PITUITARY– ADRENAL CORTEX AXIS Glucocorticoids serve a critical role in adaptation to the changing environment. Their levels are invariably increased by subconscious alterations in homeostasis as well as by what we personally define as ‘‘stressful,’’ be it an external stimulus or an internal perception. On realization of the stressor, higher order neural circuits lift the inhibition imposed by the hippocampus on the hypothalamic–pituitary–adrenal cortex (HPA) axis. As a result, corticotropin-releasing hormone (CRH) is secreted from the paraventricular nucleus of the hypothalamus, which in turn stimulates release of adrenocorticotropin (ACTH) from the anterior pituitary and subsequently secretion of glucocorticoids (CORT: cortisol in humans; corticosterone in rodents) from the adrenal cortex. During its initial phase, this elevation of CORT helps to address the challenge by improving cognition through mineralocorticoid receptors (MRs) in the hippocampus and increasing the energy availability to the brain and skeletal muscles. Once the stressor has been properly addressed, the HPA axis is inhibited by negative feedback mainly through the glucocorticoid receptors (GRs) in the hippocampus, hypothalamus, and pituitary. After the CORT levels return to baseline, MRs in the hippocampus take over the control of regulating basal CORT levels.

AGE-RELATED ALTERATION IN CORTICOSTERONE PROFILE IN RODENTS p0025

Age-related changes in the human HPA axis are susceptible to a myriad of factors that can be generally divided into genetic differences, nonpsychological environmental factors, complex daily personal interactions, various life histories, and interactions among these factors that define our attitude toward life experiences. In contrast, most of these factors are well controlled in experiments examining aging in laboratory rodents. However, even here, the data have been highly variable. After analyzing the available data on age-related alterations in plasma CORT in several different rat strains, Sapolsky concluded that there was more than a twofold increase in CORTwith age. This has been corroborated by other more recent studies. A more descriptive picture of the age-related plasma CORT levels was gained from a longitudinal study in Fischer 344 rats by Sabatino and colleagues. They

found that the free CORT levels increased twofold between middle age and old age. Furthermore, the diurnal pattern showed an expansion of the circadian peak into the trough. This suggests that with progressing age, there is more activation of GRs and for a longer period. Another important result from this study was that when the rats were young, restraint stress-induced CORT levels started to decline soon after termination of the stressor, whereas in old age the stress-induced CORT levels stayed elevated for an additional 30 min prior to the decline. Because these animals were not chronically stressed, these agerelated differences arise due to mechanisms that are not well appreciated. Age also seems to affect the expression levels of the MR and GR. The most consistent observation is the age-related decline in MR mRNA and protein levels within the hippocampus. On the other hand, changes in hippocampal GR levels appear to be strain dependent, with age–related decreases in both mRNA and protein levels in some rat strains but not in others. Interestingly, decreases in hippocampal MRs and GRs are also observed in chronically stressed animals, and the extent of reduction in these receptors may be associated with the severity of stress. Thus, the age- and stress-related modulation of MR and GR expression has a serious impact on regulation of the HPA axis. In light of the importance of these receptors in the regulation of the HPA axis, it should be mentioned that life histories play an important role in expression of these receptors in the hippocampus. Meaney and colleagues reported that maternal affection (licking) during postnatal development in rodents increases GR expression and that these pups, during their adulthood, show an attenuated CORT response to stress, have a lower age-related increase in CORT, and are less cognitively impaired than their counterparts. In contrast, pups that were deprived of maternal care or infected with endotoxins show a decrease in their hippocampal receptor levels and show an accentuated CORT response to stressors during adulthood. Furthermore, elevated CORT levels during fetal development decrease the expression of these receptors and possibly predispose these animals to age–related increases in CORT. Thus, experiences over a lifetime should be considered as an important variable in age–related CORT elevations.

CORTICOSTERONE LEVELS, AGE, AND COGNITIVE IMPAIRMENT IN RATS As with age-related increases in plasma CORT levels, the within–group variability in cognitive performance

p0040

250 among the aged animals is large. In one study by Meaney and colleagues, 20% of the aged rats were cognitively unimpaired (AU) and performed as well as young controls, whereas 28% of the aged rats showed dramatic cognitive impairment (AI). Furthermore, these two aged groups differed significantly for several hippocampus–HPA parameters. The plasma CORT levels across the diurnal cycle in the AU rats were similar to those in the young rats with a single peak of elevated CORT levels at the beginning of the dark cycle, whereas the AI rats showed an expansion of the diurnal peak such that the body was exposed to elevated CORT levels for an extended period of time (12 h); this diurnal pattern is similar to that observed by Sabatino and colleagues mentioned previously. The AI rats also had elevated plasma ACTH levels at the beginning of the acrophase, and this was probably responsible for the extended period of elevated CORT in this group. The termination of stressinduced CORT response was also different in these two aged groups, such that the total CORT exposure (area under the curve) in the AI rats was much greater. Consistent with this profile, the hippocampal MR and GR levels in the AI animals were lower than those in the AU rats. The data from McEwen’s laboratory has clearly established that stress-related increases in CORT are associated with atrophy of the dendritic tree, such that the number of branching points is decreased. This is likely to yield a dramatic reduction in the number of synapses. This is supported by Jucker and colleagues’ observation that the number of synaptic boutons in the dentate gyrus region of the hippocampus is associated with performance in the Morris water maze. An added cellular complexity should also be mentioned. Plasma CORT levels are not necessarily the ‘‘true’’ indication of total cellular CORT exposure. Intracellular CORT levels are regulated by 11bhydroxysteroid dehydrogenase (11b-HSD). The type 1 isoform, 11b-HSD1, can convert inactive CORT metabolites (11-dehydrocorticosterone, cortisone) back to active CORT. Seckl’s group has shown that old mice lacking 11b-HSD1 perform as well as young controls in the Morris water maze, whereas old control mice take twice as long to find the hidden stage. Because the levels of the CORT metabolites are in excess of the free CORT levels, regulation of this activity can prove to be extremely beneficial in decreasing CORT-related cognitive impairment.

Glucocorticoids in Aging: Relevance to Cognition

AGE-RELATED ALTERATION IN THE CORTISOL PROFILE IN HUMANS Despite all of the previously mentioned concerns regarding examination of age-related changes in CORT in humans, it is important to assess whether the human HPA axis is susceptible to dysregulation with age. However, these concerns cannot be undermined and should be evaluated further because doing so may set forth hypotheses that are directly relevant to successful aging as well as to age-related disorders in which etiology remains unclear. For example, it is possible to subdivide healthy volunteers based on their anticipatory stress responses, dexamethasone suppression of CORT levels, and ACTH responsiveness. All of these parameters are quantitative, and if they are to be followed longitudinally, they might allow identification of progressive steps that increase susceptibility to age-related disorders. The data regarding age-related alteration in basal plasma CORT levels in humans are inconsistent. Because most of these studies have also evaluated cognitive functions, the data are discussed in more detail in the next section. However, an important difference in the experimental design between the rodent studies and the human studies should be emphasized here: in the human studies, unhealthy individuals are either excluded or considered as an independent variable, whereas the same level of discretion is hard to attain, and is often ignored, in rodent studies. Proinflammatory cytokines can activate the HPA axis and may contribute to the age-related cognitive impairment observed in rodents. This may explain some of the inconsistencies observed between the human data and the rodent data. The HPA axis in humans is not spared by aging. It is now widely accepted that elderly humans also show a delayed termination of the CORT response after a stressful situation. One possible explanation for this increase in latency is the ‘‘accumulative effect of stress’’ over the life span. Allostasis refers to constant alterations in the HPA axis that are necessary to maintain homeostasis in the new environment/situation. As the frequency of HPA activation and/or total exposure to CORT increases, so do the side effects of hypercortisolemia. The HPA axis is designed to minimize this consequence by a negative feedback mechanism, but as the ‘‘allostatic load’’ on the hippocampus increases, the regulation of HPA deteriorates and CORT levels stay elevated for longer than necessary. A subgroup of elderly persons also show elevated

p0060

251

Glucocorticoids in Aging: Relevance to Cognition

nighttime CORT levels, leading to the suggestion that pharmacological intervention to reduce nighttime CORT may have therapeutic value in these individuals.

COGNITIVE IMPAIRMENTS AND PLASMA CORTISOL LEVELS IN HUMANS

p0075

The human data do not show a direct relationship between age and cognition or between age and cortisol levels. However, elderly persons tend to have a greater risk of both cognitive impairment and elevated cortisol levels. In the MacArthur Foundation Study on Successful Aging, healthy elderly individuals residing in three different communities were followed for 212 years. During this period, 20% of the women showed a progressive increase in CORT, whereas 15% showed a decline in urinary CORT levels. In the subgroup with increasing CORT, 75% of the women showed poor performance on the delayed recall of a story task as compared with their performance at the start of the study. In contrast, in the subgroup with decreasing CORT, 70% of the women showed an improvement in their performance on this task. Similarly, Lupien and colleagues found that individuals with increasing CORT over a 5-year period performed poorly on the delayed memory and spatial memory tests as compared with individuals with stable/decreasing CORT levels. Furthermore, those with increasing CORT had a mean 15% decrease in the hippocampal volume. Hippocampal atrophy of a similar extent (>10%) is also seen in AD. More than half of the studies that have examined basal cortisol levels in AD reported an association between AD and increased plasma cortisol. Furthermore, the few studies that have examined the HPA axis overactivity by the dexamethasone suppression test found that a greater number of individuals with AD failed the test as compared with controls. Thus, the age-related risk of elevated CORT levels may also increase the risk of AD. A decrease in hippocampal volume is also observed in other disorders with hypercortisolemia: depression, posttraumatic stress disorder (PTSD), and Cushing’s syndrome. In the cases of PTSD and depression, it is possible that altered brain physiology could have confounding effects on this measure. However, in Cushing’s syndrome, hypercortisolemia occurs as a consequence of peripheral tumors that secrete excessive amounts of ACTH or CRH. Moreover, on surgical correction of hypercortisolemia in Cushing’s

syndrome, the hippocampal volume increases. These observations support the notion that cortisol leads to cognitive decline by damaging the hippocampus.

GLUCOCORTICOIDS’ EFFECTS ON LONG-TERM POTENTIATION The associative, specific, and relatively long-lasting characteristics of long-term potentiation (LTP) have led to its acceptance as an excellent model of the molecular mechanism underlying learning and memory. Similar synaptic efficacy can also be elicited by a physiologically patterned, lower threshold primedburst stimulation. In contrast, low-frequency stimulation of the synapse yields long-term depression (LTD) of the synapse. Both of these mechanisms rely on Ca2þ influx and the N-methyl-D-aspartate (NMDA glutamate analogue) receptor. A well-known target of NMDA-mediated Ca2þ influx is the calcium– calmodulin-dependent kinase II (CaMKII), which on activation auto-phosphorylates and becomes autonomously active to maintain the sensitivity of the synapse to future stimulation. Similar mechanisms are also employed during a low-frequency stimulation, but because the Ca2þ influx is gradual and of low intensity, it fails to induce the Ca2þ–mediated pathways that are activated during high-frequency stimulation. CORT shifts the balance between LTP and LTD by differential activation of the MR and GR. For example, treatment of adrenalectomized rats with an MR-specific agonist, aldosterone, enhanced LTP, whereas treatment with a GR-specific agonist, RU-28362, suppressed synaptic efficacy. Under basal conditions, preferential activation of MR in the hippocampus is associated with short-lived Ca2þ currents that promote LTP. On the other hand, activation of GR increases expression of the NMDA receptor subunit NR2B. This substitution in the NMDA receptor leads to enhanced Ca2þ influx. Because this alteration is at the receptor level, the effects are observed for a longer time and lead to increased Ca2þ influx even during low-frequency stimulation. Kim and Yoon proposed that the intracellular Ca2þ ([Ca2þ]i) has a metaplastic effect on synaptic efficacy. They suggested that if the [Ca2þ]i is high, the subsequent stimulus would have to be of much greater intensity to favor synaptic efficacy. If the stimulus is smaller than the previous stimulus, it will fail to induce Ca2þ–calmodulin pathways but will still elicit other effects of Ca2þ such as activation of phosphatases and Ca2þ-dependent Kþ channels. Activation of Kþ channels increases the afterhyperpolarization

p0085

252 current (refractory period) of the hippocampal neurrons. This would block the cells’ ability to respond to high-frequency stimuli, further decreasing the cells’ ability to shift back in favor of synaptic efficacy. Consistently, the aged rats have a propensity for LTD and are susceptible to reversal of LTP. In one study, the number of potentiations observed after a primed-burst stimulation in 24-month-old Fischer 344 rats was only 40% that in 6-month-old controls.

EFFECTS OF CHRONIC DISEASES ON GLUCOCORTICOIDS One of the earliest systemic responses to immune system activation is the release of interleukin-1 (IL-1), IL-6, and tumor necrosis factor-a (TNF-a). These cytokines trigger an acute phase response in other organs to alter their function during an infection. These cytokines also activate the HPA axis at several levels; all of these cytokines can also stimulate the HPA axis through the brain; IL-1 and TNF-a can increase CRH, all three increase secretion of ACTH, and IL-6 can also activate the adrenocortical cells directly to secrete CORT. A classic function attributed to CORT is as an anti-inflammatory agent, presumably to reduce ‘‘nonspecific’’ inflammation elsewhere in the body. CORT also increases expression of macrophage migration inhibitory factor (MIF) in various organs in a tissue- and time-dependent manner. MIF exhibits an anti-CORTeffect by reversing CORT’s effects on both apoptosis and cytokine expression in T lymphocytes and macrophages. This action may serve to overcome the anti-inflammatory effects of CORT within the local area of inflammation. However, in chronic illness, elevated levels of both MIF and CORT may lead to dysregulation by decreasing CORT’s effectiveness. IL-1, IL-6, and TNF-a have been associated with aging and/or age-related disorders and may also be responsible for the observed age-related elevation of CORT in a subgroup of elderly persons. Thus, individuals in preclinical stages of a disease could

Glucocorticoids in Aging: Relevance to Cognition

have elevated cytokines, elevated CORT, and (as a consequence) cognitive impairment. This is supported by the observation that cognitive impairment is associated with mortality. Furthermore, most of the commonly used rat strains develop multiple lesions with age. Thus, the age-related increase in CORT levels in rodents may be indicative of disease pathology. In conclusion, age-related increases in CORT might not be a consequence of the aging process but rather might be a consequence of disease, and this elevation of CORT leads to cognitive impairment by damaging the hippocampus.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Aging and Longevity of Human Populations . Alzheimer’s Disease and Hormones . Brain, Effects of Steroid Hormones . Functional Genomics of Aging . Glucocorticoids, Overview . Neuroendocrine System and Aging . Stress, Aging, and Central Nervous System Interactions

Further Reading Harvey, P. D., and Mohs, R. C. (2001). Memory changes with aging and dementia. In ‘‘Functional Neurobiology of Aging’’ (P. R. Hof and C. V. Mobs, eds.), pp. 53–63. Academic Press, San Diego. Lucassen, P. J., and De Kloet, E. R. (2001). Glucocorticoids and the aging brain: Cause or consequence? In ‘‘Functional Neurobiology of Aging’’ (P. R. Hof and C. V. Mobs eds.), pp. 883–905. Academic Press, San Diego. Lupien, S. J., Nair, N. P., Briere, S., Maheu, F., Tu, M. T., Lemay, M., McEwen, B. S., and Meaney, M. J. (1999). Increased cortisol levels and impaired cognition in human aging: Implication for depression and dementia in later life. Rev. Neurosci. 10, 117–139. Patel, N. V., and Finch, C. E. (2002). The glucocorticoid paradox of caloric restriction in slowing brain aging. Neurobiol. Aging 23, 707–717. Sapolsky, R. M. (1999). Glucocorticoids, stress, and their adverse neurological effects: Relevance to aging. Exp. Gerontol. 34, 721–732. Turnbull, A. V., and Rivier, C. L. (1999). Regulation of the hypothalamic–pituitary–adrenal axis by cytokines: Actions and mechanisms of action. Physiol. Rev. 79, 1–71.

254

Glucocorticoids, Overview

11-deoxycortisol, which is further converted to cortisol by CYP11B1 (P450c11, 11b-hydroxylase) in the mitochondria. In the zona reticularis of the adrenal cortex and in the gonads, the 17,20-lyase activity of CYP17 converts 17a-hydroxypregnenolone to dehydroepiandrosterone (DHEA), a C-19 steroid and sex steroid precursor. DHEA is further converted by 3b-HSD to androstenedione. In the gonads, androstenedione is reduced by 17b-hydroxysteroid dehydrogenase. In pubertal ovaries, aromatase (CYP19, P450c19) can convert androstenedione and testosterone to estrone and estradiol, respectively. Testosterone may be further metabolized to dihydrotestosterone by steroid 5a-reductase in androgen target tissues.

deoxycorticosterone (DOC). Aldosterone, the most potent 17-deoxysteroid with mineralocorticoid activity, is produced by 11b-hydroxylation of DOC to corticosterone, followed by 18-hydroxylation and 18-oxidation of corticosterone (Fig. 1). The final three steps in aldosterone synthesis are accomplished by a single mitochondrial P450 enzyme, CYP11B2 (P450aldo, aldosterone synthase). To produce cortisol, CYP17 (P450c17, 17a-hydroxylase/17,20-lyase) in the endoplasmic reticulum of the zona fasciculata and zona reticularis converts pregnenolone to 17a-hydroxypregnenolone. 3b-HSD in the zona fasciculata utilizes 17a-hydroxypregnenolone as a substrate, producing 17a-hydroxyprogesterone. The latter is 21-hydroxylated by CYP21 to form

Cholesterol StAR protein scc 17α-Hydroxylase*

Pregnenolone

→

17-OH Pregnenolone → DHEA

3b-HSD*

3b-HSD* 17a -Hydroxylase*

Progesterone

→

17b -HSD

17,20-Lyase

21-Hydroxylase*

Androstenediol

3b-HSD*

→

21-Hydroxylase*

3b-HSD*

17b-HSD

17,20-Lyase

17-OH Progesterone

→

A

Testosterone

→

Aromatase

Aromatase

17b -HSD

Deoxycorticosterone

11b -Hydroxylase*

Corticosterone

11-Deoxycortisol

Estrone

Estradiol

11b -Hydroxylase*

Cortisol

Present in adrenal & gonadal tissue

18-Hydroxylase

18-OH Corticosterone

18-OH dehydrogenase

Aldosterone f0005

Figure 1 Schematic representation of adrenal steroidogenesis. Solid lines indicate major pathways. Dotted lines indicate major pathways in ovaries and minor pathways in adrenals. Asterisks indicate that deficient enzymatic activity results in congenital adrenal hyperplasia (CAH). StAR, steroidogenic acute regulatory protein; scc, cholesterol side-chain cleavage enzyme; 3b-HSD, 3b-hydroxysteroid dehydrogenase; 17b-HSD, 17b-hydroxysteroid dehydrogenase; DHEA, dehydroepiandrosterone; A, androstenedione.

255

Glucocorticoids, Overview

REGULATION OF CORTISOL SECRETION

p0060

Plasma glucocorticoid concentrations are regulated in ways that reflect the varying physiologic needs for the hormones under basal conditions and in response to stress. Cortisol secretion is primarily regulated by ACTH, a 39-amino-acid peptide released by the anterior pituitary. ACTH is synthesized as part of a higher molecular weight precursor peptide, proopiomelanocortin. ACTH is secreted in regular pulses of variable amplitude over 24 h, with peak concentrations attained in the early morning hours (4:00–8:00 am), thus forming the basis of the circadian pattern of cortisol secretion. The acute action of ACTH is to increase the flux of cholesterol through the steroidogenic pathway, resulting in the rapid production of steroids. ACTH also influences the remaining steps of steroidogenesis as well as the uptake of cholesterol from plasma lipoproteins, thus ensuring a continuous supply of cholesterol to the mitochondria to meet the demands of activated pregnenolone biosynthesis. It also maintains the size of the adrenal glands, stimulates melanocytes, and results in hyperpigmentation when secreted in excess. Corticotropin-releasing hormone (CRH) is the principal hypothalamic factor that stimulates the pituitary production of ACTH. It is produced in the paraventricular nuclei of the hypothalamus, but is also found in other parts of the central nervous system as well as in noncentral locations. CRH is secreted in a pulsatile fashion that results in the episodic secretion of ACTH and the circadian variation of cortisol secretion. The magnitude of cortisol response to each ACTH burst remains relatively constant; it is therefore the number of secretory periods, rather than the magnitude of each pulse of CRH or ACTH, that determines the total daily cortisol secretion. In addition to CRH, vasopressin, a peptide product of the posterior pituitary gland, stimulates ACTH release by acting synergistically with CRH. Although CRH increases the amount of ACTH secreted from each responsive corticotroph, vasopressin appears to increase the number of CRH-responsive corticotrophs. In addition to ACTH, other factors may play an important role in the regulation of the adrenal cortex. Cortisol is the primary negative regulator of basal hypothalamic–pituitary–adrenal (HPA) axis activity through negative feedback on ACTH and CRH secretion. The negative feedback effects of cortisol are exerted at the level of both the hypothalamus and the pituitary and are mediated by type II corticosteroid

receptors. Whether and to what extent direct glucocorticoid feedback on the adrenal cortex itself regulates cortisol synthesis is not clear.

SECRETION AND METABOLISM In normal subjects, the secretion of glucocorticoids follows a diurnal pattern, with peak concentrations observed between 6:00 and 8:00 am and the lowest concentrations observed at approximately 12:00 am. The cortisol production rate is approximately 12 mg/m2/day. More than 90% of circulating cortisol, and to a lesser extent aldosterone, is bound tightly to corticosteroid-binding globulin (CBG) or transcortin. The remaining (10%) of the circulating cortisol is free or loosely bound to albumin. The free and albuminbound fractions of cortisol represent the biologically active form of the hormone. When plasma cortisol concentrations exceed 20 mg/dl, CBG becomes fully saturated and most of the excess cortisol is biologically active. CBG is synthesized in the liver. Estrogens, thyroid hormones, pregnancy, and oral contraceptives are associated with increased CBG concentrations, whereas hypercortisolism, hepatic disease, or renal disease results in decreased CBG concentrations. In the presence of an intact HPA axis, alterations in CBG concentrations are likely not to affect circulating free cortisol concentrations. The primary site of cortisol metabolism in humans is the liver, and a number of cytosolic and microsomal enzymes, including cytochrome P450, 5a/5b-reductase, 3a/3b-oxidoreductase, and 11bhydroxysteroid dehydrogenase, play an important role in the hepatic metabolism of cortisol. The major routes of hepatic metabolism involve A-ring and side-chain reduction followed by conjugation with glucuronic acid and sulfate. The inactive glucuronide and sulfate metabolites are excreted by the kidneys, whereas less than 1% of cortisol is excreted unchanged in the urine. The metabolic clearance of cortisol, therefore, is influenced primarily by factors altering hepatic clearance and to a much lesser degree by factors affecting renal excretion.

MECHANISMS OF GLUCOCORTICOID ACTION At the cellular level, the actions of glucocorticoids are mediated by an intracellular receptor protein, the glucocorticoid receptor, which functions as a hormoneactivated transcription factor that regulates the expression of glucocorticoid target genes.

256

TREATMENT WITH GLUCOCORTICOIDS Natural and synthetic glucocorticoids can be used for both endocrine and nonendocrine disorders. In clinical practice, glucocorticoids are used to establish the diagnosis and cause of Cushing’s syndrome and in the treatment of adrenal insufficiency and congenital adrenal hyperplasia. Glucocorticoids are also given in pharmacologic doses to treat patients with inflammatory, allergic, or immunologic disorders. Chronic therapy has many side effects, ranging from suppression of the HPA axis and Cushing’s syndrome to infections and changes in mental status. Factors that influence both the therapeutic and adverse effects of glucocorticoids include the pharmacokinetic properties of the glucocorticoid, daily dosage, timing of doses during the day, individual differences in steroid metabolism, and the duration of treatment.

Glucocorticoid Replacement Therapy In deficiency states, physiologic replacement is best achieved with a combination of hydrocortisone and the mineralocorticoid fludrocortisone; hydrocortisone alone does not usually provide sufficient mineralocorticoid activity for complete replacement. In Addison’s disease or following adrenalectomy, hydrocortisone at 10–15 mg/m2 daily by mouth is usually required. This is given in two doses, the larger in the morning and the smaller in the evening, mimicking the normal diurnal rhythm of cortisol secretion. The optimum daily dose is determined on the basis of clinical response. Glucocorticoid therapy is supplemented by fludrocortisone 50 to 300 mg daily. In acute adrenocortical insufficiency, hydrocortisone is given intravenously (preferably as sodium succinate) at doses of 100 mg every 6 to 8 h in 0.9% sodium chloride intravenous infusion. In hypopituitarism, glucocorticoids should be given as in adrenocortical insufficiency, but since the production of aldosterone is regulated by the renin–angiotensin system, a mineralocorticoid is not usually required. Additional replacement therapy with levothyroxine and sex hormones should be given as indicated by the pattern of hormone deficiency.

Glucocorticoid Therapy In comparing the relative potencies of corticosteroids in terms of their anti-inflammatory (glucocorticoid) effects, it should be borne in mind that high

Glucocorticoids, Overview

glucocorticoid activity in itself is of no advantage unless it is accompanied by relatively low mineralocorticoid activity. The mineralocorticoid activity of fludrocortisone is so high that its anti-inflammatory activity is of no clinical relevance. The equivalent anti-inflammatory doses of corticosteroids are shown in Table I. The relatively high mineralocorticoid activity of cortisone and hydrocortisone and the resulting fluid retention make them unsuitable for disease suppression on a long-term basis. However, they can be used for adrenal replacement therapy; hydrocortisone is preferred because cortisone requires conversion to hydrocortisone in the liver. Hydrocortisone is used on a short-term basis by intravenous injection for the emergency management of some conditions. The relatively moderate anti-inflammatory potency of hydrocortisone also makes it a useful topical corticosteroid for the management of inflammatory skin conditions because side effects (both topical and systemic) are less marked; cortisone is not active topically. Prednisolone has predominantly glucocorticoid activity and is the corticosteroid most commonly used by mouth for long-term disease suppression. Betamethasone and dexamethasone have very high glucocorticoid activity but insignificant mineralocorticoid activity. This makes them particularly suitable for high-dose therapy in conditions where fluid retention would be a disadvantage. Betamethasone and dexamethasone also have a long duration of action and this, coupled with their lack of mineralocorticoid action, makes them particularly suitable for conditions that require suppression of ACTH secretion (e.g., congenital adrenal hyperplasia). Some esters of betamethasone and beclometasone (beclomethasone)

Table I Equivalent Anti-inflammatory Doses of Corticosteroids Prednisolone
Betamethasone
Cortisone acetate
Deflazacort

t0005

5 mg 750 mg 25 mg 6 mg


Dexamethasone

750 mg


Hydrocortisone

20 mg


Methylprednisolone
Triamcinolone

4 mg 4 mg

Source. Reprinted from the British Medical Association and the Royal Pharmaceutical Society of Great Britain (2002). British National Formulary No. 44, with permission. Note. This table takes no account of mineralocorticoid effects, nor does it take account of variations in duration of action.

257

Glucocorticoids, Overview

exert a considerably more marked topical effect (e.g., on the skin or the lungs) than when given by mouth; use is made of this to obtain topical effects while minimizing systemic side effects (e.g. for skin applications and asthma inhalations). Deflazacort is derived from prednisolone and has high glucocorticoid activity.

Table II Effects of Chronic Pharmacologic Use of Glucocorticoids Endocrine and metabolic Suppression of HPA axis (adrenal suppression) Growth failure in children Carbohydrate intolerance Hyperinsulinism Insulin resistance

SIDE EFFECTS OF CORTICOSTEROIDS Overdosage or prolonged use may exaggerate some of the normal physiologic actions of corticosteroids, leading to mineralocorticoid and glucocorticoid side effects. Mineralocorticoid side effects include hypertension, sodium and water retention, and potassium loss. They are most marked with fludrocortisone, but are significant with cortisone, hydrocortisone, corticotropin, and tetracosactide (tetracosactrin). Mineralocorticoid actions are negligible with the high-potency glucocorticoids betamethasone and dexamethasone and occur only slightly with methylprednisolone, prednisolone, and triamcinolone. Glucocorticoid side effects include diabetes and osteoporosis, avascular necrosis of the femoral head, mental disturbances (a serious paranoid state or depression with risk of suicide may be induced, particularly in patients with a history of mental disorder), euphoria, and muscle wasting (proximal myopathy) (Table II). Corticosteroid therapy is also weakly linked with peptic ulceration. High doses of corticosteroids may cause Cushing’s syndrome, which is usually reversible on withdrawal of treatment, but this must always be gradually tapered to avoid symptoms of acute adrenal insufficiency. In children, administration of corticosteroids may result in suppression of growth. Other complications include increased susceptibility to infection, poor wound healing, and activation of latent granulomatous infections.

Abnormal glucose tolerance test Diabetes mellitus Cushingoid features Moon facies, facial plethora Generalized and truncal obesity Supraclavicular fat collection Posterior cervical fat deposition (buffalo hump) Glucocorticoid-induced acne Thin and fragile skin, violaceous striae Impotence, menstrual disorders Decreased thyroid-stimulating hormone and triodothyronine Hypokalemia, metabolic alkalosis Gastrointestinal Gastric irritation, peptic ulcer Acute pancreatitis (rare) Fatty infiltration of liver (hepatomegaly) (rare) Hematopoietic Leukocytosis Neutrophilia Increased influx from bone marrow and decreased migration from blood vessels Monocytopenia Lymphopenia Migration from blood vessels to lymphoid tissue Eosinopenia Immunologic Suppression of delayed hypersensitivity Inhibition of leukocyte and tissue macrophage migration Inhibition of cytokine secretion/action Suppression of the primary antigen response Musculoskeletal Osteoporosis, spontaneous fractures

Adrenal Suppression During prolonged therapy with corticosteroids, adrenal atrophy develops and may persist for years after stopping. Abrupt withdrawal after a prolonged period may lead to acute adrenal insufficiency, hypotension, or death. Withdrawal may also be associated with fever, myalgia, arthralgia, rhinitis, conjunctivitis, painful itchy skin nodules, and weight loss. To compensate for a diminished adrenocortical response caused by prolonged corticosteroid treatment, any significant intercurrent illness, trauma, or surgical procedure requires a temporary increase in

Aseptic necrosis of femoral and humoral heads and other bones Myopathy Ophthalmologic Posterior subcapsular cataracts (more common in children) Elevated intraocular pressure/glaucoma Neuropsychiatric Sleep disturbances, insomnia Euphoria, depression, mania, psychosis Pseudo-tumor cerebri (benign increase of intracranial pressure) Cardiovascular Hypertension Congestive heart failure in predisposed patients

258 corticosteroid dose or, if already stopped, a temporary reintroduction of corticosteroid treatment. Anesthetists must therefore know whether a patient is taking or has been taking a corticosteroid, to avoid a precipitous fall in blood pressure during anesthesia or in the immediate postoperative period. A suitable regimen for corticosteroid replacement, in patients who have taken more than 10 mg prednisolone daily (or equivalent) within 3 months of surgery, is as follows: .

.

Minor surgery under general anesthesia—usual oral corticosteroid dose on the morning of surgery or hydrocortisone 25–50 mg (usually sodium succinate) intravenously at induction; the usual oral corticosteroid dose is recommenced after surgery. Moderate or major surgery—usual oral corticosteroid dose on the morning of surgery and hydrocortisone 25–50 mg intravenously at induction, followed by hydrocortisone 25–50 mg three times a day by intravenous injection for 24 h after moderate surgery or for 48–72 h after major surgery; the usual preoperative oral corticosteroid dose is recommenced on stopping hydrocortisone injections.

Patients on long-term corticosteroid treatment should carry a Steroid Treatment Card, which gives guidance on minimizing risk and provides details of prescriber, drug, dosage and duration of treatment.

Infections Prolonged courses of corticosteroids increase susceptibility to infections and severity of infections; clinical presentation of infections may also be atypical. Serious infections, e.g., septicemia and tuberculosis, may reach an advanced stage before being recognized and amebiasis or strongyloidiasis may be activated or exacerbated (they should be excluded before corticosteroid treatment is initiated in those at risk or with suggestive symptoms). Fungal or viral ocular infections may also be exacerbated. Chickenpox Unless they have had chickenpox, patients receiving oral or parenteral corticosteroids for purposes other than replacement should be regarded as being at risk of severe chickenpox. Passive immunization with varicella-zoster immunoglobulin is needed for exposed nonimmune patients receiving systemic corticosteroids or for those who have used them within the previous 3 months; varicella-zoster immunoglobulin

Glucocorticoids, Overview

should preferably be given within 3 days of exposure and no later than 10 days. Confirmed chickenpox warrants specialist care and urgent treatment. Corticosteroids should not be stopped and dosage may need to be increased. Topical, inhaled, or rectal corticosteroids are less likely to be associated with an increased risk of severe chickenpox. Measles Patients taking corticosteroids should be advised to take particular care to avoid exposure to measles and to seek immediate medical advice if exposure occurs. Prophylaxis with intramuscular normal immunoglobulin may be needed.

ADMINISTRATION OF CORTICOSTEROIDS Whenever possible, local treatment with creams, intra-articular injections, inhalations, eye drops, or enemas should be used in preference to systemic treatment. The suppressive action of a corticosteroid on cortisol secretion is lowest when it is given as a single dose in the morning. In an attempt to reduce pituitary–adrenal suppression further, the total dose for 2 days can sometimes be taken as a single dose on alternate days; alternate-day administration has not been very successful in the management of asthma. Pituitary–adrenal suppression can also be reduced by means of intermittent therapy with short courses. In some conditions, it may be possible to reduce the dose of corticosteroid by adding a small dose of an immunosuppressive drug. Dosage of corticosteroids varies widely in different diseases and in different patients. If the use of a corticosteroid can save or prolong life, high doses may need to be given, because the complications of therapy are likely to be less serious than the effects of the disease itself. When long-term corticosteroid therapy is used in some chronic diseases, the adverse effects of treatment may become greater than the disabilities caused by the disease. To minimize side effects, the maintenance dose should be kept as low as possible.

WITHDRAWAL OF CORTICOSTEROIDS A gradual withdrawal of systemic corticosteroids should be considered in those subjects whose disease is unlikely to relapse and who have: (1) recently received repeated courses (particularly if taken for

259

Glucocorticoids, Overview

longer than 3 weeks); (2) taken a short course within 1 year of stopping long-term therapy; (3) other possible causes of adrenal suppression; (4) received more than 40 mg daily prednisolone (or equivalent); (5) been given repeat doses in the evening; or (6) received treatment for more than 3 weeks. Systemic corticosteroids may be stopped abruptly in those whose disease is unlikely to relapse and who have received treatment for 3 weeks or less and who are not included in the patient groups described above. During corticosteroid withdrawal, the dose may be reduced rapidly to physiological doses (equivalent to prednisolone at 7.5 mg daily) and then reduced more slowly. Assessment of the disease may be needed during withdrawal to ensure that relapse does not occur.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Adrenal Cortex, Physiology . Adrenal Insufficiency . Adrenal Suppression . Glucocorticoid Receptor . Glucocorticoid Resistance Syndromes and States . Glucocorticoids and Immunity . Glucocorticoids in Aging: Relevance to Cognition . Glucocorticoids, Overview . Growth and Glucocorticoids . Nuclear Factor-kB and Glucocorticoid Receptors

Further Reading British Medical Association and The Royal Pharmaceutical Society of Great Britain. (2002). ‘‘British National Formulary No. 44,’’ Chap. 6, pp. 348–351. Pharmaceutical Press, London, UK. Clark, J. K., Schrader, W. T., and O’Malley, B. W. (1992). Mechanism of steroid hormones. In ‘‘Williams Textbook of Endocrinology’’ ( J. D. Wilson and D. W. Foster, eds.), pp. 35–90. W. B. Saunders, Philadelphia, PA. Gower, D. B. (1984). Steroid catabolism and urinary excretion. In ‘‘Biochemistry of Steroid Hormones’’ (H. L. J. Makin, ed.), pp. 349–382. Blackwell Science, Oxford, UK. Habib, K. E., Gold, P. W., and Chrousos, G. P. (2001). Neuroendocrinology of stress. Endocrinol. Metab. Clin. N. Am. 30, 695–728. Honour, J. W. (1993). The adrenal cortex. In ‘‘Clinical Paediatric Endocrinology’’ (C. G. D. Brook, ed.), pp. 434–452. Blackwell Science, Oxford, UK. Keller-Wood, M. E., and Dallman, M. (1984). Corticosteroid inhibition of ACTH secretion. Endocr. Rev. 5, 1–24. Simpson, E. R., and Waterman, M. R. (1995). Steroid biosynthesis in the adrenal cortex and its regulation by adrenocorticotropin. In ‘‘Endocrinology’’ (L. J. DeGroote, M. Besser, H. G. Burger, J. L. Jameson, D. L. Loriaux, J. C. Marshall, W. D. Odell, J. T. Potts, Jr., and A. H. Rubenstein, eds.), pp. 1630–1641. W. B. Saunders, Philadelphia, PA. Stewart, P. M. (1996). 11b-Hydroxysteroid dehydrogenase: Implications for clinical medicine. Clin. Endocrinol. 44, 493–499. Yanase, T., Simpson, E. R., and Waterman, M. R. (1991). 17aHydroxylase/17,20-lyase deficiency: From clinical investigation to molecular definition. Endocr. Rev. 12, 91–108.

264

Glucose Physiology, Normal

Figure 1 Diurnal changes in plasma glucose, insulin, and glucagon concentrations.

FACTORS THAT REGULATE GLUCOSE FLUXES Insulin Insulin plays a central role in glucose homeostasis mainly by its action on liver, kidney, adipose tissue, and skeletal muscle. In liver and kidney, it suppresses glucose production by regulating the rate-limiting key enzymes of gluconeogenesis (glucose-6-phosphatase and fructose-1,6-bisphosphatase) and glycogenolysis (glycogensynthase and phosphorylase). In skeletal muscle, its main action is to promote glucose uptake by causing the translocation of Glut-4 transporters to the cell membrane from an intracellular pool. In adipose tissue, insulin will also increase glucose transport and the resultant increase in glycolysis provides a-glycerophosphate to promote triglyceride formation but its main effect is to inhibit hormone-sensitive lipase. This will lead to a decrease in circulating FFA levels, which will affect glucose production in liver and kidney and glucose uptake in various tissues (Fig. 2).

Binding of insulin to its receptor activates a complex intracellular cascade of autophosphorylation by protein kinases. Intracellular receptor substrates and members of the phosphatidylinositol 3-kinase family seem to be the most important participants. The latter triggers the migration of glucose transporters from an intracellular pool toward the plasma membrane, thereby increasing the number of transporters located on the cellular surface and thus promoting the efficiency of glucose uptake (Fig. 3). This will also result in decreases in intracellular cyclic AMP (cAMP) levels, which mediate the effects of insulin on glucose production and lipolysis. Insulin will also alter the activity of various genes that will affect the amount of certain enzymes. Finally, via its action on amino acid transport, insulin suppresses protein degradation and lipolysis, which will diminish the availability of gluconeogenic precursors and thereby reduce glucose production indirectly. The main regulator of insulin release is the prevailing plasma glucose concentration. An increase in the

265

Glucose Physiology, Normal

Figure 2 Postprandial insulin secretion; regulation of glucose production and disposal.

plasma glucose concentration stimulates insulin secretion, whereas a decrease in plasma glucose inhibits insulin secretion so that throughout the day, plasma glucose and insulin levels change in parallel (Fig. 1). Amino acids, e.g., arginine, and to a lesser extent FFA can also stimulate insulin secretion. The small intestine produces factors called incretins (e.g., gastrointestinal peptide, glucagon-like peptide-2), which are secreted after meal ingestion and augment postprandial insulin secretion. This explains

why insulin concentrations increase to a greater extent when glucose is given orally than intravenously.

Glucagon Glucagon acts solely on the liver, having effects mediated by changes in intracellular cAMP levels that are opposite of those of insulin. Its secretion is also regulated in a reciprocal manner to that of insulin. An increase in plasma glucagon will increase hepatic glucose release within minutes via an increase in glycogen breakdown. Binding of glucagon to its receptor immediately increases intracellular cAMP levels, which increases glycogenolysis and inhibits glycogen synthase by stimulation of phosphorylase and inactivation of glycogen synthase. Prolonged elevation of plasma glucagon can increase gluconeogenesis in the liver, whereas it has no effect on renal gluconeogenesis. Glucagon secretion is suppressed by increases in plasma glucose and insulin and is increased by hypoglycemia and catecholamines. Amino acids are a potent stimulator of glucagon release. Thus, after protein-rich meals, glucagon release might not be suppressed despite increases in plasma insulin and glucose concentrations.

Catecholamines Figure 3 Insulin binding and intracellular cascade. Glut, glucose transporter; IRS, intracellular receptor substrates; PI3-Kinase, phosphatidylinositol 3-kinase.

Epinephrine and norepinephrine are released by the adrenal glands and norepinephrine is released from sympathetic nerves during exercise, various stresses (e.g., trauma, infection), and hypoglycemia. They

266 have complex effects on glucose mediated by both direct and indirect mechanisms. Such actions include stimulation of renal gluconeogenesis, hepatic and muscle glycogenolysis, adipose tissue lipolysis, and glucagon release, which are mediated by badrenergic receptors. Catecholamines also inhibit insulin release directly via a-adrenergic receptors. Indirect effects include suppression of glucose uptake in skeletal muscle due to the elevation of plasma FFA and stimulation of gluconeogenesis in liver and kidney via increases in plasma FFA and gluconeogenic precursors (mainly glycerol from lipolysis and lactate from skeletal muscle glycogenolysis). Along with glucagon, catecholamines are the most important counterregulatory factors protecting against hypoglycemia.

Growth Hormone, Cortisol, and Thyroid Hormone Growth hormone, cortisol, and thyroid hormone largely act to regulate the response of target tissues to insulin, glucagon, and catecholamines on a longterm basis, e.g., reducing responses to insulin and increasing responses to glucagon and catecholamines. Under conditions similar to those during which catecholamines are released, growth hormone and cortisol are released and within an hour or two reduce the effectiveness of insulin and enhance the action of glucagon and catecholamines. Prolonged elevation of these hormones, such as is seen in acromegaly and Cushing’s syndrome, can cause severe insulin resistance and diabetes mellitus.

FFA FFA are a major fuel used by most tissues of the body except the brain, renal medulla, and erythrocytes. Increases in plasma FFA and consequently their uptake into cells have numerous direct and indirect effects that influence glucose homeostasis. These include direct effects on hormone secretion (a moderate stimulating action on insulin secretion and a potent inhibitory action on glucagon and growth hormone) as well as stimulating effects on hepatic and renal gluconeogenesis and an inhibitory effect on muscle glucose uptake. The effects on liver, kidney, and muscle are mediated in part by changes in hormonal environment and competition with glucose as an oxidative fuel (mediated primarily by changes in pyruvate dehydrogenase and interference with insulin signaling pathways, both of these being mediated by coenzyme

Glucose Physiology, Normal

A metabolites of FFA). In general, opposite effects are observed when plasma FFA are low. Circulating FFA levels, like those of glucose, are the net result of changes in FFA entry and exit from plasma. FFA entry into plasma largely depends on the balance between the activation of hormone-sensitive lipase by catecholamines, growth hormone, and cortisol and the inhibition of lipase by insulin. The exit of FFA from plasma is stimulated by insulin.

THE POSTABSORPTIVE STATE General Considerations In the period after an overnight fast, referred to as the postabsorptive state, plasma glucose ranges between 70 and 110 mg/dl (average 90 mg/dl). This state is considered to represent a steady-state condition since the rate of appearance of glucose approximates its rate of disappearance (10 mmol kg1 min1). However, even though removal is often undetectable, the rate of removal is slightly greater than the rate of appearance so that with more prolonged fasting, plasma glucose concentrations decrease. However, even after 72 h of fasting, plasma glucose does normally not decrease below 50 mg/dl (2.9 mM).

Glucose Utilization In the postabsorptive state, plasma insulin levels are low and therefore glucose uptake in tissues is largely dependent on tissue needs. The majority of glucose is taken up by the brain (50%) and is completely oxidized; glucose taken up by muscle (20%), adipocytes (5%), erythrocytes (5%), splanchnic organs (10%), and kidney (5%) (Fig. 4) undergoes mostly nonoxidative glycolysis, resulting in the release of 3-carbon precursors (lactate, pyruvate, and alanine), which are used for gluconeogenesis, into the circulation.

Glucose Production Glucose production in the postabsorptive state is regulated to match tissue demand, which may increase during exercise or stresses such as infection and trauma. Normally, approximately 50% of the glucose released into the circulation is the result of hepatic glycogenolysis; the remaining 50% is due to gluconeogenesis (30% liver; 20% kidney). The proportion of glucose produced due to gluconeogenesis increases with the duration of the fast since glycogen stores are rapidly depleted. The liver

267

Glucose Physiology, Normal

Figure 4 Glucose production and utilization after an overnight fast. contains a total of 75 g glucose. Assuming that the liver releases glucose from glycogen at a rate of 5 mmol kg1 min1, glycogen stores would be depleted within 20 h. Thus, the proportion due to gluconeogenesis must increase so that after 72 h, glucose production by the liver is almost exclusively due to gluconeogenesis. The kidney, in contrast, contains little glycogen stores and the cells that could make glycogen lack glucose-6-phosphatase; consequently, all the glucose released by the kidney is due to gluconeogenesis. (Renal gluconeogenesis increases with fasting to a greater extent than hepatic gluconeogenesis.) Insulin suppresses both hepatic and renal glucose release; however, glucagon promptly increases hepatic glucose release, whereas catecholamines stimulate more renal glucose release.

Prolonged Fasting With the duration of fasting, plasma insulin levels decrease, whereas those of glucagon, catecholamines, growth hormone, and cortisol increase. Therefore, the oxidation of glycerol, plasma FFA and FFA products, and the ketone bodies b-hydroxybutyrate and acetoacetate increases. Hepatic glycogen stores become depleted and after 60 h virtually all of glucose released is due to gluconeogenesis. During the first 60–72 h of fasting, the decrease in glucose release is greater than the decrease in glucose uptake, so that plasma glucose levels decrease. At approximately 60 h, with plasma glucose averaging 60 mg/dl, a new pseudo-steady state is achieved (Table I). This stabilization is the basis for the 72 h fast for the diagnosis of patients with hypoglycemia due to insulin-producing tumors of the pancreas (insulinoma). In such patients, insulin

secretion is not appropriately reduced and leads to a further reduction of endogenous glucose production together with increased glucose uptake and consequently to the development of hypoglycemia with plasma glucose levels below 50 mg/dl.

THE POSTPRANDIAL STATE General Considerations The major function of meal ingestion is to replenish tissue glucose (glycogen) and lipid (triglyceride) stores that have been depleted due to fasting and physical activity. Thus, after meal ingestion, endogenous glucose and FFA release is suppressed, favoring glycogen accumulation. Glucose replaces FFA as the predominant energy fuel as plasma FFA decrease, Table I Glucose Release and Disposal after Prolonged Fasting (60 h) Glucose release (mmol kg1 min1) Overall Gluconeogenesis

6.0 5.5

Glycogenolysis

0.5

Tissues

Glucose disposal (mmol kg1 min1) Overall Oxidation

6.0 4.8

Glycolysis

1.2

Tissues

Liver

2.7

Brain

3.5

Kidney

2.8

Skeletal muscle

1.0

Splanchnic organs

0.5

Kidney

0.4

Adipose tissue Blood cells

0.2 0.4

268

Glucose Physiology, Normal

favoring FFA incorporation into triglyceride stores, so that ingested carbohydrate becomes the major fuel used by the body. Various factors, such as the size of the meal, prior physical exercise and duration since the last meal, and composition of the meal, can affect postprandial glucose homeostasis. However, from a practical point of view, the most important factors are changes in insulin and glucagon secretion and their effects on hepatic sequestration of meal carbohydrates, suppression of endogenous glucose production, and finally stimulation of the uptake, storage, glycolysis, and oxidation of glucose in hepatic and posthepatic tissues.

Postprandial Glucose Concentrations and Fluxes After a meal, plasma glucose concentrations increase since the rate of appearance of glucose in plasma exceeds the rate of disappearance. Subsequently, plasma glucose decreases when the rate of disappearance exceeds the rate of appearance. The appearance of glucose in plasma represents the sum of glucose from the meal reaching the circulation and the remaining

endogenous glucose production. The time courses of these changes are shown in Fig. 5. Endogenous glucose release is suppressed 60% after a meal. The appearance of ingested glucose is detected within 15 min after a meal, reaches a maximum at 60 min, and decreases gradually thereafter. Only 75% of the glucose in a meal reaches the systemic circulation; the remaining 25% is sequestrated by the splanchnic bed. Of a theoretical meal containing 100 g of glucose, 20–30% is initially extracted in the splanchnic bed. At least half of this is taken up by the liver and incorporated into hepatic glycogen; the remainder is probably released as lactate due to hepatic glycolysis. Of the glucose in the meal that reaches the systemic circulation, approximately 30–40% is taken up by skeletal muscle to be initially oxidized in favor of FFA and later to be stored as glycogen. Little of the glucose taken up by muscle is released as lactate or other gluconeogenic substrates into the circulation. Of the remaining glucose released into the systemic circulation, 20% is taken up by the brain, 10% by the kidney, and 5% by adipose tissue. Approximately 40% of the glucose disposed of after a meal is stored predominantly as glycogen in

Figure 5 Postprandial changes in plasma glucose, insulin, glucagon concentrations, rates of plasma glucose appearance/ disappearance, and hepatic and renal glucose production.

269

Glucose Physiology, Normal

Figure 6 Contribution of storage and glycolysis to the disposal of glucose after a meal.

liver and muscle and, to a lesser extent, as triglycerides in adipose tissue. The remaining 60% is undergoes glycolysis either oxidatively to CO2 and H2O (40%) or nonoxidatively to lactate (20%) (Fig. 6). That a substantial amount of glucose undergoes nonoxidative glycolysis is not surprising if one considers that glycolysis is needed to provide 3-carbon fragments for gluconeogenesis and precursors for the indirect pathway of hepatic glycogen formation. Renal glucose production initially increases after a meal (Fig. 5). The physiological role of this increase still needs to be elucidated. Teleologically, postprandial renal glucose release may facilitate efficient liver glycogen repletion by permitting substantial suppression of hepatic glucose release. It can be readily appreciated that three main factors regulate postprandial glucose levels: suppression of hepatic glucose release, hepatic sequestration of the ingested glucose, and uptake of glucose from the systemic circulation. Initial suppression of endogenous glucose release and hepatic sequestration depend largely on the reciprocal secretion of glucagon and insulin. Insulin increases the number of glucose transporters in muscle and thereby increases glucose fractional extraction, an index of the efficiency of glucose uptake; glucose fractional extraction by brain and other insulin independent tissues actually decreases.

SUMMARY In most circumstances, regulation of glucose production is more important than regulation of glucose utilization in determining plasma glucose

concentrations. Recall that in the fasting state, 80% of glucose utilization is insulin independent. Insulin levels are low and are needed only to suppress excessive endogenous glucose production and lipolysis. Another example of the importance of glucose production is fasting hyperglycemia in type 2 diabetes. Rates of glucose utilization are generally normal, whereas rates of glucose production are increased. Of interest is the finding that renal glucose release initially increases after a meal, whereas hepatic glucose release decreases; these reciprocal changes permit efficient repletion of glycogen stores. This finding and other observations have provided convincing evidence for the concept of hepatorenal reciprocity. According to this concept, when the release of glucose by either liver or kidney is reduced, the other organ will increase its glucose release to maintain euglycemia. Similar reciprocal changes are found during recovery from insulin-induced hypoglycemia in patients with type 2 diabetes. The strongest support for the concept of hepatorenal reciprocity has been provided by the observation that during the anhepatic stage of liver transplantation (when liver glucose release is absent), the kidney increases its release of glucose threefold so that hypoglycemia does not occur.

See Also the Following Articles Catecholamines . Diabetes, Type 2 . Glucagon and Glucagon-like Peptides . Glucose, Impaired Tolerance . Glucose Toxicity . Insulin Secretion, Physiology

Further Reading Boden, G. (1997). Role of fatty acids in the pathogenesis of insulin resistance and NIDDM. Diabetes 46, 3–10. Cahill, G. (1970). Starvation in man. N. Engl. J. Med. 282, 668–675. Dinneen, S., Gerich, J., and Rizza, R. (1992). Carbohydrate metabolism in noninsulin-dependent diabetes mellitus. N. Engl. J. Med. 327, 707–713. Ekberg, K., Landau, B., Wajngot, A., et al. (1999). Contributions by kidney and liver to glucose production in the postabsorptive state and after 60 h of fasting. Diabetes 48, 292–298. Gerich, J. (1993). Control of glycaemia. Bailliere Clin. Endocrinol. Metab. 7, 551–586. Havel, R. (1972). Caloric homeostasis and disorders of fuel transport. N. Engl. J. Med. 287, 1186–1192. Joseph, S., Heaton, N., Potter, D., et al. (2000). Renal glucose production compensates for the liver during the anhepatic phase of liver transplantation. Diabetes 49, 450–456. Kelley, D., Mitrakou, A., Marsh, H., et al. (1988). Skeletal muscle glycolysis, oxidation, and storage of an oral glucose load. J. Clin. Invest. 81, 1563–1571. Kruszynska, Y., Mulford, M., Yu, J., et al. (1997). Effects of nonesterified fatty acids on glucose metabolism after glucose ingestion. Diabetes 46, 1586–1593.

270 Landau, B., Wahren, J., Chandramouli, V., et al. (1996). Contributions of gluconeogenesis to glucose production in the fasted state. J. Clin. Invest. 98, 378–385. Magnusson, I., Rothman, D., Gerard, D., et al. (1995). Contribution of hepatic glycogenolysis to glucose production in humans in response to a physiological increase in plasma glucagon concentration. Diabetes 44, 185–189. Marin, P., Hogh-Kristiansen, I., Jansson, S., et al. (1992). Uptake of glucose carbon in muscle glycogen and adipose tissue triglycerides in vivo in humans. Am. J. Physiol. 263, E473–E480. McGarry, J. (1998). Glucose–fatty acid interactions in health and disease. Am. J. Clin. Nutr. 67, 500S–504S. Meyer, C., Dostou, J., Nadkarni, V., and Gerich, J. (1998). Effects of physiological hyperinsulinemia on systemic, renal and hepatic substrate metabolism. Am. J. Physiol. 275, F915–F921.

Glucose Physiology, Normal

Meyer, C., Dostou, J., Welle, S., and Gerich, J. (2002). Role of human liver, kidney and skeletal muscle in postprandial glucose homeostasis. Am. J. Physiol. Endocrinol. Metab. 282, E419–E427. Stumvoll, M., Chintalapudi, U., Perriello, G., et al. (1995). Uptake and release of glucose by the human kidney: Postabsorptive rates and responses to epinephrine. J. Clin. Invest. 96, 2528–2533. Taylor, R., Magnusson, I., and Rothman, D. (1997). Direct assessment of liver glycogen storage by 13C nuclear magnetic resonance spectroscopy and regulation of glucose homeostasis after a mixed meal in normal subjects. J. Clin. Invest. 97, 126–132. Taylor, R., Price, T., Katz, L., et al. (1993). Direct measurement of change in muscle glycogen concentration after a mixed meal in normal subjects. Am. J. Physiol. 265, E224–E229.

272 after as little as a few hours of exposure to high levels of glucose and, in these earlier stages, the damage is reversible. Thus, the impaired insulin secretion observed in early human type 2 diabetes can improve significantly with control of hyperglycemia. Ultimately, however, the damage to the beta cells becomes irreversible (‘‘beta cell failure’’). It is likely that different mechanisms come into play as these processes evolve. The initial events, such as desensitization to glucose, may be related to changes in the levels and/or activities of transcription factors and other key regulatory molecules involved in glucose sensing, whereas the subsequent loss of beta cell function involves down-regulation of insulin mRNA and may proceed to apoptotic cell loss. It is possible, however, that these events represent a continuum that shares a common signaling mechanism. Other workers, for example, Unger, have noted that other nutrients that are found in excess in type 2 diabetes, such as free fatty acids, can also impair beta cell function. These hypotheses—glucose toxicity and ‘‘lipotoxicity’’—are not mutually exclusive. For example, both free fatty acids and glucose can lead to chronic oxidative stress and both can activate the hexosamine signaling pathway, both being potential mechanisms underlying beta cell failure.

GLUCOSE-INDUCED INSULIN RESISTANCE IN TYPE 1 AND TYPE 2 DIABETES p0025

p0030

Another of the pathophysiologic hallmarks of type 2 diabetes is insulin resistance. Although it is considered by many to be involved in the initial pathogenesis of the disease, there are also numerous studies that illustrate that insulin resistance can be an acquired defect, the result of excess nutrient delivery to tissues. The laboratories of Rossetti, Yki-Jarvinen, and DeFronzo have pioneered work in this area and have also reviewed the topic extensively. As was the case for decreased insulin secretion, the glucose-induced insulin resistance has been convincingly demonstrated in a variety of experimental systems and in humans, nondiabetic as well as those with type 1 or type 2 diabetes. Most often, the experimental paradigm involves decreased insulin-mediated glucose disposal after a period of hyperglycemia, but the definition has been broadened to include any defect in insulin signaling or responsiveness observed after treatment with high concentrations of glucose. The ultimate mediators of glucose-induced insulin resistance are not known, although much work has

Glucose Toxicity

illuminated some of the mechanisms involved. In skeletal muscle, the effect of hyperglycemia is decreased insulin-stimulated glucose uptake caused by a failure of translocation of the specific glucose transporter GLUT4 to the plasma membrane. Why this occurs is less clear. Interest has focused on the early steps in insulin signal transduction, specifically the insulin receptor substrate/phosphatidylinositol 3-kinase/Akt (PKB) pathways, and several defects have been described in these pathways in glucose-induced insulin resistance. However, there is no consistent or universally accepted mechanism that can completely explain the phenomenon. Glucose-induced insulin resistance shares many features with that induced by excess levels of free fatty acids. This has given rise to the hypothesis that glucose-induced insulin resistance might be more appropriately termed nutrient-induced insulin resistance. A possible common denominator that could explain both glucose- and fatty acid-induced insulin resistance is hexosamine signaling, because high levels of both glucose and fatty acids result in increased hexosamine pathway products. This hypothesis is also attractive in that it would predict that insulin resistance would occur in the presence of excess nutrient flux even before nutrient levels were significantly increased in plasma. Thus, insulin resistance by this mechanism would be expected to occur before overt hyperglycemia and there could be a unifying mechanism to explain insulin resistance in obesity as well as in clinical diabetes. Glucose toxicity for insulin action is manifest in the clinical setting in two major ways that deserve mention. First, the typical patient presenting with a prolonged history of weight loss, polyuria, and severe hyperglycemia is likely to improve his or her glucose disposal rate once the hyperglycemia is brought under control; that is, the glucose toxicity will reverse. Thus, it will be harder to achieve glycemic control than to maintain glycemic control. This means that the therapy needed to attain normoglycemia may cause hypoglycemia a few days later as the glucose toxicity component resolves. Conversely, it has also been shown that an initial period of normoglycemia produced by intense insulin therapy will, by breaking the vicious cycle of glucose toxicity, allow better responsiveness to oral agents. Thus, there will be fewer oral agent ‘‘failures’’ after such a period of normoglycemia. A second major clinical consequence of glucoseinduced insulin resistance is that huge doses of insulin may be required to lead to disposal of the nutrient load from a meal and in some cases even the maximal

Glucose Toxicity

glucose disposal rate of the individual can be exceeded, meaning that no amount of insulin will lead to normoglycemia. Interestingly, Yki-Jarvinen and others have noted that in such situations, muscle glycogen content and net glucose flux into muscle are essentially normal. Furthermore, there are few if any ill effects of diabetes on skeletal muscle function. This has led to the proposal that nutrient-induced insulin resistance may be an adaptation to excess nutrient flux. That is, skeletal muscle is able to autoregulate glucose uptake by down-regulating glucose transporters in the face of hyperglycemia. If this is the case, it may be more appropriate to view glucose-induced insulin resistance in terms of normal physiology rather than as toxicity.

MEDIATION OF DIABETIC COMPLICATIONS BY GLUCOSE p0050

There is substantial evidence that most if not all of the classic complications of diabetes are mediated by hyperglycemia and/or excessive nutrient flux. This was most clearly demonstrated in the Diabetes Control and Complications Trial for type 1 diabetes and by the ‘‘Kumamoto Study’’ and the United Kingdom Prospective Diabetes Study for type 2 diabetes. Although not usually referred to as glucose toxicity, some of the pathways leading to these complications—nephropathy, neuropathy, and retinopathy—may be shared and it may be informative to group them together with beta cell failure and insulin resistance as ‘‘adverse consequences of excess nutrient flux.’’

PROPOSED MECHANISMS FOR GLUCOSE TOXICITY The precise mechanisms and mediators of glucose toxicity are not known. Given the complex regulation of glucose metabolism, the many pathways of known regulation by glucose, and the many fates of intra- and extracellular glucose, it is likely that there will be multiple pathways through which glucose exerts its adverse effects. Some of the leading proposals can be very briefly summarized as follows: . Nonenzymatic glycation of proteins. Covalent attachment of free glucose to amino groups in proteins or other macromolecules can occur via formation of a Schiff’s base. These can subsequently rearrange, leading to the elaboration of advanced glycation end-products that can have toxic effects on

273 cells. Nonenzymatic glycation is contrasted to the enzymatic linkage of carbohydrates (usually uridine diphosphate amino sugars) to specific residues of proteins. Given that the effects of high levels of glucose on insulin secretion and insulin action can be seen in hours to a few days and that these effects are likely to be mediated intracellularly, it is unlikely that this process plays a large role in beta cell failure or insulin resistance. . Aldose reductase pathway. Sorbitol, generated by the enzyme aldose reductase, has been postulated to act normally as an intracellular osmolyte to buffer changes in extracellular osmolality. In the presence of chronic hyperglycemia, however, untoward effects of sorbitol accumulation, such as depletion of myoinositol or increased oxidative stress, may occur. This pathway has not been specifically linked to insulin resistance or beta cell failure. . Protein kinase C. Hyperglycemia-induced activation of diacylglycerol/protein kinase C-dependent pathways has been postulated to lead to changes in gene regulation that could play a role in diabetic complications, although no studies provide a specific link to the classic glucose toxicity effects. . Hexosamines. Since the demonstration by Marshall that glucosamine can induce insulin resistance in adipocytes, and his discovery of the hexosamine signaling pathway, much evidence that this pathway plays an important role in nutrient sensing and adaptation to excess nutrient flux has accumulated. In particular, it has been shown in several animal and cell culture models that high concentrations of glucose lead to high levels of products of the hexosamine biosynthesis pathway that in turn can mimic very well the diabetic phenotypes of beta cell failure and glucose-induced insulin resistance. The hypothesis is further attractive in that the pathway may provide a common mechanism for the similar effects of free fatty acids, which also result in increased hexosamine flux. The signaling function of the pathway is thought to occur by the enzymatic O-linked glycosylation of cytosolic signaling proteins. This dynamic process occurs on residues otherwise used for regulatory phosphorylation, so that high levels of hexosamine flux can generate a dominant signal downstream to abrogate hormone signaling in situations of excess nutrients. Thus, insulin would no longer be able to stimulate glucose uptake or glycogen synthesis if intracellular fuel stores were already replete, protecting the cell from excess nutrient accumulation and allowing the excess calories to be shunted to adipocytes.

274 Oxidative stress. This has been shown to be a factor in a large variety of tissues that are harmed by the diabetic milieu. Brownlee has proposed that many features of glucose toxicity and diabetic complications may be functionally linked to this mechanism, including activation of the hexosamine signaling pathway. .

See Also the Following Articles Diabetes, Type 1 . Diabetes, Type 2 . Glucose, Impaired Tolerance . Glucose Physiology, Normal . Insulin-Resistant States, Role of Free Fatty Acids (FFA) . Obesity and Diabetes, Regulation of Food Intake

Further Reading Du, X.-L., Edelstein, D., Rossetti, R., Fantus, I. G., Goldberg, H., Ziyadeh, F., Wu, J., and Brownlee, M. (2000). Hyperglycemiainduced mitochondrial superoxide overproduction activates the hexosamine pathway and induces plasminogen activator inhibitor-1 expression by increasing Sp1 glycosylation. Proc. Natl. Acad. Sci. USA 97, 12222–12226.

Glucose Toxicity

The Diabetes Control and Complications Trial Research Group. (1993). The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. N. Engl. J. Med. 329, 977–986. Leahy, J. L., Bonner-Weir, S., and Weir, G. C. (1992). Beta-cell dysfunction induced by chronic hyperglycemia: Current ideas on mechanism of impaired glucose-induced insulin secretion. Diabetes Care 15, 442–455. Marshall, S., Bacote, V., and Traxinger, R. R. (1991). Discovery of a metabolic pathway mediating glucose-induced desensitization of the glucose transport system. J. Biol. Chem. 266, 4706–4712. McClain, D. A. (2001). Hexosamines as mediators of nutrient sensing: Relevance to obesity, insulin resistance, and diabetes. Curr. Opin. Endocrinol. Diabetes 8, 186–191. Robertson, R. P., Olson, L. K., and Zhang, H. J. (1994). Differentiating glucose toxicity from desensitization: A new message from the insulin gene. Diabetes 43, 1085–1089. Rossetti, L., Giaccari, A., and DeFronzo, R. A. (1990). Glucose toxicity. Diabetes Care 6, 610–630. Rossetti, L. (2000). Perspective: Hexosamines and nutrient sensing. Endocrinology 141, 1922–1925. UK Prospective Diabetes Study (UKPDS) Group. (1998). Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet 352, 837–853. Yki-Jarvinen, H. (1992). Glucose toxicity. Endocr. Rev. 13, 415–431.

261

Glucose, Impaired Tolerance

t0005

Table I Diagnostic Criteria for Impaired Glucose Levels Organization

Impaired fasting a glucose

Impaired glucose b tolerance

World Health Organization

Not defined

140–199 mg/dl

American Diabetes Association

110–125 mg/dl 6.1–6.9 mmol/liter

140–199 mg/dl 7.8–11.1 mmol/liter

7.8–11.1 mmol/liter

a After b

an 8 to 12 h overnight fast. 2 h value on 75 g OGTT.

Prospective studies in people with IGT have revealed rates of progression to diabetes (usually type 2) in the range of 5–10% per year. The rates translate to a threeto sevenfold increase in the risk of diabetes relative to people with normal glucose levels. Factors that may alter the risk of diabetes among people with IGT vary from study to study, but include glucose levels, insulin resistance, B cell defect, obesity, lipid levels, and blood pressure (greater values for each factor indicate increased risk). Clinical cardiovascular events have been reported to be more frequent in people with IGT than in people with normal glucose tolerance in some, but not all studies. In general, the excess risks were accounted for by the presence of cardiovascular risk factors other than mild hyperglycemia.

PATHOPHYSIOLOGY In general, glucose levels rise above normal when insulin secretion from the pancreas is insufficient to meet the insulin needs of tissues that make glucose (the liver) or take it up in response to insulin (skeletal muscle and adipose tissue). Diabetes reflects a large imbalance between these two factors. IGT represents a relatively milder imbalance, albeit not a trivial one. Cross-sectional data from the Insulin Resistance and Atherosclerosis Study indicate that people with IGT have 60% less insulin secretion for their degree of insulin resistance than people with normal glucose tolerance. Similar findings have been reported for women with gestational diabetes, a separate form of impaired glucose tolerance. Causes of insufficient insulin secretion vary among people with IGT. Some people have circulating autoimmune markers directed at the pancreatic islets or insulin-secreting B cells. Those individuals appear to have IGT as part of evolving type 1 diabetes. Other individuals with IGT (the majority) are obese or have other reasons to have increased insulin requirements of their tissues (‘‘insulin resistance’’). They may make considerable insulin, but less than equally insulin-resistant people

with normal glucose tolerance. Evidence from the Troglitazone in Prevention of Diabetes (TRIPOD) study indicates that, at least in Hispanic Americans, the increased insulin secretory demands that are imposed on pancreatic B cells by chronic insulin resistance cause loss of B cell function. Treatment of insulin resistance can delay or prevent progressive B cell failure and diabetes. The insulin resistance from which IGT frequently evolves is at the center of a cluster of clinical conditions that are collectively known as the insulin resistance syndrome. Like IGT, atherosclerosis is a component of this syndrome. Many mechanisms have been proposed as links between insulin resistance and atherosclerosis, including hyperinsulinemia, dyslipidemia (especially elevated triglycerides and low levels of high-density lipoprotein cholesterol), hypertension, impaired fibrinolysis, and chronic inflammation. In people with IGT, mild hyperglycemia could promote atherosclerosis as well, although solid evidence for this in humans (i.e., lowering of mild hyperglycemia to mitigate atherosclerosis) is lacking.

TREATMENT There is no standard treatment for IGT. Four different randomized trials to evaluate the effects of different interventions on the risk of diabetes in people with IGT have been completed. In two of them [the Finnish Diabetes Prevention Study (DPS) and U.S. Diabetes Prevention Program (DPP)], intensive lifestyle interventions were designed to achieve modest weight reduction (e.g., 7% of body weight in the DPP) and a modest increase in physical activity (e.g., brisk walking or the equivalent for 150 min per week in the DPP) in adult men and women with IGT. The risk of diabetes was reduced 58% compared to the control group in each study. The U.S. DPP also included an arm in which subjects were randomized to metformin 1500 mg/day instead of intensive lifestyle intervention. The risk of diabetes in the metformin arm was reduced by 31% compared to the control group. In the STOP-NIDDM trial, acarbose (100 mg three times daily) was given to adult men and women with IGT. The incidence of diabetes was reduced by 25% compared to placebo-treated subjects. Survival curves in the Finnish DPS, U.S. DPP, and STOPNIDDM trials revealed a slowing of the onset of diabetes in treated arms rather than true diabetes prevention. In the TRIPOD study cited above, a medium dose of an insulin-sensitizing thiazolidinedione drug (troglitazone, which is no longer available for clinical use) reduced the incidence of diabetes by

262 55% in Hispanic American women with prior gestational diabetes. Women who responded to the drug with reduced endogenous insulin requirements had complete stabilization of pancreatic B cell function and glucose levels for 4.5 years, indicating true diabetes prevention during that period of time. These four studies reveal that it is possible to delay or prevent the onset of diabetes in people with IGT using behavioral and/or pharmacological interventions. Efficient strategies for screening to find people with IGT and optimal approaches to clinical management remain to be defined. Likewise, whether treatment of mild hyperglycemia per se in IGT can reduce cardiovascular events independent of changes in other cardiovascular risk factors remains to be determined. People with IGT should be evaluated and treated for standard cardiovascular risk factors.

See Also the Following Articles Atherosclerosis . Cardiovascular Disease in Diabetes . Diabetes, Type 1 . Diabetes, Type 2 . Glucose Physiology, Normal . Glucose Toxicity . Insulin Secretion: Functional and Biochemical Aspects

Further Reading American Diabetes Association. (1997). Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 20, 1183–1197.

Glucose, Impaired Tolerance

Buchanan, T. A., Xiang, A. H., Peters, R. K., Kjos, S. L., Marroquin, A., Goico, J., Ochoa, C., Tan, S., Berkowitz, K., Hodis, H. N., and Azen, S. P. (2002). The TRIPOD Study: Preservation of pancreatic B-cell function and prevention of type 2 diabetes by pharmacological treatment of insulin resistance in high-risk Hispanic women. Diabetes, 51, 2796–2803. Chiasson, J. L., Josse, R. G., Gomis, R., Hanefeld, M., Karasik, A., and Laakso, M. (2002). Acarbose for prevention of type 2 diabetes: The STOP-NIDDM randomised trial. Lancet 359, 2072–2077. Diabetes Prevention Program Research Group. (2002). Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N. Engl. J. Med. 346, 393–403. Harris, M. I., Flegal, K. M., Cowie, C. C., Eberhardt, M. S., Goldstein, D. E., Little, R. R., Wiedmeyer, H.-.M, and ByrdHolt, D. D. (1998). Prevalence of diabetes, impaired fasting glucose, and impaired glucose tolerance in U.S. adults. Diabetes Care 21, 518–524. Haffner, S. M., Agostino, R. D., Jr., Saad, M. F., O’Leary, D. H., Savage, P. J., Rewers, M., Selby, J., Bergman, R. N., and Mykka¨nen, L. (2000). Carotid artery atherosclerosis in type-2 diabetic and nondiabetic subjects with and without symptomatic coronary artery disease (The Insulin Resistance Atherosclerosis Study). Am. J. Cardiol. 85, 1395–1400. Tuomilehto, J., Lindstrom, J., Eriksson, J. G., Valle, T. T., Hamalainen, H., Ilanne-Parikka, P., Keinanen-Kiukaanniemi, S., Laakso, M., Louheranta, A., Rastas, M., Saliminen, V., and Uusitupa, M. (2001). Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N. Engl. J. Med. 344, 1343–1350. World Health Organization. (1998). Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus: Provisional report of a WHO consultation. Diabetic Med. 15, 539–545.

276

Glycation- and/or Polyol Pathway-Inducing Complications

A

Amadori rearrangement

+ RNH2

Glucose

Schiff's base

Fructosamine

- RNH2 + RNH2 and/or Arg

+ RNH2 and/or Arg

+ RNH2 and/or Arg

AGEs

Glycolytic intermediates B

Lipid peroxidation

Early glycation adducts: OH OH

C O + NH2-(CH2)4-CH N H

O OH

HO OH

+ NH2

O OH

HO OH

R

Potent glycating agents:

C

O

HOCH2 H

O H

H

CH3

O H

Ne-(1-Deoxy-D-fructose-1-yl)lysine

Na-(1-Deoxy-D-fructose-1-yl)amino acid

O

Glyoxal

H

D Hydroimidazolones CO H

HC (CH2)3 NH N

HN

CH3

O

H

3-Deoxyglucos one

N

NH

O

CO

H

HC (CH2)3 NH

G-H

HN N

NH

O

MG-H

CH2 H

HC (CH2)3 NH

O

3DG-H

Monolysyl adducts

HOCH2 CO

CO HC

H

HOCH2(CHOH)2 H HN

NH

H O

Methylglyoxal

"FRUCTOSAMINE"

CO

HO H

O

OH

HC

(CH2)4 NHCH2CO2-

(CH2)4

HC

NH-C H CO2-

NH

NH

CO

CH3

(CH2)4

N

NH H

Ne-Carboxymethyl-lysine (CML)

Ne-Carboxyethyl-lysine (CEL)

Bis(lysyl)imidazolium crosslinks

CH2(CHOH)2CH2OH CH3

CO

CO (CH2)4

HC

N + N

(CH2)4

NH

Others: CO HC NH

CH NH

CO (CH2)4

N + N

(CH2)4

CH

NH

NH

CO HC

CO (CH2)4

CO

(CH2)4

HC CO CH

NH

(CH2)4 CH NH

DOLD CH3

+ N

N + N

NH

MOLD

N HN

H2C

CO HC

GOLD

(CH2)3-NH

O

Pyrraline

N (CH2)3 NH

OH N

CO HC NH

CH3

N

CH3 CO2OH

(CH2)3 NH HN

OH CH3

NH

Pentosidine crosslink (fluorophore)

Argpyrimidine (fluorophore)

Argpyrimidine (fluorophore)

277

Glycation- and/or Polyol Pathway-Inducing Complications

efficient enzymatic detoxification of these aoxoaldehydes. Glyoxal and methylglyoxal are detoxified to glycolate and D-lactate, respectively, by the glutathione-dependent glyoxalase system. 3Deoxyglucosone is detoxified by NADPH-dependent 3-deoxyglucosone reductase to 3-deoxyfructose. In addition, there is a high cysteinyl thiol pool that binds a-oxoaldehydes reversibly and thereby suppresses their irreversible reactions to form AGEs. Fructosamine residues are removed from proteins by phosphorylation to fructoseamine-3-phosphate, catalyzed by fructosamine 3-phosphokinase, and degraded to 3-deoxyglucosone. Once formed in vivo, AGEs must be repaired or replaced and degraded. The reactions of saccharide derivatives with proteins, nucleotides, and phospholipids to form AGEs occurs intra- and extracellularly. Inside cells, the impact of glycation is countered by high turnover (short half-life) of many cellular proteins, phospholipids, and RNA as well as by mechanisms of DNA repair. Degradation of extracellular glycated proteins occurs by specific recognition by receptors, internalization, and proteolytic processing of the glycated protein ligand.

PROTEIN GLYCATION IN MONITORING GLYCEMIC CONTROL Fructosamine concentrations change in response to persistent hyperglycemia but are unresponsive to short postprandial hyperglycemia. The assay of fructosamine residues is a diagnostic marker of medium-term glycemic control in diabetes mellitus. Approximately 6 to 15% of the human serum albumin (HSA) is glycated with a fructosamine in vivo, primarily at Lys-525. Glycated HSA reflects glycemic control during the 2 to 3 weeks preceding analysis. Glycated hemoglobin HbA1 accounts for approximately 7.5% of total hemoglobin in normal human individuals. The various forms are designated HbA1a1, HbA1a2, HbA1b, and HbA1c. HbA1c is the most abundant of the minor components in normal human red blood cells in vivo, accounting for approximately 5% of total hemoglobin. It too is a mixture of mostly the fructosamine adducts of the b-val-1 (60%) and a-Lys-61 (40%). HbA1a1 and HbA1a2 are the b-chain N-terminal adducts of

fructose-1,6-bisphosphate and glucose-6-phosphate, respectively. HbA1b is thought to result from a deamidation in the b-chain of HbA1c. The measurement of HbA1c reflects glycemic control over the 6 to 8 weeks preceding analysis. In diabetes mellitus, the concentration of fructosamine of serum proteins is typically increased twofold relative to that of normal human individuals (approximately 5 vs 2 nmol/mg protein). Glycated hemoglobin HbA1c is typically increased from 5% of total hemoglobin in normal controls to
7% (good glycemic control), 7 to 10% (moderate control), and >10% (poor control). There is no significant increase of fructosamine and glycated hemoglobin in prediabetic impaired glucose tolerance.

PROTEIN GLYCATION A RISK MARKER OF DIABETIC COMPLICATIONS Glycated hemoglobin is a risk factor for the development of chronic clinical complications, probably as a surrogate indicator of hyperglycemia. Increased concentrations of AGE have been associated with diabetic complications in cross-sectional studies. AGE content of skin collagen (CML, pentosidine, and others) was shown to be a risk marker for microvascular complications of diabetes. Serum protein AGE content at baseline was shown to be a risk predictor for development of early morphological kidney damage leading to diabetic nephropathy.

FUNCTIONAL CONSEQUENCES OF PROTEIN GLYCATION Glycation occurs inside and outside of cells. Glycation of nucleotides in DNA normally induces DNA repair but may induce apoptosis if glycation is excessive. Glycation of basic phospholipids may induce changes in membrane structure and lipid peroxidation. Glycation of cellular proteins produces changes in structure and loss of enzymatic activity. These effects are countered by protein degradation and renewal. Glycation of the extracellular matrix produces changes in macromolecular structure affecting matrix–matrix and matrix cell interactions associated with decreased

Figure 1 Formation of early glycation adducts and advanced glycation end products: (A) glycation adducts, (B) early glycation adducts (fructosamines), (C) reactive a-oxoaldehydes in physiological glycation process, and (D) molecular structures of AGEs. 3DG-H, Nd-[5-(2,3,4-trihydroxybutyl)-5-hydro-4-imidazolon-2-yl]ornithine; DOLD, 3-deoxyglucosone-derived lysine dimer, 1,3-di(Ne-lysino)-4-(2,3,4-trihydroxybutyl)-imidazolium salt; G-H, Nd-(5-hydro-4-imidazolon-2-yl)ornithine; GOLD, glyoxalderived lysine dimer, 1,3-di(Ne-lysino)imidazolium salt; MG-H, Nd-(5-hydro-5-methyl-4-imidazolon-2-yl)-ornithine; MOLD, methylglyoxal-derived lysine dimer, 1,3-di(Ne-lysino)-4-methyl-imidazolium salt.

278 elasticity and increased fluid filtration across arterial walls and endothelial cell adhesion. When the concentration of AGEs increases above a critical level, cell surface AGE receptors become activated. This is associated with increased expression of extracellular matrix proteins, vascular adhesion molecules, cytokines, and growth factors. Depending on the cell type and concurrent signaling, this is associated with chemotaxis, angiogenesis, oxidative stress, and cell proliferation or apoptosis. These processes are thought to contribute to disease mechanisms associated with the development of diabetic complications. Strategies for therapeutic interventions to counter the effects of glycation by preventing increased aoxoaldehyde formation in hyperglycemia, scavenging of a-oxoaldehydes, and blocking of AGE receptor activation are in experimental development, and all hold promise for the eventual prevention of diabetic complications clinically.

POLYOL PATHWAY AND DIABETIC COMPLICATIONS p0035

Aldose reductase (ALR2) is the first enzyme in the polyol pathway (Fig. 2). It catalyzes the NADPHdependent reduction of glucose and many other carbonyl compounds to sorbitol and the corresponding alcohol derivatives. ALR2 has a low affinity (high KM) for glucose. Hence, there is an increased flux of glucose metabolism via the polyol pathway in hyperglycemia; in human red blood cells, glucose metabolism increases from 3% of total glucose metabolism in normoglycemia to 11% in hyperglycemia. Sorbitol produced in the first step of the polyol pathway is oxidized by NADþ-dependent sorbitol dehydrogenase (SDH) to fructose in the second step. The complete traverse of the polyol pathway involves oxidation of NADPH to NADPþ and reduction of NADþ to NADH, that is, a net hydride transfer from NADPH in the pentosephosphate pathway to NADþ in the Embden–Meyerhof pathway. Therefore, there is a net reduction of NADþ to NADH in the Embden– Meyerhof pathway and an increase in the cytosolic NADH/NADþ ratio (Fig. 2). The effects of activation of the polyol pathway in diabetes have been attributed to osmotic stress arising from intracellular accumulation of sorbitol, decreased NaþKþ ATPase activity, in situ inhibition of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) by change in the NADH/NADþ ratio, and in situ inhibition of glutathione reductase by depletion of NADPH-exacerbating intracellular oxidative stress. Diabetes induced in homozygous knockout mice deficient in ALR2 neither decreased the

Glycation- and/or Polyol Pathway-Inducing Complications

Figure 2 The polyol pathway. glutathione content of sciatic nerve nor had decreased motor nerve conduction velocity—in contrast to wildtype mice. Studies of the inhibition of the polyol pathway in vivo with specific inhibitors have given inconsistent results. An ALR2 inhibitor prevented the development of diabetic neuropathy in dogs but failed to prevent retinopathy or thickening of the capillary basement membrane in the retina, kidney, and muscle. Several clinical trials have been performed with negative outcomes. The prevention of clinical diabetic neuropathy with a potent aldose reductase inhibitor was found in a multiple-dose, placebocontrolled trial. Effect on the redox balance is the most likely mechanism by which increased flux through the polyol pathway has deleterious consequences.

See Also the Following Articles Diabetic Nerve Disease, Neuropathy . Diabetes, Type 1 . Glucose Physiology, Normal . Glucose Toxicity . Glycoproteins

Further Reading Brownlee, M. (2001). Biochemistry and molecular cell biology of diabetic complications. Nature 414, 813–820. Carington, A. L., and Litchfield, J. C. (1999). The aldose reductase pathway and nonenzymatic glycation in the pathogenesis of diabetic neuropathy: A critical review for the end of the 20th century. Diab. Rev. 7, 275–299. Greene, D. A., Arezzo, J. C., and Brown, M. B. (1999). Effect of aldose reductase inhibition on nerve conduction and morphometry in diabetic neuropathy. Neurology 53, 580–591. McCance, D. R., Dyer, D. G., Dunn, J. A., Baiue, K. E., Thorpe, S. R., Baynes, J. W., and Lyons, T. J. (1993). Maillard reaction products and their relation to complications in insulindependent diabetes mellitus. J. Clin. Invest. 91, 2470–2478. Thornalley, P. J. (1998). Cell activation by glycated proteins: AGE receptors, receptor recognition factors, and functional classification of AGEs. Cell. Molec. Biol. 44, 1013–1023. Thornalley, P. J. (1999). Clinical significance of glycation. Clin. Lab. 45, 263–273. Williamson, J. R., Chang, K., Frangos, M., Hasan, K. S., Ido, T., Kawamura, T., Nyengaard, J. R., van den Enden, M., Kilo, C., and Tilton, R. G. (1993). Hyperglycaemic pseudohypoxia and diabetic complications. Diabetes 42, 801–813.

f0010

280

Glycoproteins

a

NH 2

52

CGb NH 2

13

LHb

13

NH 2

TSHb

NH 2

FSHb

NH 2

30

30

23

7

24

78

COOH

121 132 127 138

COOH

COOH

COOH

COOH

Figure 1 Structures of the subunits of human glycoprotein hormones. N-linked sugar chains and O-linked sugar chains are shown by ‘‘trees’’ and ‘‘boxes,’’ respectively, together with their positions. The N-linked sugar chains of a-subunit are those of hCG, and the a-subunit of other glycohormones have different sugar chains. Dotted portions of the trees in the b-subunits of LH, TSH, and FSH indicate heterogeneity in branching structures. LHb in the case of human hormone is glycosylated at the Asn 30. In contrast, a sugar chain is linked only at the Asn 13 of LHb of most other animal species. Reproduced, with slight modifications, from Bielinska and Boime (1995).

sugar chains of glycoproteins. However, comparative study of the N-linked sugar chains of hCGa and hCGb revealed that sialylated N5 is distributed mainly in hCGb and that sialylated N8 is distributed mainly in hCGa, whereas sialylated N6 is detected as a major sugar chain of both subunits. The specific distribution of different N-linked sugar chains at the four N-glycosylation sites of hCG molecule cannot be explained by our current knowledge of the biosynthetic mechanism of the N-linked sugar chains. An unknown control mechanism involving the steric effects of the polypeptide moiety may play a role in the formation of N-linked sugar chains of hCG. This assumption was supported by the study of the N-linked sugar chains of the free a-subunit. A small amount of a-subunit occurs in free form in the urine of pregnant women. Interestingly, this free a-subunit cannot bind to hCGb, in contrast to hCGa dissociated from hCG heterodimer. Because the free a-subunit has the same amino acid sequence as does hCGa, it was assumed that the free a-subunit contains different sugar chains from hCGa. Structural studies of the sugar chains of the free a-subunit revealed that it contains sialylated N5 as its major sugar chains. This evidence indicated that bulky sialylated N5 on the free a-subunit may sterically inhibit its association with hCGb. Maturation of the N-linked sugar chains of the free a-subunit to larger complex-type sugar chains might be induced because

the subunit did not bind to b-subunit. Therefore, uneven distribution of the N-linked sugar chains at the four N-glycosylation sites of hCG may be produced only when the two subunits are associated before the N-linked sugar chains start maturation at the Golgi apparatus. Later on, the structures of the N-linked sugar chains of LH and TSH of various mammals, including human (hLH and hTSH), were elucidated (as shown in Fig. 4), as were those of human FSH (hFSH) (as shown in Fig. 5). In contrast to hCG, the three pituitary glycoprotein hormones contain triantennary and tetraantennary complex-type sugar chains, and their sialic acids are linked at the C-3 and C-6 positions of galactose residues. Occurrence of bisected sugar chains was also found. The most interesting evidence is that a part of the N-linked sugar chains of hLH and hTSH contain the SO4–4GalNAcb1–4GlcNAc group as their outer chains. An N-acetylgalactosaminyltransferase, which forms the GalNAcb1–4GlcNAc group, was found to occur in the pituitary but not in the placenta. This enzyme requires presence of the –Pro–X–Arg– motif in the polypeptide portion of the substrate glycoprotein. Such sequence is present in hLH and hTSH but not in hFSH, explaining the absence of N-acetylgalactosamine residue in hFSH. The enzyme responsible for sulfation of the sugar chains of pituitary

p0040

281

Glycoproteins

-

Fuca1 Galb1 - 4GlcNAcb1 - 2Mana1 6 6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn Galb1 - 4GlcNAcb1 3 4 Mana1 Galb1 - 4GlcNAcb1 2

N1

Galb1 - 4GlcNAcb1 - 2Mana1 6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn Galb1 - 4GlcNAcb1 3 4 Mana1 Galb1 - 4GlcNAcb1 2 Fuca1 -

N2

N3

Mana1 6 6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3 4 Mana1 Galb1 - 4GlcNAcb1 2

Galb1 - 4GlcNAcb1

Mana1

6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3 4 Mana1 Galb1 - 4GlcNAcb1 2 Fuca1 -

N4

Galb1 - 4GlcNAcb1

N5

Galb1 - 4GlcNAcb1 - 2Mana1 Galb1 - 4GlcNAcb1 - 2Mana1 Galb1 - 4GlcNAcb1 - 2Mana1

N6 Galb1 - 4GlcNAcb1 - 2Mana1

6 6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3

6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3

N7

Galb1 - 4GlcNAcb1 - 2Mana1

N8

Mana1 Galb1 - 4GlcNAcb1 - 2Mana1

N9

N10

Mana1 - 3Mana1

-

Fuca1 Mana1

6 6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3

6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3

6 Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3

Galb1 - 4GlcNAcb1 - 2Mana1 Mana1 6 Mana1 3 6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn Mana1 3 Galb1 - 4GlcNAcb1 - 2Mana1

Figure 2 Structures of desialylated N-linked sugar chains isolated from various hCG preparations. All sialic acid residues are exclusively linked at the C-3 position of the galactose residues. Gal, galactose; GlcNAc, N-acetylglucosamine; Man, mannose; Fuc, L-fucose.

glycoprotein hormones was detected in the Golgi membrane preparation of pituitary gland by the transfer of sulfate residues from 30 -phosphoadenosine 50 phosphosulfate. This 4-O-N-acetylgalactosamine sulfotransferase does not require any specific peptide motif and can transfer sulfate even to the trisaccharide: GalNAcb1–4GlcNAcb1–2Man. Therefore, addition of a b-N-acetylgalactosamine residue to the sugar chains is a prerequisite for the sulfation of the sugar chains. Both FSH and LH are synthesized by gonadotrophs in the anterior pituitary. Although they share the same a-subunit polypeptide, their sugar chains

differ in branching and terminal modification. Therefore, assembly of the two subunits should control the maturation of the sugar chains, as discussed previously for the site-directed maturation of the sugar chains of hCG. Actually, assembly of the two subunits of glycoprotein hormones was found to occur when the N-linked sugar chains are still in the state of high mannose types, which are sensitive to endo-b-N-acetylglucosaminidase H digestion. In further processing and maturation steps to lead to complex-type sugar chains, association to different b-subunits will result in different sugar chain formation.

282

Glycoproteins

Neu5Aca2

O-a

6 Neu5Aca2 - 3Galb1 - 3GalNAc

Neu5Aca2 - 3Galb1 - 4GlcNAcb1

O-b

6 Galb1-3GalNAc

Figure 3 Structures of O-linked sugar chains found in hCG. Neu5Ac, N-acetylneuraminic acid; GalNAc, N-acetylgalactosamine. Other abbreviations are the same as in Fig. 2.

hCGs OF TROPHOBLASTIC DISEASES A large amount of hCG is detected in the sera and urine of patients with various trophoblastic diseases such as hydatidiform mole and choriocarcinoma. Some of the hydatidiform moles show malignant characteristics, such as invasion into the surrounding tissues and metastasis, and are discriminated by the term ‘‘invasive mole.’’ Structural studies of the sugar chains of hCG samples purified from the urine of these patients revealed many interesting structural alterations that can be used for the diagnostic purposes. The N-linked sugar chain patterns of the urinary hCG samples, obtained from patients with hydatidiform mole, were identical to those of the samples from healthy pregnant women. In contrast, quite different patterns were obtained from the hCG samples, purified from the urine of patients with choriocarcinoma. In all choriocarcinoma hCGs, eight

+ - Mana1

6

Mana1 3 6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn + - Mana1 3

R' - 4 GlcNAcb1 -2Mana1

Mana1 R or R' - 4GlcNAcb1 -2Mana1 R or R' - 4GlcNAcb1 -2Mana1 R or R' - 4GlcNAcb1 -2Mana1

6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3

6Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn 3

R - 4GlcNAcb1 -2Mana1 6 GlcNAcb1 - 4Manb1 - 4GlcNAcb1 - 4GlcNAc - Asn R - 4 GlcNAcb1 -2Mana1 3

Figure 4 Structures of N-linked sugar chains found in LH and TSH. R represents the Neu5Aca2–3 or 6Galb1– group, and R0 represents the SO4–4GalNAcb1– group. Abbreviations of the sugar units are the same as in Figs. 2 and 3.

oligosaccharides, N1 to N8 as shown in Fig. 2, were detected in either their sialylated or nonsialylated forms. Interestingly, the hCG samples purified from the urine of patients with invasive mole gave another alteration in their glycosylation pattern. These samples contained sialylated N1, N2, N5, N6, N7, and N8 but not sialylated N3 and N4. Although the structural alteration found in the sugar chains of choriocarcinoma hCGs is quite complicated, it can be explained by the changes induced in the two glycosyltransferases. An increase in the molar ratio of total fucosylated oligosaccharides (N1, N3, N5, and N7 in Fig. 2) indicated that expression of the fucosyltransferase responsible for formation of the Fuca1–6GlcNAc group is enhanced in choriocarcinoma cells. Appearance of oligosaccharide N7 may indicate that the enhanced fucosyltransferase has wider specificity than that in normal syncytiotrophoblasts. Structurally, oligosaccharides N1, N2, N3, and N4 can be formed by addition of the Galb1–4GlcNAcb1–4 outer chain to N5, N6, N7, and N8, respectively. Therefore, Nacetylglucosaminyltransferase IV (GnT-IV), which forms the GlcNAcb1–4Man group and is suppressed in syncytiotrophoblasts, must be strongly expressed in choriocarcinoma. Detection of N1 and N2 in invasive mole indicated that GnT-IV is expressed in these cells as well. However, absence of N3 and N4 indicates that the newly expressed GnT-IV can transfer an Nacetylglucosamine residue to biantennary sugar chains but not to monoantennary sugar chains. Cloning of the structural gene of GnT-IV from a human liver cDNA library revealed that two active GnT-IV genes are present. The translation products of these two genes were named GnT-IVa and GnTIVb. When activities of the five glycosyltransferases (GnT-I to GnT-V), which are related to the formation of the antennary portion of the complex-type sugar chains, were studied comparatively in normal placenta and several choriocarcinoma cell lines, only GnT-IV activity was strikingly increased in the cancer cells. Northern blot analysis revealed that the GnT-IVa gene was strongly overexpressed in the cancer cells, whereas the GnT-IVb gene was expressed at the same level as in the normal placenta. So far, no difference in the substrate specificities of GnT-IVa and GnT-IVb has been found. The data have indicated that overexpression of the GnT-IVa gene and the resulting increase in GnT-IV activity are the enzymatic basis of formation of the abnormal sugar chains in choriocarcinoma cells. An enzymatic basis to form the different sugar patterns of choriocarcinoma and invasive mole hCGs remains to be elucidated.

283

Glycoproteins

-

-

-

+ - Fuca1 + Neu5Aca2 3(6)Galb1 4GlcNAcb1 2Mana1 6 6 + - 4GlcNAcb1 - 4GlcNAc - Asn - GlcNAcb1 - 4Manb1 3 + - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 - 2Mana1 + + - Fuca1 - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 6 Mana1 2 6 + 6 - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 + - 4GlcNAcb1 - 4GlcNAc - Asn -GlcNAcb1 - 4Manb1 3 + - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 - 2Mana1 + - Fuca1 + Neu5Aca2 3(6)Galb1 4GlcNAcb1 2Mana1 6 6 + - 4GlcNAcb1 - 4GlcNAc - Asn -GlcNAcb1 - 4Manb1 + - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 4 Mana1 3 2 + - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 -

+ + - Fuca1 - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 6 Mana1 2 6 + 6 - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 + - 4GlcNAcb1 - 4GlcNAc - Asn -GlcNAcb1 - 4Manb1 + - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 4 Mana1 3 2 + - Neu5Aca2 - 3(6)Galb1 - 4GlcNAcb1 Figure 5 Structures of the N-linked sugar chains found in hFSH. Abbreviations of the sugar units are the same as in Figs. 2 and 3. Because oligosaccharides N1 and N2 were detected in invasive mole and choriocarcinoma hCGs but not in normal pregnant hCGs and hydatidiform mole hCGs, this difference can be used for the diagnosis of malignant trophoblastic diseases. Actually, a Datura stramonium agglutinin– Sepharose column, which has an affinity to the Galb1–4GlcNAcb1–4(Galb1–4GlcNAcb1–2)Man group, can discriminate hCGs in the urine of patients with malignant diseases from those in the urine of pregnant women and patients with hydatidiform mole. Alteration of the O-linked sugar chains of hCG by malignant transformation was also found. hCG samples obtained from patients with hydatidiform mole, invasive mole, and choriocarcinoma all contain both O–a and O–b in Fig. 3. However, the proportion of O–b increased prominently in choriocarcinoma. Although hydatidiform mole hCGs contain a similar proportion of the sialylated tetrasaccharide to normal hCG, the proportion increased moderately but significantly in invasive mole hCGs.

FUNCTIONAL ROLE OF THE SUGAR CHAINS OF GLYCOPROTEIN HORMONES The specific distribution of different sugar chains in the two subunits of hCG indicated that even the

smallest N-linked sugar chain, such as sialylated N8, may work as an important signal in expression of the hormonal action of hCG. That modification of the N-linked sugar chains of hCG alters its hormonal activity was proposed by many studies. Complete removal of sialic acid residues from hCG reduces its hormonal activity (as measured by cyclic AMP (cAMP) production and steroidogenesis in the target cells) to 50%, although its binding to target cells is enhanced. Removal of the whole N-linked sugar chains from hCG further increases the binding of hCG to its target cells but completely eliminates its hormonal activity. These results indicate that absence of the sugar chains dissociates receptor binding of hCG from its signal transduction. The deglycosylated hCG behaves as an antagonist to native hCG. It was also reported that the glycopeptides mixture, obtained from hCG by exhaustive pronase digestion, blocked hCG signal transduction. These results suggest that binding of hCG to a lectin-like membrane component in addition to an hCG receptor is necessary to induce the signal transduction. Comparative study of the biological activities of hCGs, in which only one of the N-glycosylation sites was eliminated by recombinant DNA technology, revealed that N-glycosylation at the Asn-52 of hCGa is essential for the expression of hormonal activity, whereas removal of the N-linked sugar chains at either Asn-13 or Asn-30 of hCGb or Asn-78 of hCGa had no effect. Other important evidence

p0080

284 shown by this line of study is that removal of Asn-78 glycosylation of hCGa markedly reduces its assembly with hCGb. Glycosylation of the two N-glycosylation sites of hCGb is not essential for its assembly with hCGa, but elimination of Asn-30 glycosylation inhibits the secretion of uncombined hCGb. These results indicated that the N-linked sugar chains of hCG are important for constructing the correct conformation of each subunit. That the presence of at least one N-acetylglucosamine residue at each of the four N-glycosylation sites of hCG is enough to keep the correct folding of the two subunits was shown two decades ago by investigating the effects of digestion with exo- and endoglycosidases on the folding and assembly of the two subunits. Although exact structures of the sugar chains essential for proper expression of the functional role of hCG were not presented, the role of sialic acid residues was further investigated. As described previously, removal of all sialic residues of the N-linked sugar chains of hCG, which exclusively occur as the Neu5Aca2–3Gal group, reduces the hormonal activity of hCG to 50%. When the desialylated hCG was resialylated by incubation with CMP–Neu5Ac and Galb1–4GlcNAc:a2–6 sialyltransferase, the isomeric hCG containing only the Neu5Aca2–6Gal group gave almost the same dose–response curve as did the natural hCG. This recovery of hormonal activity is not obtained by the addition of the Gala1–3 residue to the nonreducing terminal galactose residues. Interestingly, extensive sialylation of desialylated hCG reduced the hormonal activity of the isomeric hCG. Further investigation revealed that sialylation of the outer chain on the Mana1–3 arm, rather than the Mana1–6 arm, of the N-linked sugar chains of hCG is favorable for the signal transduction. Addition of 2 mM N-acetylneuraminic acid hexamer, obtained by incomplete sialidase digestion of colominic acid, to the mixture of hCG and MA-10 cells, a mouse Leydig tumor cell line established by Ascoli, revealed that the oligosaccharide did not inhibit the binding of hCG to the surface of the target cells but that cAMP production was reduced to 50%. This result indicated that the hexasaccharide can inhibit the interaction of the sialic acid residues of hCG with the specific binding site on the cell surface but does not influence the binding of the peptide portion of hCG to the hCG receptor. The possibility that the action of the sialic acid hexamer may be due to a nonspecific anionic polymer effect was defuted because the addition of fucoidin did not show any inhibition of the [3H]hCG binding to the cell surface receptor or cAMP production by hCG.

Glycoproteins

Presence of the sialic acid binding site on the surface of MA-10 cells was confirmed by using 30 -sialyllactoseconjugated bovine serum albumin as a probe. Based on the data indicating that sialic acid residues bind directly to the cell surface, a model of the hCG–receptor complex was proposed. A dual interaction of the peptide portion and the sialylated sugar chain of hCG with respective binding sites is essential for the signal transduction. Interestingly, a region homologous to the soybean lectin was found in the hCG receptor. Compared to hCG, the functional roles of the sugar chains of pituitary hormones were not investigated aggressively. Like other serum glycoproteins, sialylation of the N-linked sugar chains of glycoprotein hormones will protect them from clearance by hepatic galactose-binding receptor. In contrast, sulfation of the N-linked sugar chains of hLH and hTSH leads to more rapid clearance of them. It was reported that sulfated LH is cleared from circulation by binding to a receptor, which specifically recognizes the SO4–4GalNAcb1–4GlcNAc group, on the surface of hepatic endothelial cells. Fully sialylated hTSH, made by recombinant technique, showed a longer half-life in circulation than did natural TSH. Because the sulfation versus sialylation pattern of TSH is regulated by thyrotropin-releasing hormone, the rapid clearance may be important in controlling the serum level of this glycohormone.

See Also the Following Articles FSH (Follicle-Stimulating Hormone) . Glycation- and/or Polyol Pathway-Inducing Complications . LH (Luteinizing Hormone) . TSH Function and Secretion

Further Reading Baenziger, J. U., Kumar, S., Brodbeck, R. M., Smith, P. L., and Beranek, M. C. (1992). Circulatory half-life but not interaction with the lutropin/chorionic gonadotropin receptor is modulated by sulfation of bovine lutropin oligosaccharides. Proc. Natl. Acad. Sci. USA 89, 334–338. Bielinska, M., and Boime, I. (1995). The glycoprotein hormone family: Structure and function of the carbohydrate chains. In ‘‘Glycoproteins’’ ( J. Montreuil, H. Schachter, and J. F. G. Vliegenthart, eds.), pp. 565–587. Elsevier, Amsterdam. Fiete, D., Srivastava, V., Hindsgaul, O., and Baenziger, J. U. (1991). A hepatic reticuloendothelial cell receptor specific for SO4– 4GalNAc beta 1,4GlcNAc beta 1,2Man alpha that mediates rapid clearance of lutropin. Cell 67, 1103–1110. Kobata, A., and Amano, J. (1995). Structures and function of the N-linked sugar chains of glycohormones. In ‘‘Biopolymers and Bioproducts: Structure, Function, and Application’’ ( J. Svasti, ed.), pp. 210–217. Samakkhisan Publishing, Bangkok, Thailand.

Glycoproteins

Kobata, A., and Takeuchi, M. (1999). Structure, pathology, and function of the N-linked sugar chains of human chorionic gonadotropin. Biochim. Biophys. Acta 1455, 315–326. Nemansky, M., DeLeeuw, R., Wijands, R. A., and Van den Eijnden, D. H. (1995). Enzymic remodelling of the N- and O-linked carbohydrate chains of human chorionic gonadotropin: Effects on biological activity and receptor binding. Eur. J. Biochem. 227, 880–888. Skelton, T. P., Hooper, L. U., Srivastava, V. V., Hindsgaul, O., and Baenziger, J. U. (1991). Characterization of a sulfotransferase responsible for the 4-O-sulfation of terminal beta-N-acetylD-galactosamine on asparagine-linked oligosaccharides of glycoprotein hormones. J. Biol. Chem. 266, 17142–17150.

285 Takamatsu, S., Oguri, S., Yoshida, A., Minowa, T., Nakamura, K., Takeuchi, M., and Kobata, A. (1999). Unusually high expression of N-acetylglucosaminyltransferase-IVa in human choriocarcinoma cell lines: A possible enzymatic basis of the formation of abnormal biantennary sugar chain. Cancer Res. 59, 3949–3953. Varki, A., Cummings, R., Esko, J., Freeze, H., Hart, G., and Marth, J. (eds.) (1999). ‘‘Essentials of Glycobiology,’’ Cold Spring Harbor Laboratory Press, New York. Weishaar, G., Hiyama, J., and Renwick, A. G. C. (1991). Sitespecific N-glycosylation of human chorionic gonadotropin: Structural analysis of glycopeptides by one- and two-dimentional 1 H NMR spectroscopy. Glycobiology 1, 393–404.

287

Goitrogens, Environmental

Thiocyanate Isothiocyanates

Goitrin Flavonolds Resorcinol and phenolics DHBAs Pyridines

Iodide Lithium

Iodide transport

Oxidation Organic binding and coupling

Proteolysis release and dehalogenation

Thyroglobulin I-

I-

MIT DIT

T4 T3

MIT DIT

T4 T3

I-

Figure 1 Environmental antithyroid–goitrogenic compounds and their site of action in the thyroid gland. Goitrin, L,5-vinyl-2thiooxazolidone; DHBAs, dihydroxy-benzoic acids; I, iodide; MIT, monoiodotyrosine; DIT, diiodotyrosine; T4, thyroxine; T3, triiodothyronine. Reproduced with permission from the Annual Review of Nutrition, volume 10, (ß)1990 by Annual Reviews.

affected by goiter. Cassava, a staple food in these areas, has antithyroid effects in humans and experimental animals. Thus, daily consumption of cassava, in the presence of severe iodine deficiency, is thought to be the cause of endemic goiter and cretinism in these areas of Zaire. The goitrogenic action of cassava is due to endogenous release of thiocyanate from linamarin, a cyanogenic glucoside present in cassava, particularly in the tuberous roots. Thiocyanate is also present in Pearl millet, the staple food of people living in iodinedeficient endemic goiter areas of western Sudan. Pearl millet is rich in C-glycosylflavones, which in combination with thiocyanate exert additive and complementary antithyroid and goitrogenic effects. Thiocyanate is also found in high concentrations (1 g/L) in wastewater effluents of coal conversion processes and in body fluids as a metabolite of hydrogen cyanide gas consumed while smoking. Studies in Sweden indicate that cigarette smoking may produce goiter. Similarly, goiter and hypothyroidism were documented in patients receiving long-term thiocyanate treatment for hypertension. This goitrogenic effect of thiocyanate is more evident in the presence of iodine deficiency. Several observations suggest that thiocyanate crosses the human placenta and may cause both goiter and neonatal hypothyroidism.

Thiocyanate Thiocyanate and thiocyanate-like compounds primarily inhibit the iodine-concentrating mechanism of the thyroid, and their goitrogenic activity can be overcome by iodine administration (Fig. 1). Thiocyanate at low concentrations inhibits iodide transport by increasing the velocity constant of iodide efflux from the thyroid gland. At high concentrations, the iodide efflux is greatly accelerated, whereas the unidirectional iodide clearance into the gland is inhibited. Thiocyanate at these high concentrations also inhibits the incorporation of iodide into thyroglobulin by competing with iodide at the thyroid peroxidase (TPO) level. Thiocyanate is rapidly converted to sulfate in the thyroid gland. Administration of thyroid-stimulating hormone (TSH) increases the intrathyroidal catabolism of thiocyanate and is capable of reversing the block of iodide uptake produced by this ion. Isothiocyanates The isothiocyanates and cyanogenic glycosides act on the thyroid mainly by their rapid conversion to thiocyanate. However, isothiocyanates also react spontaneously with amino groups to form thiourea derivatives, which produce a thiourea-like antithyroid effect. Isothiocyanates also possess intrinsic antithyroid activity, as demonstrated by in vitro inhibition of iodide uptake in the case of methyl- and allylisothiocyanates and of both iodide uptake and organification in the case of butyl-isothiocyanate. Thio-Oxazolidone (Goitrin) The thionamide or thiourea-like goitrogens interfere in the thyroid gland with the organification of iodide and formation of the active thyroid hormones, and their action usually cannot be antagonized by iodine. Naturally occurring goitrin is representative of this category (Fig. 1). Long-term administration of goitrin to rats results in increased thyroid weight and decreased radioactive iodide uptake and hormone synthesis by the thyroid gland. Actually, goitrin possesses 133% of the potency of propylthiouracil in humans. Goitrin is unique in that it is not degraded like thioglycosides. Additive antithyroidal effects of thiocyanate, isothiocyanate, and goitrin also occur with combinations of these naturally occurring goitrogens.

Disulfides The small aliphatic disulfides (R–S–S–R; R ¼ methyl-, ethyl-, n-propyl, phenyl-), the major components of

288 onion and garlic, exert marked thiourea-like antithyroid activity. None of these disulfides inhibits in vitro the TPO enzyme, but fractions with sulfur-bearing organic compounds, possibly aliphatic disulfides from the goitrogenic well supplying a Colombian district with endemic goiter, inhibited in vitro 125 I-organification. Disulfides are also present in high concentrations (0.3–0.5 g/L) in aqueous effluents from coal conversion processes, and they have also been identified as water contaminants in the United States, where the most frequently isolated compounds are dimethyl, diethyl, and diphenyl disulfides.

FLAVONOIDS Flavonoids are important stable organic constituents of a wide variety of plants. Flavonoids are universally present in vascular plants and in a large number of food plants. Because of their widespread occurrence in edible plants such as fruits, vegetables, and grains, flavonoids are an integral part of the human diet. They are present in high concentrations in polymeric (tannins) and oligomeric (pigments) forms in various staple foods in the Third World such as millet, sorghum, beans, and ground nuts. Flavonoids are polyhydroxyphenolic compounds with a C6–C3–C6 structure. Mammalian organisms are unable to synthesize the flavone nucleus. Flavonoids are strictly exogenous food components of exclusively vegetable origin. They have high chemical reactivity with multiple important biological implications. Flavonoids are quickly metabolized in higher organisms, and that is the reason why they are not found in normal tissue constituents. Most flavonoids are present as b-glucosides that cannot be absorbed in tissues. No mammalian enzymes have been found that deglycosylate these compounds to their bioactive aglycone species. Following ingestion by mammals, flavonoid glycosides are hydrolyzed by intestinal microbial glycosidases to flavonoid aglycones. These may be absorbed and undergo metabolism by mammalian tissues, or they may be further metabolized by intestinal micro-organisms to undergo B-ring hydroxylation and middle-ring fission, with production of various metabolic monomeric compounds, including phenolic acids, phloroglucinol, resorcinol, and gallic acid. Each metabolic step is characterized by a marked increase in antithyroid effects. Flavonoid aglycones, such as apigenin and luteolin present in Fonio millet (Digitaria

Goitrogens, Environmental

exilis), and a variety of flavonoid metabolites (e.g., phloroglucinol, resorcinol, phenolic acids) are several times more potent than the parent glycosides glucosylvitexin, glucosylorientin, and vitexin present in Pearl millet (Pennisetum [L.] leeke, also known as typhoides or americanum), as inhibitors of TPO, the enzyme-catalyzing iodide oxidation and hormone synthesis in the thyroid gland. This greater inhibitory effect is further enhanced by the additive effects exerted by mixtures of flavonoid aglycones and flavonoid metabolites that are formed after ingestion of mixtures of flavonoid glycosides present in many plant foodstuffs. In addition, these metabolic products may produce adverse effects on other parameters of thyroid function not observed with the glycosides. As a result, the antithyroid effects of flavonoid glycosides in foodstuffs may be greatly enhanced by metabolic alterations after ingestion by mammals, as in the case of the flavonoids present in the Pearl millet grain, the staple food of people living in iodine-deficient endemic goiter areas of western Sudan, which make a major contribution to and are primarily responsible for its antithyroid and goitrogenic effects. Furthermore, antithyroid effects in vivo of vitexin, one of the three major flavonoids in Pearl millet, has been demonstrated to provide evidence that C-glycosylflavones are the goitrogens in this cereal grain. It is of interest that a significant portion of the flavonoids isolated from Fonio millet, the staple food of people living in the severely affected endemic goiter area of Guinea in Western Africa, are already present as the aglycones apigenin and luteolin, with more potent antithyroid activity than their parent glycosides. Flavonoids not only inhibit TPO but, acting on iodothyronine deiodinase enzymes, also inhibit the peripheral metabolism of thyroid hormones. Flavonoids also affect serum thyroid hormone binding and thyrotropin (TSH) regulation. Thus, this class of compounds alters thyroid hormone economy in a complex manner. At this point, there is substantial evidence indicating, first, that various millet species used as staple food by the populations in the semi-arid tropics are rich in flavonoids; second, that flavonoids have potent and diverse antithyroid properties; and third, that under the appropriate environmental dietary conditions of low iodine and protein–calorie intakes, which are prevalent in most countries of the Third World, flavonoids become an important etiological determinant of endemic goiter and hypothyroidism.

Goitrogens, Environmental

POLYHYDROXYPHENOLS AND PHENOL DERIVATIVES Coal is a source of a large variety of antithyroid and goitrogenic compounds such as, phenol, dihydroxyphenols (resorcinol), substituted dihydroxybenzenes, thiocyanate, disulfides, phthalic acids, pyridines, and halogenated and polycyclic aromatic hydrocarbons (PAH) (Table I). Most of these compounds have been identified in drinking water from the iodinesufficient goitrous areas of Kentucky (United States) and Colombia (South America). Phenolics are the major organic pollutants in wastewater effluents from various types of coal treatment processes. Resorcinol, substituted resorcinols, and other antithyroid phenolic pollutants are present at levels of as high as 5 g/L in coal-derived effluents. Up to 8% of shale bitumen is also composed of phenols. Phenol, dihydroxyphenols, trihydroxyphenols, and halogenated phenols are readily absorbed from the gastrointestinal tract. Phenol, resorcinol, and catechol, in suitable preparations, are readily absorbed through human skin. Essentially all phenols, polyhydroxyphenols, and halogenated phenols are readily absorbed after injection. A major route of metabolism of polyhydroxyphenols, polyhydroxyphenolic acids, and halogenated phenols is by conjugation to glucuronic or sulfuric acids. The major route of excretion of these compounds is the urinary tract, and various amounts of the free parent compound and its monoglucuronide and monosulfate conjugates are excreted in the urine. Resorcinol, the prototype of this group of compounds, is antithyroid and goitrogenic both in man and in experimental animals. During the early 1950s, the goitrogenic effect of resorcinol was demonstrated when patients applying resorcinol ointments for the treatment of varicose ulcers developed goiter and hypothyroidism. Several observations also suggest that resorcinol crosses the human placenta and may cause both goiter and neonatal hypothyroidism. Resorcinol has been shown both in vivo and in vitro to inhibit thyroidal organification of iodide. A comparison of the antiperoxidase activity of resorcinol (1,3dihydroxybenzene), catechol (1,2-dihydroxybenzene), and hydroquinone (1,4-dihydrozybenzene) (Table I) indicates the importance of hydroxyl groups in the meta position for maximal activity. Furthermore, the net antiperoxidase effects of mixtures of dihydroxyphenols, as well as dihydroxyphenols and thiocyanate (also a coal-derived pollutant), are equivalent to or greater than the sum of the effects produced by individual compounds, indicating that the true goitrogenic potential of the major water-soluble compounds present

289 in coal and shales are due to the combined effects of the individual constituents rather than to any single compound. Demonstration in vivo and in vitro of antithyroid and goitrogenic activities of coal–water extracts from iodine-sufficient goiter areas indicate that shaleand coal-derived organic pollutants appear to be a major factor contributing to the high goiter prevalence and associated disorders observed in certain areas with aquifers and watersheds rich in these organic rocks. Studies of the physical state of organic goitrogens in water indicate that the active compounds form dissociable complexes and that they are part of larger organic molecules, possible humic substances (HS). Furthermore, resorcinol and other parent antithyroid phenolic and phenolic-carboxylic compounds are degradation monomeric by-products of reduction, oxidation, and microbial degradation of HSs. HSs, high-molecularweight complex polymeric compounds, are the principal organic compounds of soils and waters. More than 90% of total organic matter in water consists of HSs, which are also present in coals and shales. Decaying organic matter becomes the substrate of lignin and flavonoid types of HS during the process of fossilization (or coalification). Actually, cyanidin, a naturally occurring flavonol used as a model subunit of flavonoid-type HS, yields the following antithyroid compounds by reductive degradation: resorcinol, phloroglucinol, orcinol, and 3,4-dihydroxybenzoic acid (Table I). Demonstration in vivo and in vitro of antithyroid effects of vitexin, a major C-glucosylflavone in Pearl millet, provides evidence that flavonoid structures are the link for phenolic goitrogens in foodstuffs (e.g., millet) and those present in coals, shales, soils, and water, all of which are an obligatory step and integral part of the biogeochemical cycle of organic–phenolic goitrogens in nature (Fig. 2). In addition to thiocyanate, cigarette smoke contains various goitrogenic resorcinol derivatives, flavonoids, and hydroxypyridines. As mentioned previously, cigarette smoking may produce goiter, and smoking increases the severity and metabolic effects of hypothyroidism, probably by alteration of both thyroid function and hormone action. The presence of halogenated organic compounds with known or potential harmful effects has prompted public health and environmental concerns. These compounds are produced by the chlorination of water supplies, sewage, and power plant cooling waters. Present at micrograms-per-liter concentrations (parts per billion) in treated domestic sewage and cooling waters, 4-chlororesorcinol and 3-chloro4-hydroxybenzoic acid possess antithyroid activities as inhibitors of TPO and thyroidal iodide organification.

Table I Environmental Agents Producing Goitrogenic and/or Antithyroid Effects Goitrogenic/Antithyroid effects In vivo Compound

Humans

Animals

In vitro

Sulfurated organics a

Thiocyanate (SCN) Isothiocyanates

þ NT

L-5-vinyl-2-thiooxazolidone (goitrin)

þ þ

þ þ

þ

þ

þ

NT

þ

0,þ(?)

Glycosides

NT

þ

þ

Aglycones

NT

þ

þ

C-ring fission metabolites (e.g., phloroglucinol, phenolic acids)

NT

þ

þ þ

Disulfides (R–S–S–R) Flavonoids (polyphenols)

Polyhydroxyphenols and phenol derivatives Phenol

NT

NT

NT

NT

þ

þ

þ

þ

Hydroquinone (1,4-dihydroxybenzene)

NT

NT

þ

m-Dihydroxyacetophenones

NT

NT

þ

2-Methylresorcinol

NT

þ

þ

5-Methylresorcinol (orcinol)

NT

þ

þ

4-Methylcatechol Pyrogallol (1,2,3-trihydroxybenzene)

NT NT

NT þ

þ þ

Phloroglucinol (1,3,5-trihydroxybenzene)

NT

þ

þ

4-Chlororesorcinol

NT

þ

þ

3-Chloro-4-hydroxybenzoic acid

NT

NT

þ

þ

þ

0

NT

NT

þ

NT

þ

þ

Diisobutyl phthalate

NT

NT

0

Dioctyl phthalate

NT

þ

0

o-Phthalic acid

NT

NT

0

Catechol (1,3-dihydroxybenzene) a

Resorcinol (1,3-dihydroxybenzene)

2,3-Dinitrophenol Pyridines 3-Hydroxypyridine Dihydroxypyridines Phthalate esters and metabolites

m-Phthalic acid

NT

NT

0

3,4-Dihydroxybenzoic acid (DHBA)

NT

NT

þ

3,5-Dihydroxybenzoic acid

NT

NT

þ

NT

þ

NT

þ

þ

NT

Polychlorinated (PCB) and polybrominated (PBB) biphenyls PCBs (Aroclor) PBBs and PBB oxides Other organochlorines Dichlorodiphenyltrichloroethane (p,p1-DDT) Dichlorodiphenyldichloroethane (p,p1-DDE) and dieldrin

NT

þ

NT

NT

þ

NT

2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD)

NT

þ

NT

3.4-Benzpyrene (BaP) 3-Methylcolanthrene (MCA)

NT NT

þ(?) þ

NT NT

7,12-Dimethylbenzanthracene (DMBA)

NT

þ

NT

9-Methylanthracene (MA)

NT

þ

0

þ

þ

þ

þ

þ

þ

Polycyclic aromatic hydrocarbons (PAHs)

Inorganics a

Excess iodine a

Lithium

Note. þ, active; 0, inactive; NT, nontested. a Agents also used as medications. b Inactive in TPO assay; active (?) in thyroid slices assay.

b

291

Goitrogens, Environmental

Oxidation and polymerization

Decomposition Plants

Polyphenol monomers

Fossilization and coalification HS

Coal and shale

(Flavonoids and lignin)

Figure 2 Simplified scheme of the biogeochemical cycle of phenolic goitrogens and interrelationships among plant flavonoids (e.g., millet), polyhydroxyphenols, humic substances (HS), coals, and shales. Reproduced with permission from the Annual Review of Nutrition, volume 10, (ß)1990 by Annual Reviews. Whether these pollutants exert additive or synergistic antithyroid effects or act as ‘‘triggers’’ of autoimmune thyroiditis, or both, requires investigation, particularly because more than 60 soluble chloro-organics have been identified in the primary and secondary effluents of typical domestic sewage treatment plants. Derivatives of 2,4-dinitrophenol (DNP) are widely used in agriculture and industry. An insecticide, a herbicide, and a fungicide, DNP is also used in the manufacture of dyes, to preserve timber, and as an indicator as well as being a by-product of ozonization of parathion. DNP is readily absorbed through intact skin and respiratory tract. DNP causes toxicity by the uncoupling of oxidative phosphorylation in the mitochondria of cells throughout the body. Administration of 2,4-DNP to human volunteers resulted in rapid and pronounced decline of circulating thyroid hormones. A decrease in TSH secretion results in decreased synthesis and release of thyroxine (T4) and triiodothyronine (T3) and possibly involution of the thyroid gland. The antithyroid effect of 2,4-DNP is due in part to an inhibition of the pituitary TSH mechanism. Once T4 and T3 are released into the circulation, they are instantaneously bound to serum carrier proteins. DNP also interferes with T4 binding, further decreasing serum T4 concentration. In addition to inhibiting the TSH mechanism and interfering with T4 binding, DNP accelerates the disappearance of T4 from the circulation; thus, the serum concentration is lowered even more. The public health impact of this pollutant on the thyroid is still unknown.

s0045

PYRIDINES Hydroxypyridines also occur in aqueous effluents from coal conversion processes as well as in cigarette smoke. Dihydroxypyridines and 3-hydroxypyridine are potent inhibitors of TPO, producing effects comparable to or greater than those of propylthiouracil. After ingestion, mimosine, a naturally occurring amino acid in the seeds and foliage of the tropical

legume Leucaena leucocephala, is metabolized to 3,4dihydroxypyridine (3,4-DHP), a potent antithyroid agent that produces goiter in mammals. 3,4-DHP crosses the placenta barrier, producing goitrous offspring. The phenolic properties of the 3-hydroxy group in various hydroxypyridines are reflected in the metabolism of these compounds in vivo. 3hydroxypyridine fed to rabbits is converted to ethereal glucuronide and sulfate conjugates. 3,4-DHP glucuronide and sulfate conjugates account for the majority of 3,4-DHP in the blood of cattle grazing on leucaena. The ring structure of dihydroxypyridines does not appear to be broken down in the body and also appears to be relatively stable to bacterial degradation.

PHTHALATE ESTERS AND METABOLITES Phthalic acid esters, or phthalates, are ubiquitous in their distribution and have been frequently identified as water pollutants. Dibutyl phthalates (DBPs) and dioctyl phthalates (DOPs) have been isolated from water-supplying areas of endemic goiter in western Colombia and eastern Kentucky. Although phthalate esters are most commonly the result of industrial pollution, they also appear naturally in shale, crude oil, petroleum, plants, and fungal metabolites and as emission pollutants from coal liquefaction plants. Phthalate esters are well absorbed from the gastrointestinal tract. Prior to intestinal absorption, there is hydrolysis to the corresponding monoester metabolite. This is particularly true of the longer chain derivatives such as DOPs. Phthalates are widely distributed in the body, with the liver being the major initial repository organ. Clearance from the body is rapid. Short-chain phthalates can be excreted unchanged or following complete hydrolysis to phthalic acid. Prior to excretion, most longer chain compounds are converted, by oxidative metabolism, to polar derivatives of the monoesters. The major route of phthalate esters elimination from the body is urinary excretion. Phthalate esters are commonly used as plasticizers to impart flexibility to plastics, particularly polyvinylchloride polymers (PVCs), which have a wide variety of biomedical and other uses such as building and construction, home furnishings, cars, clothing, and food wrappings. A small fraction of phthalate esters are used as nonplasticizers for pesticide carriers, oils, and insect repellents. Phthalates may be present in concentrations of up to 40% of the weight of the plastic.

292 Phthalate esters are known to leach out from finished PVC products into blood and physiological solutions. The entry of these plasticizers into a patient’s bloodstream during blood transfusion, intravenous fluid administration, or hemodialysis has become a matter of concern among public health officials and the medical community. A high incidence of goiter in patients receiving maintenance hemodialysis has been reported. It remains to be determined whether phthalate ester metabolites, contaminants in the water entering the patient’s bloodstream, or middle molecules (e.g., hydroxybenzoic and vanillic acids) that accumulate in uremic serum and are poorly removed by hemodialysis are responsible for this condition. Although phthalate esters and phthalic acids do not possess intrinsic antithyroid activity (Table I), they undergo degradation by gram-negative bacteria to form dihydroxybenzoid acid (DHBA). DHBAs are known to possess antithyroid properties (Table I). The 3,4- and 3.5-DHBAs also inhibit in vitro TPO and the incorporation of iodide into thyroid hormones. The proven effective role of gram-negative bacteria in phthalate biodegradation may explain, in part, the relationship established between frequency of goiter and bacterial contamination of water supplies. Furthermore, marked ultrastructural changes of the thyroid gland, similar to those seen after administration of TSH, and decreased serum T4 concentration have been observed in rats treated with phthalic acid esters. Thus, phthalates may become goitrogenic under appropriate conditions and are also actively concentrated and metabolized by several species of fish. Whether these widely distributed pollutants exert deleterious effects on the thyroids of humans has not been investigated.

POLYCHLORINATED AND POLYBROMINATED BIPHENYLS These are aromatic compounds containing two benzene nuclei with two or more substituent chlorine or bromine atoms. They have a wide variety of industrial applications, including electric transformers, capacitors, and heat transformers. There is growing evidence that atmospheric transport is the primary mode of global distribution of polychlorinated biphenyls (PCBs) from sites of use and disposal. Plant foliage accumulates the vapor of PCBs from the atmosphere. In addition to their occurrence in surface water (e.g., rivers, lakes), PCBs have been detected in drinking water. Perhaps the most significant human exposures are limited to individuals consuming freshwater fish

Goitrogens, Environmental

from contaminated streams and lakes and to occupational exposure of industrial workers. PCBs can also be found in the milk of nursing mothers who have eaten large amounts of sport fish or who have been occupationally exposed. Currently, PCBs are targeted by bioremediation strategies, and some strains of Pseudomonas spp. (Pseudomonas cepacia) can degrade these stable aromatic pollutants. PCBs and polybrominated biphenyls (PBBs) have high-lipid solubility and resistance to physical degradation. They are slowly metabolized, and their excretion is limited. Long-term low-level exposure to the organohalides results in their gradual accumulation in fat, including the fat of breast milk. PCBs have been found in the adipose tissue of 30 to 45% of the general population. The biological and toxicological properties of PCB mixtures may vary depending on their isomeric composition. Oral administration of PCBs to various mammals results in rapid and nearly complete (90%) intestinal absorption. The degradation and elimination of PCBs depend on the hepatic microsomal enzyme system. The excretion of PCBs is related to the extent of their metabolism. Those with greater chlorine content have a correspondingly longer biological half-life in mammals. This resistance to metabolism is reflected in their deposition in adipose tissue. The PCBs, however, have very low acute toxicity in all animal species tested, and PBBs have biological properties similar to PCBs. Despite the lack of evidence that dietary PCBs and PBBs have any deleterious effects on health, there is a growing concern and uncertainty about the long-range effects of bioaccumulation and contamination of our ecosystem with these chemicals. The uncertainty extends to the potential harmful effects of these pollutants on the thyroid. For instance, an increased prevalence of primary hypothyroidism (11%) was documented among workers from a plant that manufactured PBBs and PBB oxidases. These individuals had elevated titers of TPO–microsomal antibodies, indicating that hypothyroidism was probably a manifestation of lymphocytic autoimmune thyroiditis, perhaps a PBB-induced pathogenic autoimmune response or exacerbation of underlying subclinical disease. PCBs are potent hepatic microsomal enzyme inducers. Rats exposed to PCBs exhibit a greatly enhanced biliary excretion of circulating T4. The T4 is excreted as a glucuronide that is then lost in the feces. This response is probably secondary to induction of hepatic microsomal T4–uridine diphosphate– glucuronyl transferase. The enhanced peripheral metabolism and reduced binding of T4 to serum proteins

293

Goitrogens, Environmental

in PCB-treated animals result in markedly decreased serum T4 concentrations. These low levels stimulate the pituitary–thyrotropin–thyroid axis, and this eventually results in goiter formation. Although PCB-treated animals exhibit decreased serum T4, their T3 levels are unchanged. The relative iodine deficiency brought about by the accelerated metabolism of T4 may induce increased thyroidal T3 secretion as well as increased peripheral deiodination of T4 to T3. PBBs appear to act similarly to PCBs. There is, however, some indication that they may also interfere directly with the process of hormonal synthesis in the thyroid gland.

OTHER ORGANOCHLORINES DDT (2,2-bis(p-chlorophenyl)-1,1,1-trichlorethane) is polychlorinated and nondegradable. The substance is practically insoluble in water and is resistant to destruction by light and oxidation. Its stability has created difficulties in residue removal from water, soil, and foodstuffs. The dominant degradative reaction of DDT is dehydrochlorination to form DDE (2,2-bis(p-chlorophenyl)-1,1-dichloroethylene), which has the same low solubility in water and high lipid–water partitioning as did its precursor. This substance is almost undegradable, both biologically and environmentally. Dieldrin is one of the cyclodiene insecticides. It is almost insoluble in water and, like DDT and DDE, is very stable, both environmentally and biologically. DDT has been used extensively, both in malaria control and in agriculture, all over the world. Because of biomagnification and persistence, DDT and its breakdown products, DDE and DDD (dichlorodiphenyldichloroethane), are ubiquitous contaminants of water and of virtually every food product. Most of the fish from Lake Michigan in North America contain DDT residues. The substance is also present in milk; humans are at the top of the food pyramid, so human milk is especially contaminated. The situation is basically similar for dieldrin, which is found in surface waters virtually everywhere. Dieldrin is heavily bioconcentrated in the lipids of terrestrial and aquatic wildlife, humans, and foods, especially animal fats and milk. Global distribution of high concentrations of organochlorines, including DDT, DDE, DDD, and dieldrin, were recently found not only in developing countries but also in industrialized countries, which continue to be highly contaminated even though the use of many of these compounds is restricted. DDT is reductively dechlorinated in biological systems to form DDE and DDD. DDE, the predominant

residue stored in tissues (reaching about 70% in humans), is much less toxic than DDT. DDE is eliminated from the body slowly, and little is known about its degradation pathway. DDT is also eliminated from the human body slowly through reduction to DDD and other more water-soluble derivatives. DDT is known to cause marked alterations in thyroid gland structure such as thyroid enlargement, follicular epithelial cell hyperplasia, and progressive loss of colloid in birds, and DDD is known to cause goiter and increased hepatobiliary excretion of thyroid hormones in rats. All of these compounds (DDT, DDE, DDD, and dieldrin) induce microsomal enzyme activity that may affect thyroid hormone metabolism in a similar way to that of the polyhalogenated biphenyls and PAH. The impact of these pollutants on the human thyroid is unknown. Dioxin (or tetrachlorodibenzodioxin, TCDD), one of the most toxic small organic molecules, is a contaminant in the manufacturing process of several pesticides and herbicides, including Agent Orange. Also a potent inducer of hepatic microsomal enzymes, TCDD markedly enhances the metabolism and biliary excretion of T4–glucuronide. Rats treated with TCDD concomitantly develop hypothyroxinemia, increased serum TSH concentrations, and goiter, probably as a result of T4 loss in the bile. The impact on the thyroid of humans exposed to this agent is unknown, and studies of thyroid gland function and thyroid hormone metabolism in those affected are needed.

POLYCYCLIC AROMATIC HYDROCARBONS PAHs have been found repeatedly in food and domestic water supplies as well as in industrial and municipal waste effluents. They also occur naturally in coal, soils, ground water, and surface water as well as in their sediments and biota. One of the most potent of the carcinogenic PAH compounds, 3,4-benzpyrene (BaP), is widely distributed and, as in the case of other PAHs, is not efficiently removed by conventional water treatment processes. The PAH carcinogens, BaP and 3-methylcolanthrene (MCA) by enhancement of hepatic UDP– glucuronyltransferase and glucuronidation, accelerate T4 metabolism and excretion of T4–glucuronide, resulting in decreased serum T4 concentrations, activation of the pituitary–thyrotropin–thyroid axis, and (eventually) goiter formation. There is also indication that MCA interferes directly with the process of hormonal synthesis in the thyroid gland. Furthermore, MCA, as well as 7,12-dimethylbenzanthracene,

294 induces goitrous thyroiditis in the BUF rat. Thus, MCA exerts its deleterious effects on the thyroid gland by at least three different mechanisms. Finally, the coal-derived PAH methylanthracene (MA), which has been identified in drinking water from the goitrous coal-rich district of eastern Kentucky, was found to produce goiter in the BUF rat without alteration of hormone synthesis or lymphocytic infiltration of the thyroid gland.

See Also the Following Articles Graves’ Disease . Hypothyroidism, Causes of . Iodine Deficiency . Nontoxic Goiter . Smoking and the Thyroid . Thyroid Autoimmunity . Thyroid Carcinoma . Thyroid Disease, Epidemiology of . Thyroid Gland, Anatomy and Physiology . Toxic Multinodular Goiter . TSH Function and Secretion

Further Reading Cody, V., Middleton, E., Jr., and Harborne, J. B., eds. (1986). ‘‘Plant Flavonoids in Biology and Medicine: Biochemical,

Goitrogens, Environmental

Pharmacological, and Structure–Activity Relationships.’’ Liss, New York. Cody, V., Middleton, E., Jr., Harborne, J. B., and Beretz, A., eds. (1988). ‘‘Plant Flavonoids in Biology and Medicine II: Biochemical, Cellular, and Medicinal Properties.’’ Liss, New York. Dumont, J. E., Lamy, F., Roger, P., and Maenhaut, C. (1992). Physiological and pathological regulation of thyroid cell proliferation and differentiation by thyrotropin and other factors. Physiol. Rev. 72, 667–697. Gaitan, E., ed. (1989). ‘‘Environmental Goitrogenesis.’’ CRC, Boca Raton, FL. Gaitan, E. (1990). Goitrogens in food and water. Annu. Rev. Nutr. 10, 21–39. Gaitan, E. (1996). Flavonoids and the thyroid. Nutrition 12, 127–129. Gaitan, E. (1997). Environmental goitrogens. In ‘‘Diseases of the Thyroid’’ (L. E. Braverman, ed.), pp. 331–348. Humana Press, Totowa, NJ. Gaitan, E., and Dunn, J. T. (1992). Epidemiology of iodine deficiency. Trends Endocrinol. Metab. 3, 170–175. Gaitan, E., Nelson, N. C., and Poole, G. V. (1991). Endemic goiter and endemic thyroid disorders. World J. Surg. 15, 205–215. Lindsay, R. H., Hill, J. B., Gaitan, E., Cooksey, R. C., and Jolley, R. L. (1992). Antithyroid effects of coal-derived pollutants. J. Toxicol. Envir. Health 37, 467–481. Weetman, A. P., and McGregor, A. M. (1994). Autoimmune thyroid disease: Further development in our understanding. Endocrine Rev. 15, 788–830.

296

Gonadotropin-Induced Ovulation

Table 1 Comparision of Various Ovulation Induction Agents Agent

Mechanism of action

How given

Indications

Multiple pregnancy rate (percentage)

Clomiphene citrate

Oral

Blockade of estrogen receptors

Ovulatory dysfunction Superovulation

8

Dopamine agonists

Oral

Dopamine angonist

Hyperprolactinemia

1

GnRH

Subcutaneous or intramuscular pump

Restoration of GnRH pulses

Hypothalamic amenorrhea

1

Gonadotropins Human menopausal gonadotropins Follitropin

Subcutaneous or intramuscular injections

Direct stimulation of ovarian follicle growth

Ovulatory dysfunction Superovulation

Bromocriptine Cabergoline Pergolide

20–25

Note. GnRH, gonadotropin-releasing hormone.

is not discussed here because it is not used in humans. The categories of gonadotropins in clinical use and their trade names are summarized in Table II. hMG products were the first, and for many years the only, gonadotropins available for human use.

Table II Gonadotropin Preparations Drug Human menopausal gonadotropins (hMG)

Urofollitropin (urinary derived)

Trade names Humegon

Merional

Lepori Menogon

Pergonal Pertisol

Menopur

Repronex

Bravelle Fertinex Fertinorm Fostimon Metrodin

Follitropin (recombinant)

Follistim Gonal-F

Lutropin (LH) [not available–United States]

Luveris

Human chorionic gonadotropin (hCG) (urinary derived)

A.P.L. Choragon

Puregon

Novarel Pregnyl Profasi

Human chorionic gonadotropin (hCG) (recombinant)

Ovidrel Ovitrelle

They contain an equal mixture of FSH and LH and are extracted from the urine of postmenopausal donors. The medications must be reconstituted before each use and are given intramuscularly or subcutaneously. In the United States, they are sold under the trade names Pergonal and Repronex. The earliest preparations of purified FSH were extracted from urinary sources much like hMG (e.g., urofollitropin, Metrodin) and were indicated for intramuscular use only. Techniques were then developed for purification of urinary products to enable subcutaneous administration (e.g., Fertinex, Bravelle). Today, most FSH used comes from recombinant DNA technology and mammalian cell culture. Recombinant products provide a much higher degree of purity and batch-to-batch consistency and include follitropin–a (Gonal-F) and follitropin-b (Follistim). They are designed for subcutaneous use but may be injected intramuscularly as well. Until recently, all preparations required fresh daily reconstitution from powder, although one multidose formulation is currently available in the United States. The final stages of ovulation induction require a surge in circulating luteinizing hormone (LH) levels. In gonadotropin cycles, LH is substituted with human chorionic gonadotropin (hCG). Because LH and hCG work via the same receptor and hCG has a far longer circulating half-life, hCG is a logical agent to induce ovulation. hCG has been available for many years as a urinary-derived product for intramuscular injection. More recently, these older products, as well as newer recombinant products, have been given by subcutaneous injection.

p0030

297

Gonadotropin-Induced Ovulation

INDICATIONS FOR THERAPY There are essentially two categories of indications for treatment with gonadotropins. The first subset is the use of these medications in anovulatory women. This group may be further broken down into two groups. The first group consists of women with hypogonadotropic hypogonadism, also known as hypothalamic amenorrhea, or WHO class I. These women do not have regular menstrual cycles due to very low levels of endogenous gonadotropins, are hypoestrogenic, and do not respond to clomiphene. They are prime candidates for hMG injections and have excellent pregnancy rates. The other subgroup consists of anovulatory women with normal estrogen levels and includes women with polycystic ovarian syndrome and luteal phase defects. Often, these patients have failed clomiphene- and/or insulin-sensitizing agents. The goal in all anovulatory women on gonadotropins is the production and release of one or two mature oocytes. The second category for gonadotropin use is for controlled ovarian hyperstimulation (COH), also known as superovulation. With this technique, higher doses of gonadotropins are used to recruit multiple follicles, usually as an adjunct to intrauterine insemination (IUI) or in vitro fertilization (IVF). Indications for COH with IUI or IVF include unexplained, tubal, cervical, male, and endometriosis-related infertility. The objective in non-IVF gonadotropin cycles is to recruit no more than three or four mature eggs, but with IVF even more oocytes may be desirable.

THE TREATMENT CYCLE

p0055

The gonadotropin treatment cycle begins on the third day of the menstrual cycle, commonly referred to as ‘‘day 3.’’ At this time, a transvaginal ultrasound is performed, and blood is obtained for estradiol and FSH levels. On that day, gonadotropin injections are begun. Doses range from 75 to 600 IU of FSH or hMG daily or some combination thereof. They may be daily or twice a day and may be intramuscular or subcutaneous. Patients then return to the clinic on a frequent basis over the next 7 to 10 days for serial ultrasounds to monitor follicular growth. At the same visits, serum estradiol levels are obtained as an adjunct in follicular monitoring, and serum LH levels are checked to ensure that premature luteinization does not occur. One should note that this monitoring protocol may differ significantly from clinic to clinic.

When the dominant follicle(s) reaches a mature size as seen on ultrasound, usually somewhere between 16 and 20 weeks, the decision is made to trigger ovulation. Human chorionic gonadotropin at 5,000 to 10,000 IU or 250 mg is given by intramuscular or subcutaneous injection. Approximately 36 h later, ovulation will occur, and at this time IUI is performed or timed intercourse is encouraged. If an IVF procedure is planned, oocyte retrieval will be performed at this time. During the subsequent week, many clinicians will choose to provide support of the luteal phase. This may be accomplished by supplemental lower dose hCG injections or progesterone by various routes. Ovulation may be confirmed by a mid-luteal serum progesterone level. Gonadotropin cycles may be performed back to back without skipping a month in between provided that no large cysts are noted on the interval baseline ultrasound. Doses of gonadotropins may be adjusted in between or during cycles. The course of therapy is continued for three or four cycles. If it is unsuccessful, IVF or other treatments may be contemplated.

SIDE EFFECTS Gonadotropins are potent medications, and although they are generally safe in experienced hands, they clearly have the potential to cause serious problems when not used with caution. Common side effects include headache, abdominal discomfort, pain at the injection site, and mood swings. One may also see skin reactions when hMG is used subcutaneously. More troublesome is the potential for multiple gestation and ovarian hyperstimulation. The chance for multiple pregnancy among women who achieve pregnancy on gonadotropins is 20 to 25%. Although twins and triplets predominate, quadruplets and other higher order multiple gestations have become more commonplace. This phenomenon is due almost exclusively to gonadotropin use. The risks for many complications of pregnancy, including preterm delivery, diabetes, and preeclampsia, increase exponentially with increasing numbers of feti. Therefore, it behooves the clinician to limit the number of mature follicles permitted and, if necessary, to recommend cycle cancellation to the patient. Multifetal reduction may be used to reduce triplet or higher gestations down to twins. Although fraught with ethical and emotional concerns, this procedure can significantly reduce maternal and fetal morbidity

298 in such cases. However, as stated earlier, the best treatment is prevention. The other serious concern is the development of the ovarian hyperstimulation syndrome (OHSS). This disorder, which occurs only with ovulation induction, may in some cases involve massive ovarian enlargement, ascites, abdominal discomfort, hydrothorax, and/or a hypercoagulable state. The syndrome has been classified into mild, moderate, and severe, based on symptomatology, ovarian size, and the degree of hemoconcentration. Treatment is generally supportive, with maintenance of urine output, bed rest, and monitoring for increase of severity. The use of albumin infusions for prophylaxis remains controversial. More severe cases may require hospitalization, paracentesis, and/or (occasionally) prophylaxis for deep venous thrombosis, intensive care unit (ICU) admission, and dopamine drips. More extensive reviews can be found elsewhere. As with multiple gestation, the best treatment is prevention. The long-term concern with gonadotropin therapy revolves around the issue as to whether these medications increase the risk for ovarian cancer. Whittemore and colleagues demonstrated an elevated risk for ovarian tumors (malignant and low-malignant potential) in patients on various ovulation-inducing drugs, a risk that disappeared when treatment was successful. Rossing and colleagues showed an increased risk of ovarian cancer in women on long-term clomiphene but not short-term clomiphene or gonadotropins. Several studies from large IVF programs that use gonadotropins suggested no change in the risk of ovarian carcinoma. In contrast, Shushan and colleagues found a threefold increase in ovarian cancer in women on gonadotropin preparations. In summary, one can conclude that the bulk of the literature suggests no effect of these drugs on cancer risk, but the question is far from answered.

FUTURE DIRECTIONS The future of ovulation induction likely will involve a refinement in the agents available to the clinician as well as the development of protocols for optimal stimulation. Recombinant DNA technology has enabled pharmaceutical companies to develop recombinant FSH products during recent years. These new products have enabled easier delivery of the drug via the subcutaneous route and have been a great leap forward in quality control. In the future, it might even be possible to develop an orally active

Gonadotropin-Induced Ovulation

gonadotropin via emerging technologies, further enhancing ease of use. Extension of molecular engineering techniques may also enable the development of ‘‘designer gonadotropins’’ in which the potency and half-life may be altered. One might even envision a time where very specific molecules might be engineered to meet the needs of individual patients. By optimizing the stimulation of each patient, we may better succeed in the concurrent goals of improving pregnancy rates while minimizing multiple gestation and ovarian hyperstimulation.

CONCLUSION Gonadotropin-induced ovulation involves the use of FSH with or without LH to stimulate the growth of ovarian follicles, followed by injection of hCG to induce release of the oocytes. This treatment may be used for women who are anovulatory but refractory to clomiphene or for controlled ovarian hyperstimulation in preparation for insemination or IVF. These drugs are potentially dangerous and require monitoring with serial ultrasounds and estradiol levels. Women on injectable gonadotropins are at risk for multiple gestation and the ovarian hyperstimulation syndrome. For these reasons, gonadotropin therapy is best prescribed by clinicians familiar with the use of the drugs and with the management of the associated complications.

See Also the Following Articles FSH (Follicle-Stimulating Hormone) . Gonadotropin-Releasing Hormone (GnRH) Actions . In Vitro Fertilization (IVF) . Infertility, Overview . LH (Luteinizing Hormone) . Ovarian Hyperstimulation Syndrome . Ovarian-Follicular Apparatus . Superovulation and Intrauterine Insemination

Further Reading Adashi, E. Y., and McClamrock, H. C., eds. (1994). ‘‘Managing the Anovulatory State: Medical Induction of Ovulation’’ [ACOG technical bulletin]. American College of Obstetricians and Gynecologists, Washington, DC. Baird, D. T. (2001). Is there a place for different isoforms of FSH in clinical medicine? IV. The clinician’s point of view. Hum. Reprod. 16, 1316–1318. Daya, S., and Gunby, J. (2000). Recombinant versus urinary follicle stimulating hormone for ovarian stimulation in assisted reproduction cycles. Cochrane Database Syst. Rev. 4, CD002810. Dodson, W. C. (1998). ‘‘Induction of Ovarian Follicle Development and Ovulation with Exogenous Gonadotropins’’ [report]. American Society for Reproductive Medicine, Birmingham, AL. Whelan, J. G., and V lahos, N. F. (2000). The ovarian hyperstimulation syndrome. Fertil. Steril. 73, 883–896.

303

Gonadotropin-Releasing Hormone Deficiency, Congenital Isolated

Table I Summary of Gene Mutations Known to Cause Hypogonadotropic Hypogonadism in Humans Gene

Principle effects

KAL1

Hypothalamic (disrupts neuronal migration)

X-linked recessive

AHC

Hypothalamus, pituitary, adrenal

X-linked recessive

LEP

Hypothalamus—leptin deficiency

Autosomal recessive

LEPR

Hypothalamus—leptin resistance

Autosomal recessive

GNRHR

Pituitary—GnRH resistance

Autosomal recessive

PROP1

Pituitary—combined pituitary deficiency

Autosomal recessive

HESX1 FSHb

Pituitary—septo-optic dysplasia Pituitary—isolated FSH deficiency

Autosomal recessive Autosomal recessive

LHb

Pituitary—isolated LH deficiency

Autosomal recessive

neurons fail to synapse normally and Kallmann syndrome ensues. Up to about half of obvious X-linked Kallmann syndrome families possess KAL1 mutations, and of these, half also have unilateral renal agenesis. The frequency of KAL1 mutations in unselected IHH males is considerably less (5–10%), indicating that this X-linked gene is not the most common cause of human IHH, as was originally proposed. Although the sample size is small, males with KAL1 mutations uniformly tend to demonstrate the most severe gonadotropin defect—an apulsatile pattern of LH secretion. No KAL1 gene mutations have been identified in females with IHH and anosmia, suggesting that other autosomal genes may be involved.

AHC Gene Males with adrenal hypoplasia congenita (AHC) display adrenal failure during infancy or childhood because of a failure to form the permanent zone of the adrenal gland. They are deficient in both glucocorticoid and mineralocorticoid hormones. If they are properly treated, they fail to undergo puberty due to IHH. Mutations in the AHC gene have been t0010

Table II KAL1 Expression in Chick and Human Correlates with Phenotypic Findings of KAL1 Mutations (1–10) Olfactory epithelium

Anosmia

Cerebellum Corticospinal tracts

Nystagmus, abnormal balance Synkinesia (mirror movements)

Oculomotor nucleus

Visual abnormalities

Retina

Visual defects

Facial mesenchyme

Midline facial clefting

Mesonephros and metanephros

Unilateral renal agenesis

Limb buds

Pes cavus

Inheritance

demonstrated to cause both adrenal failure (AHC) and IHH. The AHC gene encodes a protein termed DAX1. DAX1 is an orphan receptor (with no known ligand) that belongs to the steroid hormone superfamily. DAX1 is a transcription factor that is important in the development of the adrenal cortex and pituitary gonadotropes and that appears to regulate gonadotropin secretion at both the hypothalamus and the pituitary. A large number of AHC gene mutations have been described in patients with AHC/IHH, most of which are nonsense and frameshift inactivating mutations throughout the gene. Missense mutations occur almost exclusively in the C terminus. Although AHC is an X-linked recessive disease affecting males, one mutation was identified in a female with normal adrenal function and IHH only. Mutations in the AHC gene probably do not cause IHH without adrenal disease given that no mutations were detected by DNA sequencing of DNAs from 100 IHH males. Although the gene was originally hypothesized to be an ovarian determinant gene because of its localization within the DSS region of the X chromosome, a region of Xp that, when duplicated, is associated with sex reversal (feminization) of 46,XY males, conditional knockout studies do not support its role in ovarian determination but do indicate a role in spermatogenesis.

GNRHR Gene The GnRH receptor (GNRHR), a G protein-coupled receptor, constituted the first autosomal-recessive gene identified to possess mutations in human IHH. Although the hypogonadal mouse had a naturally occurring GnRH gene deletion, no GnRH gene mutations had been identified in humans. Several investigators characterized mutations in the larger receptor for GnRH—the GNRHR gene. One group

304 hypothesized that partial loss-of-function GNRHR mutations could occur in patients with incomplete IHH, that is, IHH patients with evidence of puberty. Others suggested that the GNRHR gene was a likely candidate gene for mutation in IHH patients because of the absence of mutations in the ligand and the variability in response to exogenous GnRH that some IHH patients have. More than a dozen GNRHR mutations have been identified to date, but they are almost exclusively compound heterozygous missense mutations affecting binding and/or signal transduction. The phenotypic spectrum of GNRHR mutations producing GnRH resistance ranges from complete IHH (no evidence of puberty or fertility) to incomplete IHH (partial pubertal defects). Even in patients who have no evidence of puberty, GnRH administration may improve pituitary gonadotropin responses, suggesting marginal GNRHR function. The true prevalence of GnRHR mutations among all IHH patients is difficult to assess but in one study was 2.2% (1 of 46) of total IHH patients and 7.1% (1 of 14) of patients where an affected female was present. A more recent study corroborated these findings.

LEP and LEPR Genes Leptin (LEP) possesses important roles in metabolism and puberty. LEP-deficient (ob/ob) and LEP-resistant (db/db) mice manifest obesity and hypogonadotropic hypogonadism. LEP mutations have been described in only a few families with obesity, low serum LEP concentrations, and absent pubertal development. Similar to LEP-deficient ob/ob mice, patients with LEP mutations had extreme obesity, hyperinsulinemia, and hypogonadism, but unlike the mice, they did not have stunted height, hyperglycemia, or hypercortisolemia. An LEP receptor (LEPR) mutation resulted in a similar phenotype of obesity and hypogonadotropic hypogonadism except that serum LEP levels were elevated, indicating LEP resistance. These rare human autosomal-recessive diseases involving LEP and its receptor strongly implicate a role for LEP action in normal pubertal development.

PROP1 Gene Mutations in Prop1, the mouse orthologue of human PROP1, produce the Ames dwarf mouse. Human PROP1 mutations cause combined pituitary hormone deficiency (CPHD), with varying deficiencies of growth hormone, thyroid-stimulating hormone

Gonadotropin-Releasing Hormone Deficiency, Congenital Isolated

(TSH), prolactin, FSH, LH, and (occasionally) adrenocorticotropic hormone (ACTH). Affected individuals display short stature, absent puberty, and hypothyroidism. A variety of PROP1 mutations have been identified, including missense mutations and deletions, but a particular two-base pair deletion is common, being a hot spot in the PROP1 gene for mutations. In a large cohort of patients with IHH only, no PROP1 mutations were identified, indicating that PROP1 mutations generally cause a more universal pituitary failure.

HESX1 Gene Mutations in the HESX1 gene have been identified in some families with septo-optic dysplasia, a disorder characterized by panhypopituitarism, optic nerve atrophy, and other midline central nervous system (CNS) abnormalities such as agenesis of the septum pellucidum and corpus callosum. Both autosomalrecessive and autosomal-dominant HESX1 mutations have been described in this disorder. Hesx1, the mouse orthologue of HESX1, is expressed in the early forebrain and later becomes restricted to Rathke’s pouch, which ultimately becomes the anterior pituitary gland. Hypogonadotropic hypogonadism, due to pituitary gonadotropin deficiency, is a common component of the phenotype in patients with septo-optic dysplasia.

FSHb Gene FSH, LH, TSH, and human choriogonadotropin (hCG) comprise the pituitary glycoprotein hormones. Each dimeric protein consists of a common a-subunit encoded by a single gene and a specific b-subunit. No human a-subunit mutations have been described, but similar to the a-subunit knockout mouse, the phenotype would be expected to include hypogonadotropic hypogonadism and hypothyroidism. It is possible that a-subunit gene mutations could be lethal in humans given that hCG would be predicted to be deficient (mice do not have hCG). Female homozygous FSHb knockout mice had low serum FSH levels and were sterile because of arrested ovarian follicular development. Surprisingly, serum estradiol was normal in these mice despite unmeasurable serum FSH levels. Male homozygous FSHb knockout mice had normal levels of serum testosterone, small testes, and oligospermia but were fertile. The phenotype of humans with FSHb mutations is similar except that estradiol levels are low. Affected

305

Gonadotropin-Releasing Hormone Deficiency, Congenital Isolated

females presented with absent or incomplete breast development, low FSH and estradiol, high LH, and sterility. Despite the elevated LH level similar to that of women with polycystic ovary syndrome (PCOS), hirsutism is not present, suggesting that FSH might be necessary for normal LH-induced androgen production by the theca cells. Males with FSHb mutations present with azoospermia, but puberty may be normal or absent. Contrary to the FSHb knockout mice, humans with mutations suggested that FSH was essential for spermatogenesis in human males. In addition, more severe mutations affect androgen production, manifested clinically by delayed puberty and low testosterone.

LHb–hCGb Genes The LHb–hCGb gene complex, consisting of a single LHb gene and six hCGb genes, is polymorphic, but there is only one known human mutation in the LHb gene. The proband presented with pubertal delay, bilaterally small descended testes, low testosterone, and elevated gonadotropins. He responded to hCG administration, suggesting that the LH ligand might be defective. He was homozygous for a missense mutation of the LHb gene, which was detectable by immunoassay but undetectable by radioreceptor assay. Although LHb polymorphisms have been identified, it is more problematic that they are definitive causes of LH dysfunction; however, they may modify LH effects. No LHb mutations have been identified in females, but the phenotype would be expected to be normal pubertal development and anovulation without hirsutism.

CONCLUSION The genetic basis of IHH remains unknown for most patients; however, mutations in X-linked and autosomal genes may account for approximately 20% of cases. It is highly likely that additional autosomal genes that cause IHH will be identified. However, it is also possible that a large percentage of patients will have complex disease caused by a large number of modifying gene mutations and environmental factors.

Acknowledgments Portions of this work were supported by a grant from the U.S. Public Health Service, National Institute of Child Health and Human Development (HD33004).

See Also the Following Articles Congenital Adrenal Hypoplasia Syndromes . Gonadotropin-Releasing Hormone (GnRH) Actions . GonadotropinReleasing Hormone Receptor Gene, Mutation of . Leptin

Further Reading Achermann, J. C., and Jameson, J. L. (1999). Fertility and infertility: Genetic contributions from the hypothalamic–pituitary–gonadal axis. Mol. Endocrinol. 13, 812–818. Georgopoulos, N. A., Pralong, F. P., Seidman, C. E., Seidman, J. G., Crowley, W. F., Jr., and Vallejo, M. (1997). Genetic heterogeneity evidenced by low incidence of KAL-1 gene mutations in sporadic cases of gonadotropin-releasing hormone deficiency. J. Clin. Endocrinol. Metab. 82, 213–217. Hardelin, J. P. (2001). Kallmann syndrome: Towards molecular pathogenesis. Mol. Cell Endocrinol. 179, 75–81. Layman, L. C. (1999a). The genetics of human hypogonadotropic hypogonadism. Am. J. Med. Genet. 89, 240–248. Layman, L. C. (1999b). The molecular basis of human hypogonadotropic hypogonadism. Mol. Genet. Metab. 68, 191–199. Layman, L. C., Cohen, D. P., Jin, M., Xie, J., Li, Z., Reindollar, R. H., Bolbolan, S., Bick, D. P., Sherins, R. J., Duck, L. W., Musgrove, L. C., Sellers, J. C., and Neill, J. D. (1998). Mutations in the gonadotropin-releasing hormone receptor gene cause hypogonadotropic hypogonadism. Nat. Genet. 18, 14–15. Layman, L. C., and McDonough, P. G. (2000). Mutations of the follicle stimulating hormone-beta and its receptor in human and mouse: Phenotype/Genotype. Mol. Cell Endocrinol. 161, 9–17. Oliveira, L. M., Seminara, S. B., Beranova, M., Hayes, F. J., Valkenburgh, S. B., Schipani, E., Costa, E. M., Latronico, A. C., Crowley, W. F., Jr., and Vallejo, M. (2001). The importance of autosomal genes in Kallmann syndrome: Genotype–Phenotype correlations and neuroendocrine characteristics. J. Clin. Endocrinol. Metab. 86, 1532–1538. Rugarli, E., and Ballabio, A. (1993). Kallmann syndrome from genetics to neurobiology. JAMA 270, 2713–2716. Zhang, Y-H., Guo, W., Wagner, R. L., Huang, B-L., McCabe, L., Vilain, E., Burris, T. P., Anyane-Yeboa, K., Burghes, A. H. M., Chitayat, D., Chudley, A. E., Genel, M., Gertner, J. M., Klingensmith, G. J., Levine, S. N., Nakamoto, J., New, M. I., Pagon, R. A., Pappas, J. G., Quigley, C. A., Rosenthal, I. M., Baxter, J. D., Fletterick, R. J., and McCabe, E. R. B. (1998). DAX1 mutations provide insight into structure–function relationships in steroidogenic tissue development. Am. J. Hum. Genet. 62, 855–864.

300

p0025

receptors demonstrated a 421-fold preference for GnRH-II versus GnRH-I. Unlike the GnRH-II receptor and most other GPCRs, the GnRH-I receptor contains no large cytoplasmic C-terminal tail. The C termini of both types of receptors are phosphorylated in response to GnRH, leading to receptor desensitization. GnRH activation of its receptor results in stimulation of several signaling pathways. Phospholipase C transmits its signal to diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3). DAG activates the intracellular protein kinase C pathway, and IP3 stimulates release of intracellular calcium. Protein kinase C activation in response to GnRH also leads to increases in the mitogen-activated protein kinase (MAPK) pathway in pituitary cells. These pathways modulate gonadotropin synthesis and release from pituitary cells.

DEVELOPMENTAL REGULATION OF GnRH

p0035

In contrast to most cells of the CNS and all other hypothalamic-releasing factor neurons, GnRH neurons do not initially develop within the brain. GnRH neurons originate in the olfactory placode, migrate across the cribriform plate, and ultimately reside in the diencephalic region of the basal hypothalamus in a dispersed neuronal network. Although born in the olfactory placode, GnRH neurons are not derived from olfactory neurons. In rodents, GnRH neuronal migration begins on embryonic day 10.5 to 11 and final migration occurs by day 15 or 16. During this migratory phase, GnRH neurons express low levels of GnRH mRNA and protein. Identification of the factors that modulate both migration and expression of GnRH in migratory neurons is currently an area of active research. Cell adhesion molecules, including neural cell adhesion molecule (NCAM), are thought to be important. In addition, an adhesion related kinase (Ark) has been shown, in cell culture models, to modulate both neuronal migration and expression of GnRH. Soluble factors such as gamino butyric acid (GABA) and netrins have also been implicated in GnRH neuronal migration. The human hypothalamus only contains a few thousand GnRH neurons (in contrast to approximately 800 neurons in the rodent) that are also arranged in a neuronal network in the basal hypothalamus after migration. Release of GnRH from these neurons occurs in a pulsatile fashion. Immortalized GnRH neuronal cell culture models, as well as isolated

Gonadotropin-Releasing Hormone (GnRH) Actions

GnRH neurons, demonstrate that the GnRH ‘‘pulse generator’’ is intrinsic to the neuronal cells. During human development, the GnRH pulse generator and subsequent gonadotropin release is active in the early neonate but decreases by 6 months of age. Until puberty, the GnRH pulse generator is repressed. The exact mechanisms of repression are not known and may involve central GABAergic neuronal activity. Similarly, the factors that reactivate the GnRH pulse generator at puberty are not completely understood. Neuronal decreases in GABAergic activity and increased glutaminergic, norepinephrine, and neuropeptide Y activities have been implicated in reactivation of the GnRH pulse generator at the time of sexual maturation. Studies have also suggested a role for transforming growth factor-a/erbB-1 and neuregulin/erbB-4 secreted from glial cells, and nitric oxide from central endothelial cells, as factors that may influence the reactivation of GnRH pulsatile release at the time of puberty.

DISEASE STATES ASSOCIATED WITH GnRH Precocious puberty is defined as puberty that occurs prior to 8 years of age in girls and prior to 9 years of age in boys. Central precocious puberty, which is more frequent in girls, is caused by premature reactivation of the GnRH pulse generator and is usually idiopathic. Hypothalamic lesions or tumors, which are more common in boys, have been shown to induce central precocious puberty. The exact mechanism is unknown but may involve premature activation of the glial production of transforming growth factor-a to cross-talk with GnRH neurons. In contrast to precocious puberty, patients with hypogonadotropic hypogonadism (HH) display delayed or absent puberty, hypogonadism, and infertility. HH arises due to deficiency in either GnRH production or GnRH signaling to the pituitary. In Kallmann’s syndrome, GnRH neurons fail to migrate to the hypothalamus from the olfactory placode; therefore, GnRH production is compromised. Anosmia, or lack of smell, is also associated with Kallmann’s syndrome. This finding suggests that a common mechanism between GnRH neuronal migration and olfactory neuronal maturation is defective. A genetic defect in the KAL gene, whose product is anosmin, leads to incomplete migration of GnRH and olfactory neurons in an X-linked inheritance. The molecular bases of other forms of GnRH deficiency are unknown. GnRH deficiency affects 1

p0045

301

Gonadotropin-Releasing Hormone (GnRH) Actions

p0050

in 7500 males and 1 in 70,000 females. Other causes of HH involve defective signaling pathways downstream of GnRH. Mutations in pituitary GnRH receptors affect the ability of GnRH to activate gonadotropin synthesis and secretion. In addition, patients have been reported with mutations of the LHb or FSHb genes resulting in defective sexual maturation. Clinically, GnRH is used in a pump delivery system to result in an episodic delivery to restore defects in patients with HH. Paradoxically, GnRH agonists such as nafarelin, leuprolide, histrelin, and goserelin block GnRH action. The pituitary is sensitive to pulsatile stimulation by GnRH to synthesize and release the gonadotropins. Conti-nuous stimulation of pituitary GnRH receptors by exogenously administered GnRH agonists, rather than by pulsatile stimulation, down-regulates and desensitizes GnRH receptors. The ultimate effect of chronic stimulation of the pituitary GnRH receptors is to decrease LH and FSH production, with subsequent decreases in circulating estrogen or testosterone. Nafarelin, leuprolide, and histrelin are indicated for central precocious puberty. Nafarelin, leuprolide, and goserelin are indicated for treatment of endometriosis. Leuprolide and goserelin are also indicated for palliative treatment of advanced breast and prostate tumors due to the ability of GnRH agonists to decrease sex steroid hormone levels. More recently, GnRH antagonists have been used to directly block GnRH receptors on the pituitary and to inhibit LH and FSH production. They are indicated for the treatment of female infertility as adjunct therapy during ovarian hyperstimulation for in vitro fertilization.

CONCLUSION p0055

GnRH is the hypothalamic hormone that stimulates the pituitary gland to produce FSH and LH. The reactivation of GnRH expression in the hypothalamus is responsible for the initiation of the onset of puberty

and sexual maturation. Aberrant reactivation of GnRH expression is associated with either premature, delayed, or absent puberty. There-fore, the GnRH pathway is a target for several drug therapies that target either inappropriately early or delayed puberty. In addition to their role in normal reproductive development, the sex steroids may have a pathological role in hormone-dependent tumors such as breast and prostate. As the first step in the hypothalamic– pituitary–gonadal axis, the GnRH pathway is also a therapeutic target to inhibit the production of sex steroids from the gonads in the treatment of hormone-dependent malignancies.

See Also the Following Articles Gonadotropin-Induced Ovulation . Gonadotropin-Releasing Hormone Deficiency, Congenital Isolated . Gonadotropin-Releasing Hormone Receptor Gene, Mutation of . Gonadotropins and Testicular Function in Aging . Gonadotropin-Secreting Tumors . Precocious Puberty, Central (Female) . Precocious Puberty, Central (Male)

Further Reading MacColl, G., Quinton, R., and Bouloux, P. M. (2002). GnRH neuronal development: Insights into hypogonadotrophic hypogonadism. Trends Endocrinol. Metab. 13, 112–118. Neill, J. D. (2002). Minireview: GnRH and GnRH receptor genes in the human genome. Endocrinology 143, 737–743. Parhar, I. S. (ed.) (2002). ‘‘Gonadotropin-Releasing Hormone: Molecules and Receptors,’’ Progress in Brain Research, vol. 141. Elsevier, Amsterdam. Prevot, V. (2002). Glial–neuronal–endothelial interactions are involved in the control of GnRH secretion. J. Neuroendocrinol. 14, 247–255. Tobet, S. A., Bless, E. P., and Schwarting, G. A. (2001). Developmental aspect of the gonadotropin-releasing hormone system. Mol. Cell. Endocrinol. 185, 173–184. Wierman, M. E. (1996). Gonadotropin-releasing hormone. In ‘‘Reproductive Endocrinology, Surgery, and Technology’’ (E. Y. Adashi, J. A. Rock, and Z. Rosenwaks, eds.), pp. 666–681. Lippincot–Raven, Philadelphia.

318

Gonadotropin-Releasing Hormone Receptor Gene, Mutation of

detected to date and briefly discusses their impact on the phenotypic expression of HH in individuals bearing such receptor defects.

THE GnRH RECEPTOR p0020

The GnRHR belongs to the superfamily of G proteincoupled receptors (GPCRs), specifically the family related to the rhodopsin/b-adrenergic receptors, which is the best-known member in terms of its structural and functional characteristics. Seven transmembrane hydrophobic domains (TMDs) oriented roughly perpendicular to the plasma membrane plane, with an extracellular NH2 terminal, an intracellular COOH terminal, and three alternating

f0005

intra- and extracellular hydrophilic loops connecting the TMDs, characterize the structure of these receptors (Figs. 1A and 1C). The mammalian GnRHR exhibits more than 85% amino acid identity among the several species that have been cloned. Unlike other members of the rhodopsin/b-adrenergic subfamily of GPCRs, the GnRHR exhibits several unique features, including the reciprocal exchange of the conserved Asp and Asn residues in TMDs 2 and 7, the replacement of Tyr with Ser in the Asp–Arg– Tyr motif located in the junction of the TMD 3 and the intercellular loop (IL) 2, and the lack of the COOH terminal domain. This latter feature, which is not exhibited by GnRH receptors from nonmammalian vertebrate species such as fish and avian GnRH

Figure 1 The human GnRH receptor. (A) Counterclockwise orientation of a prototypical G protein-coupled receptor from TMDs 1 to 7. The closed-loop structure is representative of receptors for peptide ligands such as GnRH. In this arrangement, the core consists mainly of TMDs 2, 3, 5, and 6, whereas domains 1 and 7 are peripherally sequestered. (B) Schematic of the human GnRHR gene. The open reading frame is distributed among three exons (black rectangles), spanning 18.9 Kb, that encode amino acids 1 to 174, 175 to 248, and 249 to 328, respectively. Intron 1 is located between amino acids 174 and 175 in the putative transmembrane domain (TMD) 4, and intron 2 is located between amino acids 248 and 249 in the intracellular loop (IL) 3 (shaded inverse triangles). (C) Sequence of the human GnRHR and location of the inactivating mutations identified to date (black circles).

Gonadotropin-Releasing Hormone Receptor Gene, Mutation of

receptors, is unique among the thousands of members of mammalian GPCRs and is apparently associated with differential physiological regulation (internalization, desensitization, and cell surface expression) of the receptor in mammals and nonmammals. In humans, the GnRHR is located in 4q13.2-3 and consists of three exons and two introns that encode for a 328-amino acid protein (Fig. 1B). The GnRHR is preferentially coupled to the trimeric Gq/11 protein, localized in the cytoplasm, and associated with the intracellular domains of the receptor. Activation of the GnRHR by its ligand is associated with conformational changes in the receptor molecule that lead to activation of the Gaq/11 subunit followed by its dissociation from the Gbg complex. The activation of the GnRHR–Gaq/11 protein complex stimulates the effector enzyme phospholipase– Cb, which in turn induces phosphatidylinositol 4,5biphosphate hydrolysis, leading to formation of the second messengers inositol 1,4,5-triphosphate (IP3) and diacylglycerol. The former diffuses through the cytoplasm, promoting the release of intracellular calcium and the secretion of gonadotropins, whereas the latter activates the enzyme protein kinase C, triggering a cascade of protein–protein phosphorylation and interactions that eventually lead to the expression of biological effects.

MUTATIONS IN THE HUMAN GnRHR GENE

p0035

Structural alterations in key residues of the receptor molecule or the G proteins may potentially lead to altered function of the receptor–G protein system. Mutations in sites involved in ligand binding usually result in altered receptors unable to recognize the ligand and to become activated (loss-of-function mutations), whereas mutations in sites involved in receptor activation or G protein coupling may lead either to loss of function or to constitutive activation (activation in the absence of ligand or gain-of-function mutations) of the receptor molecule. Spontaneous mutations of this latter type have not yet been detected in the GnRHR. Resistance to GnRH by inactivating (loss-of-function) mutations in the human GnRHR gene leads to distinct forms of autosomal-recessive HH (Table I). GnRHR mutations may also occur in individuals with sporadic HH, but with a lower frequency. The detected mutations in the human GnRHR gene are distributed along the entire coding sequence of the

319 receptor, including the NH2 terminus, TMDs 2 to 7, extracellular loops 1 and 2, and the IL3 (Fig. 1C). However, two ‘‘hot spots’’ have been identified: the Gln106Arg and the Arg262Gln mutations. Expression of 15 mutated GnRH receptors in heterologous cell systems has shown that these mutations may influence ligand binding, receptor expression at the cell surface, and/or intracellular signal transduction. Although the expression of a given functional defect is apparently related to the structural modification introduced on specific microdomains involved in particular receptor functions, further studies are still necessary for this viewpoint to be accepted with certainty. In fact, in a number of these mutant receptors the function may be partially or completely rescuable in vitro by genetic or pharmacologic means, which indicates that the structural alteration may additionally lead to altered intracellular trafficking and reduced cell surface expression. Studies employing these GnRHR mutants have found that the Arg262Gln, Thr32Ile, Cys200Tyr, Leu266Arg, and Cys279Tyr mutations predominantly affect IP3 production, whereas in the Gln106Arg, Ala129Asp, Ser168Arg, Asn10Lys, and Arg139His mutations, agonist binding at the cell surface is decreased or severely impaired; all of these mutations apparently do not affect the membrane expression of the altered GnRHR in a significant manner. On the other hand, the Tyr284Cys and Glu90Lys mutations profoundly affect membrane receptor density as a consequence of either disruption of trafficking to the membrane or instability and subsequent degradation of the mutant receptor. Interestingly, deletion of K191 (which results in increased membrane expression and reduced rate of internalization of the wild-type human GnRHR) in the latter GnRHR mutant efficiently restores membrane expression and agonist-induced, receptor-mediated intracellular signaling. In the Ser217Arg mutation as well as in the Leu314X(Stop) mutation, GnRH binding is practically abolished. In this latter mutation, which leads to partial deletion of the TMD 7, the mRNA levels of the receptor were reduced considerably, suggesting that the truncated protein might not be adequately translated and expressed in vivo. A truncated nonfunctional GnRHR has also been reported for the homozygous splice junction mutation (G to A replacement) at the intron 1–exon 2 boundary, which results in a transcript showing splicing of exon 1 to exon 3 (i.e., complete deletion exon 2). Finally, the Ala171Thr mutation in the TMD 4 of the GnRHR presumptively disrupts receptor function by impeding conformational mobility of the TMD 3 and 4,

320

t0005

Gonadotropin-Releasing Hormone Receptor Gene, Mutation of

Table I Mutations in the Human GnRH Receptor Gene, Genotypes, and Impact on Receptor Function Mutation

Genotype

G1n106Arg Arg262G1n

Compound heterozygous

Arg262Gln Tyr284Cys

Compound heterozygous

Ala129Asp Arg262Gln

Compound heterozygous

Gln106Arg a Ser217Arg a Arg262Gln

Compound heterozygous

Ser168Arg

Homozygous

Gln106Arg Leu314X(Stop) Thr32Ile Cys200Tyr Leu266Arg Gln106Arg Cys279Tyr Asn10Lys Gln106Arg Arg139His

Impact on GnRHR function Decreased ligand binding NME? Decreased IP3 production Decreased IP3 production and Decreased membrane expression Undetectable ligand binding Decreased IP3 production Decreased ligand binding Undetectable ligand binding Decreased IP3 production Undetectable ligand binding

NME?

NME

Decreased ligand binding Undetectable ligand binding Absent IP3 production NME Absent IP3 production NME Absent IP3 production NME Decreased ligand binding

Compound heterozygous Compound heterozygous Compound heterozygous Homozygous Homozygous

Absent IP3 production

NME

Homozygous

Decreased ligand binding NME Decreased ligand binding Undetectable ligand binding NME

Glu90Lys

Homozygous

Absent membrane expression

Exon2, Complete deletion Gln106Arg

Homozygous

Unfunctional truncated receptor

Compound heterozygous

Undetectable ligand binding

Ala171Thr

Compound heterozygous

NME

Note. NME, normal membrane expression. aMutations in the same allele.

resulting in stabilization of the receptor in its inactive conformation. Individuals with HH due to GnRHR mutations exhibit a strikingly wide spectrum of clinical and biochemical phenotypes, including variable alterations in pubertal development, plasma gonadotropin and sex steroid levels, response to exogenous GnRH administration, and pulsatile pattern of gonadotropin release. These alterations usually occur in the absence of anatomical or functional abnormalities of the hypothalamic–gonadotrope axis. Thus, the hypogonadism due to GnRHR mutations can be complete or partial. The differences between phenotypes in HH due to inactivating GnRHR mutations could be related to the particular allelic combination of the coexisting mutations, with the functional activity of a given mutant being more or less severely affected than that exhibited by the other. Nevertheless, the fact that different phenotypes may be present within affected kindred bearing the same molecular alteration suggests that other factors or mechanisms may be implicated in the phenotypic expression of the GnRH-resistant HH.

CONCLUSION GnRH resistance due to inactivating mutations in the GnRHR gene leads to impaired synthesis and secretion of the pituitary gonadotropins and to a distinct form of hypogonadotropic hypogonadism. Several naturally occurring mutations in the GnRHR have been described. These mutations may occur along the entire coding sequence of the receptor and usually result in distinct alterations in receptor function. Naturally occurring mutations in the GnRHR represent unique models for the analysis of the structure–activity relationships of this particular receptor.

See Also the Following Articles Gonadotropin-Releasing Hormone (GnRH) Actions . Gonadotropin-Releasing Hormone Deficiency, Congenital Isolated

Further Reading Beranova, M., Oliveira, L. M. B., Be´de´carrats, G. Y., Schipani, E., Vallejo, M., Ammini, A. C., Quintos, J. B., Hall, J. E., Martin, K.

Gonadotropin-Releasing Hormone Receptor Gene, Mutation of

A., Hayes, F. J., Pitteloud, N., Kaiser, U. B., Crowleym, W. F., Jr., and Seminara, S. B. (2001). Prevalence, phenotypic spectrum, and modes of inheritance of gonadotropin-releasing hormone receptor mutations in idiopathic hypogonadotropic hypogonadism. J. Clin. Endocrinol. Metab. 86, 1580–1588. Caron, P., Chauvin, S., Christin-Maitre, S., Bennet, A., Lahlou, N., Counis, R., Bouchard, P., and Kottler, M. L. (1999). Resistance of hypogonadic patients with mutated GnRH receptor genes to pulsatile GnRH administration. J. Clin. Endocrinol. Metab. 84, 990–996. Costa, E. M. F., Bedecarrats, G. Y., Mendonca, B. B., Arnhold, I. J. P., Kaiser, U. B., and Latronico, A. C. (2001). Two novel mutations in the gonadotropin-releasing hormone receptor gene in Brazilian patients with hypogonadotropic hypogonadism and normal olfaction. J. Clin. Endocrinol. Metab. 86, 2680–2686. de Roux, N., Young, J., Brailly-Tabard, S., Misrahi, M., Milgrom, E., and Schaison, G. (1999). The same molecular defects of the gonadotropin-releasing hormone receptor determine a variable degree of hypogonadism in affected kindred. J. Clin. Endocrinol. Metab. 84, 567–572. de Roux, N., Young, J., Misrahi, M., Genet, R., Chanson, P., Schaison, G., and Milgrom, E. (1997). A family with hypogonadotropic hypogonadism and mutations in the gonadotropinreleasing hormone receptor. N. Engl. J. Med. 337, 1597–1602. Karges, B., Karges, W., Mine, M., Ludwig, L., Ku¨hne, R., Milgrom, M., and De Roux, N. (2003). Mutation Ala171Thr stabilizes the gonadotropin-releasing hormone receptor in its inactive conformation, causing familial hypogonadotropic hypogonadism. J. Clin. Endocrinol. Metab. 88, 1873–1879. Kottler, M. L., Chauvin, S., Lahlou, N., Harris, C. E., Johnston, C. J., Lagarde, J. P., Bouchard, P., Farid, N. R., and Counis, R. (2000). A new compound heterozygous mutation of the gonadotropin-releasing hormone receptor (L314X, Q106R) in a woman with complete hypogonadotropic hypogonadism: Chronic estrogen administration amplifies the gonadotropin defect. J. Clin. Endocrinol. Metab. 85, 3002–3008. Layman, L.C., Cohen, D.P., Jin, M., Xie, J., Li, Z., Reindollar, R.H., Bolbolan, S., Bick, D.P., Sherins, R. R., Duck, L. W.,

321 Musgrove, L. C., Sellers, J. C., and Neill, J. D. (1998). Mutations in gonadotropin-releasing hormone receptor gene cause hypogonadotropic hypogonadism. Nat. Genet. 18, 14–15. Lu, Z-L., Saldanha, J. W., and Hulme, E. C. (2002). Seventransmembrane receptors: Crystals clarify. Trends Pharmacol. Sci. 23, 140–146. Maya-Nu´n˜ez, G., Janovick, J., and Conn, P. M. (2000). Combined modification of intracellular and extracellular loci on human gonadotropin-releasing hormone receptor provides a mechanism for enhanced expression. Endocrine 13, 401–409. Maya-Nu´n˜ez, G., Ulloa-Aguirre, A., So¨derlund, D., Conn, P. M., and Me´ndez, J. P. (2000). Molecular basis of hypogonadotropic hypogonadism: Restoration of mutant (E90K) GnRH receptor function by a deletion at a distant site. J. Clin. Endocrinol. Metab. 87, 2144–2149. Pralong, F. P., Go´mez, F., Castillo, E., Cotecchia, S., Abuin, L., Aubert, M. L., Portmann, L., and Gaillard, R. C. (1999). Complete hypogonadotropic hypogonadism associated with a novel inactivating mutation of the gonadotropin-releasing hormone receptor. J. Clin. Endocrinol. Metab. 84, 3811–3816. Silveira, L. F. G., Stewart, P. M., Thomas, M., Clark, D. A., Bouloux, P. M. G., and MacColl, G. S. (2002). Novel homozygous splice acceptor site GnRH receptor (GnRHR) mutation: human GnRHR ‘‘knockout.’’ J. Clin. Endocrinol. Metab. 87, 2973–2977. So¨derlund, D., Canto, P., de la Chesnaye, E., Ulloa-Aguirre, A., and Me´ndez, J. P. (2001). A novel homozygous mutation in the second transmembrane domain of the gonadotropin releasing hormone receptor gene. Clin. Endocrinol. (Oxford) 54, 493–498. Ulloa-Aguirre, A., and Conn, P. M. (1998). G proteincoupled receptors and the G protein family. In ‘‘Handbook of Physiology, Section 7: The Endocrine System, Volume 1: Cellular Endocrinology’’ (P. M. Conn and H. M. Goodman, eds.), pp. 87–124. Oxford University Press, New York. Ulloa-Aguirre, A., Janovick, J. A., Lean˜os-Miranda, A., and Conn, P. M. (2003). Misrouted cell surface receptors as a novel disease aetiology and potential therapeutic target: the case of hypogonadotropic hypogonadism due to gonadotropin-releasing hormone resistance. Expert Opin. Ther. Targets 7, 175–185.

307

Gonadotropin-Releasing Hormone, Family of

f0005

p0030

Figure 1 Hypothalamo–Hypophyseal (pituitary)–Gonadal axis. Through its control of the hypothalamo–pituitary–gonadal axis, GnRH regulates reproductive development.

GnRH-II (chicken GnRH-II) is the most ubiquitous form, present in the brain in all vertebrate classes except Agnathans. The ratfish and dogfish, both fish species of ancient origin in the class Chondrichthyes, possess a GnRH-II- and GnRH-I-type isoform, respectively. The ancestral GnRH molecule likely arose prior to the protochordates. In all GnRH peptides in vertebrates and protochordates, certain regions of the molecule have been highly conserved, including the NH2 terminus (pGlu1) and Ser4, and the COOH terminus. These regions and the length of the molecule have remained unchanged during the 500 million years of evolution of the chordates. The conservation of the NH2 terminus (pGlu1) and Ser4, and the COOH terminus suggests that these regions are significant for active conformation of the peptide, effective receptor binding, resistance to enzymatic

degradation, and receptor-mediated events required for gonadotropin release. This significance has been supported by numerous activity studies of GnRH analogues in mammalian and other vertebrate systems. Further comparative studies on the GnRH family will help to provide clues on the evolution of reproductive mechanisms and insights into our understanding of gene duplication during the early development of vertebrates, structure–activity relations, and the molecular evolution and functional diversity of GnRH.

GnRH PRECURSOR The prepro-GnRH molecule is tripartite, encoding a signal peptide that is followed directly by the GnRH

308

Gonadotropin-Releasing Hormone, Family of

Table I GnRH Family of Peptides GnRH

1

2

3

4

5

6

7

8

9

10

Percentage identity to mGnRH

Mammal

pGlu

His

Trp

Ser

Tyr

Gly

Leu

Arg

Pro

GlyNH2

X

Guinea pig

pGlu

Tyr

Trp

Ser

Tyr

Gly

Val

Arg

Pro

GlyNH2

80

Chicken-I

pGlu

His

Trp

Ser

Tyr

Gly

Leu

Gln

Pro

GlyNH2

90

Rana

pGlu

His

Trp

Ser

Tyr

Gly

Leu

Trp

Pro

GlyNH2

90

Vertebrates

Seabream

pGlu

His

Trp

Ser

Tyr

Gly

Leu

Ser

Pro

GlyNH2

90

Salmon White fish

pGlu pGlu

His His

Trp Trp

Ser Ser

Tyr Tyr

Gly Gly

Trp Met

Leu Asn

Pro Pro

GlyNH2 GlyNH3

80 80

Medaka

pGlu

His

Trp

Ser

Phe

Gly

Leu

Ser

Pro

GlyNH2

80

Catfish

pGlu

His

Trp

Ser

His

Gly

Leu

Asn

Pro

GlyNH2

80

Herring

pGlu

His

Trp

Ser

His

Gly

Leu

Ser

Pro

GlyNH2

80

Chicken-II

pGlu

His

Trp

Ser

His

Gly

Trp

Tyr

Pro

GlyNH2

70

Dogfish

pGlu

His

Trp

Ser

His

Gly

Trp

Leu

Pro

GlyNH2

70

Lamprey-III

pGlu

His

Trp

Ser

His

Asp

Trp

Lys

Pro

GlyNH2

60

pGlu

His

Tyr

Ser

Leu

Glu

Trp

Lys

Pro

GlyNH2

50

Lamprey-I Invertebrates Tunicate-I

pGlu

His

Trp

Ser

Asp

Tyr

Phe

Lys

Pro

GlyNH2

60

Tunicate-II

pGlu

His

Trp

Ser

Leu

Cys

His

Ala

Pro

GlyNH2

60

Note. There are currently 16 known forms of GnRH: 14 in vertebrates and 2 in invertebrates.

decapeptide, a dibasic cleavage site, and the GnRHassociated peptide (GAP), as diagrammed in Fig. 2. Similar to other neuropeptides, GnRH is first synthesized as a larger precursor protein, called prepro GnRH, which is then processed to its final decapeptide form. Neuropeptide precursors consist of at least 60 amino acids and include a signal sequence at the amino terminus and a spacer region at the carboxy terminus. The spacer region is separated from the neuropeptide by a dibasic amino acid cleavage site and may or may not have a biological function. The amino acid sequence of prepro GnRH has been indirectly identified by the isolation of its cDNA sequence.

Table III shows a complete list of the GnRH cDNAs/ genes identified to date. Of the 16 forms of GnRH identified, the cDNA or gene sequences of 11 forms have been determined. Due to the low identity of the prepro-GnRH cDNA sequence among forms, investigators are restricted to the use of degenerate oligonucleotide primers or probes based on the amino acid sequence of the specific GnRH for identification of GnRH. Classical molecular biological techniques were used for isolating the first GnRH cDNA. This entailed screening a genomic library with degenerate oligonucleotide probes. The full cDNA sequence was then cloned

Table II The Lineages of GnRH Type GnRH

Brain distribution/origin

Primary GnRH structures identified in vertebrates

GnRH-I

Hypothalamus, diencephalon/ Olfactory origin

Mammal GnRH in mouse, primate, human, sheep, pig, eel, newt, frog; chicken GnRH-I in chicken, lizard; salmon GnRH in goldfish, salmon; catfish GnRH in catfish; dogfish GnRH in dogfish

GnRH-II

Midbrain/Ventricular ependyma origin

Chicken GnRH-II in mouse, primate, human, chicken, lizard, frog, newt, eel, goldfish, catfish, salmon, medaka, red seabream, tilapia, ratfish

GnRH-III

Telencephalon/Olfactory origin

Salmon GnRH in medaka, red seabream, tilapia

GnRH-IV

Hypothalamus, diencephalon/ ventricular origin

Lamprey GnRH-I and lamprey GnRH-III in lamprey

Note. GnRH is divided into four types based on a combination of function, location of expression, and molecular phylogenetic analysis.

309

Gonadotropin-Releasing Hormone, Family of

Gene structure Exon 1

Intron 1

Exon 2

Intron 2

Exon 3

Intron 3

Exon 4

cDNA structure

Signal peptide GnRH/Dibasic processing site GnRH Associated peptide Untranslated region

Figure 2 Prepro–GnRH tripartite structure. The GnRH precursor is made up of three regions encoding the signal peptide, the GnRH decapeptide, and the GnRH-associated peptide.

from a brain cDNA library. Salmon GnRH (type GnRH-III) was also isolated by screening a brain cDNA library with degenerate oligonucleotides. Later in the 1990s, the polymerase chain reaction (PCR) was used to amplify specific gene sequences from a messenger ribonucleic acid (mRNA) population or from cDNA or genomic libraries. This method has now been used to isolate the nine additional GnRH cDNA sequences. The first cDNA sequences for chicken GnRH-I, chicken GnRH-II, catfish GnRH, and seabream GnRH all were determined using reverse transcription–PCR (RT-PCR) and/or rapid amplification of cDNA ends (RACE). Both of these methods are based on amplification via PCR. In the African cichlid, the salmon GnRH gene was first isolated in 1991 by Bond and colleagues, and later the prepro-chicken GnRH-II gene was isolated from the same species. The presence of two cDNA sequences confirmed that multiple GnRH forms present in an organism were not produced from differential processing of one GnRH gene but rather were, in fact, transcribed from separate genes. Since that time, multiple GnRH genes have been isolated from many species. The signal peptide contains a core of hydrophobic amino acids that is followed by alternating polar and nonpolar amino acids, except in the case of seabream GnRH, which contains only polar residues at the carboxy terminus of the signal peptide. The signal peptide is considered to aid in the prohormone’s transport across the endoplasmic reticulum membrane for posttranslational processing. The characteristic hydrophobic signal peptide targets the prohormone to the proper cellular organelle. The mechanism for this is currently being elucidated. It is thought that the hydrophobic residues may be interacting with the signal recognition particle

(SRP), which targets proteins to the rough endoplasmic reticulum (RER). The SRP is a ribonucleoprotein complex containing a 7S RNA molecule and six polypeptides. One of the associated polypeptides, SRP54, is a 54-kDa protein that binds to the signal portion of a prepro peptide as it is translated at the ribosome. SRP54 contains a putative guanosine 50 -triphosphate (GTP)-binding domain as well as a methionine-rich region. GTP is required for releasing SRP from the signal peptide and ribosome. The region that is rich in methionine residues is proposed to be the region that binds to the signal peptide. The methionine-rich region is predicted to form four amphipathic helices that position the hydrophobic residues entirely on one face, with polar residues exclusively on the opposite face. This secondary structure would suggest that the hydrophobic signal peptide could bind in the hydrophobic pocket lined by amphipathic helices. The methionine residues of SRP54 are proposed to form methionine bristles to accommodate a wide variety of hydrophobic signal sequences, and this could explain the lack of apparent consensus sequence of signal peptides apart from their generally high hydrophobicity. After the signal peptide is cleaved from the prohormone, the glutamine residue at GnRH position 1 is cyclized to pyro-glutamate. Although the mechanism of this cyclization is unknown, the physiological spontaneous rate of formation occurs very slowly, and enzymatic activity catalyzing this formation has been described in brain, pituitary, and lymphocytes. Following the GnRH decapeptide in the precursor hormone structure is a glycine residue that donates its amide group to the terminal glycine residue of GnRH. This amidation is catalyzed by peptidyl glycine a-amidating mono-oxygenase only after GAP is removed. This enzyme requires that the amide donor

310

t0015

Gonadotropin-Releasing Hormone, Family of

Table III GnRH cDNAs and Genes Identified to Date Isoform Mammalian

Organism

Year

Reference

Human

1984

Norway rat

1986

Seeburg et al., Nature Adelman et al., Proc. Natl. Acad. Sci. USA

Human

1986

Adelman et al., Proc. Natl. Acad. Sci. USA

Mouse

1986

Mason et al., Science

Norway rat

1989

Bond et al., Mol. Endocrinol.

African clawed frog

1994

Hayes et al., Endocrinology

Rhesus monkey Tree shrew

1996 1996

Dong et al., Mol. Cell Endocrinol. Kasten et al., Gen. Comp. Endocrinol.

Japanese eel

1999

Okubo et al., Zool. Sci.

Bullfrog

2001

Wang et al., J. Exp. Zool.

Guinea pig

Guinea pig

1997

Jimenez-Linan et al., Endocrinology

Chicken-I

Chicken

1993

Dunn et al., J. Mol. Endocrinol.

Rana

Frog

2000

Yoo et al., Mol. Cell Endocrinol.

Medaka

Medaka

2000

Okubu et al., Biochem. Biophys. Res. Commun.

Catfish Seabream

African catfish Gilthead seabream

1994 1995

Bogerd et al., Eur. J. Biochem. Gothilf et al., Mol. Mar. Biol. Biotechnol.

Chicken-II

Salmon

Haplochromis burtoni

1995

White et al., Proc. Natl. Acad. Sci. USA

Red sea bream

1996

Okuzawa, unpublished

Striped sea bass

1998

Chow et al., J. Mol. Endocrinol.

Haplochromis burtoni

1998

White et al., Gen. Comp. Endocrinol.

European sea bass

2002

Zmora et al., J. Endocrinol

Barfin flounder

2002

Amano et al., Gen. Comp. Endocrinol.

Nile tilapia Haplochromis burtoni

2002 1994

Farahmand et al., unpublished White et al., Proc. Natl. Acad. Sci. USA

African catfish

1994

Bogerd et al., Eur. J. Biochem.

Tree shrew

1996

Kasten et al., Gen. Comp. Endocrinol.

Goldfish

1996

Lin et al., Gen. Comp. Endocrinol.

Human

1998

White et al., Proc. Natl. Acad. Sci. USA

Striped sea bass Haplochromis burtoni Human

1998 1998 1998

Chow et al., J. Mol. Endocrinol. White et al., Gen. Comp. Endocrinol. White et al., Proc. Natl. Acad. Sci. USA.

House shrew

1998

White et al., unpublished

Silver–Gray brushtail possum

1999

Lawrence et al., unpublished

Rio cauca caecilian

1999

Ebersole et al., unpublished

Rhesus monkey

1999

Urbanski et al., Endocrinology

Japanese eel

1999

Okubo et al., Zool. Sci.

Medaka

2000

Okubo et al., Biochem. Biophys. Res. Commun.

Australian bonytongue Bullfrog

2001 2001

Okubo and Aida, Gen. Comp. Endocrinol. Wang et al., J. Exp. Zool.

European sea bass

2002

Zmora et al., J. Endocrinol.

Zebra fish

2002

Whitlock et al., unpublished

Zebra fish

2002

Stevens et al., unpublished

Barfin flounder

2002

Amano et al., Gen. Comp. Endocrinol.

Common carp

2002

Li et al., unpublished

Haplochromis burtoni

1991

Bond et al., Mol. Endocrinol.

Atlantic salmon Atlantic salmon

1992 1992

Klungland et al., Mol. Cell Encrdocrinol. Klungland et al., Mol. Cell. Encrdocrinol.

Rainbow trout

1992

Alestrom et al., Mol. Mari. Biol. Biotechnol.

Cherry salmon

1992

Suzuki et al., J. Mol. Endocrinol. (continues)

311

Gonadotropin-Releasing Hormone, Family of

t0015

Table III (continued) Isoform

Organism Brook trout

Lamprey-III

Lamprey-I

Year 1992

Reference Klungland et al., Mol. Cell Endocrinol.

Chinook salmon

1992

Klungland et al., Mol. Cell Endocrinol.

Rainbow trout Brown trout

1992 1992

Klungland et al., Mol. Cell Endcrinolol. Klungland et al., Mol. Cell Endocrinol.

Plainfin midshipman

1995

Grober et al., Gen. Comp. Endocrinol.

Sockeye salmon

1995

Ashihara et al., J. Mol. Endocrinol.

Sockeye salmon

1995

Coe et al., Mol. Cell Endocrinol.

Goldfish

1996

Lin et al., Gen. Comp. Endocrinol.

Haplochromis burtoni

1998

White et al., Gen. Comp. Endocrinol.

Medaka

2000

Okubo et al., Biochem. Biophys. Res. Commun.

Zebrafish Australian bonytongue

2000 2001

Torgersen et al., unpublished Okubo and Aida, Gen. Comp. Endocrinol.

European sea bass

2002

Zmora et al., J. Endocrinol.

Zebrafish

2002

Torgersen et al., unpublished

Barfin flounder

2002

Amano et al., Gen. Comp. Endocrinol.

Mozambique tilapia

2002

Molina et al., unpublished

Sea lamprey

2002

Silver et al., Am. Zool.

Pacific sea lamprey

2002

Silver et al., Am. Zool.

Australian lamprey Pouched lamprey

2002 2002

Silver et al., Am. Zool. Silver et al., Am. Zool.

American brook lamprey

2003

Silver et al., Am. Zool.

Northern brook lamprey

2003

Silver et al., Am. Zool. Silver et al., Am. Zool.

Western brook lamprey

2003

Silver lamprey

2003

Silver et al., Am. Zool.

Sea lamprey

2001

Suzuki et al., J. Mol. Endocrinol.

Note. There are currently 57 GnRH precursor cDNAs and 17 genes (in boldface) identified and sequenced.

be at the C terminus of the peptide. It has been proposed that GAP-releasing enzyme initiates this posttranslational modification by an endoproteolytic event. The enzyme is a serine protease that recognizes the lysine12–arginine13 dibasic cleavage site sequence that typically acts in brain prohormones for proteolytic cleavage of the active peptide from the precursor molecule and cleaves the amide bond directly following the residues. The next enzyme proposed to be involved is hypothalamic carboxypeptidase E, which sequentially removes the arginine13 and the lysine12 residues. This leaves the glycine11 at the C terminus of the peptide as a substrate for peptidyl glycine a-amidating mono-oxygenase. The initial cleaving event between residues 13 and 14 is thought to occur during vesicle formation at the Golgi apparatus, whereas further processing occurs within the secretory granules where GAP and bioactive GnRH are stored until needed for release. The function of GAP remains undetermined. GAP, in conjunction with the signal peptide, most likely functions to ensure that the hormone is long enough for

insertion into the endoplasmic reticulum for proper processing to occur. Researchers have questioned a possible biological role for GAP because there is in vitro evidence that human GAP is responsible for both inhibiting prolactin release and stimulating gonadotropin hormone (GTH) release in rat pituitary as well as inhibiting prolactin release in the teleost fish tilapia. A bioactive role for GAP is also supported by rat in vivo studies, in which the first 13 residues of the GAP molecule were shown to stimulate the release of GTH. Although the molecular organization of GnRH precursors appears to be similar throughout various species, the amino acid composition is highly divergent. The sequence encoding the GnRH peptide and the following cleavage sites are highly conserved among all of the vertebrates, whereas the signal peptide and GAP remain quite divergent (Fig. 3). In contrast to other known vertebrate GnRH precursors, which typically have one transcript (and in two cases have two transcripts), three distinct transcripts have been isolated and sequenced in lampreys.

312

Gonadotropin-Releasing Hormone, Family of

Mammalian GnRH precursor identities Signal peptide

GnRH

GAP

Homo sapien mGnRH 67%

100%

70%

61%

100%

70%

61%

100%

84%

22%

100%

34%

22%

100%

26%

23%

100%

34%

Mus musculus mGnRH

Rattus norvegicus mGnRH

Tupaia glis mGnRH

Xenopus laevis mGnRH

Rana catesbeiana mGnRH

Anguilla japonica mGnRH

Figure 3 Prepro–GnRH sequence comparison. The GnRH precursor is highly conserved in structure and in the GnRH decapeptide sequence. The lamprey GnRH-I (type GnRH-IV) transcripts, termed GAP49, GAP50, and GAP58, differed in the length of the GAP coding sequence and were demonstrated to be the products of a single gene. Analysis of the lamprey GnRH-I gene intron-2 splice junction demonstrated that alternate splicing produces the different lamprey GnRH-I transcripts. Lamprey GnRHI is the first GnRH gene demonstrated to use splice sequence variants to produce multiple transcripts that may reflect an ancestral gene regulatory mechanism. The GnRH genes that have been identified have maintained a highly conserved pattern of introns and exons (Fig. 2). The coding region is dispersed over four exons, where the first exon is untranslated; the second encodes the signal peptide, the GnRH decapeptide, the cleavage site, and the N-terminal portion of GAP; the third consists entirely of the middle portion of the GAP molecule; and the fourth contains the C-terminal of GAP as well as the 30 untranslated region. The only noticeable difference between the salmon and mammalian GnRH gene structures is their relative intron sizes. Because of the similar architecture, the emergence of multiple forms of GnRH has been attributed to nucleotide base changes and not to differential splicing of message or variable processing of the precursor protein.

GnRH RECEPTOR In light of the crucial role GnRH plays in human physiology and disease, its receptor has been a subject of intense research for many years. Numerous studies

on the binding characteristics of the GnRH receptor were performed throughout the 1970sand 1980s. In 1992, the first successful cloning of a GnRH receptor from the mouse using a homology-based PCR amplification scheme was reported. Later that year, the human GnRH receptor cDNA was reported as well. Since these landmark studies, the GnRH receptor primary structure has been reported in several other organisms (Table IV). Multiple receptor forms have been reported in brains and pituitaries of primates, teleost fish (e.g., goldfish, medaka), and amphibians (e.g., bullfrog, Korean frog [Rana dybowskii]). The GnRH receptors have been classified into two major groupings, type I and type II, based primarily on sequence identity and major structural characteristics. Further subdivisions of these two groups have been proposed, but more studies on expression, activity, and pharmacological characteristics are required to establish a solid subdivision of these two major groups. Analysis of the sequences of the first identified GnRH receptors defined them as belonging to the G protein-coupled receptor (GPCR) superfamily of receptors. The members of this superfamily share a common general structure composed of seven hydrophobic a-helical transmembrane domains connected by hydrophilic intracellular and extracellular loops, an extracellular N-terminal tail, and an intracellular Cterminal tail. The GnRH receptor belongs to the class A rhodopsin-like GPCR subfamily. This subfamily is characterized by conserved amino acids, or motifs, in certain positions of the receptor. The GnRH receptors exhibit an interesting pattern of evolutionary change from the otherwise highly conserved motifs of the class A GPCRs (Fig. 4). Of the structural variations seen in the GnRH receptors, most notable is the absence of any C-terminal tail in the type I mammalian GnRH receptors. These receptors are the only GPCRs without a C-terminal intracellular tail. The implications of this drastic modification of structure are still not well understood. The identification of the GnRH receptors in the goldfish and the bullfrog marked the first time that multiple receptors had been cloned from one species. Recently, several studies have been published reporting the identification of multiple GnRH receptors in brains and pituitaries of other organisms. Two distinct receptors have been cloned in another teleost fish, the medaka, and the three receptor types identified in the bullfrog have been supported by isolation of three very similar receptors in another frog, the Korean frog R. dybowskii. A second GnRH receptor has been identified in the marmoset, and similar novel

313

Gonadotropin-Releasing Hormone, Family of

t0020

Table IV GnRH Receptors Identified to Date Organism (receptor number)

Year

Reference

Human

1992

Kakar et al., Biochem. Biophys. Res. Commun.

1993

Chi et al., Mol. Cell Endocrinol.

1992

Tsutsumi et al., Mol. Endocrinol.

Mouse

Rat

1992

Perrin et al., Biochem. Biophys. Res. Commun.

1993

Reinhart et al., J. Biol. Chem.

1992

Eidne et al., Mol. Cell Endocrinol.

1992 1993

Kaiser et al., Biochem. Biophys. Res. Commun. Kudo et al., Zool. Sci.

Cow

1993

Kakar et al., Domest. Anim. Endocrinol.

Pig

1994

Weesner and Matteri, J. Anim. Sci.

Sheep

1996

Campion et al., Gene

Marmoset

1999

Byrne et al., J. Endocrinol.

Australian brushtail possum

2000

King et al., Gen. Comp. Endocrinol.

Bonnet monkey

2000

Santra et al., Mol. Hum. Reprod.

Dog Tamar wallaby

2000 2002

Cui et al., Mol. Endocrinol. Cheung et al., Reprod. Fertil. Dev.

African catfish

1997

Tensen et al., Eur. J. Biochem.

Goldfish A (1)

1999

Illing et al., Proc. Natl. Acad. Sci. USA

Goldfish B (2)

1999

Illing et al., Proc. Natl. Acad. Sci. USA

African clawed frog

2000

Troskie et al., Endocrinology

Japanese eel

2000

Okubo et al., Gen. Comp. Endocrinol.

Rainbow trout

2000

Madigou et al., Biol. Reprod.

Striped sea bass African green monkey

2000 2001

Alok et al., Mol. Cell Endocrinol. Neill et al., Biochem. Biophys. Res. Commun.

Bullfrog (1)

2001

Wang et al., Proc. Natl. Acad. Sci. USA

Bullfrog (2)

2001

Wang et al., Proc. Natl. Acad. Sci. USA

Bullfrog (3)

2001

Wang et al., Proc. Natl. Acad. Sci. USA

Chicken

2001

Sun et al., J. Biol. Chem.

Marmoset (2)

2001

Millar et al., Proc. Natl. Acad. Sci. USA

Medaka (1)

2001

Okubo et al., Endocrinology

Medaka (2) Korean frog (1)

2001 2003

Okubo et al., Endocrinology Seong et al., Endocrinology

Korean frog (2)

2003

Seong et al., Endocrinology

Korean frog (3)

2003

Seong et al., Endocrinology

Lamprey

2001

Nucci et al., Amer. Zool.

Note. The amino acid sequences of 33 GnRH receptors have been deduced. Receptors that lack the evolutionarily conserved C-terminal tail are in boldface, whereas receptors that retain this domain are in regular type.

receptors have been described from the African green monkey and the rhesus monkey Macaca mulatta. These receptors have been shown to contain intracellular C-terminal tails, unlike all previously identified mammalian GnRH receptors. The discovery and description of these receptors opens up exciting new possibilities in the study of the importance of the Cterminal tail in receptor signaling, cycling, expression, and desensitization. The gene structure of the GnRH receptor has also been described in multiple species: human, sheep,

mouse, rat, dog, Japanese eel, African clawed frog, and medaka. With the exception of the medaka receptor 2 and the clawed frog receptor, all of these genes have shown a conserved pattern of introns and exons. This pattern is composed of three exons and two introns. Exon 1 extends from the 50 end of the transcript through the middle of the region encoding transmembrane helix 4, exon 2 encodes from the middle of transmembrane helix 4 to the middle of intracellular loop 3, and exon 3 includes the rest of the coding sequence and the 30 untranslated region.

314

Gonadotropin-Releasing Hormone, Family of

Figure 4 Class A GPCR and GnRH-R motifs. The GnRH receptors exhibit an intriguing pattern of motif change over the course of vertebrate evolution. Most notably, mammalian type I GnRH receptors are the only GPCRs that lack an intracellular tail.

Investigations into the expression level and localization of the GnRH receptor have also indicated the presence of great complexity in the function and regulation of GnRH at the level of the receptor. These studies have shown wide distribution of the GnRH receptor in the brain, particularly in the hypothalamus and midbrain, and also in the pituitary (as expected). In addition, GnRH receptors have been found in detectable levels in other organs. In humans, GnRH receptor mRNA was shown to be present in the pituitary, ovary, testis, breast, and prostate. In fish and amphibians, GnRH receptor mRNAs have been found in the brain, pituitary, ovary, testis, and retina. Interestingly, in the studies that have characterized multiple GnRH receptors in the same species, differential tissue distribution of receptor isoforms has been demonstrated. In the goldfish, differential tissuespecific expression was reported, whereas in the bullfrog, differential expression was also shown but only

within regions of the brain and pituitary rather than among many tissues. The variability of expression among the three bullfrog GnRH receptors was also shown to correspond with stages of the reproductive cycle. This is the first solid evidence of regulation of reproduction by expression variance of multiple GnRH receptors. The extensive cloning of GnRH receptors from various species and the characterization and conservation of gene structure have enabled researchers to explore the structure–function aspects of the receptors and how their differences and similarities give insight into the complex integration and function of the GnRH system. Many studies have investigated the importance of the C-terminal tail in receptor signaling, internalization, desensitization, and expression. The unusual GnRH receptor variations in the conserved motifs typically found in the class A GPCRs have been investigated as well, as have specific

315

Gonadotropin-Releasing Hormone, Family of

structural components of the GnRH receptor. This body of data, combined with modeling studies of the GnRH peptide and the hundreds of activity studies on GnRH analogues, has produced a putative picture of how GnRH binds to its receptor. In the absence of a crystallographic structure of the GnRH-bound receptor, these studies provide the best information on the mechanics of GnRH binding.

GnRH IN MAMMALS GnRH was first isolated from porcine and ovine hypothalamic extracts, giving rise to the popularly held view that a single form of GnRH (mammalian GnRH) regulates the hypothalamo–pituitary–gonadal axis in all mammals. The question that has arisen over the years is how one GnRH differentially regulates the release of two gonadotropins: LH and FSH. Although there have been a number of different models proposed, questions remain and have fueled the quest for a separate FSH-releasing factor or novel GnRH isoform. Recent studies have demonstrated that at least two different GnRH isoforms are expressed within the brain of a single vertebrate species. Current thought suggests that one GnRH (type GnRH-I) functions as a neurohormone regulating the pituitary in mediating the release of gonadotropin. The other form (type GnRH-II) may have a neurotransmitter or neuromodulatory function and is generally localized in areas outside the hypothalamus, particularly in midbrain regions. In mammals, the second GnRH isoform that has been identified to date is chicken GnRH-II. The only direct evidence of the existence of GnRH forms other than mammalian GnRH in mammals has come from the identification of GnRH complementary DNAs in three species: the tree shrew, the guinea pig, and the human. Two prepro-GnRH mRNAs identified in the tree shrew encoded mammalian and chicken GnRH-II. In the guinea pig, a prepro GnRH encodes for a unique form of GnRH, guinea pig GnRH. Most recently in the human, two genes for mammal GnRH and chicken GnRH-II were demonstrated. Using indirect methods in mammals, a limited number of species, monotremes, marsupials, rats, and primitive eutherians have been examined for variant forms of GnRH through immunocytochemistry (ICC), high-performance liquid chromatography (HPLC), and radioimmunoassay (RIA) using specific antibodies to various GnRH forms. In most of the species examined, immunoreactive chicken GnRH-II

was shown to be present generally in the midbrain. In recent studies in rats using a double-labeled immunocytochemical technique, lamprey GnRH-III neurons not only were observed in regions that control FSH release but also were colocalized with mammalian GnRH neurons in areas primarily controlling LH release. These studies used indirect techniques with various heterologous antibodies. Immunoreactive (ir) GnRH will need confirmation by identification of these ir-variant GnRH forms through determination of the primary structure by protein purification or molecular cloning followed by extensive biological activity studies. The first study to demonstrate that a lamprey-like GnRH was present in the human hypothalamus and median eminence used a combination of ICC, HPLC, and RIA. In these same studies, the hypothalamic distribution of immunopositive lamprey-like GnRH neurons was similar to that observed for those containing the mammalian GnRH. More recently, also using similar indirect methods, it was demonstrated that a chicken GnRH-II-like form was found in stumptail and rhesus monkeys but that only a few of the ir-chicken GnRH-II cells were in the posterior basal hypothalamus; most of the immunopositive neurons to chicken GnRH-II antiserum were shown to be in the midbrain. These studies did not screen for other forms of GnRH such as the lamprey GnRH forms, nor did their data suggest that chicken GnRH-II is a neurohormone. Thus, there is incomplete and contrasting data on the nature of GnRHs in primates. Confirmation of the exact nature of a second or possibly third form of GnRH in primates and other mammals will require isolation, sequence analysis, determination of localization, and biological function studies.

STATUS OF AN FSH-RELEASING FACTOR Although research has shown convincingly that GnRH regulates LH release, there are not any definitive data supporting GnRH control of FSH release. However, it has been suggested that regulation of FSH secretion involves a complex balance among stimulation by GnRH from the hypothalamus, inhibitory feedback by sex steroids (testosterone and estradiol) and inhibin B from the gonads, and autocrine/ paracrine modulation by activin and follistatin within the pituitary. As a result, for many years there has been a quest for the identification of a separate FSH-releasing factor in mammals. It has been shown

316 that lamprey GnRH-III (type GnRH-IV) is highly selective in stimulating FSH release in rats and cows. Combined with the earlier immunocytochemical data of a lamprey GnRH-like form in various species of mammals, this provides convincing evidence for a lamprey GnRH-like molecule in the mammalian hypothalamus. In light of this more recent evidence that there might be a novel GnRH in mammals and other vertebrates, the possibility exists that a lampreylike peptide might be an FSH-releasing factor. Thus, the possibility exists that there is a second GnRH-like molecule that is also a hypothalamic factor.

Gonadotropin-Releasing Hormone, Family of

of GnRH and its receptor. Our understanding of GnRH, as the central figure in the control of reproduction, is critical at the molecular, biochemical, and physiological levels. We have come to realize that the GnRH system has proven to be extremely complex and that there is still much to be learned. Importantly, GnRH has become an important subject of not only physiological research but also medical treatments.

See Also the Following Articles Corticotropin-Releasing Hormone, Family of

GnRH ANALOGUES Since 1971, when the primary structure of mammalian GnRH was determined, more than 10,000 analogues have been made to GnRH and been tested in hundreds of studies in mammals. As a result of these studies, several mammalian GnRH analogues have been shown to be highly successful and are currently being used for sterilization, conception, and other therapeutic and clinical applications. For example, Lupron Depot, a GnRH analogue, is now one of the leading chemical treatments for advanced prostate cancer and endometriosis in humans. Continuous treatment with Lupron Depot results in decreased levels of LH and FSH. In males, testosterone is reduced to castrate levels; in premenopausal females, estrogens are reduced to postmenopausal levels. GnRH antagonists are also used clinically, and the most effective GnRH antagonists have substitutions of one amino acid in position 6 as well as substitutions in positions 1, 2, and 3. However, because of the lack of systematic examination in the characterization of these various analogues, there is still critical information that is needed in our understanding of their biological activity. Information on the comparative activities of the GnRH family in vertebrates may provide important hypotheses about the role of specific amino acids in GnRH binding and receptor activation and may lead to the development of more potent GnRH analogues.

CONCLUSION Since the discovery of GnRH in 1971, research has exploded relating to the structure and function

Further Reading Dubois, E. A., Zandbergen, M. A., Peute, J., and Goos, H. J. (2002). Evolutionary development of three gonadotropin-releasing hormone (GnRH) systems in vertebrates. Brain Res. Bull. 57, 413–418. Fernald, R. D., and White, R. B. (1999). Gonadotropin-releasing hormone genes: Phylogeny, structure, and functions. Front Neuroendocrinol. 20, 224–240. Jennes, L., Eyigor, O., Janovick, J. A., and Conn, P. M. (1997). Brain gonadotropin releasing hormone receptors: Localization and regulation. Recent Prog. Horm. Res. 52, 475–490. McCann, S. M., Karanth, S., Mastronardi, C. A., et al. (2001). Control of gonadotropin secretion by follicle-stimulating hormone-releasing factor, luteinizing hormone-releasing hormone, and leptin. Arch. Med. Res. 32, 476–485. Norwitz, E. R., Jeong, K. H., and Chin, W. W. (1999). Molecular mechanisms of gonadotropin-releasing hormone receptor gene regulation. J. Soc. Gynecol. Invest. 6, 169–178. Parhar, I. S., and Sakuma, Y. (eds.) (1997). ‘‘GnRH Neurons: Gene to Behavior.’’ Brain Shuppan, Tokyo. Sealfon, S. C., Weinstein, H., and Millar, R. P. (1997). Molecular mechanisms of ligand interaction with the gonadotropin-releasing hormone receptor. Endocrinol. Rev. 18, 180–205. Shacham, S., Harris, D., Ben-Shlomo, H., et al. (2001). Mechanism of GnRH receptor signaling on gonadotropin release and gene expression in pituitary gonadotrophs. Vitamin Horm. 63, 63–90. Silver, M. R., Lee, K. J., Takahashi, A., Kawauchi, H., Joss, J., Nozaki, M., and Sower, S. A. (2001). Molecular phylogenetic analysis within the petromyzontiforme lineage using the cDNA for lamprey gonadotropin releasing hormone-III. Am. Zool. 41, 1587. Sower, S. A., and Kawauchi, H. (2001). Update: Brain and pituitary hormones of lampreys. Comp. Biochem. Physiol. B Biochem. Mol. Biol. 129, 291–302. Suzuki, K., Gamble, R. L., and Sower, S. A. (2000). Multiple transcripts encoding lamprey gonadotropin-releasing hormone-I precursors. J. Mol. Endocrinol. 24, 365–376.

327

Gonadotropin-Secreting Tumors

subgroup of clinically nonfunctioning pituitary adenomas (NFPA) in vitro, suggesting an antiproliferative effect of this locally secreted peptide. This antiproliferative effect involves an up-regulation of p21, a cyclin-dependent kinase inhibitor. It is concluded that defects in activin-mediated signal transduction pathways and dysregulation of modulating peptides may be involved in abnormal activin-mediated growth control in pituitary tumors. It is likely that an alteration in these control mechanisms may play a role in the tumorigenesis of gonadotropinomas but further studies are required.

PATHOLOGY

Figure 1 Gonadotropinoma ‘‘iceberg.’’ (1) Gonadotropinomas in which there is hormonal secretion with clinical effects; (2) gonadotropinomas in which hormonal secretion occurs without clinical effects; (3) gonadotropinomas in which there is no hormonal secretion (silent adenoma) but there is positive immunohistochemistry; and (4) silent adenomas with gonadotropin gene expression on Northern blot analysis. Reprinted from Beckers (2003), with permission.

are possible (inhibin AabA and inhibin BabB). Activin, a member of transforming growth factor-b growth hormone superfamily, is a homodimer of two disulfide-linked b-subunits of inhibin. Thus, three possible forms of activin exist: AbAbA, BbBbB, and ABbAbB. Activin is produced in many tissues, including the gonads and pituitary, where it has autocrine and paracrine effects. It stimulates FSH secretion from normal rat gonadotroph cells and human gonadotropinoma cells. Follistatin is produced in the pituitary by folliculostellate and gonadotroph cells and is a negative modulator of the effects of activin. It is believed that a decrease in follistatin could be responsible for enhanced activin activity, which could account for the increased cell division and FSH secretion characteristic of gonadotroph adenomas. Other studies found a positive correlation between the concentrations of activin, follistatin, and FSH secreted by gonadotroph adenomas in vitro and concluded that the production of activin A might explain the relatively high levels of FSH and follistatin, but the amount of follistatin was apparently insufficient to antagonize the stimulatory effects of activin A. Finally, an inhibitory effect of activin on DNA synthesis has been reported in a

As is the case with other pituitary adenomas, gonadotroph adenoma cells are usually arranged in cords and may vary in size from one adenoma to another. Before the advent of immunohistochemical techniques, NFPA were recognized by their chromophobe, agranulocytic appearance after Herlant tetrachomic staining. Electronic microscopy first revealed that these supposedly nonsecretory tumors contained cytoplasmic secretory granules. Later, immunohistochemical studies using monoclonal antibodies directed against the a-subunit or against the intact LH, FSH, hCG, and their respective b-subunits clearly identified the secretory nature of gonadotropinomas (Fig. 2). Prognostic proliferation markers, such as proliferating cell nuclear antigen (PCNA), p53, and Ki67, have been investigated in gonadotroph adenomas but with inconclusive results. The expression of the polysialylated neural cell adhesion

Figure 2 Gonadotropinoma immunohistochemistry. Staining was performed using anti-FSH antibodies. Positivity is observed at the periphery of the cell (vascular pole). Reprinted from Beckers (2003), with permission; also reprinted with kind permission from Dr. J. Trouillas, Lyon, France.

p0035

328 molecule, however, seems to be strongly associated with tumor invasion and an aggressive profile.

CLINICAL ASPECTS The clinical recognition of gonadotroph adenoma is relatively difficult, as these tumors do not secrete efficiently and their secretory products, intact gonadotropins and/or their subunits, generally do not cause a recognizable clinical syndrome. Consequently, gonadotropinomas can grow undetected until they become sufficiently large to cause symptoms due to tumor expansion, such as visual impairment, headaches, pituitary insufficiency, and, more rarely, seizures, meningitis, and cerebrospinal fluid rinorrhea. Oligomenorrhea or amenorrhea can present in premenopausal female patients and is correlated with increased secretion of intact FSH, hyperprolactinemia caused by pituitary stalk compression, and/or secondary hypogonadism attributable to the destruction of normal residual pituitary gland. However, FSH secretion can also cause in premenopausal women a sustained ovarian stimulation, inducing multiple large ovarian cysts, elevated serum estradiol, and infertility. There have been few reported cases of males with gonadotroph adenomas associated with a characteristic clinical syndrome due to hypersecretion of LH causing testis hyperstimulation (testicular enlargement, elevated serum testosterone). Most gonadotropinomas in men, however, cause secondary hypogonadism with decreased libido, normal to low serum testosterone levels, and normal testicular volume.

Gonadotropin-Secreting Tumors

Basal concentrations of gonadotropins are not helpful in postmenopausal women because it is not possible to differentiate a sellar mass from normal postmenopausal gonadotroph cells as the source of gonadotropins. Only a marked discrepancy among FSH, LH, and/or their subunits would make such measurements useful in the diagnosis of a gonadotropinoma in a postmenopausal female patient. Administration of TRH typically produces an increase in the secretion of FSH, LH, and especially the LH b-subunit and this finding could be of assistance in making a diagnosis in cases where gonadotroph adenomas are associated with basal levels of gonadotropins and/or their subunits in the normal range. However, other studies have suggested that the TRH test is unable to detect silent gonadotropinomas, but is useful in postmenopausal women and also in the confirmation of a diagnosis in patients with secreting gonadotroph adenomas. Usually, gonadotropinomas also remain sensitive to GnRH stimulation. It is difficult to interpret the response to the GnRH in NFPA, as the hormonal response may originate from the gonadotropinoma or from the remaining normal gonadotroph cells of the pituitary. Pituitary magnetic resonance imaging is by far the best radiological tool to assess the dimensions and the invasiveness of gonadotropinomas (Fig. 3). Almost all reported gonadotropinomas are macroadenomas and nearly one-third of them are considered to be invasive. The incidence of microadenoma is largely unknown but should be suspected when a pituitary microincidentaloma is found.

TREATMENT DIAGNOSIS The overwhelming majority of NFPA are gonadotropinomas but, as noted above, they seldom cause a clinical hormonal hypersecretory syndrome. Gonadotroph adenomas can be recognized by a supranormal basal serum concentration of intact gonadotropins, often FSH but less commonly intact LH and/or their subunits (a-and/or b-subunits). In relatively few cases, high levels of hCG have been found. When intact LH is elevated, male patients may also have supranormal serum testosterone concentrations. When a supranormal serum a-subunit level exists as the sole abnormality in the presence of a pituitary mass, the diagnosis may be either a thyroid-stimulating hormone-secreting adenoma or a gonadotropinoma. In these cases, thyrotropin-releasing hormone (TRH)-induced stimulation of intact FSH and/or LH or their subunits would confirm a gonadotroph origin.

The initial treatment of gonadotroph adenoma is by transsphenoidal surgery, especially if visual disturbances or hemorrhage is present. However, tumor excision is usually incomplete in large macroadenomas and recurrence rates average approximately 30% in NFPA after a mean follow-up of 6 years. Incomplete tumor removal determines a low cure rate and additional treatment may be required. Some centers advocate routine radiotherapy after surgery to reduce the risk of tumor regrowth. When conventional irradiation is administered following surgery, it is usually effective in preventing regrowth of the adenoma. The efficacy of stereotactic radiation remains to be determined. When the surgery is not curative, several drugs have been utilized in the treatment of gonadotroph adenoma, but thus far none have been found to reduce tumor size consistently and substantially.

329

Gonadotropin-Secreting Tumors

Figure 3 Gonadotroph macroadenoma extending far above the sella, elevating the optic chiasm, and invading the cavernous sinus as observed with magnetic resonance imaging in coronal (A) and sagittal (B) views. Reprinted from Beckers (2003), with permission.

p0100

p0110

Dopamine receptors have been identified on gonadotroph adenomas and dopamine agonists decrease the levels of the gonadotropins and the a-subunit in vitro and in vivo. However, the use of dopamine agonists (e.g., bromocriptine, cabergoline, and quinagolide) has not been found to be useful in the majority of gonadotroph adenomas, although 20% of treated patients did experience a response. In any case, a significant positive correlation has been reported between the uptake score measured using pituitary scintigraphy with 123I-labeled methoxybenzamide and the percentage inhibition of a-subunit levels and tumor shrinkage during dopamine agonist treatment. Somatostatin analogues have been used to treat gonadotroph adenomas, as somatostatin receptors have been identified on gonadotropinomas and somatostatin can decrease the secretion of gonadotropins and/or their subunits by gonadotropinoma cells in vitro. Several clinical trials have studied the use of somatostatin analogues in patients with gonadotroph adenomas. Although dramatic decreases in the size of gonadotroph adenomas and some associated improvements in visual field defects have been occasionally reported following somatostatin analogue treatment, the majority of patients experience little if any improvement in adenoma size or visual fields. GnRH agonists and antagonists have been considered as potential therapeutic agents for gonadotroph adenomas. However, their administration generally does not produce an effect on adenoma size. Troglitazone, a PPAR (peroxisome proliferator activated receptor)-g ligand, has been studied in mice as a novel medical therapy for NFPA with encouraging results.

CONCLUSION In conclusion, the approach to gonadotroph adenomas is guided by the tumor size, the presence of effects of the tumor mass, and hypopituitarism. Surgery remains the first curative choice for gonadotroph macroadenoma. The gonadotroph microadenoma may be observed and treated only if significant growth is documented over the course of a year or more. Radiotherapy may be useful in patients who are poor surgical candidates, in patients whose tumors are surgically inaccessible, or after surgery to reduce the risk of tumor regrowth. Several pharmacologic treatments have been tried but only a minority of patients experience significant benefits. Further studies are necessary and novel drugs that can reduce the size of the tumor and/or prevent its regrowth are greatly needed.

See Also the Following Articles Adrenal Tumors, Molecular Pathogenesis . GonadotropinReleasing Hormone (GnRH) Actions . Pituitary Adenomas, TSH-Secreting . Pituitary Region, Non-Functioning Tumors of . Pituitary Tumors, ACTH-Secreting . Pituitary Tumors, Clonality . Pituitary Tumors, Molecular Pathogenesis . Pituitary Tumors, Surgery . Testicular Tumors

Further Reading Asa, S. L., Bamberger, A. M., Cao, B., Wong, M., Parker, K. L., and Ezzat, S. (1996). The transcription activator steroidogenic factor-1 is preferentially expressed in the human pituitary gonadotroph. J. Clin. Endocrinol. Metab. 81, 2165–2170.

330 Beckers, A. (2003). ‘‘Pituitary Adenomas,’’ CD-ROM, 3rd ed. Graphmed Ltd., Lie`ge, Belgium. Beckers, A., Stevenaert, A., Mashiter, K., and Hennen, G. (1985). Follicle-stimulating hormone-secreting pituitary adenomas. J. Clin. Endocrinol. Metab. 61, 525–528. Boggild, M. D., Jenkinson, S., Pistorello, M., Boscaro, M., Scanarini, M., McTernan, P., Perrett, C. W., Thakker, R. V., and Clayton, R. N. (1994). Molecular genetic studies of sporadic pituitary tumors. J. Clin. Endocrinol. Metab. 78, 387–392. Chanson, P., De Roux, N., Young, J., Bidart, J. M., Jacquet, P., Misrahi, M., Milgrom, E., and Schaison, G. (1998). Absence of activating mutations in the GnRH receptor gene in human pituitary gonadotroph adenomas. Eur. J. Endocrinol. 139, 157–160. Colao, A., Ferone, D., Lastoria, S., Cerbone, G., Di Sarno, A., Di Somma, C., Lucci, R., and Lombardi, G. (2000). Hormone levels and tumour size response to quinagolide and cabergoline in patients with prolactin-secreting and clinically non-functioning pituitary adenomas: Predictive value of pituitary scintigraphy with 123I-methoxybenzamide. Clin. Endocrinol. 52, 437–445. Danila, D. C., Inder, W. J., Zhang, X., Alexander, J. M., Swearingen, B., Hedley-Whyte, E. T., and Klibanski, A. (2000). Activin effects on neoplastic proliferation of human pituitary tumors. J. Clin. Endocrinol. Metab. 85, 1009–1015. Heaney, A. P., Fernando, M., and Melmed, S. (2003). PPAR-g receptor ligands: Novel therapy for pituitary adenomas. J. Clin. Invest. 111, 1381–1388. Penabad, J. L., Bashey, H. M., Asa, S. L., Haddad, G., Davis, K. D., Herbst, A. B., Gennarelli, T. A., Kaiser, U. B., Chin, W. W., and Snyder, P. J. (1996). Decreased follistatin gene expression in gonadotroph adenomas. J. Clin. Endocrinol. Metab. 81, 3397–3403. Pichard, C., Gerber, S., Laloi, M., Kujas, M., Clemenceau, S., Ponvert, D., Bruckert, E., and Turpin, G. (2002). Pituitary carcinoma: Report of an exceptional case and review of the literature. J. Endocrinol. Invest. 25, 65–72. Poncin, J., Stevenaert, A., and Beckers, A. (1999). Somatic MEN 1 gene mutation does not contribute significantly to sporadic pituitary tumorigenesis. Eur. J. Endocrinol. 140, 573–576.

Gonadotropin-Secreting Tumors

Risbridger, G. P., Schmitt, J. F., and Robertson, D. M. (2001). Activins and inhibins in endocrine and other tumors. Endocr. Rev. 22, 836–858. Roncaroli, F., Nose´, V., Scheithauer, B. W., Kovacs, K., Horvath, E., Young, W. F., Lloyd, L. V., Bishop, M. C., Hsi, B., and Fletcher, J. A. (2003). Gonadotropic pituitari carcinoma: HER/neu expression and gene amplification. J. Neurosurg. 99, 402–408. Shimon, I., Rubinek, T., Bar-Hava, I., Nass, D., Hadani, M., Amsterdam, A., and Harel, G. (2001). Ovarian hyperstimulation without elevated serum estradiol associated with pure folliclestimulating hormone-secreting pituitary adenoma. J. Clin. Endocrinol. Metab. 86, 3635–3640. Snyder, P. J. (2002). Gonadotroph adenomas. In ‘‘The Pituitary’’ (S. Melmed, ed.), 2nd ed., pp. 575–591. Blackwell Science, Cambridge, UK. Trouillas, J., Daniel, L., Guigard, M. P., Tong, S., Gouvernet, J., Jouanneau, E., Jan, M., Perrin, G., Fischer, G., Tabarin, A., Rougon, G., and Figarella-Branger, D. (2003). Polysialylated neural cell adhesion molecules expressed in human pituitary tumors and related to extrasellar invasion. J. Neurosurg. 98, 1084–1093. Valdes-Socin, H., Jaffrain-Rea, M. L., Tamburrano, G., Cavagnini, F., Ciccarelli, E., Colao, A., Delemer, B., Brue, T., Rohmer, V., Wemeau, J. L., Levasseur, S., Teh, B. T., Stevenaert, A., and Beckers, A. (2002). Familial isolated pituitary tumors: Clinical and molecular studies in 80 patients. In ‘‘The Endocrine Society’s 84th Annual Meeting, San Francisco, California, 19–22 June 2002.’’ Abstract 647. Verges, B., Boureille, F., Goudet, P., Murat, A., Beckers, A., Sassolas, G., Cougard, P., Chambe, B., Montvernay, C., and Calender, A. (2002). Pituitary disease in MEN type 1 (MEN1): Data from the France–Belgium MEN1 multicenter study. J. Clin. Endocrinol. Metab. 87, 457–465. Wessels, H. T., Hofland, L. J., van der Wal, R., van Gastel, L., van Koetsveld, P. M., de Herder, W. W., and de Jong, F. H. (2001). In vitro secretion of FSH by cultured clinically nonfunctioning and gonadotroph pituitary adenomas is directly correlated with locally produced levels of activin A. Clin. Endocrinol. 54, 485–492.

323

Gonadotropins and Testicular Function in Aging

Figure 1 Serum LH (A) and testosterone (B) concentration profiles in three young and three older men sampled at 2.5-min intervals during sleeping hours. The continuous lines are predicted by deconvolution analysis. Deconvolution revealed loss of high-amplitude LH and testosterone secretory pulses in older individuals with preserved basal secretion rates. Entropy analyses disclosed more disorderly (less regular) release of LH and testosterone in older men. Reproduced with permission from the American Society of Andrology. Veldhuis, J. D. (1999). Recent insights into neuroendocrine mechanisms of aging of the human male hypothalamic–pituitary–gonadal axis. J. Androl. 20, 1–17. frequency with loss of high-amplitude LH secretory pulses, even though basal secretions are preserved. The dysfunctional GnRH–LH pulsing mechanism results in decreased LH mass per pulse values (Fig. 1). The release of LH showed more disorderliness in older men than in younger men. Moreover, sophisticated mathematical modeling techniques show that there is more disruption of synchrony between LH and FSH, as well as between LH and testosterone secretion, in older men than in their younger counterparts. This multifold synchrony disruption of neuroendocrine control of LH and subsequent testosterone secretion results in uncoupling of testosterone-related nocturnal penile tumescence activity that may be of physiological importance. The abnormal neuroendocrine regulation of LH pulse secretion is also evident in the aging rat model, where decreased LH pulse amplitude is observed. Hypothalamic GnRH pulse generator dysfunction is suggested by the decreased GnRH response to excitatory amino acid stimulation. There is evidence to show that inducible nitric oxide synthase (NOS) activity is increased in the aging GnRH secretory neurons in the hypothalamus.

The increase in inducible NOS activity might be a result of increased release of reactive oxygen species, cytokines, or other substances associated with aging. Increased NOS activity results in nitrosylation of cells and increased apoptosis of GnRH-secreting cells in the hypothalamus of aging rats. From population-based studies, cross-sectional analyses fail to show significant trends of serum prolactin with aging, but the longitudinal data show a sharp increase in morning serum prolactin levels at 5.3% per year. In detailed studies of nocturnal prolactin secretory patterns in elderly men, serum basal concentration and secretory mass per burst of prolactin are decreased at night and serum prolactin is correlated with serum testosterone concentrations. The significance of the changes in prolactin with aging is not known but has been postulated to be reflective of neuroendocrine effects of aging.

TESTICULAR FUNCTION IN AGING There is now ample evidence not only from cross-sectional studies but also from longitudinal

324

Figure 2 Longitudinal effects of aging on date-adjusted testosterone and free testosterone index. Linear segment plots for total testosterone and free testosterone index versus age are shown for men with testosterone and sex hormone-binding globulin values on at least two visits. Each linear segment has a slope equal to the mean of the individual longitudinal slopes during each decade and is centered on the median age for each cohort of men from the second to the ninth decades. Numbers in parentheses represent the numbers of men in the various cohorts. With the exception of free testosterone index during the ninth decade, segments show significant downward progression at every age, with no significant change in slopes for testosterone or free testosterone index over the entire age range. Reproduced with permission from The Endocrine Society. Harman, S. M., Melter, E. J., Tobin, J. D., Pearson, J., and Blackman, M. R. (2001). J. Clin. Endocrinol. Metab. 86, 724–731.

epidemiological studies to show that aging in men is associated with a slow but progressive decline in serum total testosterone and bioavailable or free testosterone levels (Fig. 2). Serum total testosterone concentrations decline by 1.6% and bioavailable testosterone declines by 2.3% per year, and these decreases are accompanied by significant increases (1.3% per year) in serum sex hormone-binding globulin concentrations. There is also a concomitant decline in serum estrogens and adrenal androgens, androstenedione and dehydroepiandrosterone (DHEA) and DHEA sulfate. In contrast, serum 5a dihydrotestosterone (DHT) concentrations tend to increase in older men (3.4% per year). Such increases in serum DHT are associated with stable intraprostatic DHT levels. The rise in DHT in the presence of declining testosterone may be due to increased 5a reductase activity in the liver, skin, or prostate. The decline in serum testosterone is

Gonadotropins and Testicular Function in Aging

associated with ill-defined symptoms of andropause, including asthenia, loss of motivation, failure of concentration, decreased libido and sexual activity, loss of muscle and bone mass, loss of strength, increased body fat, depressed mood, and decreased quality-of-life measures. The decrease in serum testosterone concentration is the result of defective Leydig cell steroidogenic capacity coupled with decreased responsiveness of the Leydig cells to endogenous LH. Studies in rodent models of the aging confirmed the decreased steroidogenic capacity of Leydig cells both in vitro and in vivo.Our laboratory has shown that although Leydig cell numbers are not reduced in aged animals, the Leydig cell volume is decreased. There is also evidence showing that some Leydig cells undergo apoptosis without evidence of renewal. Despite the decline in steroidogenic capacity, the ability of the seminiferous tubules to produce normal spermatozoa appears to persist in elderly healthy men. There is some evidence that older men might have slightly lower sperm concentration and motility, with little impact on their fertility. Studies in aging men showed that circulating inhibin B (a marker of Sertoli cell function) is maintained. Serum inhibin B is related to serum FSH but not to age or serum testosterone levels in older men. The maintenance of normal serum inhibin B concentrations in older men indicates normal Sertoli cell function and spermatogenic activity. Our laboratory and others have shown in the brown Norway rat model that aging is associated with decreased testis volume due to decreased seminiferous tubule volume and sperm concentration. The decrease in germ cells is nonuniform and affects some but not all tubules. We have also showed that this decline in germ cells is related to accelerated apoptosis affecting germ cells. Because the loss of germ cells appears to be patchy, it is likely that, in addition to decreased testosterone secretory capacity of Leydig cells, other factors such as increased cytokines, reactive oxygen species, and other paracrine factors induced by decreased blood flow or neuronal input may play important roles.

IMPLICATIONS IN MANAGEMENT OF AGING MEN There is general agreement that aging in men is associated with declining serum testosterone and bioavailable/free testosterone concentrations. Although serum testosterone may be decreased, seminiferous tubule

325

Gonadotropins and Testicular Function in Aging

function appears to remain relatively normal, with normal sperm output and reproductive capability even to the seventh decade. Low serum testosterone levels are implicated as a possible cause of the clinical features associated with andropause. Lack of energy, frailty, decreased general well-being, sexual dysfunction, loss of muscle, and loss of bone mass may be ameliorated by testosterone supplementation. Because of the probable abnormal functioning GnRH pulse generator in aging men, stimulation of the axis by clomiphene citrate might not result in favorable increases in LH pulse amplitude and testosterone secretion. Furthermore, because Leydig cell dysfunction appears to be the predominant cause of androgen deficiency in aging males, treatment with recombinant LH or human chorionic gonadotropin might not elicit a satisfactory response, although gonadotropin administration has not been tested in a large number of individuals. Based on our understanding of the gonadotropins and testicular dysfunction in aging men, replacement with androgens at the current time and replacement with selective androgen receptor modulators in the future appear to be the best choices for treatment of symptoms or signs of andropause. Studies are ongoing to determine the benefit versus risk ratios of androgen replacement therapy in aging men.

See Also the Following Articles Androgen Biosynthesis and Gene Defects . GonadotropinReleasing Hormone (GnRH) Actions . Hypergonadotropic Hypogonadism . Impotence and Aging . Spermatogenesis, Endocrine Control of

Further Reading Feldman, H. A., Longcope, C., Derby, C., Johannes, C. B., Araujo, A. B., Coviello, A., Brekmner, W. J., and McKinley, J. B. (2002). Age trends in the level of serum testosterone and other hormones in middle-aged men: Longitudinal results from the Massachusetts Male Aging Study. J. Clin. Endocrinol. Metab. 87, 589–598. Harman, M., Metter, E. J., Tobin, J. D., Pearson, J., and Blackman, M. R. (2001). Longitudinal effects of aging on serum total and free testosterone levels in healthy men. J. Clin. Endocrinol. Metab. 86, 724–731. Veldhuis, J. D. (1999). Recent insights into neuroendocrine mechanisms of aging of the human male hypothalamic–pituitary– gonadal axis. J. Androl. 20, 1–17. Vermuelen, A. (2001). Androgen replacement therapy in the aging male: A critical evaluation. J. Clin. Endocrinol. Metab. 80, 2380–2389. Wang, C., Sinha Hikim, A., Ferrini, M., Bonavera, J. J., Vernet, V., Leung, A., Lue, Y. H., Gonzalez-Cadavid, N., and Swerdloff, R. S. (2002). Male reproductive aging using the brown Norway rat as a model for men. In ‘‘Endocrine Facets of Aging’’ (Novartis Foundation Symposium 242, D. J. Chadwick and J. A. Goode, eds.), pp. 82–97. Wiley, Chichester, UK.

350 deposited in the dermal tissues. These deposits lead to disruption of the collagen bundles and marked edema. Lymphocytic infiltration is not always apparent.

ETIOLOGY The reason why localized myxedema occurs in Graves’ disease is unknown. The strong association with Graves’ hyperthyroidism, and especially ophthalmopathy, suggests that it should have an autoimmune pathogenesis. Although nearly all patients afflicted with myxedema have high titers of thyroid-stimulating hormone (TSH) receptor autoantibodies, there is no clear relation between these titers and the severity of the skin lesions. Nevertheless, these antibodies may very well play a role given that skin fibroblasts express the TSH receptor. However, studies in which serum immunoglobulins (IgGs) from patients with dermopathy were added to cultured skin fibroblasts did not show any effect of the immunoglobulins on the production of glycosaminoglycans by the fibroblasts. Other studies showed that serum from patients with Graves’ disease contains a factor—other than IgG—that is able to stimulate cultured skin fibroblasts to produce glycosaminoglycans. This factor may be a cytokine produced by lymphocytes. The preponderance for the lower legs in Graves’ dermopathy is unclear. One reason may be the influence of gravity and stasis, but there is also evidence that fibroblasts from different sites have a different regulation of glycosaminoglycan synthesis.

TREATMENT Treatment is often not indicated because the lesions are usually asymptomatic. Only when there is local discomfort or the lesions become unsightly should

Graves’ Dermopathy

treatment be considered. There are no controlled studies on any treatment for Graves’ dermopathy, but uncontrolled studies suggest that topical application of corticosteroids is beneficial. The corticosteroids containing cream should be applied directly to the lesions and then covered with a plastic occlusive dressing. After 3 to 10 weeks, the dose and frequency of the application may be gradually tapered. Compressive stockings during the daytime will diminish the fluid accumulation.

See Also the Following Articles Graves’ Disease . Graves’ Disease, Hyperthyroidism in . Graves’ Ophthalmopathy . Thyroid Disease, Epidemiology of

Further Reading Daumerie, C., Ludgate, M., Costagliola, S., and Many, M. C. (2002). Evidence for thyrotropin receptor immunoreactivity in pretibial connective tissue from patients with thyroid-associated dermopathy. Eur. J. Endocrinol. 146, 35–38. Fatourechi, V. (2000). Localized myxedema and thyroid acropachy. In ‘‘Werner and Ingbar’s the Thyroid: A Fundamental and Clinical Text’’ (L. A. Braverman and R. D. Utiger, eds.), 8th ed., pp. 548–555. Lippincott, Williams & Wilkins, Philadelphia. Peacey, S. R. Flemming, L., Messenger, A., and Weetman, A. P. (1996). Is Graves’ dermopathy a generalized disorder? Thyroid 6, 41–45. Schwartz, K. M., Fatourechi, V., Ahmed, D. D. F., and Pond, G. (2002). Dermopathy of Graves’ disease (pretibial myxedema): Long-term outcome. J. Clin. Endocrinol. Metab. 87, 438–446 Smith, T. J., Bahn, R. S., and Gorman, C. A. (1989). Connective tissue, glycosaminoglycans, and diseases of the thyroid. Endocrinol. Rev. 1, 366–391.

352 trigger in developing Graves’ thyroid disease. Various infectious diseases have been considered risk factors. Infections with Yersinia enterocolitica have been implicated because this pathogen has thyroid-stimulating hormone (TSH) binding sites that presumably are homologous to the TSH receptor, the prime autoantigen in Graves’ hyperthyroidism. The association between the occurrence of Graves’ disease and the therapeutic use of interferon-a (IFN-a) in patients infected with the hepatitis C virus suggests that other conditions resulting in high Interferon alpha levels, such as certain viral infections, may also increase the risk of Graves’ disease. Finally, the clear female preponderance suggests that estrogen levels may play an important role. Graves’ disease often occurs during the postpartum period.

PATHOGENESIS Although the cause of Graves’ disease is unknown, autoimmunity directed against the TSH receptor is the hallmark of Graves’ thyroid disease. Autoantibodies against this receptor mimic the action of its natural ligand, TSH, inducing hyperthyroidism and goiter. Whether these antibodies also are responsible for the other, extra-thyroidal manifestations of Graves’ disease is uncertain, although the TSH receptor has been found in fibroblasts residing in the retrobulbar tissues and in the pretibial skin. The pathogenetic importance of TSH receptor autoantibodies as the cause of Graves’ hyperthyroidism is, however, without doubt. Transfer of human immunoglobulin (IgG) samples containing these antibodies to rodents causes a prolonged release of previously radiolabeled thyroid hormone from mouse or guinea pig thyroids. This bioassay was used to make the diagnosis of Graves’ disease in the past, and the transferred antibodies were named ‘‘long-acting thyroid stimulator’’ (LATS). Nowadays, different tests are applied. Patient IgGs can be incubated with cells transfected with the human TSH receptor, and the response of these cells in terms of the release of cyclic AMP (cAMP) can be measured in a TSH receptor-stimulating immunoglobulin (TSI) bioassay. Another widely used and less cumbersome assay measures the inhibition of binding of 125I-labeled TSH to TSH receptor containing preparations by patient IgGs in the so-called TSH-binding inhibitory immunoglobulins (TBII) assay. In the majority of patients, other antithyroidal autoantibodies are also present. In approximately 70% of patients, antibodies against thyroid peroxidase (TPO) can be detected. TPO antibodies are the

Graves’ Disease

hallmark of Hashimoto’s hypothyroidism, and their titers are related to the degree of lymfocytic infiltration of the thyroid gland. The importance of these antibodies in Graves’ hyperthyroidism is unclear, but they might be related to the fact that some patients with Graves’ hyperthyroidism become hypothyroid in the long run.

VARIOUS MANIFESTATIONS OF GRAVES’ DISEASE There are three manifestations that may occur in patients with Graves’ disease: hyperthyroidism, ophthalmopathy, and pretibial myxedema. Hyperthyroidism is the most common manifestation, and some ophthalmopathy occurs in 25 to 50% of these patients, although severe eye disease is rather rare and develops in only approximately 5% of patients. Pretibial myxedema is even less common and is seen in fewer than 1% of patients. The reason why such diverse organs—thyroid, orbit, skin—can become affected in one patient is unclear, and it has even been reasoned that these three manifestations are, in fact, the expression of distinct disorders frequently occurring more or less simultaneously. This assumption cannot be disproved entirely, but it seems unlikely for a number of reasons. First, there is a close temporal relationship between the onset of Graves’ thyroid disease and the onset of Graves’ ophthalmopathy; 80% of patients develop one manifestation within 18 months of the onset of the other. Second, when sensitive methods are used to assess eye involvement in patients with Graves’ hyperthyroidism without clinically apparent ophthalmopathy, evidence for orbital tissue swelling is found in virtually all of them. Third, it has been observed that some patients with multinodular goiter develop Graves’ hyperthyroidism and ophthalmopathy several months after treatment with radioactive iodine. Multinodular goiter is not an autoimmune thyroid disease and, as such, is not associated with ophthalmopathy. Radioactive iodine leads to a slow destruction of the thyroid, and thyroidal proteins (including the TSH receptor) will leak into the circulation and be presented to the immune system. A minority of patients subsequently mount an autoimmune response to the TSH receptor, leading to hyperthyroidism, and a small number also develop ophthalmopathy. This chain of events suggests that the two disorders are linked to each other. In fact, it supports another theory explaining the multiple-organ character of Graves’ disease: autoantibodies against a thyroidal

p0040

353

Graves’ Disease

antigen cross-react with a similar antigen in the skin and the retrobulbar tissues. The nature of this antigen is unknown, but the TSH receptor is a good candidate because TSH receptor autoantibodies are the cause of the hyperthyroidism and the TSH receptor is expressed on retrobulbar fibroblasts and possibly also on fibroblasts derived from skin affected by pretibial myxedema.

See Also the Following Articles Goitrogens, Environmental . Graves’ Dermopathy . Graves’ Disease, Hyperthyroidism in . Graves’ Ophthalmopathy .

Hashimoto’s Disease . Iodine . Smoking and the Thyroid . Thyroid Autoimmunity . Thyroid Gland, Anatomy and Physiology . TSH Function and Secretion

Further Reading Davies, T. F. (2000). Autoimmune thyroid disease. Endocrinol. Metab. Clin. North America 29, 239–442. McKenna, T. J. (2001). Graves’ disease. Lancet 357, 1793–1796. Prummel, M. F., and Laurberg, P. (2003). Interferon-a and autoimmune thyroid disease. Thyroid 13, 547–551. Weetman, A. P., and McGregor, A. M. (1994). Autoimmune thyroid disease: Further developments in our understanding. Endocrinol. Rev. 15, 788–830.

355

Graves’ Disease, Hyperthyroidism in

Table I Common Clinical Manifestations of Thyrotoxicosis Symptoms

Nervousness, hyperactivity Fatigue, weakness Increased perspiration, heat intolerance Palpitations Tremor Weight loss despite increased apetite Menstrual disturbances

Signs

Hyperactivity Tachycardia, atrial fibrillation Systolic hypertension Warm moist skin Tremor Hyperreflexia Stare and eyelid retraction

Only in Graves’ disease

Diffusely enlarged thyroid (goiter) Ophthalmopathy Dermopathy

concentrations remain unaffected. Hypercalcemia and hypercalciuria are common but usually only moderate, serum magnesium levels tend to be decreased, and phosphate levels tend to be higher. There is no clear evidence for direct toxic effects of elevated thyroid hormones on the liver, although mild elevations of alkaline phosphatase and bilirubin are sometimes seen. An increased frequency of bowel movements is not uncommon and is associated with an increased motility of the intestines. Peristalsis of the esophagus is also increased, and patients occasionally may complain of swallowing difficulties. The red blood cells may be microcytic, but anemia is usually not seen. Mild thrombocytopenia due to shortened platelet survival may occur. Some muscular weakness and wasting occurs in nearly all patients, and this may lead to fatigue and weakness of the larger proximal limb muscles. Nervousness, anxiety, irritability, attention deficits, and some memory loss are common manifestations, but true psychiatric syndromes occur in only a minority of patients. In women, oligomenorrhea is common, but fertility is not severely impaired. In men, there is no evidence for an effect on spermatogenesis, but the biological activity of estrogens is increased and mild gynecomastia can often be found. Hyperthyroidism is a known risk factor for osteoporosis, and the bone mass is frequently decreased in hyperthyroidism. Most of these systemic effects of hyperthyroidism are restored to normal on successful reversal to the euthyroid state.

DIAGNOSIS The biochemical diagnosis of thyrotoxicosis is usually straightforward. Laboratory tests will show elevated levels of T4 and T3 in the presence of low or undetectable TSH values. In Graves’ disease, T3 levels are relatively higher than T4 levels, and in some patients only T3 levels may be increased (T3 toxicosis). To discriminate Graves’ disease from other forms of hyperthyroidism (e.g., toxic multinodular goiter, toxic adenoma) or thyrotoxicosis (e.g., silent or subacute thyroiditis, exogenous thyroid hormone use), additional diagnostic procedures may be useful. A thyroid scintigram will show homogenous and often increased uptake. Determination of TSH receptor autoantibodies may sometimes be useful. An assay measuring the capability of patient immunoglobulins (IgGs) to displace binding of labeled TSH to TSH receptors (TSH binding inhibitory immunoglobulin assay, TBII) has a sensitivity of 96% for the diagnosis of Graves’ hyperthyroidism. TBIIs are sometimes present in patients with multinodular goiter and destructive thyroiditis; hence, the specificity is only 84%.

MANAGEMENT Initial treatment with a beta blocking agent will alleviate most of the signs and symptoms of hyperthyroidism because many are mediated through the sympathetic nerve system. This symptomatic treatment is usually followed by a therapy aimed at diminishing the synthesis and secretion of thyroid hormones. This can be achieved effectively by antithyroid drugs such as methimazole, carbimazole, and propylthiouracil (PTU); by ablating large parts of the thyroid gland with radioactive iodine; or by partly removing the gland surgically. Each of these options has its advantages and disadvantages. Therapy with 30 mg of methimazole taken once daily, or with 300 to 400 mg of propylthiouracil in three or four divided doses, will render most patients euthyroid in 4 to 6 weeks. When euthyroidism has been achieved, one can choose between diminishing the dose (‘‘titration method’’) or maintaining the same dose and adding L-thyroxine (‘‘block and replacement’’). This antithyroid treatment is then continued for 12 to 24 months, after which time approximately 50% of patients will enter into a complete remission. In general, chances to enter into a remission increase if there is a decrease in goiter size, TSH levels become normal during treatment, TSH receptor antibodies disappear, and the patient is a nonsmoker.

p0035

356 Disadvantages of antithyroid drugs include allergic reactions consisting of an often transient pruritic rash, and agranulocytosis, a very rare but lifethreatening event. Prompt discontinuation of the antithyroid drugs will restore the leukocyte count to normal within a week. Other rare side effects include arthralgias, hepatitis, and drug fever. The advantage of subtotal thyroidectomy is the rapid and instantaneous reversal of the thyrotoxic state. It is preferably performed in patients who have been rendered euthyroid by antithyroid drugs, but in emergency situations this can also be achieved by the administration of large doses of iodine (five drops of Lugol’s solution daily for a maximum of 10 days) or by treatment with lithium carbonate or amiodarone. Although mortality rates are negligible, a subtotal thyroidectomy carries a small but unavoidable risk for damage to the recurrent laryngeal nerve leading to vocal cord paralysis and for hypoparathyroidism. Postoperative hypothyroidism occurs in a substantial minority of patients. Radioactive iodine is a simple and economical way in which to control hyperthyroidism. It can be given to patients who have been rendered euthyroid by antithyroid drugs or can be used as a first line of therapy. Its main disadvantage is the induction of hypothyroidism, which occurs in 40 to 70% of patients within 10 years after treatment. Although hypothyroidism is easily treated with L-thyroxine substitution, not all patients like the idea of lifelong substitution treatment. Radioactive iodine may aggravate concomitant ophthalmopathy, especially in smokers, in patients with high pretreatment T3 levels, and in patients whose eye disease is active. This can be prevented by concomitant administration of corticosteroids. Radiation thyroiditis occurs in a minority of patients during the first weeks after treatment and very rarely may result in a thyrotoxic storm. This is seen mainly in elderly patients with severe

Graves’ Disease, Hyperthyroidism in

hyperthyroidism; hence, the common practice is to treat these patients with antithyroid drugs first. There is no evidence of a carcinogenic or leukemogenic effect of radioactive iodine; in fact, a recent large follow-up study actually found a 10% decrease in risk for cancer mortality. Radioactive iodine and a 1-year course of antithyroid drugs are the most commonly used therapies in patients with newly diagnosed Graves’ hyperthyroidism. Subtotal thyroidectomy is reserved for young patients, especially when a course of antithyroid drugs has failed. In the United States, a preference is seen for the use of high doses of radioactive iodine aimed at induction of hypothyroidism. In Europe, patients are more often treated with a course of antithyroid drugs in light of the 50% remission rate.

See Also the Following Articles Graves’ Dermopathy . Graves’ Disease . Graves’ Ophthalmopathy . Hyperthyroidism, Subclinical . Iodine, Radioactive . Thyrotoxicosis: Diagnosis . Thyrotoxicosis, Overview of Causes

Further Reading Franklyn, J. A., Maisonneuve, P., Sheppard, M., Betteridge, J., and Boyle, P. (1999). Cancer incidence and mortality after radioiodine treatment for hyperthyroidism: A population-based cohort study. Lancet 353, 2111–2115. Lazarus, J. H. (1997). Hyperthyroidism. Lancet 349, 339–342. McDougall, I. R. (1991). Graves’ disease: Current concepts. Med. Clin. North America 75, 79–95. Pedersen, I. B., Knudsen, N., Perrild, H., Ovesen, L., and Laurberg, P. (2001). TSH receptor antibody measurement for differentiation of hyperthyroidism into Graves’ disease and multinodular toxic goitre: A comparison of two competitive binding assays. Clin. Endocrinol. 55, 381–390. Solomon, B., Glinoer, D., Lagasse, R., and Wartofsky, L. (1990). Current trends in the management of Graves’ disease. J. Clin. Endocrinol. Metab. 70, 1518–1524.

358

Graves’ Ophthalmopathy

PATHOGENESIS

Figure 1 CT scanning of the orbits of a patient with unilateral Graves’ ophthalmopathy. The eye muscles in the right orbit are small and ribbon-like, whereas one eye muscle (the medial rectus) on the left is obviously swollen.

ETIOLOGY Like Graves’ hyperthyroidism, Graves’ ophthalmopathy is an autoimmune disorder with a multifactorial background. The majority of patients with Graves’ ophthalmopathy have family members afflicted by Graves’ hyperthyroidism, but a family history of ophthalmopathy itself is uncommon. This may be partly due to the relative rareness of this complication. Another plausible explanation is that certain environmental factors are necessary to bring about the eye disease in patients with (a genetic predisposition for) Graves’ hyperthyroidism. No genetic loci have been found to be associated with an increased risk for ophthalmopathy, with the possible exception of a CTLA-4 polymorphism. Interestingly, one protective locus has been identified (HLA-DPB1*201). On the other hand, an environmental factor, smoking, has been identified beyond any doubt as a strong risk factor for the development of this eye disease. Smoking increases the risk of ophthalmopathy 7.7-fold and is especially associated with more severe eye disease. The pathogenetic importance of smoking is further underscored by the fact that the results of various therapeutic interventions are not as good in smokers as in nonsmokers. The reason behind the influence of smoking on the eye disease is unknown. Another environmental risk factor for ophthalmopathy is the treatment of the underlying hyperthyroidism with radioactive iodine. This treatment carries a small but unequivocally higher risk than does antithyroid drug treatment for the development of (usually mild and transient) ophthalmopathy.

On histology, the enlarged retrobulbar tissues show a marked lymphocytic infiltration consisting mainly of T lymphocytes and only a small number of B cells. The autoantigen responsible for attracting these immunocompetent cells is unknown, unlike the situation in Graves’ hyperthyroidism where the thyroid-stimulating hormone (TSH) receptor is the autoantigen. The current hypothesis suggests that autoreactive T lymphocytes recognize an antigen that is present in both the thyroid and the retrobulbar fibroblasts (Fig. 2). These T cells then secrete various cytokines, facilitating an amplification of the immune response by inducing adhesion molecules such as ICAM-1 on retrobulbar fibroblasts, which direct other immune cells to the target tissues. T-cell-derived cytokines may also activate B lymphocytes to differentiate into plasma cells capable of producing autoantibodies against unknown autoantigens. In addition, several cytokines possess the ability to stimulate glycosaminoglycan (GAG) production by fibroblasts and fibroblast proliferation. This chain of events leads to swelling and active inflammation of the retrobulbar tissues. As for the nature of the responsible autoantigen, one might suggest the TSH receptor given that this protein is also expressed on retrobulbar fibroblasts at certain stages of their differentiation.

Figure 2 Schematic representation of the pathogenesis of Graves’ ophthalmopathy. T lymphocytes recognize and process an antigen on the thyroid gland and mount an antithyroidal autoimmune attack, resulting in hyperthyroidism. These T cells circulate throughout the body and may become attracted to the orbit, where retrobulbar fibroblasts express homing molecules. There, the T cells recognize an antigen similar (or even identical) to the thyroidal antigen and start an autoimmune attack against the fibroblasts. The cells will produce several cytokines, which stimulate the production of glycosaminoglycans by the retrobulbar fibroblasts.

359

Graves’ Ophthalmopathy

RELATIONSHIP WITH GRAVES’ HYPERTHYROIDISM Graves’ ophthalmopathy is usually seen in patients in whom a diagnosis of Graves’ hyperthyroidism is well established. In the majority of patients, the eye disease develops concomitantly with or some months after the onset of the thyroid disease. However, in 25% of patients, the eye disease precedes a diagnosis of hyperthyroidism by a number of months. sometimes, there is considerable delay between the onset of the two manifestations that in rare cases may even be several years. In these circumstances, it can be difficult to make the correct diagnosis. In a small minority (5%), patients with ophthalmopathy present with primary hypothyroidism instead of hyperthyroidism. The pathogenesis of this thyroid failure is unclear but may be due to so-called TSH-blocking autoantibodies against the TSH receptor. These antibodies render the thyroid gland devoid of the stimulating effect of TSH, resulting in atrophy and hypothyroidism.

Graves’ ophthalmopathy causes a plethora of signs and symptoms that are summarized in Table I, the so-called NO SPECS classification. This classification system was first developed in 1969, was subsequently revised on a number of occasions, and now serves as a good memory aid. The various signs and symptoms might seem complex (Fig. 3), but all can be explained from a mechanical point of view.

The lid retraction causes stare and lid lag on downward gaze (Von Graefe’s sign) and can be due to swelling of the superior levator muscle. However, thyrotoxicosis per se can also induce this sign by increasing the sympathetic tone, so Von Graefe’s sign is sometimes also present in hyperthyroidism not caused by Graves’ disease. Sympathetic overactivity is not the only cause of lid retraction given that upper eyelid retraction frequently remains present when ophthalmopathy patients are rendered euthyroid. If severe, lid retraction may lead to lagophthalmos, that is, incomplete closure of the eyelids at night.

Class 1: Only Signs, No Symptoms

Class 2: Soft Tissue Involvement

This refers to the upper eyelid retraction frequently observed in patients with Graves’ hyperthyroidism.

This entails chemosis (edema of the conjunctiva), conjunctival injection and redness, swelling of the

CLINICAL MANIFESTATIONS p0040

t0005

Figure 3 A patient with severe Graves’ ophthalmopathy. Note the abnormal eyeball position, resulting in double vision, swelling of the eyelids, increased lid aperture, and periorbital swelling.

Table I NO SPECS Classification of Eye Changes in Graves’ Ophthalmopathy Class

Description

0

No signs or symptoms

1 2

Signs and symptoms

Cause

—

—

Only signs, no symptoms

Lid retraction, stare, lid lag

Increased sympathetic tone

Soft tissue involvement

Swelling of eyelids, chemosis, photophobia, grittiness

Impaired venous drainage, herniated orbital fat

3

Proptosis

Exophthalmos

Increased retrobulbar pressure pushing globe forward

4

Restricted eyeball motility (often with diplopia)

Swollen eye muscles

5

Extraocular muscle involvement Corneal involvement

Keratitis, corneal ulcer

Overexposure of cornea

6

Sight loss

Decreased visual acuity due to optic nerve involvement, impaired color vision, visual field defects

Pressure on optic nerve, apical crowding

360 caruncle, and swelling of the upper and lower eyelids (periorbital swelling). These findings are partly explained by impaired venous drainage as a result of the increase in volume of the retrobulbar tissues. Periorbital swelling is also due to herniation of retrobulbar fatty tissues through openings in the orbital septum covering the retrobulbar cavity.

Class 3: Proptosis Because of the confining bony surroundings, the swollen retrobulbar tissues have no other outlet than pushing the globe forward. Hence, exophthalmos may be seen as ‘‘nature’s own decompression.’’

Class 4: Extraocular Muscle Involvement One can easily imagine that swelling of the normally very thin extraocular eye muscles leads to impaired mobility. If the impairment is asymmetrical, the patient will have double vision. However, if the impairment is symmetrical, no diplopia will occur. Sometimes, the patient will keep the neck in a certain position (usually bent backward) to correct for impaired motility. This so-called ocular torticollis may lead to painful neck muscles.

Class 5: Corneal Involvement Exophthalmos, lid retraction, lagophthalmos, and less frequent blinking all contribute to an excessive exposure of the cornea to air that can lead to inflammation of the cornea (keratitis). Early signs are photophobia, a gritty sensation, intolerance to contact lenses, and blurred vision. This phenomenon is different from diplopia in that the abnormal images disappear after blinking.

Graves’ Ophthalmopathy

Other Signs and Symptoms Apart from the manifestations described in the NO SPECS classification, patients in the active stages of the eye disease often complain of a dull pain or pressure on or behind the eyeball. Pain can also be felt during attempted up, side, or down gaze. Perhaps the most important complaint is the change in appearance. Bulging of the eyes, swelling of the eyelids or even only lid retraction, and an abnormal position of the globes all contribute to a sometimes very marked change in appearance. This will be better appreciated if the patient shows photographs of himself or herself taken before onset of the disease.

p0075

QUALITY OF LIFE In view of these many and different clinical manifestations, it is not surprising that patients suffer from a diminished quality of life. The changes in appearance as a consequence of proptosis and periorbital swelling can be profound. Diplopia will hamper many activities in daily life such as reading and driving. In fact, research has established that even patients with mild to moderately severe eye disease already have a markedly decreased sense of well-being. They rate their degree of social and role functioning lower than do patients with other chronic diseases such as diabetes mellitus. Use of a disease-specific quality of life questionnaire developed with the aid of patient self-support groups has shown unequivocally that Graves’ ophthalmopathy is a seriously disabling disorder. The disease leads to feelings of social isolation in as many as 40% of these patients. Half of the patients notice unpleasant reactions from others, and many do not want to appear in photographs. As a consequence, as many as 70% of patients with mild to moderately severe ophthalmopathy report a marked decrease in selfconfidence.

Class 6: Sight Loss Sight loss can occur if the enlarged eye muscles compress the optic nerve. This can occur in the apex of the orbital cavity where the optic nerve leaves the orbit. On CT scanning, no room is seen between the optic nerve and the swollen muscles; this is called ‘‘apical crowding.’’ Early signs of optic nerve involvement are impaired color vision and visual field defects. This severe complication is more often seen in males and in patients without significant proptosis. In those patients, a tight orbital septum precludes forward displacement of the globe, causing a rise in retrobulbar pressure that is damaging to the optic nerve.

DIAGNOSIS In patients with clear evidence of Graves’ hyperthyroidism who present with bilateral proptosis, a diagnosis of Graves’ ophthalmopathy is easily made. However, this diagnosis may be less obvious in patients with unilateral disease or in patients without a thyroid disorder. Nevertheless, even in these patients, the most likely diagnosis will be Graves’ eye disease. It should be noted that there is no diagnostic procedure establishing a diagnosis of thyroid-associated eye disease unequivocally, but several imaging procedures may be helpful. CT scanning of the orbits typically

p0085

361

Graves’ Ophthalmopathy

will reveal swelling of the extraocular eye muscles, of the retrobulbar tissues, or of both (Fig. 1). Magnetic resonance imaging (MRI) will show the same but may also demonstrate a prolonged T2 relaxation time, suggestive of edematous swelling. Other diagnostic imaging procedures include octreotide scintigraphy, where labeled octreotide binds to somatostatin receptors on activated orbital lymphocytes, and ultrasound. However, enlargement of the eye muscles or connective tissues is not definitive proof of the existence of ophthalmopathy. Other diagnoses that should be entertained are lymphomas or metastases of carcinomas to the orbit and the rare disease entity of an orbital pseudotumor. It is the combination of various signs and symptoms of ophthalmopathy, together with evidence of an autoimmune thyroid disease, that will trigger the physician to make the correct diagnosis.

GENERAL THERAPEUTIC PRINCIPLES The first step in the management of a patient with ophthalmopathy is adequate treatment of the underlying thyroid disorder. Restoration of the euthyroid state frequently leads to some amelioration of soft tissue involvement and sometimes even of diplopia in 2 to 3 months. Generous application of lubricants (eye drops should be applied at least six times daily) prevents corneal damage. Lagophthalmos is frequent, and application of a protective gel at night protects the exposed cornea during sleep. Further therapeutic measures depend on the severity and stage of the eye disease.

DISEASE ACTIVITY From observations by the Australian physician, F. F. Rundle, it has become clear that the disease has a tendency toward spontaneous improvement (Fig. 4). After reaching a peak in severity, the signs and symptoms gradually ameliorate over a highly variable period of time, from several months to a number of years. However, in most patients, a complete restoration to the premorbid state is seldom reached. On histology, the active state is characterized by lymphocytic infiltration, whereas during the chronic phase, fibrotic scar tissue is found. It is conceivable that immunosuppressive therapies will be effective during the active stage only and that rehabilitative surgery is the treatment of choice when the disease is in its inactive phase. Unfortunately, reliably assessing the stage of the disease in individual patients has been difficult. Obviously, when the eye condition is rapidly

Severity GO

Dynamic phase Static phase

Time

Figure 4 Rundle’s curve showing the natural history of Graves’ ophthalmopathy and its tendency toward spontaneous improvement over a variable period of time. Note that the x axis (time) is not defined.

deteriorating, it is in its active stage. On the other hand, when the ophthalmopathy has been stable for 6 months, it is likely to have become inactive. However, in many patients with moderately severe disease, such changes are less obvious and one prefers to assess the disease stage without having to observe symptomatic patients during several months without therapy. Therefore, various methods to assess disease activity have been developed, and the ones used most frequently are mentioned in Table II. However, none of these measures is entirely satisfactory in predicting a successful outcome of immunosuppressive therapy. If one or more of the measures from Table II indicates activity, it is reasonable to assume that the disease is still amenable to immunosuppressive therapy (if the severity of the disease warrants treatment).

MANAGEMENT DURING THE ACTIVE STAGE One option during the initial phase of the disease is to observe the patient over the course of several months. When the disease is not severe, this is certainly a good option in view of the natural tendency toward spontaneous improvement and because of the side effects associated with the various treatment options. Adequate antithyroid treatment and lubrication of the cornea will alleviate some of the symptoms until the disease has become inactive and surgical rehabilitation is possible. In more severe cases, nonsurgical (immunosuppressive) treatment can be indicated. Both corticosteroids and retrobulbar irradiation have been used for this indication for many decades. The

362

t0010

Graves’ Ophthalmopathy

Table II Commonly Used Parameters for Disease Activity and Their Cutoff Values Indicating Active Disease Parameter

Cutoff value

Duration of signs or symptoms

3 months). . Give alternate-day regimen if prolonged systemic treatment is

required (growth suppression is likely with cortisone 30 mg/m2 daily, prednisone  4mg/m2 daily, betamethasone  0.6mg/m2 daily). . Use concomitantly with glucocorticoid-sparing agents

(e.g, cyclosporin-A for nephrotic syndrome). . Avoid during infancy and/or puberty.

373

Growth and Glucocorticoids

Concomitant Effects on the Hypothalamic–Pituitary–Adrenal Axis, Body Composition, and Bone Metabolism In addition to inhibiting linear growth, supraphysiological doses of oral glucocorticoids and high doses of locally administered glucocorticoids can suppress hypothalamic–pituitary–adrenal (HPA) function and may reduce bone mass. It is not clear exactly how these major side effects relate to each other in terms of dose–response and whether different glucocorticoid preparations have the same hierarchy of systemic effects.

GLUCOCORTICOID THERAPY IN SPECIFIC DISEASES Glucocorticoid Replacement for Abnormal Hypothalamic– Pituitary–Adrenal Axis Congenital Adrenal Hyperplasia and Glucocorticoid Substitution Treatment Substitution treatment with glucocorticoids is essential in the management of congenital adrenal hyperplasia (CAH) to suppress excessive hypothalamic corticotropin-releasing hormone (CRH) and pituitary adrenocorticotropic hormone (ACTH) secretion, reduce circulating levels of adrenal androgens, and provide physiological replacement. This is achieved with hydrocortisone (10–15 mg/m2/day in two or three doses, i.e., above the physiological cortisol production rate). Overtreatment, especially during the first 2 years of life, and undertreatment can result in abnormal growth during childhood and in failure to reach adult height potential. Potential future approaches with a reduced hydrocortisone dose or alternative glucocorticoid preparations along with other treatments, such as an antiandrogen (flutamide) or an aromatase inhibitor (testolactone) and luteinizing hormone-releasing hormone (LHRH) agonist with or without growth hormone (GH), may improve growth and adult height in patients with CAH.

Primary Adrenal Insufficiency The most common cause of primary adrenal insufficiency during childhood is autoimmune disease. Height is below average but not usually less than the 3rd percentile. Growth is normal with adequate replacement, and although the onset of puberty may be delayed, pubertal development is normal.

Familial Glucocorticoid Deficiency (ACTH Resistance) These autososmal-recessive disorders are characterised by a failure of the adrenal cortex to respond to ACTH but normal mineralocorticoid secretion. Tall stature and advanced skeletal maturation have been described despite glucocorticoid deficiency. Physiological replacement with glucocorticoids allows normal growth.

s0065

Hypopituitarism Deficiency of GH, ACTH, TSH, and gonadotropins contribute to the growth failure and the delayed or lack of pubertal development seen in hypopituitarism. As with other glucocorticoid-deficient states, the lowest dose of hydrocortisone necessary to prevent the symptoms of adrenal failure should be used to avoid suppression of growth.

Pharmacotherapy in Diseases That Do Not Primarily Involve the Hypothalamic–Pituitary–Adrenal Axis Atopic Diseases Asthma and Inhaled Glucocorticoid Treatment Impaired growth and a higher prevalence of short stature in children with atopic disease were reported well before the availability of glucocorticoids (8–25% of children with asthma not treated with glucocorticoids and up to 45% of those treated with daily oral glucocorticoids had a height standard deviation score [SDS]  2). Independent of glucocorticoid treatment, short stature in children with asthma is associated with delayed bone age, delay in the onset of puberty, and delay in the pubertal growth spurt by approximately 1.3 years. Although the mean adult height attained is normal, up to 20% of those with severe unremitting asthma have final height less than the 10th percentile. The degree of growth impairment is greater in those with more severe asthma and with onset of asthma before 3 years of age. Inhaled glucocorticoids are the treatment of choice for childhood asthma, and dose–response studies suggest that doses of less than 400 mg/d of any inhaled glucocorticoid are beneficial in controlling symptoms. Higher doses confer little added advantage, but the potential for adverse effects increases dramatically. Lower leg length assessed by knemometry is a highly sensitive measure of the dose-dependent systemic effects of inhaled glucocorticoids over days to weeks (short term) but does not predict longer term growth or final adult height. The potential for linear

p0075

374

Growth and Glucocorticoids

Inhaled glucocorticoid: Dose inhaled differs from dose prescribed Roughly 80−90% is deposited in the oropharynx and amount swallowed can be reduced by mouth rinsing Roughly 10−20% is deposited in the lungs, influenced by the delivery device volume and shape electrostatic charge of walls volume of dead space presence and type of valves the child tidal volume breathing pattern technique of using device Absorption from the lungs

Absorption of swallowed drug from the gut Liver

Beclomethasone dipropionate is metabolized to active beclomethasone monopropionate

90−100% of budesonide and fluticasone propionate are inactivated by firstpass metabolism

Systemic circulation

Side effects on growth, HPA axis, and other tissues

Figure 1 The pharmacokinetics of inhaled glucocorticoids.

growth impairment with inhaled glucocorticoid treatment depends on systemic bioavailability (Fig. 1), which in turn is influenced by the delivery device, inhalation technique, prescribed dose, lipophilicity, pharmacokinetics of the glucocorticoid preparation, and severity of asthma. Randomized controlled trials help to distinguish the effects on linear growth of inhaled glucocorticoid treatment from those of the asthma itself. At doses of at least 400 mg/day, beclomethasone dipropionate suppresses linear growth by approximately 1.5 cm/year but does not necessarily suppress the HPA axis. Although it is advisable to use the minimum dose necessary to control symptoms, undertreatment that compromises asthma control may also impair linear growth. In addition, there is no evidence of impairment in final adult height with prolonged inhaled or intranasal glucocorticoids in recommended doses. Growth failure associated with increased body fat with centripetal distribution and reduced lean mass indicative of iatrogenic Cushing’s syndrome has been reported in patients on high doses of inhaled glucocorticoids. Paradoxically, poor linear growth and poor weight gain associated with endogenous cortisol insufficiency but without these well-known peripheral effects of excess exogenous glucocorticoids have also been described in a small number of children treated with inhaled glucocorticoids.

Chronic Rhinitis and Intranasal Glucocorticoid Treatment Intranasal glucocorticoids are increasingly being used as a first-line treatment for conditions associated with chronic mucosal inflammation. Systemic effects, including growth impairment, can occur and depend on the dose and duration of treatment. As with inhaled treatment, growth suppression may occur without adverse effects on the HPA axis. It has been observed for lower leg growth and linear growth, for example, with budesonide nasal spray (400 mg/day) and beclomethasone dipropionate (>160 mg/day for 6 weeks to 12 months) but not with mometasone furoate (100 mg/day). Like fluticasone propionate, mometasone furoate has reduced systemic bioavailability when absorbed from the gut, owing to complete first-pass hepatic metabolism. There are no longterm studies of growth in children treated with intranasal glucocorticoids. Atopic Dermatitis and Topically Applied Glucocorticoid Treatment A number of factors influence percutaneous absorption of topically applied glucocorticoids and the potential for systemic effects such as the formulation, method of application, sites treated, and severity of dermatitis. Growth retardation in association with Cushing’s syndrome and HPA suppression, with catch-up growth when treatment was stopped or

375

Growth and Glucocorticoids

reduced, has been reported in isolated cases treated with topical glucocorticoids. There are no randomized controlled trials on the systemic effects of topical glucocorticoid treatment in children and, thus, no studies that clearly separate the contribution made by such treatment from other disease factors that might adversely affect growth. Impaired growth in children with atopic dermatitis and a higher prevalence of atopic dermatitis among children with shorter stature than those with normal stature have been reported (up to 10% of children with atopic dermatitis have height 4.5 SD in height) have GH1 gene deletions. Since gene deletions as well as frameshift and nonsense mutations have been found to cause the IGHD IA phenotype, this disorder is best described as complete GHD due to heterogeneous GH1 gene defects, rather than gene deletions alone (see Table I).

f0005

Figure 1 The growth hormone (GH) biosynthetic pathway (for details, see text).

having complicated tests to detect GHD. Provocative tests for GHD include GH stimulation tests. Deficient GH peak responses range from 7 to 10 ng/ ml. Testing for concomitant deficiencies of LH, FSH, TSH, and/or ACTH should be performed on GHD patients to detect CPHD to provide a complete diagnosis and enable planning of optimal treatment.

Types of Familial IGHD Several types of familial IGHD are associated with at least six different Mendelian disorders. These include four autosomal-recessive disorders (IGHD 1A, IGHD 1B, bioinactive GH defects, and GHRHR defects, OMIM Nos. 262400, 139250, 262650, and 139191, respectively). In addition, there is an autosomaldominant form (IGHD II, OMIM No. 173100) and an X-linked form of IGHD (IGHD III, OMIM No. 307200; see Table I). IGHD IA The most severe form of IGHD, called IGHD IA (OMIM Nos. 262400 and 139250), has an autosomal-recessive mode of inheritance. Affected neonates occasionally have mildly decreased birth lengths and hypoglycemia in infancy. All develop severe dwarfism

IGHD IB A milder form of IGHD, IGHD IB, also has an autosomal-recessive mode of inheritance. These cases differ clinically from IGHD IA in their having low but detectable levels of GH and a continued growth response due to immunological tolerance to treatment with exogenous GH. IGHD IB cases are caused by GH gene defects that result in a mutant GH protein that may not be detected by radioimmunoassay (RIA). The presence of these mutant GH protein molecules may explain the good responses that are seen with GH therapy because their presence mitigates against the production of anti-GH antibodies. IGHD IB is caused by mutations that affect splicing of the GH1 gene. This altered splicing causes loss of amino acids that affect the stability and biological activity and reduce the secretion of the mutant GH protein. IGHD II IGHD II has an autosomal-dominant mode of inheritance due to dominant-negative mutations of the GH1 gene and patients respond well to GH treatment. Almost all the GH1 gene defects reported in IGHD II are mutations that alter the splicing of GH mRNA and cause skipping or deletion of exon 3. The mechanism by which these dominant-negative mutations prevent expression of GH protein from the other, normal GH1 gene is poorly understood. Other IGHD II mutations cause skipping of exon 3 by disrupting splicing enhancer sequences (SEs) that regulate the splicing pattern of GH mRNA and, when these SEs are perturbed, exon 3 skipping occurs. An

399

Growth Hormone Deficiency, Genetic

IGHD II mutation that does not cause abnormal splicing is a G-to-A transition that results in an Arg to His substitution at residue 183 (Arg183His) of the GH molecule. This substitution is thought to alter the intracellular processing of the GH molecule by binding to zinc, thereby deranging the zincassociated presecretory packaging of GH.

IGHD III A third form of IGHD called IGHD III (OMIM No. 307200) has an X-linked mode of inheritance and there have been distinct clinical findings in different families. In some families, all cases have agammaglobulinemia associated with their IGHD, whereas in other families, all cases have only IGHD. This suggests that contiguous gene defects on the long arm of the X chromosome may cause some IGHD III cases. Duriez et al. reported that X-linked agammaglobulinemia and IGHD are caused by a mutation in the Bruton’s tyrosine kinase (BTK) gene. Laumonnier et al. studied the SOX3 gene in families with X-linked mental retardation, where the causative gene had been mapped to Xq26–q27. They showed that the SOX3 gene maps to Xq26.3 and was involved in a large family in which affected individuals had mental retardation and IGHD (OMIM Nos. 300123 and 313430; Table I). The mutation was an in-frame duplication of 33 bp encoding 11 alanines in a polyalanine tract of the SOX3 gene. The expression pattern during neural and pituitary development suggested that dysfunction of the SOX3 gene caused by this polyalanine expansion might disturb transcription pathways and the regulation of genes involved in pituitary development.

Biodefective GH There have been a number of reports of patients with the clinical features of IGHD who achieved normal plasma immunoactive GH levels following GH provocative or stimulation tests, but low levels of somatomedin (OMIM No. 139250; Table I). Less GH was detected by radioreceptor assay than by RIA analysis in some studies. In view of these patients’ clinical syndrome of IGHD, their apparently normal plasma concentrations of GH, their low basal somatomedin levels, and their normal response to exogenous GH, individuals with bioinactive GH are thought to secrete a biologically inert GH. Takahashi et al. identified a C-to-T transition at codon 77, which results in an Arg to Cys substitution in the GH1 gene of a subject diagnosed with bioinactive GH.

GHRH Receptor Defects A variety of mutations have been detected in the human GHRHR gene in individuals with IGHD (OMIM No. 139191). In a kindred with a nonsense mutation, affected family members had poor growth since infancy and were extremely short. They failed to produce GH in response to standard provocative tests and had good responses to GH replacement. Cases were homozygous for a G-to-T transversion that caused a premature termination mutation (Glu72Stop). Salvatori et al. described a large Brazilian family with many family members with IGHD due to an intronic G-to-T transition that destroys the 50 splice site of IVS1 of the GHRHR gene.

COMBINED PITUITARY HORMONE DEFICIENCY Cases with CPHD vary in their clinical findings because they have deficiencies of varying severity of one or more of the other pituitary tropic hormones (ACTH, FSH, LH, or TSH) in addition to GHD (OMIM No. 262600). Whereas most cases of CPHD are sporadic, a variety of familial forms that can have autosomal-recessive, autosomal-dominant, or X-linked modes of inheritance are known.

HESX1 Mutations HESX1 is expressed in the thickened layer of oral ectoderm that gives rise to Rathke’s pouch, the primordium of the anterior pituitary. Down-regulation of HESX1 coincides with the differentiation of pituitaryspecific cell types. Dattani et al. found a missense HESX1 mutation (Arg53Cys) in a homozygous state in a brother and sister with septo-optic dysplasia, agenesis of the corpus callosum, and CPHD (OMIM No. 182230).

LHX3 Mutations Murine Lhx3 mRNA accumulates in Rathke’s pouch, the primordium of the pituitary, and may be involved in the differentiation of pituitary cells. Netchine et al. identified two families with CPHD (OMIM No. 262600) caused by mutations in the LHX3 gene. The phenotype associated with these mutations included the following: (1) severe growth retardation, (2) complete deficiency of all but one of the anterior pituitary hormones (ACTH), (3) elevated and anteverted shoulders with a short neck associated with severe restriction of rotation of the cervical spine, and (4) an

400 enlarged anterior pituitary. The authors concluded that LHX3 is required for the proper development of all anterior pituitary cell types except corticotropes and that the rigid cervical spine phenotype is consistent with a function of LHX3 in the proper development of extrapituitary structures as well.

PIT1 Mutations Defects in the PIT1 gene cause familial CPHD cases, which have a different phenotype (OMIM No. 173110). PIT1 is an anterior pituitary-specific transcription factor that regulates the expression of GH, prolactin (PRL), and TSH. PIT1 is also required for pituitary cellular differentiation and function. PIT1 has functional domains that enable the transactivation of other genes including GH, PRL, and TSH or binding to these genes. At least six different PIT1 mutations causing autosomal-recessive CPHD and two others causing autosomal-dominant CPHD have been found in humans in a subtype of panhypopituitary dwarfism associated with GH, PRL, and TSH deficiency (see Table I).

PROP1 Mutations PROP1 is a pituitary-specific homeodomain factor that is required for the development of somatotropes, lactotropes, and thyrotropes of the anterior pituitary and for the expression of PIT1. Multiple PROP1 gene mutations cause an autosomal-recessive CPHD that has a third phenotype in humans (OMIM No. 601538; Table I). In addition to deficiencies of GH, PRL, and TSH that are seen in those with PIT1 defects, subjects with PROP1 defects also have deficiencies of LH and FSH, which prevent the onset of spontaneous puberty and, in some cases, ACTH deficiency in later life. The various PROP1 mutations include (1) a C-to-T transition at codon 120, which encodes a TGC (Arg) to CGC (Cys) substitution; (2) a T-to-A transversion that encodes a TTC (Phe) to ATC (Ile) substitution at codon 117; and (3) a 2 bp AG deletion in codon 101 (101delAG) that causes a frameshift and results in a premature stop at codon 109. The resulting protein products from all three of these different PROP1 mutations have greatly reduced DNA-binding and transactivation abilities. The 101delAG is a recurring mutation that is estimated to occur in 55% of familial and 12% of sporadic CPHD cases. A fourth PROP1 mutation is a 2 bp GA deletion at codon 51 (51delGA). Like the 101delAG mutation, the 51delGA mutation causes a frameshift that results in a premature stop codon.

Growth Hormone Deficiency, Genetic

This mutation was found in 12% of familial and 21% of sporadic CPHD cases.

X-Linked CPHD Lagerstrom-Fermer et al. described a family that included affected males suffering from variable degrees of CPHD (OMIM No. 312000). Some affected males who died during the first day of life had postmortem findings of hypoadrenalism, presumed to be due to CPHD. Others had variable combinations of hypothyroidism, delayed pubertal development, and short stature due to GHD. All surviving patients exhibited mild to moderate mental retardation. They found linkage with markers in the Xq25–q26 region. Furthermore, they found an apparent extra copy of the marker DXS102 in affected males and heterozygous carrier females, suggesting that a segment including this marker was duplicated.

MENDELIAN DISORDERS WITH ENDOCRINE ABNORMALITIES A variety of Mendelian disorders have among their pleiotropic effects endocrine abnormalities. These disorders include some abnormalities, such as achondroplasia, that have a single common mutation and others, such as hemoglobinopathies, that are caused by heterogeneous mutations.

Achondroplasia with Obstructive Sleep Apnea Achondroplasia is a common skeletal dysplasia in which the dwarfism is due to an abnormality in endochondral ossification (see OMIM No. 100800; Table I). Up to 10% of patients with achondroplasia have been reported to have serious respiratory complications. Some with achondroplasia and obstructive sleep apnea have low growth hormone secretion during sleep as a contributing cause of their growth retardation.

Borjeson-Forssman-Lehmann Syndrome The X-linked Borjeson-Forssman-Lehmann syndrome is characterized by short stature, hypogonadism, hypotonia, severe mental deficiency, and coarse facial appearance with a prominent brow ridge and large ears in affected males (see OMIM No. 301900; Table I). Markedly deficient GH responses to arginine and l-DOPA as well as low somatomedin C levels have been documented in affected individuals.

401

Growth Hormone Deficiency, Genetic

CHARGE Association

Neurofibromatosis Type 1

CHARGE is an acronym that describes a nonrandom association of anomalies: colobomas of the eye; heart disease; atresia of the choanae; retarded growth and development and/or CNS anomalies; genital hypoplasia; and ear anomalies or deafness (see OMIM No. 214800; Table I). Growth retardation, which is usually of postnatal onset, and hypogonadism are prominent features of the CHARGE syndrome and may well be due to hypothalamic defects.

A variety of endocrine disturbances have been reported in patients with neurofibromatosis, which has an autosomal-dominant mode of inheritance (see OMIM No. 162200; Table I). The most common associated endocrine disorder in children with neurofibromatosis type 1 (NF1) is sexual precocity, whereas the most common associated endocrine disorder in adults is pheochromocytoma. Marked growth retardation and GH deficiency have also been reported.

Fanconi Anemia

Pallister-Hall Syndrome

Fanconi’s syndrome is an autosomal-recessive disorder characterized by chronic pancytopenia with bone marrow hypoplasia, abnormal pigmentation, upper limb malformations, kidney anomalies, growth retardation, small genitalia, and increased frequency of chromosomal breaks in cultured lymphocytes (see OMIM Nos. 227650, 227660, 227645, 227646, and 600901; Table I). A number of investigators have documented GH deficiency in patients with Fanconi anemia and administration of GH resulted in excellent short-term and long-term responses in most of these patients.

This neonatally lethal malformation syndrome consists of hypothalamic hamartoblastoma, hypopituitarism, postaxial polydactyly, and imperforate anus (see OMIM No. 146510; Table I). An anterior pituitary gland was absent in all cases. The posterior pituitary was absent in the majority of cases.

Hemochromatosis Both male hypogonadism and pituitary hemosiderosis can occur in hemochromatosis (see OMIM Nos. 235200 and 602390; Table I) and abnormalities have also been found in gonadotropin, cortisol, GH, PRL, and TSH secretion.

Hemoglobinopathies There are well-documented cases of acquired pituitary insufficiency occurring in adults with hemoglobinopathies, presumably secondary to infarction of the gland (see OMIM No. 141900; Table I).

Histiocytosis X Histiocytosis X (also known as Letterer-Siwe disease, Hand-Schuller-Christian disease, or eosinophilic granuloma) is characterized by foamy histiocyte infiltration in many areas of the body, including the hypothalamus. When the histiocytic infiltration involves the hypothalamus, prepubertal growth retardation associated with GH deficiency and diabetes insipidus frequently occur (see OMIM No. 246400; Table I).

Rieger’s Syndrome Rieger’s syndrome (also known as iris–dental dysplasia) is an autosomal-dominant disorder associated with malformation of the iris, pupillary anomalies, and hypoplasia of the teeth, with or without maxillary hypoplasia. A large family in which multiple individuals had both Rieger’s syndrome and IGHD (see OMIM No. 180500; Table I) has been described. Siblings of the proband had Rieger’s syndrome with normal pituitary function, but GHD was not found in any member of the family who did not have Rieger’s syndrome. One subject who was treated with GH exhibited substantial enhancement of his rate of growth.

CHROMOSOMAL DISORDERS WITH ENDOCRINE ABNORMALITIES A large variety of chromosomal abnormalities are associated with endocrine disorders. These chromosomal anomalies can affect a variety of autosomes as well as the sex chromosomes.

Autosomal GH deficiency has been described with 18p and 20p chromosomal deletions (see OMIM No. 146390; Table II). In addition, molecular detection of deletions of 17q (i.e., the GH gene cluster region) has been documented in IGHD IA. Since the gene for

402

Growth Hormone Deficiency, Genetic

GHRH has been mapped to 20p, the GH deficiency in 20p could result from either the deletion of the GHRH gene or a developmental anomaly of the hypothalamus. Patients with 18p have been reported with hypopituitarism and solitary central maxillary incisor, suggesting that the pituitary insufficiency in 18p may be due to a structural malformation of the hypothalamus. GHD has been rarely associated with 47,XXY, 49,XXXXY, and ring 5. Finally a search of the London Dysmorphology Database at http:// www.hgmp.mrc.ac.uk/DHMHD/view.html identifies a series of partial chromosome deletions or duplications that are associated with short stature and pituitary abnormalities (see Table II). These include the deletions del(4)pter–p16, del(7)q32–qter, del(13)q22– qter, del(14)q22–q23, del(18)p, del(18)q21–qter, and del(22)pter–q11 and the duplications dup(1)q25–q32, dup(9)p, dup(9)pter–q22, and dup(11)q23–qter.

Turner’s Syndrome (45,X) Although GH secretion has been reported to be normal, or paradoxically increased, in most patients with gonadal dysgenesis, pituitary insufficiency has been reported in several patients (see Table II). Ross et al. studied GH secretion in 30 patients with Turner’s syndrome and found no differences in mean GH concentration or peak amplitudes throughout the day and night between patients less than 8 years of age and controls. Patients over 9 years of age had lower

t0010

mean GH levels and peak amplitudes. Reduced plasma somatomedin C levels and delayed bone age were found in patients of all ages. These abnormalities in GH secretion in Turner’s syndrome are probably secondary to the absence of sex hormones during adolescence.

TREATMENT OF PITUITARY GENE MUTATIONS Recombinant-derived GH is widely available but must be given by subcutaneous injection. To obtain an optimal outcome, children with GHD should be started on replacement therapy as soon as their diagnosis is established. The dosage increases with increasing body weight to a maximum during puberty and is usually discontinued by 17 years of age. Disorders in which GH treatment is of proven efficacy include GHD, either isolated or in association with CPHD, and Turner’s syndrome. The clinical responses of individuals with IGHD or CPHD to GH replacement therapy vary depending on (1) the severity and age at which treatment is begun, (2) the recognition and response to treatment of associated deficiencies such as thyroid hormone deficiency, and (3) whether treatment is complicated by the development of anti-GH antibodies. The outcome of Turner’s syndrome subjects varies with the severity of their (1) short stature, (2) chromosomal complement, and (3) age when treatment began.

Table II Chromosomal Disorders with Growth Hormone Deficiency Disorder

Cytogenetic change

Endocrine features

Comments

Autosomes Deletions

del(4)pter–p16 del(7)q32–qter

Pituitary abnormalities with decreased GH, CPHD, or other deficiencies

Replace appropriate hormone(s)

Pituitary abnormalities with decreased GH, CPHD, or other deficiencies

Replace appropriate hormone(s)

del(13)q22–qter, del(14)q22–q23 del(18)p del(18)q21–qter del(20)p del(22)pter–q11 Duplications

dup(1)q25–q32 dup(9)p dup(9)pter–q22 dup(11)q23–qter

Rings Sex chromosomes

Ring 5

Decreased GH or CPHD

Replace appropriate hormone(s)

45,X

Decreased GH or CPHD

Replace appropriate hormone(s); GH & estrogen for 45,X

47,XXY 49,XXXXY

Growth Hormone Deficiency, Genetic

See Also the Following Articles Constitutional Delay of Growth and Puberty (CDGP) . Gigantism: Excess of Growth Hormone . Growth and Chronic Disease . Growth Hormone (GH) . Growth Hormone Insensitivity . Growth Hormone-Binding Proteins . Postnatal Normal Growth and Its Endocrine Regulation . Short Stature and Chromosomal Abnormalities . Turner Syndrome

Further Reading Cogan, J. D., and Phillips, J. A., III (2001). Inherited defects in growth hormone synthesis and action. In ‘‘The Metabolic and Molecular Bases of Inherited Disease’’ (C. R. Scriver, A. L. Beaudet, W. S. Sly, and D. Valle, eds.), 8th ed., pp. 4159–4180. McGraw-Hill, New York. Cogan, J. D., Wu, W., Phillips, J. A., III, Arnhold, I. J. P., Agapito, A., Fofanova, O. V., Osorio, M. G. F., Bircan, I., Moreno, A., and Mendonca, B. B. (1998). The PROP1 2-bp deletion is a common cause of CPHD. J. Clin. Endocrinol. Metab. 83, 3346–3349. Dattani, M. T., Martinez-Barbera, J.-P., Thomas, P. Q., Brickman, J. M., Gupta, R., Martensson, I.-L., Toresson, H., Fox, M., Wales, J. K. H., Hindmarsh, P. C., Krauss, S., Beddington, R. S. P., and Robinson, I. C. A. F. (1998). Mutations in the homeobox gene HESX1/Hesx1 associated with septo-optic dysplasia in human and mouse. Nat. Genet. 19, 125–133. Duriez, B., Duquesnoy, P., Dastot, F., Bougneres, P., Amselem, S., and Goossens, M. (1994). An exon-skipping mutation in the btk gene of a patient with X-linked agammaglobulinemia and isolated growth hormone deficiency. FEBS Lett. 346, 165–170. Fofanova, O., Takamura, N., Kinoshita, E., Parks, J. S., Brown, M. R., Peterkova, V. A., Evgrafov, O. V., Goncharov, N. P., Bulatov, A. A., Dedov, I. I., and Yamashita, S. (1998). Compound heterozygous deletion of the PROP-1 gene in children with combined pituitary hormone deficiency. J. Clin. Endocrinol. Metab. 83, 2601–2604. Lagerstrom-Fermer, M., Sundvall, M., Johnsen, E., et al. (1997). X-linked recessive panhypopituitarism associated with a regional duplication in Xq25–q26. Am. J. Hum. Genet. 60, 910–916. Laumonnier, F., Ronce, N., Hamel, B. C. J., Thomas, P., Lespinasse, J., Raynaud, M., Paringaux, C., van Bokhoven, H., Kalscheuer, V., Fryns, J.-P., Chelly, J., Moraine, C., and Briault, S. (2002). Transcription factor SOX3 is involved in X-linked

403 mental retardation with growth hormone deficiency. Am. J. Hum. Genet. 71, 1450–1455. Netchine, I., Sobrier, M.-L., Krude, H., Schnabel, D., Maghnie, M., Marcos, E., Duriez, B., Cacheux, V., Moers, A. V., Goossens, M., Gruters, A., and Amselem, S. (2000). Mutations in LHX3 result in a new syndrome revealed by combined pituitary hormone deficiency. Nat. Genet. 25, 182–186. Online Mendelian Inheritance in Man (OMIM). (2000). McKusick– Nathans Institute for Genetic Medicine, Johns Hopkins University (Baltimore, MD) and National Center for Biotechnology Information, National Library of Medicine, National Institutes of Health (Bethesda, MD). World Wide Web URL: http:// www.ncbi.nlm.nih.gov/omim. Rimoin, D. L., and Phillips, J. A., III (1997). Genetic disorders of the pituitary gland. In ‘‘Principles and Practice of Medical Genetics’’ (D. L. Rimoin, J. M. Connor, and R. E. Pyeritz, eds.), 3rd ed., pp. 1331–1364. Churchill Livingstone, New York. Ross, J., Long, L., Loriaux, D., and Cutler, G. (1985). Growth hormone secretory dynamics in Turner syndrome. J. Pediatr. 106, 202. Salvatori, R., Gondo, R. G., de Aguirar Oliveira, M. H., Phillips, J. A., III, Souza, A. H., Hayashida, C. Y., Toledo, S. P., Conceicao, M. M., Prince, M., Baumann, G., and Levine, M. A. (1999). Familial isolated growth hormone deficiency due to a novel mutation in the growth hormone-releasing hormone receptor. J. Clin. Endocrinol. Metab. 84, 917–923. Shohat, M., Hermon, V., Melmed, S., et al. (1991). Deletion of 20p11.23–pter with growth hormone neurosecretory disorder but normal growth hormone releasing hormone genes. Am. J. Med. Genet. 39, 56–63. Takahashi, Y., Kaji, H., Okimura, Y., Goji, K., Abe, H., and Chihara, K. (1996). Brief report: Short stature caused by a mutant growth hormone. N. Engl. J. Med. 334, 432–436. Vnencak-Jones, C. L., Phillips, J. A., III, Chen, E. Y., and Seeburg, P. H. (1988). Molecular basis of human growth hormone gene deletions. Proc. Natl. Acad. Sci. USA 85, 5615–5619. Wajnrajch, M. P., Gertner, J. M., Mullis, P. E., Deladoey, J., Cogan, J. D., Lekhakula, , Kim, S., Dannies, P. S., Saenger, P., Moshang, T., Phillips, J. A., III, and Leibel, R. L. (2000). Arg183His, a new mutational ‘‘hot-spot’’ in the growth hormone gene causing isolated GH deficiency type II. J. Endocr. Genet. 1, 125–135. Wu, W., Cogan, J. D., Pfaffle, R. W., Dasen, J. S., Frisch, H., O’Connell, S. M., Flynn, S. E., Brown, M. R., Mullis, P. E., Parks, J. S., Phillips, J. A., III, and Rosenfeld, M. G. (1998). Mutations in PROP1 cause familial combined pituitary hormone deficiency. Nat. Genet. 18, 147–149.

384

GH ISOFORMS p0025

Several molecular forms of hGH exist. The most abundant form (except during pregnancy) is pituitary GH-1, also known as GH-N or 22-kDa GH. GH-2 (also known as GH-V or placental GH) differs from pituitary GH-1 at 13 of the 191 amino acid residues. In addition, GH-2 contains an N-linked glycosylation site, and indeed, GH-2 exists as both a glycosylated and a nonglycosylated form. During pregnancy, placental GH progressively supplants pituitary GH in the maternal circulation. The GH-1 gene also generates an internally deleted GH (lacking 15 internal amino acids and named 20-kDa GH) through alternative mRNA splicing. A secondary splice site within exon 3 is used for that purpose. This 20-kDa variant accounts for 5 to 10% of GH production. A second postulated splice variant arises from complete exon 3 skipping, owing to the relatively weak splice donor site in intron 3. Although the mRNA for this so-called 17.5-kDa GH has been demonstrated, it is not entirely clear whether the corresponding protein is produced in significant amounts in normal individuals. Additional GH molecular variants arise from posttranslational processing. These include glycosylated (especially placental GH), acetylated, deamidated, and (in some animals) phosphorylated GHs. Further molecular heterogeneity results from oligomeric GH forms (up to at least pentameric GH), with both noncovalently associated oligomers and disulfide-linked oligomers. The biological roles of the GH variants are largely unknown. They vary in biological activity, with 22-kDa GH-1 and GH-2 being the most biologically active.

REGULATION OF GH PRODUCTION AND SECRETION Expression of the GH-1 gene in pituitary somatotropes is positively regulated by GH-releasing hormone (GHRH), glucocorticoids, and (in rodents) thyroid hormones. It is negatively regulated by insulin-like growth factor-1 (IGF-1) and somatostatin. GHRH is also very important for normal somatotrope cell proliferation during pituitary development. GH is stored in secretory granules and released in response to GHRH; its release is inhibited by somatostatin. Ghrelin or related GH secretagogues, in pharmacological amounts, are potent releasers of GH, but the physiological role of ghrelin in GH regulation appears to be minor at best. The various GH isoforms appear to be cosecreted, and no isoform-specific stimulus has been identified.

Growth Hormone (GH)

GH secretion is under complex hypothalamic control by at least two hypophysiotropic hormones: GHRH (stimulating) and somatostatin (inhibitory). Secretion is pulsatile, with a marked ultradian rhythm of pulses of widely varying amplitude occurring every 1 to 2 h. The highest secretory pulses are linked to slow wave sleep and typically occur during the first 2 h after sleep onset. However, major pulses can also occur at other times due to stimuli such as stress, exercise, pain, and other acute events. Feedback inhibition of GH secretion is provided by IGF-1 as well as by GH itself. Daily GH production is high and largely unregulated during fetal and neonatal life, falls after birth under the regulatory influence of IGF-1, is again high during puberty, and falls progressively thereafter by about 15% per decade. Physiological and metabolic regulators of GH secretion are nutrition, estrogen (stimulatory), glucocorticoids (the inhibitory effect at the hypothalamic level predominates over the stimulatory effect at the pituitary level), and thyroid hormone (stimulatory, especially in rodents). In humans, undernutrition stimulates GH production, whereas overnutrition dampens GH production (the opposite is generally true in rodents). There is marked sexual dimorphism in the GH secretory rhythm, with women showing generally higher values, higher basal secretion, and more ‘‘noisy’’ circadian profiles. This difference can be largely attributed to an estrogen effect. In rats, the male secretion pattern is characterized by very high pulses interrupted by nearly complete quiescence, whereas the female secretion pattern is characterized by lower pulses but higher tonic secretion between pulses. In contrast to this complex regulation of GH production in the pituitary gland, the expression and secretion of placental GH during pregnancy appear to be constitutive and largely unregulated.

BLOOD TRANSPORT OF GH GH circulates in blood partly bound to two GHbinding proteins (GHBPs). The principal carrier is the high-affinity GHBP that corresponds to the soluble ectodomain of the GH receptor (GHR). In humans and many other species, the GHBP is produced by proteolytic cleavage of the GHR in its juxtamembranous stem region by a metalloproteinase(s) of the ADAM family (TACE ¼ TNF-a-converting enzyme and perhaps other ADAMs), with shedding of the ectodomain from the cell as the GHBP. In rodents, the GHBP is generated from the GHR gene as an alternative mRNA splice product. Under basal conditions, approximately half of circulating

p0045

385

Growth Hormone (GH)

GH is bound to this GHBP. A minority of GH (5%) is bound to another, low-affinity GHBP that appears to correspond to a modified form of a2-macroglobulin. The GHBPs prolong the plasma half-life of GH and act to provide a circulating GH reservoir. In addition, the high-affinity GHBP is a modulator of GH action at the cellular level. The plasma half-life of GH in humans is 15 to 20 min and is eliminated primarily through glomerular filtration and GHRmediated clearance.

THE GHR AND GH SIGNAL TRANSDUCTION p0050

p0055

p0060

The GHR is a 620-amino acid single-chain glycoprotein that belongs to the family of cytokine receptors. It has a large, 246-residue extracellular domain, a single-transmembrane domain, and a 350-amino acid cytoplasmic domain. The ectodomain is folded into two subdomains; the amino-terminal subdomain 1 contains the GH-binding site, and the carboxyterminal subdomain 2 is involved in receptor dimerization. A short (10 residue) linear stem region between subdomain 2 and the transmembrane domain serves as the substrate for cleavage by TACE to yield the GHBP. The cytoplasmic domain contains several features involved in hormone signaling. Among those, a membrane-proximal proline-rich region (Box 1) is most important; it serves as a docking site for Janus kinase 2 ( Jak 2). Another important region in the cytoplasmic portion is the internalization domain. The single GHR gene resides on the short arm of human chromosome 5 (5p13–p12), spans more than 156 kb, and consists of at least 10 exons, of which exons 2 to 10 encode the protein. The 22-residue portion encoded by exon 3 is not necessary for function, and inclusion or exclusion of exon 3 occurs as a form of GHR polymorphism. Several alternative exons 1 can be used to express the GHR in a tissue- or metabolic state-specific manner. They are only incompletely characterized. Two truncated splice variants of the GHR are known; they lack most of the cytoplasmic domain, represent a small part (1–5%) of the total GHR complement, and have an unknown function. The GHR is expressed ubiquitously, although the level of expression varies among tissues. The liver is the tissue with the highest GHR content. GH binding to the GHR results in GHR dimerization as two GHRs associate sequentially with the two binding sites on GH. (Some evidence also suggests that the GHR is predimerized and then conformationally changed by GH binding.) GHR

dimerization is followed by initiation of a signal transduction cascade by recruitment of Jak 2 to the GHR cytoplasmic domain and phosphorylation of both Jak 2 and the GHR. Following this initial step, several downstream signal transduction pathways are activated. They include, most prominently, the Jak–Stat (especially Stat 5b) pathway, but also the mitogenactivated protein (MAP) kinase pathway, the insulin receptor substrate (IRS)/phosphatidyl inositide 3kinase (PI3K) pathway, the phospholipase C (PLC)/ protein kinase C (PKC) pathway, and probably other signaling pathways. A detailed discussion of these signaling cascades is beyond the scope of this article. The various pathways may subserve different components of the GH bioactivity spectrum. For example, the growth-promoting/IGF-1-generating action is principally mediated by signaling through the Jak–Stat pathway, whereas some metabolic activities are mediated by signaling through PI3K. A whole host of genes are activated in response to GH signaling. Among those best recognized are serine protease inhibitor 2.1 (Spi 2.1), c-Fos, and IGF-1. GH signaling also has nongenomic effects in the cell, such as enhanced glucose transport. GH signaling is negatively regulated by several mechanisms to prevent runaway cellular stimulation, with some acting in a classic feedback loop. Among the factors inhibiting GH signaling once it has started are induction of SOCS (suppressors of cytokine signaling) proteins and CIS (cytokine-inducible SH2 protein), activation of phosphatases, GHR down-regulation through GHR internalization, and perhaps GHR inactivation by proteolytic GHR decapitation/GHBP shedding. The GHBP, in addition to its role in the circulation, acts as a local modulator of GH action by competing with GHRs for ligand and probably by forming biologically unproductive GHR / GHBP dimers at the cell surface. Thus, at the local tissue level, GHBP acts primarily as an inhibitor of GH action.

BIOLOGICAL ACTIONS OF GH Many, but not all, of the activities of GH are mediated by its second messenger, IGF-1. IGF-1 is generated in response to GH in liver and many other target tissues. It is primarily responsible for the growth-promoting action, serving as both a mitogen and a metabolically active hormone. For some GH actions, it is still not clear whether they are mediated by IGF-1, by GH directly, or by both. Table I lists the principal bioactivities of GH. The protean and diverse manifestations of GH action are evident from the table. All

p0070

386

t0005

Growth Hormone (GH)

Table I Biological Activities of GH A. Direct GH effects Lipolysis Prechondrocyte differentiation Preadipocyte differentiation Amino acid transport in muscle Glucose transport (acute hypoglycemic effect, so-called ‘‘insulin-like effect’’) IGF-1 generation, IGFBP3 and ALS generation Hypothalamic somatostatin secretion (short-loop feed back) Milk production (human GH only) B. IGF-1 mediated effects

retain IGFs in the circulation, thereby prolonging their half-life ( 20 h) and modulating their access to tissues and in vivo bioactivity. The other IGFBPs bind smaller amounts of IGFs and are of lesser importance as IGF transport proteins. They play a role in local regulation of IGF action in tissues. IGFBP1 is inversely regulated by insulin (i.e., insulin down-regulates IGFBP1) and acts as a circulating inhibitor of IGF action by complexing IGFs. IGF-1 acts through its own receptor (the type 1 IGF receptor), which is structurally related to the insulin receptor and expressed widely.

Sulphate and thymidine incorporation into growth cartilage Clonal expansion of chondrocytes Linear bone growth DNA and RNA synthesis Renal Na and phosphate retention C. Unknown or combined GH / IGF-1 effects Nitrogen retention a

Somatic growth

Insulin antagonism Beta cell hyperplasia Erythropoiesis Immune system stimulation a

Somatic growth involves not only bone elongation, but also concomitant muscle and organ growth. Somatic growth is possible with IGF-1 alone, but differs both quantitatively and qualitatively from that induced by GH.

tissues and organs are targets for GH action; hence, the term ‘‘somatotropin’’ describes this hormone more aptly than does the term ‘‘growth hormone.’’ The most prominent net effects of GH actions in vivo are somatic growth, loss or maintenance of fat mass, muscle anabolism, increase or maintenance of bone mineral density, and insulin antagonism.

IGF-1 AND IGF-BINDING PROTEINS p0080

IGF-1 and the related IGF-2 are 70- and 67-amino acid, proinsulin-like polypeptides of about 7.5 kDa, respectively, that are induced by GH (IGF-1 and IGF-2) and mediate many of the GH actions. IGF-1, rather than IGF-2, is the principal player in the GH– IGF axis. Circulating IGFs are bound to six IGFbinding proteins (IGFBPs). Foremost among those is IGFBP3, which complexes the great majority of IGFs in a 150-kDa ternary complex composed of IGF, IGFBP3, and another protein called acid-labile subunit (ALS). Both IGFBP3 and ALS are GH inducible and GH dependent. The ternary complex acts to

DISEASE STATES WITH ABNORMALITIES IN THE GH–IGF AXIS GH Deficiency GH deficiency can result from a variety of causes such as genetic defects, birth trauma, and organic lesions affecting the pituitary or the hypothalamus. The most common cause is a pituitary or hypothalamic tumor resulting in destruction of somatotropes or GHRH-producing neurons as well as interruption of hypothalamo–pituitary communication. In patients with pituitary tumors, GH deficiency usually develops before the other pituitary hormones are compromised. Genetic causes include inactivating mutations in the GH-1 gene, the GHRH receptor gene, or genes involved in pituitary development such as PROP-1, PIT-1, HESX1, PITX2, LHX3, and LHX4. One relatively common type of GH deficiency is called ‘‘idiopathic,’’ meaning that no cause is evident. In general, this is a diagnosis of exclusion in a child with poor growth and subnormal serum GH levels. The delineation between this entity and normal variation of growth patterns is difficult. The clinical manifestations of GH deficiency depend, in part, on whether it develops during childhood or during adult life. Childhood GH deficiency is characterized by growth retardation, a feature that does not apply to adults. Other clinical manifestations of GH deficiency include moderately delayed puberty, childhood hypoglycemia, increased adiposity (especially visceral fat), osteopenia, decreased lean body mass, low exercise tolerance and stamina, extracellular volume depletion, impaired psychosocial functioning, and decreased quality of life. In humans, immune function is not sufficiently affected to be clinically relevant. Except for growth retardation and hypoglycemia, the manifestations of GH deficiency are relatively subtle and not readily recognized unless sought out specifically. Diagnosis is suspected in the proper

p0085

Growth Hormone (GH)

clinical setting and is confirmed by the failure of serum GH to rise in response to standard pharmacological stimuli. Typical biochemical features include low levels of serum IGF-1, IGFBP3, and ALS as well as an elevated level of IGFBP2. Treatment consists of replacement therapy with recombinant GH. Deficiency of the placental GH appears to have no adverse effects on either mother or fetus, as has been learned from naturally occurring cases with a deletion of the GH-2 gene.

GH Insensitivity p0100

GH insensitivity or resistance shows a clinical picture that is similar, although not identical, to GH deficiency. There is no shortage of GH; rather, there is inability of GH to act. In its most extreme form, GH insensitivity is a genetic syndrome (Laron syndrome) caused by inactivating mutations in the GHR gene. Patients show all of the physical manifestations of GH deficiency, but serum GH levels are high, while IGF-1, IGFBP3, and ALS levels are low. In most, but not all, cases, the serum GHBP level is low or undetectable (the type of mutation in the GHR determines the presence or absence of GHBP). The phenotype is one of complete absence of GH activity and resembles that of patients with the most severe degrees of GH deficiency. Diagnosis is suspected in the proper familial setting by measuring serum GH, IGF-1, and GHBP; it is usually confirmed by genetic analysis. A unique form of GH insensitivity was reported in a single patient who carried a partial deletion of the IGF-1 gene. Treatment consists of IGF-1 replacement therapy. A mild to moderate form of acquired GH resistance is frequently seen in catabolic states. This appears to be an adaptive response, with low serum IGF-1 and elevated GH levels. The nutritional deprivation that frequently accompanies such conditions explains part, but not all, of this phenomenon. This is a reversible derangement in the GH–IGF axis that returns to normal when the underlying disease process is corrected. Treatment with GH is not recommended because GH therapy, in severe cases of illness, has been associated with increased mortality. However, one partially GH-resistant condition where GH treatment is beneficial is chronic renal failure.

GH Excess p0110

Overabundance of GH secretion leads to a condition called acromegaly in adults. When it occurs during

387 childhood, it leads to gigantism. In the latter case, overall somatic growth is accelerated; in the former case, there is only acral overgrowth in hands, feet, and facial structures. The condition is typically caused by a GH-producing pituitary adenoma. At least half of these adenomas have somatic mutations in the Gsasubunit of the signal-transducing G protein. (GHRH normally signals through this pathway.) The mutation renders the G protein constitutively active and leads to tumor formation and unchecked GH production. A germ line variant of this type of G-protein activation is seen in McCune–Albright syndrome, which has as one of its manifestations the occurrence of acromegaly. In rare cases, acromegaly can be caused by overproduction of GHRH, either by a eutopic source (e.g., a hypothalamic lesion) or (more frequently) in an ectopic site by tumors of neuroendocrine lineage (e.g., carcinoids, islet cell tumors). The high levels of systemic unregulated GHRH lead to somatotrope hyperplasia and GH overproduction. Only one convincing case of ectopic production of GH itself—by an islet cell tumor—has been reported. The clinical aspects of acromegaly include soft tissue swelling and bony overgrowth in hands and feet, prognathia and frontal bossing, dental malocclusion, a general coarsening of facial features and body habitus, general organomegaly, hypertension, carbohydrate intolerance or diabetes mellitus, and increased cardiovascular morbidity and mortality. The diagnosis is typically delayed because of the insidious onset of clinical signs. Biochemical findings include increased serum GH and IGF-1 levels. The diagnosis is supported by demonstrating a pituitary tumor and confirmed by showing that serum GH is not normally suppressible by administration of glucose (an oral or intravenous glucose tolerance test). In cases of ectopic GHRH production, a high GHRH level in peripheral blood is diagnostic. Treatment of acromegaly consists of surgical removal of the pituitary adenoma (or, rarely, the ectopic tumor). Surgical resection is frequently not curative, especially when the adenoma exceeds the confines of the sella turcica. In such cases, radiation or medical therapy can be used. Medical treatment in the form of the long-acting somatostatin analogue octreotide (or lanreotide) is often effective in reducing GH secretion toward or into the normal range. Another form of medical treatment is available in the form of the GH antagonist pegvisomant, which blocks GH action at the GHR by preventing receptor dimerization. The aim of therapy is to lower serum IGF-1 levels to the normal range. Effective treatment of acromegaly is successful in reversing the soft tissue changes and metabolic

388 derangements but is only partially effective in reversing the bony changes. Thus, early diagnosis and curative intervention are of paramount importance. Early diagnosis is linked to cure rate because the size of the pituitary tumor largely determines the chance for a surgical cure.

DIAGNOSTICS p0115

GH deficiency is diagnosed by pituitary stimulation test because a random serum GH level is largely uninformative due to the normally pulsatile nature of GH secretion. Standard and reliable pharmacological tests used for stimulation of GH secretion are: insulin hypoglycemia, GHRH–arginine infusion, and possibly GHRH–ghrelin or GHRH–hexarelin injection. Other useful but less reliable tests include arginine alone, GHRH alone, L-dopa, glucagon, and clonidine. Clonidine appears to be a more potent secretagogue in children than in adults, where it is considered a weak stimulus. The absolute GH response to these provocative stimuli (in terms of serum GH levels achieved) varies depending on the GH assay used. Typically, monoclonal immunoassays yield lower GH values than do polyclonal assays. Diagnostic guidelines have been published by the GH Research Society and by the American Association of Clinical Endocrinologists. Acromegaly is readily diagnosed in the right clinical setting by an elevated serum IGF-1 level and a high GH level. (The latter is not diagnostic by itself due to the normally pulsatile GH secretion.) The confirmatory test, both in overt acromegaly and especially after surgical treatment, is the glucose tolerance test. Serum GH should fall below 1 ng/ml after glucose. In addition, IGF-1 should be within the normal age-adjusted range. Frequently, evidence for low-grade acromegaly persists even in cases where these criteria are met as GH secretory dynamics remain disordered when examined carefully. For practical clinical purposes, an IGF-1 level in the mid-normal range is a reasonable criterion for a cure.

GH AS A THERAPEUTIC AGENT Human GH was extracted from cadaveric human pituitaries until the mid-1980s because there was no other source of GH that was biologically active in humans. Only children were recipients of this GH because supplies were limited by necessity. The occurrence of cases of Creutzfeld–Jacob disease attributed to hGH contaminated with prions and the

Growth Hormone (GH)

simultaneous arrival of recombinant hGH in unlimited quantities resulted in a universal switch to recombinant hGH. Supplies became sufficient to treat adult patients with GH deficiency as well as conditions not associated with GH deficiency (e.g., chronic renal failure, Turner syndrome). Several pharmaceutical companies manufacture 22-kDa hGH, which is highly effective despite lacking all the other GH isoforms that are normally produced by the pituitary. Indications for GH continue to expand beyond classical GH deficiency, and much has been learned about GH biology from the availability of large quantities of chemically defined, highly pure GH preparations as well as from clinical trials and more intense scrutiny of the manifestations of GH deficiency and the effects of GH replacement therapy.

Acknowledgments This research was supported in part by a grant from the National Science Foundation and a Merit Review grant from the Department of Veterans Affairs.

See Also the Following Articles Acromegaly, Clinical Features of . Acromegaly, Diagnosis of . Aging: Muscle . Gigantism: Excess of Growth Hormone . Growth and Chronic Disease . Growth and Glucocorticoids . Growth Hormone Insensitivity . Growth Hormone-Binding Proteins . Insulin-like Growth Factors . Lipoprotein(a) . Natural and Synthetic Growth Hormone Secretagogues . Pituitary Gland Anatomy and Embryology

Further Reading Anonymous. (2000). Consensus guidelines for the diagnosis and treatment of growth hormone (GH) deficiency in childhood and adolescence: summary statement of the GH Research Society. GH Research Society. J. Clin. Endocrinol. Metab. 85, 3990–3993. Baumann, G. (1991). Growth hormone heterogeneity: Genes, isohormones, variants, and binding proteins. Endocrine Rev. 12, 424–449. Baumann, G. (2001). Growth hormone binding protein 2001. J. Pediatr. Endocrinol. Metab. 14, 355–375. Baumann, G. (2002). Genetic characterization of growth hormone deficiency and resistance: Implications for treatment with recombinant growth hormone. Am. J. PharmacoGenomics 2, 93–111. Baumann, G., and Frank, S. J. (2002). Metalloproteinases and the modulation of growth hormone signalling. J. Endocrinol. 174, 361–368. Carroll, P. V., Christ, E. R., Bengtsson, B. A., Carlsson, L., Christiansen, J. S., Clemmons, D., Hintz, R., Ho, K., Laron, Z., Sizonenko, P., Sonksen, P. H., Tanaka, T., and Thorner, M. (1998). Growth hormone deficiency in adulthood and the effects of growth hormone replacement: A review—Growth Hormone

Growth Hormone (GH)

Research Society Scientific Committee. J. Clin. Endocrinol. Metab. 83, 382–395. Cunningham, B. C., Ultsch, M., De Vos, A. M., Mulkerrin, M. G., Clauser, K. R., and Wells, J. A. (1991). Dimerization of the extracellular domain of the human growth hormone receptor by a single hormone molecule. Science 254, 821–825. Giustina, A., and Veldhuis, J. D. (1998). Pathophysiology of the neuroregulation of growth hormone secretion in experimental animals and the human. Endocrine Rev. 19, 717–797.

389 Herrington, J., and Carter-Su, C. (2001). Signaling pathways activated by the growth hormone receptor. Trends Endocrinol. Metab. 12, 252–257. Kelly, P. A., Djiane, J., Postel-Vinay, M. C., and Edery, M. (1991). The prolactin/growth hormone receptor family. Endocrine Rev. 12, 235–251. Rosenfeld, R. G., Rosenbloom, A. L., and Guevara-Aguirre, J. (1994). Growth hormone (GH) insensitivity due to primary GH receptor deficiency. Endocrine Rev. 15, 369–390.

405

Growth Hormone Insensitivity

Table I Primary Classification of GH Insensitivity

Table II

Primary GH insensitivity (hereditary defects) . GH receptor defect (may be positive or negative for GHbinding protein)

Test

- Extracellular mutation - Cytolasmic mutation - Intracellular . GH signal transduction defect (distal to cytoplasmic domain of

GH receptor) . IGF-1 synthetic defect . IGF-1 gene deletion

Scoring System for the Diagnosis of GHIS Criterion

Score

Auxology

Height

2.5 ng/ml

1

Basal IGFs

IGF-1

3 bilateral cysts in a person with a positive family history

Unilateral nonvascular small kidneys

Hypertension with asymmetrical kidneys and frequently a history of recurrent urinary tract infection

Pyelographic or ultrasound scan evidence of pyelonephritic scarring, reflux nephropathy, and/or segmental hypoplasia

Pheochromocytoma

Paroxysmal hypertension with vasomotor symptoms

High urinary catecholamine metabolites, adrenal tumor

Ectopic primary reninism

Malignant tumor with hypertension and/or hypokalemia

Rare cases of lung, liver, pancreas, or ovary cancers

Renal ischemia

Renal tumors and cysts

Extrarenal tumors

576 Some large non-renin-secreting cysts and tumors may be found in cases of renin-dependent hypertension as the mechanism of renin stimulation is renal artery compression. In selected cases, with a lateralized RVR ratio, cyst drainage or tumor resection may improve or cure hypertension.

Unilateral Nonvascular Small Kidneys Pyelonephritic scarring associated with urinary tract infection and vesicoureteric reflux can cause childhood hypertension and progressive degradation of renal function. PRA is often high and has been shown to be the cause of hypertension. High renin concentrations may also occur in cases of renal hypoplasia. In cases with a very small unilateral kidney and a lateralized RVR ratio, unilateral nephrectomy may improve hypertension.

Extrarenal Tumors p0075

The high BP levels associated with pheochromocytoma are caused by high plasma catecholamine concentrations both directly, through stimulation of vascular a-adrenergic receptors, and indirectly, through renin activation mediated by the adrenergic stimulation of juxtaglomerular cells and renal vasoconstriction. Consequently, ACE inhibition may be used to control BP before surgery. As mentioned above, rare cases of primary reninism may be due to extrarenal tumors.

HYPERRENINEMIA WITH NORMAL BLOOD PRESSURE LEVELS In most conditions with normal BP levels, hyperreninemia is a homeostatic response to reduced renal perfusion pressure or plasma flow. These conditions include dehydration, hemorrhage, diuretic use or pharmacological vasodilation, cardiac failure, and reduced plasma volume due to hypoproteinemia in patients with nephrotic syndrome or kidney failure. These conditions have numerous signs and symptoms that cannot be listed here and the assessment of the renin–angiotensin system has no added diagnostic

Hyperreninemia

value. Renin levels are related to survival in patients with cardiac failure; high levels are associated with a poor prognosis. The prognostic value of atrial and brain natriuretic peptides, however, is higher than that of renin; thus, renin is rarely used to assess prognosis. Bartter’s syndrome, or inherited hypokalemic metabolic alkalosis, is an autosomal-recessive disorder characterized by salt wasting, insensitivity to the vasoconstrictive effects of angiotensin II, and consequently high renin levels with normal BP, hyperaldosteronism, and hypokalemia. Gitelman’s syndrome is a type of Bartter’s syndrome in which patients have hypomagnesemia and hypocalciuria. Advances in the field of molecular genetics have demonstrated that this disease is caused by four genetically distinct abnormalities that result from mutations in renal electrolyte transporters and channels.

See Also the Following Articles Bartter’s Syndrome . Captopril . Hypertension, Renin and . Hyporeninemic Hypoaldosteronism . Renal Vein Renin . Renin . Tissue Renin-Angiotensin-Aldosterone System

Further Reading Blaufox, M. D., Lee, H. B., Davis, B., Oberman, A., WassertheilSmoller, S. D., and Langford, H. (1992). Renin predicts diastolic blood pressure response to nonpharmacologic and pharmacologic therapy. J. Am. Med. Assoc. 267, 1221–1225. Corvol, P., Pinet, F., Plouin, P. F., Bruneval, P., and Me´nard, J. (1994). Renin-secreting tumors. Endocrinol. Metab. Clin. N. Am. 23, 255–270. Derkx, F. H. M., and Schalekamp, M. A. D. H. (1994). Renal artery stenosis and hypertension. Lancet 344, 237–239. Gaul, M. K., Linn, W. D., and Mulrow, C. D. (1989). Captoprilstimulated renin secretion in the diagnosis of renovascular hypertension. Am. J. Hypertens. 2, 335–340. Laragh, J. (2001). Laragh’s lessons in pathophysiology and clinical pearls for treating hypertension. Am. J. Hypertens. 14, 186–194. Plouin, P. F., Corvol, P., Guyene, T. T., and Me´nard, J. (1997). Clinical investigation of the renin–angiotensin–aldosterone system. In ‘‘Oxford Textbook of Clinical Nephrology’’ (A. M. Davison, S. Cameron, J. P. Grunfeld, D. N. S. Kerr, E. Ritz, and C. G. Winearls, eds.), pp. 1425–1432. Oxford University Press, Oxford, UK. Shaer, A. J. (2001). Inherited primary renal tubular hypokalemic alkalosis: A review of Gitelman and Bartter syndromes. Am. J. Med. Sci. 322, 316–332.

590

t0005

Hypertension and Diabetes

Table I Trials in Diabetic Hypertensives to Establish the Goal of Therapy Reduction Study a

UKPDS HOT

b

Blood pressure (mm Hg)

Patients

Drugs

1148

ACEI or beta blocker þ CCB, diuretic

1501

(1) CCB

CHD/CHF (percentage)

Stroke (percentage)

Mortality (percentage)

10/5 down to 144/82

21/56

44

18

4/4 down to 140/81

38

30

43

48

77

(2) ACEI (40%) (3) beta blocker (28%) (4) Diuretic (22%) c

470

ABCD

d

UKPDS

3642

Enalopril or nisoldipine

6/8 down to 132/78

Various drugs

Sistolic blood pressure from >160 (168) to 40 mmol/L]).

CLASSIFICATION OF HYPERTRIGLYCERIDEMIA Primary Hypertriglyceridemia It could be argued that during the third millennium, a classification system for TG disorders should be based on a molecular diagnosis. However, human genetics has uncovered the molecular basis of only a minority of TG disorders. Thus, the time-honored Fredrickson system of lipoprotein phenotypes is presented here because it remains an entrenched system

Hypertriglyceridemia

of diagnostic shorthand for lipidologists and clinical biochemists. Five of the six Fredrickson types contain elevated TG as an essential diagnostic feature. Familial Chylomicronemia Syndrome (MIM 238600) Familial chylomicronemia (type I hyperlipoproteinemia) is characterized by excess chylomicrons. Overnight-refrigerated plasma develops a creamy supernatant and a clear infranatant. Patients are usually diagnosed during early infancy, although occasionally some may go unnoticed until later in life, when pancreatitis or lipemia is noted. Clinical features include eruptive xanthomata, lipemia retinalis, hepatosplenomegaly, focal neurological symptoms (e.g., memory loss, inability to concentrate), and recurrent epigastric pain. Modern laboratory methods have eliminated the artifactual hyponatremia that used to be a feature of this condition. Fasting TG exceeds 1000 mg/dl (>11.3 mmol/L) and occurs together with low plasma LDL and HDL cholesterol. Causes of this syndrome include LPL deficiency, apoCII deficiency, and ill-defined heritable inhibitors of LPL. LPL deficiency is the most common molecularly defined cause of familial chylomicronemia. LPL deficiency is an autosomal recessive condition, occurring with a frequency of 1 in 1 million and resulting from two defective alleles of the LPL gene. The biochemical diagnosis of LPL deficiency is established by finding markedly reduced LPL activity in postheparin plasma. Deficiency of the LPL cofactor apoCII is an autosomal recessive condition (MIM 207750) that is even less common than LPL deficiency. ApoCII deficiency is less clinically severe than LPL deficiency. ApoCII deficiency can be identified by isoelectric focusing, DNA sequencing of the APOC2 gene, or documenting the rescue of absent ex vivo postheparin lipolytic activity by the addition of normal serum. Primary Mixed Hypertriglyceridemia (MIM 144650) Primary mixed hypertriglyceridemia (type V hyperlipoproteinemia) is characterized by elevated VLDL and chylomicrons. Some consider this condition to be a part of a chylomicronemia continuum that includes type I hyperlipoproteinemia. Type V hyperlipoproteinemia is typically an adult disease with a frequency up to 1 in 1000. Clinical features may include eruptive xanthomata, hepatosplenomegaly, lipemia retinalis, and recurrent epigastric pain with or without pancreatitis. Laboratory findings include fasting TG that exceeds 1000 mg/dl (>11.3 mmol/L),

Hypertriglyceridemia

increased total cholesterol, and low LDL and HDL cholesterol concentrations. Although heterozygosity for the mutant LPL gene has on occasion been demonstrated, disease expression is usually associated with secondary factors such as alcohol, poor diet, obesity, diabetes, and hypothyroidism. Familial Hypertriglyceridemia (MIM 145750) Familial hypertriglyceridemia (type IV hyperlipoproteinemia) is probably polygenic or possibly autosomal dominant with variable penetrance. It seems to be the most common primary hypertriglyceridemia phenotype seen in clinical practice. It certainly is a frequent cause of mild to moderate hypertriglyceridemia. The main lipoprotein abnormality is increased VLDL. The molecular basis of the phenotype is unknown in most instances. Typically, patients have moderately elevated plasma TG concentration ranging from 200 to 500 mg/dl (2.3–5.7 mmol/L), usually with low HDL. Familial hypertriglyceridemia is associated with increased CHD risk and often with obesity, insulin resistance, hyperglycemia, hypertension, and hyperuricemia.

Secondary Hypertriglyceridemia Some metabolic conditions are frequently, but not absolutely, associated with high TG. One interpretation of these associations is that patients who develop secondary hypertriglyceridemia might already have a subtle metabolic defect that is perhaps genetically determined and that creates susceptibility to clinical hypertriglyceridemia depending on the presence of a metabolic stress. Obesity is probably the most commonly associated clinical attribute in hypertriglyceridemic patients. Alcohol Ingestion Hypertriglyceridemia associated with alcohol intake is also due mainly to increased plasma VLDL, with or without chylomicronemia. Within the liver, ethanol is converted to acetate, and this has a sparing effect on fatty acid oxidation, resulting in increased hepatic TG production and enhanced VLDL secretion. In many individuals, plasma TG can remain within the normal range due to adaptive increase in lipolytic activity. However, ethanol can also impair lipolysis, leading to increased plasma TG. Renal Disease Nephrotic syndrome is characterized by an increase in apoB-containing lipoproteins, including VLDL.

627 The mechanism underlying this increase includes overproduction by the liver, which has also increased albumin synthesis to compensate for renal protein wasting. Also, with hypoalbuminemia, more FFA binds to lipoproteins, which might impair LPL. Uremia is associated with modest elevation in VLDL, reflecting impairment of lipolysis, possibly due to the toxic effect of uremic metabolites. Pregnancy Plasma TG normally rises during the third trimester of pregnancy by up to threefold due to increased hepatic secretion of VLDL and reduced LPL activity. The physiological increase in plasma TG has little clinical consequence. However, more pronounced TG increases have been reported in association with reduced or absent LPL activity, and with the apoE4/ E2 genotype, with variable pregnancy outcome. Severe hypertriglyceridemia during pregnancy due to chylomicronemia is very rare but can be complicated by pancreatitis, which carries a significant risk of mortality for both the mother and the fetus. Medications Use of specific drugs has been associated with hypertriglyceridemia, which can be profound in susceptible individuals. Medications that commonly exacerbate hypertriglyceridemia include highly active antiretroviral combination therapies, oral estrogens, isotretinoin, beta blockers, thiazides, tamoxifen, and bile acid-binding resins. If a medication is considered to be an important determinant of hypertriglyceridemia, the indications for the particular treatment should be reviewed, particularly if the hypertriglyceridemia is marked. If dose reductions, changes in route of administration, and/or substitution with another class of medication are not possible, marked TG elevation should be treated with diet and/or pharmacological agents. Other Causes of Secondary Hypertriglyceridemia Hypothyroidism is usually associated with elevated LDL, but elevated TG may also be present, perhaps due to impaired lipolysis. Paraproteinemias, such as the hypergammaglobulinemia in macroglobulinemia, myeloma, lymphoma, and lymphocytic leukemias, and autoimmune disorders, such as systemic lupus erythematosus, may cause hypertriglyceridemia, possibly through immune-mediated interference of lipolysis.

628

TRIGLYCERIDE-LOWERING THERAPIES Nonpharmacological Therapy

p0105

Patients with hypertriglyceridemia are frequently obese with insulin resistance, hypertension, and/or diabetes. Because these are also CHD risk factors, they should be identified and treated as part of a global risk factor reduction strategy. The marked sensitivity of plasma TG to energy balance means that treatment plans should include weight reduction, dietary modification, and exercise. Dietary modification should be aimed at weight loss, with decreased overall fat intake and a reduction in refined carbohydrates or so-called ‘‘high glycemic index’’ foods. In general, the severity of the hypertriglyceridemia dictates the severity of the fat restriction. For instance, in severe hyperchylomicronemia, recommendations are often restriction of fat to approximately 10 to 15% of total calories, with reductions in both saturated and unsaturated fat. In an adult, this represents 15 to 20 g/day of fat. The diet should include at least 5 g/day of polyunsaturated fat as a source of essential fatty acids, and fat-soluble vitamins must be provided. A specialized dietician can be very helpful in these circumstances. For less severe hypertriglyceridemia, restriction of saturated fat together with increased aerobic activity may lead to substantial reductions in plasma TG. The National Cholesterol Education Program (NCEP) advises that carbohydrate and protein intake should be 55 to 60% and 15 to 20%, respectively, whereas total and saturated fat should be less than 30% and less than 7%, respectively, of daily calories. Omega-3 fatty acids, such as eicosapentanoic and dicosahexanoic acid, are components of both the Mediterranean diet and fish oils. Omega-3 fatty acids can reduce hepatic secretion of TGRLP. Ingestion of 3 to 4 g/day of omega-3 fatty acids, with caloric and saturated fat restriction, may reduce plasma TG by up to 20%. However, response to omega-3 fatty acids is not uniform when used as the sole TG-lowering therapy.

Pharmacological Therapy In general, monotherapy with pharmacological agents should be attempted first, together with diet. Combination treatment may be required for refractory severe hypertriglyceridemia but should be attempted only with caution and with a plan for frequent followup and monitoring of serum creatine kinase and transaminases.

Hypertriglyceridemia

Fibric Acid Derivatives (fibrates) Fibrates, such as gemfibrozil, bezafibrate, and fenofibrate, are a mainstay of treatment of hypertriglyceridemia. Fibrates can reduce plasma TG by up to 50% and can raise plasma HDL cholesterol by up to 20%, although these percentages are variable. The mechanism of action of fibrates is complex but includes modulation of the activity of peroxisome proliferator-activated receptor-a in the liver, with reduced hepatic secretion of VLDL and increased lipolysis of plasma TG, possibly related to decreased secretion of apoCIII. Gemfibrozil reduced CHD events in the Veterans Affairs HDL Cholesterol Intervention Trial. Fibrates can sometimes increase plasma LDL cholesterol concentration, and this may require a change to another drug or the addition of a second agent. Fibrates are generally very well tolerated, with very rare reports of hepatitis and myositis. HMG–CoA Reductase Inhibitors (statins) In addition to lowering LDL, statins used at higher doses can produce clinically significant TG decreases, probably by lowering hepatic secretion of apoBcontaining lipoproteins. However, statins should not be considered as firstline therapy when TG is much above 500 mg/dl (5.7 mmol/L). An advantage of statins is the preponderance of clinical trial results indicating marked reductions in CHD end points. Like fibrates, statins are well tolerated but very rarely may cause myopathy and/or hepatic toxicity, often with sentinel elevations in creatine kinase and/or transaminases. Niacin (nicotinic acid) Niacin has pleiotropic, incompletely defined effects on lipoprotein metabolism that may include inhibition of hepatic VLDL secretion and stimulation of lipolysis. Niacin, when administered 3 g daily in divided doses, can lower plasma TG by up to 45%, raise plasma HDL cholesterol by up to 25%, and reduce plasma LDL cholesterol by up to 20%. Older clinical trials suggested a reduction in CHD events related to niacin treatment. However, niacin can cause lightheadedness, cutaneous flushing, and/or pruritus. These adverse effects can be minimized by initiating therapy at low doses and then gradually increasing the daily dose, along with the concomitant use of aspirin. Less common adverse effects include elevation of liver enzymes, gastrointestinal distress, worsening of glucose tolerance, and elevation of uric acid. Longer acting preparations of niacin may reduce the frequency and intensity of these adverse effects, but the clinical impression is that their TG-lowering effect may be less than that seen with crystalline niacin.

p0115

629

Hypertriglyceridemia

p0130

TARGET TRIGLYCERIDE GOALS

Further Reading

According to the recent NCEP guidelines, the primary aim of therapy when TG is more than 150 mg/dl (>1.7 mmol/L) is to attain the LDL cholesterol target for the CHD risk stratum. When TG is borderline high, emphasis should be placed on weight loss and exercise. For patients with moderately high TG, weight reduction, increased physical activity, and drug therapy may be considered in high-risk persons to achieve the NCEP targets. For patients with very high TG, the initial aim of therapy is prevention of acute pancreatitis through TG lowering by diet, medication, weight reduction, control of secondary metabolic factors, and increased physical activity. Once TG has been lowered to less than 500 mg/dl (5.7 mmol/L), attention may be directed to LDL lowering. The Canadian Working Group on Hypercholesterolemia and Other Dyslipidemias guidelines suggest TG of less than 266 mg/dl (3 mmol/L) in individuals with low CHD risk and TG of less than 177 mg/dl (2 mmol/L) in individuals with moderate to very high CHD risk, including those with diabetes mellitus and/or preexisting vascular disease.

Despres, J-P., Lemieux, I., and Prud’homme, D. (2001). Treatment of obesity: Need to focus on high risk abdominally obese patients. Br. Med. J. 322, 716–720. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. (2001). Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). J. Am. Med. Assoc. 285, 2486–2497. Fodor, J. G., Frohlich, J. J., Genest, J. J., Jr., and McPherson, P. R. (2000). Recommendations for the management and treatment of dyslipidemia: Report of the Working Group on Hypercholesterolemia and Other Dyslipidemias. Can. Med. Assoc. J. 162, 1441–1447. Fojo, S. S. (1998). The familial chylomicronemia syndrome. Endocrinol. Metab. Clinics North America 27, 551–567. Ginsberg, H. N. (2001). Hypertriglyceridemia: New insights and new approaches to pharmacologic therapy. Am. J. Cardiol. 87, 1174–1180. Ginsberg, H. N., and Goldberg, I. J. (1998). Disorders of lipoprotein metabolism. In ‘‘Harrison’s Principles of Internal Medicine’’ (A. S. Fauci, et al., eds.), 14th ed., pp. 2138–2149. McGraw– Hill, New York. Malloy, M. J., and Kane, J. P. (2001). Disorders of lipoprotein metabolism. In ‘‘Basic and Clinical Endocrinology’’ (F. S. Greenspan and D. G. Gardner, eds.), 6th ed., pp. 716–744. McGraw–Hill, New York. Miller, M. (2000). Current perspectives on the management of hypertriglyceridemia. Am. Heart J. 140, 232–240. Rubins, H. B., Robins, S. J., Collins, D., Fye, C. L., Anderson, J. W., Elam, M. B., Faas, F. H., Linares, E., Schaefer, E. J., Schectman, G., Wilt, T. J., and Wittes, J. (2000). Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol: Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group. N. Engl. J. Med. 341, 410–418.

See Also the Following Articles Abetalipoproteinemia . Anderson’s Disease (Chylomicron Retention Disease) . Familial Low Cholesterol Syndromes, Hypobetalipoproteinemia . Lipoprotein(a)

631

Hypocalcemia, Therapy

action. Therefore, calcium should be given in conjunction with magnesium for several days until normal PTH responsiveness is restored. Treatment of severe hypomagnesemia consists of giving 2 g magnesium sulfate intravenously over 5 to 10 min initially. If serum magnesium is still low, 6 g magnesium sulfate can be infused in 5% dextrose/0.5% saline over 24 h. For chronic hypomagnesemia, magnesium oxide (250–500 mg by mouth two to four times daily) should be adequate. Hypermagnesemia is a much less common abnormality than hypomagnesemia and is usually observed as result of magnesium overdose, for example, in obstetric practice when high-dose magnesium infusions are used for treatment of toxemia of pregnancy and in patients with renal failure who are receiving an antacid or enema containing magnesium. Hypermagnesemia can result in reversible hypocalcemia by mechanisms similar to those operating in hypomagnesemia. Tetany is unusual in this entity of hypocalcemia due to the concomitant hypermagnesemia. Treatment of moderate hypermagnesemia simply requires magnesium withdrawal to allow magnesium excretion by the kidneys. In cases of severe hypermagnesemia, intravenous calcium (100–200 mg over 5–10 min) will antagonize the toxic effects of magnesium in addition to raising serum calcium.

MANAGEMENT OF HYPOCALCEMIC EMERGENCY Most authors agree that hypocalcemic emergency exists when corrected serum calcium is less than 7.5 mg/dl or in the presence of severe symptoms or

signs (e.g., tetany, seizures, severe hypotension, heart failure). The mainstay treatment in hypocalcemic emergency is parenteral calcium therapy. There are several forms of calcium available for intravenous administration (Table I). Calcium chloride has a much higher content of elemental calcium than does calcium gluconate (272 and 90 mg/10 ml, respectively). Therefore, administration of calcium chloride can raise serum calcium more rapidly than can calcium gluconate. However, calcium chloride has the disadvantage of being extremely irritating to the veins and soft tissues in case of extravasation. As initial therapy, one or two ampoules, each containing 10 ml of 10% calcium chloride or 10 ml of 10% calcium gluconate, are diluted in 50 to 100 ml of 5% dextrose and given intravenously over 10 min. In a severe emergency situation (e.g., a patient with tetany), a 10-ml ampoule of 10% calcium gluconate can be infused directly over 4 min. If hypomagnesemia is suspected (e.g., an alcoholic patient, a patient having diarrhea, a patient receiving diuretics), magnesium administration (e.g., magnesium sulfate [2 g intravenously]) is appropriate pending the results of serum magnesium. Oral calcium and a rapidly acting form of vitamin D, such as calcitriol, should be started simultaneously with intravenous calcium. If the cause of hypocalcemia is likely to be longstanding or due to a permanent defect (e.g., hypoparathyroidism), continuous calcium infusion for a few days is indicated until the actions of oral calcium and vitamin D start. Different protocols exist regarding the concentration and rate of calcium drip. The recommended initial rate ranges from 0.3 to 2.0 mg elemental calcium/kg/h depending on the severity of the

Table I Parenteral Calcium Preparations Calcium preparation

Elemental calcium

Initial treatment

Remarks

Calcium gluconate (10%)

90 mg per 10-ml ampoule

1–2 ampoules in 50 to 100 ml of 5% dextrose over 5 to 10 min; one ampoule may be infused directly over 4 min in emergency situations

Less irritating to the veins than calcium chloride; therefore, more convenient for a prolonged infusion

Calcium chloride (10%)

272 mg per 10-ml ampoule

1–2 ampoules in 50 to 100 ml of 5% dextrose over 5 to 10 min

Raises serum calcium more rapidly than does calcium gluconate, but more toxic to the veins

Calcium gluceptate (10%)

90 mg per 5-ml ampoule

1–2 ampoules in 50 to 100 ml of 5% dextrose over 5 to 10 min

Useful in patients who cannot tolerate large volumes of fluid

a

Ten ampoules of calcium gluconate are added to 900 ml of 5% dextrose (i.e, 900 mg elemental calcium/L) to be infused at 50 ml/h (45 mg elemental calcium/h) and then rate titrated. If necessary, this solution can be infused as rapidly as over 4 to 6 h.

632

Hypocalcemia, Therapy

Table II Initial Management of Hypocalcemia 1. Confirm hypocalcemia by measurement of ionized calcium or total serum calcium corrected for the degree of hypoalbuminemia. 2. Measure serum magnesium, phosphorus, PTH, 25-hydroxy-vitamin D, and creatinine. 3. If corrected serum calcium is less than 7.5 mg/dl or in the presence of severe hypocalcemic symtoms, administer intravenous calcium (e.g., 1–2 ampoules of 10% calcium chloride or gluconate) in addition to oral calcium and calcitriol. 4. If serum calcium is still low and in long-term causes of hypocalcemia, start calcium drip for a few days until oral calcium and vitamin D actions start. 5. If hypocalcemia is mild with minimal symptoms and signs, if any start oral calcium and vitamin D.

case. The infusion rate is titrated according to serum calcium levels, which are monitored every 4 to 8 h to maintain serum calcium in the low-normal range. A simple protocol of calcium gluconate drip is shown in Table I. A total infusion of 15 mg/kg calcium gluconate over 4 to 6 h will raise the serum calcium by approximately 2 to 3 mg/dl (0.5–0.75 mmol/L). Recommended calcium infusion rates should not be exceeded so as to avoid cardiac dysfunction. Cardiac monitoring is advisable in all patients, particularly those receiving digitalis due to their propensity for arrhythmias if the serum calcium level is raised excessively. Bicarbonate or phosphate should not be infused along with calcium because of possible intravenous precipitation of those calcium salts. Initial management of hypocalcemia is summarized in Table II.

nephrolithiasis, and nephrocalcinosis. Among the oral calcium preparations available (Table III), calcium carbonate is the most commonly used because it offers the highest amount of elemental calcium (40%) per unit tablet weight. In addition, it is the least expensive. Calcium carbonate tablets are preferentially taken with food because their absorption is enhanced in an acid environment. Gastrointestinal side effects include bloating and constipation. If these occur, switching to calcium citrate is recommended. The latter formula is better absorbed and tolerated, but is more expensive and has less elemental calcium, compared with calcium carbonate. Treatment usually starts with daily doses of 1000 to 4000 mg elemental calcium divided into three or four doses. In hypocalcemia of renal failure, hyperphosphatemia is reduced by limitation of dietary phosphate and the use of phosphate-binding agents such as calcium carbonate and calcium acetate.

LONG-TERM TREATMENT OF HYPOCALCEMIA The most severe form of chronic hypocalcemia is hypoparathyroidism, which is due mostly to surgical removal or vascular compromise of the parathyroid glands. In hypoparathyroidism, the physiological calcium-retaining effect of PTH in renal tubules is lacking. Therefore, the goal of therapy in hypoparathyroid states is maintenance of serum calcium in the low-normal range rather than in the mid- or high-normal range so as to limit hypercalciuria,

SIDE EFFECTS OF ORAL CALCIUM The efficacy of calcium absorption from the gut declines as calcium intake increases, providing a protective mechanism against calcium toxicity. However, this adaptive mechanism can be overcome by high calcium intake (usually more than 4 g daily). The consequences of calcium toxicity are shown in Table IV. Milk-alkali syndrome is a form of calcium toxicity initially described in patients consuming large

Table III Oral Calcium Preparations Calcium salt

Elemental calcium percentage

Remarks

Carbonate

40

Inexpensive; may cause bloating and constipation

Citrate

30

Well tolerated and absorbed; expensive

Lactate

13

Poor in elemental calcium

Gluconate Glubionate Acetate

9 7 25

Poor in elemental calcium Available in palatable liquid form (252 mg/10 ml) Given as phosphate binders in renal failure

633

Hypocalcemia, Therapy

Table IV Toxicity

Complications of Calcium and Vitamin D

Renal Polyuria, polydipsia, hypercalciuria, nephrolithiasis, nephrocalcinosis Gastrointestinal Nausea, vomiting, constipation Cardiovascular Arrhythmias, short QT interval, hypertension Central nervous system Headache, altered mental status Others a

Metastatic calcification of soft tissues (e.g, kidney, lung) Laboratory abnormalities a

Hypercalcemia, hyperphosphatemia , high serum levels of a 1,25-dihydroxy-vitamin D a

Complications occurring primarily with vitamin D toxicity.

amounts of dairy calcium milk and absorbable antacids for treatment of peptic ulcer disease. Nowadays, the syndrome is more commonly seen with administration of excessive doses of calcium carbonate. The syndrome consists of nausea, vomiting, renal failure, metabolic alkalosis, and hypercalcemia.

VITAMIN D THERAPY The two precursor molecules of vitamin D are ergocalciferol (vitamin D2) and cholecalciferol (vitamin D3). In the liver, vitamin D is converted to 25-hydroxy-vitamin D, which requires further hydroxylation in the kidney to form the biologically active 1,25-dihydroxy-vitamin D (calcitriol). PTH activates the renal conversion of 25-hydroxy-vitamin D to 1,25-dihydroxy-vitamin D. Dihydrotachysterol is an active synthetic analogue that does not require further metabolism and is rarely used nowadays. Vitamin D is required for the treatment of most, if not all, cases of hypocalcemia for two reasons. First, it

Table V

improves the bioavailability of calcium present in food or calcium given as medication by providing optimal calcium absorption in the gut. Second, chronic hypocalcemic states are frequently associated with a decreased synthesis or function of one or more forms of vitamin D. For instance, in hypoparathyroidism, lack of PTH results in decreased synthesis of the active form of vitamin D, 1,25-dihydroxy-vitamin D. The latter is also decreased in severe renal insufficiency due to hyperphosphatemia and decreased functional renal mass. In end-stage liver disease, formation of 25-hydroxy-vitamin D is impaired. Selection of the vitamin D preparation depends on the specific abnormality in vitamin D synthesis and activation. Thus, in the hypoparathyroid states and renal disease, 1,25dihydroxy-vitamin D is the most rational therapy. In severe hepatic disease, either 25-hydroxy-vitamin D or 1,25-dihydroxy-vitamin D could be used. In malabsorption, relatively large doses of any form of vitamin D are appropriate orally or intramuscularly (every 6 months). Table V depicts the pharmacological properties of various vitamin D preparations. Vitamin D (either D2 or D3) is extremely lipophilic, has virtually extensive storage sites in fatty tissue, and has a long half-life of about 30 days. On the contrary, 1,25-dihydroxyvitamin D is a polar compound, not stored in large amounts of fat, and has a short half-life of approximately 6 h. The time periods required to normalize serum calcium can be shortened considerably by starting treatment with loading doses. This policy is particularly useful in severe hypocalcemia when rapid normalization of serum calcium is warranted. For instance, calcitriol can be given in a loading dose of 4 mg daily for a few days and then in a maintenance dose of 0.5 to 1.0 mg daily. Cost also plays a role in choosing vitamin D formulation. In general, cost parallels the biological activity. Thus, the precursor compounds, ergocalciferol (vitamin D2) and cholecalciferol (vitamin D3), are

Vitamin D Preparations

Preparation Ergocalciferol (vitamin D2), cholecalciferol (vitamin D3)

Approximate daily dose (mg) for treatment of hypocalcemia 1000–3000 a (40,000–120,000 U)

Onset of action (days) 10–14

Approximate time to normalize serum calcium 4–8 weeks

Time to reverse toxicity 4–12 weeks

Calcifediol (25-hydroxy D3)

50–225

7–10

2–4 weeks

2–6 weeks

Dihydrotachysterol

300–1000

4–7

1–2 weeks

3–14 days

Calcitriol (1,25-dihydroxy D3)

0.25–2.5

1–2

3–7 days

2–10 days

a

1 mg ¼ 40 U.

634 Table VI

Hypocalcemia, Therapy

Treatment of Hypercalcemia Resulting from Calcium or Vitamin D Toxicity

1. Aggressive hydration with normal saline 2. In patients who cannot tolerate fluids, hemodialysis against a low-dialysis calcium bath 3. Loop diuretics after adequate hydration 4. Corticosteroids, prednisone (30–40 mg daily orally) or hydrocortisone (50 mg every 8 h), particularly in vitamin D-induced hypercalcemia 5. Calcitonin (4–6 U/kg every 6–8 h subcutaneously) 6. In severe cases, biphosphonate infusion (e.g., pamidronate, 90 mg)

the least expensive, whereas 1,25-dihydroxy-vitamin D is the most expensive. 25-hydroxy-vitamin D3 (calcifediol) is intermediate in cost.

VITAMIN D TOXICITY Vitamin D toxicity virtually occurs more easily with the most potent agent, 1,25-dihydroxy-vitamin D, than with the inactive vitamin D forms. However, because of the short duration of action of 1,25dihydroxy-vitamin D, its toxicity is reversed within a few days (Table V). On the other hand, toxicity due to excessive administration of vitamin D2 and vitamin D3 can be protracted and last for several months (Table IV). Symptoms and signs of vitamin D overdose are related mainly to those of the prevailing hypercalcemia and hyperphosphatemia (Table IV). Treatment of calcium or vitamin D overdose is shown in Table VI.

MONITORING OF THERAPY FOR CHRONIC HYPOCALCEMIA Close monitoring of serum calcium is required initially every few days after starting oral therapy. Stable laboratory values are usually reached within 3 to 4 weeks. Subsequently, periodic measurement of serum calcium, phosphorus, and magnesium can be performed every 2 to 4 months. Urine calcium may be assessed shortly after initiation of therapy and then every 6 to 12 months to prevent the development of hypercalciuria and to keep urinary calcium at less than 350 mg/24 h. Frequently, hypoparathyroid patients whose serum calcium is maintained in the normal range have hypercalciuria due to the absence of PTH action on renal calcium reabsorption. Thiazide diuretics enhance calcium reabsorption in distal renal tubule and, therefore, limit the amount of urinary calcium. Thus, thiazides can be useful in many hypoparathyroid patients as adjunctive therapy to decrease hypercalciuria and calcium/vitamin requirements. In

advanced renal insufficiency, oral calcium, dialysate calcium, and calcitriol must be properly balanced to keep the calcium–phosphorus product within the normal range so as to avoid metastatic calcification. Acceptable targets are serum phosphorus of 4.5 mg/dl and total serum calcium of 10 mg/dl.

CONCLUSION Hypocalcemia is a common disorder characterized by decreased levels of ionized serum calcium, the only biologically active form of calcium. The serum magnesium level should be assessed in every case of hypocalcemia and should be corrected if it is abnormal. Hypocalcemic emergency is a life-threatening condition occurring when severe symptoms exist and/or corrected serum calcium is less than 7.5 mg/dl. Intravenous calcium is life-saving in hypocalcemic emergency, and in many cases calcium infusion for a few days is required until actions of oral calcium and vitamin D are in effect. Calcium carbonate is the oral preparation of choice due to its high content of elemental calcium and low cost. Calcitriol is the most potent and rapidly acting vitamin D form, but its widespread use is limited by its high cost. In hypoparathyroid states, the goal of therapy is to maintain serum calcium in the low-normal range to avoid hypercalciuria. The latter can be limited by the administration of thiazide diuretics.

See Also the Following Articles Hypercalcemia and Hypercalcemia Treatment . Hyperphosphatemia . Hypoparathyroidism . Kidney Stones . Magnesium Disorders . Vitamin D

Further Reading Bouillon, R. (2001). Vitamin D: From photosynthesis, metabolism, and action to clinical applications. In ‘‘Endocrinology’’ (L. J. DeGroot and J. L. Jameson, eds.), pp. 1009–1028. Saunders, Philadelphia.

Hypocalcemia, Therapy

Bushinsky, D. A., and Monk, R. D. (1998). Calcium. Lancet 352, 306–311. Heaney, R. P., and Weaver, C. M. (2003). Calcium and vitamin D. Endocrinol. Metab. Clin. North America 32, 181–194. Levine, M. A. (2001). Hypoparathyroidism and pseudohypoparathyroidism. In ‘‘Endocrinology’’ (L. J. DeGroot and J. L. Jameson, eds.), pp. 1133–1153. Saunders, Philadelphia. Marx, S. J. (2000). Hyperparathyroid and hypoparathyroid disorders. N. Engl. J. Med. 343, 1863–1875.

Hypocortisolism see Adrenal Insufficiency

635 Reber, P. M., and Heath, H. (1995). Hypocalcemic emergencies. Med. Clin. North America 79, 93–107. Shane, E. (1999). Hypocalcemia: Pathogenesis, differential diagnosis, and management. In ‘‘Primer on the Metabolic Bone Diseases and Disorders of Mineral Metabolism’’ (M. J. Favus, ed.), pp. 223–226. Lippincott–Raven, Philadelphia. Thomas, M. K., and Demay, M. B. (2000). Vitamin D deficiency and disorders of vitamin D metabolism. Endocrinol. Metab. North America 29, 611–627.

637

Hypoglycemia

Table I Effects of a 120-min Hyperinsulinemic Hypoglycemic (50 mg/dl) Clamp on Neuroendocrine and Autonomic Responses in Overnight Fasted Men (means  SEs) Basal

Final 30 min of clamp

163  15 165  10

225  24* ,† 328  20*

Euglycemic clamp

38  6

Hypoglycemic clamp

36  5

934  110*

35  5

Euglycemic clamp

168  15

86  15

Hypoglycemic clamp

132  25

1209  134*

78  10

59  3

71  12

253  35*

Euglycemic clamp

91

10  1

Hypoglycemic clamp

91

25  2*

21

11

,†

,†

21 21

0  1* † 21

Euglycemic clamp

00

8  1*

Hypoglycemic clamp

00

0.3  0.1

†

Euglycemic clamp

33  3

14  2*

Hypoglycemic clamp

38  2

37  4

440  70

119  24*

590  70

228  30*

Euglycemic clamp

705  128

1059  58*

Hypoglycemic clamp

661  50

1195  115*

Plasma free fatty acids (mmol/L) Euglycemic clamp ,†

Hypoglycemic clamp

†

,†

Lactate (mmol/L)

Cortisol (mg/dl) ,†

,†

Heart rate (beats/min)

Growth Hormone (ng/ml) Euglycemic clamp

Euglycemic clamp Hypoglycemic clamp

Glycerol (mmol/L)

Pancreatic polypeptide (pg/ml)

Hypoglycemic clamp

Final 30 min of clamp

Glucose infusion rate (mg/kg/min)

Epinephrine (pg/ml)

Glucagon (pg/ml) Euglycemic clamp

Basal Endogenous glucose production (mg/kg/min)

Norepinephrine (pg/ml) Euglycemic clamp Hypoglycemic clamp

Table II Effects of a 120-min Hyperinsulinemic Hypoglycemic (50 mg/dl) Clamp on Metabolic and Cardiovascular Responses in Overnight Fasted Men (means  SEs)

Hypoglycemic clamp 21 Muscle sympathetic nerve activity (bursts/min)

Euglycemic clamp

53  8*

,†

Hypoglycemic clamp Mean arterial pressure (mm Hg)

60  3

62  3

59  3

70  4*

Euglycemic clamp

28  3

32  3*

Euglycemic clamp

83  3

81  4

Hypoglycemic clamp

25  3

43  4

Hypoglycemic clamp

81  3

81  3

†

Source. Data from Davis, S. N., Fowler, S., and Costa, F. (2000). Hypoglycemic counterregulatory responses differ between men and women with type 1 diabetes. Diabetes 49, 65–72. * Values are significantly increased versus basal period. † Values are significantly different versus euglycemic clamp.

endogenous glucose production that occur primarily from the liver but also the kidney. Glucagon is a powerful and quick-acting stimulus for hepatic glucose production. The increase in hepatic glucose production is primarily through glycogenolysis, but gluconeogenesis becomes more important as hypoglycemia is prolonged. Glucagon responses are absent within 5 years’ duration of T1DM, causing an increased reliance on epinephrine for defense against hypoglycemia. Growth hormone from the pituitary gland, cortisol from the adrenal cortex, and norepinephrine spillover from the sympathetic nervous system (SNS) also increase in response to moderate hypoglycemia but have modest metabolic effects. Pancreatic polypeptide (an index of parasympathetic nervous system activity), oxytocin, and vasopressin are also released during hypoglycemia but have no discernible metabolic effects.

,†

Source. Data from Davis, S. N., Fowler, S., and Costa, F. (2000). Hypoglycemic counterregulatory responses differ between men and women with type 1 diabetes. Diabetes 49, 65–72. * Values are significantly different versus basal period. † Values are significantly different versus euglycemic clamp.

Symptoms of Hypoglycemia The symptoms of hypoglycemia can be separated into those that are neurogenic (autonomic) and those that are neuroglycopenic. Examples of neurogenic symptoms experienced when glucose levels fall below 60 mg/dl are tremulousness, palpitations, anxiety, sweating, and parasthesias. These symptoms are autonomic in origin and stem from increased sympathetic drive. For example, sweating and parathesias are cholinergically mediated via sympathetic nerve fibers, and tremor is correlated with increased circulating levels of epinephrine. Symptoms that are neuroglycopenic generally occur at glucose levels of 50 mg/dl or less, result directly from brain glucose deprivation, and include difficulty in thinking, confusion, weakness, fatigue, seizures, coma, and death. Impairment of cognitive function occurs at glucose levels of about

638 45 mg/dl and can be responsible for accidents that are sometimes fatal.

HYPOGLYCEMIA AND T1DM p0030

Persistent hyperglycemia has been determined to be the cause of long-term microvascular complications associated with T1DM. The Diabetes Control and Complications Trial, a landmark multicenter randomized clinical trial, was developed to determine the risks and benefits of tight glucose control. The results showed that tight glucose control, defined as hemoglobin A 1C (HbA1C) 7.2%, obtained by using either multiple insulin injections or continuous subcutaneous insulin infusion prevented or delayed the progression of diabetes complications as compared with conventional treatment (HbA1C 9.0%). However, patients with tight glucose control also suffered a threefold increase in the incidence of severe hypoglycemia (requiring outside assistance to recover) and coma. Ninety percent of all patients with T1DM experience symptoms of hypoglycemia. In T1DM, hypoglycemia is initially caused by insulin excess but can be potentiated by lack of food intake, physical activity, and autonomic dysfunction. Hypoglycemia-associated autonomic dysfunction is an acute failure of the autonomic response to hypoglycemia that is induced by prior episodes of hypoglycemia. This creates a situation where T1DM patients have reduced capability to defend against impending hypoglycemia. Unfortunately, hypoglycemia is the complication of diabetes most feared by diabetics. Consequently, the increased prevalence of hypoglycemia is the major roadblock preventing patients from realizing the benefits of tight glucose control.

HYPOGLYCEMIC UNAWARENESS One component of hypoglycemia-induced autonomic dysfunction is hypoglycemic unawareness. Hypoglycemic unawareness occurs when an individual has reduced or a complete loss of symptoms that indicate the presence of hypoglycemia. The pathogenesis of hypoglycemic unawareness is multifactorial but includes a shifting of the glycemic threshold. That is, symptoms occur only at progressively lower glucose levels. If this threshold occurs below a critical level, hypoglycemic symptoms might not be activated and a patient might slip into a coma with little or no warning. Importantly, hypoglycemia unawareness can be reversed by meticulous avoidance of iatrogenic hypoglycemia.

Hypoglycemia

Role of Antecedent Stress in Incidence of Hypoglycemia Many studies have demonstrated the importance of antecedent hypoglycemia in the pathogenesis of hypoglycemic-associated autonomic dysfunction. Patients with tight glucose control typically have low glycemic thresholds due to the repeated exposure to prior hypoglycemia. Prior episodes of hypoglycemia cause a shift in the glycemic threshold that result in a progressive diminution of autonomic and neuroendocrine responses to subsequent episodes of hypoglycemia. As stated previously, the autonomic and neuroendocrine responses to hypoglycemia are necessary to stimulate the liver to release more glucose and to inhibit peripheral uptake of glucose in an attempt to defend against the falling glycemia. With a reduction in these defenses, combined with the absence of glucagon, T1DM patients are left particularly vulnerable to repeated hypoglycemia. Tables III and IV contain data from healthy individuals exposed to either day 1 clamped euglycemia or hypoglycemia followed by a subsequent bout of hypoglycemia on day 2. These data clearly demonstrate the blunted day 2 neuroendocrine and ANS responses and reveal the considerably greater amounts of glucose that had to be infused during the day 2 experiments to maintain the desired level of hypoglycemia. Exercise is an important adjunct treatment of T1DM. Physical activity improves insulin sensitivity, helps in body weight maintenance, and can reduce postprandial hyperglycemia. Unfortunately, exercise is also associated with increased hypoglycemia in T1DM patients. This may be due in part to the fact that prior episodes of hypoglycemia also reduce the autonomic and neuroendocrine response to a subsequent bout of prolonged exercise. The reverse is also true, as prior prolonged exercise will blunt autonomic and neuroendocrine responses to next-day hypoglycemia. Also, one bout of exercise in the morning can even cause a blunted neuroendocrine and metabolic counterregulatory response to a second bout of exercise in the afternoon. Thus, insulin delivery (reduced) and carbohydrate intake (increased) during exercise following a hypoglycemic episode may have to be modified to prevent further hypoglycemia. The mechanism(s) responsible for the blunting effect of prior episodes of hypoglycemia is unknown at this time. Cortisol is known to blunt SNS responses to stress in both animal and human experimental models. Therefore, prior increases in cortisol during hypoglycemia or exercise could be one factor that

639

Hypoglycemia

Table III Effects of Day 1 Hypoglycemia (50 mg/dl) on Neuroendocrine and Autonomic Responses to Day 2 Hypoglycemia (50 mg/dl) in Overnight Fasted Men and Women (means  SEs)

Table IV Effects of Day 1 Hypoglycemic (50 mg/dl) on Metabolic and Cardiovascular Responses to Day 2 Hypoglycemia (50 mg/dl) in Overnight Fasted Men and Women (means  SEs)

Basal

Final 30 min of day 2 hypoglycemic clamp

Final 30 min of day 2 hypoglycemic clamp

Day 1 euglycemia

178  12

346  37*

Day 1 hypoglycemic

172  18

293  21*

Day 1 euglycemia

42  3

950  4*

Day 1 hypoglycemic

33  3

421  83*

Norepinephrine (pg/ml)

Endogenous glucose production (mg/kg/min) ,†

Epinephrine (pg/ml)

163  38 125  17

,†

1174  129* ,† 862  151*

Glucagon (pg/ml) Day 1 euglycemia Day 1 hypoglycemic

102  12 72  12

375  28*

81

171  25* 22  1*

Day 1 hypoglycemic

51

17  2*

Growth hormone (ng/ml) Day 1 euglycemia

21

46  6*

Day 1 hypoglycemic

21

32  5*

Day 1 euglycemia

30  7

44  3*

28  9

33  8*

21 1  2*

,†

Day 1 euglycemia

00

00

Day 1 hypoglycemic Glycerol (mmol/L)

00

4.3  2.2*

,†

Day 1 euglycemia

32  4

29  4

Day 1 hypoglycemic

35  2

24  2*

Day 1 euglycemia

384  100

178  32*

Day 1 hypoglycemic

492  43

199  20*

924  101 945  96

1747  14* ,† 1535  73*

Day 1 euglycemia

63  4

78  5*

Day 1 hypoglycemic

61  3

72  5*

83  3 82  2

84  4 82  3

,†

,†

Lactate (mmol/L) ,†

Day 1 euglycemia Day 1 hypoglycemic Heart rate (beats/min)

,†

Muscle sympathetic nerve activity (bursts/min) Day 1 hypoglycemic

21 21

Plasma free fatty acids (mmol/L) ,†

Cortisol (mg/dl) Day 1 euglycemia

Day 1 euglycemia Day 1 hypoglycemic Glucose infusion rate (mg/kg/min)

Pancreatic polypeptide (pg/ml) Day 1 euglycemia Day 1 hypoglycemic

Basal

,†

Source. Data from Davis, S. N., Shavers, C., Mosqueda-Garcia, R., and Costa, F. (1997). Effects of differing antecedent hypoglycemia on subsequent counterregulation in normal humans. Diabetes 46, 1328–1335. * Values are significantly increased versus basal period (P < 0.05). † Values are significantly reduced versus day 1 euglycemia (P < 0.05).

Mean arterial pressure (mm Hg) Day 1 euglycemia Day 1 hypoglycemic

Source. Data from Davis, S. N., Shavers, C., Mosqueda-Garcia, R., and Costa, F. (1997). Effects of differing antecedent hypoglycemia on subsequent counterregulation in normal humans. Diabetes 46, 1328–1335. * Values are significantly different versus basal period (P < 0.05). † Values are significantly different versus day 1 euglycemia ( P < 0.05).

causes blunted SNS responses to subsequent episodes of hypoglycemia or bouts of exercise.

See Also the Following Articles CONCLUSION A vicious cycle is generated for T1DM patients when intensive control results in iatrogenic hypoglycemia. An increased frequency of hypoglycemia leads to lower glycemic thresholds, blunted neuroendocrine and ANS responses, hypoglycemia unawareness, and (unfortunately) further episodes of hypoglycemia. Alterations in treatment that focus on preventing hypoglycemia but allow maintenance of tight control of blood glucose levels can add significantly to the quality of life for T1DM patients.

Beckwith-Wiedemann Syndrome (BWS) . Diabetes, Type 1 . Glucose Physiology, Normal . Hypoglycemic State, NonDiabetic

Further Reading Cryer, P. E. (1981). Glucose counterregulation in man. Diabetes 30, 261–264. Cryer, P. E. (2001). Hypoglycemia-associated autonomic failure in diabetes. Am. J. Physiol. Endocrinol. Metab. 281, E1115–E1121. Davis, S. N., and Cherrington, A. D. (1993). The hormonal and metabolic responses to prolonged hypoglycemia. J. Lab. Clin. Med. 121, 21–31.

640 Davis, S. N., Shavers, C., Costa, F., and Mosqueda-Garcia, R. (1996). Role of cortisol in the pathogenesis of deficient counterregulation following antecedent hypoglycemia in normal man. J. Clin. Invest. 98, 680–691. Diabetes Control and Complications Trial Research Group. (1991). Epidemiology of severe hypoglycemia in the Diabetes Control and Complications Trial. Am. J. Med. 90, 450–459. Diabetes Control and Complications Trial Research Group. (1993). The effect of intensive treatment of diabetes on the development and progression of long term complications in insulin-dependent diabetes mellitus. N. Engl. J. Med. 329, 977–986. Galassetti, P., Mann, S., Tate, D., Neil, T., Costa, F., Wasserman, D., and Davis, S. N. (2001a). Effect of antecedent prolonged exercise on subsequent counterregulatory responses to hypoglycemia. Am. J. Physiol. 280, E908–E917. Galassetti, P., Tate, D., Neill, R. A., Morrey, S., and Davis, S. N. (2002). Effect of gender on counterregulatory responses to

Hypoglycemia

euglycemic exercise in type 1 diabetes. J. Clin. Endocrinol. Metab. 87, 5144–5150. Galassetti, P., Tate, D., Neill, R. A., Morris, P. G., and Davis, S. N. (2001b). Effect of antecedent hypoglycemia on neuroendocrine responses to subsequent exercise in type 1 diabetes. Diabetes 50, A54. Hepburn, D. A., Deary, I. J., Frier, B. M., Patrick, A. W., Quinn, J. D., and Fisher, B. M. (1991). Symptoms of acute insulininduced hypoglycemia in humans with and without IDDM: Factor-analysis approach. Diab. Care 14, 949–957. McGregor, V. P., Banarer, S., and Cryer, P. E. (2001). Limited impact of vigorous exercise on defenses against hypoglycemia: Relevance to hypoglycemia-associated autonomic failure. Diabetes 40, A138. McGregor, V. P., Greiwe, J. S., Banarer, S., and Cryer, P. E. (2002). Elevated endogenous cortisol reduces autonomic neuroendocrine and symptom responses to subsequent hypoglycemia. Am. J. Physiol. Endocrinol. Metab. 282, E770–E777.

642 of cognitive function (2.5–3.0 mmol/L) are sometimes observed in everyday life conditions in some individuals.

VARIOUS FORMS OF NON-DIABETIC HYPOGLYCEMIA, DIAGNOSIS, AND TREATMENT As reviewed in details elsewhere by Lefe`bvre and Scheen, there are two principal forms of hypoglycemia: (1) exogenous hypoglycemia attributable to the administration (injection or ingestion) of a hypoglycemic compound and (2) endogenous hypoglycemia.

Exogenous Hypoglycemia Insulin is by far the most frequent cause of hypoglycemia. In non-diabetic as well as diabetic patients, insulin has been used for homicidal or suicidal purposes. Severe unexplained hypoglycemia in a nondiabetic individual should always raise the possibility of an exogenous insulin administration, either suicidal or criminal. Such cases are encountered more frequently in the medical milieu or in the families or neighborhoods of diabetic patients. Inadvertent insulin administration to a hospitalized non-diabetic patient has also been reported. In psychiatric patients, purposely induced insulin shock therapy sometimes leads to prolonged hypoglycemia and irreversible brain damage. Factitious hypoglycemia due to clandestine self-administration of insulin must always be considered in the differential diagnosis of hypoglycemia; again, this situation is encountered more frequently in the relatives of diabetic patients, in the medical or paramedical profession, or in diabetic patients themselves. Similarly, oral antidiabetic agents, mainly sulfonylureas, can be involved in the pathogenesis of hypoglycemia in non-diabetic individuals: inadvertent administration, accidental ingestion (mainly in children), suicide attempt, and clandestine ingestion (a variety of factitious hypoglycemia). Alcohol ingestion may lead to hypoglycemia if it occurs in fasting conditions given that accidental ingestion of alcohol in children can induce severe hypoglycemia. Sometimes, alcohol favors the ‘‘reactive hypoglycemia’’ following sugar ingestion. Numerous other agents or drugs may induce hypoglycemia, including salicylates, quinine, ß-receptor blocking agents, pentamidine, hypoglycins, ouabain, mebendazole, isoproterenol, disopyramide, tranylcypramine, haloperidol, clofibrate, and angiotensin-converting enzyme inhibitors.

Hypoglycemic State, Non-Diabetic

Endogenous Hypoglycemia Endogenous hypoglycemia may be organic or functional. Organic Hypoglycemia Insulinomas Insulinomas are uncommon neoplasms, most often benign, that derive from the B cells of the islets of Langerhans of the pancreas. An insulinoma should be suspected in any patient presenting with the triad described by Whipple: symptoms precipitated by fasting or exercise, proven hypoglycemia associated with symptoms, and relief of symptoms by glucose. The demonstration of endogenous plasma insulin (and C-peptide) levels inappropriate to the prevailing blood glucose level is the cornerstone of the diagnosis. Simultaneous determination of blood glucose and plasma insulin (and C-peptide) levels after an overnight fast and, mainly, during a 24- to 48-h fast is probably the best procedure to demonstrate relative hyperinsulinism. When one is convinced of the diagnosis, one attempts to localize the tumor before sending the patient to surgery. Preoperative localization procedures include tomodensitometry, conventional ultrasonography, selective arteriography, magnetic resonance imaging, effluent vein catheterization, and (the most useful) endoscopic transduodenal ultrasonography. Surgical removal of the tumor is the first and obvious choice of treatment. It must always be accompanied by intraoperative ultrasensitive pancreas echotomography, particularly because there may be more than one tumor. Medical management, mainly using diazoxide, is reserved for patients who do not accept surgery or in whom major contraindications for the operation exist. Streptozotocin, in association with fluorouracil or doxorubicin, is considered to be the most effective antitumor agent for treating the rare metastatic malignant insulinomas, possibly after surgical reduction of the tumor mass and/or removal of liver metastasis. Other therapeutic options have been reviewed by Lefe`bvre and Scheen. Extrapancreatic Neoplasms Extrapancreatic neoplasms secreting a large form of insulin-like growth factor-II, known as ‘‘big IGF-II,’’ are usually large tumors present as masses in the mediastinum or the retroperitoneal space. They often have a mesenchymal origin but can originate from the liver, gastrointestinal tract, or pancreas or can be associated with lymphomas and leukemias. In these patients, hypoglycemia coexists with low or

Hypoglycemic State, Non-Diabetic

undetectable insulin and C-peptide levels, whereas high circulating levels of big IGF-II are found. Surgery is the treatment of choice. Neonatal and Infancy Hypoglycemia Numerous inborn errors of metabolism can induce hypoglycemia in neonates and young infants. They include hereditary fructose intolerance, fructose-1,6diphosphatase deficiency, phosphoenolpyruvate carboxykinase deficiency, some cases of galactosemia, and some of the 11 varieties of glycogen storage disease. Nesidioblastosis, now called ‘‘persistent hyperinsulinemic hypoglycemia of infancy’’ (PHHI), is a rare disease leading to persistent hypoglycemia of infancy. It is basically histologically characterized by the budding off from duct epithelium of endocrine cells and by the presence of microadenomas in the pancreas. The onset of symptoms of beta cell hyperplasia may occur during the first days of life but most commonly within the first 6 months. A few cases beginning with symptoms beyond 1 year of age have been reported. The group of Saudubray in Paris reported the features of 52 neonates with hyperinsulinism. Of these, 30 had diffuse B-cell hyperfunction and 22 had focal adenomatous islet cell hyperplasia. Among the latter, the lesions were in the head of the pancreas in 9, in the isthmus in 3, in the body in 8, and in the tail in 2 neonates. Partial pancreatectomy has been successful in curing 19 of the 22 neonates in whom this procedure has been proposed. Recent studies have shown that congenital hyperinsulinism with focal or diffuse nesidioblastosis can be associated with several mutations affecting the beta cell such as the genes encoding for the sulfonylurea receptor, the glucokinase enzyme, and glutamate dehydrogenase. Other causes of hypoglycemia during infancy (more functional in nature) include erythroblastosis fetalis, infants of diabetic mothers, leucine-induced hypoglycemia, ketotic or ketogenic hypoglycemia, maple sugar urine disease, and adrenal hyporesponsiveness. Functional Hypoglycemia Alimentary Hypoglycemia Alimentary hypoglycemia can occur 1 to 2 h after a carbohydrate-rich meal in individuals who have had a gastrectomy or who, for other reasons, have rapid gastric emptying. It is believed that the rapid dumping of carbohydrates in the upper small intestine elicits an excessive insulin release mediated by both the release of intestinal gut factors (e.g., GLP-1, GIP) and a rapid rise in blood glucose. The treatment is identical to that of spontaneous reactive hypoglycemia.

643 Spontaneous Reactive Hypoglycemia Spontaneous reactive hypoglycemia is a poorly defined entity. The term is usually applied to a syndrome with the following features: (1) symptoms that resemble those seen in insulin-induced hypoglycemia (e.g., diaphoresis, tachycardia, tremulousness, headache) but that often are accompanied by other symptoms less typical of hypoglycemia (e.g., fatigue, drowsiness, feelings of incipient syncope, depersonalization, irritability, lack of motivation); (2) symptoms that may be episodic, sometimes aggravated by carbohydrate-rich meals; and (3) plasma glucose concentrations that drop to 45 mg/dl (2.5 mmol/L) or less at one or more of the half-hourly samples taken in a 5- to 6-h glucose tolerance test. Abnormal insulin secretory patterns have been reported in certain patients. This entity has had a widespread vogue, particularly in the United States, over the past 30 years but has been said to be diagnosed more rarely elsewhere in the world. The American Diabetes Association and the Endocrine Society issued a joint statement to the effect that this entity is probably overdiagnosed. Indeed, the very existence of this condition has been called into question following several studies demonstrating that 25 to 30% of apparently healthy individuals without any hypoglycemic symptoms may exhibit low plasma glucose values on being given a glucose load. Furthermore, the similarity of the symptoms to those of hyperventilation, and indeed to those of other functional syndromes, emphasizes the need to reevaluate the whole matter of so-called functional or reactive hypoglycemia. The question of cause and effect has not been settled. It would be reasonable at the current time to restrict the diagnosis of reactive hypoglycemia to individuals in whom hypoglycemic blood glucose levels are demonstrated in samples taken after the sort of meals that are said to induce their symptoms. Furthermore, and as has been discussed by Lefe`bvre, some patients have adrenergic responses after a meal or during oral glucose tolerance test (OGTT) without hypoglycemia. Such ‘‘adrenergic hormone postprandial syndrome’’ probably results from an altered glycemic threshold (a higher glucose level) for generating an adrenergic response. This results in confusion. A critical analysis of the reactive hypoglycemia syndrome can be found in the proceedings of an international symposium held in Rome in September 1986. Diet is the first treatment of alimentary and reactive hypoglycemia. Simple sugars should be omitted and replaced by complex carbohydrates. If symptoms persist, small but frequent high-protein, low-carbohydrate meals should be tried. The pharmacological

644 treatment of choice is a-glucosidase inhibitors such as acarbose and miglitol. Alcohol-Promoted Reactive Hypoglycemia Alcohol-promoted reactive hypoglycemia can occur when insulinotropic sugars (e.g., glucose, saccharose) are ingested together with alcohol (e.g., beer, gin and tonic, rum and cola, whisky and ginger ale). Such mixtures should be avoided in susceptible individuals. Other Causes of Functional Hypoglycemia Other causes of functional hypoglycemia include discontinuation of total parenteral nutrition, an endocrine deficiency state (glucocorticoid, growth hormone, or glucagon deficiency), severe liver disease, profound malnutrition, prolonged muscular exercise, the autoimmune insulin syndrome (where hypoglycemia is considered to be the consequence of inappropriate release of insulin from insulin–antibody complexes), and the rare syndrome of antibodies directed against the insulin receptor (where hypoglycemia is attributed to an insulinomimetic action of the antibody).

CONCLUSION Hypoglycemia in non-diabetic individuals is not a rare condition. It is diagnosed when the blood glucose level is lower than the lowest limit of normal, that is, lower than about 3 mmol/L (or 54 mg/dl), a value also corresponding to the threshold for symptoms in various experimental studies performed in healthy volunteers in whom mild hypoglycemia was induced using graded insulin infusion. Hypoglycemia can result from the administration (injection or ingestion) of a hypoglycemic compound (e.g., insulin, oral antidiabetic agents, alcohol, various drugs). It can also be endogenous in nature. Organic endogenous hypoglycemia can be due to an insulin-producing tumor (insulinoma) or an extrapancreatic neoplasm. In neonates and young infants, hypoglycemia results mainly from various inborn errors of metabolism or of nesidioblastosis, now known as the syndrome of persistent

Hypoglycemic State, Non-Diabetic

hyperinsulinemic hypoglycemia of infancy. Functional hypoglycemia is called alimentary if it is due to gastrectomy or a too rapid gastric emptying. It is recognized as spontaneous reactive hypoglycemia if it occurs without any identified cause. Caution must be exerted in the diagnosis of this type of hypoglycemia. However, the diagnosis of reactive hypoglycemia can be made on the basis of a careful clinical and biochemical strategy. In such a case, simple therapeutic measures can be applied and the patient’s quality of life can potentially be markedly improved.

See Also the Following Articles Glucose Physiology, Normal . Hypoglycemia cretion: Functional and Biochemical Aspects

.

Insulin Se-

Further Reading Andreani, D., Marks, V., and Lefe`bvre, P. J. (eds.) (1984) ‘‘Hypoglycemia.’’ Raven, New York. De Lonlay-Debeney, P., Poggi-Travert, F., Fournet, J. C., Sempoux, C., Dionisi, C., Brunelle, F., Touati, G., Rahier, J., Junien, C., Nihoul-Fe´ke´te´, C., Robert, J. J., and Saudubray, J-M. (1999). Clinical features of 52 neonates with hyperinsulinism. N. Engl. J. Med. 340, 1169–1175. Lefe`bvre, P. J. (1991). Hypoglycemia or non-hypoglycemia. In ‘‘Diabetes’’ (H. Rifkin, J. A. Colwell, and S. I. Taylor, eds.), pp. 757–761. Excerpta Medica, Amsterdam. Lefe`bvre, P. J., Andreani, D., and Marks, V. (1988). Statement on ‘‘post-prandial’’ or ‘‘reactive’’ hypoglycemia. Diabetologia 31, 68. Lefe`bvre, P. J., and Scheen, A. J. (2002). Hypoglycemia. In ‘‘Diabetes Mellitus’’ (D. J. Porte, R. S. Sherwin, and A. Baron, eds.), pp. 959–972. McGraw–Hill, New York. Lteif, A. N., and Schwenk, W. F. (1999). Hypoglycemia in infants and children. In ‘‘Endocrinology and Metabolism Clinics of North America: Hypoglycemic Disorders’’ (F. J. Service, ed.), pp. 619–646. W. B. Saunders, Philadelphia. Marks, V. (1987). Glycemic stability in healthy subjects: Fluctuations in blood glucose concentration during the day. In ‘‘Hypoglycemia’’ (D. Andreani, V. Marks, and P. J. Lefe`bvre, eds.), pp. 19–24. Raven, New York. Marks, V., and Rose, F. C. (1981). ‘‘Hypoglycemia,’’ 2nd ed. Blackwell Scientific, Oxford, UK. Mitrakou, A., Ryan, C., Veneman, T., Mokan, M., Jensen, T., Kiss, I., Durant, J., Cryer, P., and Gerich, J. (1991). Hierarchy of glycemic thresholds for counterregulatory hormone secretion, symptoms, and cerebral dysfunction. Am. J. Physiol. 260, E67–E74.

646 Table I Causes of Hypoparathyroidism Development abnormalities of the parathyroid gland X-linked Autosomal recessive hypoparathyroidism (GCMB) DiGeorge syndrome Kenney–Caffey syndrome Barakat syndrome Mitochondrial neuromyopathies Kearns–Sayre syndrome MELAS syndrome Mitochondrial trifunctional protein deficiency Damage to the parathyroid glands Surgical Infiltrative disorders Granulomatous diseases Wilson’s disease Metastases Hemocromatosis Radiation Autoimmune polyglandular syndrome type 1 Reduced parathyroid gland function due to altered regulation Primary Gain-of-function mutations in the calcium-sensing receptor gene PTH gene mutations (autosomal recessive/autosomal dominant) Secondary Maternal hyperparathyroidism Hypomagnesemia Impaired PTH secretion Pseudohypoparathyroidism

are disrupted. A deletion of 11 nucleotides of the rnex40 gene occurs at the translocation junction, and loss of function of this gene is responsible for at least part of the DiGeorge phenotype. Another partial transcript, called nex2.2–nex3, was also identified from this breakpoint. Both rnex40 and nex2.2–nex3 are deleted in all DGS patients with 22q11 deletions, and studies aimed to demonstrate hemizygosity and mutations in these genes in patients without deletions on 22q11 are required to prove their role in DGS. Deletion of the UDF1L gene (located on 22q11), encoding a protein involved in the degradation of ubiquinated proteins, is present in all patients with the 22q11 deletion syndrome, which includes patients with DGS, velo–cardio–facial syndrome (VCFS), and the conotruncal anomaly face syndrome (CAFS). A smaller deletion removing exons 1 to 3 of the UDF1L gene was found in one patient with hypocalcemia, cleft palate, small mouth, low-set ears and interrupted aortic arch, syndactyly of the toes, and deficiency of

Hypoparathyroidism

T lymphocytes. Patients with late-onset DGS have microdeletions in the 22q11 region and develop symptomatic hypocalcemia during childhood or adolescence with only mild phenotype. Hypoparathyroidism is also a component of the neuromyopathies caused by mitochondrial gene defects: Kearns–Sayre syndrome (KSS), MELAS syndrome, and a mitochondrial, trifunctional, protein deficiency syndrome. KSS is characterized by pigmentary retinopathy and progressive external ophthalmoplegia before 20 years of age and is often associated with heart block or cardiomyopathy. MELAS syndrome consists of childhood-onset mitochondrial encephalopathy, lactic acidosis, and stroke-like episodes. Both KSS and MELAS syndrome have been reported to occur with insulin-dependent diabetes mellitus and hypoparathyroidism. The molecular defects range from large deletions, rearrangement, and duplication of the mitochondrial genomes in many tissues (in KSS) to single-base mutations in one of the tRNA genes found only in a restricted range of cell types (in MELAS syndrome). Mitochondrial trifunctional protein deficiency is a disorder of fatty acid oxidation that is associated with pigmentary retinopathy, peripheral neuropathy, and acute fatty liver degeneration occurring in pregnant women carrying affected children. Hypoparathyroidism has been reported to occur in 50% of patients with Kenney–Caffey syndrome. It is a congenital anomaly and is associated with parathyroid agenesis, growth retardation, and medullary stenosis of tubular bones. Genetic analysis of the PTH gene did not reveal abnormalities. Both dominant and recessive modes of inheritance have been described. The occurrence of hypoparathyroidism, nerve deafness, and a steroid-resistant nephrosis leading to renal failure, which has been referred to as Barakat syndrome, has been reported in four brothers of one family. An association of hypoparathyroidism with congenital lymphedema, nephropathy, mitral valve prolapse, and brachytelephalangy has been found in two brothers from another family. Molecular genetic studies have not been performed in these two families. In most cases, the molecular defect responsible for isolated autosomal recessive hypoparathyroidism remains unknown. A homozygous deletion of the human homologue of the Drosophila glial cell missing gene b (GCMB located on chromosome 6p23) has been identified in a proband with parathyroid agenesis. Gcm2, the murine homologue of GCMB, encodes a transcription factor expressed exclusively in the parathyroid cells. Animal models have shown that mice deficient in Gcm2 had mild hypocalcemia and

647

Hypoparathyroidism

serum PTH levels similar to those of the wild type, with the tymus serving as auxiliary source for the peptide hormone.

Damage to the Parathyroid Glands The most common cause of hypoparathyroidism is surgical damage to the parathyroids as a result of total thyroidectomy for thyroid cancer, radical neck dissection for laryngeal or esophageal cancer, or repeated parathyroidectomies for multigland disease. The incidence of permanent hypoparathyroidism is approximately 1 to 4%. Permanent hypoparathyroidism can be caused by damage of the blood supply to the parathyroid glands but also by inadvertent removal of parathyroid tissue (a rare event). A permanent hypoparathyroidism is suggested by prolonged hypocalcemia, which may develop immediately or weeks to years after neck surgery. In patients at a higher risk for developing permanent hypoparathyroidism, parathyroid tissue may be autotransplanted into the brachioradialis or sternocleidomastoid muscle at the time of parathyroidectomy or may be cryopreserved for subsequent transplantation if necessary. Transient hypoparathyroidism occurs when the damage to the parathyroid glands is reversible and other mechanisms of hypocalcemia after thyroid surgery also intervene. After surgery for primary hyperparathyroidism, transient hypocalcemia can occur for the suppression of normal parathyroids by hypercalcemia. These glands recover quickly, and the serum calcium level returns to normal within days. More prolonged hypocalcemia can occur in ‘‘hungry bone syndrome’’ due to severe hyperparathyroidism, where there is an increased uptake of calcium into remineralizing bones. Hypoparathyroidism (rarely) may occur in patients with emochromatosis and iron overload, Wilson’s disease, and neoplastic or granulomatous involvement of the parathyroid glands. Hypoparathyroidism has been described in a small number of patients receiving extensive radiation to the neck and mediastinum. Autoimmune disease of the parathyroid glands can occur as an isolated condition or can be associated with other disorders in autoimmune polyendocrinopathy– candidiasis–ectodermal dystrophy (APECED) syndrome. APECED, an autosomal recessive disorder that is also known as autoimmune polyglandular syndrome type 1 (APS1), is more frequent in Finns and Iranian Jews. It is characterized by a variable combination of destructive autoimmune fenomena leading to a failure of the parathyroid glands, adrenal cortex, gonads, pancreatic beta cells, gastric parietal cells and

thyroid gland, chronic mucocutaneous candidiasis, alopecia, vitiligo, keratopathy, and dystrophy of dental enamel and nails. In most cases, candidiasis is the first clinical manifestation to appear, usually before 5 years of age, followed by hypoparathyroidism before 10 years of age and later by primary adrenal insufficiency before 15 years of age. The etiology of APECED has been clarified and is due to loss-of-function mutations of the AIRE (autoimmune regulator) gene. It is located on chromosome 21q22 and encodes a 57-kDa protein with characteristics of a transcription factor.

Reduced Parathyroid Gland Function Due to Altered Regulation Altered regulation of parathyroid gland function may be primary or secondary. Primary alterations now have an established genetic basis. Three rare genetic defects have been defined. The largest group involves the calcium-sensing receptor (CaR). The CaR, which belongs to a family of G protein-coupled receptors and cell surface receptors, regulates the secretion of PTH from the parathyroid glands and the reabsorption of calcium by the renal tubules in response to alterations in serum calcium concentration. Gainof-function mutations in the CaR gene, leading to a constitutive activated protein (i.e., a receptor with a decreased set point for extracellular calcium concentrations), cause a functional hypoparathyroid state with hypocalcemia and hypercalciuria. The disorder is inherited as an autosomal dominant trait. Sporadic cases due to de novo activating mutations of the CaR have also been described. The consequence of the activated parathyroid glands is chronic suppression of PTH secretion, whereas the activated CaR in the kidney induces hypercalciuria. Although some patients have symptoms of hypocalcemia, these tend to be more mild and intermittent than might be expected. Many affected individuals are asymptomatic and are detected only on family screening after an affected individual has been identified. Clinically, it is important to distinguish these patients from those with other forms of hypoparathyroidism because treatment with vitamin D to correct hypocalcemia could lead to marked hypercalciuria, nephrocalcinosis, and renal impairment. Isolated hypoparathyroidism has also been associated with abnormalities of the PTH gene. A heterozygous base substitution (T to C) in exon 2, leading to the substitution of arginine for cysteine in the signal peptide, has been described in a family with dominant isolated hypoparathyroidism. Functional studies have

648

p0075

shown that this mutation impairs the interaction between the nascent protein and the translocation machinery, and the cleavage of the mutant signal sequence by solubilized signal peptidase is ineffective. A homozygous mutation of the donor splice site of exon 2 and intron 2 of the PTH gene has been identified in one family with autosomal recessive isolated hypoparathyroidism. This mutation, involving a single base transition at position 1 of intron 2, results in loss of exon 2, which encodes the initiation codon and the signal peptide, and causes PTH deficiency. In another patient with autosomal recessive isolated hypoparathyroidism born to consanguineous marriage, a T-to-C mutation was found in the first nucleotide of codon 3 of the 25-amino acid signal peptide sequence, resulting in the substitution of serine with proline. Secondary causes of impaired PTH secretion include maternal hyperparathyroidism. In this condition, the high concentrations of calcium in the maternal serum cross the placenta and inhibit PTH secretion by the infant’s parathyroid glands. Hypocalcemia in the newborn usually develops by the third week of life and is often self-limited. Hypomagnesemia due to defective renal tubular reabsorption of magnesium or intestinal absorption may impair secretion of PTH and contribute to hypoparathyroidism. This defect is corrected by magnesium treatment.

Impaired PTH Action A bioinactive PTH able to cause hypoparathyroidism has not yet been described. More commonly, ineffective PTH action appears to be due to peripheral hormone resistance (pseudohypoparathyroidism).

CLINICAL MANIFESTATIONS p0090

The predominant clinical manifestations of the disease are those related to hypocalcemia (Table II). In the acute setting, neuromuscular irritability, including perioral paresthesias, tingling of the fingers and toes, and spontaneous or latent tetany with grand mal seizures and laryngeal spasm, can be evident. Chronic hypocalcemia can be asymptomatic and usually recognized by routine blood tests. Patients with chronic hypocalcemia may have coarse hair, dry skin, muscle cramps, extrapiramidal signs, cataracts, alopecia, mental retardation, and personality disorders. Abnormal dentition, enamel hypoplasia, and absence of adult teeth may suggest that hypocalcemia has

Hypoparathyroidism

Table II

Signs and Symptoms of Hypocalcemia

Physical signs Trousseau’s sign Chvostek’s sign Manifestations of increased neuromuscular excitability Muscle cramps Paresthesias of distal extremities Tetany Mental disorders Grand mal seizures Laryngeal spasm Calcium precipitation Basal ganglia calcifications Cataract

been present since childhood. On ECG, a prolonged QT interval is often observed, and in patients with severe hypocalcemia, reversible congestive heart failure may occur. Nonspecific electroencephalographic changes may be observed. The classic physical findings of hypocalcemia are also manifestations of increased neuromuscular excitability. Trousseau’s sign is elicited after inflation of a blood pressure cuff above systolic pressure for 3 to 5 min. A positive response consists of the development of typical carpal spasm, with relaxation occurring 5 to 10 s after the cuff is deflated. Chvostek’s sign is elicited by tapping over the facial nerve just anterior to the ear. The response ranges from twitching of the lip at the corner of the mouth to twitching of all the facial muscles on the stimulated side. Simple twitching at the corner of the mouth occurs in some normal persons, but more extensive muscle contraction is a reliable sign of latent tetany. Calcification of basal ganglia or more widespread calcium deposits in intracranial structures may be detected on routine radiographs, computed tomography (CT) scans, or magnetic resonance.

LABORATORY INVESTIGATIONS The biochemical hallmarks of hypoparathyroidism are low serum calcium and high serum phosphorus in the presence of normal renal function. Serum calcium levels are often approximately 6 to 7 mg/dl, whereas serum phosphorus levels are 6 to 9 mg/dl. Ionized calcium concentration is less than 1 mmol/L. Serum concentration of PTH is low or undetectable except in cases of PTH resistance, in which it is high-normal or elevated. Serum levels of 1,25-dihydroxy-vitamin D are usually low normal or low. The 24-h urinary

Hypoparathyroidism

excretion of calcium is decreased except in cases of gain-of-function mutations in the CaR, where it is elevated. Nephrogenous cyclic AMP (cAMP) excretion is low, and renal tubular reabsorption of phosphorus is increased. Phosphorus excretion and urinary cAMP increase after administration of exogenous PTH with the exception of PTH resistance.

TREATMENT The aim of treatment is to normalize extracellular ionized calcium concentration so as to abolish symptoms due to hypocalcemia and to prevent its long-term complications.

Chronic Treatment Despite wide differences around the world in treatment modalities due to various availability of vitamin D and its analogues, the general idea is to supplement calcium and vitamin D metabolites so as to achieve stable normocalcemia. Because of great variability in intestinal calcium absorption in adults, it is usually necessary to supply 1 to 2 g of elemental calcium daily. Calcium must be administered in divided doses with meals to minimize epigastric bloating and irritation. Calcium can be administered in several forms: 1 g of elemental calcium is present in 2.5 g calcium carbonate, 5 g calcium citrate, 8 g calcium lactate, or 10 g calcium gluconate. Calcium carbonate and calcium citrate are equally effective in obtaining appropriate serum calcium levels. Ergocalciferol (vitamin D2) and cholecalciferol (vitamin D3) have been widely used at doses ranging from 1.25 to 10 mg daily. Their duration of action is very long due to their wide distribution in body fat; thus, responses to changes in these doses are very slow to manifest, and when intoxication is present, it persists for a long time. Therefore, during long-term treatment, it is necessary to measure periodically (every 3–4 months) serum calcium and phosphorus concentrations. In selected cases (e.g., previous history of kidney stones), urinary calcium and creatinine clearance should also be evaluated. If mild hypercalcemia develops, a reduction in the dose of vitamin D (and/or of calcium) is sufficient to restore serum calcium levels to normal, usually within 1 month. In the case of more severe hypercalcemia, treatment should be stopped and serum calcium should be monitored weekly; hydration and glucocorticoids can occasionally be started. Nowadays, calcitriol

649 [1,25(OH)2D3] is considered the treatment of choice. It is initially administered orally at a dose of 0.5 to 1 mg daily in divided doses. The dose should be carefully titrated to match the individual’s need. In fact, during the first month of treatment, the required dose, when changed according to serum calcium levels as monitored every 3 to 4 days, may increase to as much as 1.5 to 2 mg per day. During chronic treatment, serum calcium and phosphorus should be checked every 3 to 4 months, keeping in mind that, because of the lack of PTH, renal tubular calcium reabsorption is decreased. Thus, hypercalciuria may occur with increased risk of nephrolitiasis. If hypercalcemia develops, treatment must be temporarily discontinued, with usual prompt restoration of normal serum levels. Careful monitoring of serum calcium concentration is particularly needed when other drugs are administered to the patient. In fact, many different preparations may interfere with calcium and vitamin D absorption and metabolism, and these effects are amplified by the absence of counterregulatory action exerted by PTH. Thiazide diuretics enhance renal tubular calcium reabsorption and may induce hypercalcemia, whereas loop diuretics cause hypocalcemia by increasing calcium excretion. Moreover, glucocorticoids reduce intestinal calcium absorption by antagonizing vitamin D effects, cholestyramine inhibits vitamin D absorption, and agents such as aluminum and magnesium hydroxide reduce calcium absorption by promoting its precipitation in the gastrointestinal tract. In pseudohypoparathyroidism, the risk of hypercalciuria is less present because the action of PTH on distal renal tubular calcium reabsorption is preserved. However, at the beginning of treatment, high doses are usually needed for the high prevalence of osteomalacia. During follow-up, while bone lesions are being repaired (as shown by a decline in serum alkaline phosphatase and a rise in serum calcium concentration), dose requirements progressively decrease. Occasionally, some patients may experience unsatisfactory control, with ample fluctuations of serum calcium, despite the absence of dose variations. Poor compliance should be suspected, but variations in calcium intake and intestinal absorption problems should also be considered. Particular care must be taken in the management of hypoparathyroidism during the third trimester of pregnancy and puerperium. The dose of vitamin D will usually have to be reduced because of the actions of placenta, which can synthesize 1,25-dihydroxy-vitamin D3, and of prolactin, which can stimulate 1a-hydroxylation.

650 In young patients, when hypoparathyroidism is diagnosed, the therapy regimen is similar to that described for adults. However, doses must be tailored to consider body weight. Calcium is usually initiated at the level of 30 to 50 mg/kg body weight and is gradually increased according to serum calcium levels. Also in children, the most used vitamin D preparation is calcitriol due to its effectiveness at nontoxic doses, shorter half-life, and more rapid correction of hypercalcemia when it occurs. It must be remembered that vitamin D, in all its forms, can be a difficult drug to use. Unpredictable changes in therapeutic regimens may occur even after a long period of stable treatment; therefore, continued monitoring is mandatory.

Hypoparathyroidism

long-lasting hypoparathyroidism will ultimately result. Bone remodeling after parathyroidectomy may cause hypomagnesemia, which contributes to the risk of tetany by inhibiting PTH secretion by the remaining glands and by exacerbating hypocalcemic symptoms. In the instance of magnesium deficiency, the administration of 2.4 mg/kg body weight of the element over a period of 10 min may be required. Once magnesium has been repleted, stores are usually maintained through a regular diet. An oral supplementation of 300 to 600 mg daily may be necessary in the presence of diarrhea, fistulous drainage, or other abnormalities.

Future Trends Treatment after Surgery for Hyperparathyroidism After surgery for primary hyperparathyroidism due to parathyroid adenoma, serum calcium usually falls to normal levels quickly, sometimes evolving to transient hypocalcemia. In general, no treatment is needed except for oral calcium supplementation if symptoms are present. Too aggressive supplemental calcium during the early postoperative phase may delay restoration of normal parathyroid function; this seems to be triggered by mild hypocalcemia. However, if preoperatively hyperparathyroid bone disease was present, so-called hungry bone disease usually develops. Serious hypocalcemia can be observed, requiring high doses of calcitriol (1–3 mg daily) and large calcium supplements for a long period of time. Symptomatic patients should also be treated by intravenous calcium (2 mg/kg of elemental calcium over 10 min) and, in the case of recurrence of symptoms, by a more prolonged infusion (15 mg/kg of elemental calcium over 24 h, with half of the total amount administered during the initial 6 h). In general, 15 mg/kg of elemental calcium infused during 4 to 6 h is capable of increasing serum calcium concentration by 2 to 3 mg/dl. To reduce the amount of fluids, the concentration of the solution can be increased to as much as 200 mg of elemental calcium per 100 ml without the risk of irritating the vein. Particular care must be taken if the patient is taking medications such as digitalis (calcium may increase sensitivity to adverse effects of this drug such as arrhythmias); additives, such as phosphate and bicarbonate, must be avoided. A dramatic fall in vitamin D and calcium requirements will be observed after bone disease healing. In the case of multiple parathyroid excision,

Studies are in progress to verify whether parenteral administration of the active form of PTH or oral administration of modified molecules could be used in the treatment of hypoparathyroidism. In patients with parathyroid hyperplasia or in those who underwent repeated neck explorations to identify the adenomas, parathyroid tissue can be autotransplanted into the brachioradialis or sternocleidomastoid muscle at the time of parathyroidectomy, or it can be cryopreserved for later transplantation if required.

See Also the Following Articles Autoimmune Polyglandular Syndrome . Hyperparathyroidism, Primary . Hyperphosphatemia . Hypocalcemia, Therapy . Parathyroid Glands, Pathology . Parathyroid Hormone (PTH) . Parathyroid Surgery . Pseudohypoparathyroid States

Further Reading Betterle, C., Dal Pra, C., Mantero, F., and Zanchetta, R. (2002). Autoimmune adrenal insufficiency and autoimmune polyendocrine syndromes: Autoantibodies, autoantigens, and their applicability in diagnosis and disease prediction. Endocrine Rev. 23, 327–364. Garfield, N., and Karaplis, A. (2001). Genetics and animal models of hypoparathyroidism. Trends Endocrinol. Metab. 12, 288–294. Golzman, D. G., and Cole, D. E. C. (1999). Hypoparathyroidism. In ‘‘Primer on Metabolic Bone Diseases and Disorders of Mineral Metabolism’’ (M. J. Favus, ed.), pp. 226–230. Lippincott Williams & Wilkins, Philadelphia. Gunther, T., Chen, Z., Kim, J., Priemel, M., Rueger, J. M., Amling, M., Moseley, J. M., Martin, T. J., Anderson, D. J., and Karsenty, G. (2000). Genetic ablation of parathyroid glands reveals another source of parathyroid hormone. Nature 406, 199–203. Marx, S. J. (2001). Hyperparathyroid and hypoparathyroid disorders. N. Engl. J. Med. 343, 1863–1875.

Hypoparathyroidism

O’Riordan, J. L. H. (1994). Treatment of hypoparathyroidism. In ‘‘The Parathyroids’’ ( J. P. Bilezikian, M. A. Levine, and R. Marcus, eds.), pp. 801–804. Raven, New York. Okano, K., Furukawa, Y., Morii, H., and Fujita, T. (1982). Comparative efficacy of various vitamin D metabolites in the treatment of various types of hypoparathyroidism. J. Clin. Endocrinol. Metab. 55, 238–243. Parkinson, D. B., and Thakker, R. V. (1992). A donor splice mutation in the parathyroid hormone gene is associated with autosomal recessive hypoparathyroidism. Nat. Genet. 1, 149–152.

651 Schinke, M., and Izumo, S. (2001). Deconstructing DiGeorge syndrome. Nat. Genet. 27, 238–240. Sunthornthepvarakul, T., Churesigaew, S., and Ngowngarmratana, S. (1999). A novel mutation of the signal peptide of the preproparathyroid hormone gene is associated with autosomal recessive hypoparathyroidism. J. Clin. Endocrinol. Metab. 84, 3792–3796. Thakker, R. V. (2001). The molecular genetics of hypoparathyroidism. In ‘‘The Parathyroids: Basic and Clinical Concepts’’ ( J. P. Bilezikian, R. Marcus, and M. A. Levine, eds.), pp. 779–790. Academic Press, London.

653

Hypopituitarism

t0005

Table I Clinically Relevant Hormones Secreted by the Pituitary Gland Anterior pituitary gland Luteinizing hormone and follicle-stimulating hormone (gonadotropins)

Posterior pituitary gland Vasopressin/Antidiuretic hormone

Growth hormone Thyrotropin

Oxytocin

Adrenocorticotropic hormone Prolactin

p0045

t0010

100,000 population, but this excludes those clinically insignificant lesions that have been reported at autopsy in up to 27% of individuals without an antemortem diagnosis of pituitary pathology. Primary pituitary carcinomas are extremely rare, but metastatic deposits at this site do occur, usually arising from malignancies of the lung or breast. Pituitary adenomas are divided into microadenomas (1 cm) on the basis of size. In addition, there are a variety of classifications that determine the degree of invasiveness. Figure 1 shows the magnetic resonance imaging (MRI) appearance of a pituitary macroadenoma with significant suprasellar extension. Pituitary adenomas may also be defined as functioning (70%) or nonfunctioning (30%). The former group produces systemic effects by secreting hormones into the bloodstream. The most common functioning pituitary tumors are prolactinomas, followed by those causing acromegaly and Cushing’s disease, with gonadotropinomas and thyrotropinomas occurring only rarely. Both functioning and nonfunctioning tumors may result in hypopituitarism when there is insufficient normal pituitary tissue remaining. The typical sequence of hyposecretion starts with GH deficiency, followed by gonadotropin loss before progressing to loss of thyrotropin-stimulating hormone (TSH) and then adrenocorticotropic hormone (ACTH) reserves. Mild hyperprolactinemia is often

present when the tumor itself does not secrete prolactin due to interruption of the influence of dopamine from the hypothalamus that normally inhibits prolactin release. Vasopressin deficiency in association with pituitary adenomas is extremely rare.

Pituitary Surgery/Radiotherapy Hypopituitarism may be iatrogenic, and the incidence is approximately 10 to 15% posttranssphenoidal surgery. The effects of conventional radiotherapy may take many years to manifest fully; consequently, it is vital that those patients who are at risk for developing new endocrine deficiencies as a result of irradiation be reassessed every 1 to 2 years. Approximately 50% of patients will have evidence of hypopituitarism 5 years postradiotherapy.

Parapituitary Tumors Parapituitary tumors may be intradural (e.g., craniopharyngiomas, gliomas, hamartomas, meningiomas) or extradural (e.g., chordomas, metastases, chondrosarcomas, Brown tumors), but all are less common than primary pituitary adenomas. Craniopharyngiomas account for up to 4% of all intracranial tumors, and although they are often thought to be childhood tumors, 50% of all craniopharyngiomas present in

Table II Causes of Hypopituitarism .

Pituitary tumors

.

Pituitary surgery/radiotherapy

. .

Parapituitary tumors (e.g., craniopharyngiomas, meningiomas, gliomas, leukemic infiltrates) Pitutary apoplexy

.

Inflammatory/infiltrative conditions (e.g., lymphocytic hypophysitis, sarcoidosis, hemochromatosis, histiocytosis X, amyloidosis)

.

Infection (e.g., tuberculosis, pituitary abscess)

.

Head injuries

.

Congenital malformations/empty sella syndrome (e.g., agenesis, hypoplasia, primary empty sella)

.

Genetic and idiopathic disorders (e.g., isolated ACTH deficiency, Kallman’s syndrome, PIT1 mutations)

654

Hypopituitarism

Infection Communicable diseases such as tuberculosis must remain part of the differential diagnosis of hypopituitarism, and focal abscesses of the pituitary gland may occur rarely.

Head Injuries

f0015

Figure 1 Visual perimetry showing typical bitemporal hemianopia associated with optic chiasm compression.

patients over 16 years of age. They appear to follow a more aggressive course, particularly if onset is during childhood, with increased morbidity and mortality compared with patients with other pituitary or parapituitary lesions.

Pituitary Apoplexy Pituitary apoplexy describes destruction of functioning pituitary tissue as a result of infarction and may arise due to ischemia or hemorrhage. The former usually arises in patients with preexisting macroadenomas, whereas the latter may occur as a complication of anticoagulant therapy or pregnancy (Sheehan’s syndrome). Sheehan’s syndrome is rare in the developed world. Early surgery has been advocated for this condition, but each individual case must be assessed separately taking into account the neurological deficit, visual loss, and medical comorbidity.

Inflammatory/Infiltrative Conditions Lymphocytic hypophysitis is an autoimmune condition characterized by diffuse infiltration of the gland by inflammatory cells that predominantly, but not exclusively, occurs in women toward the end of pregnancy. The condition may present very acutely over a few days or more insidiously, and histological confirmation is required to be certain of the diagnosis. It may result in isolated anterior pituitary hormone deficiencies. Infiltrative (e.g., hemochromatosis) and granulomatous conditions (e.g., sarcoidosis, histiocytosis X ) can cause hypopituitarism; therefore, a high index of suspicion is needed for these conditions.

Cranial trauma can cause disruption of the pituitary stalk, which results most commonly in diabetes insipidus; however, hyposecretion of any pituitary hormone is possible. Monitoring and frequent reassessment of these patients are essential because the deficiencies may resolve over time.

Congenital Malformations/Empty Sella Syndrome The so-called empty sella syndrome may be congenital secondary to arachnoid herniation through a diaphragmatic defect or may be acquired postsurgery, postradiotherapy, or post-pituitary infarction. Imaging in this condition reveals an enlarged empty pituitary fossa, but this finding does not by definition exclude the presence of a functioning tumor.

Genetic and Idiopathic Disorders Isolated defects of single pituitary hormones have been described but are rare and usually do not progress to involve hyposecretion of other components of the gland. Examples include isolated ACTH deficiency and gonadotropin deficiency (Kallman’s syndrome). Mutations in the PIT1 gene are an important genetic cause of hypopituitarism.

CLINICAL FEATURES The clinical presentation of hypopituitarism varies according to exactly which endocrine deficiencies are present and also depends on the age at which pituitary insufficiency developed, that is, pre- or postpuberty. The clinical picture may also be influenced by the underlying pathology and whether the presentation is acute or chronic. Patients with panhypopituitarism are classically pale due to lack of ACTH, have thin hair, display premature ageing of the skin with increased wrinkles, and lack secondary sexual characteristics. They frequently experience reduced muscle strength and increased fatigue, and they may exhibit psychological problems such as depression. This typical appearance is shown in Fig. 2. The symptoms and signs may be

p0075

655

Hypopituitarism

Figure 3 Patient with right third cranial nerve palsy secondary to infiltration of cavernous sinus by pituitary macroadenoma.

or cavernous sinuses, respectively. Figure 3 shows a patient who developed a third cranial nerve palsy from a pituitary macroadenoma extending into the cavernous sinus. More extensive tumor infiltration into surrounding tissues can result in seizures and personality changes very rarely. f0005

Figure 2 Sagittal MRI scan showing pituitary macroadenoma with extensive suprasellar extension.

Symptoms and Signs Related to Endocrine Deficiencies

divided into those due to direct mass effects of the tumor (if present) and those due to the endocrine deficiencies.

The typical presentation associated with deficiencies of each of the pituitary hormones is outlined in Table III.

DIAGNOSIS Tumor Mass Effects Expansion of the tumor may be associated with visual field loss. Classically, a bitemporal upper quadrantinopia occurs first, progressing to a bitemporal hemianopia as a result of the tumor compressing the optic chiasm; however, a variety of defects can occur depending on which parts of the visual pathways are affected. Ongoing visual recovery can occur up to 6 months postdecompression, but established optic atrophy secondary to long-standing pressure on the optic nerves is a bad prognostic sign. Headaches may occur and are thought to be the result of the enlarging tumor stretching the dura; more rarely, obstructive hydrocephalus may develop. Cerebrospinal fluid rhinorrhea or cranial nerve palsies affecting the III, IV, or VI nerve may be seen if the tumor invades the sphenoid

The diagnosis of hypopituitarism is dependent on the appropriate biochemical investigations being carried out in a patient where the clinical findings have raised the suspicion of pituitary insufficiency. The presentation in adult life is often subtle and insidious, with symptoms that overlap with both other medical conditions and the normal population; consequently, the diagnosis is frequently delayed unless the investigating physician retains a high index of suspicion for a possible underlying endocrine etiology. In any patient with suspected pituitary or hypothalamic pathology, imaging of the area, ideally with MRI scanning and formal visual field assessment, is mandatory. These investigations help to assess for the presence of damage to the optic pathways, may assist in elucidating the underlying cause, and will

f0010

656

Hypopituitarism

Table III Signs and Symptoms of Hypopituitarism Hormone deficiency

Clinical features

Growth hormone

Increased fatigue, reduced muscle strength and exercise capacity, impaired psychological well-being, reduced lean body mass, and increased cardiovascular risk

Luteinizing hormone and follicle-stimulating hormone

Women: oligomenorrhea/amenorrhea, infertility, dyspareunia, breast atrophy, and loss of libido

Thyrotrophin

Men: loss of libido, erectile dysfunction, testicular atrophy, infertility, and loss of secondary sexual characteristics Increased fatigue, weight gain, cold intolerance, dry skin and hair, constipation and muscle weakness

Adrenocorticotropic hormone

Increased fatigue, anorexia, nausea and vomiting, weight loss, generalised weakness, hypoglycemia, and hyponatremia

Prolactin Vasopressin

Failure of lactation Polydipsia, polyuria and nocturia

also help in planning ongoing management of the condition. Additional tests that may help to elucidate the likely etiology should be carried out as appropriate (e.g., serum ferritin, angiotensin-converting enzyme from serum or cerebrospinal fluid, pituitary biopsy). Specific endocrine tests can be divided into those that assess either anterior or posterior pituitary function, and tests of anterior pituitary function can be further subdivided according to whether they are static measurements at a single point in time or dynamic tests that assess pituitary reserve more formally.

Basal Anterior Pituitary Hormone Assays To gain the most information about each axis from a single blood test, the basal concentrations of both the pituitary hormone and the target hormone should be assayed: .

. . . .

Luteinizing hormone (LH), follicle-stimulating hormone (FSH), and estradiol or testosterone (9:00 am) TSH and free thyroxine ACTH and cortisol (9:00 am) Prolactin Insulin-like growth factor-1 (IGF-1)

Testosterone and cortisol should be measured at 9:00 am as a baseline because this is the time of peak secretion within the circadian rhythm. Most reference ranges have been devised for 9:00 am samples, so it is easier to interpret and compare results taken at this time.

Dynamic Tests of Anterior Pituitary Function The preceding basal pituitary assays can provide vital information, particularly when the pituitary component is appropriately elevated with a low target hormone indicating primary target organ failure. The situation is frequently more complex when either a low target hormone is found in conjunction with an inappropriately low pituitary hormone or ongoing monitoring is required for patients with known partial hypopituitarism. In these cases, dynamic testing may help to quantify pituitary reserve more accurately.

Assessing the Hypothalamic–Pituitary–Adrenal Axis A 9:00 am cortisol level of less than 3.6 mg/dl (100 nmol/L) is generally sufficient to diagnose adrenal insufficiency with confidence, whereas a level of 18.1 mg/dl (500 nmol/L) or more is indicative of adequate ACTH reserve. Patients with 9:00 am values between these limits generally warrant further investigation. It is important to be aware that oral estrogen therapy artificially raises serum cortisol levels by increasing cortisol-binding globulin, so it is recommended that any such medication be stopped for at least 6 weeks (or a transdermal preparation be substituted) before carrying out any biochemical assessment that relies on serum cortisol measurements. The main tests employed in assessing ACTH reserve are the insulin tolerance test (ITT) and the short synacthen test (SST), both originally described during the 1960s.

Hypopituitarism

Insulin Tolerance Test Cortisol and GH are counterregulatory hormones; consequently, artificially inducing hypoglycemia is a useful stimulus by which to provoke their secretion. In the ITT, 0.15 IU/kg of soluble insulin is usually administered intravenously, and if the required peak cortisol level (usually 18.1–21.0 mg/dl [500–580 nmol/ L]) is achieved during the test, then that patient is determined to have sufficient ACTH reserve to mount an adequate response to acute stress. The test is contraindicated in those with epilepsy, ischemic heart disease, or postcraniotomy, and the majority of endocrinologists would not recommend its use beyond 70 years of age. Despite the potential hazards, the complication rate of the test remains very low when conducted by experienced personnel. Short Synacthen Test The SST is useful in diagnosing adrenal insufficiency but does not help any further with the differential diagnosis unless a plasma ACTH is measured simultaneously. Patients with established primary or secondary hypoadrenalism should not reach the required peak cortisol level (usually 21.0 mg/dl [580 nmol/L]) 30 min following parenteral administration of 250 mg of synthetic ACTH. In patients who develop ACTH deficiency acutely (e.g., postsurgery), this test is not appropriate because it can take several weeks for the adrenals to involute following the loss of normal circulating ACTH levels. Given the excellent safety profile of the SST, most endocrinologists reserve the ITT for patients with borderline SST results or for cases where simultaneous assessment of GH is required. The low-dose SST, using 1 mg of synthetic ACTH, has become increasingly popular during recent years as the 250-mg dose is seen as being so supraphysiological that false-negative results may occur. This dose is currently not routinely available, and further assessment of the test is required.

p0175

Additional Tests The glucagon test has the advantage, as does the ITT, of being able to assess ACTH and GH reserves simultaneously, but the glucagon test provides a weaker stimulus for secretion of ACTH and GH than does the ITT and, consequently, is more often associated with false-negative results. Glucagon transiently increases the plasma glucose, and the test relies on the subsequent fall to create relative hypoglycemia as a stimulus for ACTH and GH release. The caveat to the tests outlined previously is that in critically ill patients with suspected adrenal insufficiency, treatment should not be delayed while

657 investigations are carried out. In this setting, random ACTH and plasma cortisol measurements can be taken and parenteral high-dose hydrocortisone therapy can be quickly instituted. The results can then be reviewed retrospectively and steroid replacement can be continued if indicated, with any further assessment delayed until the acute illness has resolved. Assessing the Pituitary–Thyroid Axis Baseline assessment of TSH and thyroxine (free or total) is usually sufficient for the diagnosis of primary and secondary hypothyroidism. Measuring free thyroxine is essential if the patient is pregnant or taking oral estrogen therapy due to the increase in thyroid hormone-binding globulin that occurs. The TSH level in secondary hypopituitarism is rarely undetectable and in fact may be low, normal, or even mildly elevated. The main important differential diagnosis is from the so-called ‘‘sick euthyroid’’ syndrome, which is associated with critical systemic illness. Measurement of thyroxine-binding globulin may be helpful in complex cases because congenital deficiencies of this protein can occur, giving a similar result to secondary hypothyroidism. The thyrotropin-releasing hormone (TRH) test (assessment of TSH response to administration of intravenous TRH) has been largely abandoned, due to its lack of diagnostic sensitivity and specificity, in favor of newer and more sensitive assays for TSH and thyroxine. Assessing Growth Hormone Reserve GH is secreted in a pulsatile fashion and is released in response to stress or exercise and at night during sleep. IGF-1 mediates many of the actions of GH, is released in response to GH secretion, and itself inhibits GH release at the level of the pituitary and the hypothalamus. Therefore, isolated GH measurements are generally meaningless, and approximately 30% of patients with GH deficiency would be missed if a normal IGF-1 were the sole criterion for diagnosis. Tests to further assess GH may be subdivided into those using physiological methods and those relying on pharmacological agents to stimulate GH release. Physiological Methods These methods include urinary GH assessment, 24-h profiles of serum GH, and measurement of GH response during sleep or exercise. These techniques are very labor intensive, and because the results show significant overlap with normal controls, they are now seldom used clinically.

658 Pharmacological Methods The ITT is still regarded as the gold standard for assessing GH reserve. The main second-line test until recently has been the glucagon test. Newer agents include arginine, growth hormone-releasing hormone (GHRH), and growth hormone-releasing peptides (GHRP) such as hexarelin. Clonidine testing has been largely discredited for diagnosing GH deficiency in adults. Although research is ongoing, there is increasing evidence that combinations of these newer tests may provide accurate results with the advantage of a more acceptable safety profile as compared with the ITT.

Assessing the Pituitary–Gonadal Axis Usually, measurement of baseline LH, FSH, and testosterone or estradiol is sufficient to confirm hypogonadotropic hypogonadism. Estradiol is best measured during the follicular phase of the menstrual cycle, and testosterone is ideally measured on a 9:00 am sample. Both sex steroids bind to sex hormone-binding globulin (SHBG); therefore, this may need to be assayed to accurately determine the amount of biologically active hormone present. Women who are menstruating regularly are not, by definition, gonadotropin deficient, although the cycles may clearly be anovulatory. Ovulation can be assessed separately by measurement of a progesterone level on day 21 of the menstrual cycle. Management of subfertility due to hypopituitarism is discussed later in this article. Dynamic tests (e.g., clomiphene test) are only indicated to distinguish hypothalamic disease from pituitary disease, but these are generally for academic interest and rarely alter clinical management.

Hypopituitarism

primary polydipsia needs to be excluded, a formal water deprivation test may be required. A variety of protocols for water deprivation tests have been described, but essentially they assess the response of urine output and osmolality to a period of fluid restriction followed by a dose of exogenous vasopressin. Plasma osmolalities and serial weights are also recorded. The tests must be carefully supervised to prevent patients from becoming dangerously dehydrated and to ensure that the fluid restriction is maintained. Newer assays that offer more accurate measurements of endogenous vasopressin, normally present at very low concentrations in the circulation, may provide an alternative means of diagnosing diabetes insipidus, but currently these are not widely used.

TREATMENT Treatment of pituitary insufficiency must involve addressing the underlying cause as well as commencing and optimizing appropriate hormone replacement therapy.

Treatment of Underlying Cause Determining the etiology of the hypopituitarism will help to guide management of the underlying condition. Options may include the use of medical treatments or surgical intervention and/or radiotherapy.

Hormone Replacement Therapy

Posterior Pituitary Assessment

Treatment regimens generally involve replacing the target hormone rather than the pituitary or hypothalamic component of the pathway. The main exceptions to this are the use of recombinant GH and the treatment of infertility caused by hypogonadotropic hypogonadism.

It is vital that ACTH reserve has been fully assessed, and that glucocorticoid replacement has been instituted if appropriate, prior to formally assessing posterior pituitary function because cortisol is essential for the kidneys to handle and excrete a water load normally and so diabetes insipidus may be concealed in patients who are steroid deficient. Paired plasma and urine osmolalities in combination with urea and electrolytes and accurate fluid balance may be sufficient to confirm a diagnosis of cranial diabetes insipidus (vasopressin deficiency); however, in less clear-cut cases, or where the differential diagnosis of nephrogenic diabetes insipidus or

GH Deficiency During the final few decades of the past century, evidence accumulated that, contrary to the accepted doctrine, GH may have an important role to play in adulthood. During the 1990s, with the availability of recombinant GH eliminating the risks associated with GH derived from human cadavers, many adult patients with severe GH deficiency are now on appropriate replacement. GH is administered as a daily subcutaneous injection using a pen device that most patients master quickly. Treatment is commenced at low dose and titrated up slowly according to clinical

Hypopituitarism

and biochemical (IGF-1) parameters. There is convincing evidence that GH replacement in GHdeficient adults is associated with significant improvements in mood, energy levels, and overall quality of life. An increase in lean body mass, with a reduction in fat mass and a resultant increase in exercise tolerance, is seen in patients receiving treatment. GH therapy is also associated with an improvement in the lipid profile and with long-term treatment results in improved bone mineral density in these patients at risk for osteoporosis. LH/FSH Deficiency The possible complications of estrogen therapy in postmenopausal women are well publicized and may lead women to refuse treatment. It is vital that each case is considered individually and that the patient is helped to make an informed decision because the severe long-term consequences of estrogen deficiency clearly outweigh the risks of such therapy, particularly in young women. Essentially any combined oral contraceptive pill, or any form of estrogen licensed for postmenopausal use (in combination with a cyclical progestagen to avoid endometrial hyperplasia and subsequent malignancy in women with intact uteri), is suitable, and the decision comes down to patient choice and physician familiarity. Many younger women prefer the combined oral contraceptive pill because it is more acceptable among their peers. Transdermal HRT preparations are also available. These have the advantage, compared with oral estrogen preparations, of not affecting cortisol-binding globulin and, consequently, not affecting the monitoring of hydrocortisone replacement. Androgen replacement in men is usually initiated with intramuscular depot testosterone injections. Trough and/or peak levels are monitored, and the therapy is optimized by adjusting the dose or dose interval. Subcutaneous testosterone implants provide a more physiological profile and avoid the sharp peaks and troughs associated with intramuscular injections, but this must be balanced against the risk of infection and scarring at the site of implantation. Transdermal testosterone patches are available but can cause local skin irritation and are less popular than their estrogen counterparts, whereas oral testosterone has a very short half-life and is rarely recommended as the therapy of choice. Restoration of fertility is usually possible in these patients by using gonadotropin or gonadotropinreleasing hormone (GnRH) to stimulate ovulation or spermatogenesis. If gonadotropin deficiency occurs prior to puberty, careful and gradual dose titration is

659 required to ensure that development occurs at an appropriate rate so that final height is achieved, secondary sexual characteristics develop normally, and there is sufficient time to cope with the usual psychological changes occurring during these years. TSH Deficiency Thyroxine replacement is the treatment of choice for secondary hypothyroidism, as it is for primary hypothyroidism, and the average adult replacement dose for a patient with total deficiency is approximately 100 to 150 mg/day. Treatment can usually be initiated at a dose of 75 to 100 mg/day, with the dose altered in 25-mg increments based on six weekly reassessments of the free thyroxine. In patients with severe hypothyroidism or coexisting ischemic heart disease, lower starting doses of 25 to 50 mg can be used. Triiodothyronine rarely is superior due to its more rapid onset of action and shorter half-life, but thyroxine is appropriate routinely. In primary hypothyroidism, TSH is a sensitive marker that can guide accurate replacement, but this is not the case in secondary hypothyroidism; the serum TSH may be low, normal, or even mildly elevated in this condition. Most endocrinologists aim for a replacement dose of thyroxine that results in asymptomatic patients and free thyroxine on the upper end of the normal reference range. It is essential that patients receive adequate steroid replacement, if required, prior to commencing with thyroxine; otherwise, a potentially dangerous hypoadrenal crisis may be precipitated. ACTH Deficiency Hydrocortisone, the generic pharmaceutical name for cortisol, is generally accepted to be the glucocorticoid of choice for replacement therapy because it directly replaces the deficient hormone and can be easily measured in the plasma or urine. Alternatives include cortisone acetate or the synthetic glucocorticoids prednisolone and dexamethasone. Cortisone acetate must first be metabolized to cortisol to achieve its glucocorticoid effect; therefore, its onset of action is delayed compared with that of hydrocortisone, which has largely replaced it as the treatment of choice. Prednisolone and dexamethasone have a longer duration of action than does hydrocortisone and can be administered twice and once daily, respectively, but the inability to monitor steroid replacement with these drugs outweighs any advantages and makes them second-choice agents. Traditional steroid replacement regimens have been reevaluated during recent years in light of increasing

p0285

660 evidence that they provide significantly supraphysiological doses and also poorly mimic endogenous cortisol secretion. The currently recommended starting replacement regimen is 10 mg of hydrocortisone on waking, 5 mg at lunchtime, and 5 mg during the early evening. Clearly, very few patients are average, and hydrocortisone replacement might need significant adjustment in individual patients depending on symptoms and the results of biochemical monitoring. Patients with secondary hypoadrenalism do not normally require mineralocorticoid supplementation because the renin–angiotensin–aldosterone pathway should be intact. Prolactin Deficiency Prolactin deficiency is rare because hypopituitarism is generally associated with mild hyperprolactinemia. When present, as in Sheehan’s syndrome, it is associated with failure of lactation, but it is otherwise clinically silent. Vasopressin Deficiency Hyposecretion of vasopressin from the posterior pituitary results in partial or complete cranial diabetes insipidus. Desmopressin, a synthetic analogue of arginine vasopressin, is the replacement drug of choice and is available in oral, intranasal, and parenteral forms. Compared with endogenous vasopressin, it has a significantly longer half-life and is generally administered between one and three times per day. Desmopressin should be started at low dose and gradually titrated up until the polyuria is controlled. Monitoring paired urine and plasma osmolality, urea and electrolytes, and fluid balance is essential, particularly during the early stages and with dose adjustments.

MONITORING AND LONG-TERM REPLACEMENT The aim of endocrine therapy in hypopituitarism is to achieve adequate replacement, ensuring that the patient is asymptomatic and able to lead a normal life, but also to avoid the potential complications caused by long-term exposure to supraphysiological doses of hormones. The adverse effects of excess glucocorticoid replacement are well known; consequently, the aim of therapy increasingly is to provide an individual with the lowest dose of hydrocortisone possible to alleviate symptoms. Use of 24-h collections for urinary-free cortisol in combination with hydrocortisone day curves (varying from 3 points up to 10 points) can

Hypopituitarism

help to guide exact replacement doses for a particular individual. Supraphysiological thyroxine replacement is also increasingly being recognized as a significant problem due to the increased risk of atrial fibrillation, osteoporosis, and possible associations with dementia, but it is less easy to define in secondary hypothyroidism than in primary hypothyroidism, where TSH monitoring provides an accurate guide. Excess androgen replacement can lead to clinically significant polycythemia, and monitoring of the hematocrit is recommended. Concerns also remain that testosterone replacement may exacerbate prostatic hypertrophy and malignancy. In view of this, although the current policy in the United Kingdom is not to carry out population screening for prostate cancer, most endocrinologists currently recommend biannual prostate-specific antigen (PSA) monitoring for men receiving androgen therapy. Overreplacement with GH can, via sodium and water retention, lead to weight gain, peripheral edema, and carpal tunnel syndrome. Arthralgia and myalgia are also well recognized side effects of treatment. GH replacement leads to enhanced insulin resistance that may unmask previously undiagnosed diabetes and contribute to increased cardiovascular risk. The concern that GH therapy would be associated with an increased risk of subsequent malignancy has not been borne out by the evidence. In patients with cranial diabetes insipidus and intact thirst responses, severe dehydration is usually avoided because adequate fluid intake is maintained. If patients complain of excessive thirst, 24-h records of fluid balance can help to guide a dose increase of desmopressin. Hyponatremia as a consequence of lack of free water excretion due to overtreatment with desmopressin is a potentially more serious complication of treatment. This can be avoided by regular biochemical monitoring when therapy is commenced and after any dose adjustment (paired urine and serum osmolality in addition to urea and electrolytes) in addition to advising patients that they should be aware of the effect of the drug wearing off (i.e., polyuria) 1 to 2 h before their next dose. In patients with partial hypopituitarism, the potential for new deficiencies must always be borne in mind, particularly if the patients have received pituitary radiotherapy given that the lag time between irradiation and developing clinically significant endocrine deficiencies may be many years. In these patients, ongoing monitoring is essential. Patients now take on average lower replacement doses of steroid than ever before, leaving them

661

Hypopituitarism

potentially more vulnerable to hypoadrenalism when intercurrent illness occurs. Consequently, these patients must receive clear verbal and written advice on when and how to adjust their glucocorticoid replacement. It is also vital that patients’ general practitioners and family members be made aware of the diagnosis, and most patients are advised to carry some form of identification, such as a medic alert bracelet, to assist medical staff if the patients lose consciousness. Mild intercurrent illnesses that are not associated with significant diarrhea, vomiting, or a fever generally do not warrant an alteration in steroid replacement. Patients should be advised to double their standard regimen in the event of a pyrexial illness and to seek medical assistance early if profuse vomiting or diarrhea is preventing the administration or absorption of oral hydrocortisone. Many patients have an emergency steroid pack consisting of a single vial of 100 mg of hydrocortisone that can be administered intramuscularly by a trained relative or visiting general practitioner. Severe illness or surgery usually necessitates hospital admission for regular parenteral hydrocortisone (50–100 mg intramuscularly every 6 h). Hormone replacement regimens can be complex and require patients to actively participate in their treatment to achieve the best quality of life possible. Patient education is fundamental to achieving this aim, and charities such as the Pituitary Foundation (United Kingdom) and the Pituitary Tumor Network Association (United States), which offer advice and support to patients and relatives, are invaluable.

THE FUTURE There is good evidence that patients with acromegaly and Cushing’s disease are at increased risk for premature death, but there is increasing evidence to support the hypothesis that other patients with hypopituitarism have a mortality exceeding that of the background population. Studying this population is clearly

complex because it includes patients with a variety of endocrine deficiencies due to differing underlying pathologies who have received a selection of treatment modalities, but it appears that the premature mortality might not be entirely attributable to GH and other endocrine deficiencies but rather might reflect increased mortality from respiratory and vascular disease. There is significant evidence that replacement therapy with dehydroepiandrosterone is beneficial in Addison’s disease, but replacement therapy is also postulated to be of benefit in improving well-being in secondary hypoadrenalism. Further research is required to clarify these and other contentious issues.

See Also the Following Articles Hypopituitarism, Hormonal Therapy for . Hypothalamic Disease . Hypothalamus–Pituitary–Thyroid Axis . Pituitary Gland Anatomy and Embryology . Pituitary Tumors, Molecular Pathogenesis . Pituitary Tumors, Surgery

Further Reading DeGroot, L. J., and Jameson, J. L. (2001). ‘‘Endocrinology’’ (3 vols.). W. B. Saunders, Philadelphia. Howlett, T. A. (2002). Hypopituitarism In ‘‘The Oxford Textbook of Endocrinology’’ (H. E. Turner and J. A. H. Wass, eds.), pp. 145–151. Oxford University Press. Lamberts, S. W. J. (1997). ‘‘The Diagnosis and Treatment of Pituitary Insufficiency.’’ BioScientifica, Bristol, UK. Regal, M., Pa´ramo, C., Sierra, J. M., and Garci´a-Mayor, R. V. (2001). Prevalence and incidence of hypopituitarism in an adult Caucasian population in northwestern Spain. Clin. Endocrinol. 55, 735–740. Tomlinson, J. W., Holden, N., Hills, R. K., Wheatley, K., Clayton, R. N., Bates, A. S., Sheppard, M. C., and Stewart, P. M. (2001). Association between premature mortality and hypopituitarism: West Midlands Prospective Hypopituitary Study Group. Lancet 357, 425–431. Turner, H.E., and Wass, J. A. H. (eds.) (2002). ‘‘The Oxford Handbook of Endocrinology.’’ Oxford University Press. Wass, J. A. H., and Thakker, R. V. (2001). ‘‘Endocrine Disorders (Medicine).’’ Medicine Publishing, Abingdon, UK.

p0365

663

Hypopituitarism, Hormonal Therapy for

Table I Causes of Hypopituitarism Common causes of hypopituitarism

Less common causes of hypopituitarism

Rare causes of hypopituitarism

Pituitary tumor

Other cranial tumors

Pituitary surgery

Rathke’s cleft cyst

Empty sella syndrome Trauma

Pituitary radiotherapy

Pituitary apoplexy including Sheehan’s syndrome

Granulomatous hypophysitis

Idiopathic isolated hormone deficiencies

Hemachromatosis Pituitary metastases Internal carotid artery Aneurysm

HORMONE REPLACEMENT THERAPY IN THE PATIENT WITH HYPOPITUITARISM Adrenocorticotropin Deficiency p0015

Hydrocortisone is the drug of choice and is the firstline treatment in most countries. Hydrocortisone has traditionally been given in two doses: a larger dose of 20 mg on waking and a smaller dose of 10 mg taken with the evening meal. However, Esteban and colleagues convincingly demonstrated daily cortisol production rates in healthy individuals to be less than half of the previously accepted values. Other work has highlighted potentially detrimental effects on bone density and glucose tolerance of high tissue exposure to glucocorticoids, so a more tailored approach to therapy is now the norm. Many patients on twice-daily replacement regimens report fatigue or headache in the afternoon that is improved by taking their hydrocortisone in three divided doses, although quality-of-life assessments were unchanged with lowered total doses of hydrocortisone in one study. Therefore, doses are fine-tuned according to patient well-being and serum cortisol levels by many centers. Urine-free cortisol measurements are not useful because saturation of corticosteroid-binding globulin (CBG) following oral hydrocortisone results in supraphysiological urinefree cortisol excretion. The hydrocortisone day curve used by our institution consists of serum sampling for cortisol at time 0 before the first dose is given and then at 30 min and 1, 2, 3, 5, 7, 9, 9.5, 10, and 11 h, with hydrocortisone administered at the usual times, typically at 0, 5, and 9 h. The aim is to achieve adequate circulating levels of cortisol throughout the day without excessive peaks (e.g., >1000 nmol/L) or troughs (e.g., 20 ml/min). ‘‘Pseudohyperkalemia’’ (e.g., hemolysis, thrombocytosis) should be excluded. The next step is to demonstrate a normal cortisol response to ACTH stimulation. Then, the response of renin and aldosterone levels to stimulation (e.g., upright posture, sodium restriction) should be measured. Low renin and aldosterone levels establish the diagnosis of HH.

as diabetes mellitus, atherosclerotic cardiovascular disease, and hypertension. Mineralocorticoid replacement, usually given as fludrocortisone acetate, is the mainstay of therapy. Large doses of the steroid, often as high as 0.4 to 1.0 mg daily, are frequently required to reduce the plasma level to normal. This observation suggests that the renal tubular cell is resistant to the action of mineralocorticoids, at least with respect to their effect on potassium secretion. Sensitivity to the sodiumretentive effects of fludrocortisone appears to be less affected given that marked sodium retention with edema, exacerbation of hypertension, and congestive heart failure are common consequences. In many patients, the administration of a loop diuretic may be required to prevent or treat these complications. If the hyperkalemia fails to respond to fludrocortisone acetate, or if excessive sodium retention is encountered, a trial of potassium-wasting diuretics is desirable. In many patients, particularly those with a primary tubular potassium secretory defect, thiazide diuretics are effective. In some patients, sodium bicarbonate may improve potassium excretion and correct the hyperkalemia. If these measures fail, one can always employ a sodium–potassium exchange resin such as sodium polystyrene sulfonate. In many patients, particularly those with only mild to moderate hyperkalemia, it is best not to institute any therapy but to check the plasma potassium concentration periodically to ensure that the hyperkalemia has not reached dangerous levels. In all patients, drugs known to impair renal potassium excretion should be avoided. Volume contraction, which decreases distal sodium delivery, also should be prevented.

See Also the Following Articles Aldosterone Receptors . Primary Aldosteronism (PAL) . Renal Vein Renin . Renin . Tissue Renin-AngiotensinAldosterone System

TREATMENT

Further Reading

Several different therapeutic regimens can be employed to lower the plasma potassium concentration. However, the physician should keep in mind that the majority of patients suffering from this syndrome are elderly and have associated conditions such

DeFronzo, R. A. (1980). Hyperkalemia and hyporeninemic hypoaldosteronism. Kidney Intl. 17, 118–134. DeLeiva, A., Christlieb, A. R., Melby, J. C., Graham, C. A., Day, R. P., Luetscher, J. A., and Zager, P. G. (1976). Big renin and biosynthetic defect of aldosterone in diabetes mellitus. N. Engl. J. Med. 295, 639–643.

674 Oh, M. S., Carroll, H. J., Clemmons, J. E., Vagnucci, A. H., Levison, S. P., and Whang, E. S. (1974). A mechanism for hyporeninemic hypoaldosteronism in chronic renal disease. Metabolism 23, 1157–1166. Perez, G. O., Lespier, L., Jacobi, J., Oster, J. R., Katz, F. H., Vaamonde, C. A., and Fishman, L. M. (1977). Hyporeninemia and hypoaldosteronism in diabetes mellitus. Arch. Int. Med. 137, 852–855.

Hyporeninemic Hypoaldosteronism

Romero, J. C., Dunlap, C. L., and Strong, C. G. (1976). The effect of indomethacin and other anti-inflammatory drugs on the renin–angiotensin system. J. Clin. Invest. 58, 282–288. Schindler, A. M., and Sommers, S. C. (1966). Diabetic sclerosis of the renal juxtaglomerular apparatus. Lab. Invest. 15, 877–884. Tuck, M. L., and Mayes, D. M. (1980). Mineralocorticoid biosynthesis in patients with hyporeninemic hypoaldosteronism. J. Clin. Endocrinol. Metab. 50, 341–347.

676 Table I Classification of Anovulation Caused by the Central Nervous–Hypothalamic–Pituitary System Physiologic anovulation Prepubertal period Postmenopausal period Pregnancy and postpartum Functional hypothalamic anovulation Psychogenic or stress factors Nutritional factors Exercise-related factors Pharmacologic anovulation Opiate agonists Dopaminergic antagonists Psychiatric-associated anovulation Anorexia nervosa Pseudocyesis Organic defects Isolated GnRH deficiency Pituitary tumors Sheehan’s syndrome Empty-sella syndrome Head trauma Inappropriate prolactin secretion

hypoestrogenic milieu might be seen in the genital tract. However, complains of hot flashes are uncommon. Hypothalamic anovulation may be related to some psychiatric conditions, especially anorexia nervosa. This is an uncommon psychiatric disorder that is usually diagnosed in adolescent and collegeaged females. Anorexia nervosa is much more common in women with amenorrhea and therefore should be considered and ruled out in young patients undergoing evaluation due to anovulation. Several organic defects are related to a small but significant number of hypothalamic anovulatory females. Such disorders include pituitary tumors, Sheehan’s syndrome (due to circulatory collapse), empty sella syndrome, and head trauma. Isolated gonadotropinreleasing hormone (GnRH) deficiency, which shares many biochemical features with functional hypothalamic anovulation, is characterized by a decrease in the secretion of endogenous GnRH. The disturbed secretion of GnRH leads to hypogonadotropic hypogonadism, eunuchoid features, incomplete development of secondary sexual characteristics, primary (as opposed to secondary in functional anovulation) amenorrhea, and, in some cases, anosmia. This disorder is the result of developmental or migratoryfailure of GnRH neurons. Hypoplasia of the olfactory bulb can be identified in some patients.

Hypothalamic Anovulation, Functional

CHARACTERISTIC OF FUNCTIONAL HYPOTHALAMIC ANOVULATION Most women with functional hypothalamic anovulation have a history of normal onset of menarche with regular menstrual cycles between 26 and 35 days. These patients tend to be highly motivated and involved in stressful occupations. Changing lifestyles and an increasing emphasis on exercise, nutrition, and slenderness are expected to result in an increased incidence of hypothalamic anovulation. In some women, emotional or stressful events precede the onset of anovulatory amenorrhea. Other environmental and interpersonal factors (e.g., social maladjustment and psychosexual difficulties) can be identified during the first interview. Women with functional hypothalamic anovulation usually have low to normal body weight. Several studies have shown that these women have more psychological characteristics associated with eating disorders, such as a feeling of general inadequacy, insecurity, lack of control over life, and confusion in identifying and responding to food-related body sensations and emotional states. A thorough physical examination usually reveals normal secondary sexual development. Signs of other hormonal imbalances, such as thyroid enlargement, galactorrhea, and androgenic excess, should be searched for during the examination. A finding of such an endocrinological disorder usually challenges the diagnosis of functional hypothalamic anovulation.

PATHOPHYSIOLOGYOF FUNCTIONAL HYPOTHALAMIC ANOVULATION The common underlying defect of these various disorders is the deficiency (or slowing in frequency) in the pulsated secretion of GnRH. When exogenous GnRH is administered to women with hypothalamic anovulation, luteinizing hormone (LH) and folliclestimulating hormone responses are either normal or exaggerated, suggesting that the primary defect is at a central level with preserved function of the pituitary. A growing body of evidence indicates that an increase in endogenous opiate activity plays a significant role in reducing the pulsated GnRH secretion (and consequently LH) in functional hypothalamic anovulation. Blockage of endogenous opiate receptors by naloxone may increase LH pulse and amplitude in women with hypothalamic anovulation. Long-term treatment with opiate receptor antagonists may result in a spontaneous return of gonadotropin secretion and ovulatory function. Studies have demonstrated that blockage of the dopamine receptor with metaclopramide can also induce an increase in LH secretion.

677

Hypothalamic Anovulation, Functional

Disruption of the reproductive function during chronic exposure to stress has been reported in both animals and humans. In humans, there is a functional association between chronic stress and the onset of ovulatory dysfunction and amenorrhea. Evidence links stress, chronic activation of the hypothalamic– pituitary–adrenal axis, and reproductive dysfunction with anovulation. In women with hypothalamic anovulation, there is a significant increase in daytime cortisol levels, a delayed or absent response in ACTH and cortisol secretion during noon meals, and a blunted response to corticotropin-releasing hormone. Anovulation associated with nutritional factors and weight loss is a well-characterized clinical syndrome. The incidence of weight loss-related anovulation varies widely from series to series, depending on the population studied. It is well-known that the proper amount of fat is crucial for the timely development of menarche (total fat should comprise 17% of body weight, corresponding to a body mass index of approximately 19 kg/m2). Regular ovulation and menstrual cycles necessitate the maintenance of at least 22% of body weight as fat. Weight loss by selfenforced abstinence, starvation, chronic illness, or exercise leads to impaired GnRH secretion and peripheral impaired estrogen status. Regular strenuous physical activity in women is often associated with anovulation and menstrual disturbances (e.g., delayed menarche, oligomenorrhea, amenorrhea, and luteal phase defect). Endocrinological status is affected by exercise, with some similarity to starvation-associated modifications. Two stress-related mechanisms have been implicated in amenorrheic athletes. Exercise-induced increases in b-endorphin levels influence the frequency and amplitude of LH pulses. Endurance exercise may also cause the release of corticotropin-releasing factor, inhibiting gonadotropin secretion and activating the adrenal release of corticosteroids and androgenic steroids. Although weight loss plays a crucial role in exercise-related anovulation, menstrual function may resume spontaneously during training, even if there is no change in body weight or composition.

MANAGEMENT Because of the functional nature of hypothalamic anovulation, major emphasis should be placed on carefully conducted initial interviews, focusing on lifestyle and interpersonal relationships. Since spontaneous recovery is observed in some cases, expectant management with periodic reassessment is a viable option. For women with persistent anovulation,

a major concern is the long-term effect of hypoestrogenism on bone density. When significant osteopenia is observed, estrogen replacement therapy should be initiated. For women who are interested in fertility, a trial of clomiphene citrate is indicated. For those who fail to respond to clomiphene, the recommended and logical approach is ovulation induction by physiologic replacement therapy with pulsate GnRH.

CONCLUSION The development of functional hypothalamic amenorrhea is linked to lifestyle and environmental stressors. Psychogenic, nutritional, or exercised-related factors cause a temporary shift into a ‘‘resting mode’’ of the reproductive cycle at the central level. Neuroendocrine factors, which appear to modulate reduced GnRH activity, include the opiate and dopamine neuronal system. In most women, reactivation of the hypothalamic unit occurs after lifestyle modification and accommodation of stress factors. Therapy is aimed at ensuring adequate bone density and resolving consequent infertility.

See Also the Following Articles Eating Disorders and the Reproductive Axis . Gonadotropin-Induced Ovulation . Gonadotropin-Releasing Hormone (GnRH) Actions . Osteoporosis, Overview . Ovarian Stimulation: Clomiphene Citrate . Polycystic Ovary Syndrome (PCOS) . Superovulation and Intrauterine Insemination

Further Reading Berga, S. L., Daniels, T. L., and Giles, D. E. (1997). Women with functional hypothalamic amenorrhea but not other forms of anovulation display amplified cortisol concentrations. Fertil. Steril. 67, 1024–1030. Grana-Barcia, M., Liz, J., Jimenez, E., Novo, A., and Aguilar, J. (1998). Ovulation induction with pulsatile GnRH in a patient with anovulation of hypothalamic origin and central diabetes insipidus. Gynecol. Endocrinol. 12, 203–207. Liu, J. H. (1990). Hypothalamic amenorrhea: Clinical perspectives, pathophysiology, and management. Am. J. Obstet. Gynecol. 163, 1732–1736. Marcus, M. D., Loucks, T. L., and Berga, S. L. (2001). Psychological correlates of functional hypothalamic amenorrhea. Fertil. Steril. 76, 310–316. Marshall, J. C., Eagleson, C. A., and McCartney, C. R. (2001). Hypothalamic dysfunction. Mol. Cell. Endocrinol. 183, 29–32. Mendes, M. C., Ferriani, R. A., Sala, M. M., Moura, M. D., and de Sa, M. F. (1999). Induction of ovulation with clomiphene citrate with metaclopramide in patients with amenorrhea of hypothalamic origin. Gynecol. Endocrinol. 13, 149–154. Schachter, M., Shoham, Z. (1994) Amenorrhea during the reproductive years—Is it safe? Fertil. Steril. 62, pp. 1–16.

679

Hypothalamic Disease

Figure 1 Normal hypothalamus. (A) The third ventricle is lined by the lateral hypothalamus that harbors the paraventricular nuclei. Inferiorly, the infundibulum carries nerve fibers to the pituitary stalk. (B) The nuclei of the hypothalamus are collections of ganglion cells; in this supraoptic nucleus, the neurons contain vasopressin immunoreactivity. adult human they are ill defined. From animal studies, it is known that individual nuclei have important physiological functions; for example, there are specific nuclei implicated in hunger or satiety control, temperature regulation, olfaction, circadian rhythms, sexual drive, ovarian regulation, and parenting behaviors. However, functionally, a given hormone is often produced in more than one nucleus, and in many instances a single nucleus produces more than one hormone. These data raise doubts about the concept of individual nuclei as designated functional entities.

CLINICAL MANIFESTATIONS OF HYPOTHALAMIC DISEASE Lesions of the hypothalamus cause headache, nausea, vomiting, somnolence, behavioral alterations, psychosis, and dementia. Hypothalamic destruction can result in bulimia or anorexia. Visual disturbance can result from oculomotor alterations or optic nerve damage. Hypopituitarism and diabetes insipidus are common manifestations. In severe cases, patients can develop hydrocephalus. Inflammation can result in meningitis. Diabetes insipidus is the most common and often the initial manifestation. The presentation of hypopituitarism varies with age. In children, hypothalamic dysfunction may present with dwarfism. In adults, sexual dysfunction is the most common endocrine

complaint, with impotence in males and primary or secondary amenorrhea in females. If the disease is predominantly hypothalamic or causes interruption of the pituitary stalk, pituitary hypofunction is associated with hyperprolactinemia due to destruction of the dopaminergic neurons that maintain tonic inhibition of that pituitary hormone, and stimulation confirms an intact pituitary response. If the lesion causes destruction of hypophysial tissue as well as hypothalamic disease, there is a reduction of all basal pituitary hormones or more subtle changes with reduced response to stimulation. Some tumors are associated with pituitary hormone excess. Occasionally, this is due to the production of hormones stimulating pituitary function, such as in hypothalamic gangliocytomas that secrete adenohypophysiotrophic hormones. Alternatively, the clinical manifestations may be due to the production of substances that simulate or mimic pituitary hormones. For example, germ cell tumors associated with precocious puberty produce ß-chorionic gonadotropin. Occasionally, patients manifest excessive secretion of vasopressin, resulting in the syndrome of inappropriate antidiuretic hormone.

DEVELOPMENTAL DISORDERS The failure of development of areas of the hypothalamus can lead to variable clinical manifestations. Major defects are incompatible with life.

680

p0055

Septo-optic dysplasia or de Morsier’s syndrome is a complex developmental disorder with variable manifestations of aplasia of the septum pellucidum, hypoplasia of the optic nerves, and endocrine dysfunction as well as vegetative alterations. The disorder results from mutations of a homeobox transcription factor, Hesx1, which is required for normal development of the affected regions of the central brain. Kallman’s syndrome is an X-linked developmental defect due to mutation of the KAL-1 gene that encodes an extracellular glycoprotein, anosmin-1, that is expressed during the period of human organogenesis in the early olfactory system and is required for normal migration of gonadotropin-releasing hormone (GnRH)-containing neurons. This disorder results in gonadal insufficiency in males due to hypothalamic GnRH deficiency associated with anosmia. Lawrence–Moon–Biedl syndrome is a highly polymorphic disorder that includes pigmentary retinopathy, mental retardation, spinal paraplegia and hypogonadism associated with obesity, and digital anomalies. The genetic basis is unknown and the variable clinical manifestations suggest incomplete penetrance with differential expression of the anomaly. Prader–Willi syndrome is characterized by obesity, short stature, delayed puberty, infertility, mental retardation, muscle hypotonia, hypopigmentation, and seizure disorder. It has been linked with deletions of chromosome 15q11–q13. This region contains an imprinting center that regulates paternally expressed genes whose expression is lost; some cases are attributed to maternal uniparental disomy. However, the causative gene is unknown. Wolfram’s syndrome (WS) is characterized by diabetes insipidus, diabetes mellitus, optic atrophy, and deafness (hence the acronym DIDMOAD). This rare neurodegenerative disorder exhibits autosomal recessive inheritance due to compound heterozygous mutations of the WFS1 gene, a member of a novel gene family that encodes wolframin, an endoglycosidase H-sensitive membrane glycoprotein that localizes in the endoplasmic reticulum. WFS1 is expressed predominantly in selected neurons in the hippocampus, amygdala, olfactory tubercle, and superficial layer of the allocortex—components of the limbic system or structures closely associated with this system that account for the psychiatric, behavioral, and emotional abnormalities of this syndrome. Heterozygous mutations are associated with nonsyndromic, low-frequency sensorineural hearing loss affecting only the 2000-Hz and lower range. This is an unusual disorder that worsens over time without progressing to profound deafness.

Hypothalamic Disease

Idiopathic growth hormone (GH) insufficiency is likely due to a defect in the synthesis or secretion of growth hormone-releasing hormone (GHRH) since most patients have normal pituitary GH somatotroph structure and hormone content, and they respond to GHRH administration.

HYPOTHALAMIC INFLAMMATION Infectious Lesions Acute and chronic infections of the hypothalamus are rare but they do occur, usually in association with sphenoid sinus infection, cavernous sinus thrombosis, otitis media mastoiditis, or peritonsillar abscess. Pituitary tumors have been associated with the development of pituitary abscess that can spread to the hypothalamus. It has been suggested that bony erosion by the tumor predisposes such patients to the spread of sinonasal infection. Rarely, infection results from vascular seeding of distant or systemic infection.

Noninfectious Inflammatory Lesions Sarcoidosis is a multisystem granulomatous disease of unknown etiology. It has long been attributed to an infectious agent; however, none have been identified. Disease onset usually occurs in adults and there is a predilection for blacks and females. There are usually systemic manifestations, and neural involvement is rare; however, it can occur in the hypothalamus, usually involving the meninges at the infundibulum and floor of the third ventricle. The granulomatous inflammation has a subacute or protracted course of tissue destruction that may respond to steroid suppression, and there rare reports of spontaneous resolution. Neuroinfundibulohypophysitis is a rare inflammatory condition that affects the infundibulum, the pituitary stalk, and the neurohypophysis and may be part of a range of autoimmune disorders, including lymphocytic hypophysitis. Lymphocytic hypophysitis occurs mainly in women and most often presents in the later stages of pregnancy. Infundibulohypophysitis shows no sexual predilection and usually presents with diabetes insipidus. The cause is unclear. A lesion similar to orbital pseudotumor characterized by chronic inflammation and fibrosis has been reported to involve the parasellar tissues associated with other sclerosing lesions, such as Riedel’s thyroiditis, retroperitoneal fibrosis, and sclerosing cholangitis. The etiology and appropriate management of these disorders are uncertain.

681

Hypothalamic Disease

METABOLIC LESIONS Neurodegenerative Processes Alzheimer’s disease, Parkinson’s disease, Huntington’s chorea, and other neurodegenerative diseases can involve the hypothalamus, resulting in variable endocrine, behavioral, and vegetative abnormalities.

Systemic Processes Wernicke’s encephalopathy can alter hypothalamic function, but usually the manifestations are due to mamillary body degeneration. Hemochromatosis results in hypogonadotropic hypogonadism, but this is thought to be due to iron deposition in the pituitary gonadotrophs rather than a primary hypothalamic lesion.

CYSTIC LESIONS Rathke’s Cleft Cysts p0100

These cysts originate in the remnants of Rathke’s pouch of the pituitary. Rathke’s cleft arises from the oropharynx and migrates upward with an anterior and posterior limb that ultimately give rise to the anterior and intermediate lobes of the adenohypophysis, respectively. In the human, the intermediate lobe is vestigial, and its remnants line small cystic cavities that are remnants of the cleft and usually 1 cm) suffer from a disorder of gonadal function due to both tumor mass effect and hyperprolactinemia, but in microadenoma patients or in drug-induced hyperprolactinemia, hypogonadism recognizes a functional origin due to neuroendocrine effects of elevated PRL levels that impairs pulsatile release of GnRH and ultimately that of gonadotropin. According to the hypothesis of Scanlon, it is possible that hyperprolactinemia enhances dopaminergic tone at the hypothalamic levels with a consequent inhibitory effect on gonadotropin release.

Organic Structural lesions of the hypothalamus can interfere with the neuroendocrine regulation of GnRH. In these conditions, the impairment of gonadal function is almost invariably associated with deficiency of other pituitary functions. In children, the most common form of hypothalamic hypogonadism is craniopharyngioma presenting with visual field defects, growth failure, and/or diabetes insipidus. Other common hypothalamic tumors such as meningiomas, dysgerminoma, and gliomas are accompanied by hypogonadism. Secondary hypothalamic localizations of breast tumors in women and lung tumors in men may also determine secondary gonadal impairment generally associated with diabetes insipidus. Infiltrative disorders of the hypothalamus, such as sarcoidosis, hystiocytosis, and hemocromatosis, may also cause hypothalamic hypogonadism as well as radiotherapy for central nervous system (CNS) tumors or leukemia.

Miscellaneous Disorders Secondary hypogonadism with the characteristics of HH is found in genetic conditions such as

691

Hypothalamic Hypogonadism

Prader–Willi syndrome, Bardet–Biedl syndrome, and Moebius syndrome (i.e., the association of congenital ophthalmoplegia and facial paresis). In a single patient with this latter syndrome, the association with hypogonadism and anosmia has been reported. HH resistant to treatment with GnRH has been reported in a patient with Gordon Holmes spinocerebellar ataxia.

DIAGNOSIS As a form of secondary hypogonadism, hypothalamic hypogonadism is characterized by signs and symptoms of gonadal failure, low levels of gonadal hormones, and low levels of gonadotropin. This latter finding is not absolute given that a single evaluation of follicle-stimulating hormone (FSH) and LH, particularly in the acquired forms, may display levels in the normal range. Actually, the marker of the hypothalamic disturbance is represented by alterations in gonadotropin pulsatility. In general, the most severe forms of congenital hypogonadism in patients who show no evidence of puberty are associated with no detectable pulses of gonadotropin when examined by serial samples at 10to 20-min intervals for 12 to 24 h. Pulses of normal frequency but of reduced amplitude attributed to impaired responsiveness of the GnRH receptors may also be observed. In patients with acquired hypogonadism, the unpulsatile pattern of gonadotropin strongly suggests an organic cause or is associated with severe nutritional disturbances as in anorexia nervosa. In patients with the functional form, there is generally only a decrease in pulse frequency with a pattern similar to that observed during the early follicular phase of the cycle. In patients with hypogonadotropic hypogonadism, the possibility of an organic disease of the hypothalamic–pituitary region must always be considered. Nuclear magnetic resonance imaging is always indicated in male patients and in females without a clear-cut history of functional amenorrhea. A determination of basal PRL levels must always be performed together with a gonadotropin assessment. Supranormal PRL levels, in the absence of other causes of hyperprolactinemia such as primary hypothyroidism, renal or liver failure, and treatment with antidopaminergic agents, render nuclear magnetic resonance imaging mandatory. In particular, marginally elevated PRL levels (i.e., in the range of 40–120 ng/ml) should be viewed as strongly suspected for a space-occupying lesion at the hypothalamic level that impairs the dopamine delivery to the lactotrophs. PRL levels above 120 ng/ml are diagnostic for a PRL-secreting adenoma.

When basal PRL levels have been carefully evaluated by means of at least three samples taken at bed rest and during saline infusion to avoid stress-related elevation of the hormone levels, dynamic tests such as thyrotropin-releasing hormone (TRH) injection are definitely of no value. When neuroradiology has identified an organic disease, an evaluation of the thyroid, adrenal function, and growth hormone (GH) secretion should be performed.

TREATMENT The treatment of HH depends on the nature of the disease and on the goal to be achieved in relation to pubertal stage and desire of fertility. In prepubertal patients, gonadal hormones should be administered at low initial doses that are increased gradually to obtain the development of secondary sexual characters. In men, long-acting preparations of testosterone are effective, starting with 50 mg monthly and increased by 50 mg every 3 to 6 months. Alternatively, low doses of human chorionic gonadotropin (hCG) have the advantage of stimulating testicular growth. In females, the therapy can be started with low doses of estrogens, for example, 5 mg of ethynyl estradiol to obtain breast development, followed by a cyclical therapy in combination with a progestin. During adulthood, different approaches should be followed depending on whether fertility is desired. If fertility is not desired, the goal of the treatment in men is to maintain virilization, normal sexual function, and adequate muscle and bone mass. Usually, 100 to 200 mg of long-acting testosterone esters are given every 2 to 4 weeks, although the dosage can be adjusted to obtain a testosterone level in the lower end of the normal range before the injection of the next dose. The disadvantage of this route of administration is an unfavorable pharmacokinetic profile often characterized by supranormal levels during the first week after the injection. Testosterone undecanoate, an orally administrable preparation, has an absorption that is influenced by diet composition. The transdermal testosterone preparations applied to scrotal skin or, more recently, to nonscrotal skin provide an effective method of substitutive treatment with the drawbacks of high cost and frequent skin irritations. Testosterone implants are also an effective, albeit not readily available, way in which to treat hypogonadism. In female patients, substitutive treatment is carried out with low-dose pills or with conjugated estrogens or transdermal estradiol associated with cyclic progestin to achieve endometrial protection. If fertility is desired,

692 the treatment has to be carried out with pulsatile GnRH, gonadotropins, or clomiphene citrate. In men, pulsatile GnRH therapy is practically performed by subcutaneous route of injection through a portable pump, although intravenous administration produces the most physiological pattern of gonadotropin secretion. The dose per bolus is highly variable, ranging from 25 to 600 ng/kg delivered every 120 min. The dose is individually tailored to achieve normalization of testosterone and gonadotropin levels. This route of therapy is effective in obtaining spermatogenesis in the majority of patients, with treatments lasting up to 2 years or even longer in patients with very small basal testicular volume. When testicular volume reaches 8 ml, semen analyses are obtained. Gonadotropin treatment is carried out by a combination of hCG and recombinant FSH. The starting schedule is 500 to 2000 IU of hCG two or three times weekly, followed after 6 months by FSH (75 IU three times weekly) in relation with testicular growth and induction of spermatogenesis. In patients with adultonset HH, spermatogenesis may be induced by the use of hCG alone. The main side effect of gonadotropin treatment is gynecomastia, which can be minimized by reducing hCG doses to keep testosterone levels at the lowest limit of the normal range. The therapeutic approaches to achieving fertility in women are similar to those in men, but the treatment with gonadotropin in women is slightly less effective in terms of cumulative rate of contraception and is more tolerated in terms of multiple folliculogenesis than that with GnRH. The better results obtained by GnRH are probably due to the fact that it is possible to mimic the hormone dynamics of a normal

Hypothalamic Hypogonadism

menstrual cycle and to achieve the maturation of a single follicle. When ovulation by gonadotropin therapy is needed, treatment is started with 150 IU of FSH and the dose is increased until the levels of estradiol begin to rise. Follicular growth is monitored by ultrasound, and follicular rupture is induced by 10,000 IU of hCG. In patients with the acquired form of hypothalamic amenorrhea, induction of ovulatory cycles can also be attempted by clomiphene citrate (50 mg/day for 5 days).

See Also the Following Articles Anorexia Nervosa . Delayed Puberty and Hypogonadism, Female . Delayed Puberty and Hypogonadism, Male . Diabetes Insipidus, Neurogenic . Gonadotropin-Releasing Hormone (GnRH) Actions . Gynecomastia . Hypothalamus, Anatomy of . Kallmann’s Syndrome and Idiopathic Hypogonadotropic Hypogonadism . Prader-Willi Syndrome . Undescended Testes

Further Reading Hayes, F., Seminara, S., and Crowley, F., Jr. (1998). Hypogonadotropic hypogonadism. Endocrinol. Metab. Clinics North America 27, 739–759. Quinton, R., Cheow, H., Bouloux, R., and Jacobs, H. (1999). Kallmann’s syndrome: Is it always for life? Clin. Endocrinol. 50, 481–487. Quinton, R., Duke, M., Robertson, A., Maffin, G., de Zoysa, P., Azcona, G., MacColl, G., Jacobs, H., Conway, G., Besser, M., Stanhope, R., and Bouloux, P. (2001). Idiopathic gonadotropin deficiency: Genetic questions addressed through phenotypic characterization. Clin. Endocrinol. 55, 163–174. Warren, P., and Fried, L. (2001). Hypothalamic amenorrhea. Endocrinol. Metab. Clinics North America 30, 611–623. Yen, S. (1993). Female hypogonadotropic amenorrhea. Endocrinol. Metab. Clinics North America 22, 29–58.

694

Hypothalamic Hypothyroidism

an intact hypothalamic–pituitary–thyroid axis. In contrast to HH, these conditions are generally characterized by low T3 but normal TSH and FT4. Reduction of circulating TSH and FT4 occurs only in prolonged and manifest disease states and is likely due to suppression of hypothalamic TRH gene expression.

Prevalent hypothalamic defect Prevalent pituitary defect 10

FT4 (pmol/L)

8 6

CLINICAL PRESENTATION 4

2

0

0.1

1 TSH (mU/L)

10

Figure 1 Serum levels of FT4 and TSH in patients with central hypothyroidism. Dashed lines indicate the lower and upper limits of the normal ranges. The evident lack of correlation between TSH and FT4 demonstrates that quantitative impairment of TSH secretion does not completely explain the pathogenesis of CH and indicates the involvement of qualitative alterations of hypothalamic regulation in a large portion of CH patients, particularly those with prevalent hypothalamic defects.

role played by hypothalamic factors in the generation of circadian TSH rhythm. The differential diagnosis of CH is indicated by circulating levels of immunoreactive TSH: High TSH is typical of patients with prevalent hypothalamic defects (HH), whereas low TSH indicates the almost complete lack of functional thyrotropes (pituitary or secondary hypothyroidism). In CH patients with normal TSH, important information on the entity of hypothalamic impairment can be obtained by TRH testing. Indeed, differential diagnosis of CH is one of the residual indications for TRH tests. Pure HH is characterized by delayed/exaggerated/prolonged responses of immunoreactive TSH after intravenous injections of TRH (0.2 mg), but net TSH responses >4.0 mU/liter can be considered indicative of a prevalent hypothalamic defect. Conversely, blunted TSH responses (300 kb, 48 exons, open reading frame of 2748 AA). The human phenotype is characterized by variable degrees of severity of CH and goiter development when thyroid hormone replacement is delayed. Dominant inheritance of Tg mutations has been described as well, not in patients with CH but rather in patients with the development of goiter later in life. The functional role of pendrin (PDS), an anion transporter, in the thyrocyte is the transport of iodide across the apical membrane into the follicular lumen. Pendrin is also expressed in the inner ear and kidney, where its exact functions are unkown. The phenotype of Pendred’s syndrome is characterized by congenital sensorineural hearing loss, goiter, and (in only a minority of patients) overt CH. Autosomal recessive inheritance of mutations of the pendrin gene has been reported in familial cases of Pendred’s syndrome in which some newborns presented with congenital deafness, hypothyroidism, and goiter. The thyroid oxidases 1 (THOX1) and 2 (THOX2) have been identified as components of the H2O2 generating system of the thyroid. The genes encode for two very similar proteins with a fivefold higher expression of THOX 2. The putative structure of the proteins predict seven transmembrane domains, four nicotinamide adenine dinucleotide phosphate (NADPH)-binding sites, and a flavine adenine dinucleotide-binding site. A screening of patients with CH with organification defects revealed one patient with permanent CH and a homozygous mutation resulting in a truncated protein lacking the hydrogengenerating domains. In addition, three patients with transient hypothyroidism during the neonatal period were identified with monoallelic loss-of-function mutations of THOX2.

MOLECULAR DEFECTS OF THYROID ORGANOGENESIS Even in patients with putative autosomal recessive defects of thyroid hormone biosynthesis, familial cases have been described only rarely and systematic molecular genetic studies of candidate genes of thyroid hormone biosynthesis in patients with normally developed or enlarged thyroid glands are scarce. Therefore, no data on the epidemiology of the various molecular defects of thyroid hormone biosynthesis or on the possible modes of inheritance are available.

Hypothyroidism, Congenital

Some studies have described autosomal recessive inheritance and familial dominant occurrence of thyroid dysgenesis. Moreover, an increased frequency of minor abnormalities of the development of the thyroid and pharyngeal derivatives has been described in first-degree relatives of patients with CH due to thyroid dysgenesis. Again, epidemiological data on the prevalence of familial thyroid dysgenesis are scarce, but it has become apparent that thyroid dysgenesis, at least in a subset of patients, is an inherited disorder. Studies of mouse models with targeted disruption of genes involved in the development of the thyroid gland have provided insight into the molecular mechanisms of organogenesis and, thereby, the basis for molecular genetic studies in human patients affected by thyroid dysgenesis. In mice, normal organogenesis and migration have been shown to be dependent on the normal expression and interplay of at least three different transcription factors: NKX 2.1, PAX-8, and TTF-2. Furthermore, targeted mutagenesis of these transcription factors in mice has demonstrated associated developmental defects of other organs because none of these factors is exclusively expressed in the thyroid. The PAX-8 gene belongs to a family of genes that is characterized by a highly conserved paired-box DNAbinding domain, which encodes for proteins that play an important role in the entire embryonic development. PAX-8 is expressed in the thyroid primordium, the mid- and hindbrain region, and the developing kidney. Mice homozygous for disruption of the PAX-8 gene are characterized by severe hypothyroidism and small hypoplastic thyroid remnants without follicular structures, whereas in heterozygous mice no abnormalities of thyroid development have been described. Screening for mutations in the PAX-8 gene of patients with CH has led to the identification of several patients with heterozygous mutations that have been inherited in a dominant fashion. Most patients do not present with other developmental defects, but in two unrelated male patients, one hypoplastic kidney and one renal agenesis were observed. The thyroid gland of the affected patients presents with different morphological phenotypes. Thyroid hypoplasia, cystic hypoplastic remnants, and ectopic thyroid were described, and the severity of hypothyroidism was mild to moderate. FKHL15/FOXE1 (previously named TTF-2) belongs to a family of transcription factors characterized by a forkhead DNA-binding domain. FKHL15 is expressed in the thyroid, Rathke’s pouch, pharyngeal structures, and hair follicles. In mice with homozygosity for targeted disruption of the Titf2 gene, both

725

Hypothyroidism, Congenital

thyroid agenesis and thyroid ectopy were identified, indicating that athyrosis and ectopy may be regarded as different degrees of severity of the same molecular defect in humans as well. Furthermore, these mice have a cleft palate that makes their feeding impossible; therefore, early neonatal death is unavoidable. Although screening of the FKHL15 gene of patients with CH without associated problems failed to demonstrate any mutation, the study of two siblings with so-called Bamforth syndrome, including athyrosis and CH, developmental delay, cleft palate, choanal atresia, bifid epiglottis, and spiky hair, demonstrated homozygosity of a loss-of-function mutation of the FKHL15 gene in both siblings. In another family, patients were described with a less severe and incomplete phenotype, indicating either partial activity of the gene or the presence of other modifiers. FKHL15 mutations seem to be a very rare cause of CH in humans, resulting in a specific syndrome with multiple manifestations in other organs. The NKX2.1 (TTF-1, TITF-1, or T/ebp) gene encodes for a transcription factor of the homeobox domain containing genes of the NKX2 family. NKX2.1 is expressed in the thyroid, forebrain, basal ganglia, pituitary, and lung. Targeted disruption of both Titf1 alleles leads to a complex phenotype of newborn mice. These mice die shortly after birth due to respiratory distress resulting from defective lung development with insufficient surfactant production, and they lack any thyroid tissue at birth (athyrosis). Because of the early death of homozygous mice, a study of hypothalamic–pituitary function or neurological testing could not be performed. The search

t0005

for mutations in the NKX2.1 gene in patients with CH did not reveal any abnormalities. In the investigation of the NKX2.1 gene in patients with CH, where the long-term outcome, despite an early onset of treatment and adequate doses, was unfavorable due to pulmonary complications, severe muscular hypotonia, and neurological symptoms defined as choreoathetosis, heterozygous mutations of the NKX2.1 gene were identified. The phenotype of thyroid and pulmonary manifestations covers a wide spectrum ranging from hyperthyrotropinemia to severe CH due to thyroid agenesis and severe neonatal RDS requiring ventilation to a slight increase in pulmonary infections, whereas choreoathetosis presents with less phenotypical variation. Familial benign choreoathetosis without accompanying pulmonary or thyroid disorders has been attributed to NKX2.1 mutations as well. In light of the severe phenotype of NKX2.1 knockout mice, homozygosity for NKX2.1 mutations in humans is probably not viable. The mechanism by which heterozygous mutations cause the phenotype is most likely haploinsufficiency. Although heterozygous NKX2.1 mice have been reported to be unaffected previously, a more recent study described abnormalities of thyroid function and neurological development in heterozygous mice. See Tables I to IV.

CONCLUSION Neonatal screening has resulted in early diagnosis and treatment of patients with CH. Subsequently, normal mental development and outcome have been documented in more than 90% of the patients who were

Table I Molecular Defects of Thyroid Hormone Biosynthesis NIS

PDS

Tg

Function

Transport of iodine from the blood into the thyroid cell

Transport of iodide from the cytoplasm to the follicular lumen

Synthesis and storage of iodothyrosines

Human thyroid phenotype

Goiter

Goiter

Congenital Hypothyroidism in some patients

Congenital Hypothyroidism in some patients

Inheritance

Autosomal recessive

Autosomal recessive

Autosomal recessive

Chromosome

19p13

7q31

Autosomal recessive and autosomal dominant 8q24

2p25

Autosomal recessive and autosomal dominant 15q21

OMIM

601843

274600

188450

274500

607200

600044

TPO

THOX2

Iodination of Tg (organification) and coupling of iodotyrosines

Generation of H2O2

Goiter

Goiter

Congenital Hypothyroidism in some patients

Congenital Hypothyroidism

Permanent congenital hypothyroidism (homozygous) Transient congenital hypothyroidism (heterozygous)

726

t0010

Hypothyroidism, Congenital

Table II Molecular Defects of Thyroid Development TSH receptor

PAX-8

TTF-2

NKX2.1

Protein family

G protein-coupled receptor

Paired domain

Forkhead domain

Homeodomain

Transcription factor

Transcription factor

Transcription factor

Expression pattern

Thyroid

Thyroid

Thyroid

Thyroid

Pituitary

Mid- and hindbrain

Anterior pituitary

Forebrain

Hypothalamus? Adipose tissue

Kidney

Thyroid hypoplasia CH

Thyroid hypoplasia Early death

Phenotype in knockout mice

Pituitary Lung Thyroid agenesis or thyroid hypoplasia

Thyroid agenesis Pituitary aplasia

Cleft palate Early death

Forebrain defects Disturbed lung development Neonatal death

Human thyroid phenotype

Manifestation in other human organs Inheritance

Thyroid hypoplasia Severe to moderate CH

Thyroid hypoplasia Cystic rudiments Ectopy

Thyroid agenesis

Thyroid agenesis Thyroid hypoplasia Normal thyroid Severe to moderate CH

Hyperthyrotropinemia

Severe to mild CH

Severe to moderate CH

Hyperthyrotropinemia

—

Developmental defect of the kidneys in some patients

Cleft palate Bifid epiglottis

Choreoathetosis Respiratory distress

Choanal atresia

Pulmonary infections

Spiky hair Autosomal recessive

Mental retardation Autosomal dominant

Autosomal recessive

Autosomal dominant

Chromosome

14q31

2q13–14

9q22

14q12–q21

OMIM

275200

167415

602617

600635

followed up to adolescence. In up to 10% of patients detected by screening programs, neuropsychological development below the normal range was observed. The following factors, which were correlated with a less favorable outcome, have been identified: . . .

a delay in the onset of therapy, an insufficient initial thyroid hormone dose, a poor socioeconomic environment,

. .

poor compliance with therapy, other associated defects or complications.

If therapy was initiated during the first 2 weeks of life with an adequate L-thyroxine dose (> 10 mg/kg/ day), no difference in the development of patients with severe or milder forms of CH was observed. From recent studies, there is evidence that persistent developmental delay, mental retardation, and

Table III Outcome Studies of CH Study

Number of patients

Number of controls

IQ of patients (median)

IQ of controls (median)

New England, 1990

72

144

106

109

Glorieux, 1985

36

195

102

106*

Illig, 1986 Toublanc, 1990

40 49

40 52

104 116

108 118

Illicki, 1991

60

68

101

97

Fuggle, 1991

344

443

105

112*

Rovet, 1992

95

108

107

111

Kooistra, 1994

62

133

97

103

727

Hypothyroidism, Congenital

Table IV Outcome Studies in Correlation with Severity of CH and Age at Onset of Treatment Median IQ Study Illicki, 1991

Severe CH

Moderate CH

Significance

Age at onset of therapy (median þ range)

100

101

n.s.

Glorieux, 1992

89

104

P< 0.001

37 (10–65)

Kooistra, 1994

95

101

P< 0.001

23 (14–93) 17 (1–173)

Tillotson, 1994

108

115

P< 0.001

Own study, 1997

100

102

n.s.

neurological symptoms in some patients with early treatment of CH may be caused by the same molecular defect leading to an impaired development of the thyroid gland rather than by fetal or perinatal hypothyroidism. However, further research is needed to define these molecular mechanisms. During the era of newborn screening, CH remains an apparent sporadic disease. However, with the normal outcome of patients with early diagnosis and treatment leading to a normal reproduction rate, it is possible that inheritance even of thyroid dysgenesis will become more prevalent. Virtually all industrialized countries now have neonatal screening programs for hypothyroidism in which capillary blood specimens (obtained from heel pricks) collected on filter paper soon after birth are analyzed for TSH or T4.

See Also the Following Articles Hypothyroidism, Causes of . Hypothyroidism, Congenital, Long-Term Follow-Up . Hypothyroidism, Congenital, Screening Programs . Hypothyroidism, Diagnosis of . Hypothyroidism, Systemic Manifestations of . Hypothyroidism, Treatment of . Thyroid Disease, Epidemiology of . TSH (Thyroid-Stimulating Hormone; Thyrotropin)

Further Reading Biebermann, H., Liesenkotter, K. P., Emeis, M., Obladen, M., and Gru¨ters, A. (2001). Severe congenital hypothyroidism due to

12 (8–22)

9 (1–32)

a homozygous mutation of the b-TSH gene. Pediatr. Res. 46, 170–173. Biebermann, H., Scho¨neberg, T., Krude, H., et al. (1995). Mutations of the human thyrotropin receptor gene causing thyroid hypoplasia and congenital hypothyroidism. J. Clin. Endocrinol. Metab. 82, 3471–3480. Bikker, H., Vulsma, T., Baas, F., et al. (1995). Identification of five novel mutations in the human TPO gene by denaturating gel electrophoresis. Hum. Mut. 195(6), 9–16. Clifton-Bligh, R. J., Wentworth, J., Heinz, P., et al. (1998). Mutation of the gene encoding human TTF-2 associated with thyroid agenesis, cleft palate, and choanal atresia. Nat. Genet. 18, 399–401. Coyle, B., Reardon, W., Herbrick, J. A., et al. (2001). Molecular analysis of the PDS gene in Pendred syndrome. Hum. Mol. Genet. 7, 1105–1112. Devos, H., Rodd, C., Gagne, N., et al. (1999). A search for possible mechanisms of thyroid dysgenesis: Sex ratios and associated malformations. J. Clin. Endocrinol. Metab. 84, 2501–2506. Fuijwara, H., Tatsumi, K., Miki, H., Harada, T., Miyai, K., Takai, S., and Amino, N. (1997). Congenital hypothyroidism caused by a mutation in the sodium–iodide symporter. Nat. Genet. 16, 124–125. Krude, H., Schutz, B., Biebermann, H., et al. (2002). Choreotetosis, hypothyroidism, and pulmonary problems due to NKX2.1 haploinsufficiency. J. Clin. Invest. 109, 475–480. Leger, J., Marinovic, D., Garel, C., Bonaiti-Pellie, C., Polak, M., and Czernichow, P. (2002). Thyroid developmental anomalies in first degree relatives of children with congenital hypothyroidism. J. Clin. Endocrinol. Metab. 87, 575–580. Macchia, P. E., Lapi, P., Krude, H., et al. (1998). PAX8 mutations associated with congenital hypothyroidism caused by thyroid dysgenesis. Nat. Genet. 19, 83–86. Van de Graaf, S. A., Ris-Stalpers, C., Veenboer, G. J., et al. (2001). A premature stop codon in TG mRNA results in familial goiter and moderate hypothyroidism. J. Clin. Endocrinol. Metab. 84, 2537–2542.

729

Hypothyroidism, Congenital, Long-Term Follow-Up

colleagues defined a plasma total T4 at diagnosis of 43 nmol/L as the critical point below which the IQ becomes affected by CH. Both Tillotson and colleagues and Derksen-Lubsen and Verkerk concluded that treatment factors did not appear to have an impact on developmental outcome. However, several recent more studies of children born during the 1990s and treated at a mean age of 14 days with 10 to 15 mg/kg/day of T4 showed that the developmental gap that existed between children with severe CH and controls has been closed. Both early and high-dose treatment appear to be necessary, but this remains controversial.

TREATMENT AND FOLLOW-UP Most centers are now able to start treatment at a mean age of 9 to 14 days. The rationale for increasing the initial dose of T4 was that, with 5 to 6 mg/kg/day, persistent elevations of thyroid-stimulating hormone (TSH) and persistent delay in bone maturation at 3 years were observed. The starting dose at many centers is now 10 to 15 mg/kg/day. This regimen promptly normalizes plasma TSH but is associated with plasma-free T4 levels that can be above the reference range of most laboratories. However, it is important to recognize that the normal ranges of free T4 and of total triiodothyronine (T3) extend to much higher levels in infants than in older children and adults. In addition, with these starting doses, the mean plasma level of T3 remains within the normal range, and objective signs of hyperthyroidism have not been documented. Treament must be started without waiting for the results of confirmatory tests. (In cases of doubtful diagnosis, treatment can be safely stopped at 3 years of age.) TSH, T3, and free T4 will be checked 2 to 4 weeks later to ensure that thyroid hormones have risen to normal values and that TSH has normalized. Later, clinical and biological follow-up at least at 3, 6, 9, and 12 months will allow adjustment of the treatment during the first year of life. TSH has to be kept in the normal range but does not need to be suppressed, and free T4 has to be kept near the upper limit of the normal range. Afterward, controls every 6 months will be sufficient so long as TSH and free T4 remain in the normal range. The dose of T4 will be titrated upward if the TSH rises and downward if TSH is consistently below the normal range and/or if T3 is high. Bone mass is not compromised by the large doses of T4 and high free T4 levels during the first year of life. There is anecdoctal evidence, but no published data based on objective measurements, of

excessive irritability during the first 2 years of life when T4 levels are high. Infants with dyshormonogenesis appear to generally need less T4 than those with dysgenesis. Children with CH fed with soybean formulas may require a higher dose of T4 because soy products interfere with T4 absorption. If an orthotopic thyroid gland was present on the initial nuclear medicine scan, and if there is no secondary rise in TSH requiring an increase in thyroid hormone doses after it has returned to normal, transient hypothyroidism should be suspected. In these cases, treatment should be discontinued at 3 years of age, when thyroid hormone insufficiency no longer has irreversible effects on brain development, and thyroid homones and TSH should be measured 3 to 4 weeks later. If the TSH has not risen to a level exceeding 10 U/L, it should be tested again in another 6 weeks. If it remains normal, transient hypothyroidism is confirmed.

CONCLUSION The most important aspect of the outcome of CH is developmental. Most children with CH have normal neuropsychological development (and normal physical growth) when they are treated early and with high doses of T4. Further studies of neurophysiological functions that are more sensitive to under- or overtreatment than is the measurement of IQ will probably lead to greater individualization of initial dose recommendations. In the meantime, starting as early as possible with 10 to 15 mg/kg/day appears to be safe and seems to allow all children with CH, including those with a severe form of the disease, to achieve their full intellectual potential. In view of the excellent outcome of children with CH detected and treated shortly after birth, testing of cognitive functioning can be limited to those children who are diagnosed late and/or who have school difficulties.

See Also the Following Articles Hypothalamic Disease . Hypothalamic Hypothyroidism . Hypothyroidism, Causes of . Hypothyroidism, Congenital, Screening Programs . Hypothyroidism, Diagnosis of . Hypothyroidism, Subclinical . Hypothyroidism, Systemic Manifestations of . Hypothyroidism, Treatment of . Thyroid Hormone Metabolism . TSH Function and Secretion

Further Reading American Academy of Pediatrics, Section on Endocrinology and Committee on Genetics, and American Thyroid Association

730 Committee on Public Health. (1993). Newborn screening for congenital hypothyroidism: Recommended guidelines. Pediatrics 91, 1203–1209. Bongers-Schokking, J. J., Koot, H. M., Wiersma, D., Verkerk, P. H., and de Muinck Keizer-Schrama, S. M. (2000). Influence of timing and dose of thyroid hormone replacement on development in infants with congenital hypothyroidism. J. Pediatr. 136, 292–297. Derksen-Lubsen, G., and Verkerk, P. H. (1996). Neuropsychologic development in early treated congenital hypothyroidism: Analysis of literature data. Pediatr. Res. 39, 561–566. Dubuis, J-M., Glorieux, J., Richer, F., Deal, C. L., Dussault, J. H., and Van Vliet, G. (1996). Outcome of severe congenital hypothyroidism: Closing the developmental gap with early high dose levothyroxine treatment. J. Clin. Endocrinol. Metab. 81, 222–227. Hrytsiuk, I., Gilbert, R., Logan, S., Pindoria, S., and Brook, C. G. (2002). Starting dose of levothyroxine for the treatment of congenital hypothyroidism: A systematic review. Arch. Pediatr. Adolesc. Med. 156, 485–491.

Hypothyroidism, Congenital, Long-Term Follow-Up

Klein, R. Z., and Mitchell, M. L. (2000). Neonatal screening. In ‘‘Werner and Ingbar’s The Thyroid (L. E. Braverman and R. D. Utiger, eds.), pp. 973–977. Lippincott Williams & Wilkins, Philadelphia. New England Congenital Hypothyroidism Collaborative. (1981). Effects of neonatal screening for hypothyroidism: Prevention of mental retardation by treatment before clinical manifestations. Lancet 2, 1095–1098. Rovet, J. F., and Ehrlich, R. M. (1995). Long-term effects of L-thyroxine therapy for congenital hypothyroidism. J. Pediatr. 126, 380–386. Tillotson, S. L., Fuggle, P. W., Smith, I., Ades, A. E., and Grant, D. B. (1994). Relation between biochemical severity and intelligence in early treated congenital hypothyroidism: A threshold effect. Br. Med. J. 309, 440–445. Van Vliet, G. (1999). Neonatal hypothyroidism: Treatment and outcome. Thyroid 9, 79–84. Van Vliet, G. (2001). Treatment of congenital hypothyroidism. Lancet 358, 86–87.

732 newborns with sporadic permanent CH. Sensitive TSH assays have been shown to reduce the number of controls. This method is used in most countries of Europe and Japan as well as in developing countries with recent screening programs ongoing. Some countries have a more ambitious program for screening: recognizing all forms of hypothyroidism by the measurement of T4, TSH. and thyroid hormonebinding protein (TBG), but it is expensive and offers fewer advantages of detecting hypothalamo–pituitary hypothyroidism or deficiency of TBG. T4 level is determined first, and a confirmation by TSH measurement is performed in the lowest T4 values (10–20%) (in the Netherlands, an assessment of TBG is realized on the 5% lowest T4 values). This method is also used in the United States. Furthermore, the screening results obtained by these measurements can be used to monitor the iodine supply in the newborn population. This is an important issue because there is still iodine deficiency in many European countries.

RESULTS p0045

The mean incidence of CH observed in North America, Japan, and Europe was 1 per 3000 to 1 per 4000 births, without seasonal variations. CH had an ethnic background; the highest incidence, 1 per 1600 births, was observed in the Middle East due to the high prevalence of dyshormonogenesis induced by consanguinity. Conversely, some ethnic groups were less affected, for example, people of African ethnic origin (incidence varying from 1 per 10,000 to 1 per 20,000 births in the United States). Etiologies observed in all series were similar, with the type being revealed by ultrasound or radionuclide scans: athyreosis, 30 to 40%; ectopy, 40 to 50%; dyshormonogenesis, 15%; and hypoplasia, 5%. Athyreosis and ectopies had a feminine prevalence of 4 to 1; with dyshormonogenesis being autosomal recessive, the sex ratio was 1 to 1. This genetic background is an argument for a genetic origin of CH, but only 2% were family cases in the French population. The babies underwent screening on day 3 in combination with other disease screening. They were recalled on days 8 to 10, and L-T4 treatment was started (10–15 mg/kg/day). The treatment was managed on blood test results; the thyroxine should be in the normal range within 2 weeks, and TSH should be normalized within 4 weeks and should remain in the normal range afterward to obtain the best long-term results. Probably, the first patients detected by screening 20 years ago have not reached an optimal achievement due to late onset of treatment and no

Hypothyroidism, Congenital, Screening Programs

tight biological control; nevertheless, newly detected patients with CH could have a normal outcome. Two other categories have been raised from the screening results: false-positive and false-negative cases. First, false positives are babies recalled for CH without permanent CH. Causes are multiple (e.g., transient hypothyroidism in preterm infant, iodine overload in mother or baby, autoimmune disease in mother). Isolated elevated TSH persisting over months should be investigated by ultrasonography or scintigraphy to display big ectopies or dyshormonogenesis or to be checked for mutations of TSH receptor (TSHr). Second, false negatives are babies who are genuinely CH but who are not screened and recalled at the proper time; they are the failures of screening. The main causes of screening pitfalls are contaminated samples and human errors in either processing the sample or reporting the result. Furthermore, pituitary TSH deficiency cannot be detected by the measurement of TSH alone. CH could be missed by both methods in dyshormonogenesis with iodine overload or in infants having an exchange transfusion prior to the screening.

PHYSIOPATHOLOGY The physiopathology of CH by dysgenesis is still unknown. Recently, four genes—TTF1 and TTF2 (thyroid transcription factors 1 and 2), PAX8, and TSHr—have been shown to be mutated in some very specific instances. Mutations of these genes have been searched extensively in sporadic cases and were not found. Therefore, thyroid dysgenesis might be due either to a cascade of genes controlled by TTF1, TTF2, and PAX8 or to a multifactorial origin related to genetic background. Dyshormonogenesis is related to mutation in genes of the biochemical pathway: NIS (symporter Na/iodine), TPO (thyroid peroxidase), thyroglobulin, desiodases (DI and DII), and Pendred syndrome (Pendrine).

See Also the Following Articles Hypothalamic Hypothyroidism . Hypothyroidism, Causes of . Hypothyroidism, Congenital, Long-Term Follow-Up . Hypothyroidism, Diagnosis of . Hypothyroidism, Subclinical . Hypothyroidism, Systemic Manifestations of . Hypothyroidism, Treatment of

Further Reading Castanet, M., Polak, M., Bonaiti-Pellie, C., Lyonnet, S., Czernichow, P., and Leger, J. (2001). Nineteen years of national screening for congenital hypothyroidism: Familial cases with

Hypothyroidism, Congenital, Screening Programs

thyroid dysgenesis suggest the involvement of genetic factors. J. Clin. Endocrinol. Metab. 86, 2009–2014. Derksen-Lubsen, G., and Verkerk, P. H. (1996). Neuropsychologic development in early treated congenital hypothyroidism: Analysis of literature data. Pediatr. Res. 39, 561–566. Dubuis, J. M., Glorieux, J., Richer, F., Deal, C. L., Dussault, J. H., and Van Vliet, G. (1996). Outcome of severe congenital hypothyroidism: Closing the developmental gap with early high dose levothyroxine treatment. J. Clin. Endocrinol. Metab. 81, 222–227.

733 Kopp, P. (2002). Perspective: Genetic defects in the etiology of congenital hypothyroidism. Endocrinology 143, 2019–2024. Toublanc, J. E. (1992). Comparison of epidemiological data on congenital hypothyroidism in Europe with those of other parts in the world. Horm. Res. 38, 230–235. Working Group on Neonatal Screening of the European Society for Paediatric Endocrinology. (1999). Revised guidelines for neonatal screening programs for primary congenital hypothyroidism. Horm. Res. 52, 49–52.

735

Hypothyroidism, Diagnosis of

t0005

Table I Accuracy of 12 Symptoms and Signs in the Diagnosis of Primary Hypothyroidism (Percentages) Symptoms and signs

Sensitivity

Specificity

Positive predictive value

Negative predictive value

Hearing impairment

22

98

90

53

Diminished sweating

54

86

80

65

Constipation

48

85

76

62

Parasthesia

52

83

75

63

Hoarseness

34

88

73

57

Weight increase Dry skin

54 76

78 64

71 68

63 73

Slow movements

36

99

97

61

Periorbital puffiness

60

96

94

71

Delayed ankle reflex

77

94

92

80

Coarse skin

60

81

76

67

Cold skin

50

80

71

62

Symptoms

Physical signs

Note. Hypothyroid, 6 points; intermediate, 3–5 points; euthyroid, 2 points. Source. Zulewski, H., et al. (1997). Estimation of tissue hypothyroidism by a new clinical score: Evaluation of patients with various grades of hypothyroidism and controls. J. Clin. Endocrinol. Metab. 82, 771–776. ß 1997, The Endocrine Society.

best single assay for detection of hypothyroidism. Test characteristics of the serum TSH assay for the diagnosis of abnormal thyroid function are 98.8% sensitivity, 94.3% specificity, 83.9% positive predictive value, and 99.7% negative predictive value. The high diagnostic accuracy of the TSH assay is caused by the exquisite sensitivity of the pituitary for small changes in serum thyroid hormone concentrations. Because of the negative feedback of thyroid hormone on the TSH release from the pituitary, a fall in serum thyroxine will result in elevated TSH levels in serum, and a rise in serum thyroid hormone concentrations will suppress serum TSH. Figure 1 depicts a flow diagram for the biochemical diagnosis of hypothyroidism. If TSH is normal, euthyroidism is nearly certain and no further tests are necessary. However, central hypothyroidism (due to TSH deficiency) may be overlooked because serum

TSH assay  -------------------------------------------------↓ ↓ TSH normal TSH elevated Euthyroidism*  FT4 assay ↓ ---------------------------------------------------↓ ↓ ↓ FT4 decreased FT4 elevated FT4 normal Primary hypothyroidism

Subclinical Thyroid hormone resistance hypothyroidism TSH-producing adenoma

Figure 1 Flow diagram for the biochemical diagnosis of hypothyroidism. *, see text.

TSH in this condition is usually normal or decreased. Fortunately, clinical examination provides sufficient clues to suspect hypothalamic or pituitary disease such as symptoms arising from space-occupying lesions in the sella or from overproduction of pituitary hormones. As a rule, lack of gonadotropins occurs before the onset of TSH deficiency; therefore, the presence of regular menstrual periods in women or normal potency in men renders central hypothyroidism unlikely. The low incidence of central hypothyroidism (2.7 per 100,000 persons per year) does not warrant routine free thyroxine (FT4) measurements after a normal TSH test result. If TSH is elevated, a decreased serum FT4 value indicates overt hypothyroidism and a normal FT4 points to subclinical hypothyroidism. In both instances, the thyroid gland itself is at fault (primary hypothyroidism). The very rare combination of elevated TSH and elevated FT4 allows the diagnosis of thyroid hormone resistance or TSH-producing adenoma.

NOSOLOGIC DIAGNOSIS The cause of the hypothyroidism may reveal itself in many instances from the history (e.g., recent delivery, exposure to iodine excess, family members with autoimmune thyroid disease, use of antithyroid drugs, thyroid surgery or 131I therapy) and physical examination (although most patients will have no goiter). The presence of thyroid peroxidase antibodies in serum indicates chronic autoimmune thyroiditis.

736 Thyroid scans usually show low and inhomogenous uptake of the radioisotope. Preserved thyroidal radioiodine uptake and homogenous distribution of the tracer increases the likelihood of reversible hypothyroidism (due to iodine excess). Spontaneous recovery of hypothyroidism occurs in the course of subacute and postpartum thyroiditis and sometimes during the first 6 months after subtotal thyroidectomy or 131I therapy. It is exceptional in chronic autoimmune thyroiditis.

See Also the Following Articles Hypothalamic Hypothyroidism . Hypothyroidism, Causes of . Hypothyroidism, Congenital . Hypothyroidism, Subclinical . Hypothyroidism, Systemic Manifestations of . Hypothyroidism, Treatment of

Hypothyroidism, Diagnosis of

Further Reading Billewicz, W. L., Chapman, R. S., Crooks, J., et al. (1969). Statistical methods applied to the diagnosis of hypothyroidism. Q. J. Med. 38, 255–266. Hollowell, J. G., Staehling, N. W., Flanders, W. D., et al. (2002). Serum TSH, T4, and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J. Clin. Endocrinol. Metab. 87, 489–499. Seshadri, M. S., Samuel, B. U., Kanagasabapathy, A. S., and Cherian, A. M. (1989). Clinical scoring system in hypothyroidism: Is it useful? J. Genet. Int. Med. 4, 490–492. Waise, A., and Belihetz, P. E. (2000). Unsuspected central hypothyroidism. Br. Med. J. 321, 1275–1277. Wiersinga, W. M. (1989). The value of sensitive TSH measurements in clinical practice. J. Endocrinol. Invest. 9(Suppl. 4), 67–76. Zulewski, H., Mu¨ller, B., Exer, P., et al. (1997). Estimation of tissue hypothyroidism by a new clinical score: Evaluation of patients with various grades of hypothyroidism and controls. J. Clin. Endocrinol. Metab. 82, 771–776.

738

Hypothyroidism, Subclinical

SCORE:

Thyrotropin (TSH)

80

(n = 82)

Intermediate

Hypothyroid

100

60

P < 0.0001

40

80

TSH > 6-12mU/L

20 TSH ≤ 6mU/L

0 0

5

10

15

6060 Microsomal Thyroid Antibodies (MAB )

MAB +

Percentage of patients

Incidence of overt hypothyroidism (%)

Euthyroid

TSH > 12mU/L

60

40

20

(n = 82)

0

P < 0.05

3030

MAB -

C

Subclinical Overt hypothyroidism hypothyroidism Normal T3 Low T3

Figure 2 Clinical assessment of patients and controls with the 00 0

5

10

15

Years of follow-up f0005

Figure 1 Kaplan-Meier estimates of incidence of overt hypothyroidism according to TSH and microsomal thyroid antibody levels. Graphs show natural course and spontaneous evolution without treatment. Adapted from Huber, G., Staub, J. J., Meier, C., Mitrache, C., Guglielmetti, M., Huber, P., and Braverman, L. E. (2002). Prospective study of the spontaneous course of subclinical hypothyroidism: prognostic value of TSH, thyroid reserve, and thyroid antibodies. J. Clin. Endocrinol. Metab. 87, 3221–3226. later. In this survey, 97 females with subclinical or overt hypothyroidism have been detected with a prevalence of 9.3% and a calculated incidence of 0.41% per year. The risk factors for the development of hypothyroidism were a raised TSH level and/or positive antithyroid antibodies. For females with a raised serum TSH, a mean annual risk for developing hypothyroidism of 26% over 10 years could be calculated. In a prospective study of 82 female patients with known subclinical hypothyroidism, we investigated the natural course of this syndrome with regular evaluations at yearly intervals. Over a mean period of 9.2 years, 28% of the patients progressed to overt hypothyroidism, 68% remained in the subclinical state, and 4% reverted to a normal TSH. The incidence of overt hypothyroidism was correlated with the initial serum TSH concentrations. The calculated 10-year rate of overt hypothyroidism was 0% for TSH levels of 4 to 6 mU/L but was 43% for values of 6 to 12 mU/L and 77% for TSH levels greater than 12 mU/L (Fig. 1).

clinical score by Zulewski and colleagues in subclinical hypothyroidism (n ¼ 93), overt hypothyroidism (n ¼ 50), and age-matched controls (n ¼ 80). Only the patients with the most severe hypothyroidism (with low triiodothyronine [T3]) are clearly hypothyroid. The patients with subclinical hypothyroidism or with less severe hypothyroidism (normal T3) show just a few clinical signs. Adapted from Zulewski, H., Mu¨ller, B., Exer, P., Miserez, A. R., and Staub, J. J. (1997). Estimation of tissue hypothyroidism by a new clinical score: Evaluation of patients with various grades of hypothyroidism and controls. J. Clin. Endocrinol. Metab. 82, 771–776.

CLINICAL MANIFESTATIONS AND BENEFITS OF TREATMENT Symptoms One of the most controversial aspects concerning subclinical hypothyroidism is whether affected patients are symptomatic and so may benefit from thyroid hormone replacement. Based on case control studies, nearly 30% of patients with subclinical hypothyroidism may have symptoms that are suggestive of thyroid hormone deficiency (Fig. 2). Symptoms and signs of hypothyroidism may be very vague and nonspecific and so are easily overlooked in an individual patient. Five randomized, placebo-controlled intervention trials over a period of 6 to 12 months were published. These studies support the finding that patients with subclinical hypothyroidism may indeed have, in part, specific clinical signs and symptoms of hypothyroidism (Table I). Three of these studies reported significant improvement in signs and symptoms of hypothyroidism assessed by various clinical scores of hypothyroidism, whereas two other studies found no benefit of thyroxine therapy.

739

Hypothyroidism, Subclinical

t0005

Table I Effect of Thyroxine Replacement Therapy on General Symptoms in Patients with Subclinical Hypothyroidism: Randomized Placebo-Controlled Trials Treatment duration (months)

TSH level (mU/L) n

before L-T4

on L-T4

12

17

10.8

2.6

Symptom score (Billewicz; P < 0.05)

Cooper et al. Ann. Int. Med. (1984)

6

17

7.7

1.9

Symptom score (Billewicz; P < 0.01)

10 12

18 31

12.1 12.8

4.6 3.1

HRQL questionnaire (P ¼ n.s.) Symptom score (Billewicz, Zulewski; P < 0.05)

Nystro¨m et al. Clin. Endocrinol. (1988) Jaeschke et al. J. Gen. Int. Med. (1996) Meier et al. J. Clin. Endocrinol. Metab. (2001)

6

23

8.0

3.4

HRQL questionnaire (P ¼ n.s.)

Symptomatic response

Reference

Kong et al. Am. J. Med. (2002)

Note. The symptomatic response to thyroxine treatment was assessed by clinical scores based on Billewicz, W. Z., Chapman, R. S., Crooks, J., et al. (1969). Statistical methods applied to the diagnosis of hypothyroidism. Q. J. Med. 38, 255–266, and Zulewski, H., Mu¨ller, B., Exer, P., et al. (1997). Estimation of tissue hypothyroidism by a new clinical score: Evaluation of patients with various grades of hypothyroidism and controls. J. Clin. Endocrinol. Metab. 82, 771–776, as well as by quality-of-life scales.

Cognitive Function and Neuropsychiatric Dysfunction Several studies suggest that subclinical hypothyroidism is associated with neuropsychiatric disease, with higher scores on scales of depression and anxiety in affected patients. Neuropsychiatric parameters (i.e., reaction time and figure identification), as well as memory scores, were significantly improved in four intervention trials assessing the effect of thyroxine treatment in patients with subclinical hypothyroidism (mean TSH levels at baseline: 7 to 12 mU/L).

Atherosclerosis and Cardiovascular Risk Factors The relationship between subclinical hypothyroidism and atherosclerotic risk factors has been widely investigated, with major interest in serum lipid abnormalities. Several cross-sectional studies found serum lipid concentrations, mainly total cholesterol levels, to be within the normal range, whereas others detected significant elevations of total cholesterol and/or lowdensity lipoprotein (LDL) cholesterol concentrations, especially in smokers. In addition, reduced highdensity lipoprotein (HDL) cholesterol levels were reported in some studies. Two meta-analyses reanalyzed the effect of several intervention trials and could demonstrate a beneficial effect of thyroxine on serum cholesterol levels. In analyzing 13 studies, Danese and co-workers reported favorable reductions of total cholesterol and LDL cholesterol levels, with mean decreases of 0.20 and 0.26 mmol/L, respectively. These findings were confirmed by two randomized, placebo-controlled studies reporting

significant reductions of total and LDL cholesterol concentrations in thyroxine-treated women with subclinical hypothyroidism (Fig. 3). Our ‘‘Basel Thyroid Study’’ is the first randomized trial that demonstrates a significant effect of thyroxine on total cholesterol and atherogenic LDL cholesterol levels. Based on this study, we calculated a relevant risk reduction of cardiovascular mortality of 9 to 31% in relation to the observed improvement in LDL cholesterol levels. Hence, subclinical hypothyroidism must be considered as a risk factor for the development of atherosclerosis and coronary heart disease. Many cross-sectional studies have suggested an association between subclinical hypothyroidism, or autoimmune thyroid disease, and atherosclerosis. A recent population-based survey reported that increased serum TSH is an independent risk factor for the development of aortic atherosclerosis and myocardial infarction. In this study, the risk of having atherosclerotic disease was twice as high in women with subclinical hypothyroidism than in euthyroid controls. The difference persisted after adjustment for age, body mass index, blood pressure, smoking status, and cholesterol levels. Further mechanisms are thought to be involved in the association between mild thyroid failure and cardiovascular disease. These include smoking, a hypercoagulable state, elevated lipoprotein and homocysteine levels, and endothelial effects of thyroid hormones.

Others Myocardial function has been shown to be slightly impaired in patients with subclinical hypothyroidism. The identified functional abnormalities include impaired myocardial contractility and diastolic function,

740

Hypothyroidism, Subclinical

Jaeschke (1996)

(n = 31)

Caron (1990)

(n = 29)

Kong (2002)

(n = 40)

Miura (1994)

(n = 15)

Nilsson (1976)

(n = 29)

Bell (1985)

(n = 18)

Nyström (1988)

(n = 17)

Meier (2001)

(n = 63)

Paoli (1998)

(n = 15)

Cooper (1984)

(n = 33)

Franklyn (1993)

(n = 11)

Arem (1995)

(n = 14)

Caraccio (2002)

(n = 49)

Arem (1990)

(n = 13)

Powell (1989)

(n = 15)

Bogner (1993)

(n = 7)

Overall

(n = 388) −2.0

−1.5

−1.0

−0.5

0.0

0.5

Change in total cholesterol (mmol/L)

Figure 3 Effect of thyroxine treatment on total cholesterol levels in subclinical hypothyroidism. Randomized placebocontrolled trials are shown in boxes. Adapted from Danese, M. D., Ladenson P. W., Meinert, C. L., and Powe, N. R. (2000). Clinical review 115: Effect of thyroxine therapy on serum lipoproteins in patients with mild thyroid failure—A quantitative review of the literature. J. Clin. Endocrinol. Metab. 85, 2993–3001. at rest or with exercise, with beneficial results after thyroid hormone replacement therapy. Furthermore, abnormalities in peripheral nerve function and neuromuscular activity, skeletal muscle abnormalities, intraocular pressure, and ovulatory dysfunction have been described.

DIAGNOSTIC AND THERAPEUTIC RECOMMENDATIONS Screening Screening for thyroid disease is still a controversial issue. TSH screening in women over 35 years of age has been shown to be cost-effective. Some authors favor TSH screening for thyroid dysfunction in asymptomatic adults, although others do not favor

this policy because the effects of subsequent thyroxine therapy are not beneficial in all patients with subclinical hypothyroidism. Subclinical hypothyroidism is a frequent finding in women over 40 years of age (affecting about 10% of the female population in this age group), and its clinical presentation may be subtle. Therefore, TSH screening should be advocated at least by a case-finding approach, focusing on patients visiting their physicians for unrelated reasons. Because smoking impairs both thyroid hormone secretion and thyroid hormone action, smoking status should be considered in the evaluation of patients in whom hypothyroidism is suspected.

Replacement Therapy with Thyroxine The goal of treating patients with mild thyroid failure is to reverse clinical and metabolic alterations by

741

Hypothyroidism, Subclinical

TSH level elevated, fT4 level normal (confirmed on follow-up testing)

TSH level < 10mU/L

TSH level ≥ 10mU/L

Elevated thyroid autoantibody levels Hypothyroid symptoms, smoking, hypercholesterolemia Goiter, infertility, pregnancy, endocrine ophthalmopathy

−

Annual follow-up

+

Thyroxine treatment

Figure 4 Algorithm for the management of patients with subclinical hypothyroidism.

hormone supplementation and to prevent progression of the subclinical form to the overt stage of hypothyroidism, with its considerable morbidity and possible mortality. Thyroxine treatment should be used in patients who have elevated TSH levels greater than 10 mU/L and measurable circulating thyroid autoantibodies. Furthermore, thyroxine therapy is indicated for patients at risk with special clinical conditions such as goiter, thyroidectomy, depression, infertility, and endocrine ophthalmopathy and hypercholesterolemia, particularly in the presence of other cardiovascular risk factors such as smoking and hypertension. In patients with minimal or moderate TSH elevations ( 2.5 cm) adrenal masses that warrant discussion for excision based on their malignant potential. Adrenal Vein Sampling Although technically difficult and potentially hazardous, this remains the gold standard for diagnosis of adrenal adenomas. Given its risks, it is generally reserved for those in whom adrenal surgery is being considered, but it may be performed in others in whom there is diagnostic doubt. Diagnosis of adrenal adenoma can be made if aldosterone levels are elevated in one adrenal vein compared with the other. Simultaneous cortisol measurements ensure correct positioning of the cannula in the adrenal vein; a gradient of at least 300% between peripheral and central cortisol measurements confirms that this is the case. In summary, the diagnosis of primary aldosteronism is problematic in that there are a number of potential diagnostic tests, many of which are affected by posture, electrolyte balance, and medications. Figure 4 outlines a possible approach to the initial investigation of mineralocorticoid excess and PA. Management Aldosterone-Producing Adenoma Surgical excision in suitably fit individuals can be curative. Preoperatively, patients should be treated with spironolactone to lower blood pressure and normalize plasma potassium. Spironolactone also reduces the risk of postoperative hypoaldosteronism because it allows recovery of the RAA axis, resulting in stimulation of the previously atrophic contralateral zona glomerulosa. After removal of an aldosterone-producing adenoma, serum potassium returns to normal in 100% of cases and blood pressure is normalized in 60% within a month and 75% within a year. Laparoscopic techniques keep morbidity to a minimum. In patients who are unfit for or decline surgery, medical treatment is preferred. Idiopathic Aldosteronism The most appropriate treatment for idiopathic aldosteronism is blockade of the MR with aldosterone

250

Mineralocorticoids and Mineralocorticoid Excess Syndromes

A

B Plasma renin activity

Normal

Resistant hypertension Unprovoked hypokalemia Family history of hypertension

No investigation

Low

Plasma aldosterone

High

ARR

?Primary aldosteronism

Normal

Excluded

High Low Withdraw therapy for 2 weeks Ensure sodium intake > 200 mmol/4 days

Cortisol/ cortisone

High

Syndrome of apparent mineralocorticoid excess Supine and ambulant Renin/aldosterone Cortisol

Normal

Deoxycorticosterone

High

Normal

Excluded

High ARR

?Cause

Low

Image adrenals

?Liddle's Syndrome Adrenal lesion Consider surgery

Equivocal imaging Consider GRA ?Vein sampling

Normal imaging Consider GRA if aldo falls on ambulation consider vein sampling

Figure 4 (A) Approach to mineralocorticoid excess. (B) Investigation of primary aldosteronism.

receptor antagonists such as spironolactone. Patients may require relatively high doses (occasionally up to 400 mg/day), although many do show a good response to much lower doses. There is evidence that patients with high ARRs, who do not have distinct adenomas, respond well to a relatively low dose of spironolactone. Dose titration is limited by dose-dependent side effects, including nausea, painful gynecomastia, and menstrual irregularities. The selective aldosterone receptor antagonist (SARA) eplerenone may offer considerable advantages, although results comparing the blood pressure lowering efficacy of this with spironolactone in PA are not yet available. Alternatives include the potassium-sparing diuretics, amiloride and triamterene, which act on the distal tubule of the nephron, inhibiting sodium–potassium exchange. In practice, combination drug therapy using spironolactone or amiloride with one or more agents is

usually required for optimal blood pressure control. Additional agents include calcium channel blockers, angiotensin-converting enzyme (ACE) inhibitors, and angiotensin II antagonists. Glucocorticoid Remediable Aldosteronism Suppression of ACTH with dexamethasone provides a possible long-term therapeutic option. Patients can often be well controlled using small doses (e.g., 0.25 mg/day). Spironolactone is an alternative treatment. Other Uses of Selective Aldosterone Receptor Antagonists The occurrence of secondary aldosteronism in other circumstances, such as CCF, was discussed earlier. The concept of aldosterone-mediated cardiac injury has led to studies that explored the use of SARAs in patients with cardiac disease. The Randomized

251

Mineralocorticoids and Mineralocorticoid Excess Syndromes

Aldactone Evaluation Study (RALES) reported substantial benefit (30% reduction in mortality) in patients with advanced cardiac failure who were given spironolactone in addition to conventional treatment. More recently, the Eplerenone Post-Acute Myocardial Infarction Heart Failure Efficacy and Survival Study (EPHESUS) showed that eplerenone could provide substantial benefit (15% reduction in mortality) in patients following acute myocardial infarction. These are new and exciting data that illustrate the key importance of MR antagonism in cardiovascular pathophysiology and identify important therapeutic opportunities.

CONCLUSION p0280

PA is now recognized as the most common cause of secondary hypertension and is thought to exist in up to 10% of unselected hypertensives. Aldosteroneproducing adenomas are no longer the most common cause of PA but remain an important diagnosis as surgical removal offers the prospect of cure of hypertension. The majority of patients with PA are now thought to have bilateral adrenal hyperplasia. Development of new, more selective aldosterone receptor antagonists is the major therapeutic challenge to optimize blood pressure control and minimize side effects in this increasing number of patients.

See Also the Following Articles ACTH (Adrenocorticotropic Hormone) . Aldosterone in Congestive Heart Failure . Angiotensin, Evolution of . Atrial Natriuretic Factor and Family of Natriuretic Peptides . Catecholamines . Primary Aldosteronism (PAL) . Tissue Renin-Angiotensin-Aldosterone System

Further Reading Bubien, J. K., Ismailov, I. I., Berdiev, B. K., et al. (1996). Liddle’s syndrome: Abnormal regulation of amiloride-sensitive sodium channels by beta subunit mutation. Am. J. Physiol. 270, C208–C213. Farman, N., and Bocchi, B. (2000). Mineralocorticoid selectivity: Molecular and cellular aspects. Kidney Intl. 57, 1364–1369. Lim, P. O., Dow, E., Brennan, G., Jung, R. T., and Macdonald, T. M. (2000). High prevalence of primary aldosteronism in the Tayside hypertensive clinic population. J. Hum. Hypertension 14, 311–315. Lim, P. O., Young, W. F., and Macdonald, T. M. (2001). A review of the medical treatment of primary aldosteronism. J. Hypertension 19, 353–361. Stewart, P. M. (1999). Mineralocortcoid hypertension. Lancet 353, 1341–1347. Vallotton, M. B. (1996). Primary aldosteronism: I. Diagnosis. Clin. Endocrinol. 45, 47–52. White, P. C., Mune, T., and Agarwal, A. K. (1997). GRA: Diagnosis, variability of phenotype, and regulation of potassium homeostasis. Steroids 60, 48–51. Young, M. J., and Funder, J. W. (2002). Mineralocortcoid receptors and pathophysiological roles for aldosterone in the cardiovascular system. J. Hypertension 20, 1465–1468.

253

Mitogen-Activated Protein (MAP) Kinases and Receptors

Growth factors, cytokines CR PM

Hormones, stress, DNA damage

RTK

GPCR Ras

Raf

Mos

MEK1/2

ERK1/2

DSP

Tpl-2

MEKK

MEK4/7

MEK3/6

JNK1/2/3

Cell proliferation, differentiation, meiosis and cell cycle

p38α/β/γ/δ

Apoptosis, inflammation, cell cycle

Figure 1 Schematic of the three major MAP kinase signaling pathways. Growth factors, cytokines, and other hormones activate receptor tyrosine kinases (RTK), cytokine receptors (CR), or G protein-coupled receptors (GPCR) on the plasma membrane (PM) and downstream MAP kinase pathways. Cell stress or DNAdamaging agents may activate signaling pathways independent of a receptor. Inactivation of MAP kinases occurs through dephosphorylation by dual-specificity phosphatases (DSP). Solid and dashed lines indicate direct and indirect interactions, respectively.

MEK1 (44 kDa) and MEK2 (45 kDa) are dualspecificity kinases that can phosphorylate both the threonine and the tyrosine residues, which are separated by a glutamate residue within the tripeptide activation motif on ERKs. Although full activation of ERK proteins requires dual phosphorylation of both threonine and tyrosine residues, single phosphorylations on either of these residues may confer partial ERK protein activity. The MEK proteins are also activated through dual phosphorylation of two serine residues primarily by Raf kinases in somatic cells and the Mos kinase in germ cells. However, the MEK proteins may also be activated by other MEK kinases, including the tumor progression locus-2 (Tpl-2) serine/threonine kinase, which is involved in inflammatory responses; the MEK kinase-1, which is involved in apoptotic responses; and mixed lineage kinase-3, which may function as a promoter of proliferation or cell death depending on the extracellular stimulus. Raf kinases are activated through phosphorylation by a variety of serine/threonine and tyrosine kinases and by recruitment to the cell membrane through interactions with Ras G proteins. Additional phosphorylations of Raf proteins may result in inhibition

regulate gene expression and protein translation. Following activation, a fraction of ERK1/2 translocates to the nucleus and directly regulates gene expression by phosphorylating several transcription factors, including the p62 ternary complex factor (p62TCF; also called Elk-1), Myc, and proteins belonging to the Ets family. ERK proteins can also regulate the shape and motility of cells by phosphorylating structural proteins associated with microtubules, which are important determinants of the cell shape and architecture. Following activation, ERK proteins may also inhibit their own activity by phosphorylating the signaling protein son of sevenless (SOS), which inhibits further growth factor receptor activation of the ERK signaling pathway.

ERK Structure and Activation The ERK proteins contain both N- and C-terminal lobes with a conserved catalytic region containing the regulatory phosphorylation sites in the activation loop (Fig. 2). These regions are important for ATP nucleotide and substrate binding as well as phosphoryl transfer onto substrate proteins. Currently, the upstream kinases, MAP or ERK kinases-1 and-2 (MEK1/2), are the only known activators of ERK1/2.

Figure 2 Diagram of the 3D structure of ERK2 showing the spatial relationship of the ERK2 N- and C-terminal lobes. The regulatory threonine and tyrosine residues within the activation loop are shown as spheres on the left. The ATP-binding region resides in the N-terminal lobe. Docking domains important for substrate interactions are represented by aspartate residues as depicted by the darker spheres on the right, and threonine residues as depicted by the lighter spheres on the right are also shown. This bilobed structure is common to many kinases.

254 of Raf catalytic activity and downstream ERK pathway signaling. For example, protein kinase B (also called Akt) proteins phosphorylate a serine residue in the Raf-1 catalytic domain and inhibit Raf activation of MEK and ERK. Ras proteins are coupled to the activated membrane receptors through adaptor proteins and provide a major link between plasma membrane receptors and activation of the ERK signaling pathway. The activity of the Raf–MEK–ERK signaling module is also regulated through interactions with other binding proteins. For example, a protein called kinase suppressor of Ras may act as a scaffolding protein that binds to Raf, MEK, and ERK in a complex and determines the degree of ERK activation depending on the stimulus. In addition, a specific MEK-binding protein, MEK partner-1, may be important for directing MEK1 interactions with ERK1 but not ERK2. Lastly, a Raf-interacting protein, the Raf kinase inhibitor protein, may function by preventing Raf activation of MEK1/2. Although these and other ERK pathway binding proteins have been identified, their functional role in regulating ERK signaling events and biological functions remains largely unknown.

JNK/SAPK MAP KINASES The c-jun NH2-terminal kinase (JNK) proteins are an evolutionarily conserved family of serine/threonine protein kinases. The events leading to the identification and cloning of this family began in 1990, when a mammalian stress-activated protein kinase (SAPK) was discovered. These experiments identified a 54kDa protein kinase that was activated in cells that were treated with the protein synthesis inhibitor cycloheximide. At the same time, a protein was discovered that had affinity for the transcription factor cjun and phosphorylated c-jun on N-terminal residues when cells were exposed to ultraviolet (UV) radiation. In 1994, isolation and cloning experiments confirmed that the JNK and SAPK proteins were identical and belonged to the MAP kinase superfamily. Three isoforms, JNK-1,-2, and-3 (also known as SAPK-a, SAPK-b and SAPK-g), have been shown to exist as at least 10 alternatively spliced variants. Whereas JNK-1 and JNK-2 are ubiquitously expressed, JNK3 has a restricted expression pattern and is found mainly in the brain, heart, and testis. The JNK proteins have multifunctional roles and have been proposed to be involved in tumor development, cell growth and differentiation, apoptosis, survival, and cytokine production. JNK proteins may also

Mitogen-Activated Protein (MAP) Kinases and Receptors

mediate cardiac hypertrophic responses during hypertension and ischemia injury in the heart and kidney, and they may be involved in several neurodegenerative diseases. The JNK signaling pathway therefore represents a potential target for therapeutic intervention.

JNK Structure and Activation JNKs have a similar core structure as that of ERK proteins but there are differences in the conformation of their activation loop, resulting in differences in the mechanism of regulation. The small N-terminal lobe helps in the orientation and binding of ATP, whereas the large C-terminal lobe aids in substrate recognition. JNKs, like ERK proteins, are activated by phosphorylation on threonine and tyrosine residues, which are separated by a proline within the tripeptide activation loop in the kinase domain. Mammalian JNK proteins are activated by various extracellular stimuli, including growth factors, cytokines, and cellular stresses such as heat shock, hyperosmolarity, UV radiation, and ischemia/reperfusion. Although the stress-mediated activation mechanism is not clear, it has been hypothesized that stress factors induce receptor clustering and internalization, which lead to JNK activation. The organization of the JNK activation cascade is conserved between the other MAP kinase members, beginning with receptor activation and followed by recruitment of adaptor molecules and activation of small GTP-binding proteins. This is followed by three to four tiers of dual-specificity protein kinases, which culminate in the activation of the JNK proteins. Similar to the ERK proteins, JNK proteins target a variety of substrates and can translocate to the nucleus, where they regulate gene expression by phosphorylating transcription factors. JNK proteins are activated through dual phosphorylation on threonine and tyrosine residues by MEK4 (also called SEK1 or MKK4) or MEK7 (also called MKK7). Three MEK4 and six MEK7 isoforms have been identified. These different isoforms demonstrate selectivity depending on the extracellular stimulus. For example, MEK7 is primarily activated by cytokines such as tumor necrosis factor (TNF) and interleukins (ILs). In contrast, MEK4 is primarily activated by environmental stress stimuli such as osmotic changes and DNA-damaging agents. Although both MEK4 and MEK7 proteins can activate JNK proteins, MEK4 can also activate the p38 MAP kinases. Another difference is in the specificity of the two MEKs toward the threonine and tyrosine residues within the activation site of JNK; MEK4

255

Mitogen-Activated Protein (MAP) Kinases and Receptors

preferentially phosphorylates the tyrosine residue, whereas MEK7 has a higher affinity toward phosphorylating the threonine residue on JNK. The JNK-specific MEK proteins are activated by dual phosphorylation of a serine and threonine residue in their activation loop, and they are found in both the nucleus and the cytoplasm. Thus, the JNK proteins may be activated in both the nucleus and the cytoplasm. A wide variety of upstream kinases phosphorylate the MEK proteins responsible for activating both the JNK and p38 MAP kinases. Some of these kinases include proteins belonging to the MEK kinases (MEKK1–4), the mixed lineage protein kinases (MLK1–3, DLK, and LZK), the apoptotic stimulating kinase (ASK1 and-2), and transforming growth factor-b (TGF-b)-activating kinase (TAK1) and Tpl2, which also activate MEK1/2. These proteins are in turn activated by the p21-activated kinases (Paks), the germinal center kinases (GCKs), and the hematopoietic progenitor kinases (HPKs). The activity of Paks, GCKs, and HPKs is regulated by G proteins in a manner analogous to Ras G protein activation of Raf kinases in the ERK MAP kinase pathway. In the case of the JNK and p38 MAP kinases, members of the Rho family of G proteins are involved in the initiation of the signaling cascades. Similar to the ERK MAP kinases, several JNK pathway binding proteins have been identified that function in regulating the assembly and activation of the JNK pathway. Two such proteins have been identified as JNK-interacting proteins (JIPs), which lack enzymatic activity but act as important organizers of JNK pathway complexes. JIP1 and JIP2 are closely related proteins that can bind to JNK proteins, MEK7, and MLK proteins. JIP proteins may be involved in potentiating JNK activation in response to activation by MLK proteins.

P38 MAP KINASES The p38 MAP kinase [also known as the cytokine suppressive anti-inflammatory drug-binding protein (CSBP), reactivating kinase (RK), or SAPK] pathway is responsible for mounting a cellular response to many types of stress signals. Stress stimuli that activate p38 MAP kinase include osmotic and temperature shock, proinflammatory cytokines, hypoxia, reactive oxygen species, and irradiation. In some circumstances, p38 may be activated in response to certain growth factors. There are at least six isoforms of p38 that have alternative names, which are given in parentheses: p38 a1/a2 (CSBP1/2, Mpk2, RK, and

SAPK2a), p38 b1/b2 (SAPK2b), p38 g (ERK6 and SAPK3), and p38 d (SAPK4). The p38 proteins a1, b1, g, and d are encoded by four separate genes, whereas the a2 and b2 isoforms are the result of alternative splicing of the messenger RNA of the a1 and b1 isoforms. Although all of these isoforms are similar enough to be considered members of the p38 family, they are expressed at different levels depending on cell type and have different affinities for p38 substrates. Thus, expression of these isoforms in varying amounts in different cell types may allow cells and tissues to fine-tune their responses to various stimuli.

p38 Structure and Activation The topological structure of p38 MAP kinase is similar to that of the ERK proteins, although some differences exist in the activation loop, substrate-binding regions, and the ATP binding site, which helps explain the differences in activation and regulation between the various MAP kinases. The direct activators of p38 MAP kinases, through dual phosphorylation of threonine and tyrosine residues, include primarily MEK3 and-6 and, in some cases, MEK4. Activators of MEK3, -4, and -6 were previously described. Like the ERK and JNK MAP kinase pathways, activation of p38 MAP kinase proteins occurs through a kinase cascade, which is often initiated by activated membrane receptors. Upon activation, p38 MAPK phosphorylates other kinases in the cytosol and translocates to the nucleus, where it phosphorylates and activates transcription factors as well as proteins that modify the topological structure of DNA. Downstream targets of p38 MAP kinases regulate the expression of genes responsible for the inflammatory response, including cytokines such as TNF-a, IL-1b, and IL-6. Because of its key role as a regulator of the inflammatory response, p38 is currently the target of anti-inflammatory drugs designed to treat acute and chronic inflammatory diseases, such as rheumatoid arthritis and osteoarthritis.

MAP KINASE ACTIVATION BY RECEPTOR TYROSINE KINASES The ERK MAP kinase pathway is activated by several growth factor ligands that stimulate RTK activity, including epidermal growth factor (EGF), fibroblast growth factor, hepatocyte growth factor, insulin or insulin-like growth factors, platelet-derived growth factor (PDGF), and vascular endothelial growth factor. Ligand binding induces receptor dimerization

p0075

256 and the activation of intrinsic tyrosine kinase activity, which causes autophosphorylation of tyrosine residues in the cytoplasmic domain. The phosphorylated tyrosine residues of the receptors provide docking sites for Src homology-2 (SH2) domain-containing adaptor proteins, such as SH2-containing collagenrelated proteins and growth factor receptor-bound protein-2 (Grb2). Grb2 recruits the guanine nucleotide exchange factor protein SOS, which promotes the active GTP-bound form of Ras and subsequent activation of the MAP kinase signaling pathways. The JNK and p38 MAP kinases are activated mainly under cellular stress conditions, such as heat shock, irradiation, or hyperosmolarity. Although it is not clear which mechanisms are involved in mediating the stress response and activation of the JNK and p38 pathways, it has been proposed that cell stress may induce membrane receptor clustering and internalization, which could facilitate receptor activation. Nonetheless, similar to the ERK proteins, activation of the JNK and p38 MAP kinases pathways may also be through RTKs or cytokines such as TNF and IL-1b. RTK stimulation of Ras G protein, in addition to activating Raf-1 kinase and the ERK pathway, may activate MEK kinases involved in activating the JNK and p38 pathways.

ACTIVATION OF MAP KINASES THROUGH G PROTEIN-COUPLED RECEPTORS G protein-coupled receptors (GPCRs) are integral membrane proteins that contain regions that pass through the membrane seven times; thus, the GPCRs are also referred to as the seven-transmembrane spanning or serpentine receptors. GPCRs are coupled to heterotrimeric G proteins that contain a, b, and g subunits that determine which signaling pathways will be used. The G protein a subunit binds to GTP and contains the GTP-hydrolyzing activity, whereas the b and g subunits contain regulatory information. A wide variety of GPCRs, including the adenosine A1, a-adrenergic, and muscarinic acetylcholine receptors, as well as endothelin, angiotensin, histamine, glucagon, and thrombin receptors, initiate physiological responses through activation of MAP kinase pathways. GPCR activation of MAP kinases may occur through G protein activation of phospholipases, which hydrolyze membrane phospholipids and generate second messengers such as diacylglycerol (DAG) and inositol triphosphate (IP3). IP3 stimulation of

Mitogen-Activated Protein (MAP) Kinases and Receptors

intracellular Ca2+ and DAG can activate protein kinase C, which phosphorylates and activates Raf kinase. Alternatively, activated GPCRs may transactivate RTKs, which feed into downstream MAP kinase signaling pathways. For example, MAP kinase activation by endothelin or thrombin receptors is coupled to coactivation of the EGF receptor family of RTKs, and the RTK activity is required for endothelin- or thrombin-induced MAP kinase activity. Other mechanisms for activating MAP kinase pathways may be through the G protein b and g subunits, which can activate nonreceptor tyrosine kinases belonging to the Src protein family. Src kinases can phosphorylate tyrosine residues on the cytoplasmic domains of receptors and enhance Ras activity or directly phosphorylate and regulate Raf kinase activation. Activation of JNK and p38 MAP kinases by GPCRs has been shown to involve the Ga and the Gbg subunits as well as the activation of the Rho and Ras families of GTPases. Depending on the cell type and the receptor stimulated, Ga or Gbg or both may be involved in the activation of the MAP kinase pathways.

OTHER MECHANISMS FOR ACTIVATING MAP KINASES Activation of the MAP kinase signaling pathways may occur through cell-permeable factors such as steroid hormones, which bind to soluble cytoplasmic and nuclear receptors. For example, estradiol, progesterone, and testosterone activate MAP kinase indirectly through steroid receptor interactions with membrane-associated receptors. One proposed mechanism is that estradiol receptors are coupled to membranebound receptors such as the EGF receptor, which then facilitate the activation of the MAP kinase signaling pathway in a manner similar to transactivation by GPCRs described previously. Moreover, GPCRs may mediate steroid hormone activation of RTK and MAP kinase signaling. Some membrane-bound receptors belonging to the receptor serine/threonine kinase (RSTK) family that undergo serine or threonine autophosphorylation on the cytoplasmic domains regulate MAP kinase signaling pathways. For example, the TGF-b receptors are RSTKs that regulate a wide variety of biological responses, including cell proliferation, differentiation, extracellular matrix production, and cell death, through the p38 MAP kinases. Activated TGF-b receptor stimulates TAK1, which is a novel MEK kinase that stimulates MEK3/6 and p38 and may inhibit cell proliferation.

257

Mitogen-Activated Protein (MAP) Kinases and Receptors

MAP KINASE PHOSPHATASES Inactivation of MAP kinase signaling pathway is accomplished largely through dephosphorylation of threonine and tyrosine residues by unique protein tyrosine phosphatases called dual-specificity phosphatases (DSPs). Of the DSP family, at least six MAP kinase phosphatases (MKPs) have been identified. The MKP proteins show some specificity toward the MAP kinase family members. For example, although MKP1 and MKP2 can target activated ERK, JNK, and p38, MKP3 and MKP4 are specific toward ERK proteins, and MKP5 and MKP6 specifically dephosphorylate JNK and p38. Growth factor activation of the ERK proteins stimulates the expression of MKP1, thus providing a mechanism by which ERK proteins downregulate their own activity. Other phosphatases belonging to the serine/threonine phosphatases may also play an important role in down-regulating MAP kinases. For example, protein phosphatase 2A (PP2A) is able to inhibit MEK1/2 and ERK1/2 activity in the absence of growth factor-induced DSPs. Similarly, PP2A has been identified as a JNK phosphatase. MAP kinase pathways are also regulated by protein serine/threonine phosphatases of the type 2C (PP2C) family. At least two members of the PP2C family, PP2Ca and-b, dephosphorylate and inactivate the JNK MAP kinases.

CONCLUSION The MAP kinase signaling pathways regulate most physiological functions throughout the life span of a diverse range of organisms, from yeast to humans.

Activation of MAP kinases occurs through plasma membrane receptor-dependent and -independent mechanisms in response to a variety of extracellular stimuli. Importantly, dysregulation of MAP kinase pathways is involved in many human diseases. Therefore, understanding of how MAP kinases are regulated and function is an important goal for the development of new therapies.

See Also the Following Articles Janus Kinases and Cytokine Receptors . Lipid Second Messengers and Receptors . Receptor Serine/Threonine Kinases

Further Reading Cobb, M. H., and Goldsmith, E. J. (1995). How MAP kinases are regulated. J. Biol. Chem. 270, 14843–14846. Davis, R. J. (2000). Signal transduction by the JNK group of MAP kinases. Cell 103, 239–252. Keyse, S. M. (2000). Protein phosphatases and the regulation of mitogen-activated protein kinase signaling. Curr. Opin. Cell. Biol. 12, 186–192. Lewis, T. S., Shapiro, P. S., and Ahn, N. G. (1998). Signal tranduction through MAP kinase cascades. Adv. Cancer Res. 74, 49–139. Lowes, V. L., Ip, N. Y., and Wong, Y. H. (2002). Integration of signals from receptor tyrosine kinases and G protein-coupled receptors. Neurosignals 11, 5–19. Luttrell, L. M., van Biesen, T., Hawes, B. E., Koch, W. J., Krueger, K. M., Touhara, K., and Lefkowitz, R. J. (1997). G-proteincoupled receptors and their regulation: Activation of the MAP kinase signaling pathway by G-protein-coupled receptors. Adv. Second Messenger Phosphoprotein Res. 31, 263–277. Ray, L. B., and Sturgill, T. W. (1987). Rapid stimulation by insulin of a serine/threonine kinase in 3T3-L1 adipocytes that phosphorylates microtubule-associated protein 2 in vitro. Proc. Natl. Acad. Sci. USA 84, 1502–1506.

259

Mixed Lipemias

p0030

p0035

particles that contain apo’s B100, CII, CIII, and E. After coupling to VLDL receptors in extrahepatic tissues, the triglycerides in VLDL undergo lipolysis by lipoprotein lipase (LPL), which requires apoCII as a cofactor. The liberated free fatty acids (FFA) are taken up by muscle or adipose tissue (through CD36) for oxidation or energy storage, respectively. In this process, VLDL remnants become intermediate density lipoproteins (IDL) that are enriched with apoE and then low-density lipoproteins (LDL). The latter step is also mediated by hepatic lipoprotein lipase (HL), which is attached to the endothelial surface in the liver by heparan sulphate proteoglycans (HSPG) and is facilitated by apoE. VLDL remnants and IDL are removed from the circulation by the lowdensity lipoprotein receptor (LDLR) and LDLRrelated protein (LRP), largely through recognition of apoE. LDL is removed by the LDLR through recognition of apoB100. In the exogenous pathway, dietary cholesteryl esters and triglycerides are secreted from the intestines in chylomicrons containing apoB48, A-I, and A-IV. They lose A-I and A-IV but acquire apo’s CII, CIII, and E. As in VLDL, chylomicron triglycerides are hydrolyzed by LPL in the circulation and the chylomicron remnants are removed from the circulation by LDLR and LRP mainly in the liver through recognition of apoE. In the reverse cholesterol transport pathway, cholesterol is removed from peripheral tissues by high-density lipoprotein (HDL), which binds to cholesterol-rich cells, at least partly through the SRB1 receptor. The uptake of cholesterol from these cells by HDL is facilitated by the recently characterized ABC1 transporter. HDL-associated cholesterol is then esterified by plasma lecithin cholesterol acyltransferase (LCAT), enlarging the size of the HDL particle. Cholesterol from HDL can be transported into the liver through the action of HL, by the action of the SRB1 receptor, or by the uptake of entire HDL particles. Alternatively, cholesteryl ester transfer protein (CETP) mediates HDL cholesteryl ester to triglyceride-rich particles instead of triglycerides, which are then removed from the circulation by liver LDLR and LRP. In the liver, cholesteryl ester is hydrolyzed and the cholesterol is excreted in the bile as bile acids or free cholesterol.

LIPID-LOWERING DRUGS p0040

Hydroxy-methylglutaryl coenzyme A (HMG–CoA) reductase inhibitors inhibit the rate-limiting step in hepatic cholesterol synthesis, causing an increase in LDL receptor levels in hepatocytes and enhancement

of remnant and LDL cholesterol removal from the circulation. In addition to lowering LDL cholesterol, they lower triglyceride levels (by decreased hepatic apoB production) and increase HDL cholesterol. Large intervention trials have demonstrated the potency of these drugs in primary or secondary prevention of cardiovascular disease in patients with or without lipid disorders. Fibric acids stimulate the activity of the liver transcription factor PPARa that increases LPL activity, thereby enhancing VLDL degradation. PPARa stimulation also reduces apoCIII, which facilitates VLDL remnant uptake. VLDL production may also be reduced. Stimulation of peroxisomal fatty acid oxidation by fibrates may also contribute to the triglyceride-lowering actions. In addition to reducing triglycerides, fibrates increase HDL cholesterol levels. In several intervention trials, fibrates have been demonstrated to reduce CHD events, but conflicting data on mortality are reported. Nicotinic acid derivatives decrease both total and LDL cholesterol, reduce VLDL cholesterol levels, and raise HDL cholesterol levels. The mechanism of action is not fully understood, but it appears to inhibit the secretion of apoB-containing lipoproteins from the liver as well as lipolysis in peripheral adipose tissue. In the Coronary Drug Project, niacin significantly reduced CHD events.

PRIMARY MIXED LIPEMIAS Type III Hyperlipoproteinemia (Familial Dysbetalipoproteinemia) Type III hyperlipoproteinemia is a lipoprotein disorder with a prevalence of 1 to 4 in 10,000. The disorder is usually diagnosed in adults and has a male predominance. Individuals have increased plasma cholesterol and triglyceride levels resulting from increased triglyceride-rich remnant concentrations. Plasma LDL levels are usually reduced, and HDL levels are usually normal. A typical clinical feature is the presence of planar xanthomas at the palmar surface. Patients with familial type III hyperlipoproteinemia have increased susceptibility to CHD and peripheral vascular disease. The typical disorder is caused by a mutation in the apoE gene. The apoE gene has three alleles coding for E4, E3, and E2 that occur in Caucasians with frequencies of 15, 77, and 8%, respectively. The apoE polymorphisms account for approximately 10% of the population variance in LDL cholesterol levels. More than 95% of type III hyperlipoproteinemia individuals have the E2/E2

s0025

260 phenotype. This common E2 mutation, representing an Arg158Cys mutation, causes a nearly total defect of apoE in LDLR binding. The result is a less efficient clearance of VLDL and chylomicron remnants in the liver, leading to increased triglyceride and cholesterol levels. LDL is decreased as a result of the decreased conversion of IDL to LDL. Individuals with heterozygosity for the E2 mutation do not exhibit the phenotype, indicating that the disease is inherited in an autosomal-recessive fashion. Interestingly, only 1 to 4% of individuals with the E2/E2 genotype have the typical phenotype. Hence, type III hyperlipoproteinemia with E2/E2 homozygosity displays a low rate of penetrance. Additional factors that are associated with increased VLDL or chylomicron production or with decreased clearance, such as gender, obesity, hypothyroidism, diabetes mellitus and certain dietary factors, are thought to contribute to manifesting the disease. The autosomal-dominant inheritance pattern of type III hyperlipoproteinemia has been described for some rare apoE variants. For instance, individuals with heterozygosity for apoE2Lys146Gln or apoE3Leiden exhibit the phenotype. Nearly all carriers of these rare alleles have the hyperlipoproteinemic phenotype, indicating that this dominant inheritance pattern associates with a high rate of penetrance. In the treatment of this disorder, a thorough search for underlying disorders should be made. Usually, dietary measures (restriction of calories) may normalize lipid levels. If this cannot be achieved, drug therapy may be necessary. HMG–CoA reductase inhibitors have been applied successfully, and fibric acid derivatives may also be useful.

Familial Combined Hyperlipidemia p0065

Familial combined hyperlipidemia (FCH) is a common lipid disorder affecting 0.5 to 2.0% of the population. The disease occurs in both adults and children. Biochemically, FCH is characterized by familial occurrence of increased cholesterol and triglyceride levels. The FCH lipoprotein pattern includes elevated apoB and LDL cholesterol levels, diminished HDL cholesterol levels, and the presence of small dense LDL. The phenotype may vary within one affected family and even within one individual. Xanthomas are not observed in FCH. Patients with FCH have an increased risk for CHD. FCH is associated with increased body mass index (BMI), waist/hip ratio, and fasting glucose and insulin levels, which are (together with the lipid abnormalities) characteristics of the metabolic syndrome (syndrome X). Thus,

Mixed Lipemias

knowledge of the pathogenesis of FCH will provide insight into the metabolic syndrome and vice versa. Initially, the dominant mode of inheritance was thought to indicate a monogenetic cause. Several candidate genes or chromosomal regions have been suggested. The apoA-I/CIII/A-IV gene cluster, LCAT (chromosome 16q22.1), and 11p were found in Dutch FCH families. However, this could not be confirmed in Finnish FCH families. In contrast, Finnish studies identified candidate regions on chromosomes 1q, 2q, 10p, 10q, and 21q, but in turn, the strong chromosome 1 locus was not found in the Dutch families. These varying results suggest that FCH is a complex disorder where many genes are involved and a large heterogeneity in gene–gene and gene–environment interactions is present. In the treatment, dietary and lifestyle interventions should be attempted and may be combined with drugs when necessary. Drug therapy should be aimed at the predominating lipid disorder. HMG–CoA reductase inhibitors are a logical choice, and fibric acid derivatives and niacin may be useful as well. Bile acid sequestrants should be avoided because they tend to raise triglyceride levels.

SECONDARY MIXED LIPEMIAS Diabetic Dyslipidemia Worldwide, diabetes mellitus (DM), and especially type 2 DM, is a rapidly expanding health problem, with estimates suggesting a doubling of its prevalence during the coming years. Cardiovascular disorders are the major cause of mortality in DM. Cardiovascular disorders, in turn, are considered the result of the combined effects of hypertension and alterations in lipid and carbohydrate metabolism, the fibrinolytic system, and inflammatory cascades. The importance of lipid abnormalities in DM is underscored by the fact that in large studies of type 1 and type 2 DM, optimization of glycemic control could prevent microvascular complications but not macrovascular ones. In contrast, in large lipid intervention studies, subgroup analyses revealed favorable effects of lipidlowering drugs on CHD risk. Lipoprotein disorders in DM are related primarily to abnormalities in triglyceride metabolism, and this may be explained in part by the role of insulin in lipid metabolism. In the absence of insulin action (e.g., insulin deficiency in untreated type 1 DM or insulin resistance), lipolysis from adipose tissue is increased by the action of hormone-sensitive lipase. The increased FFA flux to the liver and the decreased degradation of apoB, also a

261

Mixed Lipemias

result of deceased insulin action, lead to enhanced VLDL production. As insulin facilitates LPL activity, VLDL (and chylomicron) remnant clearance is decreased when insulin action is reduced, thereby further increasing triglyceride (and VLDL-derived) cholesterol levels. Although LDL levels are expected to be lower (due to decreased conversion of VLDL remnants), LDL clearance may also be decreased. In the presence of triglyceride-rich lipoproteins, enhanced CETP activity may result in a net transfer of cholesteryl ester from LDL and HDL to VLDL remnants. As a result, LDL particles become smaller and denser, and this is thought to be unfavorable with respect to atherosclerosis. The removal of cholesteryl ester from HDL leads to reduced plasma HDL cholesterol measurements. In addition, HDL clearance in the liver is enhanced. Another factor in decreased HDL is the diminished transport of surface lipids from VLDL to nascent HDL caused by decreased LPL activity; this impairs HDL maturation. Decreased HDL cholesterol is associated with increased risk for CHD. Lipoprotein abnormalities in type 1 DM are the result of insulin deficiency and can be corrected completely by insulin therapy. The situation in type 2 DM is complex; in contrast to type 1 DM, lipid abnormalities usually are not corrected completely with glycemic control. Moreover, this dyslipidemia is often found in patients with insulin resistance without overt diabetes and is one of the features of the metabolic syndrome. Abnormalities in insulin action rather than hyperglycemia are associated with this lipid abnormality, and evidence suggests that a pathological FFA flux from visceral fat to the liver even precedes defective insulin action. Given the high CHD risk in DM, aggressive lipid-lowering therapy is of major importance. Apart from dietary intervention, from a theoretical point of view, HMG–CoA reductase inhibitors, fibrates, and nicotinic acid are rational. Because HMG–CoA reductase inhibitors have been demonstrated to reduce CHD risk in DM patients in large randomized studies, this category of drugs should be the first choice.

Nephrotic Syndrome The nephrotic syndrome is almost always accompanied by hyperlipidemia. Plasma VLDL, IDL and LDL cholesterol, and total triglycerides may be increased, whereas HDL cholesterol is decreased. Mortality from CHD is particularly high in the nephrotic syndrome. The increased VLDL and IDL concentrations

may result from decreased clearance due to reduced LPL activity on the vascular endothelium that, in turn, may result from either decreased synthesis or inadequate binding of this enzyme by HSPG to endothelial surfaces. The mechanism of the increased LDL levels is not clear. Synthesis of apoB100 is not related to that of albumin, suggesting a mechanism different from the increased synthesis of nonlipoproteins. HDL concentrations in the nephrotic syndrome are normal, but maturation is impaired, leading to a shift from the larger HDL2 to the smaller HDL3 and resulting in decreased plasma HDL cholesterol concentrations. Because of the high CHD risk, lipid-lowering therapy—with HMG–CoA reductase inhibitors as first choice—must have high priority.

Hypothyroidism Because thyroid hormone plays a role in the LDLR expression, clinical or subclinical hypothyroidism is associated with decreased LDL clearance and, consequently, increased LDL cholesterol levels; few patients with hypothyroidism have normal lipid profiles. In addition, hypothyroidism is present in approximately 5% of patients who present primarily with lipid disorders. Hypothyroidism may also be associated with increased triglyceride levels, and this is thought to be the result of decreased LPL activity. Whether hyperlipidemia in hypothyroidism is associated with increased risk for CHD is still being debated. Substitution with thyroid hormone usually corrects the lipid abnormalities.

Other Conditions Hypopituitarism In patients with hypopituitarism, mixed lipemia may be present. This is likely the result of the combined contribution of hypothyroidism and growth hormone deficiency, the latter of which is accompanied by increased LDL cholesterol levels. Glucocorticoid Excess Excess of glucocorticoids, either in the context of Cushing’s syndrome or by steroid hormone therapy, may be associated with increased LDL and VLDL concentrations. Increased VLDL production and increased conversion of VLDL to LDL appear to play a role. Because glucocorticoid excess is also associated with impaired glucose tolerance, lipoprotein abnormalities (as observed in DM) may also be present.

262 Cushing’s syndrome is associated with increased CHD risk; however, the independent contribution of dyslipidemia has not been identified.

Human Immunodeficiency Virus Patients with human immunodeficiency virus (HIV) infection who are treated with HIV-1 protease inhibitors can develop hyperglycemia, hypertriglyceridemia, or hypercholesterolemia. The mechanism appears to be multifactorial, but data indicate that this effect may be at least partly accounted for by decreased degradation of apo-B and, hence, increased VLDL synthesis. Decreased PPARa activity may lead to effects opposed to those of fibrates, whereas accumulation of the active portion of sterol regulatory element-binding protein-1c are also involved.

See Also the Following Articles Diabetes, Type 1 . Diabetes, Type 2 . Dysbetalipoproteinemia and Type III Hyperlipidemia . Familial Low Cholesterol Syndromes, Hypobetalipoproteinemia . Hypertriglyceridemia . Hypopituitarism . Hypothyroidism, Subclinical . Lipoprotein(a) . Low HDL/High HDL Syndromes

Mixed Lipemias

Further Reading Aouizerat, B. E., Allayee, H., Bodnar, J., Krass, K. L., Peltonen, L., deBruin, T. W. A., and Rotter, J. I. (1999). Novel genes for familial combined hyperlipidemia. Curr. Opin. Lipidol. 10, 113–122. Breslow, J. L. (2000). Genetics of lipoprotein abnormalities associated with coronary heart disease susceptibility. Annu. Rev. Genet. 34, 233–254. Faergeman, O. (2000). Hypertriglyceridemia and the fibrate trials. Curr. Opin. Lipidol. 11, 609–614. Goldberg, I. J. (2001). Diabetic dyslipidemia: Causes and consequences. J. Clin. Endocrinol. Metab. 86, 965–971. Gotto, A., and Pownall, H. (1999). ‘‘Manual of Lipid Disorders.’’ Williams & Wilkins, Baltimore, MD. Mahley, R., and Rall, S. C. (1995). Type III hyperlipoproteinemia: The role of apolipoprotein E in normal and abnormal lipoprotein metabolism. In ‘‘The Metabolic and Molecular Bases of Inhereted Disease’’ (C. Scriver, A. L. Beaudet, W. S. Sly, and D. Valle, eds.), pp. 1953–1980. McGraw–Hill, New York. Steiner, G. (2000). Diabetes, lipids, and coronary heart disease: What have we learned from lipid lowering trials? In ‘‘Lipids and Vascular Disease’’ (D. J. Betteridge, ed.), pp. 159–172. Martin Dunitz, London. Thompson, G. R., and Barterba, P. J. (2000). Therapeutic approaches to reducing the LDL- and HDL-associated risks of coronary heart disease. Curr. Opin. Lipidol. 11, 567–570. Wilson, P. W., Schaefer, E. J., Larson, M. G., and Ordovas, J. M. (1996). Apolipoprotein E alleles and risk of coronary disease: A meta-analysis. Arterioscl. Thromb. Vascular Biol. 16, 1250–1255.

264 in Merck’s laboratories followed a different strategy by first cloning and sequencing a large variety of G protein-coupled receptors and then looking at the nature of ligands for these receptors. Starting from an orphan receptor found in the thyroid with a structure close to that of the receptor for the growth hormone secretagogue (GHS), mass screening of more than 500 peptide and nonpeptide molecules was performed to obtain positive results for motilin. Much of our knowledge of the motilin receptor has been derived from structure-activity studies of peptide analogues and binding of radioactive ligands on various tissue membranes prepared in vitro. Nuclear magnetic resonance showed that the motilin molecule (Fig. 1) is shaped like a golf club or pastoral staff. Its N-terminal portion is responsible for receptor binding and in vitro bioactivity. Peptide fragments containing the first 12 amino acids of the molecule have demonstrated full binding capacity in vitro, whereas removal of its first amino acid is enough to abolish its activity. Amino acids 4 and 7 also seem to play a major role in this N-terminal sequence. Interestingly, the C-terminal structure, not required for

Motilin

in vitro bioactivity, was found to be essential for in vivo bioactivity. N-terminal fragments of 15 amino acids, fully active in vitro, were indeed inactive in vivo, whereas longer fragments with 19, 20, or 21 amino acids could mimic the motor activity of the 22-amino acid peptide. This is an interesting concept in the field of peptide pharmacology, where the a-helix-shaped C-terminal structure seems to be important in protecting the whole molecule, probably against circulating degrading enzymes. Receptors have been studied in membrane solutions prepared from various species (mostly humans or rabbits) and from neural or muscle elements of the intestinal wall. The binding of motilin analogues was different in membranes prepared from the human or rabbit antrum, suggesting that motilin receptors’ structure is, as for the motilin substance discussed previously, different among species. Solutions enriched in muscle or in neural elements of the intestinal wall have allowed the identification of specific and distinct responses to various motilin synthetic analogues, indicating the existence of structurally heterogeneous motilin receptor subtypes specific to muscle or neural elements (M and N receptor subtypes).

PHARMACOLOGICAL ACTION

Figure 1 Structure of the 22-amino acid motilin peptide revealed by nuclear magnetic resonance. The N-terminal curved sequence binds to the receptor, whereas the a-helix C-terminal structure protects the molecule for in vivo bioactivity.

As its name implies, motilin acts on digestive motility. Figure 2 shows the expected effect in humans after the administration of motilin or motilin receptor agonists. Stimulation of antro-duodenal motility remains the dominant action of the peptide. Initial in vitro studies with intestinal tissues from rabbits and humans revealed that motilin can stimulate smooth muscle contraction by a direct effect on muscle cells; the peptide action was seen despite the addition of all neural blockers, including tetrodotoxin. Experiments with isolated muscle cells have confirmed the presence of motilin receptors on muscle cell membranes. However, in vivo studies in dogs and humans clearly indicated that the motor action of motilin is mediated by muscarinic transmitters. As discussed previously, binding experiments supported the existence of motilin receptor subtypes (M and N) that are functionally and structurally different. Currently, the most interesting hypothesis proposes that the interdigestive motor action of motilin (induction of phase III contraction) is elicited at low doses through an action via neural cholinergic receptors, whereas postprandial stimulation of antral motility is evoked at higher doses via muscle receptors.

265

Motilin

Figure 2 Biological effects obtained in the human gastrointestinal tract in response to the injection of motilin or a motilin receptor agonist such as erythromycin.

PHYSIOLOGICAL ROLE The characteristic action of motilin is the induction of phase III contraction of the migrating motor complex (MMC). The MMC is the basic organization of motor activity of the gut during the fasting interdigestive period. It lasts 80 to 120 min and consists of three successive phases. During phase I, no significant contraction is seen for 20 to 60 min. During phase II, intermittent and irregular contractions start to occur 20 to 60 min before phase III, during which strong peristaltic contractions, lasting 3 to 10 min, start from the stomach and lower esophagus to migrate distally to the duodenum, jejunum, and ileum until the colon. This phase III peristaltic wave has been proposed to clean, from the GI tract, bacteria or nutrients that could accumulate during the digestive period, with deleterious effects on gut (e.g., bacterial overgrowth). Feeding interrupts the intermittent cyclical fasting motility (MMC) and induces a more constant motor activity of moderate amplitude to allow the optimal absorption of ingested nutrients; accelerated transit during phase III would impair nutrient absorption. Regulatory peptides can exert their influence via different pathways: endocrine (hormonal), neurocrine, paracrine, or autocrine. Motilin is in the unique position, as a GI regulatory peptide, of being an interdigestive hormone. This has been well identified in the dog, where all of Morton Grossman’s criteria establishing the endocrine contributions of peptides have been fulfilled. First, regulation of the cyclical pattern of the MMC is under the control of circulating

factors, as shown in various experiments where the MMC persisted in animals when their stomachs had been completely denervated. Second, there is a perfect correlation between circulating peak levels of motilin and the initiation of phase III contractions from the stomach or proximal duodenum. Third, exogenous motilin, given in small doses reproducing physiological plasma variations, induces phase III contraction of the MMC. Fourth, inhibition of circulating motilin by the administration of specific motilin antisera blocked phase III contractions from the upper gut. Although the situation could not be explored under such perfect experimental conditions in humans, most evidence suggests that circulating motilin also plays a key role in the regulation of phase III initiated from the antrum of our species.

RELEASE MECHANISMS Most gastrointestinal hormones are released after a meal to allow or facilitate the digestion and absorption of nutrients. Motilin is a unique hormone. It is released periodically during the interdigestive fasting period, and its cyclical release is abolished after a meal, as shown schematically in Fig. 3. Therefore, a ‘‘biological clock’’ somewhere in the organism periodically signals motilin cells to release the peptide into the circulation. In vitro preparations of intestinal mucosal cells enriched in motilin cells showed that muscarinic receptors are present on the motilin cell membrane and that protein kinase C activators are the most potent

266

Figure 3 Schematic representation of plasma motilin variations in dog. During the fasting interdigestive period, motilin is released cyclically every 80 to 120 min (lower panel) to induce phase III contraction of the migrating motor complex from the stomach to the ileum (indicated by clear boxes in the upper panel). After eating, the motilin cyclical peak increases are abolished for 2 to 8 h (depending on the content and nature of the meal) while the fed pattern motility profile is taking place.

second messengers eliciting motilin release. In the ex vivo perfused canine intestine, bombesin has been identified as a direct stimulant of motilin release, whereas the effect of opiates is mediated by acetylcholine. Phenylephrine and somatostatin appear to act on M-cell membrane receptors to block release of the peptide. In humans, meal ingestion is followed by a very early and brief increase in plasma motilin before the interdigestive release cycle is interrupted. This early release can be mimicked by central stimulation with modified sham-feeding and by distension of the fundus with an air-filled balloon. The contribution of this postprandial motilin release (not present in the dog) in the process of nutrient digestion remains to be characterized.

CONTRIBUTION TO CLINICAL MEDICINE Some GI hormones, such as gastrin and vasointestinal polypeptide (VIP), are important to clinicians because of the disease symptoms they can generate. Up to now, there is no clinical phenotype attributed to motilin hypersecretion. High levels of circulating motilin have been documented in some patients with pancreatic tumors and Zollinger-Ellison syndrome as well as in patients with carcinoi¨d tumors of the gut. Although it could be tempting to speculate on the

Motilin

role of motilin in the diarrhea found in these patients, its contribution to the biological alterations remains unknown. Because motilin hypersecretion could be expected to generate GI hypermotricity and hypersecretion with probable diarrhea, it is logical to expect that hypomotilinemia will induce GI hypomotility. Some investigators have indeed found low levels of plasma motilin in patients with idiopathic intestinal pseudo-obstruction and idiopathic or postoperative gastroparesis. To this point, however, plasma motilin measurement has not been shown to be useful for inclusion in the workup diagnoses of any clinical situations. On the other hand, motilin was recently of major interest to medical clinicians because of the capacity of motilin receptor agonists to act as powerful stimulants of GI motor activity in patients with hypokinetic disorders. Itoh was the first to observe that erythromycin could mimic the motor effect of motilin when injected in dogs. It was soon established that erythromycin was in fact acting on motilin receptors, and Janssens and colleagues made the capital observation that erythromycin was the most potent gastrokinetic ever tested to stimulate gastric emptying in diabetic patients with gastroparesis. Since then, erythromycin, whether administered intravenously (i.v.) or by mouth (p.o.), is used by many clinicians for the treatment of patients with gastroparesis or intestinal pseudo-obstruction. Motilides, or motilin receptor agonist substances derived from the erythromycin macrolide and with improved gastrokinetic activity but devoid of antibiotic properties, have been developed by many pharmaceutical companies. At least three motilides have been tested in humans, but for various reasons (e.g., rapid tachyphylaxis, no significant clinical benefit, potential side effects), clinical trials with these newly derived molecules have failed to substantiate the impressive pharmacological potential seen with i.v. erythromycin. Whether the gastrokinetic capacity of motilin receptor agonists will be amenable to commercial development and clinical exploitation remains to be seen.

THE CENTRAL NERVOUS SYSTEM AND THE PEPTIDE FAMILY Most GI peptides are found in the brain and/or act as neuropeptides, but the situation remains unclear for motilin. Motilin mRNA has been identified in brain tissues, but RIA determination of motilin content brought ambiguous results. Motilin administration in

267

Motilin

this organ induced actions (suggesting the existence of motilin receptors in the brain) that were quite unexpected: appetite stimulation, growth hormone (GH) release, anxiety suppression, and so forth. Yet no valid data support its role as a neuropeptide. Most peptides are members of a ‘‘peptide family’’ that includes structurally related compounds; gastrinCCK, and secretin-VIP are typical examples. Motilin remained alone; however, recently a new peptide discovered in the gastric mucosa shows 25% similitude to motilin. This new peptide has been called motilinrelated peptide (MTL-RP) by some, but it is more often recognized as ghrelin because of its effect on GH release. Interestingly, ghrelin administration induces central actions that are similar to those described previously for motilin (e.g., GH release, appetite stimulation). Peripherally, ghrelin mimics motilin and appears to be the most potent gastrokinetic agent we have ever tested in the rodent. Future studies should

tell us more about the importance of this new family of peptides.

See Also the Following Articles CCK (Cholecystokinin) . Gastrin . Ghrelin . GI Hormone Development (Families and Phylogeny) . GI Hormones Outside the Gut: Central and Peripheral Nervous System . GI Tract, General Anatomy (Cells)

Further Reading Brown, J. C., Cook, M. A., and Dryburgh, J. H. (1973). Motilin, a gastric motor activity stimulating polypeptide: The complete amino-acid sequence. Can. J. Biochem. 51, 533–537. Itoh, Z. (1997). Motilin and clinical application. Peptides 18, 593–608. Peeters, T. L. (1993). Erythromycin and other macrolides as prokinetic agents. Gastroenterology 105, 1886–1899. Poitras, P. (1984). Motilin is a digestive hormone in the dog. Gastroenterology 87, 909–913.

269

Mu¨llerian Inhibiting Substance: New Insights

female genitalia. Evidence indicates that ALK2 is the MIS type I receptor. Expression of ALK2 in coelomic epithelium as well as in the circumferential mesenchyme of the Mu¨llerian duct may explain the antineoplastic effects of MIS on epithelial ovarian cancers, which are derivatives of the coelomic epithelium. Receptor or R-SMADs, the co-SMAD SMAD-4, and the inhibitory or I-SMADs function downstream of the TGF-b superfamily type I receptors. Once phosphorylated, R-SMADs dimerize with SMAD-4 and translocate to the nucleus, where they modulate SMAD-responsive gene expression either directly or through recruitment of additional DNA-binding proteins, such as CREB, CBP/p300, or FAST-1. The repressors of SMAD signaling such as SNIP1 interfere with the interaction between SMAD-4 and CBP/p300; I-SMADs 6/7 impair phosphorylation/activation of R-SMADs.

MIS DURING DEVELOPMENT In males, serum MIS is present at birth, peaks in infancy, and declines at puberty. Excluding a short time period after birth, serum MIS and testosterone are reciprocally related. This is best illustrated in male puberty, when a precipitous decline in MIS occurs during the characteristic increase in testosterone. In contrast, female ovaries produce low levels of MIS at birth, increasing slightly at puberty to levels similar to those of the adult male.

OBSERVATIONS FROM TRANSGENIC MICE AND HUMAN SYNDROMES WITH MIS DEFECTS Human Mu¨llerian duct regression is complete 51 days after ovulation. Failure of MIS signaling in males results in persistent Mu¨llerian duct syndrome (PMDS), in which affected individuals show normal male external/ internal genitalia development but manifest a retained cervix, uterus, and Fallopian tubes. In PMDS, findings of crossed ectopia of the contralateral gonad and Mu¨llerian structures contained within an inguinal hernia are common. Unilateral or bilateral testicular maldescent is also common and many PMDS patients are infertile. The phenotype seen in MIS and MIS RII knockout mice recapitulates that seen in human PMDS. Approximately half of human PMDS patients have undetectable MIS levels and harbor MIS gene mutations; most of the remaining patients show

normal/elevated serum MIS levels and have MIS RII gene mutations. Normally virilized MIS-deficient mice develop focal Leydig cell hyperplasia and tumors. In contrast, transgenic mice overexpressing MIS show undervirilization and cryptorchidism secondary to the effect of MIS on Leydig cell function. Interestingly, in addition to the expected Mu¨llerian regression, female overexpressors develop masculinized ovaries with seminiferous tubules, Sertoli cells, and a paucity of germ cells. MIS is also expressed in the granulosa cells surrounding the small and preantral follicles where MIS activity may function in oocyte development.

¨ LLERIAN EFFECTS EXTRA-MU Antiproliferative Effects Human ovarian epithelium derives from the coelomic epithelium, which expresses both type I and type II receptors, suggesting MIS as a potential therapeutic for the most common ovarian cancers that originate from these cells. Preclinical in vitro and in vivo studies of a human ovarian cancer cell line transplanted beneath the renal capsule of immunosuppressed mice showed that delivery of highly purified human recombinant MIS (hrMIS) via intraperitoneal injection or MIS-producing tissue implants could suppress tumor growth. We found that human breast cancer cells also express the MIS type II receptor, as does the normal involuting but not rapidly growing lactating breast. Furthermore, growth of human breast cancer lines is inhibited by hrMIS in vitro. Similar observations have been made of MIS growth inhibition against human prostatic cell lines, indicating additional targets for MIS as a cancer therapeutic against other reproductive tumors expressing MIS RII.

Regulation of Steroidogenesis The role of MIS as a modulator/regulator of Leydig cell differentiation/function was suggested by the phenotypes of transgenic mice that either over- or under-express MIS or MIS RII. The highest MISproducing male transgenics were undervirilized, had Leydig cell hypoplasia with low testosterone, and ultimately developed gonadal exhaustion. Conversely, those mice underexpressing MIS or MIS RII showed Leydig cell hyperplasia and high testosterone. A plausible explanation was derived when it was shown that Leydig cells express MIS RII and that

270 MIS suppresses testosterone production and Cyp17 expression, which catalyzes the committed step of testosterone synthesis. Cyp17 mRNA levels were highest in the gonads of mice with the lowest MIS levels and Cyp17 mRNA decreased after MIS treatment. When added to cAMP-stimulated MA-10 cells, a mouse Leydig cell line, MIS inhibited testosterone biosynthesis 10-fold, coincident with a decrease in the transcriptional activity at the Cyp17 promoter. MIS suppression of androgens as well as the detection of MIS type II receptor in normal prostate and prostatic cancers indicate that testing MIS in the setting of benign prostatic hypertrophy and prostate cancers may be warranted. MIS inhibition of aromatase activity has also been documented in cultured fetal gonads. It is of interest to determine if MIS has a therapeutic role in hyperandrogenic women with polycystic ovarian syndrome.

DIAGNOSTIC APPLICATIONS Normal human gender-specific MIS levels have been established from virtually all developmental stages using an enzyme-linked immunoabsorbant assay, and this assay is helpful in the evaluation of gonadal abnormalities and tumors. MIS is an excellent marker of testicular activity, specifically of the Sertoli cell compartment. Accordingly, in prepubertal boys with nonpalpable gonads, MIS can be used to differentiate anorchia from intraabdominal testes. MIS is also a useful diagnostic tool in intersex states, with levels correlating with the mass of testicular tissue. Consequently, MIS is abnormally low in gonadal dysgenesis, in which functional testicular tissue is subnormal. In androgen-insensitivity syndromes, MIS is elevated; thus, MIS levels can differentiate abnormal testicular determination from defects in androgen biosynthesis or sensitivity. MIS is also extremely valuable as a marker of tumor burden or recurrence in patients with granulosa or sex-cord tumors.

Mu¨llerian Inhibiting Substance: New Insights

CONCLUSION MIS plays an important role in normal sexual development and differentiation, and it has proven diagnostic value for intersex disorders, undescended testis, and gonadal tumors. Furthermore, MIS may have important therapeutic applications in a number of common reproductive tumors in both sexes. The success of these applications will likely depend on the efficiency and cost-effectiveness of MIS production. A complete understanding of MIS downstream signaling pathways and the induced gene products essential for its effect may lead to the development of therapeutics capable of modulating this pathway, which could be of benefit.

See Also the Following Articles Agonadism, Male and Female . Androgens, Gender and Brain Differentiation . Endocrine Disrupters and Male Sexual Differentiation . In Vitro Fertilization (IVF) . Testes, Embryology of

Further Reading Behringer, R. R., Finegold, M. J., and Cate, R. L. (1994). Mu¨llerianinhibiting substance function during mammalian sexual development. Cell 79(3), 415–425. Gustafson, M. L., et al. (1992). Mu¨llerian inhibiting substance as a marker for ovarian sex-cord tumor. N. Engl. J. Med. 326(7), 466–471. Lee, M. M., et al. (1997). Measurements of serum Mu¨llerian inhibiting substance in the evaluation of children with nonpalpable gonads. N. Engl. J. Med. 336(21), 1480–1486. Mishina, Y., et al. (1996). Genetic analysis of the Mu¨llerian-inhibiting substance signal transduction pathway in mammalian sexual differentiation. Genes Dev. 10(20), 2577–2587. Teixeira, J., et al. (1999). Mu¨llerian-inhibiting substance regulates androgen synthesis at the transcriptional level. Endocrinology 140(10), 4732–4738. Teixeira, J., Maheswaran, S., and Donahoe, P. K. (2001). Mu¨llerian inhibiting substance: An instructive developmental hormone with diagnostic and possible therapeutic applications. Endocr. Rev. 22(5), 657–674.

272

CLINICAL DESCRIPTION OF APS TYPES I AND II In 1980, Michael Neufeld, Noel Maclaren, and Robert Blizzard presented data from a survey of patients with APSs and proposed a classification for the syndromes observed: APS type I was defined as present in patients who have at least two of the triad of Addison’s disease, hypoparathyroidism, and chronic mucocutaneous candidiasis. Other associated autoimmune disorders were also allowed to be present. APS type II was defined as Addison’s disease with autoimmune thyroid disease and/or type 1 diabetes mellitus. Autoimmune disorders other than hypoparathyroidism and candidiasis were also sometimes present. APS type III was defined as patients presenting with thyroid autoimmune disease and any other autoimmune disease except Addison’s disease or hypoparathyroidism. APS type III could be grouped separately or together with type II. APS type IV was defined as patients presenting with two or more organ-specific autoimmune diseases that do not otherwise fall into the category of type I, II, or III. This form of APS is frequently present in patients with non-endocrine autoimmune disorders. In this article, we consider APS type IV to be part of APS type III.

APS TYPE I Prevalence APS type I is a rare disorder. The highest prevalence, 1 in 9000, has been reported in the Iranian Jewish population. It is also more common in the Finnish and Sardinian populations, with prevalences of 1 in 25,000 and 1 in 14,500, respectively. APS type I may appear in a sporadic or familial form and is now known to be an autosomal recessive disease caused by mutations of the autoimmune regular (AIRE) gene. According to Michael Neufeld and colleagues, females exceed males at all ages in a ratio of 1.6 to 1.

Presentation Of the classic features, chronic mucocutaneous candidiasis occurs at the youngest age, often during infancy. Hypoparathyroidism may occur shortly thereafter, usually before adolescence. Addison’s disease tends to appear after the onset of hypoparathyroidism and as late as the fourth decade of life. Chronic active

Multiple Autoimmune Endocrinopathy

hepatitis and malabsorption due to celiac disease tend to occur in APS type I and not in APS type II. Chronic active hepatitis is particularly important because it has been reported as the cause of death in a significant number of patients with APS type I. In addition, gonadal failure and diffuse vitiligo may be found in patients with APS type I, and in 1985, it was recognized that hypopituitarism and diabetes insipidus also occur. In 1990, P. Ahonen and colleagues reviewed their data from 68 patients in 54 families in Finland (coining the term ‘‘autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy [APECED]) and found that all of their patients had candidiasis at some time with hypoparathyroidism (79%), adrenocortical failure (72%), and/or gonadal failure (60% in females and 14% in males). Several of their patients also had dental enamel hypoplasia (77%) and/or keratopathy (35%) that was not attributable to hypoparathyroidism and whose endocrine dysfunction did not manifest until the fifth decade of life. These observations underscored the need for lifelong surveillance for new components of the disease. In 1998, Corrado Betterle and colleagues from Italy published their experience following 41 patients for 20 years (1967–1996). They compared the clinical manifestations in their patients with those of Neufeld and Ahonen. The occurrence of clinical manifestations and the age of onset were similar. They reported calcifications of the basal ganglia in 17 of their patients, most likely secondary to hypoparathyroidism. As in earlier published reports, Betterle and colleagues showed that 20% of their patients had chronic active hepatitis, again emphasizing the importance of continued surveillance of liver function in these patients (Table I).

Genetics of APS Type I APS type I is the only autoimmune disease known to be caused by a defect in a single gene. This disease is inherited as a simple recessive Mendelian trait. It occurs primarily in the familial form but can occur sporadically. The defective gene in this disorder was identified by positional cloning and was called AIRE (Fig. 1). The AIRE (autoimmune regulator) gene is located on chromosome 21q22.3. More than 40 mutations have been described. All of the mutations described are localized to the coding portion of the gene and are present in all ethnic populations, with the exception of one mutation that was found only in the homozygous form of APS type I in Iranian Jewish patients. The Iranian Jewish population also seems to have a milder clinical expression of this disorder.

273

Multiple Autoimmune Endocrinopathy

t0005

Table I ages)

Autoimmune Polyglandular Syndrome (Percent-

Type I Endocrine disorders Hypoparathyroidism

90

Addison’s disease Gonadal failure

60 45

Autoimmune thyroid disease

10

Type 1 diabetes mellitus

1

Hypopituitarism Diabetes insipidus Nonendocrine disorders Mucocutaneous candidiasis

75

Malabsorption syndromes Alopecia

25 20

Pernicious anemia

15

Chronic active hepatitis

10

Vitiligo

5

Sjo¨rgen’s syndrome Dystrophy of nails and dental enamel Progressive myopathy Type II Endocrine disorders Addison’s disease

100

Autoimmune thyroid disease

70

Type 1 diabetes mellitus

30

Gonadal failure

10

Hypopituitarism

11 mmol/L) was shown to decrease mortality. Recently, in a study by Gaede and colleagues published in the New England Journal of Medicine, the intensive concurrent treatment of various risk factors for cardiovascular disease, including diabetes, was shown to decrease the risk of stroke. Although nonfatal stroke was not the primary end point, and intensive treatment was simultaneously targeted at a number of risk factors such as diabetes, there were only 3 nonfatal strokes in the intensive treatment group and 11 in the control group. However, the specific contribution of glycemic control to this overall risk reduction is not clear. Diabetes mellitus and acute hyperglycemia are associated with a poor outcome of stroke. Although it appears increasingly likely that hyperglycemia causes this poorer outcome, this has not been definitely established

in humans. Animal studies have shown convincingly that hyperglycemia, as compared with euglycemia, increases the extent of ischemic damage in rats and monkeys. However, in focal ischemia models, this effect was shown only for reperfused brain tissue, albeit less consistently. The negative effects of hyperglycemia on outcome of brain ischemia are probably mediated through increased lactic acidosis and increased release of excitatory amino acids (glutamate in particular) that may contribute to neuronal cell death (excitotoxicity), exaggeration of edema formation, blood–brain barrier disruption, and hemorrhagic transformations. In humans, the data underlining a causal relation between hyperglycemia and stroke are still only circumstantial. This is explained partly by the intricate relations between the pathogenesis of hyperglycemia and the evolution and type of the stroke. First, reperfused brain tissue may be especially vulnerable to hyperglycemia, whereas anoxic tissue may even benefit from hyperglycemia. Recently, it was shown that the ischemic penumbra in stroke patients, represented as a mismatch between perfusion and diffusion on magnetic resonance imaging (MRI) (Fig. 1), is more likely to progress to infarction when

Figure 1 Diffusion weighted imaging (DWI) sequence 6 h after the onset of a stroke showing a circumscribed area of high signal indicating disturbed diffusion. This represents infarction of the brain. The magnetic resonance (MR) angiography after 6 h shows absent flow in the right middle cerebral and carotid artery. The mean transit time (MTT) MR shows hypoperfusion of the total right middle cerebral artery territory. The area of hypoperfusion on the MMT MR minus the area of disturbed diffusion on the DWI MR is generally considered to be the penumbra. The DWI MR after 3 days shows that the hypoperfused area has not progressed to infarction. The MR angiography shows reperfusion of the right middle cerebral artery through the circle of Willis. Courtesy of Geoffrey Donnan, National Stroke Research Institute.

p0040

335

Stroke

p0045

p0050

the patient had hyperglycemia at presentation. Also, hyperglycemia levels were related to lactate levels, which in turn were independently related to salvage of mismatch tissue. In a previous study, a poorer outcome of strokes had been reported when hyperglycemia occurred in nonlacunar strokes that may show reperfusion but not in lacunar strokes with little or no reperfusion. A special vulnerability of reperfused brain tissue to hyperglycemia is also suggested by the experiences with thrombolysis in stroke patients. In the first trial showing the benefit of thrombolysis in acute stroke patients, the negative effects of hyperglycemia on outcome of thrombolysis were also reported. Further studies have shown that it was specifically in patients with reperfusion after tissue plasminogen activator (TPA) that hyperglycemia was associated with poor outcome. In these patients, there is a relation with hemorrhagic transformation of the infarct. Interestingly, diabetes mellitus has actually been reported to carry a decreased risk of primary intracerebral haemorrhage, so it seems less likely that diabetic vasculopathy explains the increased incidence of hemorrhagic transformation in thrombolyzed strokes. The fact that stroke type determines the association between hyperglycemia and stroke outcome is also underlined by the fact that, in hemorrhagic strokes, hyperglycemia is not related to a poorer outcome in both nondiabetic and diabetic patients. Second, it seems that acute hyperglycemia may be more strongly related to poor outcome than is hyperglycemia due to diabetes mellitus. In a large meta-analysis, hyperglycemia was found to have a relative risk of mortality after ischemic stroke of 3.1 in nondiabetic patients and only 1.3 in diabetic patients. In conclusion, the literature suggests a causal relation between hyperglycemia and stroke outcome rather than hyperglycemia as a paraphenomenon of strokes with more severe outcome. It seems sensible to withhold intravenous fluids containing glucose during the acute phase of a stroke, and hyperglycemia should be controlled with the usual measures, especially when blood sugar levels higher than 16 mmol/L are measured. However, whether tight control of hyperglycemia should be the goal in patients with stroke cannot be determined with certainty until currently conducted trials of treatment of hyperglycemia in acute stroke patients are concluded. Hyperglycemia occurs in 20 to 50% of stroke patients. One-third of these hyperglycemic patients were known to have diabetes mellitus. In another third, the hyperglycemia is the initial presentation of de novo diabetes, demonstrated by an elevation of glycosylated hemoglobin (HbA1c) levels. In the remainder (with

a normal HbA1c at the time of hyperglycemia), the hyperglycemia is considered to be a result of the stroke, but the mechanism for this is unclear. Although a general stress response may lead to hyperglycemia, this is probably not the predominant cause in patients with a stroke. Other parameters of the stress response, such as levels of catecholamines, have been shown not to be related to blood sugar levels after stroke. Because the effects of focal ischemic events on neurotransmitter release at a distance from the focal ischemia can operate in the entire ipsilateral hemisphere of the stroke, the neuroendocrine axis may well be influenced by these alterations in neuronal excitation. Alternatively, the focal ischemic brain may mediate hyperglycemia directly through as yet unclarified mechanisms. If this were to occur preferentially in inadequately reperfused brain, the association of hyperglycemia and poor outcome of stroke would represent a paraphenomenon and not a causal relationship. Although hyperglycemia will rarely be confused with stroke, the differential diagnosis between stroke and hypoglycemia can be problematic. The blood sugar level in the patient presenting with coma or focal neurological deficit is obviously of crucial importance. If it is normal, hypoglycemia is unlikely. However, in the case of focal neurological deficit, the hypoglycemia may have been corrected by the patient’s oral glucose intake before the neurological signs and symptoms resolve. When the patient is comatose, established hypoglycemia will be the likely cause. More difficulties arise when the patient presents with hypoglycemia and focal neurological deficits. Hypoglycemia can occasionally present with focal neurological deficit without any accompanying symptoms of hypoglycemia. Correction of blood glucose levels will generally lead to a rapid recovery of the deficit, but excluding a transient ischemic event might be impossible in these cases. In patients with previous strokes, hypoglycemia reproduces or worsens the previous deficit. When there are accompanying signs and symptoms of hypoglycemia, the balance will definitely shift toward hypoglycemia as a cause of the neurological deficit.

p0055

THYROID DISEASE Both hyperthyroidism and hypothyroidism can contribute to a cardioembolic source for stroke. The most frequent event is hyperthyroidism leading to atrial fibrillation, a factor that generally leads to a fivefold increase in risk of subsequent stroke and

p0060

336

p0065

p9000

also a factor that, if it cannot be cured, leads to anticoagulation. Hypothyroidism can affect cardiac function and, thereby, the risk of intracardiac thrombus formation. Graves–Basedow disease is rarely associated with cerebral vasculitis, an entity also referred to as Hashimoto’s encephalopathy. Thyroid disease has not been firmly established as affecting the outcome of strokes. However, in general, a range of systemic diseases affect the outcome of stroke, and it seems appropriate to diagnose and treat thyroid disease promptly in patients in these circumstances. Strokes have not been shown to lead to thyroid disease. Thyrotoxicosis may feature in the differential diagnosis of stroke. Delirium and coma may be a feature of both, but the fever, tachycardia, hypotension, vomiting, and diarrhea should point toward thyrotoxicosis. Hyperthyroidism has been reported to lead to isolated corticospinal tract dysfunction, the mechanism of which is not known. Hypothyroidism may present with limb and gait ataxia as signs of cerebellar dysfunction. The onset is often more gradual than in strokes.

PARATHYROID DISEASE p0075

p9005

p0080

Hyperparathyroidism is known to contribute to hypertension. Patients with osteoporosis and compensatory hyperparathyroidism may have an increased risk of stroke due to this contribution. Hypoparathyroidism has not been identified as a risk factor for stroke, but it is related to MELAS syndrome. MELAS (mitochondrial myopathy, encephalopathy, lactic acidosis, stroke) is a mitochondrial disorder related to a point mutation in the transfer RNALeu(UUR) gene. It is typified by mitochondrial (ragged red fiber) myopathy, epilepsy, lactate acidosis, and stroke-like episodes. Frequently, this syndrome is associated with hypothalamo–pituitary axis dysfunction, diabetes mellitus, and/or hyperthyroidism. Parathyroid dysfunction has not been shown to affect the outcome of stroke or to be the result of stroke. Hypercalcemia, whether as a result of hyperparathyroidism or secondary to malignancy, may be confused with a stroke when a patient presents with confusion, nausea, and vomiting. In hypercalcemia, there is fatigue, anorexia, constipation, increased urination, a short QT interval, and generally a serum calcium level higher than 2.9 mmol/L (11.5 mg/dl).

Stroke

HYPOTHALAMO–PITUITARY AXIS DISEASE Hypopituitarism is associated with an increased risk of stroke, reflected in an excess mortality rate for cerebrovascular disease of 2.4. Because of the complexity of hormonal and metabolic disturbances in hypopituitarism, the causes of this excess mortality are poorly defined. A prothrombotic tendency due to growth hormone deficiency (which tends to persist even when hormones are adequately substituted), Cushing’s syndrome, and secondary hyperthyroidism are among the factors that may be implied in the pathogenesis of strokes. Pituitary disease has not been shown to affect the outcome of stroke, although it may affect the chances of recurrence of stroke. Pituitary apoplexy is the term that refers to a stroke in the pituitary gland. Such strokes can be either ischemic or hemorrhagic and are often complications of pituitary adenomas or the surgery or radiotherapy thereof. The condition is probably underdiagnosed given that the pathological correlate of apoplexy has a prevalence of 1 to 3% in autopsy studies. The clinical presentation is highly variable but should be suspected in any patient with severe headache, visual field defects, ophthalmoplegia, and/or altered mental status. In some cases, there is only headache at first. This headache is accompanied by nausea, vomiting, nuchal rigidity, fever, stupor, and coma (when blood and necrotic tissue leak and cause the features of subarachnoid hemorrhage or aseptic meningitis). In tumorous apoplexy, the destruction of the pituitary in most cases leads to hypopituitarism because most underlying adenomas are endocrinologically silent. In cases of endocrinologically active adenomas, there can be spontaneous resolution of preexisting endocrinopathy after apoplexy. The slowly evolving expansion of adenomas rarely leads to cranial nerve deficit because the nerves slowly lengthen in response to this expansion. In apoplexy (Fig. 2), there is sudden compression of cranial nerves due to hemorrhage or infarction and swelling. Compression of the optic nerve leads to visual field defects and decreased visual acuity, whereas compression of nerves III, IV, and VI leads to oculomotor disturbances. Compression of the trigeminal nerve may result in facial paresthesias and absent corneal reflex. Compression of the cavernous sinus may result in proptosis and eyelid edema. Finally, there can be Horner’s syndrome due to compression of the sympathetic chain and hyperpyrexia, and there can be diabetes insipidus or SIADH

p0085

p9015

p0090

337

Stroke

f0010

Figure 2 (A) Posterior coronal view showing anatomical relationship of pituitary gland to optic chiasm superiorly, sphenoid sinus inferiorly, and cavernous sinus laterally. (B) Mechanism of acute compression of structures within cavernous sinus from sudden expansion of pituitary adenoma due to hemorrhage or infarction and edema. Note that the further the tumor has eroded the floor of sella turcica prior to apoplectic episode, the more likely it is that multiple structures within the cavernous sinus will be involved. Reproduced from Reid, R., Quigley, M., and Yen, S. (1985). Pituitary apoplexy. Arch. Neurol. 42, 712–719.

p0095

(syndrome of inappropriate antidiuretic hormone) when the hypothalamus is compressed. The radiological diagnosis can be made on computed tomography (CT) if the stroke is hemorrhagic or if the adenoma is sufficiently large. The investigation of choice is MRI, which shows hemorrhage and underlying adenomas in great detail (Fig. 3). Infarction can be difficult to diagnose on routine magnetic resonance sequences. Diffusion-weighted images greatly increase the accuracy of the diagnosis of pituitary infarction. Peripheral enhancement of intrasellar masses can be another less specific sign. The hypothalamic–pituitary unit can be tested as outlined elsewhere. Pregnant women have a special predisposition to apoplexy due to the prothrombotic state related to pregnancy (and the resulting 12-fold increase in risk of stroke) and the dramatic enlargement of the pituitary gland due to the proliferation of prolactin-secreting cells. The classical features of postpartum pituitary apoplexy are absence of lactation, persistent amenorrhea, and lethargy. Other predisposing factors to pituitary apoplexy include bleeding disorders, anticoagulation, upper

respiratory tract infections, trauma, carotid angiography, Cushing’s disease, diabetes mellitus, adrenalectomy, atherosclerosis, sickle cell trait, and acromegaly. Surgical intervention is advocated in the majority of patients, with radiotherapy a less likely option. Less catastrophic is the effect of nonpituitary stroke on the hypothalamic–pituitary–adrenal (HPA) axis. Stroke is associated with increased activity of the HPA, manifested particularly by hypercortisolism. The normal regulation of cortisol secretion by adrenocorticotropic hormone (ACTH) is disturbed after stroke, a process in which cytokines seem to be implicated. Such a pathological HPA axis may exhibit an overriding function on established risk factors for cardiovascular disease, diabetes mellitus, and stroke such as abdominal obesity, hypertension, cholesterol, and triglycerides. Whether there is a causal relation between stroke and increased HPA axis activity needs to be further elucidated given that depression, anxiety, alcohol consumption, and smoking all have been shown to have similar effects. Also, changes in growth hormone, prolactin, and thyrotropin response to

p0100

338

Stroke

Figure 3 Axial and sagittal T1-weighted and transverse T2-weighted MRIs showing a large hemorrhage into the pituitary gland. Courtesy of G. Fitt.

thyrotropin-releasing hormone have been reported, and some of these changes may be a consequence of the hypersensitive HPA axis. They may explain some of the insulin resistance after stroke. It has been postulated that stroke in the caudate nucleus interrupts neurotransmitter pathways involved in the control of secretion of gonadotropins. Although hypercortisolism and some of the other disturbances have been related to disorientation and levels of motor impairment, the clinical relevance of many of the changes remains uncertain. In general, stroke leads to an increase in antidiuretic hormone (ADH) levels. However, this usually does not lead to hyponatremia. Patients with subarachnoid hemorrhage are especially likely to develop a syndrome of inappropriate ADH secretion (SIADH). However, it is important to distinguish SIADH from cerebral salt wasting, which is the more likely explanation for hyponatremia in patients with subarachnoid hemorrhage.

ADRENAL GLAND DISEASE p0105

The intermittent hypertension in pheochromocytoma increases the risk of stroke. In some cases, the pheochromocytoma is actually not identified until the patients present with a stroke. Also, Cushing’s disease

has been postulated to have an increased risk of stroke due to hypercortisolism or to the effects of treatment of Cushing’s disease such as external pituitary irradiation and posttreatment hypopituitarism.

See Also the Following Articles Cardiovascular Disease in Diabetes . Graves’ Disease, Hyperthyroidism in . Hypercalcemia and Hypercalcemia Treatment . Hypercorticolism and Cushing’s Syndrome . Hypertension and Diabetes . Hypopituitarism . Lactic Acidosis . Pheochromocytoma . Thyrotoxicosis: Diagnosis

Further Reading Gaede, P., Vedel, P., Larsen, N., Jensen, G., Parving, H-H., and Pedersen, O. (2003). Multifactorial intervention and cardiovascular disease in patients with type 2 diabetes. N. Engl. J. Med. 348, 383–393. Parsons, M., Barber, A., Desmond, P., Baird, T., Darby, D., Byrnes, G., Tress, B., and Davis, S. (2002). Acute hyperglycemia adversely affects stroke outcome: A magnetic resonance imaging and spectroscopy study. Ann. Neurol. 52, 20–28. Reid, R., Quigley, M., and Yen, S. (1985). Pituitary apoplexy. Arch. Neurol. 42, 712–719. Rosmond, R., and Bjorntop, P. (2000). The hypothalamic– pituitary–adrenal axis activity as a predictor of cardiovascular disease, type 2 diabetes, and stroke. J. Int. Med. 247, 188–197.

340

p0025

shortly after birth with the sensory neurotoxin capsaicin lack this extrinsic innervation to the alimentary tract, with the result that there is a marked depletion of SP fibers in the gastric mucosa and around gastric submusocal blood vessels. A scattered population of SP-immunoreactive endocrine-like cells is present in the mucosa of the human small intestine and colon. In the pancreas, a sparse distribution of SP fibers, which are probably of extrinsic origin, innervate blood vessels and acini and such fibers are also found surrounding acini and along blood vessels in the salivary glands. In the hepatobiliary system, immunoreactive fibers are found in the parenchyma of the liver and hepatic vasculature and are localized to the ganglionated and mucosal plexi of the gallbladder. In the trachea and bronchi, extrinsic SP-containing nerve fibers are found within the smooth muscle layer and around local ganglion cells and, in the nasal mucosa, within and under the epithelium and around arterioles, venules, and exocrine glands. In the human heart, SP fibers are found in close proximity to arterioles and are localized to the adventitia and to the border between the adventitia and media in a wide range of blood vessels. Within the urogenital system, immunoreactive nerve fibers of extrinsic origin are present in the urinary bladder throughout the ureter close to smooth muscle cells and around blood vessels in the kidney cortex often close to renal tubules and glomeruli. The female (uterus, oviduct, and vagina) and male (seminal vesicle, testis, epididymis, and vas deferens) genital organs are also innervated by extrinsic SP fibers. Consistent with predominantly neuronal localization of the peptide and its rapid rate of clearance from the circulation (half-life 20) in the literature are sporadic; i.e., the mutation is found only in the index case and not in the parents (de novo mutation). The classical phenotype is characterized by severe congenital or neonatal hyperthyroidism and goiter. TSHR antibodies are usually negative and this helps differentiate the disorder from the other relevant cause of neonatal hyperthyroidism, the transplacental passage of maternal TSHR-stimulating antibodies. Treatment with anti-thyroid drugs can usually control the hyperthyroidism, but relapses are the rule. Thyroid growth is not contrasted by antithyroid drugs and the goiter evolves from initially diffuse to multinodular early during childhood, requiring surgery. Milder cases have been described, with later onset of hyperthyroidism and goiter. Several mutations of the TSHR have been identified as being responsible for the syndrome. The relationship of genotype to phenotype has been found to be somewhat inconsistent in large families with multiple

p0020

Thyroid Disease, Genetic Factors in

members carrying the same mutation. This indicates that other factors (genetic or environmental) play a role in determining the phenotype, at least with some of the known mutations. More rarely, inherited mutations of the TSHR cause an inordinate responsiveness of the receptor to the placental hormone b human chorionic gonadotropin (b-HCG), which is closely related to TSH. In these cases, the effect of the mutation becomes evident only when high levels of b-HCG are present, i.e., during pregnancy, hence the appellation of familial gestational hyperthyroidism for this peculiar disease (OMIM 603373). Other than in pregnancy, the carriers of this rare mutation have a completely normal thyroid function. Inactivating Mutations of the TSHR Inactivating mutations of the TSHR manifest themselves with the clinical phenotype of resistance to TSH (OMIM 275200). Patients have often been detected at birth with severe primary congenital hypothyroidism (elevated TSH and low thyroxine) and a normally located but hypoplastic thyroid. In other cases, only mild, subclinical hypothyroidism has been found. Some of these patients have been noted to have an increased thyroglobulin level, an unexplained finding. Mutations leading to this disease have a variable effect on the receptor function, when studied in vitro. As expected with such a variable phenotype, several different mutations have been found; some completely abolish the expression of the receptor, whereas others only incompletely impair its responsiveness to TSH. The disease was initially described as autosomal-recessive, as predicted in many ‘‘loss-of-function’’ mutations. Initially reported cases have most often been compound heterozygotes, i.e., carrying two different mutations on each copy of the gene, but true homozygotes have also been found, usually in offspring of consanguineous marriages. However, since the first recognition of the syndrome, more and more heterozygous cases are being reported, with mild subclinical hypothyroidism, no goiter, and no evidence of thyroid autoimmunity. In one report of 10 such cases, 4 of the patients did indeed harbor mutations of the TSHR, 3 of which were heterozygous.

Inherited Defects of Thyroid Hormone Biosynthesis and Processing Defective synthesis of thyroid hormone from the thyroid results in overt or compensated primary hypothyroidism. Since thyroid growth is largely independent

411 of the enzymatic pathways involved in the biosynthesis of thyroid hormone, ongoing TSH elevation as a consequence of low circulating thyroid hormone results in goiter. Therefore, the clinical phenotype of this group of diseases is characterized by hereditary goiter, usually early in onset, and various degrees of impairment of thyroid function, ranging from euthyroidism to overt primary hypothyroidism. Additional clinical features may be observed when the defect has consequences on other organs, as in Pendred’s syndrome. The phenotype may be further characterized by the application of sophisticated thyroid function tests, such as the perchlorate challenge test or the measurement of thyroid hormone by-products. Because of their great impact on the overall health and survival of affected individuals, some of these diseases are quite rare and mostly recessive. However, their recognition has greatly enhanced the understanding of thyroid gland metabolism. Pendred’s Syndrome Pendred’s syndrome (OMIM 274600) is an autosomal-recessive disease characterized by congenital deafness, early-onset goiter, and euthyroidism (occasionally subclinical hypothyroidism). The prevalence of Pendred’s syndrome is estimated at 1/1000. The sensory deafness is often due to a malformation of the inner ear in which the cochlea is replaced by a single cavity (Mondini’s defect). However, an enlarged vestibular aqueduct, endolymphatic sac, and endolymphatic duct on magnetic resonance imaging (MRI) of the ear have been shown to be more specific signs. Except for its unusually early onset, the goiter is clinically not distinguishable from the more common endemic multinodular goiter. The pathology of the thyroid, however, displays marked hyperplasia and nodularity, sometimes suggestive of cancer. The incidence of true thyroid cancer, however, does not seem to be increased in these patients. The diagnosis of the disease relies on the classical perchlorate discharge test, which is highly sensitive but relatively nonspecific as other defects of thyroidal iodine organification result in a positive test. However, the combination of a positive perchlorate discharge test with typical malformations on MRI imaging of the inner ear establishes the diagnosis. The disease has been mapped by linkage analysis to chromosome 7q31. The cloned gene (termed PDS) encodes a protein (pendrin) that has features consistent with a transmembrane protein and is largely expressed in the adult human thyroid and (to a lesser extent) in the human cochlea and kidney. Although its function has not been fully clarified, pendrin seems to be an iodine–chloride transporter.

412 Sodium–Iodide Symporter Defect (OMIM 601843) This extremely rare disease is characterized by an autosomal-dominant mode of inheritance, severe congenital hypothyroidism, and multinodular goiter. The diagnosis is clinically established on the basis of a very low thyroidal uptake of radioiodine in the presence of goiter. An additional feature is a partial response in terms of thyroid hormone production to high oral doses of potassium iodide. The disease is caused by homozygous inactivating mutations of the sodium– iodide symporter gene, located on chromosome 19p33.2–p12. Thyroglobulin Defects (OMIM 274900) Patients with this autosomal-recessive condition display a wide range of thyroid dysfunction, ranging from severe hypothyroidism to euthyroidism. Goiter is invariably present and the thyroidal radioiodine uptake is elevated. Thyroglobulin is a very large protein whose gene is located on chromosome 8q24.2– q24.3. Several structural (qualitative) and quantitative defects of thyroglobulin have been detected in patients with the disease. Usually, patients with qualitative defects have elevated circulating thyroglobulin levels, whereas patients with quantitative defects have low thyroglobulin levels. Thyroperoxidase Defects (OMIM 274500) Patients with thyroperoxidase defects have severe congenital hypothyroidism and various degrees of goiter. The disease is autosomal-recessive and is diagnosed by a positive perchlorate discharge test, in the absence of clinical features of Pendred’s syndrome. The gene for thyroperoxidase has been mapped to chromosome 2p25 and has been found to be mutated in many but not all patients with this clinical presentation, suggesting some degree of heterogeneity in the syndrome. Thyroid Hormone Coupling Defect (OMIM 247700) This is a poorly defined group of disorders, also characterized by goiter and various degrees of hypothyroidism in which metabolic studies seem to suggest an altered coupling of iodotyrosil residues on thyroglobulin. Because part of this reaction is probably mediated by thyroperoxidase, there is a partial overlap with deficits of that enzyme and the phenotype could also be explained by alterations in thyroglobulin structure.

p0065

Dehalogenase Defects (OMIM 274800) Deiodinases (or dehalogenases) are a group of enzymes capable of deiodinating metabolites of thyroid hormone degradation, such as diiodotyrosine (DIT)

Thyroid Disease, Genetic Factors in

and monoiodiotyrosine (MIT). As such, they greatly contribute to the intrathyroidal pool of iodine by deorganifying ‘‘used’’ organic iodine and making it available for new hormone synthesis. Patients with a dehalogenase defect are incapable of reusing MIT or DIT and develop hypothyroidism and goiter secondary to urinary loss of iodine. The defect is inherited in an autosomal-dominant fashion and is diagnosed by an elevated urinary excretion of administered labeled diiodotyrosine and monoiodotyrosine. Also characteristic is a prompt resolution of hypothyroidism when dietary supplementation with high doses of iodine is given.

Developmental Defects of the Thyroid and Congenital Hypothyroidism Congenital hypothyroidism (CH) has an incidence of approximately 1 in 3000–4000 newborns. Universal screening for the disease is available in most countries and allows early detection and treatment. Although CH can be observed in any severe form of the defects of thyroid synthesis described above, it is most often due to an abnormal in utero development of the thyroid gland, varying from its complete absence (thyroid agenesis) to various degrees of ectopy and hypoplasia (thyroid dysgenesis). Hereditary forms of the latter are very rare, most likely due to the severity of the consequences of CH, which include infertility. A number of genes involved in thyroid organogenesis have been identified and their role in the pathogenesis of CH has been partially clarified. The thyroid transcription factor 1 (TTF-1; OMIM 600635), mapped to chromosome 14q21, has been shown in animal models to be necessary for normal early thyroid organogenesis. Despite its central role, mutations in TTF1 have been identified in only a few patients with congenital hypothyroidism due to thyroid dysgenesis. TTF-2 (on chromosome 9q22; OMIM 241850) is also highly expressed during thyroid ontogenesis. A familial case of thyroid agenesis associated with cleft palate and spiky hair has been shown to be due to heterozygous missense mutations of TTF-2 and was named Banforth-Lazarus syndrome. Paired-box gene 8 (PAX-8; OMIM 167415) is another transcription factor involved in thyroid organogenesis and regulation of the transcription of thyroid-specific genes, such as thyroglobulin and thyroid peroxidase. Nonsense and missense mutations of PAX-8 have been observed in two sporadic cases and one familial case of thyroid dysgenesis. Despite these encouraging findings, the etiology of the large majority of cases

Thyroid Disease, Genetic Factors in

of CH and thyroid dysgenesis remains unknown. It is likely that, in addition to other genetic factors, environmental factors play an important role.

GENETIC FACTORS IN THYROID AUTOIMMUNE DISEASES The Clinical Phenotype It is important to recognize that these diseases represent a group of diseases rather than a homogenous entity and that several clinical features may occur independently of one another, defining a variety of possible phenotypes. In general, the AITD are defined by the presence of a thyroid lymphocytic infiltrate, associated with serological evidence of thyroid autoimmunity in the form of circulating antibodies reactive to thyroid antigens and various degrees of thyroid dysfunction, ranging from profound hypothyroidism, as in the case of atrophic Hashimoto’s thyroiditis (HT; OMIM 140300), to severe hyperthyroidism, as in the typical Graves’ disease (GD; OMIM 275000). A number of other features may or may not be associated, such as the presence of Graves’ ophthalmopathy, an autoimmune disease of the orbital tissues, typically observed in a relevant number of patients with hyperthyroid GD but occasionally also found in patients with HT. Pretibial myxedema is a puzzling inflammatory process of the dermis, localized to the pretibial regions, that is also associated (more rarely than ophthalmopathy) with GD. Finally, depending on various factors that are not completely understood, the thyroid gland may be enlarged as a consequence of either massive lymphocytic infiltration or of ongoing stimulation, it may be normal in size, or it may be strikingly reduced in size, as in the case of primary myxedema.

Evidence for a Role of Genetic Factors in the Pathogenesis of Autoimmune Thyroid Disease It is a common experience in the practice of endocrinology to observe familial clustering of HT and GD. These diseases are quite prevalent in the general population (prevalence of approximately 1%). Therefore, given this high prevalence, random occurrence of the disease in more than one family member could account for the clinical observation, without necessarily indicating the presence of a genetic predisposition. However, a number of lines of evidence strongly

413 indicate an important influence of genetic factors in the etiology of these diseases. One indication of the presence of a genetic influence in the predisposition to AITD comes from family studies. In general, when there is such an influence, the prevalence of the disease in first-degree relatives of patients with the disease (probands) is significantly higher than in the general population. In early studies conducted in the late 1950s and 1960s, a high incidence of thyroid autoantibodies was observed in first-degree relatives of patients with HT or GD. The prevalence of positive anti-thyroglobulin or anti-microsomal antibody tests in relatives ranged from 45 to 55%, compared to a general population prevalence of approximately 15%. Although those studies were exposed to a number of potential selection biases, they provided the first indication that at least one component of the disease phenotype (e.g., the formation of thyroid autoantibodies) had a possible inherited cause. Indeed, more stringent classical segregation analysis in families with autoimmune thyroid disease confirmed these earlier results, indicating a Mendelian dominant mode of inheritance for thyroid autoantibodies, at least in some families. It is interesting that such a model of inheritance would in fact result in a prevalence of the phenotype in firstdegree relatives of patients of approximately 50%, as in the initial studies. In a further refinement of these findings, some investigators have indicated that in some families not only the predisposition to form thyroid autoantibodies is strongly hereditary, but also that even the fine molecular specificity of such antibodies can be inherited. In summary, there seems to be a clear role for genetic factors in the formation of thyroid antibodies, at least in families in which the clinical disease exists in one member. In these families, the thyroid autoantibody trait seems to be inherited in a classical Mendelian fashion as an autosomal-dominant trait, with high penetrance. The situation was different when normal subjects with thyroid autoantibodies but without clinical HT were selected as probands. In one such study, only 30% of relatives had a positive result, compared to approximately 16% expected from population data, indicating at best a polygenic inheritance with low penetrance. Thus, the relatively common anti-thyroid autoantibody phenotype may represent the consequence of a strong genetic influence only in families with overt AITD. The presence of circulating thyroid autoantibodies represents only one aspect of the AITD phenotype. Less clear data are available when one looks at the inheritance of the full AITD phenotype. Only

414 approximately 18% of relatives of patients with HT were found to be affected in one study, although the percentage also includes second-degree relatives. Another study observed that 33% of siblings of patients with HT were affected. Even lower percentages have been reported in GD, where a prevalence of 5 to 10% has been reported in siblings of patients. By dividing the prevalence of a disease in first-degree relatives of probands with the disease by the prevalence of the same disease in the population of origin, one can obtain the relative risk for the disease in sibs, also termed l(s). This measure provides an estimate of the importance of genetic factors in a disease. In nongenetic diseases, l(s) equals 1, whereas in highly inheritable and penetrant, monogenic disorders, it can rank as high as several hundred. The l(s) for HT can be estimated from available data to range from 20 to 45, quite a high number, indicating that indeed a significant genetic component in the etiology of HT exists. A somewhat lower value of 7.5 to 10 has been estimated for GD. It should be noted, however, that these numbers are somewhat artificial in that they are extrapolated from different studies, whereas no large population-based studies directly aimed at obtaining unbiased data are available. Another way of estimating the role of the genetic contribution to the etiology of a disease is studying twins. In theory, although both dizygotic and identical twins share as much of the environmental influences as possible, including the intrauterine milieu, only dizygotic twins share approximately 100% of the genome, whereas dizygotic twins share on the average only 50% of the genome. For example, an autosomal-dominant, fully penetrant disease should be 100% concordant (i.e., present in both twins) in identical twins and only 50% concordant in dizygotic twins. By measuring the concordance rate of a disease in identical twins and comparing it to the rate observed in dizygotic twins, investigators can obtain very precise data on the relevance of genetic factors, the penetrance, and the mode of inheritance of disease. In both GD and HT, studies in twins have shown concordance rates well below 100% in identical twins (approximately 30% in GD and 55% in HT). However, much lower concordance rates (close to 0) have been observed in dizygotic twins. Thus, whereas a lower concordance rate in dizygotic twins indicates the presence of genetic factors, the lessthan-100% concordance rate observed in identical twins indicates a role for environmental factors as well (incomplete penetrance). Interestingly, 80% of identical twins and 40% of dizygotic twins of HT patients had circulating thyroid autoantibodies. Again, these data suggest a dominant mode of

Thyroid Disease, Genetic Factors in

inheritance with high (80%) penetrance for the thyroid autoantibody trait, when part of a general predisposition to thyroid autoimmune disease. Family and twin studies confirm the clinical impression of a genetic predisposition to AITD. However, the mode of inheritance of the full clinical phenotype seems to be complex rather than simple Mendelian inheritance. Admittedly, knowledge of the mode of inheritance of AITD is still very limited. In the search for the susceptibility genes, it is reasonable to consider two hypotheses. According to one hypothesis, the overall genetic susceptibility is provided by a relatively (>10) large number of disease-associated gene variations at many genetic loci (risk factors). When a sufficient number of these variants, in any combination, are inherited by an individual and appropriate environmental events take place, a threshold is reached and the disease develops. Given the small contribution to the overall susceptibility provided by these putative genes, they are more easily detected by sensitive association studies. If, in contrast, only a few (10%. Epidemiologists have traditionally distinguished ‘‘endemic’’ goiter, observed in geographic areas of iodine deficiency, from ‘‘sporadic’’ goiter, observed in areas of iodine sufficiency. The accuracy of the distinction is, however, questionable, as the clinical phenotype of the two forms of the disease is indistinguishable, except for areas of extreme iodine deficiency, where large goiters are associated with cretinism and hypothyroidism. In addition to iodine deficiency, other factors are likely involved in the pathogenesis of this common disease. This hypothesis stems from several circumstantial observations. In

418

Thyroid Disease, Genetic Factors in

areas with iodine deficiency, not all exposed subjects develop goiter. Universal iodine supplementation strikingly reduces but does not abolish the disease and goiter is observed in the absence of iodine deficiency, as mentioned above. Additional environmental factors, such as tobacco smoking, naturally occurring goitrogens, and pharmacologic goitrogens, have been demonstrated. Sex hormones are also likely to be involved as goiter is more prevalent in women living in areas with mild iodine deficiency. The observation of familial clustering of the disease is a daily experience of all endocrinologists, especially in areas with mild or no iodine deficiency, and this has led to the hypothesis that inherited factors coincide in the predisposition to the development of goiter. The familial clustering of the disease has also been reported in several epidemiological surveys, mostly from areas of endemic iodine deficiency. As in the case of the AITD, a number of twin studies have been performed. These have shown that the environmental factors are largely predominant when cases are drawn from endemic regions, as shown by similar concordance rates between dizygotic and monozygotic twins. However, a large, population-based study in twins has shown significantly different and overall higher concordance rates between monozygotic and dizygotic twins (Table II). Interestingly, the intrinsic population prevalence of goiter in this latest study is much lower than in the two older studies, performed in the 1960s in areas of iodine deficiency. One possible explanation for these apparently discrepant data is that once the effect of the major environmental factor (iodine deficiency) is removed from the population, cases with a stronger genetic effect emerge. In this view, the distinction between endemic and sporadic goiter seems to be justified. Despite evidence in favor of relevant genetic influences, these have not been clarified yet. As in many common diseases, there is no identifiable mode of inheritance and multiple, relatively frequent gene variants are likely to play a role. Thus far, only a few studies have addressed this problem on a large scale. Studies in single large kindreds

t0010

with highly prevalent and apparently dominant transmission of multinodular goiter have indicated linkage of the phenotype to a locus (termed MNG-1) located on chromosome 14q31. The same studies have ruled out the TSHR (which resides in the same chromosomal region) from the linked region. MNG-1 has been confirmed as a susceptibility locus in at least another large family. Interestingly, MNG-1 overlaps with GD-2, a susceptibility locus for GD. It is therefore possible that the region contains a novel thyroidspecific growth factor and/or antigen. Similarly, a point mutation in exon 10 of the thyroglobulin gene causing a substitution of histidine for glutamine at codon 870 has been described and linked to goiter in three families, in the absence of hypothyroidism. Finally, an Italian family with suggestions of X-linked transmission has been described. Indeed, linkage analysis has shown evidence of linkage of goiter to chromosome Xp22. These findings are encouraging but were obtained in a small subset of families with a seemingly large genetic effect and their epidemiological relevance to goiter found in the general population is unknown. Monogenic disorders of thyroid hormone metabolism represent a rare cause of hereditary goiter, with or without hypothyroidism. These conditions, described in some detail above, are distinct from the commonly encountered sporadic or endemic goiter in that there is usually an earlier appearance of goiter, a more distinct familial pattern (usually autosomal-recessive), and several degrees of thyroid dysfunction. Although it is conceivable that more subtle defects in any of these genes might be involved in the pathogenesis of goiter, extensive screening of the general population for association and/or linkage to known genes is lacking. In summary, genetic predisposition is probably important in the etiology of simple goiter, both sporadic and endemic. In endemic areas, though, the effect of environmental factors (iodine deficiency) is predominant and widespread, making goiter very prevalent. In contrast, in areas with normal iodine intake, genetic factors are more relevant, but with a smaller

Table II Available Studies in Twins in Populations with Different Prevalences of Goiter Geographic location

Concordance rate in monozygotic twins

Concordance rate in dizygotic twins

1967

United Kingdom

24%

12%

25.3%

1967 1999

Greece Denmark

89% 42%

73% 13%

53.0% 1.5%

Year of the study

Population prevalence of goiter

Note. In populations with a lower prevalence, a higher concordance rate is observed in monozygotic twins when compared with dizygotic twins, suggesting a stronger genetic predisposition.

Thyroid Disease, Genetic Factors in

prevalence. As a consequence, the prevalence of goiter is strikingly reduced and the phenotype more clearly clustered in families. It is likely that multiple different genes are involved in the predisposition to simple goiter, but thus far only a few of these have been identified.

GENETIC FACTORS IN NONMEDULLARY THYROID CANCER p0160

The etiology of thyroid cancer of follicular origin (OMIM 188550) is poorly understood, although several somatic mutations consistently associated with the neoplasm have been described. Exposure to ionizing radiation, especially during childhood, is a well-recognized environmental risk factor, accounting for only a minority of cases. Most other cases appear to be sporadic. In some instances, familial clustering has been observed. Although in many such cases, random occurrence of this relatively common disease in members of the same family can explain the observation, in some extended families, a clear pattern of inheritance has been observed, suggesting the existence of genetic factors. Clinically, familial nonmedullary thyroid cancer (NMTC) has been reported to be more aggressive and more often multifocal than its sporadic counterpart. A few population-based studies have shown an increased risk of NMTC in relatives of propositi. For example, the Connecticut Tumor Registry study has demonstrated a fivefold increased risk in relatives of index cases with thyroid cancer. Such studies, however, are subject to several selection biases and may also reflect environmental influences shared by families. It is the opinion of many experts that a genetic predisposition to isolated nonmedullary thyroid cancer exists in only a few families. In two members of the family in which MNG-1 was identified, thyroid cancer was found. It is tempting to speculate that the same gene driving follicular cell hyperplasia in that family may also increase the risk for thyroid malignancy. Moreover, distinct loci have been identified in a few families. In a large multigeneration family with familial NMTC, linkage to a 20 cM segment on chromosome 1q21 was established. In two affected members, papillary renal neoplasms were also observed, suggesting a syndromic effect of the identified locus. In another large kindred with several cases of papillary thyroid cancer only, linkage was suggested with chromosome 2q21. The authors were able to confirm their results in a data set of 80 smaller families with more than one case of NMTC, obtaining significant evidence for linkage at the same locus, although the presence

419 of benign thyroid disease was included in the analysis. Even stronger evidence was found when the analysis was limited to families with the follicular variant of papillary thyroid cancer. A rare form of thyroid neoplasm, thyroid tumors with oxyphilia, has also been linked to chromosome 19p31. The phenotype in this case is represented by benign or malignant thyroid nodules, with prominent oxyphilia. The same locus (termed TCO1) was found to be linked to papillary thyroid carcinoma in a subsequent study in several families. Several genes are known to reside within that chromosomal location, but none has been found to be mutated in patients with NMTC thus far. An increased risk of thyroid cancer has also been observed as a component of known hereditary syndromes. Cowden’s disease (OMIM 158350) is an autosomal-dominant disease, characterized by multiple hamartomas of the skin, breast, thyroid, brain, and endometrium. Facial trichilemmomas are a distinctive feature of the syndrome. Other findings include craniomegaly, scrotal tongue, and cerebellar neoplasias. There is some overlap with related syndromes such as Lermitte-Duclos syndrome and Ruvalcalba’s syndrome, so that some investigators believe they are in fact synonyms. A significant increase in the incidence of breast cancer is seen in Cowden’s syndrome families. Thyroid cancer, of both the papillary and follicular types, is also highly prevalent, as is benign thyroid nodular disease, although the magnitude of the relative risk is not known. Cowden’s syndrome has been mapped to chromosome 10q23.31 and the mutated gene, termed PTEN for phosphatase and TENsin homologue (OMIM 601728), has been isolated. Several mutations of PTEN have been found in families with Cowden’s syndrome, without a clear-cut genotype–phenotype relationship. Moreover, in several families with many features of the syndrome, no mutation of the gene has been found, indicating that some degree of heterogeneity exists. Several in vitro studies suggest that PTEN acts as a tumor suppressor gene. This observation is in keeping with an autosomal-dominant mode of inheritance and with the observation that most of the mutations identified thus far cause a loss of function. The PTEN protein is expressed in several tissues and is believed to function as a tyrosine phosphatase. As such, the protein is believed to be able to dephosphorylate activated regulator proteins, thus down-regulating proliferation pathways. An increased risk of NMTC has also been observed in familial adenomatous polyposis (FAP; OMIM 175100), also termed Gardner’s syndrome. FAP is inherited in an autosomal-dominant fashion. The most

p0170

420 prominent feature of the syndrome is early-onset colonic cancer and multiple widespread, preneoplastic polyps of the colon. Extracolonic manifestations include congenital hypertrophy of the retinal pigment epithelium in up to 90% of affected persons and has been used as a marker of the disease before genetic testing became available. Other features of the syndrome include benign bone tumors and hepatoblastomas. In women with the syndrome, the risk of papillary thyroid cancer has been estimated to be 160-fold higher than in the general population. Papillary thyroid cancer in the setting of FAP is often detected before age 30, is multifocal, and displays unusual pathological features. The gene responsible for FAP has been mapped to chromosome 5q21 and was designated adenomatous polyposis of the colon (APC). As expected by the autosomal mode of inheritance, APC functions as a tumor suppressor gene and detected mutations associated with the phenotype cause a loss of function of the gene product. The adenomatous polyposis of the colon (APC) gene product (b-catenin) is expressed on the membrane of several epithelial cell types. It has been hypothesized that loss of b-catenin function causes a loss of contact inhibition or a loss of cell adhesion mechanisms, leading to uncontrolled cell growth. In summary, although the vast majority of cases of NMTC are sporadic, a few cases present in the setting of recognized hereditary syndromes. Knowledge of these clinical conditions is important to the clinician, in order to perform appropriate screening for associated conditions and to provide sound genetic counseling to individuals at risk.

FAMILIAL MEDULLARY THYROID CANCER p0180

Approximately 7% of thyroid cancers originate from the parafollicular cell lineage. These cancers are termed medullary thyroid cancer (MTC) and account for approximately 15% of all thyroid cancer-related deaths. In contrast to NMTC, MTC is quite often familial (in approximately one-fourth of the cases). Familial medullary thyroid cancer (FMTC) arises in the setting of multiple endocrine neoplasia (MEN) type II syndromes or as an isolated disease. In MEN IIA (OMIM 171400), FMTC is associated with pheochromocytomas and parathyroid adenomas, and in MEN IIB (OMIM 162300), patients present a marfanoid habitus, mucosal neuromas, and ganglioneuromatosis, in addition to the above-described features (see Fig. 1). In isolated FMTC (OMIM 155240), there are no

Thyroid Disease, Genetic Factors in

Figure 1 Typical lingual neuromas in a patient with MEN IIB.

extrathyroidal manifestations. The inheritance of these closely related conditions is autosomal-dominant, with high penetrance. In MEN IIA and MEN IIB, MTC is found in almost 100% of carriers by the end of third decade of life if sought by biochemical screening, whereas pheochromocytomas occur in 50% of carriers and parathyroid tumors in 25% of carriers. Despite these clearly distinct phenotypes, all three conditions are due to germ-line mutations in the RET protooncogene sequence. The RET proto-oncogene maps to chromosome 10q11.2 and encodes a membranebound tyrosine kinase receptor. Causative mutations detected thus far are almost entirely limited to a few cysteine residues located in the putative extracellular domain. It is hypothesized that these mutations induce critical conformational changes in the gene product, causing constitutive activation of the receptor with ongoing production of intracellular proliferative signals. Interestingly, there is quite a strict genotype– phenotype relationship. Mutations involving codons 609, 611, 618, 620, and 634 are found exclusively in families with MEN IIA or with FMTC and mutations involving codons 768, 804, and 891 have been found only in families with FMTC, although even in these families occasional cases of pheochromocytoma and hyperparathyroidism are found. Since the molecular genetics of the two conditions partially overlap, other genes must be involved in determining the phenotype. In contrast, mutations at codon 918 are found in almost all families with MEN IIB. Somatic missense mutations at codon 918 are also found in one-half of cases of sporadic MTC. The RET gene is also mutated in many cases of Hirschsprung’s disease, although most mutations detected in this disorder induce the loss of function of the mutant allele. Surprisingly,

421

Thyroid Disease, Genetic Factors in

Hirschsprung’s disease has been found in some families with MEN IIA. Detailed knowledge of the molecular genetics of these aggressive diseases has greatly enhanced the ability of clinicians to detect persons at high risk for multiple endocrine neoplasias. Genetic testing for RET mutations has become widely available for family members of patients and allows early, preventive treatment, which is expected to greatly increase the life expectancy of patients.

See Also the Following Articles Iodine . Sodium Iodide Symporter . Thyroid Autoimmunity . Thyroid Carcinoma . Thyroid Disease and Pregnancy . Thyroid Disease, Epidemiology of . Thyroid Disorders in the Elderly . Toxic Multinodular Goiter

Further Reading Alsanea, O., and Clark, O. H. (2001). Familial thyroid cancer. Curr. Opin. Oncol. 1, 44–51. Barbesino, G., and Chiovato, L. (2000). The genetics of Hashimoto’s disease. Endocrinol. Metab. Clin. N. Am. 2, 357–374. Braverman, L. E., and Utiger, R. D. (2000). ‘‘The Thyroid,’’ 8th ed. Lippincott Williams & Wilkins, Philadelphia, PA. Brix, T. H., and Hegedus, L. (2000). Genetic and environmental factors in the aetiology of simple goiter. Ann. Med. 3, 153–156. Kopp, P. (2001). The TSHR and its role in thyroid disease. Cell Mol. Life Sci. 9, 1301–1322. Macchia, P. E. (2000). Recent advances in understanding the molecular basis of primary congenital hypothyroidism. Mol. Med. Today 1, 36–42. Malchoff, C. D., and Malchoff, D. M. (2002). The genetics of hereditary nonmedullary thyroid carcinoma. J. Clin. Endocrinol. Metab, 6, 2455–2459. Tomer, Y., and Davies, T. F. (1997). The genetic susceptibility to Graves’ disease. Bailliere Clin. Endocrinol. Metab. 3, 431–450.

423

Thyroid Disorders in the Elderly

the elderly. Hypercholesterolemia is present in the majority of overt hypothyroid patients. Cardiovascular system abnormalities include cardiomegaly secondary to pericardial effusion, bradycardia, diastolic hypertension, and atherosclerosis. The myopathy of hypothyroidism causes proximal muscle pain and stiffness and weakness. Hypothyroidism also causes sleep apnea. Anemia—microcytic, normocytic, or macrocytic—is a well-characterized hematologic feature of hypothyroidism. Thyroxine requirements are decreased in the elderly because of a decline in the metabolism of thyroid hormone. In elderly patients with coexisting cardiovascular disease, starting treatment with full replacement doses can result in exacerbation of angina and worsening of underlying heart disease. The starting dose of thyroxine should be small, 12.5–25 mg/day. The dose should be adjusted at 6- to 8-week intervals by an increment of 12.5–25 mg until the patient is euthyroid and the serum TSH is in the mid-normal range. Patients with central hypothyroidism should be monitored by FT4 measurement and not serum TSH. Drugs that block thyroxine absorption and increase thyroxine requirements are calcium carbonate, ferrous sulfate, cholestyramine, colestipol, sucralfate, and aluminum hydroxide. Rifampin, carbamazepine, phenytoin, and sertaline (Zoloft) accelerate T4 clearance and increase the serum TSH level in patients on previously ideal replacement doses. Estrogen may also increase thyroxine requirements.

Subclinical Hypothyroidism In an English survey, the prevalence of subclinical hypothyroidism was 17.4% in women older than age 75. In a survey in the United States, the prevalence of serum TSH elevation was 8.5% in women and 4.4% in men older than age 55. The causes of subclinical hypothyroidism are the same as those of overt hypothyroidism. Subclinical hypothyroidism increases the risk for myocardial infarction two- or threefold and increases low-density lipoprotein cholesterol significantly when serum TSH is increased threefold above normal. The potential benefits of treating mild thyroid failure include prevention of progression to overt hypothyroidism, reduction of elevated serum cholesterol, and improvement in the quality of life. Replacement therapy is recommended for all patients with serum TSH concentrations greater than twice the upper limit of normal. It is also recommended if there are any clinical features of depression, fatigue, hyperlipidemia, or goiter. The goal of therapy is to normalize the serum TSH concentration. In older

people, 12.5–25 mg of levothyroxine is recommended as the initial dose. With minimal TSH elevation and the absence of clinical features, patients should be followed at intervals of 6 months for worsening of the condition, indicating a need for treatment.

HYPERTHYROIDISM The prevalence of hyperthyroidism varies from 0.5 to 2.3% in the elderly. Approximately one-sixth of all hyperthyroid patients are older than age 60. Graves’ disease is the most common cause of hyperthyroidism in the elderly. TSH receptor antibodies are detectable in approximately 80% of untreated Graves’ patients. Toxic multinodular goiter is more common in the elderly and has been reported in almost half of older patients with hyperthyroidism, especially in regions of relative iodine deficiency. Thyrotoxicosis can be precipitated in patients with nontoxic multinodular goiter by administration of a large iodine load such as radiocontrast agent. Some autonomously functioning thyroid adenomas have a mutation in the TSH receptor that results in chronic activation of the follicular cell. Amiodarone may cause destructive thyroiditis and thyrotoxicosis. Hyperthyroidism resulting from a TSH-secreting pituitary adenoma or pituitary resistance to thyroid hormone is very rare. Elderly patients often lack typical features of hyperthyroidism, including goiter. The absence of the typical hypermetabolic manifestations of hyperthyroidism is termed apathetic hyperthyroidism. The dominant clinical findings may be weight loss, atrial fibrillation, or heart failure. Serum TSH level is suppressed and FT4 and FT3 are elevated in hyperthyroidism. However, the serum T3 level was found to be increased in only half of hyperthyroid patients between 75 and 95 years of age because of reduced conversion of T4 to T3 in peripheral tissues. A thyroid radioiodine uptake is useful to detect the conditions causing thyrotoxicosis with low thyroid uptake: thyroiditis, exogenous intake of thyroid hormone, or iodine-containing drugs. Radioactive iodine-131 is the most common therapy of Graves’ hyperthyroidism and toxic multinodular goiter. The usual doses are 5–15 mCi of 131I for Graves’ disease and 15–50 mCi for large multinodular glands. Beta-adrenergic blockers are used to control symptomatic tachycardia, tremor, anxiety, and muscle weakness, and they are discontinued when the patient is euthyroid. Hypothyroidism is a common consequence after radioiodine treatment, with an eventual incidence of more than 50%.

424 Definitive therapy with an antithyroid drug, propylthiouracil or methamizole, is appropriate for otherwise healthy elderly patients. The recurrence rate is significantly less with advanced age than in younger patients. Surgical thyroidectomy is only advised if there are obstructive symptoms from a large goiter or the presence of a nodule that is suspicious for malignancy. Although the mortality from subtotal thyroidectomy is very low, the complications of recurrent laryngeal nerve damage and hypoparathyroidism can result in lifelong disability.

Subclinical Hyperthyroidism Subclinical hyperthyroidism is defined as a state of suppression of serum TSH with normal free thyroxine and triiodothyronine levels in a patient who lacks clinical features of thyrotoxicosis. The causes of subclinical thyrotoxicosis are the same as those of overt thyrotoxicosis. In a study of patients older than 55 years of age, 0.7% had endogenous subclinical hyperthyroidism. Two meta-analyses found a significant loss of bone density in postmenopausal women with suppressed serum TSH. Subclinical hyperthyroidism has been associated with an increased frequency of nervous symptoms and an increased risk of dementia and Alzheimer’s disease. Therapy should be considered for any patient with subclinical hyperthyroidism who has mental symptoms, osteoporosis, atrial fibrillation, or cardiac disease. A trial of antithyroid drugs to normalize the serum TSH level is warranted. In patients with more severe features, such as atrial fibrillation, ablation of the hyperfunctioning thyroid with radioactive iodine is preferable.

Thyroid Nodules Thyroid nodules, either solitary or multiple, increase in frequency with advancing age. Ninety percent of women older than age 60 and 60% of men older than age 80 have a nodular thyroid gland. Thyroid nodules in asymptomatic individuals (incidentalomas) are identified more frequently by ultrasonography rather than by examination of the gland by palpation. Most thyroid nodules do not cause symptoms. Pain may occur with a hemorrhage into a preexisting colloid nodule or a benign adenoma. Rapid growth over a period of weeks is suspicious of malignancy, and persistent hoarseness may indicate recurrent laryngeal nerve invasion by tumor. A hard and fixed nodule is more likely to be malignant, but many papillary carcinomas or

Thyroid Disorders in the Elderly

follicular tumors are soft or cystic. Lymphadenopathy is strongly suggestive of malignancy. Low serum TSH concentration in the setting of a nodular goiter suggests the presence of either an autonomously functioning adenoma or a toxic multinodular goiter. Positive anti-peroxidase antibody indicates lymphocytic thyroiditis that may present as a nodule. Thyroid ultrasound is capable of identifying impalpable nodules as small as 2 mm. The clinical significance of small nodules detected by ultrasonography is uncertain. Solitary incidentalomas larger than 1.5 cm should probably be biopsied under ultrasound guidance. Fineneedle aspiration (FNA) biopsy is the most important diagnostic test, with accuracy, sensitivity, and specificity of 98 or 99%. In a large series of FNA biopsy of thyroid nodules, benign cytology was found in 69% (mainly colloid goiter), malignant cytology in 3.5%, and suspicious cytology in 10%. The suspicious category consists of variants of follicular neoplasm, but follicular adenomas are approximately 10-fold more common than follicular carcinomas. In a patient with a follicular nodule, a radioiodine scan may be helpful. ‘‘Hot’’ or functional nodules are rarely malignant. The presence of nuclear atypia in a follicular lesion has a 44% prevalence of malignancy, and the absence of nuclear atypia denotes a benign lesion. Positive immunostaining for galectin-3 correlates with malignancy; immunostaining for galectin-3 and other proteins may improve the differential diagnosis of suspicious lesions. Treatment of the thyroid nodule depends on the functional state of the nodule and cytologic diagnosis. If the cytology indicates malignancy or is strongly suspicious for malignancy, the nodule should be removed surgically. In the 10% of suspicious cytologic findings, approximately one-fourth of patients who go to surgery are found to have a malignant lesion. Altogether, only 6% of thyroid nodules are malignant. The hyperfunctioning hot nodule is treated with radioiodine ablation or surgery. The vast majority of thyroid nodules are benign and should be managed medically. Medical management with thyroxine suppression therapy is based on the assumption that growth of the nodule depends on TSH. Spontaneous regression of thyroid nodules may occur. Use of suppressive therapy of benign thyroid nodules has been challenged in the past few years due to the failure of some studies to show a significant decrease in nodule size and concern about reducing mineral bone density. However, several studies have shown >50% reduction in nodule size in 40% of patients with a single nodule. Generally, patients are followed by palpation at intervals of 3 months. Ultrasound examination

425

Thyroid Disorders in the Elderly

provides more objective assessment of growth or shrinkage of a nodule.

THYROID CANCER Thyroid cancer accounts for 1.6% of all new cancers in the United States and causes 0.4% of all cancer deaths. It is classified into five major types: papillary, follicular, medullary, anaplastic, and thyroid lymphoma. Most thyroid cancers are indolent and grow slowly over years, whereas a few grow aggressively and cause death within 1 year. Thyroid carcinomas tend to be more aggressive and poorly differentiated in the elderly compared to younger patients. Papillary carcinoma accounts for 80% of all thyroid cancers, and follicular carcinoma accounts for 10%. These differentiated cancers are more aggressive in older patients. Hurthle cell carcinoma is considered a variant of follicular thyroid carcinoma and carries an even worse prognosis. Extension of the tumor through the thyroid capsule and into the surrounding structures is associated with poorer prognosis. Cervical lymph node metastases occur in approximately 50% of patients with papillary carcinoma and are associated with only a slightly higher rate of recurrence and mortality. Surgery, either near total or total thyroidectomy, is the initial treatment of choice for patients with differentiated carcinoma. Near total thyroidectomy is performed for extensive unilateral tumors with local metastases. Total thyroidectomy is performed for patients with extensive multifocal disease with metastases to the cervical lymph nodes, contiguous neck structures, or distant sites. The main disadvantage of total thyroidectomy is the higher incidence of hypoparathyroidism. Radioiodine therapy is used as an adjunct to surgery to treat patients with residual or recurrent papillary cancer in the neck. Thyroid hormone in a suppressive dose is given after thyroidectomy to reduce the recurrence rate. TSH stimulates growth of thyroid tumors that contain TSH receptors. The dose of thyroxine should be adjusted to keep TSH suppressed without causing clinical thyrotoxicosis. The degree of suppression is based on the staging of the patient. In patients with a good prognosis, TSH should be suppressed to the slightly subnormal range. In patients with worse prognosis, which includes many of the elderly, TSH should be suppressed to TR-a

Bone development

TR-a > TR-b

TSH suppression

TR-b > TR-a

Ligand-independent TSH elevation Cochlear development and function

TR-a > TR-b TR-b

Maturation of small intestine Cardiac gene expression

TR-a ¼ TR-b

TR-a > TR-b

Heart rate

TR-a

Retinal development

TR-b

Growth

TR-a > TR-b

Immune function

TR-a > TR-b

Temperature regulation

TR-a > TR-b

hormone. The TR-a1 and-a2 isoforms are expressed in a wider area but are still clustered around the central vein. These findings indicate that when studying TR isoform-dependent effects, it is important to consider these effects not only in vitro but also in vivo in relation to the local expression of the different TR isoforms. The elucidation of the TR isoform-dependent effects was greatly helped by the advent of mice devoid of one or more specific isoforms. The TR-a1 knockout is fertile and shows a mild hypothyroid phenotype, reduced body temperature, and reduced heart rate. Selective ablation of TR-a2 results in overexpression of TR-a1, low levels of thyroid hormones,

Figure 4 Zonal expression of TR-b1 in rat liver. Rat liver slices were incubated with polyclonal anti-TR-b1. Monoclonal antiglutamine synthetase was used as a control to stain the central veins on a consecutive slice. It can be clearly seen that TR-b1 is expressed in the same subset of cells that express GS. A schematic drawing depicting the position of the central (CV) and portal (PV) veins in the slices is shown on the right. Magnification, 100.

494 and normal levels of TSH. Interestingly, the phenotype of TR-a2 mutant mice also shows signs of hyperthyroidism, such as decreased body weight, elevated heart rate, and increased body temperature. These data suggest that the balance between TR-a1 and TR-a2 may provide an additional level of adjustment of hormone responsiveness in certain tissues. In mice missing the full-length TR-a1 and TR-a2 (but that have the Da isoforms), the thyroid gland develops abnormally, there is arrested maturation of the intestine and reduced bone growth, and the mice die within a few weeks after birth. Interestingly, when all TR-a isoforms are deleted (both full-length and D), the phenotype is less severe. The selective inactivation of the TR-b gene results in thyroid hyperplasia, increased serum thyroid hormones and TSH, impaired T3-dependent regulation of cholesterol metabolism, and defects in cochlear function (similar to the resistance to thyroid hormone syndrome in humans). The TR knockout animal models emphasize two important features. First, the mice without all known isoforms are still viable. The existence of an unknown receptor isoform is a possibility that has not been clarified. Second, the deletion of a particular isoform can be partly compensated by other receptors. However, certain genes or processes will be influenced by the deletion of a particular isoform, as seen in the case of genes involved in lipid metabolism or inner ear development.

TR AGONISTS AND ANTAGONISTS Solving the X-ray structure has allowed the development of TR isoform-specific agonists and antagonists. As for the knockouts discussed previously, it is not an all-or-nothing phenomenon but it is a matter of preferential stimulation or inhibition. One of the first isoform-specific agonists to be synthesized was GC-1, which showed a clear preference for TR-b1 both in vitro and in vivo. The X-ray structure suggested that compounds with a 50 -aryl extension could act as antagonists because they would interfere with the proper folding of helix 12 (the lid on the box). However, many compounds based on GC-1 that have large extensions act as agonists with the exception of GC14, which acts as a partial TR-b1 antagonist. This can probably be explained by the fact that the side chains are not rigid enough and therefore allow helix 12 (the lid) to assume its proper position. Other synthetic ligands have been developed that behave as antagonists (DIBRT and NH-3) or partial antagonists (NH-4). Of these, NH-3 appears to be the first

Thyroid Hormone Receptors

high-affinity TR antagonist that also inhibits TR action in a Xenopus development model. It has been shown to block both coactivator and corepressor binding. The latter is strange since all known receptor antagonists promote the binding of these corepressors because although they block the proper positioning of helix 12, they leave the corepressor binding site intact. This means that as a result of NH-3 binding, helix 12 assumes a position that precludes both coactivator and corepressor binding. In light of the fact that the TRs are also able to bind without hormone and thus inhibit gene expression, this may be a beneficial property.

CONCLUSION Solving the structure has allowed the design and understanding of the mechanism of action of antagonists and antagonists. With the current knowledge of receptor structure, it can be envisaged that it may be possible to design receptor agonists that will correct the receptor defect in thyroid hormone resistance. For instance, it has been shown that the shift of 0.3 A˚ of helix 6 of the TR due to mutation of alanine-317 to threonine is the cause of decreased T3 binding. If an agonist were found that could ‘‘live’’ with this small shift and thus activate the receptor, patients harboring this particular mutation could be treated. Furthermore, the fact that TRs are not homogeneously expressed in target tissues and the design of novel agonists and antagonists open up exciting possibilities for a directed interference in specific cellular processes.

See Also the Following Articles Amiodarone and Thyroid . Drug Effects and Thyroid Function . Iodine . Resistance to Thyroid Hormone (RTH) . Thyroid Hormone Action . Thyroid HormoneBinding Proteins . Thyroid Hormone Metabolism

Further Reading Cheng, S. Y. (2000). Multiple mechanisms for regulation of the transcriptional activity of thyroid hormone receptors. Rev. Endocr. Metab. Dis. 1, 9–18. Flamant, F., and Samarut, J. (2003). Thyroid hormone receptors: Lessons from knockout and knock-in mutant mice. Trends Endocrinol. Metab. 14, 85–90. Jungermann, K., and Kietzmann, T. (1996). Zonation of parenchymal and nonparenchymal metabolism in liver. Annu. Rev. Nutr. 16, 179–203. Lazar, M. A. (1993). Thyroid hormone receptors: Multiple forms, multiple possibilities. Endocr. Rev. 14, 184–193. O’Shea, P. J., and Williams, G. R. (2002). Insight into the physiological actions of thyroid hormone receptors from genetically modified mice. J. Endocrinol. 175, 553–570.

Thyroid Hormone Receptors

Sachs, L. M., Damjanovski, S., Jones, P. L., Li, Q., Amano, T., Ueda, S., Shi, Y. B., and Ishizuya-Oka, A. (2000). Dual functions of thyroid hormone receptors during Xenopus development. Comp. Biochem. Physiol. B 126, 199–211. Wagner, R. L., Apriletti, J. W., McGrath, M. E., West, B. L., Baxter, J. D., and Fletterick, R. J. (1995). A structural role for hormone in the thyroid hormone receptor. Nature 378, 690–697. Webb, P., Nguyen, N. H., Chiellini, G., Yoshihara, H. A., Cunha Lima, S. T., Apriletti, J. W., Ribeiro, R. C., Marimuthu, A., West, B. L., Goede, P., Mellstrom, K., Nilsson, S., Kushner,

495 P. J., Fletterick, R. J., Scanlan, T. S., and Baxter, J. D. (2002). Design of thyroid hormone receptor antagonists from first principles. J. Steroid Biochem. Mol. Biol. 83, 59–73. Yen, P. M. (2001). Physiological and molecular basis of thyroid hormone action. Physiol. Rev. 81, 1097–1142. Zandieh Doulabi, B., Platvoet-ter Schiphorst, M., van Beeren, H. C., Labruyere, W. T., Lamers, W. H., Fliers, E., Bakker, O., and Wiersinga, W. M. (2002). TR(beta)1 protein is preferentially expressed in the pericentral zone of rat liver and exhibits marked diurnal variation. Endocrinology 143, 979–984.

475

Thyroid Hormone-Binding Proteins

Table I Human Thyroid Hormone-Binding Proteins Bound hormone (%) Proteins

T4

Characteristic

a

Thyroxine-binding globulin

66

75

Transthyretin

11

10

Albumin

20

9

Minor THBPs Lipoproteins

a

Protein

T3

Major THBPs

Immunoglobulin M Immunoglobulin G

Table II Characteristics of the Major Thyroid HormoneBinding Proteinsa

3 ? ?

6 ? ?

THBPs, thyroid hormone-binding proteins.

oligosaccharide chains with an average of 10 sialic acid residues. Carbohydrates affect the half-life of TBG in serum since deglycosylation is associated with rapid clearance of the protein by the liver. In addition, carbohydrate removal slightly reduces TBG immunoreactivity and T4-binding activity and decreases its stability. TBG has only one binding site for iodothyronines, and it binds T4 with a higher affinity than T3. The very high affinity of TBG for thyroid hormones explains why TBG, although present in serum at a much lower concentration than TTR or HSA, carries approximately 65% of T4 and 75% of T3 (Table I). TBG concentration in normal adult human serum ranges from 12 to 20 mg/liter (Table II), with a maximal T4-binding capacity of 0.14–0.25 mg T4/liter. TBG is detectable in the 12-week-old fetus; concentrations in newborns are higher than in adults, decline until midadulthood, and increase thereafter.

Molecular weight (kDa)

TBG 54

TTR

Albumin

56

66

Structure

Monomer

Tetramer

Monomer

Carbohydrates (%)

20

0

0

Association constant (M-1) T4 T3 Serum concentration (mg/liter) Half-life (days)

1  1010 9

2  105 6

1.5  106

1  10

1  10

2  105

12–20

250

40,000

5

2

15

Gene location (chromosome)

Xq22.2

18

4 (humans)

Site of synthesis

Liver

Liver

Liver

5 (mice) Choroid plexus Retina Pancreas

a

Abbreviations used: TBG, thyroxine-binding globulin; TTR, transthyretin.

ment because it may help maintain the appropriate T4 concentration in the central nervous system and favor its uniform distribution in different areas of the central nervous system. The TTR gene exists in a single copy located on chromosome 18 and is composed of four exons spanning 7.3 kilobase pairs. The 50 -flanking region has a highly conserved DNA sequence among species, suggesting a crucial role in the regulation of TTR gene expression.

Transthyretin TTR is a 56-kDa protein composed of four identical subunits, each containing 127 amino acids; it does not contain carbohydrates (Table II). It has two identical thyroid hormone binding sites, but normally only one of them is occupied. The normal serum TTR concentration is 250 mg/liter (Table II), corresponding to maximal binding capacity of 2 mg T4/liter. TTR binds approximately 10% of T4 and 10% of T3. In addition to thyroid hormone, TTR also binds retinol-binding protein and is therefore involved in vitamin A transport. Synthesis of TTR occurs mostly in the liver, but the protein is also expressed in the pancreatic islet cells, the retina, and the epithelial cells of choroid plexus in both rats and humans. TTR synthesized in the choroid plexus may play an important role in brain develop-

Albumin Albumin is a 66-kDa protein composed of 585 amino acids (Table II). It has a relatively strong binding site for thyroid hormone and several additional sites with much lower affinity. It does not contain carbohydrates. Its serum concentration is very high (40 g/liter), and the percentage of thyroid hormone bound to albumin is approximately 20% of T4 and 10% of T3. The human albumin gene consists of a single copy and is located on the long arm of chromosome 4, linked to vitamin D-binding a2-globulin, whereas in mice the gene is located on chromosome 5, close to the a-fetoprotein gene. There is 90% homology between

476 the human albumin gene and the corresponding gene in rodents.

Lipoproteins Lipoproteins are complex molecules composed of a protein moiety (apolipoproteins) and a lipid (both polar and nonpolar) moiety. They bind approximately 3% of T4 and 6% of T3. High-density lipoproteins are the major lipoprotein plasma carriers of thyroid hormones through a specific interaction with their apolipoproteins (A-I, A-II, A-IV, C-I, C-II, C-III, and E). These apolipoproteins have a single thyroid hormone binding site encoded by exon 3 (exon 2 for apolipoprotein A-IV) of the respective gene. The thyroid hormone binding site on apolipoproteins is distinct from the apolipoprotein portion that binds to cell lipoprotein receptors. The physiological role of thyroid hormone binding to lipoproteins remains to be defined, but lipoproteins may facilitate enterohepatic circulation, transplacental passage, and central nervous system distribution of thyroid hormones, and they may be involved in thyroid hormone delivery to target tissues with cell surface receptors for apolipoproteins.

VARIATIONS IN THYROID HORMONE-BINDING PROTEINS Acquired Variations TBG Many drugs and pathophysiologic conditions (Table III) are associated with changes in serum TBG concentration related to variations in either TBG synthesis or metabolic clearance rate. Hyperthyroidism and hypothyroidism cause a slight decrease and increase, respectively, in serum TBG levels due to an effect on liver synthesis of the protein. Pregnancy and estrogen therapy cause an increase in serum TBG concentration. This appears to be related to the longer half-life of TBG in the circulation because of estrogen-induced increased sialylation of the protein. Serum TBG values are also increased in patients with acute or chronic hepatitis and in a significant proportion of cases of hepatocarcinoma. Whereas in hepatitis the increase in TBG is probably the consequence of TBG release from damaged liver cells, in hepatocarcinoma the underlying mechanism may be increased liver synthesis of TBG. Patients with nephrotic syndrome have a reduced TBG concentration due to massive renal protein loss. Losses of TBG through peritoneal membrane are likely to account for the

Thyroid Hormone-Binding Proteins

Table III Acquired Thyroxine-Binding Globulin (TBG) Variations Condition/drug

Serum TBG concentration

Hyperthyroidism

Decreased

Hypothyroidism

Increased

Pregnancy

Increased

Acute and chronic hepatitis

Increased

Hepatocellular carcinoma

Increased

Nephrotic syndrome Chronic renal failure

Decreased Decreased

AIDS

Increased

Diabetic ketoacidosis

Decreased

Starvation

Decreased

Cushing’s syndrome

Decreased

Acromegaly

Decreased

Oat cell carcinoma

Increased

Estrogens Androgens

Increased Decreased

Anabolic steroid

Decreased

Glucocorticoids

Decreased

Perphenazine

Increased

5-Fluorouracil

Increased

Heroin, methadone

Increased

Clofibrate

Increased

l-Asparaginase Interleukin-6

Decreased Decreased

decrease in TBG concentration observed in patients with chronic renal failure undergoing regular peritoneal dialysis. Serum TBG (but not CBG) levels are increased in AIDS patients, possibly due to associated hepatitis or to a specific enhancement of TBG hepatic synthesis. Patients with diabetic ketoacidosis often have decreased serum TBG levels, which might be related to the lack of stimulation of liver protein synthesis by insulin. Starvation or extreme protein-calorie malnutrition cause a decrease in serum TBG concentration likely related to decreased hepatic synthesis of the protein. These effects, as well the decrease in TBG that occurs in severe terminal illness, may be mediated by inhibition of TBG synthesis caused by interleukin-6. Minor variations in serum TBG concentration have been reported in several other pathophysiologic conditions (Table III). In addition to estrogens, other drugs cause an increase in serum TBG concentration, including 5-fluorouracil, clofibrate, heroin, and methadone (Table III). Conversely, administration of androgens, anabolic steroids, glucocorticoids, and l-asparaginase has been associated with decreased TBG levels in the circulation (Table III).

477

Thyroid Hormone-Binding Proteins

TTR Serum TTR concentration is often decreased in patients with severe nonthyroidal illness, particularly during protein-calorie malnutrition, nephrotic syndrome, liver diseases, and cystic fibrosis (Table IV). In such circumstances, serum TTR levels decrease, whereas TBG and albumin concentrations remain normal. Both decreased liver synthesis of TTR (possibly mediated by interleukin-6) and its accelerated degradation contribute to these changes. TTR may be increased in patients with pancreatic endocrine tumors (insulinomas or glucagonomas) or gastrointestinal carcinoids, probably due to TTR synthesis by the neoplasm. TTR levels are increased in the central nervous system but not in the serum of patients with endogenous depression or with Parkinson’s disease (after adrenal medullary autotransplantation). These changes probably reflect an increased TTR synthesis by the choroid plexus. Many drugs affect serum TTR concentration, and the effect is often the converse of that on TBG. Thus, estrogens decrease serum TTR concentration and androgens, anabolic steroids, and glucocorticoids increase serum TTR concentration (Table IV). Although the underlying mechanisms are not completely understood, variations in TTR synthesis likely contribute to these changes. Albumin Albumin concentration is decreased in many acute and chronic nonthyroidal illnesses. These variations occur concomitantly and are always associated with the previously mentioned similar changes in serum TBG and TTR concentrations.

Table IV Acquired Transthyretin (TTR) Variations Condition/drug

Serum TTR concentration

Protein-calorie malnutrition

Decreased

Nephrotic syndrome

Decreased

Liver diseases

Decreased

Cystic fibrosis

Decreased

Insulinoma Glucagonoma

Increased Increased

Gastrointestinal carcinoids

Increased

Depression

Increased

Parkinson’s disease

Increased

Estrogens

Decreased

Androgens

Increased

Anabolic steroids

Increased

Glucocorticoids

Increased

a

a a

TTR concentration is increased in cerebrospinal fluid but not in serum.

Inherited Variations TBG Familial forms of TBG deficiency and TBG excess, both inherited as X-linked traits, exist. These defects involve the TBG gene rather than the rate of TBG disposal. Complete TBG deficiency, partial TBG deficiency, and TBG excess are distinguished according to serum TBG levels in hemyzygous subjects. When TBG deficiency is complete, affected males have no detectable TBG in serum, whereas carrier females have half the normal serum TBG levels. In partial TBG deficiency, serum TBG concentration in heterozygous females is usually higher that half the normal value. In the presence of excess TBG, serum concentration of the protein is usually two- to fourfold higher than normal. Complete TBG deficiency occurs in approximately 1 in 15,000 newborn males. Eleven TBG variants account for complete TBG deficiency. In most cases, a single nucleotide substitution, a frameshift due to nucleotide deletion, or multiple nucleotide deletions are the mechanisms leading to early termination of translation and truncation of the TBG molecule. Mutations may also occur outside the coding region of the TBG gene. In a family with complete TBG deficiency, no mutations were detected either in the coding or in the promoter regions of the gene. Partial TBG deficiency occurs in 1 in 4000 newborns. Six different TBG variants cause variable degrees of decreases in serum TBG concentration. Some of these variants are unstable, have a reduced binding affinity for T4 and T3, or show an abnormal migration pattern on isoelectric focusing. A Japanese family with partial TBG deficiency has been reported with normal thyroid hormone-binding affinity, normal isoelectric focusing pattern, normal heat stability, and no mutations in the TBG gene coding region. In this family, the hereditary transmission appeared to be autosomal dominant. Inherited TBG excess is a rare condition, occurring in approximately 1 in 25,000–30,000 newborns. The pathophysiological basis of TBG excess has been shown to be TBG gene amplification (duplication and triplication), whereas no mutations in the coding and promoting regions have been detected. TTR Many TTR variants characterized by single amino acid substitutions have been described, most in patients with familial amyloidotic polyneuropathy, amyloidotic cardiomyopathy, or senile systemic amyloidosis. Some of these TTR variants have a reduced binding

478 affinity for thyroid hormone. A different TTR variant characterized by an increased affinity for T4 is responsible for a pattern of euthyroid hyperthyroxinemia (i.e., TTR-associated hyperthyroxinemia). Albumin A well-characterized inherited albumin variation transmitted as an autosomal dominant trait is familial dysalbuminemic hyperthyroxinemia (FDH), which is characterized by the presence in serum of an albumin variant with increased affinity for thyroid hormones. In many cases, the albumin variant has increased affinity for T4 only; in other instances, an increased affinity for T3 and/or reverse T3 is also present. Three different single nucleotide substitutions have been identified as the molecular basis for the increased albumin affinity for thyroid hormone. The inherited absence of albumin (analbuminemia) and the polymorphism called bisalbuminemia have negligible effects on thyroid hormone transport because the decrease in albumin levels is partially compensated for by a slight increase in TBG and TTR levels.

EFFECTS OF VARIATIONS IN THYROID HORMONE-BINDING PROTEINS ON THYROID FUNCTION TESTS Variations in THBP concentration or affinity profoundly affect serum total thyroid hormone concentrations. This is particularly true for TBG because it has a major role in thyroid hormone binding. Accordingly, a decrease or an increase in serum TBG concentration lead to a decrease or an increase, respectively, in serum total thyroid hormone levels. Although the latter changes are similar to those found in hypothyroidism and hyperthyroidism, respectively, they do not reflect thyroid hypofunction or hyperfunction because they are not associated with variations in the metabolically active, free (unbound) thyroid hormone fraction. Similar considerations are tenable for FDH and TTRassociated hyperthyroxinemia. Therefore, TBG excess, FDH, and TTR-associated hyperthyroxinemia are among the most important causes of euthyroid hyperthyroxinemia. The latter may be independent of THBP variations and caused by drugs (e.g., amiodarone, propranolol, iodinated contrast agents, and l-thyroxine), resistance to thyroid hormones, or the acute phase of some psychiatric disorders (Table V). Thus, should serum total thyroid hormone measurement provide results that are in contrast with the clinical picture, a THBP abnormality should be suspected and searched for. The correct definition

Thyroid Hormone-Binding Proteins

Table V

Causes of Euthyroid Hyperthyroxinemia

TBG excess Familial dysalbuminemic hyperthyroxinemia Transthyretin-associated hyperthyroxinemia Amiodarone Propranolol Iodinated contrast agents l-Thyroxine Resistance to thyroid hormone Acute phase of psychiatric disorders

of thyroid status requires measurement of serum free thyroid hormones and thyrotropin concentrations. This approach is particularly useful when THBPs (e.g., TBG excess or FDH) coexist with thyroid disorders, such as Graves’ disease or Hashimoto’s thyroiditis. In these circumstances, serum total thyroid hormone levels may be normal in hypothyroid patients, whereas the increased levels of hyperthyroid patients may not easily be distinguished from the increased concentrations due to THBP abnormalities. Because serum free thyroid hormone determination is crucial for the assessment of thyroid status and to avoid inappropriate treatment for hyperthyroidism or hypothyroidism, it is essential to select methods for free thyroid hormone measurement that are not affected by the abnormal THBP concentration or affinity. The two-step methods in which free hormone is first separated from protein-bound hormone by dialysis, ultrafiltration, column adsorption chromatography, or immunoadsorption provide the most reliable results. In fact, in the second step (immunoassay) the tracer is not in contact with THBP, thus preventing interaction between the two and the consequent artifactual results.

CONCLUSION THBPs exert functions that are important for thyroid physiology. They provide a buffering action, preventing abrupt changes in serum thyroid hormone levels; function as a storage system for thyroid hormones; and are involved in targeted delivery of thyroid hormone at the tissue level, thus facilitating thyroid hormone cellular distribution. TBG is the major THBP in serum since it binds approximately two-thirds to threefourths of T4 and T3. Both inherited and acquired variations of the major THBPs (TBG, TTR, and albumin) have been demonstrated. These variations do not modify thyroid status but do affect the results of serum total thyroid hormone measurement and

Thyroid Hormone-Binding Proteins

may lead to incorrect diagnosis and inappropriate treatment for hyperthyroidism or hypothyroidism. Thus, for a correct definition of thyroid status, determination of free T4 and T3 by assays that are not influenced by THBPs is required.

See Also the Following Articles Resistance to Thyroid Hormone (RTH) . Thyroid Hormone Action . Thyroid Hormone Metabolism . Thyroid Hormone Receptors

Further Reading Bartalena, L. (1990). Recent achievements in studies on thyroid hormone-binding proteins. Endocr. Rev. 11, 47–74.

479 Bartalena, L., and Robbins, J. (1993). Thyroid hormone transport proteins. In ‘‘Clinics in Laboratory Medicine: Pathophysiology of Thyroid Disease’’ (G. C. Klee, ed.), pp. 583–598. Saunders, Philadelphia. Bartalena, L., Bogazzi, F., Brogioni, S., Burelli, A., Scarcello, G., and Martino, E. (1996). Measurement of serum free thyroid hormone concentrations: An essential tool for the diagnosis of thyroid dysfunction. Hormone Res. 45, 142–147. Benvenga, S., and Robbins, J. (1993). Lipoprotein–thyroid hormone interactions. Trends Endocrinol. Metab. 4, 194–198. Refetoff, S. (1990). Inherited thyroxine-binding globulin (TBG) abnormalities in man. Endocr. Rev. 10, 275–293. Robbins, J. (2000). Thyroid hormone transport proteins and the physiology of hormone binding. In ‘‘Werner’s and Ingbar’s The Thyroid’’ (L. E. Braverman and R. D. Utiger, eds.), 8th ed., pp. 105–120. Lippincott Williams & Wilkins, Philadelphia. Schussler, G. C. (2000). The thyroxine-binding proteins. Thyroid 10, 141–149.

497

Thyroid Imaging

t0005

Table I Nuclides of Iodine with Clinical Relevance Nuclide of iodine (I)

Half-life

127

I

Nonradioactive

123

I

13 h

Dose

Clinical uses Necessary for life; can be imaged by fluorescent scanning

100–300 mCi

For measurement of uptake, scanning of thyroid

3.7–7.4 MBq 123

I

13 h

1–5 mCi

For imaging thyroid cancer after surgical thyroidectomy

37–185 MBq 131

131

I I

8 days

1–10 mCi

For diagnostic images of thyroid cancer after surgical thyroidectomy

8 days

37–370 MBq 30–>200 mCi

For imaging thyroid cancer days after radioiodine treatment

1.1–7.4 MBq 124

I

4 days

5 mCi

Positron emitter for imaging; provides functional and volumetric information

185 MBq

tetrafosmin, 18F fluorodeoxyglucose (FDG), 111In octreotide, and 99mTc demercaptosuccinate (DMSA). These are shown in Table II and their roles in imaging thyroid cancer are described later.

(PET) cameras have better resolution than hybrid PET–gamma cameras.

NORMAL THYROID GLAND INSTRUMENTS The optimal instrument is a gamma camera fitted with a pinhole collimator. This provides high-resolution scans and allows anterior and oblique views to be produced. Tomographic images, or single photon emission computed tomography, provide better resolution but do not provide additional relevant information. For whole-body imaging, a dual-headed whole-body camera is recommended. For imaging positrons, dedicated positron emission tomography

t0010

The adult thyroid gland has an average weight of 15–20 g and appears as two pear-shaped lobes connected by an isthmus. The appearance on scan is variable, with some degree of asymmetry being common (Fig. 1). Iodide that is not trapped by the thyroid gland is primarily excreted in the urine, with small amounts trapped by the salivary glands, stomach, choroid plexus, and lactating breast tissue. Iodine trapped by nonthyroidal tissues is not organified. Recent intake of iodine-rich foods and drugs decreases radioiodine trapping by the thyroid gland

Table II Properties of Radionuclides and Radiopharmaceuticals Used in the Diagnosis of Thyroid Disorders Radionuclide

Usual administered dose

Main clinical use

Thallium 201

3–5 mCi 111–185 MBq

Has been used to differentiate malignant from benign nodules

99m

Tc-sestamibi

20–25 mCi

Whole-body scan for thyroid cancer as above

740–925 MBq

Has been used to differentiate malignant from benign nodules

99m

Tc-tetrafosmin

20–25 mCi

Whole-body scan for thyroid cancer as above

740–925 MBq

Has been used to differentiate malignant from benign nodules

1–10 mCi 37–370 MBq for scan

Minimal value for differentiated thyroid cancer Moderately useful in medullary cancer

15 mCi

Whole-body scan for thyroid cancer useful in Tg-positive, 131I-negative patients

555 MBq

Has been used to differentiate malignant from benign nodules

111

18

In-octreoscan

F-fluorodeoxyglucose

Whole-body scan for thyroid cancer, usually when Tg positive and 131 I scan negative

Useful in medullary cancer 99m

TcV dimercaptosuccinate (pentavalent DMSA)

15–20 mCi 555–740 MBq

Useful for medullary cancer but not available in the United States

p0030

498

Thyroid Imaging

HYPERTHYROIDISM (THYROTOXICOSIS) Thyrotoxicosis refers to the effects of excess thyroid hormone; hyperthyroidism implies that the excess hormones are produced and secreted by the thyroid. In the setting of thyrotoxicosis, diagnostic radioiodine studies are primarily used to differentiate low-uptake thyrotoxic conditions (e.g., silent thyroiditis), which are self-limited without treatment, from high-uptake conditions that persist unless treated (Table III). Uptake measurement of radioiodine can help determine optimal treatment doses of 131I in the latter case.

p0040

Thyrotoxicosis with High Uptake of Radioiodine (Hyperthyroidism) Figure 1 A normal thyroid scan. The imaging was completed

24 h after the oral ingestion of 7.4 MBq 123I. There is homogeneous uptake in both lobes, and the isthmus shows less uptake. The normal 24-h uptake is approximately 10–30% when the dietary iodine intake is approximately 500 mg daily.

and lowers the calculated uptake value. Similarly, administration of intravenous radiographic contrast agents impairs thyroidal iodine trapping.

NONTOXIC GOITER The most common cause of goiter worldwide is iodine deficiency. In iodine-deficient regions, the uptake of radioiodine is increased. Imaging of a diffuse (endemic) goiter in regions of low iodine intake reveals a uniformly enlarged gland with a relatively homogeneous pattern of uptake. The most common form of goiter in the United States is Hashimoto’s disease. The scan is seldom required for Hashimoto’s disease, but when obtained the appearance can vary from a diffusely enlarged gland with normal uptake to that similar to Graves’ disease, patchy uptake, or a gland with significantly reduced uptake. With increasing numbers of emigrants to the United States from iodine-deficient regions, multinodular goiter is becoming more common, especially in women and those of advanced age. Imaging typically demonstrates a markedly heterogeneous radioiodine distribution due to the presence of multiple nodules with varying degrees of function. This condition can progress to thyrotoxic nodular goiter if one or more of the nodules enlarge and develop autonomous function.

Graves’ Disease Graves’ hyperthyroidism is caused by autoantibodies to the receptor for TSH (TSI). These autoantibodies cause continuous production of thyroid hormone, and the thyroid is unresponsive to normal inhibitory feedback mechanisms (i.e., it is nonsuppressible). Imaging of the thyroid in Graves’ disease reveals a diffusely enlarged gland with uniformly increased accumulation of tracer throughout both lobes (Fig. 2). The 24-h uptake value is elevated, often in the range of 60–80%. Visualization of the pyramidal lobe is more

Table III

Causes of Thyrotoxicosis

Thyrotoxicosis with high uptake Graves’ disease

Thyrotoxicosis with low uptake Excess thyroid hormone Thyrotoxicosis factitia Thyrotoxicosis medicamentosa Hamburger thyrotoxicosis

Single toxic adenoma Single hot nodule Functioning nodule

Thyroiditis Subacute thyroiditis De Quervain’s

Toxic multinodular goiter

Silent thyroidits Postpartum thyroidits

Functioning pituitary tumor secreting TSH Excess iodine Contrast agents Pregnancy-associated tumor Hydatidiform mole Choriocarcinoma Struma ovarii

Medications, amiodarone

t0015

499

Thyroid Imaging

Figure 3 Examples of functioning hot nodules (A) in a euthyroid patient, (B) in a mildly hyperthyroid patient, and (C and D) in hyperthyroid patients. In C and D, there is suppression of normal thyroid. Figure 3C shows central degeneration, and Fig. 3D shows markers where edges of the nodule are palpated. Figure 2 A thyroid scan 24 h after ingestion of 7.4 MBq 123I in Graves’ disease. The uptake is increased (usually > 40%) and the gland is larger and plumper.

common in cases of autoimmune thyroid disease, perhaps due to increased stimulation of otherwise minimally functioning remnant tissue, and it is seen in more than 50% of patients. Toxic Multinodular Goiter Toxic multinodular goiters are more common in older patients and in those from areas of endemic iodine deficiency. Unlike euthyroid multinodular glands, these are accompanied by symptoms and signs of hyperthyroidism due to the unchecked, autonomous overproduction of thyroid hormone by one or more of the nodules. Scans reveal heterogeneous uptake of tracer by nodules with varying degrees of functionality. At least one nodule demonstrates increased uptake, accounting for the toxic aspect of this condition. Because the uptake tends to be less elevated in this condition than in Graves’ disease, and because the condition is more resistant to radiation than Graves’ disease, higher doses of therapeutic radioiodine are required to achieve successful ablation. Solitary Toxic Adenoma In single toxic adenoma (‘‘hot’’ nodule), imaging reveals high uptake within this nodule, with little or no uptake throughout the remainder of the gland due to suppression by low TSH (i.e., the remainder of the gland remains sensitive to the normal feedback mechanisms). Scans of toxic adenomas can also reveal

‘‘cold’’ (i.e., relatively photopenic) regions within the otherwise intense uptake of the adenoma due to necrosis (degeneration). Figure 3 shows the range of scan findings in functioning thyroid nodules. The uptake in a solitary toxic adenoma is often only mildly elevated or near the upper limit of normal. A functioning euthyroid nodule is likely benign, and the rate of progression to thyrotoxicosis is usually slow but increases when the nodule is large (>3 cm) and the patient is advanced in age.

Thyrotoxicosis with Low Uptake of Radioiodine Thyroiditis is a general term for a group of disparate conditions, several of which have thyrotoxic symptoms and signs and biochemical findings despite low iodine uptake by the thyroid gland. The excess thyroid hormones result from uncontrolled release of previously stored hormone from disrupted follicles. Imaging reveals minimal trapping of radioiodine by the thyroid gland, and uptake measurements are depressed (Fig. 4). Included in this category are subacute thyroiditis, silent thyroiditis, postpartum thyroiditis, and martial arts thyroiditis, which is also called traumatic thyroiditis. These conditions are usually transient and selflimited. Acute thyroiditis (thyroid abscess) also shows reduced uptake. Thyrotoxicosis with low uptake can also result from oversupply of thyroid hormone by exogenous sources. This can be intentional (thyrotoxicosis medicamentosa) in patients with thyroid cancer

500

Figure 4 A thyroid scan in a patient with subacute thyroiditis. There is no uptake by the thyroid. A radioactive marker over the sternal notch shows the expected site of the thyroid (arrow). or factitious when the patient conceals ingesting the medication. Patients who need thyroid hormone commonly take slightly more than a physiological dose because it makes them feel better. This is not factitious thyrotoxicosis because both patient and physician recognize the deception. Thyroid imaging reveals a pattern similar to that seen in thyroiditis, with minimal uptake of radioiodine and a markedly suppressed uptake value. Low thyroglobulin is an indicator of factitious thyrotoxicosis. In addition, hyperthyroidism can be secondary to ectopic overproduction of thyroid hormone (as in functioning thyroid metastases, trophoblastic disease, struma ovarii, or teratomas). Scanning outside the thyroid determines the source of thyroid hormone production. Excess intake of iodine can cause hyperthyroidism ( Jod Basedow effect). This is more common in patients who have been chronically iodine deficient and then are exposed to excess iodine.

THYROID NODULES Thyroid imaging using radiopharmaceuticals has a limited role in patients with a solitary thyroid nodule. Scintigraphy can be used to determine the functional status of such nodules, recognizing that nonfunctioning (cold) nodules have a greater likelihood of being malignant than do functioning (hot) nodules. The vast majority of nodules imaged, however, are nonfunctional (approximately 90%), and 10–15% of these

Thyroid Imaging

are malignant. Thus, the positive predictive value of scintiscan for thyroid cancer is low (on the order of 15%). This raises concern about the cost-effectiveness of scintigraphy. Because of the low positive predictive value of thyroid scan to diagnose cancer, patients with thyroid nodules who are euthyroid are better evaluated by fine needle aspiration (FNA). In contrast, when patients have a nodule and are hyperthyroid, scintigraphic imaging can be performed first to discriminate between a benign functioning nodule and a nonfunctioning nodule in the setting of Graves’ disease. The latter case is relatively uncommon, but it warrants further evaluation by FNA. This discussion regarding the relative malignant potential of hot and cold nodules applies to adults only and does not hold for children, in whom functional nodules have a relatively high likelihood of malignancy. 123I is preferred to 99mTcO4 because it provides more physiologic information and predicts response to therapeutic 131I for functional nodules. There are disparate results for 123 I and 99mTcO4. A cold 123I nodule that traps pertechnetate has a significant probability of being malignant.

ECTOPIC THYROID TISSUE The thyroid begins its embryologic development in the posterior oral cavity and subsequently migrates downward toward its final destination in the neck base. This migration may be arrested, leaving thyroid tissue in ectopic locations ranging from the base of the tongue to the pericardium. Ectopic thyroid tissue is usually hypofunctional, and the resultant elevation in TSH stimulates the ectopic tissue to grow larger over time. When ectopic thyroid tissue is present (above the neck base), a normal cervical thyroid gland is almost always absent. Diagnosis should be made using a combination of thyroid function tests and radioiodine scintigraphy. The embryologic descent is marked by the thyroglossal duct, and cysts can develop along this route. Thyroglossal duct cysts generally do not take up radioiodine and are thus better confirmed with ultrasound or other anatomic imaging studies. The pyramidal lobe is a normal remnant of the thyroglossal duct tract and can rarely be seen scintigraphically in normal scans and in approximately two-thirds of scans in autoimmune hyperthyroidism.

THYROID CANCER After thyroidectomy for thyroid carcinoma, radioablation with 131I can be used to eradicate any

p0080

501

Thyroid Imaging

remaining thyroid tissue within the thyroid bed or metastases to lymph nodes or distant sites. Wholebody imaging has an important role in the postoperative management of the thyroid cancer patient. A diagnostic scan prior to treatment is useful to determine the dose of therapeutic 131I to be prescribed based on uptake in the thyroid region and the presence or absence of uptake in lymph nodes or other metastatic sites. Opinions vary regarding the necessity of the diagnostic scan prior to the first radioablation, but follow-up using the diagnostic scan is well established. Imaging of the patient following ablation allows the physician to verify uptake of the radiopharmaceutical by the thyroid tissue and to identify any areas of uptake not seen on the diagnostic scan. Periodic follow-up imaging is performed to monitor for recurrence of thyroid cancer; the general guidelines proposed by some authorities recommend annual 131I imaging until two consecutively negative studies are found. A TSH level >25–30 mU/liter facilitates the detection of small amounts of thyroid tissue. We advise using a level of 50 mU/liter or higher. Levo-thyroxine is withdrawn for 4 weeks; alternatively, triiodothyronine (T3), which has a shorter half-life, is substituted for 4 weeks, allowing time for T4 to be metabolized, and then T3is discontinued for 2 weeks. The introduction of recombinant human TSH (rhTSH) has made it possible to image (and treat) with 131I, without rendering the patient hypothyroid. Studies show that rhTSH is almost equivalent to endogenous TSH stimulation for determining the presence or absence of cancer, provided scan and serum thyroglobulin values are obtained. Peak serum TSH levels using rhTSH can be higher (the mean value in our experience with >100 patients is 140 mm/liter) than those obtained after conventional withdrawal of thyroid hormone, but the time of stimulation is shorter. Patients prefer the rhTSH protocol because hypothyroidism is avoided. Since rhTSH has been studied extensively only in the diagnostic setting and patients scanned after surgery frequently require 131I therapy, it may be prudent to reserve rhTSH for follow-up when it is anticipated that the scan will be negative. There are reports of the value of rhTSH in therapy. In any situation in which prolonged hypothyroidism and sustained elevation of TSH would be disadvantageous, rhTSH should be considered (e.g., when the metastases are in confined anatomic spaces, such as the spinal cord, and expansion of these could cause clinical problems). The plasma inorganic iodine level is an important factor with regard to the amount of radioiodine that is trapped by the thyroid. Decreasing the intake of

iodine to 30–50 mg/day over 7–14 days increases the uptake two or three times, thus theoretically increasing the effectiveness of radioablation. A low-iodine diet is recommended for 2 weeks prior to radioiodine scanning (details of a low-iodine diet are available at www.Thyca.org). Diagnostic whole-body scanning is conducted 2–4 days after administration of 37–370 MBq 131I. 131I is used because its long halflife enables imaging after 48–96 h or more. Anterior and posterior whole-body images and spot scans of the neck with corresponding uptake measurements are obtained. Normal thyroid traps significantly more iodine than metastases; therefore, when there is a normal remnant it might have to be ablated prior to treating metastases. The importance of a skilled thyroid surgeon is emphasized. Lymph node metastases are usually in the lateral neck and less commonly in the mediastinum. Pulmonary metastases can be focal or diffuse, Whereas skeletal lesions are focal in nature. The sensitivity of diagnostic 131I scan for papillary cancer and follicular cancer has been reported to be 45–80%. The sensitivity of posttherapeutic 131 I scans is higher. Controversy exists regarding whether use of 131I for diagnostic scan can cause ‘‘stunning,’’ which is the inability of the thyroid tissue to take up a therapeutic dose of 131I secondary to radiation on the thyroid tissue by the diagnostic dose. Some investigators have not found the stunning effect after administration of 74 or 185 MBq of 131I. Stunning appears to be occur when larger diagnostic doses are prescribed and when there is a delay between testing and treatment. Reasons for lack of iodine uptake by cancers include genetic changes in the Naþ/I symporter, Hurthle cell types, and poorly differentiated follicular and papillary carcinomas. Retinoic acid has been used to promote redifferentiation and induce 131I uptake in thyroid cancers with previously 131I-negative papillary, follicular, and mixed cell-type tumors. These studies have shown mixed results, but overall this strategy appears to have minimal clinical impact.

Diagnostic Scanning with 123

123

I

I emits gamma rays at lower energies than does 131I, and it does not emit beta particles and is unlikely to induce thyroid stunning. At least one study has demonstrated a higher rate of ablation after 131I treatment when 123I was used in the diagnostic scan, indicating the possibility that 123I may replace 131I for wholebody scintigraphy (Fig. 5). A does of 74 MBq 123I has been shown to have the same overall effectiveness in diagnostic imaging as a dose of 74 MBq 131I.

502

f0025

Thyroid Imaging

Figure 5 (A, right) Anterior and posterior whole-body scan 24 h after a dose of 74 MBq 123I in a patient with thyroid cancer who had thyroidectomy and prior 123I therapy. There are two metastases in the posterior image (arrows). These are seen faintly in a rib and sacroiliac joint on bone scan (left, arrows). (B) A whole-body scan 10 days after a therapeutic dose of 7.4 GBq 123I. The two lesions seen on diagnostic scan are shown. In addition, there is a faint lesion in the low thoracic spine that was not imaged with 123I (arrow). There is significant uptake in the salivary glands. Liver uptake is due to metabolism of radioactive thyroid hormones.

Posttherapy Scanning A posttherapy 131I scan is usually performed 5–7 days following radioiodine ablation. Figure 6 shows wholebody 131I scans—first a diagnostic scan followed by a posttherapy scan and a second diagnostic scan 12 months later to determine whether 131I treatment was successful. There is a higher sensitivity and clearer delineation of lesions using posttherapy scans due to the higher dose of 131I. Some authors report that as many as one-third of patients had metastases to the lymph nodes and lungs seen on posttherapeutic study that were not seen on the diagnostic scan. In fact, some clinicians perform only a posttherapeutic scan. We do not recommend this approach because in our experience, the diagnostic scan determines how much therapeutic 131I to prescribe and the posttherapy scan seldom shows additional clinically relevant information.

Figure 6 (A) A diagnostic scan made 72 h after a dose of 74 MBq 123 I. There is uptake in the thyroid bed. Physiological uptake is present in the stomach and intestines. (B) Whole-body scan in the same patient 1 week after 3.7 GBq 123I therapy. There is uptake in the thyroid bed and a left cervical node, and there is also uptake in the liver and gut. There is no stunning. (C) Whole-body scan in the same patient 1 year after I-131 treatment. The scan was made 48 h after a dose of 74 MBq 123I. RhTSH was used to stimulate uptake. There is no evidence of disease. There is physiological uptake in the gut and nasopharynx.

503

Thyroid Imaging

False-Positive

131

I Scans

Physiologic uptake is seen in the salivary glands, nasal mucosa, gastric mucosa, small bowel, and colon. Excretion can also be identified in the bladder and bowel. These should not be confused with thyroid metastases. However, contamination by secretions or excretions can be misinterpreted as metastases. Other false positives have been reported in sinusitis, dental disease, tracheostomy, bronchiectasis, thymus, gallbladder, Meckel’s and Zenker’s diverticulum, psoriatic plaque, rheumatoid arthritis, hiatal hernia, and achalasia. Nonthyroidal cancers that have been mistaken as recurrent thyroid disease or metastases include salivary adenocarcinoma, meningioma, lung cancer, ovarian cancer, breast cancer, teratoma, neurilemoma, and gastric adenocarcinoma. Labeled thyroid hormone may also collect diffusely in the liver in the postablation scan, after sufficient production by residual thyroid tissue and concentration by hepatocytes.

Other Radionuclide Scanning Techniques p0105

Other scanning procedures that do not depend on iodine trapping can be employed for cancers that do not trap iodine. PET appears to be the first choice. When used in patients who are Tg positive and 131 I scan negative, PET has a sensitivity of approximately 60–80%. False-positive results may be due to uptake of FDG in the tense cervical muscles of anxious patients and in brown fat. Thallous-201 chloride is taken up by all types of thyroid cancer. The maximal cancer to background ratios are obtained 10–15 min after injection. Anterior and posterior whole body images are obtained. Neither a low-iodine diet nor thyroid hormone cessation are necessary. The sensitivities range from 45 to 94%. In one study, the detection rate of recurrent or metastatic thyroid carcinoma using 201Tl was similar to that of FDG PET, and the two modalities are mostly concordant as well as complementary to 131I scintigraphy. However, FDG PET is capable of providing better image quality. Sestamibi has been used to identify thyroid cancers and metastases; anterior and posterior whole-body images are obtained 10–20 min after injection. More than 90% of the tracer is found in the inner mitochondrial matrix. The sensitivity is 70–90%. FDG PET was found to be more sensitive in the detection of recurrent thyroid cancer than 99mTc sestamibi. 99mTc tetrafosmin has properties similar to those of sestamibi. This radiopharmaceutical is used most often in the detection of local recurrence and cervical lymph node metastases. The sensitivity of tetrafosmin scanning for

Figure 7 PET scan 1 h after intravenous injection of 555 MBq 18

F fluorodeoxyglucose. The patient had an elevated Tg but negative 123I diagnostic and three negative posttherapy scans. Abnormal uptake is seen in the left supraclavicular node. This node was removed surgically and Tg became undetectable.

the detection of metastases has been reported to be 70–90%. 111In octreotride, an analogue of somatostatin, has been most useful in imaging residual or metastatic medullary thyroid cancer since these neuroendocrine tumors express somatostatin receptors. 111 In octreotide scans have occasionally been useful in imaging differentiated thyroid carcinomas, especially in the case of Hurthle cell carcinoma. Figure 7 shows a positive PET scan in a patient who had an elevated Tg but negative diagnostic and posttherapy scans. Specificities are higher with Tg 5–10 mg/liter and have been correlated with increasing thyroglobulin levels. Misinterpretation of tense or active muscles in the neck and larynx has led to false positives; therefore, it is important that the patient remain relaxed during the procedure and avoid speaking or chewing. Incidental focal uptake seen in the thyroid during PET performed for other indications is highly suspicious for primary thyroid cancer.

IMAGING MEDULLARY THYROID CANCER When calcitonin remains high following surgery, residual medullary cancer is present. Noninvasive methods of detecting and imaging medullary cancer include 201Tl, 99mTc sestamibi, 111In octreotide, radioiodinated meta-iodobenzylguanadine, 131I antiCEA, and 99Tcm-labeled DMSA. PET is superior to computed tomography or magnetic resonance imaging in detecting metastatic and recurrent medullary thyroid cancer, and sensitivities as high as 76% have been reported. Patients with rapidly rising serum calcitonin levels during the first postoperative year benefit most from a PET scan.

504

See Also the Following Articles Graves’ Disease . Nontoxic Goiter . Thyroid Carcinoma . Thyroid Fine Needle Aspiration Cytology . Thyroid Nodule . Thyrotoxicosis: Diagnosis . Toxic Adenoma . Toxic Multinodular Goiter

Further Reading Gulzar, Z., Jana, S., Young, I., et al. (2001). Neck and wholebody scanning with 5-mCi dose of 123I as diagnostic tracer in patients with well-differentiated thyroid cancer. Endocr. Practice 7, 244–249. Haugen, B. R., Pacini, F., Reiners, C., et al. (1999). A comparison of recombinant human thyrotropin and thyroid hormone withdrawal for the detection of thyroid remnant or cancer. J. Clin. Endocrinol. Metab. 84, 3877–3885. McDougall, I. R. (1995). Whole body scintigraphy with radioiodine-131. A comprehensive list of false positives with some examples. Clin. Nucl. Med. 20, 869–875. McDougall, I. R. (1997). 131I treatment of 131I negative whole body scan and positive thyroglobulin in differentiated thyroid carcinoma: What is being treated? Thyroid 7, 669–672.

Thyroid Imaging

McDougall, I. R. (1997). 74 MBq radioiodine 131I does not prevent the uptake of therapeutic doses of 131I (i.e., it does not cause stunning in differentiated thyroid cancer). Nucl. Med. Commun. 18, 505–512. McDougall, I. R., Davidson, J., and Segall, G. M. (2001). Positron emission tomography and the thyroid with an emphasis on thyroid cancer. Nucl. Med. Commun. 22, 485–492. Pellegriti, G., Scollo, C., Giuffrida, D., Vigneri, R., Squatrito, S., and Pezzino, V. (2001). Usefulness of recombinant human thyrotropin in the radiometabolic treatment of selected patients with thyroid cancer. Thyroid 11, 1025–1030. Schlumberger, M. J. (1998). Papillary and follicular thyroid carcinoma. N. Engl. J. Med. 338, 297–306. Shankar, L. K., Yamamoto, A. J., Alavi, A., and Mandel, S. J. (2002). Comparison of 123I scintigraphy at 5 and 24 hours in patients with differentiated thyroid cancer. J. Nucl. Med. 43, 72–76. Siddiqi, A., Foley, R. R., Britton, K. E., et al. (2001). The role of 123 I-diagnostic imaging in the follow-up of patients with differentiated thyroid carcinoma as compared to 131I-scanning: Avoidance of negative therapeutic uptake due to stunning. Clin. Endocrinol. 55, 515–521. Singer, P. A., Cooper, D. S., Daniels, G. H., et al. (1996). Treatment guidelines for patients with thyroid nodules and well-differentiated thyroid cancer. Arch. Intern. Med. 156, 2165–2172.

522

Thyroid Nodule

EPIDEMIOLOGY

Figure 1 Patient with a clinically solitary thyroid nodule located in the midline of the anterior neck.

role of specific candidate genes in the etiology of nodular thyroid disease have not provided a clear picture, mostly because too small and too few families have been studied. Although single genes may play a role in certain families, it is thought that genetic heterogeneity (i.e., no single gene is either necessary or sufficient for disease development) is highly likely. The following sequence of events appears to lead to the development of nodular thyroid disease: First, iodine deficiency or other goitrogenic factors induce thyroid hyperplasia. Second, due to increased proliferation during this stage, mutagenesis is increased. In the case of hot or toxic nodules, these mutations confer constitutive activation of the cyclic AMP cascade (e.g., thyroid-stimulating hormone receptor and Gsa protein mutations). This eventually leads to stimulation of iodine uptake and metabolism, thyroid hormone synthesis and release, and hyperthyroidism. In the case of cold thyroid nodules, a similar mechanism, but with mutations in genes that favor dedifferentiation (e.g., ras oncogene), is suggested. These latter mutations initiate growth but not function of the affected thyroid cells.

Unfortunately, our knowledge is hampered by a lack of population-based longitudinal studies using sensitive diagnostic imaging (e.g., ultrasound) allowing distinction between uninodular and multinodular disease and morphologic as well as functional characterization. Despite these shortcomings, there is a clear pattern of increased thyroid nodularity with decreasing iodine intake. The fact that nodules exist in the face of iodine sufficiency and even iodine excess emphasizes the importance of other environmental etiologic factors acting in concert with genetic factors. In the Whickham survey (United Kingdom), a solitary thyroid nodule was present in 5.3% of women and 0.8% of men (6.6:1 ratio). Size and function of the nodules were not indicated. In the Framingham, Massachusetts, study, a solitary thyroid nodule was present in 6.4% of women and 1.6% of men. Both investigations used clinical evaluation (palpation) and were performed in an iodine-sufficient area. If ultrasound is used, the prevalence of thyroid nodules >10 mm is usually 20–30%, increasing with age and in areas with insufficient iodine intake. In autopsy studies, 50% or more have either single or multiple thyroid nodules. If investigated using isotope scintigraphy, approximately 10–15% of all nodules are autonomously functioning (taking up the isotope, hot or toxic), whereas 85–90% are nonfunctioning (cold, no isotope uptake). The incidence of clinical disease is estimated to be 0.1% by palpation, corresponding to a lifetime risk of 5–10%.

NATURAL HISTORY The natural history with respect to growth and function varies and is difficult to predict in a given patient since no specific growth parameters exist. Therefore, it is difficult to decide whether a patient can be monitored without treatment or should be offered treatment before the nodule grows any more. In the Framingham survey, new nodules appeared with an incidence of 1 per 1000 individuals per year, resulting in an estimated lifetime risk for developing a nodule of 5–10%. After exclusion of the minority of patients who have rapid growth and symptoms and clinical suspicion of malignancy, and who are therefore offered treatment, nodules on average do not change significantly over time. Nodules that increase in size are predominantly solid and carry a higher risk of harboring thyroid malignancy than those predominantly cystic, which are more prone to decrease in size or even disappear. In many patients, ultrasound

523

Thyroid Nodule

will identify additional nodules not evident at clinical investigation. Given time, many patients will be classified as having multinodular goiter. In the subgroup of hot nodules, the rate of evolution into a toxic nodule is approximately 4% annually. The risk is closely related to nodule size. If the nodule is >3 cm, the risk is 20% within 6 years, whereas the risk is only 2–5% if nodule size is 4 cm and partially cystic Compression symptoms (dysphagia, hoarseness, dyspnea)

thyroxine (T4) and free triiodothyronine (T3) in serum. If serum TSH is decreased on repeat examination, treatment of this hypermetabolic state should be offered independent of whether free T4 and/or free T3 are elevated, especially in the elderly. Isotope scintigraphy is recommended and will most likely demonstrate a functioning nodule. Most patients have normal serum TSH, including those with thyroid malignancy. Elevated serum TSH with or without decreased serum free T4 suggests that the patient has chronic autoimmune thyroiditis (Hashimoto’s thyroiditis). This can be verified by demonstrating thyroid autoantibodies in serum. Thyroid autoantibodies against thyroid peroxidase or thyroglobulin do not aid in the differentiation between malignant and benign nodules. However, they are markers of an increased risk of developing hypothyroidism (Hashimoto’s thyroiditis) and hyperthyroidism (Graves’ disease) spontaneously or secondary to surgery or treatment with radioactive iodine. Calcitonin, a hormone produced by the parafollicular C cells of the thyroid gland, is the only clinically relevant biochemical marker of medullary thyroid carcinoma, which accounts for approximately 5% of all thyroid carcinomas. Basal or pentagastrin-stimulated serum calcitonin measurement is more sensitive than thyroid biopsy in detecting medullary thyroid carcinoma. However, there is no consensus on its routine use in patients with thyroid nodules.

Laboratory Investigations The only relevant biochemical test that is routinely needed is serum thyrotropin (TSH), which if normal indicates normal thyroid function. Subnormal serum TSH values should lead to determination of free

Diagnostic Imaging Neck palpation is very imprecise with regard to the determination of thyroid nodule morphology and size. For this reason, imaging methods are increasingly

t0005

Thyroid Nodule

used, although no imaging method can accurately differentiate benign and malignant nodules.

Ultrasonography This very sensitive technique with a high resolution has had a dramatic impact on clinical practice. When used in patients with a goiter, it has been shown to alter management in more than half of these patients. The increasing and widespread use, whether initially or during follow-up, is related to high availability, low cost, little discomfort to the patient, and its nonionizing nature. It allows determination of total thyroid volume, individual nodule size and echogenicity, and morphology of extranodular tissue and the evaluation of regional lymph nodes. Color-flow Doppler provides additional information regarding regional blood flow and nodule vascularity (Fig. 3). It distinguishes solid from cystic lesions and aids in the performance of accurate biopsies, punctures, and therapeutic procedures, such as percutaneous ethanol injection and laser therapy. Although there is no ultrasonographic pattern, alone or in combination with other techniques that may be considered specific for thyroid malignancy, characteristics such as hypoechogenicity, microcalcifications, and increased nodular flow are all predictive of malignancy to some extent. However, fine needle aspiration biopsy, preferably guided by ultrasonography, is far more accurate for this distinction.

525 Scintigraphy Although the resolution of isotope imaging can be enhanced to 5–10 mm by tomography, this resolution is still far below that of ultrasonography. Therefore, scintigraphy is not so much used for evaluation of morphology as for evaluation of the regional uptake of the isotope and thereby the determination of functionality of the thyroid nodules (Fig. 4). Nodules with uptake by scintigraphy (hot or toxic) almost never harbor clinically important malignancy, although rare exceptions do exist. In an unselected population of patients with a thyroid nodule, 80–90% of the nodules were nonfunctioning (cold). The a priori risk of malignancy is probably no higher than 5% for such a nodule. Scintigraphy is inaccurate in estimating thyroid and nodule size as well as in diagnosing malignancy. Computed Tomography and Magnetic Resonance imaging Computed tomography (CT) and magnetic resonance imaging (MRI) are expensive, time-consuming, and not readily available for imaging thyroid nodules. Their major strength is their ability to diagnose and assess the extent of substernal/intrathoracic thyroid tissue much more precisely than any other method. Both methods are well suited for visualizing the trachea and demonstrating narrowing of the tracheal

Figure 3 Various appearances (morphological patterns) of a thy-

Figure 4 Various appearances (morphological patterns) of a thy-

roid nodule using ultrasonography. (A) Normal thyroid tissue. (B) Solitary solid hypoecchoic (dark) thyroid nodule in the right thyroid lobe (left) surrounded by normal thyroid tissue (medium gray). (C) Solitary cyst (black area) surrounded by normal thyroid tissue. (D) Multiple nodules with varying ecchogenicity (multinodular gland).

roid nodule using scintigraphy. (A) Normal uptake in two thyroid lobes. (B) Solitary nonfunctioning (cold) nodule in the right thyroid lobe (left). (C) Solitary functioning (hot or toxic) nodule in the left thyroid lobe (right). (D) Multiple nodules with varying degrees of isotope uptake (multinodular gland).

526 area or a decrease in its volume. However, this measure correlates poorly with lung function.

Thyroid Nodule

Table II Etiology of Thyroid Nodules and the Relative Distribution of Fine Needle Aspiration Biopsy Resultsa Etiology

Fine Needle Aspiration Biopsy Fine needle aspiration biopsy (FNAB) provides the most direct and specific information about a thyroid nodule and is used by virtually all thyroid specialists in the initial evaluation of a patient with a solitary thyroid nodule or a dominant nodule in a multinodular goiter. It is without complications, inexpensive, and easy to learn to perform. Use of FNAB has reduced the number of thyroid surgeries by approximately 50%, doubled the surgical yield of thyroid cancer, and reduced the overall cost of medical care for these patients by 25%. The technique involves the use of a 5- to 20-ml plastic syringe with a 22- to 27-gauge needle. The skin is cleaned with alcohol, and sometimes skin infiltration with 1 or 2 ml of 1% lidocain is used. The needle attached to the syringe is inserted perpendicular to the anterior surface of the neck. Negative pressure is applied, and as soon as bloody fluid in the hub of the needle appears, pressure is released and the needle withdrawn. No fluid should enter the syringe. If the nodule is a cyst or partly cystic, the aspiration should be followed by FNAB of any residual solid component. Investigation of the cyst sediment rarely aids in the diagnosis of malignancy. After withdrawal, the needle is detached and the specimen is evacuated onto a slide. The specimen should be smeared immediately. Often, air drying is used and a number of staining methods are available. Diagnostically useful FNAB specimens are obtained in approximately 80% of the cases and rebiopsy typically reduces the number of insufficient samples by half. The number of sufficient samples increases with operator experience, use of ultrasound guidance, the number of aspirations, when the nodule is solid, and with increasing cytopathologist experience, but it is highly dependent on the criteria used for adequacy of a sample. The relative distribution of FNAB results is given in Table II. Diagnostic accuracy of FNAB at large depends on the classification of the 10–15% of suspicious lesions, of which 15–25% are malignant. If regarded negative, sensitivity will decrease and specificity will increase. If regarded positive, the converse is true. Patients with suspicious, malignant, and nondiagnostic FNAB results (after reaspiration) should be operated on (Fig. 2). If this strategy is followed, the risk of postponing the diagnosis of malignancy in the approximately

Benign (no evidence of malignancy)

Distribution (%) 70

Colloid nodule Cyst Thyroiditis (acute, subacute, or chronic) Suspicious Follicular neoplasia Malignant

10 4

Follicular carcinoma Papillary carcinoma Medullary carcinoma (C-cell carcinoma) Undifferentiated carcinoma (anaplastic) Lymphoma Metastasis Nondiagnostic (insufficient)

b

16

a

Data are representative of the author’s institution. The number of nondiagnostic results can be halved by rebiopsy.

b

70% of cases in which nonsurgical therapy is an option can be reduced to 1%. Repeat FNAB during follow-up of nodules left untreated will virtually eliminate the risk of overlooking thyroid malignancy. Neither elaborate classification systems for suspicious FNAB findings nor the use of large-needle biopsy increase diagnostic accuracy. Attempts to include biochemical analysis of thyroid cyst fluid or immunodetection of various candidate molecules, such as thyroid peroxidase or lectin-related molecules, in the evaluation of thyroid cytology are still in the experimental stage.

TREATMENT There is no ideal treatment for the thyroid nodule. The optimal therapy varies depending on the size and morphology of the nodule and whether it is functioning (Fig. 2; Table III). Although nodules 1–1.5 cm or larger should undergo FNAB, treatment is often not necessary once malignancy has been ruled out. In the subcentimetric nodule, FNAB need not routinely be performed and treatment is rarely indicated.

Levothyroxine Therapy Although on the decline, it is still common practice to use thyroid suppression with levothyroxine (L-T4) in the management of solid thyroid nodules in the euthyroid patient. The aim is to shrink existing nodules, considered to be a favorable sign indicating that the

527

Thyroid Nodule

Table III

Advantages and Disadvantages of the Treatment Options for the Solitary Thyroid Nodule

Treatment

Advantages

Levothyroxine

Disadvantages

Outpatient

Low efficacy

Low cost

Lifelong treatment

May slow nodule growth

Regrowth after cessation

Possibly prevents new nodule formation

Adverse effects (bone and heart)

Prompt relief of symptoms

Inpatient

Nodule ablation Definite diagnosis

High cost Anesthesiological risk

Not feasible with TSH suppressed a

Surgery

Surgical risk Vocal cord paralysis Hypoparathyroidism Hypothyroidism Bleeding and infection Scar Radioiodine

b

Outpatient Low cost

40% size reduction Contraceptives needed in fertile women Side effects Radiation thyroiditis Graves’ disease Hypothyroidism Long-term cancer risk unknown

Ethanol injection

Outpatient

Repeat injections needed

Relatively low cost Thyroid function preserved

Low efficacy in large nodules Operator dependency c

Side effects Pain

Transient dysphonia Thyroiditis Extranodular fibrosis Complicates subsequent cytological interpretation a

In this case, unilateral operation limiting the risk of side effects. can only be used in the nodule with uptake, whether thyroid function is increased or not. Except for various degrees of pain, side effects are rare.

b It c

nodule is benign. TSH suppression seems most beneficial in the subgroup of patients with small, solid nodules. Approximately 20% of solitary solid nodules actually regress as a result of L-T4 therapy, and cessation of therapy leads to rapid regrowth. On average, long-term therapy is without significant nodulereducing effect. Growth can be suppressed or slowed, and the formation of new nodules may be prevented. However, this necessitates that serum TSH is suppressed to subnormal values, which may have adverse effects. This degree of TSH suppression, called mild or subclinical hyperthyroidism, is associated with an increased risk of atrial fibrillation, other cardiac side effects, and reduced bone density, potentially leading to osteoporosis. It is without effect in the cystic nodule

and in patients with spontaneously low serum TSH with or without elevated thyroid hormone levels. For these reasons, its use is questionable; at most, it can be used in younger patients with small nodules, in whom treatment is least necessary.

Surgery When there is malignant or suspicious cytologic features and/or symptoms due to the nodule, surgery is often recommended, especially for younger patients and in cases in which there are large nodules. The preferred operation is a unilateral removal of the affected lobe. The frequency of complications decreases with increasing experience and specialist

528 training and is generally low. Complications include temporary and permanent unilateral vocal cord paralysis (1–2 and 0.5–1.0%, respectively), temporary and permanent hypocalcemia (1.0 and 0.5%, respectively), and wound hematomas and infections (0.5 and 0.3%, respectively). The risk of complications increases with the extent of operation. In the patient with normal thyroid function postoperatively, there is no indication for routine L-T4 treatment since this does not seem to hinder thyroid growth in the long term, at least in iodine-sufficient regions. Although an option, surgery is rarely used in the hyperthyroid patient with a toxic nodule. Radioactive iodine treatment is the preferred treatment.

Radioactive Iodine If the patient has hyperthyroidism (toxic nodule), antithyroid drugs (propylthiouracil or methimazole) can normalize thyroid function but disease recurrence is the rule when medication is stopped. With the exception of a few patients who have a large nodule, in which case surgery may be indicated, radioactive iodine is the treatment of choice. This is also the case for the clinically euthyroid patient with a functioning (hot) nodule without hyperthyroidism, in whom treatment may be dictated by the nodule size, which may cause compression or cosmetic disturbances. In addition, radioactive iodine treatment is used to prevent hyperthyroidism (annual risk of approximately 4%). A cure rate (i.e., normalization of thyroid function and the appearance on a thyroid isotope scintigram) of 75% is seen, and the nodule shrinks 30–40% following a single dose of radioactive iodine. Side effects are few, with rare cases of radiation thyroiditis and transition to Graves’ disease. The risk of hypothyroidism is approximately 10% after 5 years and unrelated to the dose of radioactivity. The long-term risk of malignancy is unknown but considered negligible. Radioactive iodine has no effect in the nonfunctioning (cold) thyroid nodule, whether solid or cystic. In the future, the possibility of stimulation with recombinant human TSH before radioactive iodine treatment may lead to an increased iodine uptake and also an effect in the solid, cold nodule.

Thyroid Nodule

autonomously functioning thyroid nodules and nonfunctioning thyroid nodules, whether solid or cystic. If multiple injections are used, complete cure (normal serum TSH and isotope scintigraphy) can be achieved in 60–70% of patients with toxic nodules and 70–80% with hot nodules. A single ethanol instillation (after aspiration) in thyroid cysts reduces recurrence to approximately 20% compared to approximately 50% after aspiration alone. In solitary solid, nonfunctioning thyroid nodules, approximately 50% of patients are relieved of their clinical symptoms based on a 50% nodule volume reduction. Additional injections have little effect. It is an option for patients who do not wish to undergo radioiodine treatment or surgery. However, it often necessitates repeat treatment to obtain complete cure. The long-term effects are unknown, and the treatment is not devoid of side effects. The procedure, often used in Italian centers, is not a routine option, should still be classified as experimental, and requires the special technical skill that can be obtained only at a center familiar with interventional ultrasound. Ultrasound-guided interstitial laser photocoagulation for solid solitary, benign, nonfunctioning thyroid nodules was recently introduced. One treatment lasting approximately 10 min resulted in a nodule reduction of approximately 40% and significant reduction of pressure symptoms. These results are similar to those obtained using ethanol therapy. The fact that the spread of energy with a laser (thermal destruction) can be controlled, as opposed to chemical destruction by injection of ethanol, may favor laser therapy in the long term. This treatment option is experimental.

Acknowledgments This work was supported by the Agnes and Knut Mørk Foundation, the A. P. Møller Relief Foundation, and the Novo Nordisk Foundation.

See Also the Following Articles Iodine Deficiency . Iodine, Radioactive . Smoking and the Thyroid . Thyroid Carcinoma . Thyroid Disease, Genetic Factors in . Thyroid Disorders in the Elderly . Thyroidectomy . Thyroid Fine Needle Aspiration Cytology . Thyroid Imaging . Toxic Multinodular Goiter

Ethanol Injection

Further Reading

Ethanol (70–100%) causes local small vessel thrombosis and coagulative necrosis, leading to fibrosis and permanent tissue ablation. It has been used in both

Bennedbæk, F. N., and Hegedu¨s, L. (2000). Management of the solitary thyroid nodule. Results of a North American survey. J. Clin. Endocrinol. Metab. 85, 2493–2498.

Thyroid Nodule

Bennedbæk, F. N., Perrild, H., and Hegedu¨s, L. (1999). Diagnosis and treatment of the solitary thyroid nodule. Results of a European survey. Clin. Endocrinol. 50, 357–363. Døssing, H., Bennedbæk, F. N., and Hegedu¨s, L. (2002). Benign solitary solid cold nodules: US-guided interstitial laser photocoagulation—initial experience. Radiology 225, 53–57. Eszlinger, M., Krohn, K., and Paschke, R. (2003). Complementary DNA expression array analysis suggests a lower expression of signal transduction proteins and receptors in cold and hot thyroid nodules. J. Clin. Endocrinol. Metab. 86, 4834–4842. Hegedu¨s, L., and Bennedbæk, F. N. (2002). Management of the single thyroid nodule. In ‘‘Oxford Textbook of Endocrinology’’

529 ( J. A. H. Wass and S. M. Shalet, eds.). Oxford Univ. Press, Oxford. Hegedu¨s, L., Bonnema, S. B., and Bennedbæk, F. N. (2003). Management of simple nodular goiter: Current status and future perspectives. Endocr. Rev. 24, 102–132. Krohn, K., and Paschke, R. (2001). Progress in understanding the etiology of thyroid autonomy. J. Clin. Endocrinol. Metab. 86, 3336–3345. Ridgway, E. C. (2000). Clinical evaluation of solitary thyroid nodules. In ‘‘Werner and Ingbar’s The Thyroid: A Fundamental and Clinical Text’’ (L. E. Braverman and R. D. Utiger, eds.), pp. 949–956. Lippincott Williams & Wilkins, Philadelphia.

531

Thyroid Peroxidase

reactions: Iodotyrosine is formed at the active site of TPO through the reaction of iodide radicals with tyrosine radicals. In particular, TPO catalyzes twoelectron oxidation of tyrosine and iodide and forms an iodinating intermediate (TPO–Iþ) that reacts with thyrosine and produces monoiodothyrosine (3-iodotyrosine; MIT) with TPO. As another possibility, Taurog proposed a reaction between oxidized TPO and I
to produce hypoiodite (OI
), which also involves a two-electron change.

Oxidative Coupling by TPO TPO catalyzes two-electron oxidation of iodide and tyrosine to form MIT; further reaction of the MIT radical with iodide (one-electron oxidation) gives diiodotyrosine (3,50 -diiodotyrosine; DIT) through formation of DIT radicals. No matter the precise nature of the iodinating species, it is clear that iodide is oxidized by H2O2 and TPO and transferred to the tyrosil group within the Tg peptide chain. Not all of Tg’s tyrosyls are equally accessible to iodination. Isolated from the thyroid, Tg rarely contains more than 1% iodine, or approximately 56 atoms of iodine per 660kDa molecule of Tg. The molecule has approximately 134 tyrosyl residues among its two identical chains; thus, at most, only one-third of the tyrosyls are iodinated. After DIT and MIT formation, two residues of DIT couple to make l-thyroxine (tetraiodothyronine; T4) or one DIT and one MIT make l-triiodothyronine (triiodothyronine; T3), all still within the Tg molecule. In this reaction, further oxidation mediated by TPO and H2O2 produces an iodophenyl free radical, leaving T4 or T3 at the acceptor site and dehydroalanine at the donor position.

Regulation of Thyroid Hormone Synthesis by TPO From the kinetic data on the iodinating and coupling reactions of free tyrosines catalyzed by TPO, it has been concluded that the mechanism of the enzyme action fits the preferential formation of T4, even though the formation started from free tyrosine. The native structure of Tg plays an important role in the preferential formation of T4. This is supported by the fact that a specific peptide structure of Tg is involved in the biosynthesis of thyroid hormone: The hormonogenic tyrosine residues are iodinated in rigid sequential order, resulting in the formation of DIT derivates via MIT, which subsequently undergo oneelectron oxidation to form T4. It is important to note

that thyroid hormones formed in Tg are prevented from further oxidation by TPO, whereas free iodothyronines are readily oxidized by TPO.

GENE STRUCTURE OF TPO The cDNAs encoding TPO have been isolated in man, pig, rat, and mouse. Kimura et al. cloned two different cDNAs of hTPO: one of 3048 nucleotides [base pairs (bps)] called TPO1and the second (TPO2) of 2877 bps. TPO1 coded for a protein of 933 residues and a molecular mass of 103 kDa, whereas TPO2 was identical except that it lacked exon 10 and had 1 bp change, coding for a protein of 96 kDa. Both forms occur in normal and abnormal human thyroid tissue. TPO2 appears enzymatically inactive because it does not bind heme, degrades rapidly, and failed to reach the cell surface in experiments with stable cell lines. There are different degradative pathways for the two forms. The TPO gene resides on chromosome 2p13, spans more than 150 kbps, and has 17 exons. It contains domains similar to those of acetylcholinesterase, low-density lipoprotein receptor, and insulin-like growth factor receptor. Several types of mutations of the TPO gene cause diminished iodide organification. TPO shares with NIS, Tg, and the TSH receptor the regulation of its gene expression by thyroid-specific transcription factors (TTFs), such as TTF-1, TTF-2, and Pax-8. Tg and TPO have the same binding sites for TTF-1, TTF-2, and Pax-8 in their promoters, and the genes for both have TTF-1 binding sites in their enhancer regions. TPO is synthesized on polysomes and transported to the Golgi, where it undergoes glycosylation; it is then packaged into exocytotic vesicles along with Tg. These vesicles fuse with the apical membrane in a process stimulated by TSH, and TPO is then found in the membrane associated with microvilli. Yokoyama and Taurog suggest that the C-terminal portion of the TPO molecule is in the cell cytoplasm, that the portion from residues 845 to 870 is in the apical membrane, and that the remainder, including residues 1–844, lies in the thyroid follicular lumen. TPO enzymatic activity is restricted to the apical membrane, but most of the thyroid’s total amount of TPO is intracellular at the endoplasmic reticulum and perinuclear membrane. This intracellular protein is inactive due to improper folding and contains only high-mannose-type carbohydrate units; in contrast, the membrane TPO has complex carbohydrate units, essential for enzymatic activity. Chronic TSH stimulation increases the amount of TPO and its concentration at the apical membrane.

532

MUTATIONS OF THE TPO GENE Defective organification of iodine is due to abnormalities of Tg and TPO synthesis or H2O2 production. Organification defect in iodine caused by abnormal H2O2 production is rare, so abnormal Tg and TPO synthesis is thought to be the major cause of defective organification of iodine. The prevalence of neonatal hypothyroidism is approximately 1/4000, one-fifth of which is caused by genetically determined thyroid dyshormonogenesis. Because defective TPO activity is one of the two major causes of defective organification of iodine, approximately 1 in 40,000 newborns has hypothyroidism due to defective TPO. The results of clinical observations, however, suggest that compensatory hyperplasia of the thyroid tissue partially compensates hypothyroidism when TPO activity is borderline. Therefore, more than 1 in 40,000 newborns may have milder forms of hypothyroidism caused by defective TPO. Defects of TPO are both quantitative and qualitative; the latter include impaired binding to heme, impaired binding to Tg or iodine substrates, abnormal localization in thyrocyte, and abnormal susceptibility to inhibition. More than 110 cases of hereditary defective organification of iodine ascribable to defective TPO have been reported, and the hereditary form is autosomal recessive. Pedigree maps of these families show consanguineous marriage in many cases. The TPO gene mutations responsible for congenital goitrous hypothyroidism have been identified, causing either abnormal TPO with low or absent enzymatic activity or complete absence of TPO protein formation. It has been shown that Pendred’s syndrome (an autosomal recessive disease characterized by defective iodine organification with goiter and congenital neurosensory deafness) is due to mutations of the pendrin gene. The decreased organification activity of this condition may be due to abnormal pendrin–TPO interactions of the apical membrane of thyroid follicular cells.

TPO AS AUTOANTIGEN TPO, Tg, and TSH receptor are the major thyroid autoantigens identified at the biochemical and molecular levels. In particular, TPO was identified in 1985 as the ‘‘thyroid microsomal antigen’’ reacting with thyroid microsomal autoantibodies, first described in the late 1950s. Anti-TPO autoantibodies (TPO-Ab) are detected together with anti-Tg (TgAb) and anti-TSH receptor (TR-Ab) autoantibodies in sera of patients with autoimmune thyroid diseases (AITD); the majority of TPO-Ab are IgG1 or IgG4.

Thyroid Peroxidase

Epitopes recognized by autoantibodies have been extensively studied using synthetic peptides, recombinant DNA molecules, and recombinant Fabs specific for TPO. B-cell epitopes are generally conformational, whereas T-cell epitopes are short, linear peptide fragments. In studies of B-cell epitopes with recombinant TPO, it has been found that many B-cell epitopes are located on the latter half of the COOH end of the TPO molecule and that the region between amino acids 590 and 767 contains an epitopic hotspot; approximately 80% of TPO-Ab recognize two conformational epitopes. TPO-Ab are detected in almost all patients with Hashimoto’s thyroiditis and in the majority of those with Graves’ disease. Although TPO-Ab mediate in vitro complement-dependent lysis and antibody-dependent, cell-mediated cytolysis of thyroid cells, they are probably devoid of cytolytic activity in vivo due to the inaccessibility of the apical membrane of the thyroid follicular cell. Some, but not all, TPO-Ab are able to inhibit TPO enzymatic activity in vitro, but the relevance of this phenomenon in vivo is probably minimal. Taken together, these data strongly support the concept that serum TPO-Ab are markers of thyroid autoimmunity but are not directly involved in thyroid cell damage. Sera of AITD patients also contain autoantibodies that recognize cross-reacting epitope(s) between Tg and TPO (TGPO-Ab). The exact nature of TGPOAb remains to be clarified. Results of primary lymphocyte cultures with synthetic peptides have produced a degree of consensus regarding the TPO T-cell epitopes recognized in patients. Three regions of TPO protein probably contain T-cell epitopes: amino acid residues 110–250, 414–589, and 841–901. Amino acid residues 119–126 of TPO have been predicted from algorithms to be a T-cell epitope common to both TPO and Tg, and they have been proven to aid in the activation of thyroiditogenic cells. Porcine TPO can induce thyroiditis in mice. Interestingly, murine strains with a high incidence of thyroiditis induced by porcine TPO are quite different from those developing thyroiditis after immunization with Tg.

See Also the Following Articles Thyroid Autoimmunity . Thyroid Hormone Action . Thyroid Hormone Metabolism

Further Reading De Vijlder, J. J. M., Dinsart, C., Libert, F., Geurts van Kessel, A., Bikker, H., Bolhuis, P. A., and Vassart, G. (1988). Regional localization of the gene for thyroid peroxidase to human chromosome 2pter p12. Cytogenet. Cell Genet. 47, 170–172.

p0055

Thyroid Peroxidase

Dunn, J. T., and Dunn, A. D. (2001). Update on intrathyroidal iodine metabolism. Thyroid 11(5), 407–414. Kimura, S., Kotani, T., McBridge, O. W., Umeki, K., Hirai, K., Nakayama, T., and Ohtaki, S. (1987). Human thyroid peroxidase: Complete cDNA and protein sequence, chromosome mapping, and identification of two alternately spliced mRNAs. Proc. Natl. Acad. Sci. USA 84, 5555–5559. Kopp, P. (1999). Pendred’s syndrome: Identification of the genetic defect a century after its recognition. Thyroid 9(1), 65–69. Mariotti, S., Caturegli, P., Piccolo, P., Barbesino, G., and Pinchera, A. (1990). Antithyroid peroxidase autoantibodies in thyroid diseases. J. Clin. Endocrinol. Metab. 71, 661–669. Mclachlan, S. M., and Rapoport, B. (1995). Genetic and epitopic analysis of thyroid peroxidase (TPO) autoantibodies: Markers of

Thyroid-Stimulating Hormone see TSH

533 the human thyroid autoimmune response. Clin. Exp. Immunol. 101, 200–206. Ohtaki, S., Nakagawa, H., Nakamura, M., and Kotani, T. (1996). Thyroid peroxidase: Experimental and clinical integration. Endocr. J. 43(1), 1–14. Rapoport, B., and McLachlan, S. M. (2001). Thyroid autoimmunity. J. Clin. Invest. 108, 1253–1259. Taurog, A. M. (2000). Hormone synthesis. Thyroid iodine metabolism. In ‘‘The Thyroid. A Fundamental and Clinical Text’’ (L. E. Braverman and R. D. Utiger, eds.), 8th ed. Lippincott Williams & Wilkins, Philadelphia. Yokoyama, N., and Taurog, A. (1988). Porcine thyroid peroxidase: Relationship between the native enzyme and an active, highly purified tryptic fragment. Mol. Endocrinol. 2, 838–844.

370 amiodarone, iopanoic acid, propranolol, and glucocorticoids), or decreased TSH secretion (e.g., dopamine and its agonists, octreotide and glucocorticoids). Hypothyroidism and thyrotoxicosis may develop during therapy with iodine-containing drugs, and subclinical or overt hypothyroidism may be observed with long-term lithium therapy. Cytokines (e.g., interferon-a or interleukin-2) may precipitate hypothyroidism, thyrotoxicosis, or the biphasic pattern of silent thyroiditis, especially in the presence of preexistent thyroid autoimmunity. In hypothyroid patients taking l-T4, drugs such as ferrous sulfate, cholestiramine, cholestipol, and soybean formulations may interfere with l-T4 absorption.

THYROID DISEASES IN THE ELDERLY Hypothyroidism The prevalence of hypothyroidism in the elderly is high (0.5–6% for overt and 4–15% for subclinical hypothyroidism). Autoimmune thyroiditis is the main cause, but iatrogenic hypothyroidism (radioiodine administration, thyroid surgery, head and neck radiation, and antithyroid drugs) is also common. Excess iodine from amiodarone or iodinated radiographic contrast agents may induce hypothyroidism, preferentially in glands with preexisting autoimmunity. Hypothyroidism in the elderly develops insidiously and often lacks classic clinical features, and some manifestations may be erroneously attributed to ‘‘normal’’ aging or ageassociated diseases. Unexplained increases in serum cholesterol, macrocytic anemia, severe constipation, congestive heart failure with restrictive cardiomyopathy, and subtle neurologic signs are the most common manifestations. Severe depression, lethargy, memory loss, apathy, and, rarely, psychosis or irreversible dementia may be observed. Elderly patients are more susceptible to myxedema coma, which may be precipitated by intercurrent NTI or cold exposure. The diagnosis of primary hypothyroidism is based on increased serum TSH, although the nonspecific effects of NTI and/or drugs must be taken into account. Decreased thyroid hormone concentrations may be observed in both hypothyroidism and NTI, but low FT4 is more frequent in thyroid failure. Anti-thyroid antibody tests help to differentiate autoimmune from nonautoimmune causes of hypothyroidism. Subclinical hypothyroidism (increased serum TSH with normal FT4 concentration) occurs frequently in the elderly. Like overt thyroid failure, most cases are due to autoimmune thyroiditis or to previous treatment of hyperthyroidism. Progression to overt

Thyroid, Aging and

hypothyroidism occurs in 2–18% of cases per year, with the highest percentages observed for subjects with high serum thyroid antibody titers. Subclinical hypothyroidism is associated with a slight but significant increase in serum lipids. Therapy should be initiated with low doses of l-T4 (12.5–25 mg/day) followed by incremental increases (12.5–25 mg/day every 4–8 weeks) until full replacement (1.1 or 1.2 mg/kg body weight) after several months. Particular attention should be paid to patients with coronary disease to avoid angina or myocardial infarction. Indication for therapy in subclinical hypothyroidism is controversial. Serum TSH >10 mU/liter and/or high thyroid antibody titers favor treatment, particularly when hypercholesterolemia and hypothyroid symptoms are present. Careful consideration of potential adverse reactions is always required.

Hyperthyroidism The prevalence of hyperthyroidism in the elderly is 0.5–2%, mostly due to Graves’ disease, toxic nodular goiter (TNG), and toxic adenoma. The relative frequency of TNG is higher than in young patients, especially in areas of iodine deficiency. Iodine-induced hyperthyroidism occurs frequently in elderly patients using iodine-containing medications or radiographic contrast media, and amiodarone-induced thyrotoxicosis is the most frequent form. Elderly hyperthyroid patients frequently display few signs and symptoms, hence the terms apathetic or masked hyperthyroidism. Eye signs are often lacking, but Graves’ ophthalmopathy, when present, is usually worse. Tachycardia is less common, but a high prevalence (25–35%) of atrial fibrillation is found in thyrotoxic elderly male patients. Weight loss is often associated with anorexia rather than increased appetite; muscle wasting and weakness are common, and the high risk of osteoporosis and bone fracture typical of old age is increased. Neuropsychiatric symptoms may be reported as primary manifestations. Elevated serum FT4 and/or FT3 and low TSH levels by sensitive assay establish the diagnosis. Serum TSH may be low in euthyroid severely ill patients with coexistent NTI, but undetectable rather than low serum TSH by third-generation assays strongly suggests hyperthyroidism. However, discrimination between the two conditions may not be possible without concomitant FT3 and FT4 assay. Subclinical hyperthyroidism (low TSH and normal FT3 and FT4) must be distinguished from other

371

Thyroid, Aging and

causes of low serum TSH, such as NTI. The prevalence of low serum TSH among elderly people is high (1.5–12.5%), but the number of cases progressing to overt hyperthyroidism is low (2–10%). Subclinical hyperthyroidism may be due to ‘‘subclinical’’ Graves’ disease, autonomously functioning thyroid nodules, or excessive/inappropriate thyroid hormone therapy. Subclinical hyperthyroidism in the elderly is associated with increased risk of atrial fibrillation and mortality from cardiovascular diseases, as well as reduced bone density in postmenopausal women. Long-term therapy with antithyroid drugs is not recommended for elderly patients because of the high relapse rate after withdrawal and the increased incidence of adverse effects. Radioiodine (131I) is the treatment of choice since it results in a definitive cure and avoids the risks of surgery. Euthyroidism should be restored with antithyroid drugs before 131I therapy, and control of heart rate with beta-blockers (or calcium channels blockers) should be obtained before and after therapy, as long as the patient remains thyrotoxic. Long-term follow-up of thyroid function is mandatory and eventual hypothyroidism must be corrected. Most cases of subclinical hyperthyroidism in the elderly require active treatment due to cardiovascular and bone complications.

Nontoxic Goiter and Thyroid Carcinoma An age-dependent increase in thyroid volume and nodularity has been documented by echography, reaching a prevalence after 60 years of age of approximately 50% in iodine-deficient areas. Because surgical risks are higher in the elderly, most nontoxic nodular goiters are managed conservatively, unless there is strong suspicion of malignancy or significant airway obstruction. In patients with contraindications to surgery, radioiodine has been used successfully, with

partial reduction in thyroid size and relief of pressure symptoms. The ratio of papillary/follicular thyroid carcinoma is lower in the elderly (2:1) than in younger patients (3 or 4:1). Differentiated thyroid carcinoma is more aggressive in older patients, especially males. Anaplastic thyroid carcinoma is almost exclusively observed in patients older than 65 years of age, similar to other rare thyroid neoplasms, such as sarcomas and primary thyroid lymphomas. The therapeutic approach for differentiated thyroid cancer is similar to that followed for younger persons.

See Also the Following Articles Aging and Longevity of Human Populations . Graves’ Disease, Hyperthyroidism in . Graves’ Ophthalmopathy . Hyperthyroidism, Subclinical . Hypothyroidism, Treatment of . Nontoxic Goiter . Thyroid Autoimmunity . Thyroid Carcinoma . Toxic Adenoma . Toxic Multinodular Goiter

Further Reading Chiovato, L., Mariotti, S., and Pinchera, A. (1997). Thyroid diseases in the elderly. Baillie`res Clin. Endocrinol. Metab. 11, 251–270. Mariotti, S., Franceschi, C., Cossarizza, A., and Pinchera, A. (1995). The aging thyroid. Endocr. Rev. 16, 686–715. Mariotti, S., Chiovato, L., Franceschi, C., and Pinchera, A. (1998). Thyroid autoimmunity and aging. Exp. Gerontol. 33, 535–541. Mokshagundam, S., and Barzel, U. S. (1993). Thyroid disease in the elderly. J. Am. Geriatric Soc. 41, 1361–1369. Parle, J. V., Maisonneuve, P., Sheppard, M. C., Boyle, P., and Franklyn, J. A. (2001). Prediction of all-cause and cardiovascular mortality in elderly people from one low serum thyrotropin result: A 10-year cohort study. Lancet 358, 861–865. Sawin, C. T., Geller, A., Wolf, P. A., Belanger, A. J., Baker, E., Bacharach, P., Wilson, P. W., Benjamin, E. J., and D’Agostino, R. B. (1994). Low serum thyrotropin concentrations as a risk factor for atrial fibrillation in older persons. N. Engl. J. Med. 331, 1249–1252.

Thyroidectomy

Figure 1 The skin incision in a thyroidectomy is made 1 cm below the cricoid and follows normal skin lines. A 00 silk suture is pressed against the neck to mark the incision site. Reprinted from Clark (1985), with permission.

crease for optimal cosmetic effect. Thus, the incision is directly over the isthmus of the thyroid gland. The gentle curve of the Kocher incision is then marked using a taut silk suture (Fig. 1). The length of the incision is dependent on the existing pathology and the size of the patient’s neck. It may vary from approximately 3 to 5 cm. The skin incision is extended through subcutaneous fat and the platysma muscle. The superior flaps are created by dissection in an avascular plane just deep to the platysma muscle and superficial to the anterior jugular veins up to the level of the thyroid cartilage. The lower flap is mobilized in a similar fashion to the level of the suprasternal notch. The wound edges are protected with moistened drapes, and one or two self-retaining retractors are placed.

427 Attention is turned to the superior pole. Retracting the thyroid in a caudal direction identifies the superior thyroid artery and veins. The tissue adjacent to the superior pole vessels can usually be swept from the thyroid with a peanut sponge. The space between the thyroid gland and cricothyroid muscle is opened, allowing the superior pole vessels to be skeletonized, triple clamped, ligated, and divided. The external branch of the superior laryngeal nerve is at risk of injury during this maneuver; there is a 10% reported injury rate. It has a variable course and is a thin nerve. The external branch of the superior laryngeal nerve is best preserved by not attempting to directly identify it but, rather, individually ligating the superior pole vessels close to the thyroid as opposed to mass ligation of the pedicle (Fig. 2). The tissue posterior to the pole can now be easily swept from the thyroid gland by blunt dissection. The upper parathyroid gland is usually identified at

MIDLINE DISSECTION

p0040

Excellent exposure is obtained by a midline incision through the median raphe of the superficial layer of the deep cervical fascia between the strap muscles. The midline is most easily identified low in the neck. The incision is extended superiorly to the thyroid cartilage and inferiorly to the suprasternal notch. Crossing veins in the lower neck may need to be ligated. The sternohyoid muscle is dissected from the sternothyroid muscle laterally to provide better exposure of the thyroid gland. This can usually be done with blunt dissection. The sternothyroid muscle is then dissected from the underlying thyroid. The middle thyroid vein(s) is identified, divided, and ligated. This dissection is facilitated by medial retraction of the thyroid gland and lateral retraction of the strap muscles and carotid sheath.

Figure 2 The superior pole vessels are identified and individually ligated using three right-angle clamps placed with tips away from the external branch of the superior laryngeal nerve. Reprinted from Clark (1985), with permission.

428 the level of the cricoid cartilage, where the RLN enters the larynx posterior to the cricothyroid muscle. It should be preserved on its vascular pedicle. The area cephalad to the cricoid cartilage is considered relatively safe because the recurrent laryngeal nerve enters the cricothyroid muscle below the cricoid cartilage. The pyramidal lobe, which is present in approximately 80% of patients, should be identified and mobilized to the level above the thyroid cartilage with the dissection plane immediately adjacent to this lobe. There are often numerous small vessels supplying the pyramidal lobe. When the pyramidal lobe is freed of all its lateral attachments, it is gently avulsed or divided from its superior attachment at the hyoid bone.

IDENTIFICATION OF THE RECURRENT LARYNGEAL NERVE Both right and left RLNs enter the larynx posterior to the cricothyroid muscle just above the cricoid cartilage. The right RLN takes a more oblique course in the neck and may pass anterior or posterior to the inferior thyroid artery (Fig. 3). In approximately 1% of patients, the right RLN is nonrecurrent and may

Thyroidectomy

enter the thyroid from a superior or lateral direction. The left RLN almost always runs in the tracheoesophageal groove because of its deeper origin from within the thorax. Both recurrent nerves may branch before entering the larynx; this occurs more frequently on the left. Preservation of the medial branch is of utmost importance because it usually contains the motor fibers. Retraction of the carotid sheath laterally and the thyroid medially and anteriorly places tension on the inferior thyroid artery, which helps to identify the RLN where it crosses the midportion of the thyroid gland. It is usually safe to identify the RLN low in the neck and then follow it to where it enters the cricothyroid muscle through the ligament of Berry. The tertiary branches of the inferior thyroid artery are individually ligated with fine ties and divided, mobilizing the thyroid lobe medially, away from the RLN. The most difficult part of the dissection during a thyroidectomy usually involves the ligament of Berry. The ligament is situated at the posterior lateral portion to the thyroid gland just caudal to the cricoid cartilage. A small branch of the inferior thyroid artery traverses the ligament, as do one or more veins from the thyroid gland. These vessels are usually readily identified and ligated. Should bleeding occur, it should be controlled by pressure with no clamping until the RLN is identified. A small amount of persistent bleeding at the end of the case can be controlled with the placement of a small pledget of thrombin-soaked gel foam. A tubercle of Zuckerkandl may extend over the RLN at the ligament of Berry. In addition to having a consistent relationship with the location of the RLN, the upper parathyroid gland may be situated at the tip of this protruding portion of thyroid tissue.

PRESERVATION OF THE PARATHYROID GLANDS

Figure 3 The recurrent laryngeal nerve may run anterior, posterior, or between the branches of the inferior thyroid artery. The inferior parathyroid gland is usually identified within 1 cm of this junction. Reprinted from Clark (1985), with permission.

Eighty-five percent of people have four parathyroid glands usually situated immediately adjacent to the thyroid gland on the posterior lateral capsule. The upper parathyroid glands are most commonly lateral to the recurrent laryngeal nerve at the level of the ligament of Berry in a posterior position and are usually the easiest to preserve during thyroidectomy. The lower parathyroid glands are almost always situated anterior to the RLN and within 1 cm caudal to where the RLN crosses the inferior thyroid artery. If not observed here, they are usually in the thymus or

p0070

429

Thyroidectomy

perithymic fat. Preservation of the blood supply to the parathyroid glands is best achieved by meticulous dissection from the thyroid capsule. Should it not be feasible to preserve the blood supply to a parathyroid gland, it should be biopsied to confirm it is parathyroid and autotransplanted into the contralateral sternocleidomastoid muscle. In general, parathyroid glands should be preserved rather than transplanted if possible. The parathyroid glands should not be mobilized extensively because they may be devascularized. The thyroid gland should also be examined after its removal to ensure that a parathyroid gland has not been removed.

REMOVAL OF SPECIMEN Always keeping in mind the location of the RLN, the inferior thyroid vessels are dissected free, clamped, and ligated. The thyroid gland is retracted medially to expose the trachea. The thyroid lobe and isthmus are then easily dissected off the trachea with electrocautery. For a lobectomy, the isthmus is divided between two Colodny clamps on the side contralateral to the thyroid pathology, and the isthmus is suture ligated. The operative field is irrigated with warm saline and any bleeding is controlled. The strap muscles and then platysma muscle are reapproximated with interrupted 4–0 absorbable suture, and the skin is closed with butterfly clips. The wound is dressed with sterile gauze bandages.

FUTURE DIRECTIONS Except for improved lighting and magnification, the fundamental approach to thyroid surgery, with

preservation of the RLN and parathyroid glands, has remained relatively unchanged during the past century. Minimally invasive surgery has evolved to include thyroid surgery. Endoscopic parathyroidectomy using CO2 insufflation has been demonstrated to be technically feasible, and similar techniques have been applied to thyroidectomy. The development of highly sensitive nerve probes is under investigation. Its utility in avoiding injury to the external branch of the superior laryngeal nerve and preservation of the RLN in redo surgery has had promising preliminary results. Regardless of these new developments, the importance of adequate training, expertise, and knowledge of anatomy and techniques are fundamental to ensuring good outcomes.

See Also the Following Articles Medullary Thyroid Carcinoma . Nontoxic Goiter . Parathyroid Cancer . Parathyroid Glands, Pathology . Thyrotoxicosis, Treatment

Further Reading Clark, O. H. (1985). ‘‘Endocrine Surgery of the Thyroid and Parathyroid Glands.’’ Mosby, St. Louis. Clark, O. H., and Noguchi, S. (2000). ‘‘Thyroid Cancer—Diagnosis and Treatment.’’ Quality Medical, St. Louis. Doherty, G. M., and Skogseid, B. (2001). ‘‘Surgical Endocrinology.’’ Lippincott Williams & Wilkins, Philadelphia. Duh, Q.-Y., and Clark, O. H. (1997). ‘‘Textbook of Endocrine Surgery.’’ Saunders, Philadelphia. Jossart, G. H., and Clark, O. H. (1996). Perioperative management and surgical technique. Sci. Am. 1–8. Wartofsky, L. (2000). ‘‘Thyroid Cancer. A Comprehensive Guide to Clinical Management.’’ Humana Press, Totowa, NJ.

506

Thyroiditis, Infectious

Table I Differential Diagnosis of the Painful Neck Mass Nonthyroidal Infected thyroglossal duct cyst Infected branchial cleft cyst Infected cystic hygroma Cervical adenitis Cellulitis of the anterior neck Thyroidal Subacute thyroiditis Infectious thyroiditis Acute hemorrhage into a cyst Acute hemorrhage into a benign or malignant nodule Rapidly enlarging thyroid carcinoma Painful Hashimoto’s thyroiditis Radiation thyroiditis Globus hystericus

Clinical Presentation and Diagnosis Bacterial thyroiditis is often preceded by an upper respiratory infection, which may induce inflammation of the fistula and promote the transmission of pathogens to the thyroid, and it is more common in the late fall and late spring months. More than 90% of patients present with thyroidal pain, tenderness, fever, and local compression resulting in dysphagia and dysphonia, and signs and symptoms of systemic toxicity may be present. The thyroid is tender to palpation, with unilateral or bilateral lobar enlargement, and it is associated with erythema and warmth of the skin. Cervical lymphadenopathy is not a prominent feature unless there is a predisposing pharyngitis. The differential diagnosis of bacterial thyroiditis can be divided into nonthyroidal and thyroidal causes (Table I). Essentially all of the nonthyroidal causes are

infectious in origin and present as discrete painful masses. Subacute thyroiditis is the most common cause of the painful thyroid and often results in both local and systemic symptoms similar to those seen in bacterial thyroiditis. Thyroid function tests in the normal range are the most common finding in patients with bacterial thyroiditis (Table II), although thyrotoxicosis and hypothyroidism have been reported. Fine needle aspiration biopsy is the best laboratory test in the evaluation of infectious thyroiditis and is diagnostic in most cases, especially when tenderness is limited to a solitary nodule or a localized area and subacute thyroiditis has been ruled out. Gram stain and culture of the fine needle aspirate will reveal the causative organism in more than 90% of cases. Most imaging studies are adjunctive and are best reserved for patients for whom the diagnosis is unclear. In the adult, Staphylococcus aureus and Streptococcus pyogenes are the offending pathogens in more than 80% of patients and are the sole pathogen in more than 70% of cases. In children, a- and b-hemolytic Streptococcus and a variety of anerobes account for 70% of cases, whereas mixed pathogens are identified in >50% of cases.

Management and Prognosis Treatment of acute bacterial thyroiditis requires admission to the hospital, drainage of any abscess, and parenteral antimicrobial therapy aimed at the causative agent. If no organisms are seen on the gram stain, nafcillin and gentamycin or a third-generation cephlosporin is appropriate initial therapy in adults, whereas a secondgeneration cephlosporin or clindamicin is reasonable in children. Since a pyriform sinus fistula is the most

Table II Diagnostic Tests in the Evaluation of Infectious Thyroiditis Test Fine needle aspiration biopsy

Comments Diagnostic in 90% of cases; test of choice; special stains considered necessary

Thyroid function tests

Usually normal; rare case reports of hypothyroidism and thyrotoxicosis

Erythrocyte sedimentation rate

Usually elevated; nonspecific test

Leukocyte count

Usually elevated; nonspecific test

Radionucliide imaging Neck radiograph

Adjunctive test; radioiodine best; provides information regarding overall gland function Adjunctive test; presence of gas indicates abscess with anerobic organisms

Magnetic resonance imaging

Adjunctive test; helpful in identifying pyriform sinus fistulae

Ultrasonography

Adjunctive test; helpful in identifying pyriform sinus fistulae

Computed tomography

Adjunctive test; helpful in identifying pyriform sinus fistulae

Barium swallow

Adjunctive test; helpful in identifying pyriform sinus fistulae

507

Thyroiditis, Infectious

common route of infection in bacterial thyroiditis, a barium swallow, computed tomography, or magnetic resonance imaging of the neck should be performed to search for communicating fistulae in most patients, especially children, with the first episode and all patients with recurrent episodes. Such fistulae must be surgically excised for definitive cure and prevention of recurrent infection. Mortality from acute bacterial thyroiditis has markedly improved from the 20–25% reported in the early 1900s. A 1983 review by Berger et al. estimated an overall mortality of 8.6%. Reviews involving more than 100 patients failed to list mortality as a complication of acute bacterial thyroiditis. However, the mortality is close to 100% if the diagnosis is delayed and antimicrobial therapy is not initiated. In survivors, complete recovery is the norm, although there have been reports of transient hypothyroidism, vocal cord paralysis (which may also be transient), and recurrence of infection as sequelae of acute bacterial thyroiditis.

FUNGAL INFECTIONS Although rare, fungal infections of the thyroid are the next most common cause of infectious thyroiditis, comprising 15% of cases reported through 1980. The predominant offending organism is Aspergillus species, with at least 26 documented cases. Virtually all of the affected patients were immunocompromised and had disseminated disease; most cases of thyroidal infection were determined postmortem. Asymptomatic infection of the thyroid with Pneumocystis carinii is found in up to 20% of AIDS patients with disseminated Pneumocystis infection at autopsy. The diagnosis of P. carinii infection is made by performing Gomori’s silver methenamine stain on specimens obtained by fine needle aspiration biopsy. Case reports of fungal infections of the thyroid have included Coccidioides immitis, Histoplasma capsulatum, Candida albicans, Allescheria boydii, and Nocardia asteroids.

nerve paralysis has been described, resolution without sequelae usually follows appropriate antituberculous therapy. Infections with atypical mycobacteria, including M. cheloni, M. intracellulare, and M. avianintracellulare, have also been described—the latter in patients with AIDS in the setting of widely disseminated disease. Although acid fast organisms have been found in the thyroid of individuals with disseminated M. leprae, symptomatic thyroid infection has not been described.

PARASITIC INFECTIONS Several parasitic agents have involved the thyroid on rare occasions. Echinococcus granulosus has been reported in the setting of a chronic goiter, with the diagnosis being made at the time of surgery. If echinococcal infection is suspected, biopsy of the lesion is contraindicated due to spillage and rupture of the cyst contents, and specific serologic testing should be performed. Surgical removal is the preferred mode of treatment, with antiparasitic agents useful as adjunctive therapy and for inoperable cases. Involvement of the thyroid with Strongyloides stercoralis has been described only in the setting of disseminated disease in immunocompromised patients. Mortality with disseminated Strongyloides infection is high due to both the infection and the immunocompromised status of the patient.

SYPHILITIC INFECTION Historically, secondary syphilis was frequently associated with pain and swelling of the thyroid that responded to antisyphilitic therapy, which commonly included iodides. However, microbiologic evidence of syphilitic infection of the thyroid is lacking in most of these cases; thus, a direct relationship between treponemal infection and thyroid dysfunction cannot be determined. Indeed, only seven cases of gummata of the thyroid, presenting as painless nodules, have been reported in the world’s literature.

MYCOBACTERIAL INFECTIONS The true incidence of infection of the thyroid with Mycobacterium tuberculosis is difficult to determine. Using strict pathological criteria, only 19 cases have been reported in the literature. Thyroidal tuberculosis is associated with disseminated or miliary disease and symptoms are usually present for months. Although at least three of the reported patients with tuberculous thyroiditis died and recurrent laryngeal

VIRAL INFECTIONS The most common infectious organism found in the thyroid at postmortem examination in patients with AIDS is cytomegalovirus (CMV), occurring in the setting of disseminated CMV infection. However, symptomatic thyroidal infection with CMV has not been reported. Thyroiditis has been associated with mumps parotitis, although this is rare.

508

CONCLUSION Infectious thyroiditis is uncommon and the diagnosis often requires a high index of suspicion. A rational approach to such patients, including history, physical examination, laboratory evaluation, and fine needle aspiration biopsy, will allow the appropriate diagnosis to be made in the vast majority of cases.

See Also the Following Articles Pediatric HIV Infection and Hypothalamic–Pituitary– Adrenal Axis . Thyroid Carcinoma . Thyroid Fine Needle Aspiration Cytology . Thyroiditis, Postpartum . Thyroiditis, Subacute

Further Reading Battan, R., Mariuz, P., Raviglione, M. C., Sabatini, M. T., Mullen, M. P., and Poretsky, L. (1991). Pneumocystis carinii infection of the thyroid in a hypothyroid patient with AIDS: Diagnosis by fine needle aspiration biopsy. J. Clin. Endocrinol. Metab. 72, 724–726.

Thyroiditis, Infectious

Berger, S. A., Zonszein, J., Villamena, P., and Mittman, N. (1983). Infectious diseases of the thyroid gland. Rev. Infect. Dis. 5(1), 108–122. Farwell, A. P. (2000). Infectious thyroiditis. In ‘‘The Thyroid’’ (L. E. Braverman and R. D. Utiger, eds.), 8th ed., pp. 1044–1050. Lippincott William & Wilkins, Philadelphia. Miyauchi, A., Matsuzuka, F., Kuma, K., and Takai, S. (1990). Piriform sinus fistula: An underlying abnormality common in patients with acute suppurative thyroiditis. World J. Surg. 14(3), 400–405. Nonomura, N., Ikarashi, F., Fujisaki, T., and Nakano, Y. (1993). Surgical approach to pyriform sinus fistula. Am. J. Otolaryngol. 14(2), 111–115. Park, S. W., Han, M. H., Sung, M. H., Kim, I. O., Kim, K. H., Chang, K. H., and Han, M. C. (2000). Neck infection associated with pyriform sinus fistula: Imaging findings. Am. J. Neuroradiol. 21(5), 817–822. Rich, E. J., and Mendelman, P. M. (1987). Acute suppurative thyroiditis in pediatric patients. Pediatr. Infect. Dis. J. 6(10), 936–940. Shah, S. S., and Baum, S. G. (2000). Diagnosis and management of infectious thyroiditis. Curr. Infect. Dis. Rep. 2, 147–153. Singh, S. K., Agrawal, J. K., Kumar, M., and Shukla, H. S. (1994). Fine needle aspiration cytology in the management of acute suppurative thyroiditis. Ear Nose Throat J. 73(6), 415–417. Takai, S.-I., Miyauchi, A., Matsuzuka, F., Kuma, K., and Kosaki, G. (1979). Internal fistula as a route of infection in acute suppurative thyroiditis. Lancet 1, 751–752.

510 pregnancy, with postpartum increases in helper Tcells and cytotoxic T cells. The hormonal changes that occur during pregnancy have a profound effect on the production of immune tolerance. Increased corticotropin-releasing hormone production by the placenta stimulates the maternal adrenal glands to produce a state of hypercortisolism. There is activation of the sympathetic system with increased production of catecholamines. A moderate increase in 25-hydroxy vitamin D3 and more significant increases in 1,25-dihydroxy vitamin D3 also occur. Through intermediate mechanisms, these changes suppress proinflammatory cytokine formation and promote a shift from Th1 to Th2. The increase in plasma estrogen and progesterone levels may enhance this effect. These changes help to conserve the fetus and reduce chances of rejection. Maternal immune mechanisms return to a normal nonpregnant state in the first few months to 1 year after delivery. However, there may be a rebound increase in some elements of the autoimmune reaction (e.g., antibody production or CD4:CD8 ratio), which may aggravate existing autoimmune disease or precipitate it for the first time in predisposed women.

Women with PPT The ‘‘immune rebound’’ hypothesis may explain some of the autoimmune mechanisms that form the basis of PPT. The behavior of antibodies to thyroglobulin (TgAb) and thyroid peroxidase (TPOAb) during pregnancy and the postpartum period has been well documented. This heightened immune response in the first few months of the postpartum period may be caused by several factors. Microchimerism There is a considerable influx of fetal cells into the maternal circulation at the time of delivery. These cells may persist for short or long periods of time in the maternal host. Evidence suggests that these ‘‘chimeric’’ cells may induce an immune reaction in the host by causing a breakdown in immune tolerance. One shortterm effect may be the rebound immune enhancement seen in the postpartum period, which may cause an exacerbation or new onset of some autoimmune diseases.

p0050

T-Cell Changes Evidence suggests that activation of both circulating and intrathyroidal Tcells occurs in PPT. Both Walfish and Stagnaro-Green demonstrated the expression of MHC class II molecules and a higher percentage of increased CD4:CD8 ratios in subjects who developed

Thyroiditis, Postpartum

PPT. In a prospective study of TPOAb-positive women, Kuijpens showed a higher percentage of MHC II-expressing T cells in subjects who subsequently developed PPT compared to those who did not. However, Jansson’s group did not demonstrate a difference in circulating subsets of T lymphocytes in thyrotoxic and hypothyroid PPT subjects compared with normal controls. This group demonstrated a relative increase in intrathyroidal B cells and a relative decrease in CD8 cells (resulting in an increased CD4:CD8 ratio) in subjects with PPT.

PPT as an Immune-Mediated Disease Several features of PPT point to the central importance of immune mechanisms in the pathogenesis of PPT (Table I). The majority of women who develop PPT are positive for serological markers of thyroid autoimmunity (i.e., TPOAb and TgAb). In our experience, all such women have TPOAb during early pregnancy. Evidence suggests that antibodies with a dual specificity for Tg and TPO may also be found in PPT at a higher prevalence than in normal control subjects. However, several investigators have reported PPT in TPOAb-negative women. In such women, the etiology of PPT is unclear. The histological changes occurring within the thyroid gland, with immune cell infiltration typical of autoimmune thyroid disease, give further credence to the immune pathogenesis of the disease. Biopsy of the thyroid gland in women with PPT shows lymphocytic infiltration and follicle formation reminiscent of those of Hashimoto’s thyroiditis. Several studies have shown HLA haplotype restrictions in PPT, which are commonly seen in autoimmune thyroiditis such as Hashimoto’s and Graves’ disease. The subclasses of TPOAb that are able to activate the complement cascade (IgG1–IgG3) increase during Table I

Immunohistological Features of PPT

Presence of thyroid antibodies

TPOAb (majority) Thyroglobulin antibodies and antibodies with dual specificity (minority)

Postpartum increase in TPOAb subclasses capable of activating complement

IgG1–IgG3

Histological changes

Lymphocyte infiltration and follicle formation within the thyroid gland

HLA haplotype restriction

Similar to autoimmune thyroiditis

511

Thyroiditis, Postpartum

Bioactive TPOAb (kIU/I) 140 120 100 Euthyroid 80

PPT

60

Persistent hypothyroidism

40 20

methodological discrepancies in studies reported from different locations. Variability of factors such as diagnostic criteria, length of follow-up after delivery, frequency of postpartum blood sampling, and differences in hormone assay methodology may have contributed to this variation. An average prevalence of 5–7% is acceptable for unselected pregnant women from most iodine-replete populations.

Predisposition to PPT

0 0 1 2 3 4 5 6 7 8 9 10 11 12

Months

Figure 1 Biologically active TPOAb levels in antibody-positive women who were euthyroid or had transient or persistent postpartum hypothyroidism. Antibody concentrations returned to normal in those with PPT but remained elevated in subjects with persistent hypothyroidism.

the postpartum period (Fig. 1) and are associated with both phases of PPT. Jansson reported an increase in IgG1 in hypothyroid PPT. Hall demonstrated an increase in IgG2 and IgG3 in biphasic PPT, with the increase in IgG3 coinciding with the thyrotoxic phase. Briones-Urbina confirmed the IgG1 and IgG2 changes but found low IgG3 levels. However, these investigators also confirmed that IgG4, which is incapable of influencing the complement cascade, remains relatively unchanged. This raises the interesting possibility of a pathogenetic role for these antibodies in PPT, perhaps through complement activation. A sublethal antibody-directed, complement-mediated attack on thyroid cells may result in increased secretion of thyroid hormones, producing the thyrotoxic phase of the disease. However, a more severe complementmediated attack may produce damage to the follicular architecture of the gland and produce hypothyroidism. Conclusive evidence for such complement activation in PPT is lacking; studies have been unable to demonstrate terminal complement complexes (markers of complement activation) in TPOAb-positive women who developed PPT and who were followed weekly during the postpartum period.

CLINICAL FEATURES AND MANAGEMENT OF PPT Prevalence There is a wide variation in the reported worldwide prevalence of PPT. This variation may be explained by true geographic differences in prevalence (reflecting genetic heterogeneity and other factors) or by

Women with TPOAb (and TgAb alone in less than 5%), type 1 diabetes mellitus, and previous PPTare at increased risk of developing PPT. We previously mentioned this increased risk in TPOAb-positive women. Studies in Cardiff show that approximately 50% of women with TPOAb during the early stages of pregnancy develop PPT. Other studies have shown this proportion to vary between 30 and 52%. Therefore, TPOAb is a marker of risk for the development of PPT but remains a weak predictor. There is a higher prevalence of PPT in subjects with type 1 diabetes mellitus. Gerstein followed up 40 of 51 pregnant subjects with type 1 diabetes, of whom 10 developed thyroid dysfunction (1 due to Graves’ disease), and Alvarez-Marfany followed up 28 of 41 similar women, of whom 7 developed thyroiditis. Therefore, the incidence is approximately 25% in this group of women. However, the highest incidence of PPT is found in women who have had a previous episode of the disease. Sixty-nine percent of women who have TPOAb and had PPT following a previous pregnancy will develop a similar disease during the next pregnancy. However, only 25% who are TPOAb positive and remain euthyroid will develop thyroid dysfunction during the next pregnancy. There is evidence of a role for environmental factors, such as the presence of a goiter and smoking and a family history of thyroid disease, in the causation of PPT. However, additional studies are needed to confirm this finding.

Types of PPT and Clinical Features Classically, PPT occurs after a full-term delivery. However, there are reports of postpartum thyroiditis occurring after early loss of pregnancy between 5 and 20 weeks of gestation. PPT is typically a biphasic disease (Fig. 2). A transient thyrotoxic phase is followed by a period of recovery and then a hypothyroid phase. The thyrotoxic phase occurs at a median of 13 weeks postpartum and lasts 1 or 2 months; it is probably due to the release of

p0070

512

Thyroiditis, Postpartum

20−25% 25−35%

Biphasic Hypothyroid

excess of symptoms of minor and major depression. However, a study from Spain was unable to confirm this finding. The mechanism of depression in hypothyroid PPT is speculative but may be related to the reduced 5-hydroxytryptamine drive seen in this condition or to known cytokine release associated with this phase, affecting neurotransmission.

Hyperthyroid

Management of PPT 35−40%

Figure 2 Clinical types of PPT. The thyrotoxic phase usually precedes the hypothyroid phase.

preformed thyroid hormone resulting from destruction of thyroid follicles. The hypothyroid phase that follows occurs at a median of 19 weeks postpartum, lasts longer (approximately 4–6 months), is accompanied by significant symptoms, and results from autoimmune follicular destruction and reduced hormone synthesis. Some women may require thyroxine replacement therapy during this phase. Rarely, the hypothyroid phase may precede the thyrotoxic phase. In some women, the two phases of PPT may occur independently of each other and either clinical or biochemical evidence of thyrotoxicosis or hypothyroidism alone develops during approximately the same postpartum periods as described previously. Significantly, as many as 30% of women who have TPOAb and PPT may develop permanent hypothyroidism requiring thyroxine replacement therapy by the end of the first postpartum year. The symptoms of the thyrotoxic phase are mild and self-limiting. Fatigue, palpitations, weight loss, irritability, and heat intolerance are more commonly found in subjects with thyrotoxic PPT than in euthyroid postpartum women. They may also have tremor, nervousness, and psychological symptoms. The mild and nonspecific nature of these symptoms may cause diagnostic confusion, and they may be missed if a high index of suspicion is not maintained. The hypothyroid phase lasts longer and may cause considerable morbidity. Fatigue, loss of concentration, constipation, muscle and joint pains, and stiffness are common complaints. Some of these symptoms may occur before the abnormalities in biochemical thyroid function become evident and also persist after euthyroidism is achieved. Studies from The Netherlands and United Kingdom have demonstrated a significant

The temporal relationship to pregnancy and delivery, the presence of thyroid antibodies in the majority of women who develop symptoms, and the pattern and timing of biochemical thyroid dysfunction should alert clinicians to PPT in women at risk. Symptoms and the presence of a goiter are unhelpful in differentiating PPT from other causes of thyroid dysfunction in the postpartum period. The presence of eye signs in the thyrotoxic phase, however, favors a diagnosis of Graves’ disease. The thyrotoxic phase is distinguished from an exacerbation of quiescent or a new onset of Graves’ disease relatively easily by radioiodine uptake scanning. Uptake is consistently low in PPT but high in Graves’ disease. When postpartum thyrotoxicosis occurs in women with previously known Graves’ disease, a low uptake confirms PPT (on the background of quiescent Graves’ disease). However, care needs to be taken in the use of radioiodine as a diagnostic tool in nursing mothers. Technetium scans may be preferable. The presence of thyroid-stimulating hormone (TSH) receptor antibodies favors the diagnosis of Graves’ disease. Thyroid ultrasonography, thyroglobulin estimation, and IL-6 measurement are of little practical value. Specific antithyroid drug therapy is not required in the thyrotoxic phase because symptoms and biochemical thyroid function return to normal in a few weeks. Occasionally, a beta-blocker may be indicated for symptom relief. The hypothyroid phase usually follows the thyrotoxic phase and should be anticipated with periodic follow-up. An elevated TSH level at the appropriate time postpartum in women who are most likely TPOAb positive should alert clinicians to the diagnosis. This phase lasts longer and is associated with considerable morbidity. It may result in early permanent hypothyroidism, as described previously. Thyroxine therapy is indicated with a trial of withdrawal at 9–12 months. It is possible to withdraw thyroxine therapy in the majority of women at the end of this period, but a recurrence of symptoms associated with increased TSH levels on follow-up indicates the need for permanent replacement therapy.

513

Thyroiditis, Postpartum

LONG-TERM OUTCOME FOLLOWING PPT The long-term outcome of PPT has been examined in several studies (Table II). In our series, permanent hypothyroidism occurred as early as 9 months postpartum in approximately 30% of subjects who were TPOAb positive and had PPT. These women required thyroxine replacement therapy to maintain normal clinical and biochemical thyroid function. A review of long-term follow-up studies of PPT from geographically different locations reported a 12–61% prevalence of permanent hypothyroidism. The variability may in part be due to differences in the definition of PPT and long-term thyroid dysfunction, length of follow-up, and ascertainment. We followed 98 TPOAb-positive women (of whom 48 developed PPT) and 70 TPOAb-negative controls for 66–140 months. Forty-six percent of women who developed PPT were hypothyroid (some subclinically) at the end of the follow-up period compared to only 4% of women who were TPOAb positive but did not develop PPT and 1.4% of women who were TPOAb negative. The rate of conversion to hypothyroidism in women who were TPOAb positive and developed PPT was 7.1% per year, higher than that reported for women in community-based follow-up studies. Investigators from different areas of the world have confirmed a high prevalence of hypothyroidism at the end of variable follow-up periods in women who developed PPT. In a study of Japanese women with a similar length of follow-up after PPT (mean, 8.7 years), Tachi found a 29% prevalence of permanent hypothyroidism. In a Swedish study, Jansson found a 30% prevalence of hypothyroidism at 5 years. In Brazil, Barca found a 61% prevalence of hypothyroidism at the end of a 2-year

period of follow-up after PPT. The reason for this high prevalence of relatively early hypothyroidsm is unclear. These studies raise the interesting issue of the nature of thyroid damage following the initial episode of PPT. Several investigators have indicated the distinct possibility of a persistent but subtle abnormality (probably autoimmune in nature) of thyroid function and morphology in these women. Iodine perchlorate discharge tests were abnormal in 41% of Italian and 64% of Welsh women studied 3 and 7 years, respectively, following PPT, suggestive of a persistent organification defect. Furthermore, we found a significantly higher prevalence of thyroid ultrasound hypoechogenicity (due to autoimmune destruction) at 4–8 weeks postpartum in women who were TPOAb positive and developed PPT (45%) compared to antibodypositive women who did not develop PPT (17%) and to antibody-negative women (1.5%). There was a significantly higher prevalence of persistent abnormalities in the first group after 66–144 months of follow-up (although mildly reduced from the postpartum period), indicative of persistent autoimmune destruction following the initial episode of PPT. It seems likely that a persistent but low-grade immune destructive process severe enough to produce echogenic changes in the thyroid gland continues to occur in these women, although maximal damage occurs at the time of PPT. These findings indicate the need for long-term follow-up of women who are TPOAb positive and who develop PPT. A relatively high incidence of early permanent hypothyroidism in these women and a higher than normal annual conversion rate to hypothyroidism (compared to that of TPOAb-positive

Table II Long-Term Follow-Up of PPT Author

Follow-up (years)

No. of subjects

Thyroid dysfunction (includes subclinical hypothyroidism/abnormal TRH test) 6

Nikolai

3

25

Vargas

1

42

8

Tachi

5–16

44

10

Lervang

2

23

2

Jansson

5

50

13

Othman

2–4

43

10

Solomon

0.9–3.7

55

20

Kuijpens Premawardhana

2.5–3 5–11.5

14 98

6 24

Lucas

3.3

42

5

Barca

2

49

30

514 women who do not develop PPT and women from community surveys), indicate the need for long-term follow-up. An increased TSH level with or without symptoms of hypothyroidism on annual (or more frequent) thyroid function testing is an indication for thyroxine replacement therapy.

SCREENING FOR PPT p0130

p0135

There is no consensus about screening for PPT. This relates as much to the absence of a highly sensitive and specific marker for prediction as to the lack of appreciation of the clinical problem and the long-term effects of PPT. The significant morbidity of PPT in the first postpartum year, the likelihood that approximately three-fourths of women who have PPT will have an episode in a future pregnancy, the high prevalence of long-term thyroid dysfunction following PPT, and the availability of effective treatment should make a screening strategy useful. Some authorities recommend a selective screening strategy aimed at only those women who have a high risk of developing PPT (i.e., women with type 1 diabetes mellitus and those who have had PPT in a previous pregnancy). However, a screening strategy should take into account the fact that PPT may occur in TPOAbnegative women. It is salutary to remember that the prevalence of diseases for which antenatal screening is currently recommended is considerably lower than that of PPT in women of childbearing age. TPOAb has been proposed as a marker for the prediction of PPT. As mentioned previously, when measured in early pregnancy, TPOAb is present in approximately 10% of women. However, only approximately half of these develop PPT, raising questions about the sensitivity of TPOAb as a predictor. Ten studies have examined TPOAb as a predictor of PPT. No firm conclusions can be drawn from these studies for several reasons. Most investigators measured microsomal antibodies, whereas some measured antibodies to TPO (the specific microsomal antigen). The assay methods used and the timing of antibody measurement were variable. Whereas some investigators measured antibodies in the antepartum period, others measured them at delivery and in the postpartum period. These studies, however, showed that thyroid antibodies have a sensitivity of 0.45–0.89 and specificity of 0.91–0.98. Positive predictive value and relative risk were 0.31–0.78 and 20–50.7, respectively. A report confirmed the cost-effectiveness of a screening program. It is not known whether screening

Thyroiditis, Postpartum

will be improved by thyroglobulin estimation, measuring ultrasound thyroid volume, or assessment of complement activation.

CONCLUSION PPT is a common endocrine disorder affecting young women. The exact mechanisms of cellular damage have not been determined, although the autoimmune destructive nature of the disease has long been recognized. The immune perturbations of pregnancy and the postpartum period account for the modulation of thyroid autoimmunity, which is the hallmark of the disease, and the timing and nature of clinical and biochemical changes. Although the majority of patients have a short and self-limited illness, some have a prolonged and symptomatic disorder that requires specific therapy. The recognition of short- and longterm morbidity, and the need for permanent thyroxin supplementation in a significant minority of subjects following PPT, has raised the important but unsettled issue of screening for PPT, targeted perhaps at those at highest risk. We await the discovery of a sensitive and specific screening tool.

See Also the Following Articles Depression, Thyroid Function and . Thyroid Autoimmunity . Thyroiditis, Infectious . Thyroiditis, Subacute . Thyrotoxicosis, Overview of Causes

Further Reading Gerstein, H. C. (1990). How common is postpartum thyroiditis? A methodologic overview of the literature. Arch. Intern. Med. 150, 1397–1400. Lazarus, J. H., Premawardhana, L. D. K. E., and Parkes, A. B. (1998). Postpartum thyroiditis; 83–97. In ‘‘Endocrine Autoimmunity and Associated Conditions’’ (A. P. Weetman, ed.). Kluwer, London. Muller, A. F., Drexhage, H. A., and Berghout, A. (2001). Postpartum thyroiditis and autoimmune thyroiditis in women of childbearing age: Recent insights and consequences for antenatal and postnatal care. Endocr. Rev. 22(5), 605–630. Premawardhana, L. D., Parkes, A. B., Ammari, F., John, R., Darke, C., Adams, H., and Lazarus, J. H. (2000). Postpartum thyroiditis and long term thyroid status: Prognostic influence of thyroid peroxidase antibodies and ultrasound echogenicity. J. Clin. Endocrinol. Metab. 85, 71–75. Smallridge, R. C. (1999). Postpartum thyroid disease through the ages. Thyroid 9, 671–673. Smallridge, R. C. (2000). Postpartum thyroid disease: A model of immunologic dysfunction. Clin. Appl. Immunol. Rev. 1, 89–103.

516 Table I Differential Diagnosis of Subacute Thyroiditis Subacute thyroiditis Acute hemorrhage into cyst or nodule Infected thyrogossal duct cyst Infected branchial cleft cyst Acute pyogenic thyroiditis Cellulitis of the anterior neck Rapidly enlarging thyroid cancer Painful Hashimoto’s thyroiditis Cervical adenitis Radiation thyroiditis Infected cystic hygroma Rare infections (e.g., Pneumocystis carinii ) or other inflammatory disorders (e.g., amyloidosis) Globus hystericus

all of whom were positive for HLA-Bw35. Two of the patients lived near each other but developed the disorder 1 year apart, and the third sibling lived several hundred miles away and had not been in contact with either of the other two when they had subacute thyroiditis. Thus, HLA-Bw35 appears to render individuals genetically susceptible to the development of the disorder. An interesting cluster of subacute thyroiditis was reported in 12 patients in a single town in The Netherlands. Affected individuals bore clinical similarities to subacute thyroiditis, although only 1 patient was positive for HLA-Bw35, much less than would be expected. However, 5 of 11 patients tested positive for HLA-B15/62, a markedly greater frequency than expected. This suggests both genetic and clinical heterogeneity in subacute thyroiditis. Although thyroid autoimmunity is not believed to play a role in subacute thyroiditis, autoimmune abnormalities associated with the disorder have been described. Antibodies directed against the thyroidstimulating hormone (TSH) receptor have been reported. Also, a report of sensitization of T lymphocytes against thyroid antigen suggests the possibility of an autoimmune component, although it is believed that this is likely the result of released antigen during the active inflammation phase. Low titers of thyroid autoantibodies are sometimes detectable in patients with subacute thyroiditis.

CLINICAL FEATURES Subacute thyroiditis is more common in women (80% of cases) between the ages of 40 and 50 years. As mentioned previously, a viral prodrome is common,

Thyroiditis, Subacute

as are symptoms of sore throat, weakness, low-grade fever, myalgias, and, frequently, dysphagia. Symptoms may develop gradually over a few weeks, although patients frequently describe the disorder as coming on within a few days. Anterior neck pain is usual and is generally more significant on one side. It frequently radiates to the ear, the mandible, the occiput, or even the upper chest. Commonly, the patient initially consults an ear, nose, and throat physician or a dentist because of pain in the area of the throat or mandible. Frequently, the pain will migrate from one side to the opposite thyroid lobe after a few weeks. Symptoms of hypermetabolism are frequent since thyrotoxicosis occurs in approximately 50% of affected individuals; symptoms may include diaphoresis, palpitations, tachycardia, and weight loss. Rarely, the clinical presentation of subacute thyroiditis may be so dramatic in onset and so pronounced in severity that obstructive symptoms may develop (Fig. 1). Physical examination may disclose signs of hypermetabolism, with tachycardia, diaphoresis, and tremor. Palpation of the neck generally reveals a very firm to hard, exquisitely tender, ill-defined mass in one lobe of the thyroid gland, although there is frequently thickening and tenderness of the contralateral lobe as well. Tenderness may be so pronounced that patients may try to prevent palpation by the examiner. The overlying skin is occasionally erythematous.

DIFFERENTIAL DIAGNOSIS The differential diagnosis of subacute thyroiditis is shown in Table I and includes both thyroid and nonthyroid disorders. As mentioned previously, it is the most common thyroid etiology of anterior neck pain, with hemorrhage into a cyst or other benign nodule the second most common thyroidal cause of neck pain. Clinically, the two can readily be distinguished from each other since hemorrhage into a nodule is not accompanied by either viral symptoms or thyrotoxicosis. Moreover, hemorrhage is generally sudden in onset, and the mass tends to be very smooth on palpation. In addition, the radioactive iodine uptake in patients with hemorrhage is normal, and a radionuclide scan would reveal a filling defect. Pyogenic thyroiditis is characterized by features of an abscess, with a tender, palpable, fluctuant mass. In addition, the patient would likely have a high fever and a significantly higher white blood cell count than would the patient with subacute thyroiditis, and a radionuclide scan would reveal similar radionuclide findings as those for the patient with acute hemorrhage into a cyst. Patients with pyogenic thyroidtis are not thyrotoxic, nor are those with

517

Thyroiditis, Subacute

Figure 1 A 52-year-old man with severe subacute granulomatous thyroiditis with threatened airway obstruction. Note the pronounced retropharyngeal edema (A) and its dramatic resolution following intravenous administration of methylprednisolone (B). From Singer, P. A. (1991). Thyroiditis: Acute, subacute, and chronic. Med. Clin. North Am. 75, 61–77.

cellulitis of the anterior neck. Also, patients with cellulitis have features of infection and do not have discreet masses on palpation. The patient with an infected thyroglossal duct cyst or branchial cleft cyst may present with a painful mass, which may be fluctuant. Such lesions also may be distinguishable from subacute thyroiditis by virtue of their more superior (thyroglossal duct cyst) or lateral (branchial cleft cyst) locations. In addition, patients with such lesions are euthyroid. Rapidly growing thyroid cancer is an unusual cause of anterior neck pain, but when it occurs patients complain more of an achy-type pain, often accompanied by symptoms of tracheo-esophageal pressure. Patients do not have features of acute inflammation, and on palpation the mass is nontender and usually rock hard. Painful Hashimoto’s thyroiditis is also very uncommon and is accompanied by elevated titers of thyroid antibodies and, frequently, hypothyroidism. Radioactive iodine administration, either for treatment of thyrotoxic Graves’ disease or for remnant ablation in patients with differentiated thyroid cancer, may cause anterior neck pain and thyroid tenderness. The etiology of the inflammation is always obvious. Several years ago, infection of the thyroid with Pneumocystis carinii was confused with subacute thyroidtis. We reported on three patients with biopsy-proven P. carinii infection who presented with anterior neck

pain and suppressed radioiodine uptake values; two of the patients were thyrotoxic. Other workers also reported similar cases at approximately the same time. What made most of the cases unique was the history of using inhaled aerosolized pentamidine for pneumocystis pneumonia prophylaxis, which may well have resulted in the pneumocystis organism seeking another primary target. Since pentamidine is no longer used for prophylaxis, no further cases of pneumocystis thyroiditis have been reported. Thyroid amyloidosis associated with systemic amyloid has been reported, with tender goiters, low thyroid radioactive iodine uptake values, and elevated erythrocyte sedimentation rates but without thyrotoxicosis.

LABORATORY FINDINGS Biochemical thyrotoxicosis occurs in approximately half of patients with subacute thyroiditis. Serum T4 (free T4) and T3 (free T3) levels are elevated and serum TSH concentrations are suppressed. There is a disproportionate elevation of T4 relative to T3 since serum levels of thyroid hormones are due to ‘‘dumping’’ of preformed hormone into the circulation and therefore reflect intrathyroidal T4 and T3 content. This is in contrast to the thyroid hormone levels in

518 patients with Graves’ disease, in which there is a disproportionate increase of serum T3 relative to serum T4. In addition, impaired peripheral conversion of T4 to T3 due to the illness may contribute to the relatively lower T3 concentrations. In general, the thyrotoxicosis associated with subacute thyroiditis is mild or at most modest in its severity. Radioactive iodine uptake is always suppressed, usually to less than 2% after 24 h. Indeed, if the radioactive iodine uptake is more than 5% after 24 h, the diagnosis of subacute thyroiditis is unlikely. The absent uptake of iodine by the thyroid is due to destruction of the iodine-trapping mechanism from the inflammatory process as well as from inhibition of TSH secretion by excess circulating thyroid hormone. It is important to perform a radioactive iodine uptake test in patients with suspected subacute thyroiditis in order to rule out other causes of anterior neck pain as well as other types of thyrotoxicosis. The erythrocyte sedimentation rate is almost always >50 mm/h. A normal sedimentation rate in a patient with a painful anterior neck mass places the diagnosis of subacute thyroiditis in question. A mild normochromic, normocytic anemia is frequently present, as is a slightly elevated total white blood cell count. The serum thyroglobulin concentration is always elevated during the acute phase of subacute thyroiditis, reflecting the destruction of the thyroid follicular architecture and release of thyroglobulin into the circulation. However, the serum thyroglobulin is not recommended as a routine test in the evaluation of patients with suspected subacute thyroiditis since it is elevated in virtually all other thyroid disorders as well. It may be a helpful test when anterior neck discomfort is present, and when the diagnosis of subacute thyroiditis is in question, a normal serum thyroglobulin would cast serious doubt on the diagnosis. Fine needle aspiration is not routinely recommended in the evaluation of suspected subacute thyroiditis but may be employed in patients in whom the diagnosis of a painful neck mass is not clear. The principal disorders to exclude in such circumstances include acute pyogenic thyroiditis and hemorrhage into either a malignant or benign thyroid nodule. Fine needle aspiration in subacute thyroiditis shows giant cells and pseudogranulomas. Ultrasonography has been employed as an adjunct in the diagnosis of subacute thyroiditis; findings show diffuse areas of hypoechogenicity. However, ultrasound is not routinely used in the evaluation of patients with suspected subacute thyroiditis. Table II summarizes typical clinical and laboratory manifestations of subacute thyroiditis.

Thyroiditis, Subacute

Table II Clinical and Laboratory Characteristics of Subacute Thyroiditis Symptoms Viral-type prodrome (fever, myalgias, sore throat) Dysphagia Anterior neck pain (frequently with radiation) Symptoms of hypermetabolism (palpitations, tachycardia, weight loss) Signs Fever Tachycardia Tender, hard thyroid mass a Laboratory findings Serum free T4 elevated Serum TSH suppressed Erythocyte sedimentation rate elevated Thyroidal radioactive iodine uptake suppressed Serum thyroglobulin elevated a

All of the tests on this list, except the thyroglobulin, should be included in the routine evaluation of subacute thyroditis. The thyroglobulin is helpful when it is unclear if the neck pain is thyroid in origin.

TREATMENT AND CLINICAL COURSE Treatment of subacute thyroiditis is directed toward relief of pain and inflammation and control of thyrotoxic symptoms. Salicylates and nonsteroidal antiinflammatory agents have been advocated by some physicians as preferred treatment, although they tend to be effective in only the mildest cases of the disorder. Most thyroid specialists recommend the use of glucocorticoids, which are effective at relieving pain within hours after oral administration. A divided dose of 30–60 mg of prednisone daily usually suffices. Lack of significant improvement within 24 h after initiation of steroids is uncommon since dramatic relief of pain usually occurs within hours. Indeed, the absence of rapid improvement would call into question the original diagnosis. The prednisone can begin to be tapered after approximately 1 week and discontinued within 3 or 4 weeks. Pain and swelling recur, often in the contralateral lobe, in approximately 20% of patients. When this occurs, prednisone should be resumed but with a lower dose. Approximately half of the original starting dose usually suffices. The tapering process can then begin anew. Thyrotoxic symptoms may be controlled by the use of beta-blocking agents, with the dose depending on severity of symptoms. Propranolol in a dose of 10–40 mg three or four times per day or atenolol in a dose of 25–50 mg once daily is usually sufficient to control the hypermetabolic symptoms.

519

Thyroiditis, Subacute

the thyrotoxic phase, that treatment with glucocorticoids is necessary. Following the thyrotoxic phase, a several-week interval of euthyroidism occurs. A hypothyroid period then follows, which may last from a few weeks to a few months. During the hypothyroid phase, levo-thyroxine therapy may be necessary. An asymptomatic recovery phase then follows, during which the thyroid is restored to normal function. Not all patients with subacute thyroiditis progress through all four phases of the disorder since only approximately 50% of patients develop transient hypothyroidism. In most cases, the entire episode of subacute thyroiditis lasts no more than 6 months. Late recurrences of subacute thyroiditis are uncommon but may occur, even years later. Typically, a repeat bout of subacute thyroiditis is milder than the original occurrence.

Thionamide agents are of no use in the management of subacute thyroiditis and are not recommended since the thyrotoxicosis results from release of preformed hormone into the circulation rather than from increased synthesis. Therefore, drugs that inhibit thyroid hormone synthesis, such as thyroid-blocking agents, have no beneficial effect. Sodium ipodate has been shown to be effective in correcting thyrotoxicosis more rapidly, although few physicians have reported using it. Cases have been reported in which radioactive iodine ablation, or even thyroidectomy, has been employed for patients with prolonged, disabling pain, but such circumstances are rare. The clinical course of subacute thyroiditis generally follows four phases (Fig. 2). The initial or ‘‘acute’’ phase is characterized by pain, tenderness, and thyrotoxicosis, and it may last 2–12 weeks. It is during this phase, called

T3 ng/dl

T4 µg/dl Thyrotoxic

Recovery

Hypothyroid

20

300

Euthyroid T3

250

T4

15

70%

200

Normal T4 & T3 range

10

150

100

5 50

30%

I131 uptake 24 hr.

TSH µU/ml TRH test 25

60

I131uptake

15 TRH test

7.5

40

TSH

20

Normal range TSH & I131 uptake 0

1

2

3

4

5

6

7

8

9

10

Months

Figure 2 Representation of changes observed during a typical bout of subacute thyroiditis. From Emerson, C. H., and Farwell, A. P. (2000). Sporadic silent thyroiditis, postpartum thyroiditis, and subacute thyroiditis. In ‘‘Werner & Ingbar’s The Thyroid: A Fundamental and Clinical Text’’ (L. E. Braverman and R. D. Utiger, eds.), 8th ed., pp. 578–589. Lippincott Williams & Wilkins, Philadelphia.

520 Virtually all patients with subacute thyroiditis achieve restoration of normal thyroid function, although permanent hypothyroidism has been reported in 1–5% of patients. Moreover, it has been shown that patients who have suffered a bout of subacute thyroiditis may have subtle, permanent thyroid abnormalities. For example, patients have been shown to be sensitive to the inhibitory effects of exogenously administered iodides by exhibiting elevations in serum TSH concentrations, even years after having had subacute thyroiditis. Thus, perhaps patients with a history of subacute thyroiditis should be screened with serum TSH levels a few weeks after receiving exogenous iodides in pharmacologic quantities.

Acknowledgment I gratefully acknowledge the expert secretarial assistance of Elsa Ahumada.

See Also the Following Articles Thyroid Autoimmunity . Thyroid Disease, Genetic Factors in . Thyroid Fine Needle Aspiration Cytology . Thyroiditis, Infectious . Thyroiditis, Postpartum . Thyrotoxicosis, Overview of Causes

Further Reading De Bruin, T. W. A., Riekhoff, F. P. M., and DeBoer, J. J. (1990). An outbreak of thyrotoxicosis due to atypical subacute thyroiditis. J. Clin. Endocrinol. Metab. 70, 396–402. Emerson, C. H., and Farwell, A. P. (2000). Sporadic silent thyroiditis, postpartum thyroiditis, and subacute thyroiditis. In

Thyroiditis, Subacute

‘‘Werner & Ingbar’s The Thyroid: A Fundamental and Clinical Text’’ (L. E. Braverman and R. D. Utiger, eds.), 8th ed., pp. 578–589. Lippincott Williams & Wilkins, Philadelphia. Farwell, A. P., and Braverman, L. E. (1996). Inflammatory thyroid disorders. Otolaryngol. Clin. North Am. 29, 541. Guttler, R., Singer, P. A., Axline, S. G., Greaves, T. S., and McGill, J. J. (1993). Pneumocystis carinii thyroiditis. Arch. Intern. Med. 153, 393–396. Iltaka, M., Momotani, N., Ishil, J., and Ito, K. (1996). Incidence of subacute thyroiditis recurrences after a prolonged latency: 24 year survey. J. Clin. Endocrinol. Metab. 81, 466. Kitaoka, H., Sakurada, T., Fukazawa, H., et al. (1985). An epidemiological and subacute thyroiditis in northern Japan. Nippon Naibumpi Gakk. Zasshi 61, 554. Ohsako, N., Tamai, H., Sudo, T., et al. (1995). Clinical characteristics of subacute thyroiditis classified according to human leukocyte antigen typing. J. Clin. Endocrinol. Metab. 80, 3653. Roti, E., Minelli, R., Gardini, L., et al. (1990). Iodine-induced hypothyroidism in euthyroid subjects with a previous episode of subacute thyroiditis. J. Clin. Endocrinol. Metab. 70, 1581–1585. Rubin, R. A., and Guay, A. T. (1991). Susceptibility to subacute thyroiditis is genetically influenced: Familial occurrence in identical twins. Thyroid 1, 157. Singer, P. A. (1991). Thyroiditis: Acute, subacute, and chronic. Med. Clin. North Am. 75, 61–77. Smallridge, R. C., De Keyser, F. M., Van Herle, A. J., et al. (1986). Thyroid iodine content and serum thyroglobulin: Cues to the natural history of destruction-induced thyroiditis. J. Clin. Endocinol. Metab. 62, 1213. Volpe, R. (1981). Subacute thyroiditis [Review]. Prog. Clin. Biol. Res. 74, 115. Volpe, R. (1999). The management of subacute (de Quervain’s) thyroiditis [Review]. Thyroid 3, 253. Volpe, R., Row, V. V., and Ezrin, C. (1967). Circulating viral and thyroid antibodies in subacute thyroiditis. J. Clin. Endocrinol. Metab. 27, 1275–1284. Yamamoto, M., Saito, S., Sakurada, T., et al. (1987). Effect of prednisone and salicylate on serum thyroglobulin level in patients with subacute thyroiditis. Clin. Endocriol. 27, 339.

Thyrotoxic Bone Disease

phosphatase (ALP), osteocalcin, and osteopontin. Thyroid hormone receptors are central to conferring T3 responsiveness to cells by binding to target genes either as homodimers or as heterodimers complexed with the cis-acting factor 9-cisRXR. Transcriptional regulation by thyroid hormones is mediated by ligand-dependent transcription factors called TRs (e.g., TRa1, TRß1, TRß2). In osteoblasts, TRs are coexpressed with the cis- and trans-acting factors 9cisRXR (RXR) and all-trans RAR (RAR) where they modify the regulation of endogenous gene expression (e.g., ALP, osteocalcin, osteopontin) by T3. The cellular osteoblast-like lineage hOb displays lower cytokine secretion, reduced immunostaining of TR- and T3-binding sites, and decreased thyroid receptor function than does another osteoblast-like cell line, BMS; therefore, it seems that hOb cells play a lesser role and BMS cells play a greater role than previously envisaged in T3 regulation of bone remodeling. Although it is premature to extend these observations to the situation in vivo, this nonetheless highlights the potential importance of human bone marrow cells in future studies of T3 action on bone.

Osteoblast/Osteoclast Function Interference of Thyroid Hormones in Osteoblast/Osteoclast Function Thyroid hormones increase osteoclastic bone resorption by acting indirectly on osteoblasts; a direct response of osteoclasts to T3 is disputed, and most actions of T3 in bone are thought to be mediated via osteoblasts. Osteoclasts are the primary cell type that can resorb bone. They are highly motile, giant, multinucleated cells derived from hemopoietic tissue. The osteoclast precursor cells are closely related to cells of monocyte macrophage lineage. The formation of osteoclasts takes place only in the close vicinity of mineralized bone, and the multinucleated osteoclasts never appear in the circulation. The activity of osteoclasts is essential for the physiological resorption of bone during skeletal remodeling as well as for the maintenance of calcium homeostasis. Enhanced osteoclast activity is apparent in the excessive bone loss seen in a variety of pathological conditions such as thyrotoxicosis. Studies on the expression of the receptors for the stimulators of bone resorption, parathyroid hormone (PTH), and calcitriol [1,25(OH)2 D3, vitamin D3] reveal, surprisingly, that they are not present on osteoclasts or on their precursor cells but rather are expressed in bone-forming osteoblasts. The osteoblast/stromal cell ratio is crucial for the development and differentiation of the osteoclast via

535 a mechanism involving cell-to-cell contact. The osteoblastic stromal cells produce osteoclast-differentiating factors in response to a variety of stimuli that are essential for the formation of mature osteoclasts. One of these factors has been identified as RANKL (receptor activator of nuclear factor kappa B [NF-kB] ligand), which is a membrane protein expressed in the osteoblastic stromal cells and belongs to the tumor necrosis factor (TNF) family of growth factors. Osteoclast progenitors express RANK, the receptor for RANKL, at their cell surfaces, and the interaction between RANK and RANKL stimulates the development and differentiation of osteoclasts. T3 induces the expression of RANKL mRNA in primary osteoblastic cells (POB). This effect is amplified when cells are costimulated with calcitriol. In addition to RANKL, the osteoblastic stromal cells produce osteoprotegerin, (OPG), which serves as a decoy receptor for RANKL, inhibiting osteoclastogenesis by preventing the interaction between the osteoblastic stromal cell and the osteoclast progenitor (Fig. 1). The regulation of osteoclast apoptosis is a further mechanism by which osteoclast activity can be controlled. In fact, many resorptive agents are able to alter osteoclast life span and/or apoptotic rate. Osteoclast viability is increased by agents that stimulate resorption (e.g., macrophage colony-stimulating factor [M-CSF], interleukin-1 [IL-1]) that are produced by monocyte macrophage cell lines and also by the effect of thyroid hormones. In conclusion, thyroid hormones, like other resorption agents, stimulate osteoclast activity in cocultures with osteoblasts but do not stimulate activity in highly purified osteoclast preparations. The primary effect of thyroid hormones is to stimulate osteoblastic production of downstream effectors that activate the osteoclast. These downstream effectors can act either in a paracrine fashion to directly activate osteoclast activity or in an autocrine fashion to further stimulate osteoblasts to produce the paracrine factors that directly activate the osteoclasts. Local Bone Marrow Factors (Cytokines) Modulating Bone Remodeling in Thyrotoxic Bone Disease The release of cytokines by osteoblasts provides a mechanism by which osteoblasts could mediate the action of T3 on osteoclast resorption. In the microenvironment of the bone marrow, the cytokines IL1, M-CSF, and TNF can stimulate bone formation and resorption and play a critical role in regulating osteoclast formation and activity. Furthermore, it has been

s0030

536

Thyrotoxic Bone Disease

Figure 1 Schematic representation of the role of the RANK/RANKL/OPG system in osteoclast development. RANKL is expressed on the surface of the osteoblastic stromal cell and serves as a ligand for RANK, which is expressed on the surface of the osteoclast progenitor cell. The interaction between RANK and RANKL triggers the development of the mature osteoclast. OPG, produced by the osteoblastic stromal cell, serves as a decoy receptor for RANKL and inhibits osteoclastogenesis, preventing the interaction between the osteoblastic stromal cell and the osteoclast progenitor.

shown that cell cultures of human bone marrow stromal cells, containing osteoblast progenitor cells, release IL-6 and IL-8 in response to T3. The former regulates osteoclast proliferation and recruitment, whereas the latter regulates osteoclast development and activity. In addition, IL-8 receptors have been identified in osteoclast cells. Other cytokines induced by thyroid hormones are also implicated in high-turnover bone resorption. Serum IL-6 and IL-8 are produced by a number of sources, including blood monocytes and bone tissue. IL-6 mRNA is present in thyroid follicles, but high levels of cytokines, independent of the thyroid gland, are also reported in patients with ablated thyroids. The soluble serum receptor for IL-6 (sIL-6R) regulates the biological activity of IL-6, and the levels of this receptor correlate well with those of serum thyroid hormones and may prove to be a better determinant of thyrotoxic bone resorption. High levels of serum IL-6 have been demonstrated previously in thyrotoxicosis, including Graves’ disease, and also in other conditions unrelated to autoimmunity such as toxic nodular goiter and iatrogenic

subclinical thyroid excess. The elevations in serum IL-6 and IL-8 in thyrotoxicosis seem to result from the chronic effect of thyroid hormone excess rather than from the accompanying autoimmune inflammatory response produced by Graves’ thyroid or eye disease. In hyperthyroidism, as in osteoporosis from estrogen deficiency, there is an inverse relationship between the IL-6 concentration and BMC values as well as between cytokines and deoxypyridinoline (Udpd) excretion. It is possible that the same mechanism, mediated by IL-6, induces osteopenia in two conditions with high bone turnover as in postmenopausal osteoporosis and thyrotoxicosis. Serum levels of IL-6 and IL-8, as well as markers of bone remodeling, are elevated in untreated thyrotoxicosis but fall as thyroid hormone levels normalize with treatment. The normalization of Udpd precedes that of serum cytokine concentrations; although many studies have not found a correlation between serum cytokines and Udpd, the interaction of T3 with IL-6 and other cytokines should continue to be a focus in future investigations.

537

Thyrotoxic Bone Disease

Evidence of a role for cytokines in T3-associated bone resorption is lacking; this is due partly to an interference of their effects by a variety of other growth systemic factors such as PTH, calcitriol, and prostaglandin E2 (PGE2). Systemic Factors Regulating Osteoblast/ Osteoclast Function in Thyrotoxic Bone Disease Thyroid hormones, at concentrations approaching those that occur in thyrotoxicosis, also stimulate bone resorption and increase sensitivity to PTH. This causes a marked increase in bone resorption in bone organ cultures and stimulates osteoclast formation in both murine and human marrow cell cultures. Many of the calciotropic hormones and cytokines appear to act through a dual capacity to inhibit production of OPG and to stimulate the RANK system. The RANKL/RANK interaction is essential for the effect of PTH, calcitriol, and PGE2. This is demonstrated by the ability of anti-RANKL antibodies to inhibit the bone resorption activity of these known inducers of bone resorption; in contrast, estrogen appears to inhibit RANKL production. Studies have demonstrated that PTH regulates systemic levels of IL-6 in experimental animals. Furthermore, IL-6 has been shown to be an important mediator of the bone-resorbing activity of PTH in vivo and also plays a role in coupling PTH-induced bone resorption; IL-6 can stimulate proliferation of early osteoblast precursors, and PTH induces the differentiation and fusion of the precursors to form multinucleated osteoclasts. The treatment of bone cell cultures with thyroid hormones produces a gradual increase in the concentration of PGE2. This effect is abolished by indomethacine, which also reduces thyroid hormone-induced bone resorption. In mouse bone cell cultures, the effect of calcitriol on the IL-6-dependent formation of osteoclast-like cells is also subject to modulation by T3. The mechanism of interaction of these two hormones appears to involve the joint stimulation of the prostaglandin system. T3 alone does not induce osteoclast formation in cocultures of marrow cells with POB; rather, it enhances the calcitriol-induced osteoclast formation. In addition, calcitriol induces the expression of deiodinase, an enzyme that converts the prohormone 3,5,30 ,50 -tetraiodothyronine (T4) into its active form (T3). These data facilitate an understanding of the mechanism of osteoclast formation and suggest a novel interaction between thyroid hormones and calcitriol.

The addition of calcitriol to unfractionated bone cells produces a dose-dependent increase in osteoblast survival. Calcitriol also acts indirectly on osteoclasts via the production of IL-1 and IL-6 by osteoblasts, stimulating osteoclast resorption.

THYROTOXIC BONE DISEASE Biochemical Markers of Bone Remodeling in Thyrotoxic Bone Disease Thyroid hormones stimulate both osteogenesis and osteolysis to induce an acceleration in bone remodeling. Early biochemical studies attempted to quantify the calcium balance in hyperthyroidism. Mean serum calcium and phosphorus concentrations were found to be higher than those in normal controls. The relative hypercalcemia reduces serum PTH and calcitriol levels, both of which are negatively correlated with free T4 (FT4), but these levels are normalized when the hyperthyroidism is treated. Biochemical markers of bone formation and bone resorption, such as osteocalcin, ALP, bone isoenzyme, and urinary collagen pyridinoline (Upyr) or Udpd cross-links, are elevated in thyrotoxic patients and indicate increased bone formation and osteoclastic bone resorption. The binding of T3 to its nuclear receptor in osteoblasts directly stimulates the osteoblasts to produce ALP, osteocalcin, and the propeptide of type I collagen. Furthermore, osteoblast activity mediates the T3 activation of osteclasts, leading to bone resorption and the release of markers such as collagen pyridinoline and Udpd cross-links. Total ALP and osteocalcin are elevated in 30% and 65 to 90% of thyrotoxic patients, respectively. Early studies found a significant correlation between ALP and T4 serum levels and osteoid volume. Conflicting data have been published on the potential correlation between osteocalcin and bone ALP and free T3 (FT3), so the precise relationship of osteocalcin to bone ALP remains unclear. Novel markers related to collagen have been established as specific markers of bone resorption: the nonreducible cross-links of mature collagen, serum and urinary pyridinoline and deoxypyridinoline (Udpd/ creatinine), and serum carboxyterminal telopeptide type I collagen. Urinary pyridinoline levels are raised in 99% of thyrotoxic patients but are less specific for bone than are urinary Udpd cross-links. Udpd accurately indicates thyrotoxic and subclinical advanced thyrotoxic bone resorption. Humoral markers of bone metabolism, such as bone ALP, osteocalcin, and Udpd, are more important to judge the early

538 effects of antithyroid treatment than is BMD, whose changes occur too slowly. It is important to predict which patients are likely to suffer a long-term deficit in bone mineral mass so as to increase the efficacy of the therapy. Longitudinal and cross-sectional studies confirm that bone turnover is balanced within 2 to 3 weeks of therapy, and after 4 to 8 weeks of euthyroidism, a peak in the level of ALP occurs and Udpd levels return to within the normal range. However, continued increased serum ALP levels after 1 year of therapy may indicate continuing bone formation and, therefore, provide a marker for low bone mass, even in the presence of normalized Udpd excretion levels. This should be related to high levels of PTH; prior to treatment, endogenous PTH is slightly suppressed but increases significantly during the first year of therapy. The elevation of PTH seems to play a role in maintaining the high bone turnover rate despite the euthyroidism, conserving the bone calcium levels mediated by calcitriol, and increasing bone formation by an anabolic effect on osteoblasts. Patients with Graves’ disease exhibit suppressed serum thyroid-stimulating hormone (TSH) levels and often display elevated levels of TSH receptor antibody (TRAB). Cross-sectional studies indicate that markers of bone metabolism (e.g., ALP, Udpd, Upyr) are more strongly correlated with TRAB than with TSH in Graves’ clinical and subclinical patients as well as in Graves’ patients with normal TSH levels. These data could support the clinical usefulness of TRAB as a marker of bone metabolism in Graves’ patients. Because TRAB correlates with neither FT3 nor FT4 but is closely correlated with biochemical markers of bone metabolism, TRAB might directly affect bone metabolism independently of thyroid function. This hypothesis is supported by a recent study demonstrating that osteoblasts possess functional TSH receptors.

Bone Histomorphometry in Thyrotoxic Bone Disease After the first report by von Recklinghausen, the histological bone changes in thyrotoxicosis were described as being similar to those of osteite fibrosa, osteoporosis, and osteomalacia. More recently, histomorphometric analysis has shown that the bone hyperthyroid changes occurring in thyrotoxicosis are, in fact, specific and characterized by increased turnover in trabecular bone and increased remodeling and porosity in cortical bone. The predominant changes are related to bone

Thyrotoxic Bone Disease

destruction where the osteoclast activity is increased and the osteolytic activity induces and surpasses the osteogenic activity. In Graves’ bone disease, bone mineralization is initially slightly greater than bone formation; however, subsequently, the equilibrium between the increase in bone mineralization and bone formation is restored. Osteoid production can even exceed the rate of mineralization; consequently, an osteomalacia-like, osteoid-rich pattern can be observed on histological analysis of bone biopsy specimens. Because the rate of bone formation does not always equal that of bone resorption, trabecular thickness may decrease, giving rise to trabecular perforations and higher cortical porosity. In conclusion, the thyrotoxic histomorphometric pattern is unique and characterized by increased osteoblast and osteoclast activity, giving rise to a net loss of bone volume. These changes are present in both cortical and trabecular bone, although they are more evident in cortical bone, whereas a reduction of the absolute bone volume occurs less often in cortical bone.

Bone Mineral Density Markers in Thyrotoxic Bone Disease Between 1970 and 1980, BMD was measured only in peripheral sites in patients with hyperthyroidism. However, a newer and more superior method, dual energy X-ray absorptiometry (DEXA), has been developed for evaluating bone mass, where measurements are taken in lumbar vertebrae or femoral neck, typical sites of osteoporosis-related morbidity. Longitudinal studies of thyrotoxic patients demonstrate that even a slight excess of thyroid hormones, regardless of the cause, is associated with decreased BMD in cortical and trabecular bone. In thyrotoxic patients, the baseline percentage BMD of vertebrae is 92.6% as compared with that of normal controls (0.91  0.03 g/cm2 and 0.85  0.2 g/cm2, respectively). After 12 months of antithyroid therapy, the mean lumbar spine BMD increases from an initial value of 1.01 g/cm2 to 1.07 g/cm2 and increases by 6.6% per year (P < 0.001). The femoral neck BMC increases by 1.2% per year, and increases in femoral trocanter bone mineral of 3.2% per year have also been documented. No significant difference is observed between the BMDs of male and female patients. A significant reduction in BMD is seen in postmenopausal women affected by thyrotoxicosis or long-term treatment with L-T4 suppressive therapy. Some authors, but

p0165

Thyrotoxic Bone Disease

not others, have found a significant correlation between markers of bone formation and BMD. Conflicting results have also emerged regarding the correlation between BMD and the duration of hyperthyroidism, with some findings sustaining that percentage BMD is inversely correlated with levels of the TRAB. In addition, premature hair graying may be a weak marker for reduced BMD in women with a history of Graves’ disease, although that is not the case in normal women. Hyperthyroid patients display a general reduction of bone mass in the axial skeleton that is only partially corrected, mainly at the femoral neck, after 5 to 7 months of biochemical euthyroidism. In children, severe osteopenia is observed on diagnosis of Graves’ disease, with a preferential loss of cortical bone and the BMD reduced significantly, but this is rapidly corrected after 1 year of euthyroidism. After surgical treatment for Graves’ disease, thyrotoxicosis-associated bone loss in premenopausal women is fully restored, and subclinical hypothyroidism postthyroidectomy may even result in higher BMD than in controls. In contrast, surgical thyroidectomy does not produce a similar result in postmenopausal women due to the influence of the menopause on bone resorption. T4 therapy alone does not represent a significant risk factor for bone loss or, therefore, for osteoporotic fracture. Nonetheless, it is clear that there is a potential risk of bone loss in postmenopausal females.

CLINICAL ASPECTS OF THYROTOXIC BONE DISEASE Clinical studies have shown that hyperthyroidism is one of the major causes of secondary osteoporosis; however, unlike osteoporosis, it affects more than trabecular bone. The imbalance of the resorption/ formation ratio reduces BMD and, therefore, constitutes a risk of thyrotoxicosis-associated osteoporotic fractures. The etiology and duration of thyrotoxicosis do not seem to play a role in the severity of thyrotoxic bone disease, but the clinical manifestations of Graves’ disease for bone may differ depending on the age of the patient, particularly for pre- and postmenopausal female patients. Several humoral markers reflect osteogenesis. T3 enhances the functional activity of mature osteoblasts but suppresses the differentiation of osteoprogenitor cells to osteoblasts. The marker of osteoblast activity at all stages (ALP) and the marker of mature osteoblast activity (osteocalcin) are increased in thyrotoxic bone disease.

539 The osteolytic humoral marker, Udpd cross-links/ creatinine excretion, is a highly sensitive marker of increased bone metabolism in thyrotoxicosis, whereas there are only relatively small increases in osteocalcin and bone ALP determinants. After 1 year of euthyroidism, thyrotoxicosis-associated bone loss may be reversible. The normalization of thyroid hormone levels with antithyroid drugs is followed by a significant increase in lumbar spine BMD, which is preceded by a significant attenuation of increased bone turnover. This recovery is sometimes incomplete in the lumbar spine and Ward’s triangle after 5 to 7 months euthyroidism; therefore, it is important to identify hyperthyroid patients who are at risk for insufficient recovery. Some longitudinal studies suggest that high ALP levels 1 year after the initiation of antithyroid therapy are associated with reduced BMD values and could predict poor restoration of BMD. During this phase, PTH rises and insulin-like growth factor-1 (IGF-1) bioactivity could be insufficient to restore bone mass despite euthyroidism. In addition, the use of antiresorptive drugs could be encouraged in these patients. In postmenopausal thyrotoxic women, many factors in serum and the bone microenvironment unit interact to increase osteoclast activity and subsequent bone loss. This is due to the effect of increased thyroid hormone production together with a lack of skeletal protection by estrogen. The relative dehydroepiandrosterone (DHEA) and IGF-1 insufficiency that occurs during the postmenopausal period may constitute additional risk factors for developing enhanced bone loss. The high-turnover state of the early postmenopausal stage could predispose bone to the detrimental effects of hyperthyroidism, and postmenopausal women are most sensitive to accelerated bone loss from excessive T4 therapy. Conversely, other studies have demonstrated no significant difference in BMD expressed as a Z score between pre- and postmenopausal female hyperthyroid patients. This could suggest that the impact of thyrotoxicosis is great enough to surpass the effect of the menopause on bone mass, at least during the late postmenopausal period. The effect of thyroid hormones could aggravate the evident high bone turnover state during the early postmenopausal period. In contrast, during the late postmenopausal period, characterized by slow bone loss, thyroid hormone could produce a less detrimental effect on bone remodeling. This argues in favor of the early use of hormonal replacement therapy in early postmenopausal thyrotoxic women and, moreover, for the thyroid hormone replacement or suppressive therapy to be of

540 the smallest possible dose to produce the desired clinical effect. Special discussion is required for the impact of thyrotoxicosis on bone metabolism in three specific situations: thyrotoxicosis in men, bone disease in subclinical hyperthyroidism, and thyrotoxicosis in children.

Thyrotoxic Bone Disease in Men There is little information on the effect of thyroid hormones on bone mass and the risk of fractures in men; however, it is clear that thyroid hormones have a smaller effect on bone in men than in women. Nonetheless, many experts believe that hyperthyroidism is one of the most important causes of osteoporotic fractures in men occuring after alcohol abuse, glucocorticoid excess, and hypogonadism. Previous studies have shown that thyrotoxic patients display a lower radial BMD, compared with age-matched controls, and have a twofold increased risk of hip fractures. In male patients with recent-onset Graves’ disease, BMD values are reduced to levels similar to those reported in female hyperthyroid patients. As in women, this is more marked in cortical bone where there is a significant relationship with T4 levels. This is consistent with an improvement of BMD in Graves’ patients on recovery from the hyperthyroidism and with partial recovery of bone mass after attainment of euthyroidism. The presence of greater concentrations of total ALP and osteocalcin in men affected by Graves’ disease suggests the beginning of the bone mass recovery period during the early months of effective treatment. The effect of long-term suppressive T4 therapy for thyroid cancer on BMD is similar to that produced by Graves’ disease. Nonetheless, the bone loss in Graves’ disease is reversible to some extent, whereas the bone loss in patients on suppressive therapy is reasonably continuous. The prior administration of pamidronate could be used to prevent a thyroid hormone-induced increase in bone resorption. The effect of excess endogenous and exogenous thyroid hormones is mildly deleterious in the axial bone mass in male patients. In another study, the same suppressive therapy caused only a small increase in the markers of bone metabolism without detectable changes in the BMD. In hypothyroidism, thyroid hormone replacement therapy does not produce a difference in BMC, thereby excluding the possibility of a significant loss of cortical bone mineral by thyroid hormones.

Thyrotoxic Bone Disease

Normal BMD is reported in male patients on T4 substitutive therapy. Major evidence is emerging for the existence of male osteoporosis; therefore, male patients with a history of TSH suppression and thyrotoxicosis should be included in the preventive program of skeletal status assessment.

Bone Disease in Subclinical Hyperthyroidism Endogenous subclinical hyperthyroidism is identified by suppressed values of TSH in the presence of normal T4 levels. This is also observed during suppressive T4 treatment and is referred to as exogenous subclinical hyperthyroidism. BMD and bone biochemical parameters may be influenced by the duration of subclinical hyperthyroidism and by the menopausal status of the patients. For subclinical thyrotoxicosis, it appears that the appendicular, rather than the axial, skeleton is more susceptible to minor thyroid hormone excess. In premenopausal women affected by the endogenous subclinical hyperthyroidism associated with Graves’ disease or multinodular goiter, biochemical bone markers are not increased and BMD is not reduced. Conversely, long-term endogenous subclinical hyperthyroidism (2 years or more) may be a contributing factor for the development of osteoporosis and accelerated bone loss in postmenopausal women, mostly at sites where cortical bone predominates. A subtle increase in thyroid hormone, together with the lack of skeletal protection by estrogen and relative postmenopausal insufficiency of DHEA and IGF-1, is an additional risk factor for bone loss. Antithyroid treatment of endogenous subclinical hyperthyroidism has a beneficial effect on bone loss in postmenopausal women. This is substantiated by significantly higher BMD values at the distal site of the forearm in treated patients, as compared with untreated patients, during the second year of treatment. The difference is small but cumulative over many years and could result in a decreased risk of fracture. The stabilization of the decline in BMD is remarkable considering the postmenopausal status. In premenopausal women affected by subclinical hyperthyroidism for Graves’ disease, the antithyroid drug treatment produces a significant increase in bone mass and reduces the risk of secondary osteoporosis. It still appears to be important to achieve normal TSH levels in Graves’ patients during the treatment so as to normalize their bone metabolism.

541

Thyrotoxic Bone Disease

The T4 treatment could produce exogenous subclinical hyperthyroidism .The T4 dose that renders a patient euthyroid, as shown by normal TSH values, rarely produces an adverse effect. The slightly suppressive T4 administration can activate bone turnover but does not constitute a risk factor for bone loss or for osteoporosis in pre- and postmenopausal women with nontoxic goiter. Two meta-analysis studies have shown a significant reduction in BMD, but only in postmenopausal women on long-term L-T4 suppressive therapy; this was more marked in cortical bone than in trabecular bone. In postmenopausal women, the bone loss appears early during the suppressive treatment. Longterm suppressive doses of TSH produce BMD reduction at various skeletal sites and increase the risk of fractures. TSH suppressive doses should be prescribed only when appropriate and no longer than necessary so as to minimize the adverse effects of excessive doses of thyroid hormone on bone. The magnitude of bone loss, also during long-term therapy, depends on the serum levels of thyroid hormones as well as on the functional state of thyroid hormone receptors in bone tissue. However, other risk factors should be studied to prevent the possible loss of bone mass (e.g., age, weight, calcium intake in postmenopausal women, low physical activity in premenopausal women). T4 therapy represents a significant risk factor for BMD loss only in postmenopausal women with a previous history of thyrotoxicosis. In this condition, the smallest possible dose of thyroid hormone to produce the desired clinical effect must be used. Postmenopausal women taking both thyroid and estrogen hormones exhibit BMD values comparable to those observed in women taking only estrogen. The estrogen replacement therapy abolishes the reduction in femoral and vertebral BMD in postmenopausal women on L-T4 therapy. The potential beneficial influence of estrogen replacement therapy suggests that estrogen administration should be encouraged in those patients. Antiosteoclastic agents may also be appropriate as a preventive treatment in postmenopausal women at high risk for osteoporosis during T4 therapy.

Thyrotoxic Bone Disease in Children Graves’ disease is a rare condition during childhood and adolescence, with only 1 to 5% of all patients being children. The incidence of juvenile Graves’ disease ranges from 0.1 in 100,000 patients in young children to 3.0 in 100,000 patients in adolescents. Only a few investigations in children and adolescents

have been published during the past decade or so. Accelerated growth and bone maturation is observed in prepubertal children; growth acceleration is present for several months before diagnosis, and the bone age is advanced by 1.5 to 2.5 years compared with the chronological age, irrespective of weight. During this phase, the bone maturation is affected by growth hormone (GH) and thyroid hormone, whereas at puberty, it is influenced mainly by sex hormones. This may explain the failure to observe similar growth acceleration and bone maturation in the pubertal patients for whom bone age corresponds to anagraphic age. In prepubertal children, the epiphysis is also significantly altered due to the exposition to high levels of thyroid hormones. In growing children, untreated thyrotoxicosis and inappropriate T3 replacement therapy can increase osteogenesis in the short term but generally results in short-stature adults relative to predicted heights. The severe osteopenia in children that is observed on diagnosis of Graves’ disease is rapidly corrected after 1 to 2 years of treatment. Furthermore, antithyroid treatment dramatically reduces the bone resorption in pubertal girls and increases significantly both spinal and total body BMD, providing the physiological conditions to obtain optimal peak bone mass.

OUTLINE OF THERAPY FOR THYROTOXIC BONE DISEASE Experimental animal studies demonstrate that excess thyroid hormone induces cortical bone loss associated with high bone turnover that is higher in tibia than in vertebra. This is due to the effect of alendronate, evident in tibia but not in vertebra, that interferes with the recruitment of osteoclasts and increases bone volume by inhibiting osteoclast activity. In vertebra, the lack of effect of thyroid hormone and alendronate may be ascribed to a lower basal bone turnover. The pamidronate is effective at preventing bone mineral loss in ovariectomized rats, both T4 treated and untreated. This finding may have clinical relevance in estrogen-depleted patients for whom a treatment other than the reduction of T4 administration would be desirable. In normal men subjected to mild thyroid hormone excess for 8 days, the prior administration of pamidronate is useful in the prevention of thyroid hormone-induced increased bone resorption or induced osteopenia. Alendronate also produces an increase in BMD and a corresponding decrease in serum osteocalcin levels in both pre- and postmenopausal women.

542 A recent study assessed the effect of antiresorptive therapy with nasal calcitonin on bone metabolism in recently diagnosed hyperthyroid patients. The effect of exogenous calcitonin was greater in the patients than in normal controls. The significant reduction in axial BMD in thyrotoxic patients was partially restored after attainment of the euthyroid state. Nonetheless, recovery was incomplete, with a 5% deficit compared with controls. The treatment with nasal calcitonin had no additional effect after attainment of the euthyroid state. In postmenopausal women, estrogen replacement therapy is effective at preventing an increase in bone mineral metabolism in high-dose T4 treatment as in thyrotoxicosis. For the treatment of bone thyrotoxic disease in children, the antithyroid drug therapy dramatically reduces the bone resorption and increases significantly both spinal and total BMD, providing physiological conditions for the achievement of the peak bone mass. In summary, published evidence indicates that TSH suppresser therapy for subclinical Graves’ disease, and even minimally excessive thyroid hormone replacement or chronic suppression therapy, can be accompanied by a decrease in cortical bone mass. This raises important questions about the use of lifelong replacement therapy, especially for young patients. These data should also provoke caution in the use of traditional suppressive therapy for thyroid nodules. The correct interpretation of studies such as these requires keeping in mind the concept of remodeling space. Predictable changes in bone density will be observed simply by increasing or restricting the remodeling space. However, after some months, a new equilibrium will be achieved. To evaluate the clinical impact of such changes, it is necessary to carry out longer term studies than those that have been

Thyrotoxic Bone Disease

reported, but for the present time, the smallest possible dose of thyroid hormones to achieve an euthyroid state or to suppress thyroid gland activity is advisable. The critical effect of suppressive T4 treatment on bone in thyroid cancer could be balanced by pamidronate administration. In these patients, it is also critical to assess bone mineral metabolism.

Acknowledgement I am indebted to Tracy Williams for her collaborative support.

See Also the Following Articles Bone Mass Measurement . Bone Turnover Markers . Hyperthyroidism, Subclinical . Interleukin-6 . Osteoporosis, Overview . Paget’s Disease of Bone . Parathyroid Hormone (PTH) . Tumor Necrosis Factor (TNF)

Further Reading Greenspan, S., and Greenspan, F. (1999). The effect of thyroid hormone on skeletal integrity. Ann. Int. Med. 130, 750–758. Harper, K. D., and Weber, T. (1998). Secondary osteoporosis: Diagnostic consideration. Endocrinol. Metab. Clinics North America 27, 325–348. Lerner, U. H. (2000). Osteoclast formation and resorption. Matrix Biol. 19, 107–120. Mosekilde, L., Melsen, F., Bragger, J. P., Myhre-Jensen, O., and Schwartz Soerensen, N. (1977). Bone changes in hyperthyroidism: Interrelationship between bone morphometry, thyroid function, and calcium–phosphorus metabolism. Acta Endocrinol. 85, 515–525. Motomura, K., and Brent, A. G. (1998). Mechanism of thyroid hormone action: Implications for the clinical manifestation of thyrotoxicosis. Endrocrinol. Metab. Clinics North America 27, 1–23. Pantazi, H., and Papapetrou, D. P. (2000). Changes in parameters of bone and mineral metabolism during therapy for hyperthyroidism. J. Clin. Endocrinol. Metab. 85, 1099–1106.

574 2-h radioiodine uptake, although it should be feasible to obtain the result of serum T4 determination within a few hours on an emergency basis in most hospitals today. However, initiation of therapy should not be postponed when there is a high index of suspicion merely because one is awaiting laboratory confirmation of the diagnosis. Other laboratory abnormalities often include modest hyperglycemia in the absence of diabetes mellitus. Moderate leukocytosis with a mild shift to the left is common even in the absence of infection. Increased serum calcium levels may be seen, perhaps due to both hemoconcentration and the known effects of thyroid hormone on bone resorption, but other serum electrolytes are usually normal. Hepatic dysfunction in thyrotoxic storm will result in elevated levels of serum lactate dehydrogenase, aspartate aminotransferase, and bilirubin. Because serum cortisol levels are usually elevated in thyrotoxic individuals, a normal value may be interpreted as being inappropriately low. In view of the known coincidence of adrenal insufficiency with Graves’ disease, one should maintain a reasonably high index of suspicion for this disorder, particularly if there is hypotension and suggestive electrolyte abnormalities. It would be prudent to obtain a serum sample for cortisol determination prior to the administration of corticosteroid. Even in the absence of adrenal insufficiency, adrenal reserve may be exceeded in thyrotoxic crisis due to the inability of the adrenal gland to meet the demand placed on it as a result of the accelerated turnover and disposal of glucocorticoids that occur in thyrotoxicosis.

PATHOGENESIS The precise pathogenesis underlying the precipitation of thyrotoxic storm is likely not to be the same in all cases and remains incompletely understood, although the magnitude of serum hormone levels per se does not appear to be critical. However, acute discharge of hormone resulting in rather sudden changes of its concentration, in the appropriate clinical setting, certainly can trigger crisis. This may be seen after 131-I therapy, withdrawal of propylthiouracil (PTU) therapy, vigorous palpation of the thyroid, or administration of lithium, stable iodine, or iodinated contrast dyes. A possible interaction between the effects of excessive levels of circulating thyroid hormone and the catecholamines has been proposed. This is evidenced by the dramatic clinical improvement that follows the use

Thyrotoxic Storm

of agents that either deplete their tissue levels, such as reserpine, or block b-adrenergic receptors, such as propranolol. Although these agents are useful adjuncts to therapy, they should not be used by themselves because they might not prevent the occurrence of storm.

TREATMENT We believe that to avoid a fatal outcome, it is important to implement a four-pronged approach to the management of thyrotoxic storm. The relative importance for survival of each arm of therapy will vary in a given patient. First, specific antithyroid drugs must be used to decrease the increased thyroid production and release of thyroxine (T4) and T3. The second part of management consists of treatment intended to block the effects of the remaining, but excessive, circulating concentrations of free T4 and T3. The third component addresses any underlying precipitating illness such as infection or ketoacidosis. The final arm of therapy is composed of those several specific treatments that are directed against the underlying systemic decompensation that may be characterized by fever, congestive failure, shock, and the like. In view of the poor prognosis associated with incompletely treated thyrotoxic storm, no one component of this four-pronged therapeutic approach should be neglected.

Therapy Directed against the Thyroid Gland Inhibition of new synthesis of the thyroid hormones is achieved by administration of thionamide antithyroid drugs, such as PTU and methimazole (Tapazole), given orally. There are no available parenteral preparations of these compounds. Either methimazole or PTU may also be administered via the rectum if necessary. In view of the gravity of thyrotoxic storm, thionamide doses are much higher than those for otherwise uncomplicated thyrotoxicosis. Some experienced clinicians believe that PTU will provide more rapid clinical improvement because it has the additional advantage of inhibiting conversion of T4 to T3, a property not shared by methimazole. Separate treatment must be administered to inhibit the continuing release of T4 and T3 into the blood because thionamides act to reduce new hormone synthesis but have no effect on thyroidal secretion of preformed stores of hormone. Either inorganic iodine or lithium carbonate may be used for this purpose. Iodides may be given either orally as Lugol’s solution or as a saturated solution of potassium iodide (eight

s0025

575

Thyrotoxic Storm

drops every 6 h). When iodine is administered together with full doses of antithyroid drugs, dramatic decreases in serum T4 can be seen. The sequence of administration of these agents is extremely important. Use of iodine without prior thionamide dose is contraindicated because the iodine will provide extra substrate to enrich hormone stores within the gland, thereby generating the potential for further exaggeration of thyrotoxicosis. In patients who may be allergic to iodine, lithium carbonate may be used as an alternative agent to inhibit hormonal release, although some caution has been raised in regard to its use in the setting of storm. This drug also may be used in thyrotoxic patients who are known to have serious toxic reactions to the thionamides.

Therapy Directed against Ongoing Effects of Thyroid Hormone in the Periphery For the purpose of acutely reducing the circulating hormones, peritoneal dialysis and plasmapheresis have been employed, as has experimental hemoperfusion through a resin bed or charcoal columns. Such aggressive management should be considered in severe cases. Beta blockers are also commonly used. Propranolol is the agent most commonly used in the United States today. Large doses, such as 60 to 120 mg every 6 h, are used in crisis or impending crisis. Indeed, because of the more rapid metabolism of the drug in severe thyrotoxicosis, even larger oral doses, or preferably intravenous doses, should be given. Initial intravenous doses should be given cautiously, whereas the patient’s cardiac rhythm is monitored continuously. Added benefits of b-adrenergic blockade in these patients include improvement in agitation, convulsions, psychotic behavior, tremor, diarrhea, fever, and diaphoresis.

s0035

broad-spectrum antibiotic coverage may be warranted while awaiting results of cultures. In most patients who survive thyrotoxic crisis, clinical improvement is dramatic and demonstrable within 12 to 24 h.

Therapy Directed against Systemic Decompensation Fluid depletion caused by the hyperpyrexia, diaphoresis, vomiting, or diarrhea must be vigorously replaced to avoid vascular collapse. Fluid management must be individualized. Shock may be refractory to cautious fluid resuscitation in younger patients, whereas judicious replacement of fluids is necessary in elderly patients with congestive heart failure or other cardiac compromise. Intravenous fluids containing 10% dextrose in addition to electrolytes will allow repletion of the depleted hepatic glycogen. For fever, acetaminophen, rather than salicylates, is the preferred antipyretic because salicylates inhibit thyroid hormone binding and could increase free hormone, thereby transiently worsening the thyrotoxic crisis. Vasopressor therapy may become necessary on a temporary basis to provide adequate hemodynamic support if hydration with intravenous fluid replacement is not effective. Stress dose glucocorticoids have been given on empirical grounds on the basis of postulated relative adrenal insufficiency. This approach has the added benefit of further blockade of peripheral conversion of T4 to T3, and this is an additional justification for their use.

See Also the Following Articles Graves’ Ophthalmopathy . Thyrotoxicosis: Diagnosis . Thyrotoxicosis Factitia . Thyrotoxicosis, Overview of Causes . Thyrotoxicosis, Systemic Manifestations . Thyrotoxicosis, Treatment

Therapy Directed against the Precipitating Illness

Further Reading

In most cases of thyrotoxic storm, therapy is not complete unless a diagnosis of the possible precipitating event is made and early treatment as indicated for that underlying illness is implemented. It is important to be alert to the fact that conditions such as ketoacidosis, pulmonary thromboembolism, and stroke may underlie thyrotoxic crisis, particularly in the obtunded or psychotic patient, and require the same vigorous management ordinarily indicated. In the patient with thyrotoxic crisis in whom none of the latter precipitating factors is apparent, a diligent search for some focus of infection must be carried out. Empirical,

Aiello, D. P., DuPlessis, A. J., Pattishall, E. G., III, and Kulin, H. E. (1989). Thyroid storm presenting with coma and seizures. Clin. Pediatr. 28, 571–574. Ashkar, F. S., Katims, R. B., Smoak, W. M., and Gilson, A. J. (1970). Thyroid storm treatment with blood exchange and plasmapheresis. J. Am. Med. Assoc. 214, 1275–1279. Burman, K. D., Yeager, H. C., and Briggs, W. A. (1976). Resin hemoperfusion: A method of removing circulating thyroid hormones. J. Clin. Endocrinol. Metab. 42, 70–78. Eriksson, M. A., Rubenfeld, S., Garber, A. J., and Kohler, P. O. (1977). Propranolol does not prevent thyroid storm. N. Engl. J. Med. 296, 263–264. Feely, J., Forrest, A., and Gunn, A. (1980). Propranolol dosage in thyrotoxicosis. J. Clin. Endocrinol. Metab. 51, 658–661.

576 Herrmann, J., Hilger, P., and Kruskemper, H. L. (1973). Plasmapheresis in the treatment of thyrotoxic crisis (measurement of half-concentration tissues for free and total T3 and T4). Acta Endocrinol. (Copenh) 173(Suppl.), 22. McDermott, M. T., Kidd, G. S., Dodson, L. E., and Hofeldt, F. D. (1983). Radioiodine-induced thyroid storm. Am. J. Med. 75, 353–359.

Thyrotropin see TSH (Thyroid-Stimulating Hormone)

Thyrotropin Receptor see TSH Receptor

Thyrotoxic Storm

Wartofsky, L., Ransil, B. J., and Ingbar, S. H. (1970). Inhibition by iodine of the release of thyroxine from the thyroid glands of patients with thyrotoxicosis. J. Clin. Invest. 49, 78–86. Yeung, S. C., Go, R., and Balasubramanyam, A. (1995). Rectal administration of iodide and propylthiouracil in the treatment of thyroid storm. Thyroid 5, 403–405.

552

Thyrotoxicosis Factitia

Table II Differential Diagnosis of Main Forms of Thyrotoxicosis with Low/Suppressed Thyroidal Radioactive Iodine Uptake Subacute thyroiditis

Iodine-induced thyrotoxicosis

Thyrotoxicosis factitia

Serum-free thyroid hormones

Increased

Increased

Serum TSH

Suppressed

Suppressed

Increased Suppressed

Serum thyroglobulin

Increased

Increased

Low/Undetectable

Serum interleukin-6

Increased

Increased

Low

Urinary iodine excretion

Normal

Increased

Normal

Thyroid echogenicity Thyroid vascularity

Markedly reduced and not homogeneous Reduced

Slightly reduced Reduced

Normal Reduced

Thyroid pain/tenderness

Present

Absent

Absent

CLINICAL FEATURES The clinical picture of thyrotoxicosis factitia does not differ from that of classical spontaneous hyperthyroidism. Patients complain of tachycardia, tremors, loss of weight, increased perspiration, heat intolerance, extreme anxiety and nervousness, increased bowel activity, and/or insomnia (Table I). Goiter and ophthalmopathy, as seen in Graves’ disease, are absent. Likewise, thyroid pain and tenderness, commonly observed in subacute thyroiditis, are absent (Table I). Cardiovascular complications (e.g., tachyarrhythmias, heart failure, myocardial infarction) as well as bone loss (osteopenia) may occur as a consequence of prolonged excess thyroid hormone ingestion. These complications are more likely to take place in older patients. Laboratory evaluation reveals the typical increase in free T4 (FT4) and free T3 (FT3) concentrations associated with undetectable serum thyrotropin (TSH) levels. An isolated increase in serum FT3 concentration associated with suppressed serum FT4 levels indicates that the ingested thyroid hormone preparation contained only T3. Circulating autoantibodies to thyroglobulin or thyroperoxidase, as well TSH receptor autoantibodies responsible for Graves’ disease, are usually absent. Thyroidal radioactive iodine uptake (RAIU) is characteristically very low or suppressed (Table I). Serum thyroglobulin concentration is typically markedly reduced or undetectable in thyrotoxicosis factitia (Table I). Accordingly, its measurement is a useful tool for differentiating thyrotoxicosis factitia from other thyrotoxic conditions associated with low RAIU values (Table II). Serum interleukin-6, a marker of thyroidal destructive processes, is also undetectable in thyrotoxicosis factitia (Table II). Measurement of thyroid hormones in stools may be useful for identifying the abnormally high fecal excretion of ingested thyroid hormones.

Urinary iodine excretion is normal (Table II). Color flow Doppler sonography of the thyroid shows an absent vascularity and normal–low peak systolic velocity in spite of the thyrotoxic state (Table II). If the patient denies the surreptitious thyroid hormone intake, it may be necessary to hospitalize him or her to be sure that the deliberate ingestion does not continue. Under strict medical controls, a rapid improvement in the clinical and laboratory features of thyrotoxicosis is usually observed. However, recurrency of thyrotoxicosis is frequently observed after hospital release unless reasons for surreptitious thyroid hormone intake are identified.

TREATMENT Treatment of thyrotoxicosis factitia obviously requires withdrawal of thyroid hormones. It may be useful to associate a short-term treatment with b-adrenergic blocking drugs to control tachycardia and tremors promptly. However, for a full recovery of the patient, psychiatric aid or counseling is mandatory in all cases.

See Also the Following Articles Anorexia Nervosa . Depression, Thyroid Function and . Thyroglobulin . Thyrotoxicosis: Diagnosis . Thyrotoxicosis, Overview of Causes . Thyrotoxicosis, Systemic Manifestations . Thyrotoxicosis, Treatment

Further Reading Bartalena, L., Bogazzi, F., and Martino, E. (1996). Adverse effects of thyroid hormone preparations and antithyroid drugs. Drug Safety 15, 53–63. Bartalena, L., and Vitti, P. (2001). Hyperthyroidism and thyrotoxicosis. In ‘‘Endocrinology and Metabolism’’ (A. Pinchera, ed.), pp. 161–172. McGraw–Hill International, London. Bogazzi, F., Bartalena, L., Scarcello, G., Campomori, A., Rossi, G., and Martino, E. (1999). The age of patients with thyrotoxicosis

Thyrotoxicosis Factitia

factitia in Italy from 1973 to 1996. J. Endocrinol. Invest. 22, 128–133. Bogazzi, F., Bartalena, L., Vitti, P., Rago, T., Brogioni, S., and Martino, E. (1996). Color flow Doppler sonography in thyrotoxicosis factitia. J. Endocrinol. Invest. 19, 603–606. Bouillon, R., Verresen, L., Staels, F., Bex, M., De Vos, P., and De Roo, M. (1993). The measurement of fecal thyroxine in the diagnosis of thyrotoxicosis factitia. Thyroid 3, 101–103. Cohen, J. H., Ingbar, S. H., and Braverman, L. E. (1989). Thyrotoxicosis due to ingestion of excess thyroid hormone. Endocrine Rev. 10, 113–124.

553 Hamolsky, M. W. (1982). Truth is stranger than factitious. N. Engl. J. Med. 307, 436–437. Mariotti, S., Martino, E., Cupini, C., Lari, R., Giani, C., Baschieri, L., and Pinchera, A. (1982). Low serum thyroglobulin as a clue to the diagnosis of thyrotoxicosis factitia. N. Engl. J. Med. 307, 410–412. Rose, E., Sanders, T. P., Webb, W. L., Jr., and Hines, R. C. (1969). Occult factitial thyrotoxicosis. Ann. Int. Med. 71, 309–315. Roti, E., Minelli, R., Gardini, E., and Braverman, L. E. (1993). The use and misuse of thyroid hormone. Endocrine Rev. 14, 401–423.

Table I Classification of Known Causes of Thyrotoxicosis, Epidemiology, and Distinctive Diagnostic Features

Group

Disease

Relative frequency (percentage)

Distinctive features

Neck RAIU

Serum TSH

Serum thyroglobulin

Thyrotoxicosis of thyroidal origin Associated with hyperthyroidism

Graves’ disease

70

Diffuse goiter

High

Undetectable

High

High

Ophthalmopathy Positive TRABs

Associated with thyroid destruction

Thyrotoxicosis of nonthyroidal origin

Toxic adenoma

5

Single ‘‘hot’’ nodule at thyroid scan

High

Undetectable

Toxic multinodular goiter

20

Multiple ‘‘hot’’ nodules at thyroid scan

High

Undetectable

High

Iodine-induced thyrotoxicosis TSH-secreting adenomas