Principles of Developmental Genetics [Volume 6, 1 ed.] 0123695481, 9780123695482, 9780080550718

Citation preview

PRINCIPLES OF DEVELOPMENTAL GENETICS

This page intentionally left blank

PRINCIPLES OF DEVELOPMENTAL GENETICS

SALLY A. MOODY George Washington University

AMSTERDAM • BOSTON • HEIDELBERG • LONDON NEW YORK • OXFORD • PARIS • SAN DIEGO SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO Academic Press is an imprint of Elsevier

Academic Press is an imprint of Elsevier 30 Corporate Drive, Suite 400, Burlington, MA 01803, USA 525 B Street, Suite 1900, San Diego, California 92101-4495, USA 84 Theobald’s Road, London WC1X 8RR, UK

This book is printed on acid-free paper. Copyright © 2007, Elsevier Inc. All rights reserved.

Notice No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone: (þ44) 1865 843830, fax: (þ44) 1865 853333, E-mail: [email protected]. You may also complete your request on-line via the Elsevier homepage (http://elsevier.com), by selecting “Support & Contact” then “Copyright & Permission,” and then “Obtain Permission.”

Library of Congress Cataloging-in-Publication Data 2007926494 British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-369548-2

For all information on all Elsevier Academic Press publications visit our Web site at www.books.elsevier.com

Printed in the United States of America 07 08 09 10 11 12 9 8 7 6 5 4 3 2 1

CONTENTS

PREFACE

XI

Sally A. Moody

I

THE IMPACT OF GENETIC AND GENOMIC TOOLS ON DEVELOPMENTAL BIOLOGY

1 Untangling the Gordian Knot: Cell Signaling Events That Instruct Development 2 RENE´E V. HOCH AND PHILIPPE SORIANO

2 Finding Gene Expression Changes Using Microarray Technology 32 TADAYOSHI HAYATA AND KEN W. Y. CHO

3 Chemical and Functional Genomic Approaches to Study Stem Cell Biology and Regeneration 45 WEN XIONG AND SHENG DING

4 Assessing Neural Stem Cell Properties Using Large-Scale Genomic Analysis 69 SOOJUNG SHIN, JONATHAN D. CHESNUT, AND MAHENDRA S. RAO

5 Epigenetic Influences on Gene Expression Pathways 92 SUNDEEP KALANTRY AND TERRY MAGNUSON

6 New Insights into Vertebrate Origins 114 BILLIE J. SWALLA

7 Understanding Human Birth Defects Through Model 129 Organism Studies FEYZA ENGIN AND BRENDAN LEE

v

vi

CONTENTS

II

EARLY EMBRYOLOGY, FATE DETERMINATION, AND PATTERNING

8 Germ Line Determinants and Oogenesis 150 KELLY M. HASTON AND RENEE A. REIJO PERA

9 Patterning the Anterior–Posterior Axis During Drosophila Embryogenesis 173 KRISTY L. KENYON

10 Anterior–Posterior Patterning in Mammals 201 SIGOLE`NE M. MEILHAC

11 Signaling Cascades, Gradients, and Gene Networks in Dorsal/Ventral Patterning 216 GIRISH S. RATNAPARKHI AND ALBERT J. COUREY

12 Early Development of Epidermis and Neural Tissue 241 KEIJI ITOH AND SERGEI Y. SOKOL

13 Formation of the Embryonic Mesoderm 258 LISA L. CHANG AND DANIEL S. KESSLER

14 Endoderm 295 DE´BORA SINNER, JAMES M. WELLS, AND AARON M. ZORN

15 Notch Signaling: A Versatile Tool for the Fine Patterning of Cell Fate in Development 316 AJAY B. CHITNIS

16 Multiple Roles of T-box Genes 341 L. A. NAICHE AND VIRGINIA E. PAPAIOANNOU

III

MORPHOGENETIC AND CELL MOVEMENTS

17 Gastrulation in Vertebrates 360 LILIANNA SOLNICA-KREZEL AND DIANE S. SEPICH

18 Regulation of Tissue Separation in the Amphibian Embryo 392 HERBERT STEINBEISSER

vii

CONTENTS

19 Role of the Basement Membrane in Cell Migration 404 KIYOJI NISHIWAKI AND YUKIHIKO KUBOTA

20 Epithelial Morphogenesis 424 RONIT WILK AND HOWARD D. LIPSHITZ

21 Branching Morphogenesis of Mammalian Epithelia 448 JAMIE DAVIES

22 The Roles of Ephrin–Eph in Morphogenesis 467 IRA O. DAAR

IV

ECTODERMAL ORGANS

23 Neural Cell Fate Determination 500 STEPHEN N. SANSOM AND FREDERICK J. LIVESEY

24 Pathfinding and Patterning of Axonal Connections 525 STEPHANIE A. LINN, STEPHANIE R. KADISON, AND CATHERINE E. KRULL

25 Retinal Development 548 KATHRYN B. MOORE AND MONICA L. VETTER

26 Neural Crest Determination 574 ROBERTO MAYOR

27 Determination of Preplacodal Ectoderm and Sensory Placodes 590 SALLY A. MOODY

28 Molecular Genetics of Tooth Development 615 IRMA THESLEFF

29 The Inner Ear 631 DONNA F. FEKETE AND ULRIKE J. SIENKNECHT

30 Craniofacial Formation and Congenital Defects 656 S. A. BRUGMANN AND J. A. HELMS

V

MESODERMAL ORGANS

31 Induction of the Cardiac Lineage 680 ANDREW S. WARKMAN AND PAUL A. KRIEG

viii

CONTENTS

32 Heart Patterning and Congenital Defects 698 JOHN W. BELMONT

33 Blood Vessel Formation 721 KARINA YANIV AND BRANT M. WEINSTEIN

34 Blood Induction and Embryonic Formation 755 XIAOYING BAI AND LEONARD I. ZON

35 Topics in Vertebrate Kidney Formation: A Comparative Perspective 778 THOMAS M. SCHULTHEISS

36 Development of the Genital System 805 HONGLING DU AND HUGH S. TAYLOR

37 Diaphragmatic Embryogenesis and Human Congenital Diaphragmatic Defects 829 KATE G. ACKERMAN AND DAVID R. BEIER

38 Formation of Vertebrate Limbs 847 YINGZI YANG

39 Skeletal Development 866 PETER G. ALEXANDER, AMANDA T. BOYCE, AND ROCKY S. TUAN

VI

ENDODERMAL ORGANS

40 Patterning the Embryonic Endoderm into Presumptive Organ Domains 908 BILLIE A. MOORE-SCOTT AND JAMES M. WELLS

41 Developmental Genetics of the Pulmonary System 932 THOMAS J. MARIANI

42 Pancreas Development and Stem Cells 946 MAUREEN GANNON

ix

CONTENTS

43 Early Liver Development and Hepatic Progenitor Cells 982 JAY D. KORMISH AND KENNETH S. ZARET

44 Intestinal Stem Cells in Physiologic Regeneration and Disease 1004 DAVID H. SCOVILLE, XI C. HE, GOO LEE, TOSHIRO SATO, TERRENCE A. BARRETT, AND LINHENG LI

INDEX

1023

This page intentionally left blank

PREFACE Developmental Genetics, or What Can Genetics and Genomics Tell Us About Evolution, Development, Stem Cells, Human Birth Defects, and Disease? Sally A. Moody Department of Anatomy and Cell Biology, George Washington University

The ability of researchers to answer experimental questions greatly depends on the available technologies. New technologies lead to novel observations and field-changing discoveries and influence the types of questions that can be asked. Today’s recently available technologies include sequencing and analyzing the genomes of human and model organisms, genome-wide expression profiling, and high-throughput genomic and genetic analyses. The information provided by these approaches is enabling us to begin to understand the complexity of many biological processes through the elucidation of gene regulatory networks, signaling pathway networks, and epigenetic modifications. This book describes many lines of research that are being impacting by these new technologies, including developmental genetics and the related fields of clinical genetics, birth defects research, stem cell biology, regenerative medicine, and evolutionary biology. The field of developmental genetics, or the study of how genes influence the developmental processes of an organism, has been influenced by new technologies and by interactions with other fields of study throughout its history. The concept of a genetic basis of development began in “modern” times at the intersection of descriptive embryology and cytology. Modern histological techniques were developed in the mid-19th century, largely by Wilhelm His so that he could study cell division in the neural tube, which enabled visualization of the cell nucleus, chromosomes, and the discrete steps of mitosis. Theodor Boveri cleverly applied these improved microscopic techniques to transparent marine embryos to demonstrate that each parent contributes equivalent groups of chromosomes to the zygote, and that each chromosome is an independently inherited unit. Importantly, he noted that if an embryo contains the incorrect number or improper combination of chromosomes, it develops abnormally. However, many early embryologists rejected the idea that development is driven by prepackaged heritable particles because it seemed too similar to the idea of “preformation”: the concept that development is driven by predetermined factors or “forces” (sometimes described in rather mystical terms).

xi

xii

PREFACE

Wilhelm Roux, an advocate of studying the embryo from a mechanistic point of view, was a leader in the approach of manipulating the embryo with microsurgical techniques to elucidate cause and effects between component parts (experimental embryology). By using an animal model whose embryos were large, developed external to the mother, could be surgically manipulated with sharpened forceps and cultured in simple salt media (i.e., amphibians), he rejected the role of predetermined factors and demonstrated the importance of external (epigenetic) influences and cell–cell interactions in regulating developmental programs. Experimental embryologists further refined their skills at dissecting small bits of tissue from the embryo, recombining them with other tissues in culture or transplanting them to ectopic regions in the embryo. This work led to the invention of tissue culture by Ross Harrison and the discovery of tissue inductions by Hans Spemann. While experimental embryology was thriving, T. H. Morgan founded the field of Drosophila genetics. Also trained as an embryologist, Morgan was skeptical of Boveri’s idea of heritable packets, and directed his studies towards understanding the principles of inheritance. For several decades, the two fields had little impact on one another. Interestingly, however, after a few decades of study of the fruit fly, Morgan’s work supported the idea of discrete intracellular particles that directed heritable traits, which he named “genes.” Nonetheless, the fields of experimental embryology and genetics remained fairly separate entities with distinct goals and points of view. Embryologists were elucidating the interactions that are important for the development of numerous tissues and organs, whereas geneticists were focused on the fundamentals of gene inheritance, regulation of expression, and discovering the genetic code. Indeed, elucidating the genetic basis of vertebrate development was delayed until new technologies in molecular biology and cloning were devised. From the field of bacterial and viral genetics came the techniques for cloning eukaryotic genes and constructing vectors for controlling expression. From the classical genetic studies in fly and nematode came the rationale for mutagenizing the entire genome and screening for developmental abnormalities. Important regulatory genes were discovered in these invertebrates, and their counterparts were discovered in many other animals by homology cloning approaches. Thus was born the modern field that we call developmental genetics. An important advance in the past decade is the demonstration that genes that regulate developmental processes in invertebrate species have important developmental functions in vertebrates. The wealth of information concerning the molecular genetic processes that regulate development in various animals demonstrates that developmental programs and biological processes are highly conserved, albeit not identical, from yeast to human. Indeed, the Human Genome Project has made it possible to identify the homologues in humans and demonstrate that many of these regulatory genes underlie human developmental disorders and aspects of adult diseases in which differentiation processes go awry. Currently, researchers are studying the fundamentals of developmental processes in the appropriate animal model and screening humans for mutations in the genes identified by the basic research to be likely causative candidates. Researchers are mutagenizing vertebrate animal models and screening for mutants that resemble known human syndromes. This cross-fertilization of fields is also impacting concepts in evolutionary biology,

PREFACE

xiii leading to a better understanding of “ancestral” species via gene expression profiles, and paradigms in stem cell biology in which naı¨ve cells may be directed to “designer” lineages. Most recently, there have been significant technological advances in genetic, genomic, and protein expression analyses that are having a major impact on experimental approaches and analytic design. The intersection of developmental biology with these technologies offers a new view of developmental genetics that is only beginning to be exploited. It is this new intersection at the onset of the genomic era that is the focus of this book. The book is organized into sections focused on different aspects of developmental genetics. Section I discusses the impact of new genetic and genomic technologies on development, stem cell biology, evolutionary biology, and understanding human birth defects. Section II discusses several major events in early embryogenesis, fate determination, and patterning, including cellular determinants (Boveri revisited?), gene cascades regulating embryonic axis formation, signaling molecules and transcription factors that regulate pattern formation, and the induction of the primary germ layers (ectoderm, mesoderm, and endoderm). Section III describes the reorganization of the embryo via different types of morphogenetic and cellular movements that result in the foundation of organ systems, and discusses the many signaling and adhesion molecules that are involved in regulating these complex processes. The final three sections focus on the signaling cascades and transcriptional pathways that regulate organogenesis in representative systems derived from the embryonic ectoderm, mesoderm, and endoderm. These chapters illustrate how embryonic rudiments become organized into adult tissues, and how defects in these processes can result in congenital defects or disease. Each chapter demonstrates the usefulness of studying model organisms and discusses how this information applies to normal human development and clinical disorders. Several chapters also discuss the utility of stem cells to repair damaged organs and the application of developmental genetics to the manipulation of stem cells for regenerative medicine. The goal of this book is to provide a resource for understanding the critical embryonic and prenatal developmental processes that are fundamental to the normal development of animals, including humans. It highlights new technologies to be used, new questions to be answered, and the important roles that invertebrate and vertebrate animal models have had in elucidating the genetic basis of human development. Developmental genetics has reemerged from its birth a century ago as a nexus of diverse fields that are using the common language of gene sequence and function. This is influencing what questions are posed and how the answers are used. New technologies are making it relatively easy to study gene expression and regulation at single cell, tissue, and embryonic levels. The conservation between the genomes of species that are separated by vast evolutionary time encourages us to more fully utilize animal models to gain important insights into the clinical relevance of the animal model data. It is our hope that this book will stimulate even more cross-fertilization and interactions between evolutionary biology, developmental biology, stem cell biology, basic scientists, and clinical scientists. I wish to thank all of the authors for contributing such exciting and excellent chapters, and Pat Gonzalez for keeping all of us on schedule.

xiv

PREFACE

RECOMMENDED RESOURCES Baltzer F: Theodor Boveri, life and work of a great biologist, Los Angeles, CA, 1967, University of California Press. Hamburger V: The heritage of experimental embryology: Hans Spemann and the organizer, New York, 1988, Oxford University Press. Model Organisms for Biomedical Research (Web site): http://www.nih.gov/science/models. Morgan TH: The theory of the gene, New Haven, CT, 1926, Yale University Press. Spemann H: Embryonic development and induction, New York, 1967, Hafner Publishing Co. Willier BH, Oppenheimer JM: Foundations of experimental embryology, Englewood Cliffs, NJ, 1964, Prentice-Hall.

I THE IMPACT OF GENETIC AND GENOMIC TOOLS ON DEVELOPMENTAL BIOLOGY

1

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT RENE´E V. HOCH and PHILIPPE SORIANO Program in Developmental Biology and Division of Basic Sciences, Fred Hutchinson Cancer Research Center, Seattle, WA

INTRODUCTION The developmental cell signaling field has evolved out of convergent work in developmental genetics and biochemistry. Landmark studies were performed during the 1980s and 1990s, when genetic screens identified mutants that enhanced or suppressed receptor tyrosine kinase (RTK) loss-of-function phenotypes in Drosophila melanogaster (sevenless, torso) and Caenorhabditis elegans (Egfr) embryos. Such mutants were arranged into hierarchies on the basis of epistatic relationships and cell-autonomous versus cell-nonautonomous effects on RTK functions (reviewed in Furriols and Casanova, 2003; Moghal and Sternberg, 2003; Shilo, 2003; Nagaraj and Banerjee, 2004). Concurrently, biochemical experiments validated the results of these screens and explored the molecular mechanisms underlying the observed genetic interactions. Thus, genetically defined hierarchies were translated into a molecular signal transduction cascade connecting RTKs to the activation of mitogen-activated protein kinase (MAPK; Figure 1.1, A; reviewed in Porter and Vaillancourt, 1998; Schlessinger, 2000). These efforts collectively demonstrated that RTKs signal through an evolutionarily conserved biochemical pathway that is required during development and that includes several proteins previously implicated in growth and oncogenesis. The RTK studies set the stage both conceptually and experimentally for subsequent studies of cell–cell signaling. Similar approaches have subsequently identified and characterized the components of several pathways that are activated by cell–cell contact and/or secreted molecules, and mutant phenotypes in model systems have revealed their developmental roles. From these studies, we now know that major developmental signaling pathways such as

2

Principles of Developmental Genetics

© 2007, Elsevier Inc. All rights reserved.

INTRODUCTION

3

FIGURE 1.1 Basic overview of the major cell–cell signaling pathways discussed in this chapter. A, Receptor tyrosine kinases (RTK): Extracellular ligand (red) binds the receptor and, often by facilitating dimerization, induces the activation of cytoplasmic kinase domains. The receptors then autophosphorylate on several tyrosine residues, generating docking sites for effector proteins (yellow). Effector proteins initiate various signal transduction pathways when engaged by the receptors at the plasma membrane. Three pathways commonly activated by RTKs are shown, although members of the superfamily differ in their effector pathway usage and regulation. B, Hedgehog (Hh): In the absence of ligand, the Hh receptor Patched (Ptc) inhibits the Smo-initiated signaling pathway. In this state, Costal-2 (Cos2), Fused (Fu), and Gli/Ci form a complex, and Gli/Ci is preferentially proteolyzed to a repressive form (GliR) that translocates to the nucleus and blocks transcription. When Hh binds Ptc, Smo inhibition is relieved. Smo localizes to cilia (vertebrates) or clusters at the plasma membrane (invertebrates), is phosphorylated, and binds the Cos2–Fu complex. This releases Gli/Ci, which is then preferentially processed to a different product, GliA, that enters the nucleus and activates target gene transcription. C, TGFb/BMP: Ligand binding to the heterotetrameric Activin receptor induces the type II subunits of this complex to serine/threonine phosphorylate the type I subunits, which then phosphorylate associated Receptor-Smads (R-Smads). SARA facilitates the interaction between R-Smad and the receptor. Phosphorylation of R-Smads increases their affinity for co-Smads and decreases their affinity for SARA, which is then released. Heterotrimers (R-Smad, co-Smad) or homotrimers (R-Smad) then form and translocate to the nucleus, where they regulate the transcription of target genes with help from DNA-binding cofactors and transcriptional coactivators or corepressors. D, Wnt/Wingless (Wg)–b-catenin pathway: In the absence of ligand, a destruction complex comprised of GSK3, Axin, APC, and other proteins (not shown) binds and phosphorylates b-catenin, targeting it for ubiquitin-mediated proteosomal degradation. When Wnt binds the Frizzled receptor, Axin is engaged by the coreceptors LRP-5/6, Dishevelled (Dsh) is activated, and the destruction complex no longer phosphorylates b-catenin. b-catenin is released and enters the nucleus, where it activates target gene transcription together with TCF/LEF proteins. (A more complete diagram of this pathway can be found on the Wnt home page: http://www.stanford.edu/rnusse/wntwindow.html). E, Notch signaling: In the absence of ligand binding, CSL transcription factors (CBF-1, Suppressor of Hairless, LAG-1) interact with a corepressor complex and inhibit the transcription of Notch target genes. The Notch receptor can be activated either by interaction with ligands (Delta, Delta-like, Serrate, Jagged) or the internalization of ligand into adjacent cells. Notch activation induces two cleavage events: TNFa converting enzyme (TACE) sheds the ectodomain, and g-secretase releases the Notch intracellular domain (NICD) into the cytoplasm. NICD translocates to the nucleus, recruits a coactivator complex, and displaces the corepressor complex. The NICD complex then activates target gene transcription. (See color insert.) (Continued)

4

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

(Continued)

INTRODUCTION

5

6

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

Hedgehog (Hh), transforming growth factor b/bone morphogenetic proteins (TGFb/BMP), Wnt/Wingless, and Notch (Figure 1.1) have been evolutionarily conserved and are used reiteratively during embryonic development to instruct cell behavior and fate and to coordinate tissue growth and patterning. Mutations that disrupt these pathways are associated with developmental and proliferative disorders that, in humans, include neurocristopathies and numerous forms of cancer. In recent years, developmental signaling studies have begun to illuminate the mechanisms by which different pathways promote, antagonize, and/or synergize with one another in responding cells. These studies have been aided substantially by the development of increasingly sophisticated tools for genome-wide analysis and genetic manipulation. Genomic sequence data and high throughput assays have enabled classic genetic and biochemical screens to be performed and analyzed more efficiently. Importantly, by providing platforms for systematic genome-wide analysis, these technological advances have enabled screens to be less biased toward known genes and less dependent on specific phenotypic outcomes. However, genetic approaches remain essential complements to genomic studies because they are instrumental in addressing questions of mechanism and consequence (i.e., how specific proteins and interactions contribute to signaling and development). In this chapter, we will discuss progress in four areas of developmental signaling: (1) the identification and characterization of novel signaling pathway components; (2) the distribution of ligand and the localization of signal transduction; (3) the mechanisms of signal transduction; and (4) the transcriptional targets of cell signaling events. As a result of space limitations, we are unable to provide a complete and comprehensive review of recent literature for any single signaling pathway (we refer interested readers to recent reviews for in-depth discussions of individual pathways: Schlessinger, 2000; Shi and Massague´, 2003; Kadesch, 2004; Lum and Beachy, 2004; Huangfu and Anderson, 2005b; Nusse, 2005). Instead, we focus on studies that illustrate novel conceptual advances and/or have used new approaches to address longstanding questions in the field.

I. IDENTIFICATION OF NOVEL SIGNALING PATHWAY COMPONENTS A. Phenotype-Driven Screens In Vivo Phenotype-driven screens, such as those used in early RTK studies, continue to prove invaluable for the identification of novel pathway components and modifiers in multiple developmental systems. The availability of annotated genome sequence data has made it possible to monitor the saturation of these screens, which are aided by collaborative efforts currently underway to mutate all coding genes in mouse, fly, and worm using various methods (see the databases and Web resources listed in Table 1.1). Screens in the different model systems have complemented one another and generated data sets that are overlapping but not identical. This may be in part because of the sensitivity of the phenotypes scored and the types of mutation (e.g., loss-of-function versus hypomorphic alleles) introduced in each system. However, studies of the Hh signaling pathway have suggested that some species-specific mechanisms are used to transduce cell–cell signals.

7

IDENTIFICATION OF NOVEL SIGNALING PATHWAY COMPONENTS

TABLE 1.1

Web-Based Resources for the Developmental Signaling Community

Web Site

Organisms

Contents

Model Organisms for Biomedical Research (NIH) http://www.nih.gov/science/ models/ Database of Interacting Proteins http://dip.doe-mbi.ucla.edu/

Fly, worm, mouse, fish, frog, and others

Links to Web resources for researchers, arranged by organism (including many sites listed below) Physical interactions

Enhancer Element Locator http://www.cs.helsinki.fi/u/ kpalin/EEL/

Human, fish, mouse, rat, and dog

Enhancer elements (transcription factor binding sites) predicted on the basis of the comparative analysis of vertebrate genomes

FlyBase http://flybase.bio.indiana.edu/

Fly

PIMRiderW (Hybrigenics’ functional proteomic software) http://pim.hybrigenics.com/ pimriderext/common/

Fly

Expression patterns, genetic interactions, gain- and lossof-function mutation phenotypes, available alleles Protein interaction mapping including genome-wide interactome data for Drosophila and a TGFb-specific interactome data set

WormBase http://www.wormbase.org/

Worm

Worm Interactome http://vidal.dfci.harvard.edu/ interactomedb/i-View/ interactomeCurrent.pl

Worm

Predictions of C. elegans Genetic Interactions (v 140) http://tenaya.caltech. edu:8000/ predict

Worm

Predicted functional interactions between C. elegans genes based on expression data, genetic and physical interactions in yeast, fly, and worm

International Gene Trap Consortium http://www.genetrap.org Mouse ENU databases The Sloan-Kettering Mouse Project http://mouse.ski.mskcc.org/ mutant/mutantBase.php Baylor College of Medicine Mouse Genome Project: Mouse Mutagenesis for Development Defects http://www.mouse-genome. bcm.tmc.edu/ENU/ ENUMutantSources.asp

Mouse

Available gene trap lines, expression data

Mouse

Available mutants, phenotype data

GenePaint http://www.genepaint.org/ Frameset.html

Mouse

Atlas of developmental expression patterns

Fly, worm, and mouse

Developmental expression patterns, loss-of-function mutation phenotypes, available alleles Protein–protein interaction database

Mouse (Continued)

8

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

Web Site

Organisms

Contents Atlas of developmental expression patterns

Edinburgh Mouse Atlas Project: EMAGE Gene Expression Database http://genex.hgu.mrc.ac.uk/ Emage/database/ emageIntro. html The Jackson Laboratory: Mouse Genome Informatics http://www.informatics.jax. org/menus/allsearch_menu. shtml

Mouse

Available mutant lines/alleles and phenotype data

The Zebrafish Information Network http://zfin.org/cgi-bin/ webdriver?MIval¼aaZDB_home.apg Zebrafish Enhancer TRAP lines database (ZETRAP) (Parinov et al., 2004; Choo et al., 2006) http://plover. imcb.a-star.edu.sg/zetrap/ ZETRAP.htm

Fish

Available mutant and transgenic lines, phenotype data, genetic maps, developmental gene expression patterns Available GFP lines (and patterns of GFP expression in each) generated by transposon-mediated enhancer trapping

ZF-MODELS: Zebrafish Models for Human Development and Disease http://www.zf-models.org/ data/databases.html

Fish

Fish

Microarray (expression profiling) data, developmental expression patterns, loss-offunction phenotypes, available GFP lines

Key components of the Hh pathway are highly conserved (reviewed in Lum and Beachy, 2004; Huangfu and Anderson, 2005b). Hh signal transduction is controlled by the actions of the Patched (Ptc) receptor and the sevenpass transmembrane protein, Smoothened (Smo). In the absence of Hh, Ptc inhibits Smo from transducing signals. Hh binding to Ptc relieves this inhibition and enables Smo to activate a cytoplasmic signal transduction pathway that culminates in the proteolysis and nuclear translocation of an activating transcription factor (known as Gli in vertebrates and Ci in Drosophila). In mice, ENU mutagenesis screens identified cilia and intraflagellar transport proteins as essential components of the Hh pathway that act downstream of Ptc and Smo (Huangfu et al., 2003; Huangfu and Anderson, 2005a). Functional studies have demonstrated that activation of the Hh pathway in vertebrates induces the localization of Smo, Gli2 and Gli3, and other relevant proteins to cilia; a cilia localization motif on Smo is essential for normal Hh responses in cultured cells and zebrafish (Corbit et al., 2005; Haycraft et al., 2005). By contrast, intraflagellar transport mutations do not cause Hh-like phenotypes in Drosophila, and, in this organism, Hh-responsive cells do not have cilia (Ray et al., 1999; Han et al., 2003; Avidor-Reiss et al., 2004). Drosophila Smo accumulates at the plasma membrane upon Hh stimulation, whereas the vertebrate ortholog gets internalized. Furthermore, mammalian and fly Smo proteins are phosphorylated on different residues in response to Hh. Phosphorylation is required for the internalization of mammalian Smo

IDENTIFICATION OF NOVEL SIGNALING PATHWAY COMPONENTS

9

and for downstream signal transduction (Denef et al., 2000; Zhu et al., 2003; Chen et al., 2004a; Zhang et al., 2004; Apionishev et al., 2005). Likely as a result of such differences, antagonists of Hh signaling have disparate effects in flies and mice (Incardona et al., 1998; Taipale et al., 2000; Chen et al., 2002). These studies pose the challenge of discriminating evolutionarily conserved mechanisms from species-specific mechanisms of cell–cell signaling. B. Systems Biology Approaches to the Identification of Signaling Pathway Components Random, phenotype-driven mutagenesis screens are now being supplemented with sequence-driven genome-wide screens that do not rely on chance to reach saturation. These new approaches provide several advantages over classical techniques. Importantly, they are not reliant on phenotypic output, and so are capable of identifying genes that contribute to multiple cellular processes or pathways. These genes would likely have pleiotropic mutant phenotypes and therefore be discarded in screens for pathway-specific phenotypes. In addition, genome-wide screens can identify factors that have an impact on cell–cell signaling but are not essential for a normal developmental outcome (e.g., because of redundant or compensatory pathways). Three types of genome-wide screens that have been used in signaling studies include in vitro RNA interference (RNAi) screens, protein interaction mapping (genome-wide yeast two hybrid [Y2H]), and developmental synexpression analysis. None of these approaches in isolation is sufficient to define signaling pathways and the requirements for individual components or interactions in vivo. However, each provides a platform for comprehensively scanning the genome and generating new models of cell–cell signaling. 1. RNAi Screens in Cultured Cells RNAi uses short, double-stranded RNAs to trigger the degradation of target mRNAs species. This was developed as an experimental tool for work with C. elegans, in which it is now widely used for loss-of-function studies and phenotype-driven screens (Fire et al., 1998; Wang and Barr, 2005). Recently, genome-wide screens have been developed that use RNAi in Drosophila embryonic imaginal disc cell cultures (clone-8 cells) to identify novel signaling pathway components (Lum et al., 2003). In these screens, clone8 cells are cotransfected with a pathway-responsive luciferase reporter and a comprehensive library of RNAi constructs. The products of known Drosophila coding genes are systematically tested for their ability to affect signaling pathway output as assayed by reporter activity. The original clone-8 screens used Hh-responsive transcriptional reporters. RNAi of known Hh pathway genes altered luciferase activity in this system, validating the approach. In addition, numerous genes previously unassociated with Hh signaling were found to modify Hh reporter activity and to interact genetically with known Hh pathway members (Lum et al., 2003; Nybakken et al., 2005). Some of these genes belong to classes traditionally associated with cell–cell signaling; these include a heparan sulfate proteoglycan (Dally-like, which was previously implicated in Wnt signaling), a homeodomain gene, three kinases (CK1a, Pitslre1, and Cdk9), and a phosphatase (PP2A). Interestingly, the screens also indicated that the Hh pathway is affected by factors involved in more general cellular processes, including ribosome and proteosome function, RNA regulation and splicing, and vesicle

10

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

trafficking. Although the disruption of such genes would likely cause pleiotropic phenotypes in vivo, several lines of evidence suggest that they are bona fide components or modifiers of the Hh pathway. They were independently identified in two clone-8 RNAi screens, although not all genes required for splicing, transcription, and so on altered Hh reporter activity in these experiments. Furthermore, such genes have been identified (albeit at a low frequency) in vivo in screens that rely on hypomorphic alleles and/or clonal analysis (Eggenschwiler et al., 2001; Huangfu et al., 2003; Collins and Cohen, 2005; Huangfu and Anderson, 2005a). The results of clone-8 RNAi screens greatly expand the known landscape of Hh signaling. Further studies are now needed to determine how the novel Hh modifier genes fit into current models of the signaling pathway. Similar RNAi screens with different transcriptional reporters have been used to scan the genome for genes that impact JAK/STAT and Wnt signaling. Like the Hh studies, these screens also identified proteins used in other signaling pathways as well as factors involved in general cellular processes (Baeg et al., 2005; DasGupta et al., 2005; Mu¨ller et al., 2005). Parallel screens in this system may prove useful for identifying points of crosstalk between pathways. 2. Interactome Mapping Tewari et al. recently used Y2H assays in a genome-wide screen for C. elegans proteins that interact physically with known members of the TGFb pathway (the basic pathway is diagrammed in Figure 1.1, C). They thus generated an “interactome” map describing physical interactions among 59 proteins, only four of which had previously been assigned to the TGFb signaling pathway. Novel components of this biochemically defined interactome were then analyzed in vivo expression studies determined whether they are expressed in TGFb-dependent contexts, and double RNAi experiments identified genetic interactions with previously known TGFb pathway genes. Thus, several new proteins were modeled into the TGFb signaling network, including filamin, the TTX-1 homeobox protein, Swi/Snf chromatin remodeling factors, and Hsp90 (Tewari et al., 2004). Additional biochemical and functional studies are needed to characterize the roles of the interactome components in TGFb signal transduction and development. An important feature of interactome mapping is that it is not hindered by compensatory mechanisms that may mask roles of pathway members in other assays. In addition, novel components identified using this approach can be directly modeled into known signal transduction pathways on the basis of physical and genetic interactions. Genome-wide Y2H analyses have now been reported for C. elegans and Drosophila, and protein–protein interaction data for multiple systems have been compiled into an interactive public database (Table 1.1, Database of interacting proteins; Xenarios et al., 2002; Giot et al., 2003; Li et al., 2004a; Formstecher et al., 2005). Thus, interactome mapping can now be done to some degree in silico as a starting point or modeling tool for signaling studies. The selection of different bait proteins in future Y2H screens will continue to enrich pathway-specific data sets. However, a challenge for future studies is to develop methods for mapping physical interactions in cell systems that are more representative of biological contexts. Phosphorylation events, which are known to figure prominently into signal transduction, are not recapitulated in yeast. Although phosphomimetic amino acids can be substituted into bait proteins

IDENTIFICATION OF NOVEL SIGNALING PATHWAY COMPONENTS

11

for Y2H studies, the results of these studies are limited to proteins that do not require phosphorylation for the assayed interaction. Furthermore, interactome components are likely to be cell type specific; comparative studies in different cell systems may illuminate context-specific mechanisms of signal transduction. 3. Identification of Synexpression Groups Signaling networks have increased in complexity during evolution as a result of gene duplication events and the incorporation of redundant or compensatory signaling events. Many proteins in these networks have modular and conserved protein interaction domains (e.g., phosphotyrosine-binding domains, src homology domains) that are fairly promiscuous in biochemical assays. Furthermore, in vivo analyses have indicated that many pathways use context-specific mechanisms of signal transduction during development (discussed in section III). It has therefore become a significant challenge to determine which proteins are functionally associated in distinct biological contexts. Developmental synexpression analysis has proven useful for generating models of ligand-receptor relationships, signal transduction pathways, and regulatory events that comprise signaling modules in vivo. Genome-wide expression screens performed predominantly in zebrafish led to the identification of an evolutionarily conserved Fgf8 synexpression group that contains several regulators of the RTK-Ras-MAPK pathway, namely Sprouty proteins, the transmembrane protein Synexpressed with FGF (Sef), and MAPK phosphatase 3 (Mkp3; Kudoh et al., 2001; Fu¨rthauer et al., 2002; Tsang et al., 2002; Kawakami et al., 2003; Tsang et al., 2004). As might be expected for antagonists of a broadly used signal transduction cascade, these proteins do not exhibit strict RTK specificity in biochemical assays (Camps et al., 1998; Reich et al., 1999; Tsang et al., 2002; Kovalenko et al., 2003; Preger et al., 2004; Torii et al., 2004). However, synexpression suggests that they are required in Fgf8-expressing tissues, and functional studies have indicated that they antagonize Fgf signaling in vivo (Kramer et al., 1999; Fu¨rthauer et al., 2002; Tsang et al., 2002; Kawakami et al., 2003). This does not preclude the possibility that they inhibit signaling by other RTKs at sites of Fgf8 expression. Indeed, in Drosophila, Sprouty and Mkp3 also regulate Egfr signals, and Mkp3 is expressed in contexts that are dependent on multiple RTKs (Kramer et al., 1999; Kim et al., 2004; Go´mez et al., 2005). Genetic interaction studies are needed to determine the targets of Sef, Mkp3, and Sprouty regulation in vertebrates. The functions of some signaling pathways are conserved across species: for example, the Fgf/Fgfr pathway is required for branching morphogenesis during lung and trachea development in mammals and flies, respectively (Reichman-Fried et al., 1994; Sutherland et al., 1996; Min et al., 1998; Sekine et al., 1999). Conservation of expression patterns across species is highly suggestive of functional conservation, and so expression profiling in different model organisms can help to identify gene functions. One member of the Fgf8 synexpression group in both planaria and vertebrates is the secreted Fgfr-like protein Isthmin (also known as nou-darake, Fgfr-Like 1). The roles of this protein in vertebrates have been elusive in loss-of-function studies, perhaps as a result of compensatory or redundant regulatory pathways (Cebria et al., 2002; Pera et al., 2002). However, loss of isthmin/nou-darake in planaria results in an expansion of anterior neural tissues during regeneration; this is suppressed by the simultaneous silencing of Fgfr1 and Fgfr2 (Pera et al., 2002).

12

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

These results implicate isthmin as an Fgf antagonist that restricts neural proliferation and/or fate, and they suggest that the vertebrate ortholog may have similar roles in restricting Fgf signals during neural stem cell and/or anterior central nervous system development. The integration of synexpression data with biochemical and loss-of-function data could notably expedite future studies of developmental cell–cell signaling. Several Web-based resources that detail developmental expression patterns are currently available to the community (see Table 1.1). These data can provide clues as to the context-specific usage of signaling proteins and thus help to refine models of in vivo signal transduction.

II. DISTRIBUTION/LOCALIZATION OF LIGAND AND SIGNAL TRANSDUCTION Many components and/or modifiers of signaling pathways function within cells or the extracellular space to ensure the proper localization of signals and their biochemical responses. Heparan sulfate proteoglycans (HSPGs) contribute significantly to this aspect of cell signaling by modulating the distribution and/or activity of Wnt, TGFb/BMP, Fgf, and Hh proteins. A number of studies have addressed the roles of HSPG core proteins and synthesis or modifying enzymes in developmental cell signaling. In the mouse, an ENU-induced mutation in UDP-glucose dehydrogenase (Ugdh, a glycosaminoglycan synthesis factor) was found to cause recessive mesodermal phenotypes reminiscent of Fgf8 and Fgfr1 null embryos (Garcı´a-Garcı´a and Anderson, 2003). Similarly, mutations in Ugdh (sugarless) and other HSPG synthesis and processing enzymes disrupt Fgf-dependent development in Drosophila (Lin et al., 1999). Biochemical studies have demonstrated that heparan sulfate is essential for high-affinity Fgf–Fgfr binding and that Fgfs and Fgfrs have distinct affinities for different types of HSPGs (Ornitz, 2000; Mohammadi et al., 2005). Additional roles of heparan sulfates have been identified in Drosophila imaginal wing disc studies. In this context, HSPGs including Dally and Dally-like are required for long-range Hh signaling, cell surface accumulation and tissue distribution of Wnt and Hh, and stability or transport of Decapentaplegic (Dpp, a Drosophila BMP ortholog) as it travels across the wing disc epithelium (reviewed in Ha¨cker et al., 2005). Posttranscriptional and/or posttranslational modifications of ligands can also restrict movement within a tissue and thus enhance local signaling. For example, the diffusion of some mammalian Pdgf/Vegf ligands is regulated by alternative splicing of a “retention signal” motif, which is a C-terminal stretch of positively charged residues that can keep these ligands associated with producing cells (Eriksson and Alitalo, 1999; Heldin and Westermark, 1999). The Pdgfb retention motif is essential in vivo for its local actions: genetic ablation of the motif in Pdgfbret/retmice leads to defects in pericyte number, vascular remodeling, and the association of Pdgfrb-expressing pericytes with the Pdgfb-expressing vascular endothelium (Lindblom et al., 2003). However, the phenotypes of these mice are less severe than those of Pdgfb / and Pdgfrb / mice; this suggests that some roles of Pdgfb do not require local retention (Leve´en et al., 1994; Soriano, 1994; Lindblom et al., 2003). Intrinsic ligand structure and posttranslational modifications, such as lipid conjugation, tether some signaling proteins to cell membranes. Pathways leading to the synthesis, conjugation, and release/cleavage of membrane-associated moieties likely have an impact on the activities of these signals, which

DISTRIBUTION/LOCALIZATION OF LIGAND AND SIGNAL TRANSDUCTION

13

include Notch ligands and ephrins. The ephrins are ligands that are associated with plasma membranes by GPI groups (A class) or transmembrane domains (B class). This restricts their activities to signaling between adjacent cells, and it enables them to function both in “forward” signaling to their cognate receptors as well as in “reverse” signaling in cells in which they are expressed (reviewed in Davy and Soriano, 2005; see Chapter 21). Genetic studies have demonstrated that reverse signaling via a PDZ interaction domain is essential for a subset of ephrin B1 roles during mouse embryogenesis (Davy et al., 2004). Membrane tethering of other types of ligand may similarly facilitate reverse signaling either directly or through associated proteins. Mechanisms of localizing signaling proteins and their responses have been extensively studied in Drosophila imaginal wing discs, where secreted morphogens form gradients that induce different fates at different activity thresholds. Several models have been proposed to explain how secreted signaling molecules form gradients and reach target cells several cell diameters away from their sites of origin. According to one model, morphogens diffuse through the extracellular space; local concentration and activity are determined by factors that modulate ligand secretion, diffusion, stability, and receptor-mediated uptake. In the case of BMP signaling, ligand diffusion and stability are notably affected by auxiliary factors, including Short gastrulation (Sog), Twisted gastrulation (Tsg), and Tolloid (Tld; reviewed in O’Connor et al., 2006). Tsg facilitates Sog/Dpp binding in a trimeric complex that enables Sog, which is a Dpp antagonist, to keep the ligand inactive for extracellular transport across a tissue. At target sites, the protease Tld cleaves Sog, releasing Dpp to act locally. Combining mathematical modeling with experimental genetics, Mizutani et al. (2005) demonstrated that a diffusion model incorporating the effects of these proteins can recapitulate the BMP ligand gradient as well as the nonsynonymous BMP activity (phospho-Smad) gradient in Drosophila wing discs. Although some secreted signaling proteins may be distributed by extracellular diffusion, imaging studies in Drosophila imaginal wing discs have suggested that more active mechanisms also contribute to signal localization. In unfixed wing discs expressing GFP, “cytonemes” (thin, actin-based membrane extensions that are several cell diameters long) extend from the apical surface of wing disc cells toward sites of either Dpp or Wg expression (Figure 1.2, A; Ramirez-Weber and Kornberg, 1999). Cytoneme formation in wing disc epithelia is Dpp-dependent, and the extensions are polarized toward Dpp or Wg only in regions where these factors act as morphogens. Interestingly, the Dpp receptor Thickveins (Tkv) is expressed on and moves directionally within Dpp-oriented cytonemes (Hsiung et al., 2005). Together, these data suggest that long-range actions of Dpp are mediated, at least in part, by the extension of receptor-expressing cytonemes toward sites of Dpp production. Similar structures may also guide chemotaxis in some contexts: actin-based cytonemes that contain Breathless (Btl, an Fgfr) extend toward sources of Branchless (Bnl, an Fgf) during the third instar larval migration of Drosophila tracheoblasts (Sato and Kornberg, 2002). Imaging studies using a GFP-Dpp transgene led to a third model of Dpp localization in imaginal disc epithelia. Using this transgene, Teleman and Cohen directly visualized the ligand and found that it localizes to endocytic vesicles and is concentrated basally, whereas cytonemes protrude apically from wing disc cells. On the basis of these findings, the authors proposed that the Dpp gradient is formed via cycles of endocytosis and secretion that

14

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

FIGURE 1.2 Mechanisms of signaling protein transport observed in Drosophila. A, Tkv, the receptor for Dpp, is expressed at the tips of cytonemes, which are long filamentous protrusions that extend from apical cell surfaces in imaginal wing discs toward the source of Dpp patterning signals (representative individual cells are diagrammed in green). B, Transcytosis (i.e., repeated cycles of endocytosis and secretion) moves Dpp across the wing disc epithelium. Dpp-containing vesicles are concentrated basally. C, Btl/Bnl signaling is required for tracheoblast migration. During branching morphogenesis, Btl induces the formation of short, cytoplasmic extensions on Bnl-expressing cells at the tips of tracheal branches. The filopodia-like structures observed in this context contain both actin and microtubules, and are not polarized toward ligand. (See color insert.)

transport Dpp within cells across the epithelial sheet (Figure 1.2, B; Teleman and Cohen, 2000). The cytoneme and GFP-Dpp studies may highlight distinct aspects of Dpp gradient formation. In support of this, ligand (Dpp) and response (phospho-Smad) gradients in the wing disc differ from one another, indicating that Dpp signaling activity is modulated at or downstream of the receptor (Teleman and Cohen, 2000). Live imaging studies revealed that Bnl/Btl signaling induces another type of membrane extension in tracheal cells during branching morphogenesis. In this context, cells at the tips of tracheal branches extend numerous fine protrusions in response to Bnl (Figure 1.2, C). Unlike cytonemes, these filopodia-like structures contain both actin and microtubules, are relatively short, and are not polarized toward Bnl ligand (Ribeiro et al., 2002). It is not yet clear whether these structures are involved in Btl/Bnl signal transduction or localization. Many aspects of developmental signaling are highly conserved, and so it is likely that the mechanisms of signal relay observed in Drosophila are used in other model systems. However, experiments in vertebrates have not yet

MECHANISMS OF SIGNAL TRANSDUCTION

15

validated this hypothesis. In Xenopus, embryo cocultures were used to examine mechanisms of long-range signaling using a fluorescently tagged TGFb ligand (Xnr2). This ligand induced transcriptional responses at a distance from secreting cells, but no cytoneme-like extensions were observed that were of sufficient length to explain the range of ligand action. Furthermore, Xnr2 was not observed in vesicles, and its transport did not rely on endocytosis. Thus, the authors concluded that Xnr2 is distributed in Xenopus embryos by diffusion rather than by cytonemes, filopodia, argosomes (vesicular structures), or transcytosis (Williams et al., 2004). This may reflect differences in the experimental systems, ligand- or context-specific mechanisms of signal relay, or the ability of fixation and imaging techniques to capture and preserve delicate membranous or vesicular structures. Further knowledge of the composition, formation, and mechanisms of action of ligand/receptor transport structures will greatly facilitate future studies in different model systems.

III. MECHANISMS OF SIGNAL TRANSDUCTION Whereas loss-of-function alleles have revealed essential functions of many cell–cell signaling factors during development, more subtle and directed mutations are required to analyze signaling mechanisms. These mutations eliminate or alter specific protein–protein interactions and/or sites through which protein activity is regulated, and they are often designed after biochemical models. Recent in vivo studies using these types of alleles have begun to shed light on how functional specificity is achieved within protein families and how different signaling pathways intersect within a responding cell. A. Specificity of Signal Transduction by Related Receptors Two longstanding aims in the developmental signaling field have been to determine how closely related signals drive distinct responses in vivo and how individual receptors elicit context-specific responses over the course of development (discussed in Tan and Kim, 1999; Simon, 2000). Among related growth factor receptors, functional specificity could be achieved through differential utilization of and/or affinity for effector proteins; differences in the localization, duration, or amplitude of signal activation; or the context-specific availability of factors that modulate cellular responses. Several lines of evidence have demonstrated that, despite biochemical similarities observed in vitro, members of the RTK superfamily drive nonequivalent signals in vivo. In Drosophila, the signaling domains of Torso and DER (Drosophila Egfr) drive migration responses to Btl activation incompletely and to different degrees in chimeric receptor rescue experiments (Dossenbach et al., 2001). A molecular explanation for this was suggested by the recent finding that, during tracheal branching morphogenesis, Btl and DER differ in their requirements for the downstream transcriptional effector Pointed (despite common activation of the Ras-MAPK pathway; Cabernard and Affolter, 2005). Similarly, chimeric receptor experiments performed in the mouse have shown that RTKs have distinct developmental potentials and transduce non-equivalent signals. The Drosophila Torso RTK signaling domain incompletely rescues Pdgfra functions in vivo and activates only a subset of Pdgfra-activated transduction pathways in primary cells (Hamilton et al., 2003). Likewise, the Pdgfrb signaling domain is unable to drive Fgfr1

16

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

responses during embryonic development (Hoch, 2005). By contrast, the Fgfr1 signaling domain activates more potent signaling responses than Pdgfra or Torso, and Pdgfra/Fgfr1 chimeric receptor-expressing embryos exhibit dominant gain-of-function phenotypes (Hamilton et al., 2003). Pdgfr and Vegfr studies have demonstrated that even RTKs within subfamilies transduce distinctive signals. The Pdgfra signaling domain drives weaker MAPK responses than that of Pdgfrb in cultured embryonic cells. In addition, the Pdgfrb signaling domain can fully rescue Pdgfra-dependent development in vivo, whereas the converse is not true (Klinghoffer et al., 2001). The differential recruitment of effector proteins may contribute to the disparity in Pdgfr signaling potential. Pdgfra transduces signals predominantly via a single effector (PI3K) recruitment site during embryogenesis, despite its biochemical ability to engage proteins at additional sites (Klinghoffer et al., 2002). In contrast, multiple effector pathways contribute additively to Pdgfrb functions in mice, as has also been reported for Torso in Drosophila (Gayko et al., 1999; Tallquist et al., 2003). The selective use of one pathway may limit the amplitude and variability of Pdgfra responses, and may reflect the affinity or availability of effector proteins for this receptor. Within the Vegfr subfamily of RTKs, Vegfr2 is thought to be the principal activator of signal transduction. This isoform responds to ligand with heightened receptor kinase and MAPK activities as compared with Vegfr1. These two receptors are coexpressed in vivo, and chimeric receptor studies have shown that Vegfr1 serves to regulate the activity of Vegfr2. Interestingly, different Vegfr1 ligands specify distinct modes of Vegfr2 regulation (inhibition versus potentiation; Rahimi et al., 2000; Autiero et al., 2003; Meyer and Rahimi, 2003; Roberts et al., 2004). The functional specialization of Vegfrs has been attributed to an amino acid change in the activation loop of Vegfr1 at a residue that is highly conserved among other class III RTKs (Meyer et al., 2006). Within several RTK subfamilies, homo- and heterodimers can form in vitro, but the significance of this observation in vivo and the consequences for downstream signaling are not known. The Vegfr findings introduce the possibility that subunits within other heterodimers have distinct functions that differentiate the signals transduced by homo- versus heterodimers. Fgfr1 and Fgfr2 have been shown to signal through adaptors (Frs2,3) that are distinctive among RTK effectors in that they interact constitutively with these receptors instead of being recruited after ligand-dependent activation (Wang et al., 1996; Kouhara et al., 1997; Xu et al., 1998). Biochemical studies implicated Frs adaptors in MAPK and PI3K signaling downstream of Fgfrs (Wang et al., 1996; Xu et al., 1998; Hadari et al., 2001). However, the Fgfr1–Frs interaction is required only for a subset of Fgfr1 functions during mouse embryogenesis (Hoch and Soriano, 2006). Furthermore, in primary embryonic cells, this signaling event affects basal Fgfr2 activity but is not essential for MAPK activation responses to Fgf (Hoch and Soriano, 2006). Recently, Frs adaptors have been implicated in crosstalk and feedback regulation among Fgfrs and other RTKs. Activated Frs2 can recruit Cbl and instigate the ubiquitin-mediated degradation of Frs2 and Fgfrs (Wong et al., 2002). Frs2 is also threonine phosphorylated in response to Fgfs and other RTK-mediated signals; this inhibits Frs-mediated signaling to the MAPK and PI3K pathways (Lax et al., 2002). Finally, SHP2 and Src, which can both be activated downstream of Frs2, modulate the tyrosine phosphorylation of Sprouty proteins, which could impact signaling by several RTKs (Hanafusa et al., 2002; Fong et al., 2003; Hanafusa et al., 2004;

MECHANISMS OF SIGNAL TRANSDUCTION

17

Li et al., 2004b; Jarvis et al., 2006). Further studies are needed to assess the contribution of these regulatory events to Frs functions in vivo. The uniquely constitutive association of Frs adaptors with Fgfr1 and Fgfr2 may confer preferential regulation of or sensitivity to Frs-mediated feedback regulation. B. Crosstalk Between Signaling Pathways Occurring in the Cytoplasm We are only beginning to understand the molecular mechanisms by which signaling pathways interact, although crosstalk has long been suggested by the results of tissue explant and recombination experiments. One mechanism of crosstalk that has been identified in BMP/TGFb studies involves the combinatorial control of pathway intermediates. BMP/TGFb family members signal through a small number of receptors that phosphorylate C termini of Smad proteins, thus activating these effectors to form trimers, translocate into the nucleus, and regulate transcription (Figure 1.1, C). MAPK (Erk, Jnk, p38) antagonizes this pathway by phosphorylating Smads at residues in their linker domains (Figure 1.3, A). This inhibits Smad nuclear translocation thereby blocking transcriptional responses to BMP/TGFb, and can also target Smads for ubiquitin-mediated degradation (reviewed in Massague´, 2003; Sapkota et al., 2007). BMP and TGFb signals can also induce phosphorylation of the Smad linker domain, but this event is delayed compared to the C terminal phosphorylation and does not disrupt nuclear signaling (Sapkota et al., 2007). The roles of phosphatases in cell signaling are generally understudied as compared with kinases, but these two classes of enzymes are of equal importance in the regulation of signal transduction pathways. An RNAi screen in Drosophila S2 cells identified pyruvate dehydrogenase phosphatase as a phosphatase for BMP/TGFb-responsive sites on the fly Smad ortholog (MAD) and the mammalian Smad1 (Chen et al., 2006). Similar screens could conceivably identify additional Smad kinases and phosphatases through which other pathways impinge on Smad phosphorylation and localization. The combinatorial control of Smads may enable the relatively simple BMP/TGFb pathway to drive different cellular responses depending on the availability of other signals. Recent Xenopus studies suggest that the MAPK-mediated antagonism of BMP signaling underlies neural fate induction in the early vertebrate embryo (Pera et al., 2003; Kuroda et al., 2005). Similarly, this crosstalk may redirect TGFb/BMP responses in other developmental contexts receiving concomitant RTK-mediated signals. For example, Fgf signaling and BMP antagonism have been implicated in neural crest induction (LaBonne and Bronner-Fraser, 1998; Steventon et al., 2005; Wawersik et al., 2005). Additionally, in the limb bud, p38MAPK signaling is essential for some responses to BMP (Zuzarte-Luis et al., 2004). An analysis of Smad1 phosphorylation site mutants revealed that MAPK-responsive residues on the Smad1 linker domain are essential only in select contexts in mice; these include the development of the reproductive tract and germ cells and postnatal digestive tract homeostasis. By contrast, C terminal residues phosphorylated downstream of BMP signals are required broadly for Smad1-dependent development (Aubin et al., 2004). The discrepancy between the Xenopus and mouse results may reflect species–specific roles of Smad phosphorylation, or, alternatively, may be caused by the activities of other Smad isoforms in the two experimental systems. Future studies are needed to distinguish between these models and to determine the developmental requirements for crosstalk mediated by other Smad proteins.

18

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

Mechanisms of crosstalk between signaling pathways. A, Convergence of pathways on common intermediates: BMP/TGFb signaling leads to the C-terminal phosphorylation of Smad proteins, thereby promoting their trimerization and nuclear translocation. MAPK can phosphorylate serine residues in linker regions between the Mad homology domains (MH1, MH2) . This inhibits nuclear translocation of Smads and thus blocks transcriptional responses to BMP signaling. B, Transcriptional induction of proteins that modify cell signaling: In C. elegans vulval precursor cells (VC), antagonistic interactions between the Notch and Egfr pathways cause the primary and secondary VCs to be differentially responsive to these pathways. Egfr signaling (yellow) in the presumptive primary VC induces the transcription-dependent internalization and degradation of Notch, thus activating lateral inhibition signaling by Notch ligands (Delta, Serrate, LAG-1 [DSL]). This activates the Notch pathway in the secondary VC, which culminates in transcription of MAPK antagonists that block Egfr signal transduction. C, Combinatorial control of transcription: In Drosophila eye cone cells, expression of Pax2 requires transcription factors activated by Notch (Suppressor of Hairless [Su(H)]) and Egfr signaling (Pointed [Ptd], Yan) as well as a regional transcription factor, Lozenge (Lz). Each of these transcription factors binds a distinct site in a Pax2 enhancer element.

FIGURE 1.3

IV. TRANSCRIPTIONAL TARGETS OF SIGNALING PATHWAYS A. Crosstalk Between Pathways Mediated by Transcriptional Regulation Multiple mechanisms of crosstalk have been identified that involve transcriptional regulation. Transcriptional profiling studies have indicated that cell–cell signaling events commonly induce the expression of signaling and regulatory proteins that alter the responding cell’s interactions with its environment. This

TRANSCRIPTIONAL TARGETS OF SIGNALING PATHWAYS

19

form of feedback regulation has been shown to modulate signaling both within and across pathways. During C. elegans vulva induction, Egfr and Notch signaling induce the transcription and/or activation of factors that establish reciprocal responsiveness to these pathways in neighboring cells (Figure 1.3, B). Egfr activation induces the internalization and degradation of Notch in the primary vulva cell. This transcription-dependent event enables Notch ligands to activate the receptor on neighboring cells and thus initiates lateral inhibition signaling (Shaye and Greenwald, 2002, 2005). Then, in secondary vulva cell precursors, Notch signaling induces the transcription of several MAPK pathway antagonists, thus inhibiting Egfr–Ras–MAPK signal transduction (Yoo et al., 2004). In Drosophila eye studies, two additional mechanisms of crosstalk between Egfr and Notch signaling have been elucidated. In this context, proteins from the two pathways converge to coordinately regulate target gene expression. In one set of studies, Groucho, a transcriptional corepressor that acts downstream of Notch (and Wnt), was found to be a point of crosstalk with the Egfr–MAPK pathway: MAPK can phosphorylate Groucho and thus weaken its corepressor activity. In this way, Egfr signaling can derepress the transcription of Notch target genes. These results provided a mechanistic model to explain the previous observation that Groucho interacts genetically with both Notch and Egfr (Price et al., 1997; Hasson et al., 2005). In cone cells of Drosophila eyes, transcription factors activated by Notch and Egfr signaling converge on an enhancer element to regulate Pax2 expression. A regional transcription factor, Lozenge, also binds this enhancer, and the transcriptional response requires the occupancy of all three sites (i.e., coregulation by factors activated by the Notch and Egfr pathways as well as the context-specific factor; Figure 1.3, C). Prospero is also coordinately but distinctly regulated by transcription factors downstream of the Egfr and Sevenless RTK pathways in Drosophila (Xu et al., 2000). There is evidence that this mode of crosstalk has been conserved in vertebrates: the Sox2 and Cdx3 genes in Xenopus are coordinately regulated by Wnt and Fgf signals (Haremaki et al., 2003; Takemoto et al., 2006). Recently, Hallikas et al. (2006) devised a computational tool to identify transcription factor binding sites, and they used it to scan the vertebrate genome in silico for targets of RTK, Hh, and Wnt signaling. Several putative targets were identified and validated through subsequent expression studies and cross-referencing with published work. Interestingly, this analysis indicated that there is significant overlap in the targets of Tcf (Wnt) and Gli (Hh) transcriptional regulation. The combinatorial control of enhancers may thus be a common means of crosstalk between these two pathways. These and other results were compiled in a searchable database of predicted enhancer elements for vertebrate genes (Enhancer Element Locator in Table 1.1). B. Transcriptional Profiling of Signaling Responses Many studies have characterized the transcriptional responses to signaling events since array technology was developed. For example, Fambrough et al. (1999) addressed the question of signaling specificity by comparing the transcriptional responses downstream of RTKs in cultured cells. Kit, Pdgfrb and Fgfr1 were found to induce the transcription of the same set of genes in this system (with some quantitative differences), whereas Egfr

20

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

induced transcriptional responses that differed both qualitatively and quantitatively from these other RTKs. The disruption of effector binding sites on Pdgfrb did not significantly affect its transcriptional response in these experiments, consistent with what was subsequently observed in vivo (Tallquist et al., 2003). Biologically relevant transcriptional profiling relies on the selection of informative tissue or cell samples. In a recent screen for Wnt target genes, comparative expression analysis was performed using gastrulation-stage wild-type and b-catenin mutant mouse embryos. In addition to known Wnt target genes, several novel targets were identified in this study, including components of other signaling pathways (e.g., Notch) and genes expressed in domains of Wnt reporter activity during gastrulation. Some target genes (Grsf1, Fragilis) were further validated as Wnt-associated genes in vivo: embryos derived from RNAi knockdown embryonic stem cells recapitulated aspects of Wnt mutant phenotypes (Lickert et al., 2005). In whole-embryo analyses, it is difficult to discern direct targets of signaling pathways from transcriptional changes that are secondary to developmental aberrations. Furthermore, different cell types and developmental contexts may respond to signals with distinct responses. For these reasons, profiling experiments would ideally use homogeneous cell populations that have not been immortalized or otherwise modified from their native state. High-fidelity cDNA amplification techniques are being developed to enable the profiling of single cells and small cell populations. This and similar technical advances will enable researchers to identify the transcriptional targets of signaling events in spatially or temporally restricted niches within developing embryos. The results of profiling studies need to be substantiated in functional assays that demonstrate the significance of identified targets in mediating relevant cellular responses. To facilitate the transition from expression analysis to functional validation, Chen et al. (2004b) generated a microarray of cDNAs representing genes that were randomly mutated by retroviral gene trapping in ES cells. This chip—or gene trap array—can be used to profile transcriptional changes in wild-type versus mutant cells/tissues, uninduced versus induced cells, or cells at different stages of differentiation. Mutant mice can then be generated from archived mutant ES cells for the analysis of putative target genes in vivo. In an initial study, the gene trap array was used to assess transcriptional responses of mouse embryonic cells to Pdgfra versus Pdgfrb stimulation (Chen et al., 2004b). The functions of several novel Pdgf target genes identified, and their genetic interactions with Pdgfrs were then addressed in vivo. Results of these studies implicated Pdgfs in the modulation of signaling by other secreted molecules (e.g., sphingosine) identified the transcriptional targets required for specific aspects of Pdgf-dependent development, and suggested novel postnatal roles of Pdgf signaling (Schmahl et al., 2007).

V. CHALLENGES FOR FUTURE STUDIES OF DEVELOPMENTAL SIGNALING We have highlighted four major areas of developmental signaling in which recent advances have been made using a combination of genetic and genomic tools. First, we discussed approaches that have been used to generate a global overview of factors that impact specific cell–cell signaling pathways. Next, we discussed the mechanisms underlying distinct steps in a signaling event, progressing from the secretion and transport of the signal to the signal

CHALLENGES FOR FUTURE STUDIES OF DEVELOPMENTAL SIGNALING

21

transduction events initiated by ligand-receptor binding. Finally, we discussed studies that address the outcome of cellular responses to environmental signals by examining transcriptional responses. Many key responses to cell–cell signaling (e.g., cell migration, adhesion, and cell cycle progression) may not require transcription; high-throughput assays need to be developed with which the molecular events underlying these responses can be explored. One major challenge for future studies is to further elucidate the mechanisms of cell–cell signaling in developing organisms. An important result that has emerged from recent in vivo work is that different mechanisms are used to transmit and transduce signals in different cellular and developmental contexts. Results are not always transferable between systems as defined by organism, tissue, or cell type. Therefore, future studies of signaling mechanisms will require the use of genetically mosaic embryos and inducible alleles that are activated in a restricted manner by heat, irradiation, or locally expressed recombinases. To date, many such studies have used loss- or gain-of-function alleles, but more directed alleles are needed to isolate the roles of promiscuous signaling proteins. Analyses of signaling pathways in isolation are instrumental for the elucidation of core pathway components and prominent signal transduction mechanisms. However, cells in vivo are commonly exposed to multiple concomitant cues. Thus, to fully understand cell–cell signaling, we need to transition from thinking of individual signaling pathways to considering how they are interwoven to form comprehensive signaling networks within responding cells. Recent advances in different areas of developmental signaling (particularly those that incorporate systems biology approaches) have begun to illuminate some mechanisms of crosstalk between major pathways. These include the convergence of signal transduction onto shared intermediates, transcriptional feedback loops, and the combinatorial regulation of transcription. Integration, convergence, synergy, or antagonism between signaling pathways can dramatically affect a cell’s interactions with its environment. In addition, these and other mechanisms may confer preferential responsiveness to a particular signal or enable a cell to respond differently to distinct combinations of signals. As new tools are developed for addressing developmental questions at a systems biology level, the amount of data generated in the field is growing exponentially. Consequently, biologists are growing increasingly reliant on computer scientists and computational biologists for data analysis, management, and access. Results from many experiments can no longer be contained within a standard journal article and instead require Web-based data supplements. Many collaborative efforts have been undertaken to centralize vast amounts of data in public databases and Web sites. However, several of these resources remain underused, largely as a result of insufficient publicity and a lack of infrastructure linking related data sets. Within the mouse community, this is especially apparent. Whereas the fly, worm, and fish communities have developed Web sites that comprehensively include expression, phenotype, genetic, and physical interaction data as well as available alleles and publication links, the mouse data sets are currently dispersed in several unlinked Web sites. It will take an enormous effort to integrate the information contained in these sites, but such an undertaking would create a tremendously valuable resource for the scientific community. A comprehensive mammalian database incorporating multiple types of mouse data as well as human genetic and phenotypic data could bridge developmental and medical research and make the networking of Web sites for different model organisms a far more accessible goal.

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

22

A wealth of information is contained within current Web-based resources, but the full significance of this data lies waiting to be unveiled in computational analyses that integrate different types of data. The power of this approach was demonstrated recently by Zhong and Sternberg (2006), who generated genomewide predictions of functional interactions in C. elegans by integrating expression, phenotype, and physical and genetic interaction data from multiple model systems. Computational and experimental systems biology approaches provide exciting and essential complements to genetic and biochemical investigations of cell–cell signaling. Together, these different types of studies will elevate our understanding of developmental signaling to a new level in coming years.

SUMMARY

 Several cell–cell signaling pathways are used reiteratively to instruct devel-











opmental processes, and form complex networks within cells that we are only beginning to understand thanks to convergent genetic, genomic, and biochemical studies. Components and modifiers of developmental signaling pathways have been identified in several types of screens. These screens have revealed that specific pathways are affected by proteins involved in general cellular processes as well as factors that belong to more traditional signal transduction classes. Now, the challenge is to understand how newly implicated factors affect signaling and development in vivo. Signal (and signal transduction) localization is highly regulated in vivo by a variety of mechanisms, including regulated stability and/or diffusion, facilitated transport, the protrusion of cytoplasmic filaments containing receptors, and cycles of endocytosis and secretion. Recent studies using directed signaling alleles have identified molecular mechanisms by which closely related proteins drive distinct responses, and have shown that context-specific signaling mechanisms are used in vivo. Several mechanisms of antagonistic and synergistic crosstalk between pathways have been identified, including the coregulation of signaling intermediates, transcriptional feedback regulation, and the convergence of transcriptional effectors at target enhancer or promoter elements. Expression profiling studies to characterize the transcriptional responses to specific signaling events are constantly being improved by the use of increasingly relevant sample sources as well as amplification techniques. However, technologies still need to be developed that enable researchers to study other responses such as cell migration and proliferation in largescale experiments.

ACKNOWLEDGMENTS We sincerely thank our laboratory colleagues and Susan Parkhurst for their comments on this manuscript. We apologize to the many authors whose work we were unable to cite because of space limitations and the large scope of this chapter’s subject area. Work in the author’s laboratory is supported by NIH grants HD 24875 and HD 25326.

GLOSSARY OF TERMS

23

GLOSSARY OF TERMS Argosome A type of vesicle that is derived from the basolateral membranes in the Drosophila wing disc epithelium and transports a signaling protein across a field of cells; proposed to traverse the wing disc by repeated cycles of transcytosis. Cell-autonomous Affecting only the cell of origin. Cell-nonautonomous Having effects that are not restricted to the cell of origin, as does a secreted protein. Cytoneme A thin, actin-based cellular protrusion several cell diameters long that extends from the apical surface of a cell toward a source of signaling protein (ligand); first observed in Drosophila wing disc cells extending toward Dpp and Wg. Dominant phenotype A phenotype that results when a single mutant copy of a given gene functionally dominates over the second wild-type allele. Effector protein/pathway A signaling protein or pathway that drives a biochemical or cellular response to a stimulus or signaling event; for example, in RTK signaling, a protein or pathway that is activated in response to receptor activation through the recruitment of an adaptor or another protein to the active receptor. Enhancer element A region of DNA that affects gene transcription in cis through the recruitment of transcription factors or other DNA binding/modifying proteins. ENU N-ethyl N-nitrosourea; a chemical mutagen that induces point mutations in DNA in a dosage-dependent manner. Epistasis A functional interaction between nonallelic genes; the ability of one allele to suppress the phenotypic consequences of a second mutation, which typically indicates that the epistatic mutation is dominant or is downstream in a common genetic pathway. Feedback regulation A mechanism by which a signaling pathway regulates its own activity; for example, by activating a regulatory factor that alters signal transduction, by altering the sensitivity of the pathway to upstream signals, and/or by modifying the activity or interactions of proteins in the pathway. Filopodia Thin, short cellular protrusions that contain both actin and microtubules and are not polarized toward a source of signaling protein. Gain-of-function mutation A mutation that results in a hyperactive gene product due to deregulated expression or function; for example, a mutation that renders the gene product resistant to the effects of regulatory enzymes.

24

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

Genetic interaction Functional synergy between two mutations that is suggestive of the gene products acting together in a given process or pathway. A genetic interaction is manifested through a compound mutant phenotype that is more pronounced than the sum of the two single mutant phenotypes; for example, disrupted development in an animal that is heterozygous for two mutations for which either heterozygous mutation (in isolation) does not result in a developmental phenotype. Glycosyl phosphatidylinositol (GPI) A type of phospholipid that is often conjugated to proteins and used to tether them to the plasma membrane. Heparan sulfate proteoglycan (HSPG) A macromolecule comprised of a core protein and glycosaminoglycan side chains of the heparan sulfate (polysaccharide) family; HSPGs are abundant in the extracellular matrix and are sometimes associated with plasma membranes via lipid moieties. They are important for many signaling events as revealed by the effects of mutations in HSPG core proteins and synthesis enzymes (e.g., those involved in appending the HS side groups). Hypomorphic allele/mutation A mutation that incompletely disrupts gene function and causes a phenotype that is less severe than a null (complete loss-of-function) mutation. In silico Computational, using informatics and computer-based resources. Interactome A large-scale protein interaction map, based on the results of biochemical assays testing all known coding gene products for their ability to interact physically with one or more protein(s) of interest. Kinase An enzyme that transfers a phosphate group to a substrate protein in an adenosine-triphosphate (ATP)–dependent reaction; often used in signal transduction to alter the activity or binding properties of a protein in a cascade. Lateral inhibition A signaling-mediated process by which one cell restricts the developmental potential or fate of its neighbor. Loss-of-function mutation An inactivating mutation that blocks gene expression or impairs the function of a gene product. Morphogen A protein that acts on target cells at a distance from its cell of origin, that forms an expression or activity gradient over a field of responsive cells, and that drives different cellular responses at different concentrations or activity thresholds. Niche A specific milieu defined temporally, spatially, and in some cases functionally; often during development, a given niche (e.g., a stem-cell niche) possesses

GLOSSARY OF TERMS

25

specialized characteristics as a result of the cellular composition of the niche itself as well as its interactions with nearby cells or proteins. Phosphatase An enzyme that removes a phosphate group from a substrate protein; like kinases, phosphatases are often used to regulate signal transduction. Physical interaction A direct binding interaction between two proteins. Posttranscriptional modification The modification of an mRNA after gene transcription but before translation into protein; for example, a splicing event. Posttranslational modification The modification of a protein; for example, phosphorylation, lipid conjugation, or cleavage. Recessive phenotype A phenotype that results from a mutation whose consequences can be functionally suppressed by a single wild-type allele; a mutation that only disrupts normal gene function in the homozygous state. RNA interference (RNAi) A means of knocking down the levels of one or more transcripts by introducing double-stranded or short-interfering RNAs to a cell and thus inducing the degradation of sequence-homologous mRNA. Screen, expression A screen based on development gene expression patterns. Screen, phenotype-driven A screen for proteins that affect the same developmental processes as assessed by developmental outcome, often performed using random mutagenesis approaches. Screen, sequence-driven A screen that uses genome sequence and annotation information together with gene-directed approaches to scan an entire genome for genes of interest. Synexpression Developmental coexpression. Systems biology The use of unbiased, high-throughput methods to simultaneously analyze all components of a biological system, thus providing a description of the whole system rather than its isolated components; for example, analyzing genome-wide changes in transcript or protein levels. Transcriptional profiling The analysis of mRNA expression, often using microarrays; in cell-signaling studies, comparative transcriptional profiling is often used to assess the transcriptional targets of a signaling pathway. Transcriptional reporter An experimental tool used to monitor the activity of a gene promoter or enhancer element; a protein of measurable activity or intensity (e.g.,

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

26

luciferase, b-galactosidase) driven by the transcriptional control elements of a gene of interest. Transcytosis The internalization, vesicular transport, and exocytosis of a secreted signaling protein that moves it through a cell and releases it into the extracellular space distal to its site of origin. Yeast two hybrid (Y2H) An experimental method used to assay for direct protein–protein interactions; with this method, a “bait” protein is fused to the DNA-binding domain of a transcription factor (TF), and a series of “fish” proteins are fused to the activation domain of the TF. When the bait and fish proteins physically interact, the proximity of the two TF domains render the complex capable of driving the expression of a reporter gene.

REFERENCES Apionishev S, Katanayeva NM, Marks SA, et al: Drosophila Smoothened phosphorylation sites essential for Hedgehog signal transduction, Nat Cell Biol 7:86–92, 2005. Aubin J, Davy A, Soriano P: In vivo convergence of BMP and MAPK signaling pathways: impact of differential Smad1 phosphorylation on development and homeostasis, Genes Dev 18:1482–1494, 2004. Autiero M, Waltenberger J, Communi D, et al: Role of PlGF in the intra- and intermolecular cross talk between VEGF receptors Flt1 and Flk1, Nat Med 9:936–943, 2003. Avidor-Reiss T, Maer AM, Koundakjian E, et al: Decoding cilia function: defining specialized genes required for compartmentalized cilia biogenesis, Cell 117:527–539, 2004. Baeg G-H, Zhou R, Perrimon N: Genome-wide RNAi analysis of JAK/STAT signaling components in Drosophila, Genes Dev 19:1861–1870, 2005. Cabernard C, Affolter M: Distinct roles for two receptor tyrosine kinases in epithelial branching morphogenesis in Drosophila, Dev Cell 9:831–842, 2005. Camps M, Chabert C, Muda M, et al: Induction of the mitogen-activated protein kinase phosphatase MKP3 by nerve growth factor in differentiating PC12, FEBS Lett 425:271–276, 1998. Cebria F, Kobayashi C, Umesono Y, et al: FGFR-related gene nou-darake restricts brain tissues to the head region of planarians, Nature 419:620–624, 2002. Chen HB, Shen J, Ip YT, Xu L: Identification of phosphatases for Smad in the BMP/DPP pathway, Genes Dev 20:648–653, 2006. Chen JC, Taipale J, Cooper MK, Beachy PA: Inhibition of Hedgehog signaling by direct binding of cyclopamine to Smoothened, Genes Dev 16:2743–2748, 2002. Chen W, Ren XR, Nelson CD, et al: Activity-dependent internalization of smoothened mediated by beta-arrestin 2 and GRK2, Science 306:2257–2260, 2004a. Chen WV, Delrow J, Corrin PD, et al: Identification and validation of PDGF transcriptional targets by microarray-coupled gene-trap mutagenesis, Nat Genet 36:304–312, 2004b. Choo BG, Kondrichin I, Parinov S, et al: Zebrafish transgenic Enhancer TRAP line database (ZETRAP), BMC Dev Biol 6:5, 2006. Collins RT, Cohen SM: A genetic screen in Drosophila for identifying novel components of the Hedgehog signaling pathway, Genetics 170:173–184, 2005. Corbit KC, Aanstad P, Singla V, et al: Vertebrate Smoothened functions at the primary cilium, Nature 437:1018–1021, 2005. DasGupta R, Kaykas A, Moon RT, Perrimon N: Functional genomic analysis of the Wnt-wingless signaling pathway, Science 308:826–833, 2005. Davy A, Aubin J, Soriano P: Ephrin-B1 forward and reverse signaling are required during mouse development, Genes Dev 18:572–583, 2004. Davy A, Soriano P: Ephrin signaling in vivo: look both ways, Dev Dyn 232:1–10, 2005. Denef N, Neubuser D, Perez L, Cohen SM: Hedgehog induces opposite changes in turnover and subcellular localization of patched and smoothened, Cell 102:521–531, 2000.

REFERENCES

27 Dossenbach C, Rock S, Affolter M: Specificity of FGF signaling in cell migration in Drosophila, Development 128:4563–4572, 2001. Eggenschwiler JT, Espinoza E, Anderson KV: Rab23 is an essential negative regulator of the mouse Sonic hedgehog signalling pathway, Nature 412:194–198, 2001. Eriksson U, Alitalo K: Structure, expression, and receptor-binding properties of novel vascular endothelial growth factors, Curr Top Microbiol Immunol 237:41–57, 1999. Fambrough D, McClure K, Kazlauskas A, Lander ES: Diverse signaling pathways activated by growth factor receptors induce broadly overlapping, rather than independent, sets of genes, Cell 97:727–741, 1999. Fire A, Xu S, Montgomery MK, et al: Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans, Nature 391:806–811, 1998. Fong CW, Leong HF, Wong ES, et al: Tyrosine phosphorylation of Sprouty2 enhances its interaction with c-Cbl and is crucial for its function, J Biol Chem 278:33456–33464, 2003. Formstecher E, Aresta S, Collura V, et al: Protein interaction mapping: a Drosophila case study, Genome Res 15:376–84, 2005. Furriols M, Casanova J: In and out of Torso RTK signalling, EMBO J 22:1947–1952, 2003. Fu¨rthauer M, Lin W, Anget S-L, et al: Sef is a feedback-induced antagonist of Ras/MAPKmediated FGF signalling, Nat Cell Biol 4:170–174, 2002. Garcı´a-Garcı´a MJ, Anderson KV: Essential role of glycosaminoglycans in Fgf signaling during mouse gastrulation, Cell 114:727–737, 2003. Gayko U, Cleghon V, Copeland T, et al: Synergistic activities of multiple phosphotyrosine residues mediate full signaling from the Drosophila Torso receptor tyrosine kinase, Proc Natl Acad Sci USA 96:523–528, 1999. Giot L, Bader JS, Brouwer C, et al: A protein interaction map of Drosophila melanogaster, Science 302:1727–1736, 2003. Go´mez AR, Lopez-Varea A, Molnar C, et al: Conserved cross-interactions in Drosophila and Xenopus between Ras/MAPK signaling and the dual-specificity phosphatase MKP3, Dev Dyn 232:695–708, 2005. Hadari YR, Gotoh N, Kouhara H, et al: Critical role for the docking-protein FRS2 alpha in FGF receptor-mediated signal transduction pathways, Proc Natl Acad Sci USA 98:8578–8583, 2001. Hallikas O, Palin K, Sinjushina N, et al: Genome-wide prediction of mammalian enhancers based on analysis of transcription-factor binding affinity, Cell 124:47–59, 2006. Hamilton TG, Klinghoffer RA, Corrin PD, Soriano P: Evolutionary divergence of platelet-derived growth factor alpha receptor signaling mechanisms, Mol Cell Biol 23:4013–4025, 2003. Han YG, Kwok BH, Kernan MJ: Intraflagellar transport is required in Drosophila to differentiate sensory cilia but not sperm, Curr Biol 13:1679–1686, 2003. Hanafusa H, Torii S, Yasunaga T, et al: Shp2, an SH2-containing protein-tyrosine phosphatase, positively regulates receptor tyrosine kinase signaling by dephosphorylating and inactivating the inhibitor Sprouty, J Biol Chem 279:22992–22995, 2004. Hanafusa H, Torii S, Yasunaga T, Nishida E: Sprouty1 and Sprouty2 provide a control mechanism for the Ras/MAPK signalling pathway, Nat Cell Biol 4:850–858, 2002. Haremaki T, Tanaka Y, Hongo I, et al: Integration of multiple signal transducing pathways on Fgf response elements of the Xenopus caudal homologue Xcad3, Development 130:4907–4917, 2003. Hasson P, Egoz N, Winkler C, et al: EGFR signaling attenuates Groucho-dependent repression to antagonize Notch transcriptional output, Nat Genet 37:101–105, 2005. Haycraft CJ, Banizs B, Aydin-Son Y, et al: Gli2 and Gli3 localize to cilia and require the intraflagellar transport protein Polaris for processing and function, PLoS Genet 1:e53, 2005. Heldin CH, Westermark B: Mechanism of action and in vivo role of platelet-derived growth factor, Physiol Rev 79:1283–1316, 1999. Hoch RV Distinctive mechanisms of receptor tyrosine kinase signal transduction required during mouse embryogenesis, Molecular and Cellular Biology, University of Washington, Seattle, WA, 2005 (doctoral dissertation). Hoch RV, Soriano P: Context-specific requirements for Fgfr1 signaling through Frs2 and Frs3 during mouse development, Development 133:663–673, 2006. Hsiung F, Ramirez-Weber F-A, Iwaki DD, Kornberg TB: Dependence of Drosophila wing imaginal disc cytonemes on Decapentaplegic, Nature 437:560–563, 2005. Huangfu D, Anderson KV: Cilia and Hedgehog responsiveness in the mouse, Proc Natl Acad Sci USA 102:11325–11330, 2005a. Huangfu D, Anderson KV: Signaling from Smo to Ci/Gli: conservation and divergence of Hedgehog pathways from Drosophila to vertebrates, Development 133:3–14, 2005b.

28

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

Huangfu D, Liu A, Rakeman AS, et al: Hedgehog signalling in the mouse requires intraflagellar transport proteins, Nature 426:83–87, 2003. Ha¨cker U, Nybakken K, Perrimon N: Heparan sulphate proteoglycans: the sweet side of development, Nat Rev Mol Cell Biol 6:530–541, 2005. Incardona JP, Gaffiend W, Kapur RP, Roelink H: The teratogenic Veratrum alkaloid cyclopamine inhibits Sonic hedgehog signal transduction, Development 125:3553–3562, 1998. Jarvis LA, Toering SJ, Simon MA, et al: Sprouty proteins are in vivo targets of Corkscrew/SHP2 tyrosine phosphatases, Development 133:1133–1142, 2006. Kadesch T: Notch signaling: the demise of elegant simplicity, Curr Opin Genet Dev 14:506–512, 2004. Kawakami Y, Rodriguez-Leon J, Koth CM, et al: MKP3 mediates the cellular response to FGF8 signalling in the vertebrate limb, Nat Cell Biol 5:513–519, 2003. Kim M, Cha GH, Kim S, et al: MKP-3 has essential roles as a negative regulator of the Ras/mitogen-activated protein kinase pathway during Drosophila development, Mol Cell Biol 24:573–583, 2004. Klinghoffer RA, Hamilton TG, Hoch R, Soriano P: An allelic series at the PDGFaR locus indicates unequal contributions of distinct signaling pathways during development, Dev Cell 2:103–113, 2002. Klinghoffer RA, Mueting-Nelsen PF, Faerman A, et al: The two PDGFRs display conserved signaling in vivo despite divergent embryological functions, Mol Cell 7:343–354, 2001. Kouhara H, Hadari YR, Spivak-Kroizman T, et al: A lipid-anchored Grb2-binding protein that links FGF-receptor activation to the Ras/MAPK signaling pathway, Cell 89:693–702, 1997. Kovalenko D, Yang X, Nadeau RJ, et al: Sef inhibits fibroblast growth factor signaling by inhibiting FGFR1 tyrosine phosphorylation and subsequent ERK activation, J Biol Chem 278:14087–14091, 2003. Kramer S, Okabe M, Hacohen N, et al: Sprouty: a common antagonist of FGF and EGF signaling pathways in Drosophila, Development 126:2515–2525, 1999. Kudoh T, Tsang M, Hukriede NA, et al: A gene expression screen in zebrafish embryogenesis, Genome Res 11:1979–1987, 2001. Kuroda H, Fuentealba L, Ikeda A, et al: Default neural induction: neuralization of dissociated Xenopus cells is mediated by Ras/MAPK activation, Genes Dev 19:1022–1027, 2005. LaBonne C, Bronner-Fraser M: Neural crest induction in Xenopus: evidence for a two-signal model, Development 125:2403–2414, 1998. Lax I, Wong A, Lamothe B, et al: The docking protein FRS2alpha controls a MAP kinasemediated negative feedback mechanism for signaling by FGF receptors, Mol Cell 10:709–719, 2002. Leve´en P, Pekny M, Gebre-Medhin S, et al: Mice deficient for PDGF B show renal, cardiovascular, and hematological abnormalities, Genes Dev 8:1875–1887, 1994. Li S, Armstrong CM, Bertin N, et al: A map of the interactome network of the metazoan C. elegans, Science 303:540–543, 2004a. Li X, Brunton VG, Burgar HR, et al: FRS2-dependent SRC activation is required for fibroblast growth factor receptor-induced phosphorylation of Sprouty and suppression of ERK activity, J Cell Sci 117:6007–6017, 2004b. Lickert H, Cox B, Wehrle C, et al: Dissecting Wnt/beta-catenin signaling during gastrulation using RNA interference in mouse embryos, Development 132:2599–2609, 2005. Lin X, Buff EM, Perrimon N, Michelson AM: Heparan sulfate proteoglycans are essential for FGF receptor signaling during Drosophila embryonic development, Development 126:3715–3723, 1999. Lindblom P, Gerhardt H, Liebner S, et al: Endothelial PDGF-B retention is required for proper investment of pericytes in the microvessel wall, Genes Dev 17:1835–1840, 2003. Lum L, Beachy PA: The hedgehog response network: sensors, switches, and routers, Science 304:1755–1759, 2004. Lum L, Yao S, Mozer B, et al: Identification of Hedgehog pathway components by RNAi in Drosophila cultured cells, Science 299:2039–2045, 2003. Massague´ J: Integration of Smad and MAPK pathways: a link and a linker revisited, Genes Dev 17:2993–2997, 2003. Meyer RD, Mohammadi M, Rahimi N: A single amino acid substitution in the activation loop defines the decoy characteristic of VEGFR-1/FLT-1, J Biol Chem 281:867–875, 2006. Meyer RD, Rahimi N: Comparative structure-function analysis of VEGFR-1 and VEGFR-2: what have we learned from chimeric systems? Ann NY Acad Sci 995:200–207, 2003.

REFERENCES

29 Min H, Danilenko DM, Scully SA, et al: Fgf-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless, Genes Dev 12:3156–3161, 1998. Mizutani CM, Nie Q, Wan FYM, et al: Formation of the BMP activity gradient in the Drosophila embryo, Dev Cell 8:915–924, 2005. Moghal N, Sternberg PW: The epidermal growth factor system in Caenorhabditis elegans, Exp Cell Res 284:150–159, 2003. Mohammadi M, Olsen SK, Goetz R: A protein canyon in the FGF-FGF receptor dimer selects from an a al carte menu of heparan sulfate motifs, Curr Opin Struct Biol 15:506–516, 2005. Mu¨ller P, Kuttenkeuler D, Gesellchen V, et al: Identification of JAK/STAT signalling components by genome-wide RNA interference, Nature 436:871–875, 2005. Nagaraj R, Banerjee U: The little R cell that could, Int J Dev Biol 48:755–760, 2004. Nusse R: Wnt signaling in disease and in development, Cell Research 15:28–32, 2005. Nybakken K, Vokes SA, Lin T-Y, et al: A genome-wide RNA interference screen in Drosophila melanogaster cells for new components of the Hh signaling pathway, Nat Genet 37:1323–1332, 2005. O’Connor MB, Umulis D, Othmer HG, Blair SS: Shaping BMP morphogen gradients in the Drosophila embryo and pupal wing, Development 133:183–193, 2006. Ornitz DM: FGFs, heparan sulfate and FGFRs: complex interactions essential for development, Bioessays 22:108–112, 2000. Parinov S, Kondrichin I, Korzh V, Emelyanov A: Tol2 transposon-mediated enhancer trap to identify developmentally regulated zebrafish genes in vivo, Dev Dyn 231:449–459, 2004. Pera EM, Ikeda A, Eivers E, De Robertis EM: Integration of IGF, FGF, and ani-BMP signals via Smad1 phosphorylation in neural induction, Genes Dev 17:3023–3028, 2003. Pera EM, Kim JI, Martinez SL, et al: Isthmin is a novel secreted protein expressed as a part of the Fgf-8 synexpression group in the Xenopus midbrain-hindbrain organizer, Mech Dev 116:169–172, 2002. Porter AC, Vaillancourt RR: Tyrosine kinase receptor-activated signal transduction pathways which lead to oncogenesis, Oncogene 16:1343–1352, 1998. Preger E, Ziv I, Shabtay A, et al: Alternative splicing generates an isoform of the human Sef gene with altered subcellular localization and specificity, Proc Natl Acad Sci USA 101:1229–1234, 2004. Price JV, Savenye ED, Lum D, Breitkreutz A: Dominant enhancers of Egfr in Drosophila melanogaster: genetic links between the Notch and Egfr signaling pathways, Genetics 147:1139–1153, 1997. Rahimi N, Dayanir V, Lashkari K: Receptor chimeras indicate that the vascular endothelial growth factor receptor-1 (VEGFR-1) modulates mitogenic activity of VEGFR-2 in endothelial cells, J Biol Chem 275:16986–16992, 2000. Ramirez-Weber F-A, Kornberg TB: Cytonemes: cellular processes that project to the principal signalling center in Drosophila imaginal discs, Cell 97:599–607, 1999. Ray K, Perez SE, Yang Z, et al: Kinesin-II is required for axonal transport of choline acetyltransferase in Drosophila, J Cell Biol 147:507–518, 1999. Reich A, Sapir A, Shilo B: Sprouty is a general inhibitor of receptor tyrosine kinase signaling, Development 126:4139–4147, 1999. Reichman-Fried M, Dickson B, Hafen E, Shilo BZ: Elucidation of the role of breathless, a Drosophila FGF receptor homolog, in tracheal cell migration, Genes Dev 8:428–439, 1994. Ribeiro C, Ebner A, Affolter M: In vivo imaging reveals different cellular functions for FGF and Dpp signaling in tracheal branching morphogenesis, Dev Cell 2:677–683, 2002. Roberts DM, Kearney JB, Johnson JH, et al: The vascular endothelial growth factor (VEGF) receptor Flt-1 (VEGFR-1) modulates Flk-1 (VEGFR-2) signaling during blood vessel formation, Am J Pathol 164:1531–1535, 2004. Sapkota G, Alarco´n C, Spagnoli FM, et al: Balancing BMP signaling through integrated inputs into the Smad1 linker, Mol Cell 25:441–454, 2002. Sato M, Kornberg TB: FGF is an essential mitogen and chemoattractant for the air sacs of the Drosophila tracheal system, Dev Cell 3:195–207, 2002. Schmahl J, Raymond CS, Soriano P: PDGF signaling specificity is mediated through multiple immediate early genes, Nat Genet 39:52–60, 2007. Schlessinger J: Cell signaling by receptor tyrosine kinases, Cell 103:211–225, 2000. Sekine K, Ohuchi H, Fujiwara M, et al: Fgf10 is essential for limb and lung formation, Nat Genet 21:138–141, 1999.

30

UNTANGLING THE GORDIAN KNOT: CELL SIGNALING EVENTS THAT INSTRUCT DEVELOPMENT

Shaye DD, Greenwald I: Endocytosis-mediated downregulation of LIN-12/Notch upon Ras activation in Caenorhabditis elegans, Nature 420:686–690, 2002. Shaye DD, Greenwald I: LIN-12/Notch trafficking and regulation of DSL ligand activity during vulval induction in Caenorhabditis elegans, Development 132:5081–5092, 2005. Shi Y, Massague´ J: Mechanisms of TGF-beta signaling from cell membrane to the nucleus, Cell 113:685–700, 2003. Shilo B-Z: Signaling by the Drosophila epidermal growth factor receptor pathway during development, Exp Cell Res 284:140–149, 2003. Simon MA: Receptor tyrosine kinases: specific outcomes from general signals, Cell 103:13–15, 2000. Soriano P: Abnormal kidney development and hematological disorders in PDGF beta-receptor mutant mice, Genes Dev 8:1888–1896, 1994. Steventon B, Carmona-Fontaine C, Mayor R: Genetic network during neural crest induction: from cell specification to cell survival, Semin Cell Dev Biol 16:647–654, 2005. Sutherland D, Samakovlis C, Krasnow MA: branchless encodes a Drosophila FGF homolog that controls tracheal cell migration and the pattern of branching, Cell 87:1091–1101, 1996. Taipale J, Chen JK, Cooper MK, et al: Effects of oncogenic mutations in Smoothened and Patched can be reversed by cyclopamine, Nature 406:1005–1009, 2000. Takemoto T, Uchikawa M, Kamachi Y, Kondoh H: Convergence of Wnt and FGF signals in the genesis of posterior neural plate through activation of the Sox2 enhancer N-1, Development 133:297–306, 2006. Tallquist MD, French WJ, Soriano P: Additive effects of PDGF receptor b signaling pathways in vascular smooth muscle cell development, PLoS Biol 1:288–299, 2003. Tan PBO, Kim SK: Signaling specificity: the RTK/RAS/MAP kinase pathway in metazoans, Trends Genet 15:145–149, 1999. Teleman AA, Cohen SM: Dpp gradient formation in the Drosophila wing imaginal disc, Cell 103:971–980, 2000. Tewari M, Hu PJ, Ahn JS, et al: Systematic interactome mapping and genetic perturbation analysis of a C. elegans TGF-beta signaling network, Mol Cell 13:469–482, 2004. Torii S, Kusakabe M, Yamamoto T, et al: Sef is a spatial regulator for Ras/MAP kinase signaling, Dev Cell 7:33–44, 2004. Tsang M, Friesel R, Kudoh T, Dawid IB: Identification of Sef, a novel modulator of FGF signalling, Nat Cell Biol 4:165–169, 2002. Tsang M, Maegawa S, Kiang A, et al: A role for MKP3 in axial patterning of the zebrafish embryo, Development 131:2769–2779, 2004. Wang J, Barr MM: RNA interference in Caenorhabditis elegans, Methods Enzymol 392:36–55, 2005. Wang JK, Xu H, Li HC, Goldfarb M: Broadly expressed SNT-like proteins link FGF receptor stimulation to activators of Ras, Oncogene 13:721–729, 1996. Wawersik S, Evola C, Whitman M: Conditional BMP inhibition in Xenopus reveals stage-specific roles for BMPs in neural and neural crest induction, Dev Biol 277:425–442, 2005. Williams PH, Hagemann A, Gonza´lez-Gaita´n M, Smith JC: Visualizing long-range movement of the morphogen Xnr2 in the Xenopus embryo, Curr Biol 14:1916–1923, 2004. Wong A, Lamothe B, Lee A, et al: FRS2 alpha attenuates FGF receptor signaling by Grb2mediated recruitment of the ubiquitin ligase Cbl, Proc Natl Acad Sci USA 99:6684–6689, 2002. Xenarios I, Salwinski L, Duan XJ, et al: DIP, the Database of Interacting Proteins: a research tool for studying cellular networks of protein interactions, Nucleic Acids Res 30:303–305, 2002. Xu C, Kauffman RC, Zhang J, et al: Overlapping activators and repressors delimit transcriptional response to receptor tyrosine kinase signals in the Drosophila eye, Cell 103:87–97, 2000. Xu H, Lee KW, Goldfarb M: Novel recognition motif on fibroblast growth factor receptor mediates direct association and activation of SNT adapter proteins, J Biol Chem 273:17987–17990, 1998. Yoo AS, Bais C, Greenwald I: Crosstalk between the EGFR and LIN-12/Notch pathways in C. elegans vulval development, Science 303:663–666, 2004. Zhang C, Williams EH, Guo Y, et al: Extensive phosphorylation of Smoothened in Hedgehog pathway activation, Proc Natl Acad Sci USA 101:17900–17907, 2004. Zhong W, Sternberg PW: Genome-wide prediction of C. elegans genetic interactions, Science 311:1481–1484, 2006. Zhu AJ, Zheng L, Suyama K, Scott MP: Altered localization of Drosophila Smoothened protein activates Hedgehog signal transduction, Genes Dev 17:1240–1252, 2003.

FURTHER READING

31

Zuzarte-Luı´s V, Montero JA, Rodriquez-Leo´n J, et al: A new role for BMP5 during limb development acting through the synergistic activation of Smad and MAPK pathways, Dev Biol 272:39–52, 2004.

FURTHER READING Bangi E, Wharton K: Dual function of the Drosophila Alk1/Alk2 ortholog Saxophone shapes the Bmp activity gradient in the wing imaginal disc, Development 133:3295–3303, 2006. Crickmore MA, Mann RS: Hox control of organ size by regulation of morphogen production and mobility, Science 313:63–68, 2006. Echeverri CJ, Perrimon N: High-throughput RNAi screening in cultured cells: a user’s guide, Nat Rev Genet 7:373–384, 2006. Gandhi TK, Zhong J, Mathivanan S, et al: Analysis of the human protein interactome and comparison with yeast, worm and fly interaction datasets, Nat Genet 38:285–293, 2006. Ochi H, Pearson BJ, Chuang PT, et al: Hhip regulates zebrafish muscle development by both sequestering Hedgehog and modulating localization of Smoothened, Dev Biol 297:127–140, 2006. Singla V, Reiter JF: The primary cilium as the cell’s antenna: signaling at a sensory organelle, Science 313:629–633, 2006.

2

FINDING GENE EXPRESSION CHANGES USING MICROARRAY TECHNOLOGY TADAYOSHI HAYATA1 and KEN W. Y. CHO2 1 Department of Molecular Pharmacology, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan 2

Department of Developmental and Cell Biology, University of California, Irvine, CA

INTRODUCTION Fertilized eggs give rise to many millions of cells that represent hundreds of different cell types and that eventually form the complex structures of the adult. Each dividing cell makes numerous specific decisions about which genes to express among the tens of thousands of genes in its genome. If this tightly regulated process goes awry, an embryo may develop abnormalities or subsequently develop diseases as an adult. The regulatory programs that turn specific genes on or off are embedded in the genome of the organism. Therefore, a current major effort in biology is to understand the mechanisms by which gene networks are coordinated, thereby specifying paths of cellular differentiation. DNA microarray approaches provide biologists with the opportunity to generate expression inventories of all genes used to create the embryo. This information will lead to an improved understanding of how tissues and organs develop, which has clear implications for biomedicine. Until recently, the task of cataloging the expression of thousands of genes to understand how they control animal development was an unrealistic goal for biologists. However, advances in genome sequencing and the development of high-throughput microarray approaches are moving this formidable task into the realm of reality. Although it is still challenging, scientists have begun to organize the expression of numerous genes into various groups sharing similar expression patterns. This holistic, genome-level view that involves the examination of the simultaneous readouts of all of the components is beginning to transform the way we view biological processes. Data from such studies has enabled researchers to link previously unsuspected genes to particular developmental

32

Principles of Developmental Genetics

© 2007, Elsevier Inc. All rights reserved.

MICROARRAY PRINCIPLES

33

pathways, mutations, and diseases. We will discuss the basic theory behind microarrays and how the technology is useful for biomedical research.

I. MICROARRAY PRINCIPLES A. Basic Theory Behind Microarray Technology The principle underlying the ability of microarrays to detect specific transcripts is solid support nucleic acid hybridization as first developed during the 1970s by Ed Southern for the classical Southern blot hybridization. Microarray chips are distant cousins of Southern blots in the sense that a microarray can be seen as an orderly arrangement of many miniaturized DNA “dot blots.” Unlike standard dots blots, each dot of immobilized target DNA represents a single DNA species rather than a complex mixture of genomic DNA. Modern microarrays use DNA-bound platforms (slides or quartz wafers) containing very large numbers of genes arranged in an array distributed over a very small area (mm to cm) (Schena et al., 1995; Lockhart et al., 1996). This miniaturization makes it possible to examine the expression of large numbers of genes in a small sample size. Therefore, scientists are able to address increasingly complex questions and to perform intricate experiments to address gene expression profiling, genotyping, and the global effects of gain- and loss-of-function for specific genes. Several methods are used to fabricate microarrays (Schena, 2000). One popular method, known as spotted (glass) arrays, uses DNA generated from polymerase chain reaction (PCR) amplicons (size range, 0.5 kb to 2 kb) or long oligonucleotides (size range, 50 bp to 70 bp) mechanically deposited on the surface of chemically coated glass substrates (Figure 2.1). Oligonucleotide arrays are designed to match specific subsequences in known or predicted mRNAs. Both types of arrays use a robotically driven device containing a set of metal pins, each of which is dipped into wells containing different DNA samples. These pins are then used to deposit a small amount of these DNAs onto the surface of coated microscope slides in serial order with a distance of 100 mm to 200 mm between each DNA spot, thereby generating a high-density (up to 80,000 spots) microarray slide. Alternatively, inkjet technology has been successfully used to print oligonucleotides onto glass slides, which are then used for microarray hybridization. Another widely used array format, the Affymetrix GeneChipW (Affymetrix, Santa, Clara, CA), is generated on a quartz wafer by photolithography, which is the same process that is used in semiconductor manufacturing (Figure 2.2). One can think of the Affymetrix GeneChipW construction as in situ oligonucleotide synthesis on the chip. Initially, a mask is aligned with the wafer, and light is directed through the mask to activate (deprotect) exposed substrate. Next, chemically modified nucleosides are introduced and chemically coupled, and then a capping step blocks uncoupled sites. This process is repeated with different masks until the probes are synthesized to full length (20 to 25 nucleotides in length) to generate high-density microarrays. To generate probes for spotted arrays, RNA samples isolated from two different cell types, embryonic stages, tissues, or treatments are reverse transcribed and labeled with either of two fluorescent molecules, Cy3 (green) or Cy5 (red) (Eisen and Brown, 1999). The labeled cDNAs (probes) are denatured, mixed, applied to the microarray containing glass slides, and allowed to hybridize competitively. Hybridized slides are washed and subjected to laser

34

FINDING GENE EXPRESSION CHANGES USING MICROARRAY TECHNOLOGY

FIGURE 2.1 The basic principle behind spotted microarrays. A, RNA is extracted from a cell or tissue sample and then converted to cDNA. B, Fluorescent tags (usually Cy3 and Cy5) are enzymatically incorporated into the newly synthesized cDNA, or they can be chemically attached to the new strands of DNA. C, The dye-labeled cDNAs are mixed and hybridized to the microarray, on which DNA has been spotted. D, A dye-labeled cDNA species that contains a sequence that is complementary to one of the single-stranded probe sequences on the array will hybridize to the corresponding spot. After hybridization, the red and green fluorescent signals from each spot are measured using a confocal laser scanner. The intensities indicate the level of expression of a particular gene. The ratio of red to green reflects the relative expression of each gene between the experimental and reference samples. E, A magnified view of panel C. (See color insert.)

FIGURE 2.2 The basic principle behind Affymetrix GeneChipsW. A, Millions of distinct oligonucleotides are directly synthesized on the surface of a quartz wafer by photolithography. About 20 to 25 distinct oligonucleotides, which are printed as individual features, represent the partial sequence of one gene. B, Perfectly matched and single base mismatched oligonucleotide probes against reference mRNA are designed. Each gene on the chip is typically represented by several different probe pairs on the chip. Scanned arrays produce raw data that consist of the intensities of the individual probe pairs (i.e., the perfectly matched and mismatched probes).

MICROARRAY PRINCIPLES

35

excitation, and then the resultant fluorescence of each dye is measured. The relative abundance of each mRNA present in the two samples being compared is inferred by measuring the ratio of the red and green fluorescence intensities for each spot on the array. The binding of specific fluorescent probes indicates whether genes are expressed. The intensity of the signals provides the abundance of the transcripts relative to control (frequently referred to as the “reference” probe; see later section) (Figure 2.1). Affymetrix GeneChipsW use a single probe for each chip. A biotin-labeled cDNA probe is prepared for hybridization. After hybridization, the chip is stained with streptavidin– phycoerythrin to detect the probe; it is then washed and scanned with a confocal laser, and the distribution pattern of signal in the array is recorded. B. Advantages and Disadvantages of Different Array Systems Spotted microarrays are useful for comparing relative mRNA expression levels between two populations of cells. For instance, if the main goal is to identify genes that are up- or down-regulated after a particular treatment (e.g., growth-factor stimulation), one would label and simultaneously hybridize probes prepared from both a reference control (unstimulated) sample and from a growth-factor–treated sample. Microarray spots preferentially fluorescing for the treated sample probe represent growth-factor upregulated genes and, likewise, spots preferentially fluorescing for the untreated control probe represent genes that are downregulated by the growth factor. Although this approach is useful for rapidly identifying changes in gene expression within slides, the experimental design does not allow for the easy comparison and analysis of expression levels among multiple different samples (experiments). For example, it is not straightforward to compare changes in gene expression (between control and treated samples) over multiple time points to chart temporal responses after a defined alteration to cells or their environment. To alleviate the problems of this type of analysis, one can design each hybridization experiment using a common reference sample. In the example above, the reference would be used with all time points. Thus, it is possible to indirectly compare the expression levels of two samples that are measured separately on two different slides. The ideal common reference should ensure consistent and nonzero values for all probes on the array so that no information is lost when the fluorescence ratios are calculated. For this purpose, mRNAs from whole embryos pooled from several developmental stages are frequently used. This ensures that mRNAs corresponding with each and all cDNA spots are represented at some level. This internal reference allows for the direct comparison of array experiments performed across tissues and different developmental states. Alternatively, hybridization can be spiked with a known amount of specific probes throughout the experiments, and the overall intensity can be normalized on the basis of the hybridization efficiency of the spiked probes. Methods are now available for standardizing global gene expression analysis among different platforms (Bammler et al., 2005). Advantages of spotted cDNA microarrays include their relative affordability and increased detection sensitivity resulting from longer target sequences, which enhance the hybridization efficiency. Disadvantages include the difficulties associated with monitoring the expression levels of differentially spliced transcripts from a single genomic locus and distinguishing among closely related genes that may potentially cross hybridize. These handicaps do not exist

36

FINDING GENE EXPRESSION CHANGES USING MICROARRAY TECHNOLOGY

FIGURE 2.3 Clustering analyses. A, The hierarchical activity performs a clustering analysis on the basis of pairwise comparison between expression values. All pairs of genes are computed to identify the level of correlation between experiments. Genes or/and groups are connected with lines and presented as dendrograms, which are trees that are based on the similarity of their gene expression patterns throughout numerous experiments. Then, the next most similar groups are connected with a line and correlated with their nearest neighbors. Gene expression levels are shown in a color gradient. B, In K-means clustering, the number of clusters (K) is first assigned, and here a K-value of 10 was used. K-means clustering finds 10 clusters of genes with similar expression profiles. Genes giving a similar expression profile to cluster 6 are shown in the bottom panel. (See color insert.)

in oligonucleotide arrays, including Affymetrix GeneChipsW. Oligonucleotide probes are designed to be relatively short and perfectly complementary to a target gene sequence. Therefore, a mismatched cDNA probe will not anneal efficiently to the oligonucleotide sequence. Additionally, in some cases, oligonucleotide chip design purposely incorporates a single base mismatch in the central region (Figure 2.3, B). Bona fide hybridization is distinguished from cross hybridization by comparing the signal obtained from the perfectly matched probe with that of the one mismatch probe. This sequence mismatch strategy, along with the use of multiple sequences for each gene, increases the specificity of the signal and minimizes the effects of nonspecific hybridization (Lipshutz et al., 1999). Although microarray technology is a powerful addition to “genomics” research, gene expression analysis in eukaryotic cells poses a significant challenge to the current microarray technology as a result of complexity and sensitivity issues. For example, human cells are thought to express 80,000 to 100,000 different transcripts derived from only 20,000 to 25,000 estimated genes. Of these, 99% are rare, occurring at a frequency of less than 1 copy per 20,000 transcripts, and half of mRNA populations in a cell are the product of about 300 genes (Bishop et al., 1974; Davidson and Britten, 1979). To detect both the rare and abundant messages at the same time, microarray technology must be both sensitive and have a large enough dynamic range. Using current

MICROARRAY PRINCIPLES

37

technology, typical cDNA arrays can reliably detect mRNA levels equivalent to 1 transcript in 300,000 using conventional direct incorporation of fluorescent dyes. The amount of RNA required to perform microarray analysis is often a limiting factor. Small sample sizes can make the recovery of sufficient mRNA difficult and thus prevent the synthesis of sufficient amounts of fluorescent cDNA probe. To overcome this limitation, synthesized cDNAs must first be amplified. Several different technologies are available for cDNA amplification. Examples include an adaptor-ligation–mediated PCR amplification method that allows one to obtain enough RNA from single cells to perform DNA microarray analysis. However, a major concern of such amplification is the fidelity of the representation of the original RNA population in the resulting amplified material. C. Data Analysis and Bioinformatics Microarrays have many variable experimental steps, including the design of the experiments themselves, RNA collection, probe preparation and labeling, and slide hybridization and scanning. Without the proper controls and quality checkpoints, the resulting data has the potential of being highly variable with an excess background level. When they are not properly monitored and controlled, these variations will drastically affect further data analysis and interpretation. In the past, the lack of standardization in arrays has presented problems during the exchanging and comparing of the array data sets. To facilitate this process, the scientific community has adopted the “Minimum Information About a Microarray Experiment” (MIAME) standard for describing a microarray experiment (Brazam et al., 2001). The MIAME standard has been adopted by many journals as a requirement for the submission of papers that incorporate microarray results. When handling microarray data, several issues should be carefully considered (Church, 2002; Stoeckert et al., 2004). A large set of spot identification information and acquired array hybridization data must be organized and cataloged into a usable form. Because tens of thousands of data points are acquired simultaneously, the reliability of the array data (minimizing the randomness) needs to be confirmed so that only reproducible data can be processed during subsequent data-mining analyses. Otherwise, poor-quality data could poison the array analysis, and faulty interpretations could then be made. Before actual array data analysis can be conducted, raw data must go through a “normalization” process to determine the quality of the data sets. This process attempts to compensate for technical differences among chips. For example, because the incorporations of Cy3 and Cy5 dyes are often different among probe samples, to properly compare the signal-intensity data between reference and experimental conditions, a global normalization (“correction”) process must be applied to the data sets to equalize the mean values of the expression levels for all genes between the experimental and control samples. Normalized data are compared using an analysis of variance test, which measures the difference between the means (averages) of two or more groups. Alternatively, a Bayesian probabilistic statistical method based on the t test can be used for comparing gene expression differences among different samples (Long et al., 2001). Statistically treated data can then be clustered using a variety of different methods (Quackenbush, 2001). We will discuss two commonly used

38

FINDING GENE EXPRESSION CHANGES USING MICROARRAY TECHNOLOGY

approaches: hierarchical clustering and K-means clustering. Hierarchical clustering is the most popular method for microarray data analysis. In hierarchical clustering, genes with similar expression behaviors shown in numerous microarray experiments are grouped together and connected by a series of branches, which produces a dendrogram (or clustering tree; Figure 2.3, A). With this bottom-up method, experiments are grouped together to identify genes that may behave similarly or that have been coexpressed throughout experiments. K-means clustering is an algorithm to classify a given data set through a certain number (K) of clusters (Figure 2.3, B). Because the number of group sets must be determined before an analysis can be initiated, this can be considered to be a top-down approach. One of the powerful outcomes of these clustering analyses is the possibility to infer the probable functions of new genes based on similarities in expression patterns with those of known genes or to link unsuspected genes to specific biological responses (e.g., developmental function, occurrence of disease states). For example, coregulated genes may have tight relationships (i.e., similar expression patterns) because they are regulated by the same transcription factor or by the same signaling pathway. This ability to infer the activity of the gene by microarray screening is one of the strengths of microarray technology, and it is a major difference from traditional hypothesis-driven approaches.

II. APPLICATIONS OF DNA MICROARRAY TECHNOLOGY A. Gene Expression Profiling by Microarrays Microarray-based studies attempt to monitor transcript levels in differentiating cells, genetic mutants, complex diseases, and others. Microarrays have been successfully used to link a previously unsuspected gene to a particular disease. For example, by comparing and following the diagnosis of hundreds of tumor samples, marker genes linked to metastatic tumors were isolated (Laura et al., 2002). In another case, examining genes that were differentially expressed between insulin-resistant and normal strains of rat identified a gene linked to the regulation of glucose metabolism (Aitman et al., 1999). Finally, by examining gene-expression profiles in a family affected by sudden infant death syndrome, a gene linked to the disease was discovered (Puffenberger et al., 2005). A number of clinical trials using microarrays are currently underway for the prognosis and therapeutic guidance of these and other diseases. Microarrays are also useful for identifying genes that respond to specific perturbations. For example, cells or embryos may be challenged with a growth factor or a chemical, and genes that are induced or inhibited can be identified. Gain-of-function studies involving mRNA overexpression and loss-of-function (knockdown) studies involving antisense morpholino oligonucleotides, siRNA, or the expression of dominant-negative variants of proteins will assist with the rapid identification of the genes that are affected by such manipulations. This will lead to the identification of potential gene function and the discovery of downstream target genes (Piano et al., 2002). It is interesting to note that, although an enormous number of microarray studies have been performed so far, only a fraction have concerned developmental biology. This disparity was partly the result of the limited availability of high-quality, high-density microarray chips for some model organisms

APPLICATIONS OF DNA MICROARRAY TECHNOLOGY

39

together with difficulties in isolating sufficient amounts of RNA to carry out microarray experiments. An additional layer of complexity is added, because developmental biologists prefer to study gene expression in populations of identical cells rather than in a whole tissue, an organ, or another heterogeneous population of cells so that cell fate changes can be better studied. One ingenious way to overcome this difficulty is to use RNA isolated from green-fluorescent-protein–marked tissues, from cells of transgenic embryos, or from cells separated by fluorescence-activated cell sorting using antibodies to cell-specific surface antigens. Because cell sorting is unlikely to produce enough cells to perform a traditional microarray experiment, RNA extracted from purified cells must be amplified. B. Comparative Genomic Hybridization Using Tiling Arrays Microarrays are useful for studying DNA variants, with the primary application being the identification and genotyping of mutations and polymorphisms. These applications pose different challenges than monitoring RNA expression does, because the focus is not the quantitation of the transcripts but rather discriminating a single nucleotide mismatch on the basis of differential hybridization. For this reason, oligonucleotide arrays (as opposed to spotted arrays) have been the method of choice. A tiling array consists of short overlapping oligonucleotides that are “tiled” across a sequence of interest and that differ only by having a nucleotide substitution at a single position. An amplified cDNA product containing the expected sequence will hybridize best to the exactly matched probe, whereas the rest of the probes will hybridize weakly as a result of sequence mismatches. This type of relatively simple tiling array has been successfully used to map single nucleotide polymorphisms (SNPs; pronounced “snips”). This technology is seeing increased use in new applications, such as comparative genomic hybridization and chromatin immunoprecipitation (ChIP), which are both discussed in more detail later in this chapter. C. Genotyping Single Nucleotide Polymorphisms More than 99% of the genome sequence is identical across the human population. SNPs reflect the small sequence variations (often a single base change) that can occur within an individual gene. It is important to note that SNPs do not change much from generation to generation. Human genome sequencing projects have identified more than 2 million SNPs as genetic markers. Most SNPs are found outside of coding sequences, but some SNPs found within a coding sequence are of particular interest to researchers, because the change may alter the biological function of a protein. Because of the enormous potential to associate SNP maps with the development of complex diseases such as cancer, Alzheimer’s, diabetes, and hypertension, researchers are feverishly working to identify thousands of useful SNP markers. For example, SNPs in the breast cancer genes 1 and 2 are associated with the development of breast cancers (Freedman et al., 2005). SNPs in the apolipoprotein E gene have been linked to a higher risk of developing Alzheimer’s disease (Bullido et al., 1998). One of the goals in this field is to generate arrays that are capable of genotyping thousands of polymorphisms in a single hybridization. This may someday allow for the construction of an individual’s genetic fingerprint that will be able to predict individual risk for developmental disorders and diseases. After

40

FINDING GENE EXPRESSION CHANGES USING MICROARRAY TECHNOLOGY

this technology emerges, we can also anticipate the discussion of social and ethical issues associated with the use of this information. D. Comparative Genomic Hybridization The completion of various genome projects coupled with improved array printing technology has led to the generation of more sophisticated tiling arrays. A new generation of high-density tiling array methods use millions of DNA probes that are evenly spaced across the genome, including in both coding and noncoding regions. Some of the genome tiling arrays consist of a multiple-array set, with each set containing more than a few million probe pairs. These oligonucleotide probes, which usually ranging in size from 25 to 50 nucleotides long, are tiled at a discrete resolution (e.g., 10 nucleotides apart) to cover the entire genomic sequence (Ishkanian et al., 2004). To perform a typical tiling array experiment, total genomic DNA is isolated from test and reference cell populations, fragmented by shearing, and then differentially labeled and simultaneously hybridized to DNA microarrays. The relative hybridization intensity of the test and reference signals at a given location should be proportional to the relative copy number of those sequences in the test and reference genomes. If the reference genome is normal, then increases and decreases in the intensity ratio directly indicate DNA copy-number variation in the genome of the test cells. This type of microarray experiment (Mantriparagada et al, 2004), which detects chromosomal imbalances and variation in DNA copy number, is known as comparative genetic hybridization, and it would be useful for mapping the regions of a genome containing deletion mutations, chromosome translocation, and rearrangements. E. Transcriptome Analysis The transcriptome can be considered as the complete collection of transcribed elements of the genome. Broadly speaking, transcriptome mapping attempts to define regions of transcription, transcription factor binding sites, sites of chromatin modification, sites of DNA methylation, and chromosomal origins of replication. The tiling array covering both coding and noncoding regions has proven to be extremely valuable for transcriptome analysis. For example, probe sets on the normal microarrays used for gene expression profiling are based on information about known transcribed genes. This means that we can only analyze the genes that have already been isolated and that some genes that may have important functions but that have not yet been cloned will escape unnoticed. The tiling array will overcome this shortfall and detect novel genes or microRNAs with sequence information that is not yet available in expressed sequence tag databases. Additionally, tiling arrays can detect alternatively spliced variants of genes. An interesting finding that has emerged from transcriptome mapping is that large regions of the genome beyond the coding segments are often transcribed (Kapranov et al., 2002; Stolc et al., 2004; Katayama et al., 2005). This has provided new insights into the basic understanding of how transcriptional regulation may occur in animals and the function of so-called “junk” DNA. Scientists are currently pushing the limits of transcriptome research by building whole-genome tiling arrays that interrogate the genome at resolutions at the levels of individual nucleotides.

APPLICATIONS OF DNA MICROARRAY TECHNOLOGY

41

F. Chromatin Immunoprecipitation Analysis ChIP detects interactions between specific transcription factors and the regions of genomic DNA to which they bind (Horak and Snyder, 2004; Ren and Dynlacht, 2004). The early generation of microarrays for ChIP analysis was produced by PCR-amplifying promoter regions of limited sets of genes. This powerful technology has been rapidly evolving after the completion of various genome projects. Current promoter arrays can range from spotting synthetic oligonucleotide arrays covering the upstream and downstream of tens of thousands of promoter regions to tiling arrays covering the entire genome at nucleotide resolutions. For ChIP analysis, proteins are crosslinked to DNA by treating whole cells (or nuclei) with formaldehyde, and this is followed by the fragmentation of the protein-bound DNA by sonication or endonuclease digestion (Figure 2.4). Antibodies against a specific chromatin-associated protein are then used to immunoprecipitate protein–DNA complexes. This results in enrichment for fragments bound to the immunoprecipitated proteins. These fragments are amplified, and fluorescent-labeled probes are generated both from an experimental and a control sample; they are then hybridized to the promoter-array chips. For control, a parallel experiment can be performed after omitting the

FIGURE 2.4 ChIP-on-chip analysis. A, The ChIP-on-chip method aims at determining the DNA sequences to which given proteins are bound. These binding sites may indicate the functions of various transcriptional regulators. B and C, Cells are treated with formaldehyde to immobilize protein–DNA complexes, which are later sheared by sonication and precipitated using specific antibodies against the protein of interest. D, After the removal of crosslinked samples, the DNA sample is subjected to PCR amplification. ChIP-enriched DNA and control genomic DNA samples are independently labeled with fluorescent dyes (Cy5 and Cy3). For a single-color scan, two different DNA pools are hybridized to two separate whole-genome tiling arrays, and the data are compared. For a two-color analysis, two differentially labeled samples are combined and hybridized to a single array. The DNA sequences enriched are thereby identified by hybridization on DNA chips. (See color insert.)

42

FINDING GENE EXPRESSION CHANGES USING MICROARRAY TECHNOLOGY

antibody or using nonspecific antibodies. By comparing signals between experimental and control (reference) probes, the enrichment of specifically bound fragments can be directly detected. The bound fragments are expected to reside adjacent to or within genes that are regulated by the specific transcription factor. Genome-wide ChIP-on-chip analysis is gaining popularity, and it has been successfully used to identify target genes regulated by individual transcription factors or in combination. For example, using ChIP-on-chip analyses of MyoD, Myogenin, and MEF2, it was possible to successfully construct a blueprint of a gene regulatory network for myogenic differentiation (Blais et al., 2005). Likewise, target gene networks have been identified for HNF transcription factors for hepatocyte and pancreas differentiation and for Oct4, Nanog, and Polycomb for stem cell differentiation (Odom et al., 2004; Lee et al., 2006; Loh et al., 2006). ChIP experiments were also employed to identify different histone modifications associated with active as compared with inactive chromatin (Bernstein et al., 2004). The key reagent for the success of this approach is the access to high-quality antibodies specific for a protein of interest.

III. CONCLUDING REMARKS Microarray technology is valuable in the gene expression profiling of a variety of cell types. Although the technology has been predominantly used to study gene expression patterns, refined approaches enable scientists to address increasingly complex questions and to perform more intricate experiments. We envision that the data obtained from microarray work will become essential for the detection of various diseases and developmental abnormalities and for inferring the probable functions of new genes. Furthermore, applications of microarray to developmental biology will enlarge our understanding of the gene regulatory programs that govern embryogenesis, oncogenesis, and diseases that affect humans.

SUMMARY

 Microarrays are experimental tools that were originally designed to mea  

sure the levels of transcripts in different cells in response to experimental manipulation. Microarrays allow us to simultaneously examine the expression of thousands of genes, thereby permitting the linking of previously unsuspected genes to particular developmental processes and diseases. Microarrays are useful for the study of DNA variants, with the primary applications being the identification and genotyping of mutations and polymorphisms. The availability of whole genome tiling arrays allows us to detect SNPs, mutations, and chromosome abnormalities and to perform transcriptome mapping and ChIP-on-chip analyses.

ACKNOWLEDGMENTS We thank Drs. Ira Blitz and Bruce Blumberg for their helpful discussions.

43

REFERENCES

GLOSSARY OF TERMS ChIP-on-chip Chromatin immunoprecipitation (ChIP) is an experimental approach for the identification of the transcription factors associated with specific regions of the genome. Chromatin associated with a specific protein or proteins is precipitated using appropriate antibodies. The enrichment of the DNA fragments relative to the input is measured at each genomic location on the basis of array hybridization. Comparative genomic hybridization Comparative genomic hybridization measures DNA copy number differences between two different samples using a spotted microarray or a high-density tiling array. Microarrays The technique explores the ability of DNA or RNA molecules to hybridize specifically to DNA-probe templates spotted on a glass slide or a quartz wafer. In a single experiment, the expression levels of hundreds or thousands of genes within a cell can be measured to determine the amount of mRNA bound to each site on the array. Single nucleotide polymorphism (SNP) A SNP is a small genetic change that occurs within a person’s DNA sequence. SNPs may fall within coding sequences of genes, noncoding regions of genes, or in the intergenic regions between genes.

REFERENCES Aitman TJ, Glazier AM, Wallace CA, et al: Identification of Cd36 (Fat) as an insulin-resistance gene causing defective fatty acid and glucose metabolism in hypertensive rats, Nat Genet 21:76–83, 1999. Baldi P, Hatfield GW: DNA microarrays and gene expression: from experiments to data analysis and modeling, Cambridge, UK, 2002, Cambridge University Press. Bammler T, Beyer RP, Bhattacharya S, et al: Standardizing global gene expression analysis between laboratories and across platforms, Nat Methods 2:351–356, 2005. Bernstein BE, Mikkelsen TS, Xie X, et al: A bivalent chromatin structure marks key developmental genes in embryonic stem cells, Dev Cell 6:145–155, 2004. Bishop JO, Morton JG, Rosbach M, et al: Three abundance classes in HeLa cell messenger RNA, Nature 250:199–204, 1974. Blais A, Tsikitis M, Acosta-Alvear D, et al: An initial blueprint for myogenic differentiation, Genes Dev 19:553–569, 2005. Brazma A, Hingamp P, Quackenbush J, et al: Minimum information about a microarray experiment (MIAME)-toward standards for microarray data, Nat Genet 29:365–371, 2001. Bullido MJ, Artiga MJ, Recuero M, et al: A polymorphism in the regulatory region of APOE associated with risk for Alzheimer’s dementia, Nat Genet 18:69–71, 1998. Churchill GA: Fundamentals of experimental design for cDNA microarrays, Nat Genet 32 (Suppl):490–495, 2002. Davidson EH, Britten RJ: Regulation of gene expression: possible role of repetitive sequences, Science 204:1052–1059, 1979. Eisen M, Brown P: DNA arrays for analysis of gene expression, Methods Enzymol 303:179–205, 1999. Freedman ML, Pearce CL, Penney KL, et al: Systematic evaluation of genetic variation at the androgen receptor locus and risk of prostate cancer in a multiethnic cohort study, Am J Hum Genet 76:82–90, 2005.

44

FINDING GENE EXPRESSION CHANGES USING MICROARRAY TECHNOLOGY

Horak CE, Snyder M: ChIP-chip: a genomic approach for identifying transcription factor binding sites, Methods Enzymol 350:469–483, 2002. Ishkanian AS, Malloff CA, Watson SK, et al: A tiling resolution DNA microarray with complete coverage of the human genome, Nat Genet 36:299–303, 2004. Kapranov P, Cawley SE, Drenkow J, et al: Large-scale transcriptional activity in chromosomes 21 and 22, Science 296:916–919, 2002. Katayama S, Tomaru Y, Kasukawa T, et al: Antisense transcription in the mammalian transcriptome, Science 309:1564–1566, 2005. Lee TI, Jenner RG, Boyer LA, et al: Control of developmental regulators by Polycomb in human embryonic stem cells, Cell 21:301–313, 2006. Lipshutz RJ, Fodor SP, Gingeras TR, et al: High density synthetic oligonucleotide arrays, Nat Genet 21:20–24, 1999. Lockhart DJ, Dong H, Byrne MC, et al: Expression monitoring by hybridization to high-density oligonucleotide arrays, Nat Biotechnol 14:1675–1680, 1996. Loh YH, Wu Q, Chew JL, et al: The Oct4 and Nanog transcription network regulates pluripotency in mouse embryonic stem cells, Nat Genet 38:431–440, 2006. Long AD, Mangalam HJ, Chan BY, et al: Improved statistical inference from DNA microarray data using analysis of variance and a Bayesian statistical framework. Analysis of global gene expression in Escherichia coli K12, J Biol Chem 276:19937–19944, 2001. Mantripragada KK, Buckley PG, de Stahl TD, et al: Genomic microarrays in the spotlight, Trends Genet 20:87–94, 2004. Odom DT, Zizlsperger N, Gordon DB, et al: Control of pancreas and liver gene expression by HNF transcription factors, Science 303:1378–1381, 2004. Piano F, Schetter AJ, Morton DG, et al: Gene clustering based on RNAi phenotypes of ovaryenriched genes in C. elegans, Curr Biol 12:1959–1964, 2002. Puffenberger EG, Hu-Lince D, Parod JM, et al: Mapping of sudden infant death with dysgenesis of the testes syndrome (SIDDT) by a SNP genome scan and identification of TSPYL loss of function, Proc Natl Acad Sci U S A 101:11689–11694, 2004. Quackenbush J: Computational analysis of microarray data, Nat Rev Genet 2:418–427, 2001. Ren B, Dynlacht BD: Use of chromatin immunoprecipitation assays in genome-wide location analysis of mammalian transcription factors, Methods Enzymol 376:304–315, 2004. Roepman P, Wessels LF, Kettelarij N, et al: An expression profile for diagnosis of lymph node metastases from primary head and neck squamous cell carcinomas, Nat Genet 37:182–816, 2005. Schena M: Microarray biochip technology, Natick, MA, 2000, Bio Techniques Books. Schena M, Shalon D, Davis RW, et al: Quantitative monitoring of gene expression patterns with a complementary DNA microarray, Science 270:467–470, 1995. Stoeckert CJ Jr, Causton HC, Ball CA: Microarray databases: standards and ontologies, Nat Genet 32(Suppl):469–473, 2002. Stolc V, Gauhar Z, Mason C, et al: A gene expression map for the euchromatic genome of Drosophila melanogaster, Science 306:655–660, 2004. van ’t Veer LJ, Dai H, van de Vijver MJ, et al: Gene expression profiling predicts clinical outcome of breast cancer, Nature 415:530–536, 2002. Yu Y, Khan J, Khanna C, et al: Expression profiling identifies the cytoskeletal organizer ezrin and the developmental homeoprotein Six-1 as key metastatic regulators, Nat Med 10:175–181, 2004.

3

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION WEN XIONG and SHENG DING The Scripps Research Institute Department of Chemistry and the Skaggs Institute for Chemical Biology, La Jolla, CA

INTRODUCTION Regeneration—a complex process involving the restoration of cells, tissues, and structures that are lost or damaged during disease, injury, or aging— is more commonly observed among the lower organisms (e.g., amphibians) than it is among mammals. Nevertheless, the regenerative process typically involves stem or progenitor cells. Stem cells are cells that can self-renew to produce themselves and differentiate to generate lineage-restricted progenies. They have the remarkable potential to develop into essentially all of the cell types found in the organism. Therefore, in addition to their potential therapeutic value, they provide excellent model systems for basic studies of cell behaviors. Tremendous efforts have been concentrated on the identification, isolation, and characterization of embryonic and tissue-specific stem cells in various organisms. More recently, the concept of cancer stem cells has been implicated and strengthened in a number of tumor types, and knowledge of stem cell biology may also provide new therapeutic strategies to cure cancers. Despite the enormous progress made so far, stem cell isolation, the maintenance of self-renewal, and directed differentiation remain challenging. Consequently, the realization of their therapeutic potential would require an improved ability to control their fate and a better understanding of the precise molecular mechanisms underlying their proliferation, differentiation, migration, and survival at the systems level. The completion of genome sequencing projects in several organisms in conjunction with advances in high-throughput

Principles of Developmental Genetics © 2007, Elsevier Inc. All rights reserved.

45

46

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

technologies has allowed studies at the systems level. Genome-wide approaches have been emerging for various biological investigations, and they should ultimately provide us with precise information about the abundance and modification state of all molecules in a cell at a given time or under a certain condition. A small molecule approach has historically proved to be a useful tool for modulating cell fate and probing the underlying molecular mechanisms. The application of chemical and genomic approaches in stem cells will greatly advance our understanding of fundamental questions in stem cell and developmental biology, and it may ultimately facilitate the development of novel therapeutic strategies to treat human diseases or to stimulate tissue or organ regeneration in vivo. This chapter will focus on the current advances in the chemical and functional genomic approaches in stem cell research.

I. OVERVIEW OF STEM CELLS A. Definitions of Various Types of Stem Cells Stem cells are cells that have the ability to self-renew for long periods of time and to differentiate into specialized cell types in response to appropriate signals. Traditionally, stem cells are classified as either embryonic or tissue specific. Embryonic stem cells (ESCs) are typically derived from the inner cell mass of the blastocyst. They possess an unlimited capacity for self-renewal, and they have the potential (i.e., pluripotency) to develop into any cell types found in the three primary germ layers of the embryo (i.e., endoderm, mesoderm, and ectoderm), as well as germ cells and extraembryonic cells (Hubner et al., 2003; Toyooka et al., 2003; Geijsen et al., 2004). By contrast, tissuespecific stem cells are multipotent, and they are found in differentiated tissues. They are capable of self-renewal, but they generally can only differentiate into restricted cell types of the tissue from which they originate. These cells are believed to function as the “reservoir” for cell and tissue renewal during homeostasis or tissue regeneration. The most studied tissue-specific stem cells include hematopoietic stem cells (HSCs), mesenchymal stem cells (MSCs), neural stem cells (NSCs), epidermal stem cells, and skeletal muscle stem cells. HSCs are probably the best-characterized somatic stem cells. They can be prospectively isolated from bone marrow, but they are difficult to expand without differentiation in vitro. HSCs have the capacity to provide the lifelong reconstitution of all blood–cell lineages after transplantation. Although the in vivo origin of MSCs remains elusive, MSCs are multipotent progenitor cells that can be isolated from multiple tissues, expanded substantially in vitro, and differentiated into a variety of cell types, including osteocytes, adipocytes, chondrocytes, skeletal muscle cells, and neurons (Pittenger et al., 1999; Dezawa et al., 2004; 2005). The discoveries of NSCs in the adult central nervous system (CNS) and their regenerative roles in brain damage have suggested approaches involving cell replacement therapy and the stimulation of in vivo regeneration for the treatment of neurodegenerative diseases and CNS injury. In contrast with HSCs, NSCs can be expanded in the presence of growth factors (e.g., basic fibroblast growth factor [bFGF]), but the noninvasive isolation and purification of significant numbers of NSCs from the brain remain challenging.

OVERVIEW OF STEM CELLS

47

More recent evidence supports a longstanding notion that cancers are initiated and maintained by cancer stem or progenitor cells (Beachy et al., 2004). There is a growing body of evidence that suggests a close relationship between normal stem cells and cancer stem cells: the self-renewal mechanisms of normal stem cells and cancer stem cells are similar; the deregulation of the developmental signaling pathways involved in stem cell self-renewal is associated with oncogenesis; and tumors contain “cancer stem cells” that may arise from normal stem cells or progenitor cells via cellular transformations (Pardal et al., 2003). A better understanding of stem cell regulation will not only provide novel therapeutic approaches to tissue regeneration, but it may also help identify the molecular triggers and potential cures for various types of cancers. B. Stem Cell Behaviors and Functions As a result of their unique properties of self-renewal and differentiation into various mature cell types, stem cells not only provide great opportunities for studying tissue and organ development, but they also hold important therapeutic potentials for regenerative medicine. Encouraging therapeutic examples include the transplantation of HSCs or specific neural stem and progenitor cells for the treatment of hematologic diseases or Parkinson disease in humans, respectively. More efficient and highly selective methods of controlling stem cell fate for producing homogenous populations of particular cell types will be essential for the therapeutic use of stem cells; this will facilitate studies of the molecular mechanisms of development. Stem cell fate is determined by both intrinsic regulators and the extracellular microenvironment (niche). It is postulated in one model that, to maintain the homeostasis in vivo, the stem cell undergoes asymmetric division, and the daughter cell that is proximal (i.e., still attached) to the “niche” remains undifferentiated whereas the one distal to the niche differentiates into a specific cell type (Kiger et al., 2001; Tulina and Matunis, 2001). After stem cells are isolated from where they reside and cultured in vitro outside of the “niche,” they are likely to differentiate spontaneously into a heterogeneous population of cell types unless instructed by specific signals. Therefore, it is a great challenge to maintain stem cell self-renewal in culture and to induce homogenous lineage-specific differentiation. Stem cell expansion and differentiation ex vivo are generally controlled by culturing the cells in a specific configuration, either attached monolayer or suspended aggregates, with “cocktails” of growth factors, signaling molecules, or genetic manipulations. However, most of these conditions are either incompletely defined or nonspecific with regard to regulating the desired cellular process. Undefined conditions often result in an inconsistency in cell culture and heterogeneous populations of cells that would not be useful for cell-based therapy and that would complicate the biological study of a particular cellular process. Consequently, current challenges of stem cell research remain in two areas: first, the lack of precise and highly selective methods for homogeneous stem cell self-renewal and differentiation into specific cell types, and, second, the lack of a complete understanding of these processes at the molecular level. Advances in both aspects will enable better control of stem cell fate and thus facilitate clinical applications. Chemical and genomic approaches—including genomics, functional genomics, and proteomics—have been undertaken to address the above challenges.

48

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

II. GENOMIC APPROACHES IN STEM CELL RESEARCH A. Genome-Wide Expression Analysis Any given state of a cell can be perceived as the phenotypic display of a unique gene expression pattern of that cell. Global expression analysis at the mRNA level, which is also referred to as transcriptome profiling, is a widely used method to assign molecular signatures to a given cell population or tissue. In addition, profile comparison between two distinct cell populations can identify the key molecules that contribute to the cell type differences or to a specific cellular process (e.g., self-renewal, differentiation). The main purposes of stem cell transcriptome profiling are twofold: to identify the signature genes that define the cell type (e.g., stemness) and to identify the genes that can maintain stem cell self-renewal or direct its differentiation to a specific cell type of interest. Three major methods of transcriptome analysis—microarrays, short sequence tags, and expressed sequence tags—have been successfully applied in stem cells to discover signature and essential genes through the systematic comparison of profiles among different cell types and species (Fortunel et al., 2003; Sato et al., 2003; Sperger et al., 2003; Brandenberger et al., 2004; Ng et al., 2005). More recently, chromatin immunoprecipitation on a chip platform has been developed. This technology goes beyond gene expression to explore gene regulation and to determine the precise binding sites of proteins on DNA sequences (Boyer et al., 2005). As compared with the traditional one-gene-at-a-time approach, the transcriptome profiling approach greatly speeds up gene discovery, and it provides new ways of thinking about biological questions. However, there are several limitations associated with genome-wide expression analysis: 1. 2. 3.

4.

5.

False positives and negatives could be mixed in every transcriptome data set as a result of experimental, technical, or analytical flaws; It is difficult to draw a causal relationship between a gene’s expression and a phenotype; Follow-up studies of transcriptome profiling, including loss- and gain-offunction studies, typically still require the conventional one-gene-at-atime approach; Gene regulation achieved at the posttranscriptional, translational, and posttranslational levels also play significant roles in a cellular event, and these are not reflected in the transcriptome profiling; and The dynamics of protein function affected by cellular localization, degradation rate, protein–protein interactions, and so on can only be determined at the protein level.

Additional approaches, including functional genomics and proteomics (as described later) are needed to complement the transcriptome genomics approach. B. Functional Genomics Functional genomics aims not only to determine the complex roles of vertebrate genes during development but also to screen for molecules that are involved in a biological process on a genome-wide scale. There are two main approaches in the functional genomics that are currently applied to stem cell research: gene trap and high-throughput screens. Gene trap was designed for

GENOMIC APPROACHES IN STEM CELL RESEARCH

49

the systematic knockout of gene function on a genomic scale to determine the roles of mammalian genes during embryonic and postembryonic development, but recently it has also been modified for functional screens during stem cell development. A more important advancement, however, has been the application of high-throughput screens, in which functions of a genome-scale collection of genes are simultaneously evaluated in a biological process. 1. Gene Trap Gene trap is a form of random intragenic insertional mutagenesis that was designed to perturb gene function. There are promoter and enhancer trap and polyadenylation trap systems. An important feature of gene trap is that the disrupted gene can be easily identified by the rapid amplification of cDNA ends. As compared with the targeted gene inactivation approach based on homologous recombination, gene trap is simple, rapid, and cost-effective. However, as a result of the diploid nature of the mouse genome, recessive mutations created by gene trap cannot elicit phenotypic consequences to reflect the function of trapped genes unless homozygous transgenic mice are generated. A sophisticated system has been developed recently to circumvent this problem (Guo et al., 2004). Taking advantage of the highly efficient mitotic recombination in blm-deficient ESCs, a genome-wide library of homozygous mutant cells was generated by gene trap, which enabled a direct phenotypic genetic screen in ESCs. Often, a reporter gene is also included in the trapping vector to capture the endogenous expression pattern of the disrupted gene. ESC clones carrying traceable insertional mutations can be assayed in vitro for reporter gene activity under various cell lineage specification conditions, which allows for the identification of developmentally regulated genes. A collection of mouse mutations has been generated by gene trap and organized in a searchable database (To et al., 2004). To facilitate the functional categorization of trapped genes, a responder mouse ESC (mESC) line carrying a dominant selection marker has been used (Chen et al., 2004). Recently, gene trap has also been carried out in human ESCs (hESCs; Dhara and Benvenisty, 2004) and adult rodent neural progenitor cells (Scheel et al., 2005). 2. High-Throughput Screening Technologies Conventional genome-scale functional screens have been performed largely on libraries of pooled cDNA clones. As a result of the complexity of such libraries and the need to oversample them to find the rare clones, the assay is often limited by exceedingly simple readouts (which is typically a selection method), and it requires the subsequent deconvolution of clone identities. A pluripotency-associated master gene, Nanog, has been discovered using this technique (Mitsui et al., 2003). With recent advances in automation and detection technologies for high-throughput screening and the development of individually arrayed molecular libraries (cDNAs, siRNAs, miRNA, or small molecules, each of which has members that are spatially separated in different wells of multiwell plates), more complex assays can be used for real functional screens instead of selection, such as monitoring morphological changes or dynamic cellular events without the need for clone rescue and deconvolution. In contrast with the typical random mutagenesis screen, the high-throughput functional screen has several advantages: first, the identities of the hits are already known; second, the genome can be saturated; and

50

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

third, the maintenance and replication of the screen are much more convenient and efficient. There are generally two types of high-throughput functional screens based on the assay used: reporter activity-based screens and high content phenotype-based screens. Hits identified are further characterized using traditional biochemical and cellular methods. 3. Large-Scale Genetic Approaches: cDNA and RNAi Libraries Perturbation of gene function can be achieved by overexpression (cDNA) or gene knockdown. RNA interference (RNAi), which is a highly conserved gene-silencing event that functions through the targeted destruction of the individual mRNA with the introduction of a homologous double-stranded RNA (dsRNA), is a powerful tool to knock down gene expression. Both vector-encoded short hairpin RNAs (shRNAs) and chemically synthesized double-stranded siRNAs have been demonstrated to be effective RNAi tools (Elbashir et al., 2001; Paddison et al., 2004). Screens using both cDNA and shRNA libraries have been successfully carried out in mammalian cells (Elbashir et al., 2001; Michiels et al., 2002; Paddison et al., 2004;). Recently, proof-of-concept screens using an arrayed synthetic siRNA library targeting more than 5000 human genes have been carried out in hMSCs to identify the endogenous repressors of osteogenic or adipogenic specification, which, upon silencing, could initiate the differentiation of hMSCs into osteoblasts or adipocytes, respectively. Such screens yielded a large number of novel hits and provided a foundation for studying the genetic network that controls self-renewal, the osteogenesis and adipogenesis of hMSCs, and, potentially, the molecular rationale for treating certain bone and metabolic diseases. However, the cost and effort required to generate the arrayed libraries and the availability of screening technology have been constraints. The development of cell microarrays, which use a microarray format to substitute for the multiwell plate format, could potentially drive down the cost of highthroughput studies; however, this technique is still in its infancy (Wheeler et al., 2005). In addition, there are significant technical challenges to applying the arrayed high-throughput screening technologies in primary cells and stem cells, which are more difficult to transfect and susceptible to side effects when transfection methods are used. Gene delivery methods based on viral transduction may provide an alternative solution to this problem (Michiels et al., 2002; Berns et al., 2004). Furthermore, on the basis of the identified putative “stemness” genes or differentially regulated genes during stem cell differentiation through transcriptome profiling, selected cDNA or siRNA clones could be generated and collected to systematically evaluate the function of these hits in relevant biologic contexts. Although high-throughput screening is a highly productive and promising technique to complement transcriptome profiling, perturbation of a single molecule may often not be sufficient to induce a particular biologic event. In this case, sensitized screen conditions will need to be sought out. Moreover, the careful design and validation of constructed libraries and screen assays will be essential, especially in the case of RNAi, because off-target effects would be hard to pursue. C. Proteomic Technologies and Mass Spectrometry There has been tremendous interest in developing and applying protein profiling technologies for examining protein–protein interactions and protein

GENOMIC APPROACHES IN STEM CELL RESEARCH

51

activities. More conventional tools (e.g., yeast two-hybrid and protein microarray systems) are limited as a result of sample analysis in artificial environments and their inability to identify interactions that are stabilized by more than two partners. Mass spectrometry (MS) enabled by new instrumentation coupled with various protein-separation techniques has emerged as a driving force in proteomics for analyzing protein abundance and modifications. Although protein tagging and pull-down followed by MS present a generic approach for the analysis of protein complexes, the two-dimensional difference gel electrophoresis (2D-DIGE) separation of proteins labeled with fluorescent dyes and the capillary separation of proteins labeled with isotopes (SILAC and ICAT) allow for comparative protein quantification by MS. 1. Fluorescent-Dye–Labeled 2D-DIGE Two-dimensional electrophoresis (2-DE) separates complex mixtures of proteins (e.g., whole-cell lysates) according to isoelectric point and molecular mass. 2-DE data reflect changes in protein expression level, isoforms, and posttranslational modifications. Advances in prefractionation methods and the application of narrow-overlapping immobilized pH gradients yield a greater resolution of protein spots and sensitize the detection of low-abundance proteins, as demonstrated in the attempt at developing a comprehensive proteomic analysis of undifferentiated murine R1 ESCs using 2-DE contigs (Elliott et al., 2004). Of the proteins resolved from the 2-DE contigs, a large proportion was identified as DNA repair enzymes in addition to ribosomal, transcriptional, and translational proteins. These findings may reflect the properties of ESCs to resist DNA damage while maintaining the undifferentiated state and to quickly change phenotype as seen during differentiation. Quantification data from 2-D gels initially relied on the intergel comparison of sample populations using traditional stains (e.g., silver, Coomassie blue) with low dynamic ranges. Intragel comparisons can now be accomplished with the assistance of Cy dyes, which confer greater detection sensitivity and allow for the multiplexing of samples. Because fluorescent Cy dyes have the same chemical reactivity but distinct excitation and emission spectra, multiple samples can be run on the same gel, and relative protein abundance can be assessed from differential fluorescence intensities. Using this technique, a comparative proteomic analysis between two hESC populations with differential motility was performed (Evans et al., 2004). The greatest strength of 2-DE is its ability to distinguish proteins with varying posttranslational modifications (e.g., phosphorylation, glycosylation, ubiquination). Advancements in protein separation and labeling in 2-DE technology combined with MS have enabled the generation of large-scale proteome profiles to occur and provided a more reliable means of collecting relative protein expression data. Detection sensitization will require further improvements in staining sensitivity and sample preparation to unmask the “unseen proteome.” 2. Capillary Separation with Isotope Labeling 2D-DIGE has been the traditional means for quantitative proteomic data collection, but the limitations of protein compatibility with gel electrophoresis have motivated efforts to develop capillary-based separation techniques, such as capillary liquid chromatography and capillary electrophoresis. The benefits

52

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

of capillary-based separation over 2-DE are its dynamic range and sensitivity. 2-DE requires large amounts of the sample to visually detect protein spots, whereas capillary-based separation requires a very small amount of sample with greater detection sensitivity. Recently, a large-scale identification of protein expression in mESCs was carried out (Nagano et al., 2005). Using automated 2D liquid chromatography coupled with MS analysis, proteins— including transcription factors characteristic of ESC and those previously reported as ESC-specific or “stemness” genes—were identified. New techniques that have granted quantitative information to be derived from capillary-based separation are isotope-coded affinity tag (ICAT) and stable isotope labeling by amino acids in cell culture (SILAC; Gygi et al., 1999; Ong et al., 2002). The difference between ICAT and SILAC is the means by which isotopes are added to the protein mixtures. In SILAC, mammalian cells are cultured in media that lack a standard essential amino acid but that are supplemented with a nonradioactive isotopically labeled form of that amino acid. Over time, the labeled amino acid is completely incorporated into the proteome. In ICAT, isotopes are chemically added to samples by covalent attachment to each cysteinyl residue. After chemical incorporation, ICAT samples require an additional affinity step to collect only labeled proteins for MS analysis. SILAC is a simple, cost-effective approach to quantitative proteomics, but it requires cells to be alive and cultured until the isotope is completely incorporated. Alternatively, ICAT can be applied to both living and dead cells, but it depends on the presence of cysteinyl residues. After the isotopes have been completely incorporated into samples either via ICAT or SILAC, samples labeled with various isotopes are first mixed together and then analyzed together with MS. MS data give not only multiple peaks per peptide (corresponding with the relative heavy and light isotope samples), but they also give comparative protein abundance that is based on peak amplitudes. This technology can be used to observe differences between different developmental stages (Kratchmarova et al., 2005). SILAC with MS was employed to comprehensively compare proteins that were tyrosine phosphorylated in response to epidermal growth factor and platelet-derived growth factor (PDGF) for the purpose of deriving the basis of the differential induction of hMSCs into bone-forming cells by epidermal growth factor but not by PDGF. Although the types of signaling proteins that are modified during stimulation by both ligands largely overlapped, the phosphatidylinositol 3-kinase pathway was exclusively activated by PDGF. Chemical inhibition of this pathway allowed for the PDGF-induced osteogenic differentiation of hMSCs. This work illustrates the ability of quantitative proteomics to discover critical differences that are capable of changing cell fate by directly comparing two differential MS protein profiles.

III. CHEMICAL TECHNOLOGIES IN STEM CELL STUDIES Cell-permeable small molecules that can modulate the function of specific proteins with exquisite precision provide convenient and efficient spatial and temporal control of gene function in a biological system, and they are powerful tools that complement genetic techniques. Small molecules, such as dexamethasone (a glucocorticoid receptor agonist), ascorbic acid, 5-azacytidine (5-aza-C; a DNA demethylation agent), trichostatin A (an histone deacetylase

CHEMICAL TECHNOLOGIES IN STEM CELL STUDIES

53

inhibitor), and all-trans retinoic acid (RA), have proven to be extremely useful for modulating and studying the differentiation of various stem cells. For example, studies with 5-aza-C, which induces the myogenic differentiation of a mouse mesenchymal progenitor cell line, led to the discovery of a master transcription factor, MyoD, that is responsible for skeletal myogenic fate determination. Although such chemical tools have historically been used to investigate biological systems, advancements in chemical synthesis, highthroughput screening, and molecular profiling technologies have rejuvenated chemical approaches in biology. Chemical libraries composed of millions of discrete compounds can be efficiently generated, assembled, and “mined” through high-throughput functional screens. Screening for compounds that generate a desired phenotype in cells or animals and then characterize their mechanism of action may thus serve as an alternative approach for identifying key players in a biological process. A. Rational Design of Chemical Libraries and Combinatorial Technology 1. Chemical Libraries One approach to generating functional small molecules that control stem cell fate involves the use of phenotypic or pathway-specific screens of synthetic chemical or natural product libraries. The size and diversity of a given purified chemical library as well as the selection method determine the chance of finding a desired “hit” compound. With recent advances in automation and detection technologies, millions of discrete compounds can be screened rapidly and cost-effectively. However, although combinatorial technologies allow for the synthesis of a large number of molecules with immense structural diversity, it is impossible to saturate the chemical space. Because the diversity of chemical libraries is largely constrained by the synthetic tractability, new synthetic technologies are the driving force to expand current chemical diversities for filling the chemical space. In addition, introducing a high level of structural variability to increase the molecular diversity of a chemical library drastically reduces the average fitness of the library to a given biologica selection or screen, thereby resulting in most molecules being inactive (analogous to population genetics). Consequently, the careful design of a chemical library becomes a critical aspect of combinatorial synthesis (diversity vs. fitness). The concept of “privileged structures” describes selected structural motifs that can provide potent and selective ligands for multiple biologic targets by introducing different substitutions onto the same scaffold. Privileged structures typically exhibit good “drug-like” properties, such as good solubility, membrane permeability, oral bioavailability, and metabolic stability, which make the further development of “hits” into “leads” less problematic. Given the success of privileged structures, the diversification of these scaffolds using combinatorial techniques provides not only large numbers of compounds but also highly enriched “functional” molecules. Using key biologic recognition motifs as the core scaffolds may represent one of the most straightforward and productive ways to generate “privileged” chemical libraries. Previously, we developed the concept of using privileged molecular scaffolds themselves as a diversity element for combinatorial synthesis (Ding et al., 2001; 2002a; 2002b; Wu et al., 2001) to maximize the diversity while retaining a minimal threshold of fitness to biological screens. With this approach, a variety of

54

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

naturally occurring and synthetic heterocycles that are known to interact with proteins involved in cell signaling (e.g., kinases, cell surface and nuclear receptors, enzymes) were used as the core molecular scaffolds. These included substituted purines, pyrimidines, indoles, quinazolines, pyrazines, pyrrolopyrimidines, pyrazolopyrimidines, phthalazines, pyridazines, pyridines, triazines, and quinoxalines (the first diversity elements). A general synthetic scheme (Figure 3.1) was then developed that could be used in parallel reactions to introduce a variety of substituents into each of these scaffolds to create a diverse chemical library. The library synthesis involved introducing a second diversity element into these heterocyclic scaffolds using solution-phase alkylation or acylation reactions. This was followed by the capture of the modified heterocycles onto solid support using different immobilized amines (introduced as a third diversity element). The resin-bound heterocycles could then be further modified (introduced as a fourth diversity element) through a variety of chemistries, including acylation, amination, and palladium-mediated cross-coupling reactions with amines, anilines, phenols, and boronic acids. Using these chemistries in conjunction with the “directed-sorting” method, we have generated diverse heterocycle libraries consisting of more than 100,000 discrete small molecules (representing more than 30 distinct structural classes), with an average purity of greater than 90%. These libraries have been proven to be a rich source of biologically active small molecules targeting various proteins involved in a variety of signaling pathways. 2. High-Throughput Screens To systematically identify small molecules that can generate a cellular phenotype of interest, high-throughput screens of these large and diverse chemical libraries are carried out in a desired model system, such as a cell line (Figure 3.2) or a simple organism (e.g., Xenopus, zebrafish) with an appropriate readout, such as luminescence (e.g., a luciferase reporter), fluorescence (e.g., an enhanced green fluorescence protein [eGFP] reporter), or absorbance (e.g., enzymatic reactions to generate chromophores). However, such assays only provide limited information, and they require a battery of secondary assays to determine the precise cellular pathways or processes being affected. With recent advances in high-content imaging technologies (e.g., autofocus and image analysis), high-resolution microscopy/image-based screens allow for the capture of multiple parameters from a single reading at the single-cell level, thereby facilitating the identification of molecules with a desired biologic activity. 3. Lead Optimizations Typically, the initial hits from the primary cellular screens may not be ideal for probing their mechanism of action (MOA) using affinity-based and functional genomic approaches or for serving as clean research tools in biological assays both in vitro and in vivo. Consequently, there is a need to improve the hit compound’s properties (e.g., potency, specificity, solubility, bioavailability) via detailed structure-and-activity-relationship studies. These involve reiterated rounds of testing structurally related compounds that are modified via medicinal chemistry around substituent and scaffold. Such studies may also identify a linkage position on the molecule for attachment to a solid support without adversely affecting its activity (for affinity pull-down assay).

55

CHEMICAL TECHNOLOGIES IN STEM CELL STUDIES

OMe CHO

O

R3 NH2

A

OMe R3

O

R1

N OMe H

OMe

A Cl

R2

OMe

R1 Cl chloroheterocycle

"R2"

Cl

O N

R1

R3

R2

OMe

R4

R4

R4

R5

N R1

HN R3

R1

HN R3

R2

R2

Cl

R1

chloroheterocycle

X

N H

X

X = Cl, Y = H X = H, Y = Cl X = Cl, Y = Br

Cl

X

N Cl X=H X = NH2

Cl

Cl

N

Cl

R6

Cl N Cl

N

R6

X

X = Cl, Y = H X = H, Y = Cl X = COOH, Y = H X = Cl, Y = COOH

Cl R7

N H

N H

COOH

COOH

X

X Z

N

Cl

Y

X = S, Y = H, Z = Cl X = S, Y = Cl, Z = H X = NH2, Y = Cl, Z = H X = NH2, Y = H, Z = Cl

Cl NCO

R6

N

Br N

Cl

Cl N

Cl

R6

R6

N

Y N

X

Y = Z = A = B = H, X = C l X = Z = A = B = H, Y = C l X = Y = A = B = H, Z = C l X = Y = Z = B = H, A = C l X = Y = Z = A = H, B = C l

N N

Cl

N

N

B

Cl N N

R6

Y

N

A

X = Cl X = NH2

Cl N

R6 N

R2

Cl

Z

N

Y N

Y

Cl N

N

=

R2

R5

N R1

HN R3

R1

HN R3

Cl

A

R4

O

Y

R6

X = Br, Y = H, X = H, Y = Cl,

X = Cl, Y = H, X = H, Y = Cl,

Br

N

N N

N

X

N H

FIGURE 3.1 Combinatorial synthesis of heterocyclic library.

Cl

N NH Cl

Br

S N

N Cl

56

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

R3

Self-renew

X R5

N R1

X N

N

Differentiation

N X

R4

R2

De-differentiation Stem cells

R1 = Heterocycles X = O, N, S, C

FIGURE 3.2

Differentiated cells

Cell-based high throughput screening. (See color insert.)

4. Target and Pathway Identifications The target identification of bioactive small molecules remains one of the major challenges in the field of molecular pharmacology. Various systemic approaches have been developed and used to facilitate target and mechanism determination, including affinity chromatography, genome-wide mRNA expression analysis, proteomic profiling by mass spectroscopy or protein array, and large-scale gene complementation assays using arrayed cDNA and siRNA libraries. Among these approaches, conventional affinity pulldown using small-molecule–immobilized solid matrixes is still the most straightforward and unbiased biochemical approach in target identification. However, new tricks have recently been implemented (Burdine and Kodadek, 2004), such as affinity linker optimization, blocking/soaking nonspecific binding sites in the matrix, reverse affinity chromatography (using a generic form of small-molecule–affinity resin to capture interested proteome and applying the specific small molecule to elute and compete with the compound targets in a dose-dependent manner), combinatorial chromatographic analysis with multiple optimized positive and negative affinity resins, and in situ affinity labeling. DNA microarrays have been extensively used to generate global gene expression profiles in which differentially and temporally modulated gene clusters may reveal primary responding signaling pathways. Expression profiling in conjunction with systematic pathway analysis has been shown to provide useful information in small-molecule target and signaling identification, validation, and MOA studies. Additional emerging tools, such as proteomic profiling, add another dimension of information to dissecting the small molecules’ MOA. Furthermore, large-scale complementation assay using spatially addressable arrayed cDNA (Michiels et al., 2002; Carpenter and Sabatini, 2004; Huang et al., 2004) and small interfering RNA (siRNA) (Elbashir et al., 2001; Aza-Blanc et al., 2003; Zheng et al., 2004) libraries have been explored for the functional identification of proteins involved in the compoundtargeted pathway. The fundamental notion of this technique is that the effect of a given bioactive molecule can be modulated (shift of the dose–response curve) by the overexpression or suppression of a particular set of genes that is involved in the signaling pathway, which the drug targets. Such an assay has been widely used in yeast and mammalian systems as described in the functional genomic approaches in the previous section. Although one particular approach discussed previously may be more informative than others for revealing the target and pathway of a specific compound, information combined from different approaches may ultimately shed light on the MOA of the small molecule. Most importantly, genes,

CHEMICAL TECHNOLOGIES IN STEM CELL STUDIES

57

signaling pathways, and other knowledge gained from these systematic analyses may serve as new entry points for additional investigations of the biologic phenomenon of interest. B. Small-Molecule Regulators of Stem Cell Fate 1. Self-Renewal Sustained stem cell self-renewal requires the combined forces of proliferation, the inhibition of differentiation, and the prevention of apoptosis. ESCs are conventionally maintained on feeder cells or in mixtures of exogenous factors. mESCs can be expanded in the pluripotent state in a defined medium supplemented with leukemia inhibitory factor (LIF) and bone morphogenetic protein (BMP), whereas the self-renewal of hESCs requires bFGF. Interestingly, 6-bromoindirubin-3’-oxime (BIO), a natural product derived from mollusk Tyrian purple, has been shown to maintain mESCs in the undifferentiated state in the conventional serum-containing media without feeder cells and LIF (Sato et al., 2004). BIO was proposed to function by inhibiting glycogen synthase kinase 3 (GSK3) and activating the canonical Wnt signaling pathway, although the precise mechanism and relevance of canonical Wnt signaling to ESC self-renewal remain to be determined. In mESCs, BMP functions by activating transcription factor Id through Smad. In addition, BMP inhibits both extracellular signal-regulated kinase (ERK) and p38 mitogen-activated protein kinases (MAPKs), which have been shown to be negative regulators for mESC self-renewal. Consistent with those findings, both MAPK/ERK kinase (MEK) inhibitor PD98059 (Burdon et al., 1999) and p38 inhibitor SB203580 (Qi et al., 2004) have positive effects on the promotion of the self-renewal of mESC. Recently, an imaging-based high-throughput chemical screen that combined both gene expression (Oct4, a pluripotency marker) and morphologic (undifferentiated ESCs grow as compact colonies) analyses was carried out in an Oct4-GFP reporter mESC line to identify small molecules that control the self-renewal of mESCs. A novel pyrimidine derivative, pluripotin, was discovered, and it is sufficient to propagate mESCs in the pluripotent state under chemically defined conditions in the absence of feeder cells, serum, and LIF. Long-term pluripotin-expanded mESCs can be differentiated into cells in the three primary germ layers in vitro, and they can also generate chimeric mice and contribute to the germ line in vivo. Interestingly, pluripotin does not operate through the known signaling pathways (i.e., LIF, BMP, and Wnt) that control the self-renewal of mESCs. Affinity chromatography using a pluripotin-immobilized matrix identified ERK1 and RasGAP as the molecular targets of pluripotin. Additional biochemical and genetic experiments suggest that pluripotin is a dual-function small-molecule inhibitor of both ERK1 and RasGAP and that the simultaneous inhibition of both protein activities is necessary and sufficient for pluripotin’s effects on mESCs. ERK activation has been implicated in the differentiation of mESCs. Consequently, the inhibition of ERK1 by pluripotin would be expected to contribute to the self-renewal of mESCs. RasGAP modulates Ras signaling by stimulating the guanosine triphosphate (GTP)–hydrolysis activity of Ras to form the inactive RasGDP complex. By inhibiting RasGAP, pluripotin may activate signaling by Ras or Ras-like GTPases, which in turn may enhance self-renewal through phosphatidylinositol 3-kinase or other signaling pathways. In addition to serving as a useful chemical tool for the control of the self-renewal

58

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

of ESCs, pluripotin also represents a novel class of dual inhibitors of a protein kinase and a small GTPase activating protein (Chen et al. 2006). This discovery provided the first evidence that one small molecule is sufficient to sustain long-term self-renewal of mouse ESCs by modulating more than one signaling pathways. Traditionally, ESCs are maintained in self-renewal by mitogens and cytokines through activation of plasma membrane associated receptors. Since expression of stem cell specific genes is active at the specific developmental stage when ESCs are isolated and established in culture, we believe the activation of these pathways functions mostly to inhibit various differentiation potentials and balance cells in a multipotent self-renewal state. The balance is achieved by cross suppression of the lineage specific differentiation related genes. Our results indicated that the activation of these receptors is not absolutely required. Cell membrane permeable small molecules can bypass the upstream molecules and directly activate the downstream signaling, and thus maintain ESCs in a self-renewal state. Such discovery may provide new insights into the self-renewal mechanism and facilitate the practical applications of ESCs in research and therapy. 2. Lineage-Specific Differentiation The most commonly used method for inducing the differentiation of ESCs involves growing them in suspension (in the presence of serum and the absence of supplemented LIF) to form aggregates called embryoid bodies (EBs), which begin to differentiate into various cell lineages, including hematopoietic, endothelial, neuronal, and cardiac muscle cells. However, such uncontrolled differentiation is a poorly defined, inefficient, and relatively nonselective process, and it therefore leads to heterogeneous populations of differentiated and undifferentiated cells, which are not useful for cell-base therapy and which also complicate biologic studies of a particular differentiation program. Consequently, dissecting stem cell signaling pathways and identifying critical factors that are involved in tissue specification are essential for the development of stem cell therapy and related small-molecule therapeutics. A number of small molecules have been identified that modulate specific differentiation pathways of embryonic or adult stem cells. a. Neural and neuronal differentiation RA is a widely used small molecule for the neural and neuronal differentiation of ESCs and neural cells. The effect of RA is dose and developmental stage dependent. It was recently demonstrated that subtype-specific neurons can be generated from mouse and human ESCs in a stepwise fashion. For example, to generate motor neurons, mESCs were first neuralized through EB formation with concomitant RA treatment. The generated neural cells were further caudalized by RA, and this was followed by treatment with a specific small molecule agonist (Hh-Ag1.3) of Sonic hedgehog signaling to ventralize the caudalized neural cells to become the desired motor neurons. This experiment suggests that multiple sequential signals, a combination of signals, or both may be required to generate a terminally differentiated, subtype-specific cell type. TWS119, a synthetic disubstituted pyrrolopyrimidine, was recently identified from a reporter-based screen as a potent inducer of neuronal differentiation in pluripotent mESCs and P19 murine embryonal carcinoma cells (Ding et al., 2003). A panel of affinity matrices prepared from

CHEMICAL TECHNOLOGIES IN STEM CELL STUDIES

59

representative TWS analogs were used to pull-down target proteins from P19 cell extracts. Proteins specifically bound to all positive resins derived from active molecules but not to the negative resins derived from inactive molecules were considered to be the putative targets of TWS119. Consequently, GSK3b was identified as one target of TWS119 and confirmed by additional biochemical and cellular assays. This target identification may provide yet another link between neuronal differentiation and the Wnt signaling pathway. Additional studies also indicated that TWS119 (like BIO) is not entirely specific against GSK3b. Alternatively, TWS119 might promote neuronal differentiation of mESCs via novel mechanisms other than the canonical Wnt signaling pathway. Such mechanisms might include the compound’s inhibition of other proteins that were not apparent in the affinity experiments (possibly as a result of low abundance or other factors) or cross-talk of the Wnt pathway with other signals present in the media. Neuropathiazol (a substituted 4-aminothiazole) was recently identified from a high-content imaging-based screen of chemical libraries that specifically induces the neuronal differentiation of multipotent adult hippocampal neural progenitor cells. Treatment of the neural progenitor cells with neuropathiazol significantly slowed cell proliferation without visible cytotoxic effects; more than 90% of cells differentiated into neuronal cells as determined by immunostaining with bIII tubulin and the characteristic neuronal morphology (Warashina et al., 2006). In addition, reverse transcription polymerase chain reaction (RT-PCR) of marker genes showed that Sox2 (a neural progenitor marker) was downregulated and that NeuroD1 (a neuronal differentiation marker) was upregulated after treatment with neuropathiazol. Interestingly, neuropathiazol can also inhibit the astroglial differentiation induced by LIF and BMP2 whereas RA cannot, which suggests that neuropathiazol functions by a different mechanism and that it has more specific neurogenicinducing activity than RA. In addition to the unbiased screening approach, modulating defined molecular targets that have been implicated in self-renewal and the differentiation of neural stem and progenitor cells by selective small molecules has provided a rationalized means of generating desired cell types in a controlled manner. For example, Hedgehog (Hh) pathway agonists were used to promote the proliferation of adult hippocampal neural progenitors, whereas histone deacetylase inhibitors were shown to specifically induce their differentiation into neurons. b. Cardiomyogenic differentiation The mammalian adult heart, like the brain, is mainly composed of postmitotic and terminally differentiated cells. Although there is evidence suggesting a resident population of self-renewing cardiac stem cells that is able to contribute to heart repair, the scarcity of these cells and their intrinsically poor regenerative response to heart injury remain obstacles for their therapeutic application. Alternatively, pluripotent ESCs represent a possible unlimited source of functional cardiomyocytes. A recent study has shown that hESCderived cardiomyocytes can form structural and electromechanical connections with a primary culture of neonatal rat ventricular myocytes in vitro and pace the heart of pigs that had complete heart block in vivo, suggesting that hESC-derived cardiomyocytes may act as potential rate-responsive biologic pacemakers for myocardial repair (Kehat et al., 2004).

60

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

The cardiomyogenesis of ESCs can occur after EB differentiation in vitro, but such a condition is very inefficient and nonspecific. Consequently, the development of new approaches for the directed differentiation of ESCs into cardiomyocytes will facilitate the therapeutic applications of ESCs and increase our understanding of the molecular mechanism underlying cardiomyocyte differentiation and heart development. Using mESCs that are stably transfected with the cardiac-muscle–specific myosin heavy chain promoterdriven eGFP, ascorbic acid (vitamin C) was identified from a screen of known drugs that can enhance the cardiac differentiation of mESCs in the monolayer culture (Takahashi et al., 2003). Interestingly, other antioxidants (e.g., Nacetylcysteine, vitamin E) do not have a similar effect, which suggests that the cardiomyogenesis-inducing activity of ascorbic acid may be independent of its antioxidative property. A similar screening strategy has identified cardiogenol (a substituted diaminopyrimidine) from large combinatorial chemical libraries as a compound that can selectively and efficiently induce the differentiation of mESCs to cardiomyocytes in monolayer cultures (Wu et al., 2004a). The differentiated cells expressed multiple cardiac muscle markers, including GATA-4, Nkx2.5, MEF2, and myosin heavy chain, and the differentiated culture formed large areas of spontaneous contracting patches. c. Differentiation of mesenchymal stem and progenitor cells MSCs are multipotent cells with significant cellular plasticity. They can differentiate into a variety of mesenchymal tissues, such as osteocytes, adipocytes, and chondrocytes; and they can also differentiate into other tissue types, such as neuronal and skeletal muscle cells, under specific differentiation conditions. A number of small molecules have been found that can be used to control the differentiation of mesenchymal stem or progenitor cells for a variety of applications. For example, 5-aza-C (a DNA demethylation chemical) can induce C3H10T1/2 cells (a mouse mesenchymal progenitor cell line) to differentiate into myoblasts, osteoblasts, adipocytes, and chondrocytes. 5-Aza-C does not directly activate a specific differentiation program, but rather it converts the cells into a competent differentiation state. Dexamethasone (a glucocorticoid receptor agonist), ascorbic acid, b-glycerophosphate, isobutylmethylxanthine (a nonspecific phosphodiesterase inhibitor), and peroxisome proliferator-activated receptor g gonists (e.g., rosiglitazone) have been widely used to modulate the osteogenesis or adipogenesis of MSCs under specific conditions. Interestingly, treatment with a JAK inhibitor (WHI-P131) followed by trophic factor induction was recently shown to be able to convert rat MSCs into neuronal cells (Dezawa et al., 2004). Purmorphamine, a 2,6,9trisubstituted purine compound, was identified as a potent osteogenic-differentiation–inducing molecule through a high-throughput chemical screen in C3H10T1/2 cells (Wichterle et al., 2002). Genome-wide expression profiling in conjunction with systematic pathway analysis was used to reveal that the Hh signaling pathway is the primary affected biologic network and that purmorphamine is a selective Hh pathway agonist. This was further confirmed by chemical epistasis using two different Hh pathway antagonists: cyclopamine, which binds and inhibits Smoothened (Smo), and forskolin, which activates protein kinase A to convert Gli proteins to transcriptional repressors by phosphorylation (Wu et al., 2004b). Additional biochemical assays have indicated that purmorphamine targets Smo (Sinha and Chen, 2006).

CHEMICAL TECHNOLOGIES IN STEM CELL STUDIES

61

3. Proliferation Reactivation and Dedifferentiation Terminally differentiated postmitotic mammalian cells are thought to have little or no regenerative capacity, because they are already committed to their final specialized form and function, and they have permanently exited the cell cycle. Their inability to regenerate (i.e., to divide and replace damaged tissue) may constitute a biomedical problem. Consequently, the stimulation of adult postmitotic cells to reenter the cell cycle and proliferate may provide new therapeutic approaches for treating degenerative diseases and injuries. Mammalian cardiomyocytes remain proliferative during fetal development. Shortly after birth, the cell-cycle–perpetuating machinery shuts down, and cardiomyocytes lose their proliferative capacity. p38 MAPK was identified as a key negative regulator of mammalian cardiomyocyte proliferation through the regulating genes required for mitosis (including cyclin A and cyclin B). Recently, it was reported that a p38 inhibitor, SB203580, increased the growth-factor–induced S-phase progression and mitosis in both neonatal and adult cardiomyocytes indicated by BrdU incorporation and histone 3 phosphorylation. The proliferation in adult cardiomyocytes was also associated with the transient dedifferentiation of the contractile apparatus (Engel et al., 2005). A longstanding notion in developmental biology has been that organ- and tissue-specific stem cells are restricted to differentiating into cell types of the tissue in which they reside. However, recent studies suggest that tissue-specific stem or progenitor cells may overcome their intrinsic lineage restriction after exposure to a specific set of signals in vitro and in vivo, although such reprogramming may not reflect potentials that are normally exercised in vivo. An extreme example is the reprogramming of a somatic cell to a totipotent state by nuclear transfer during which the nucleus of a somatic cell is transferred into an enucleated oocyte. The ability to dedifferentiate or reverse lineagecommitted cells back to multipotent or even pluripotent cells might overcome many of the obstacles associated with using ESCs and nonautologous stem cells in clinical applications. However, the cellular processes involved in dedifferentiation remain poorly understood, and methods for the control and study of dedifferentiation are lacking. To identify small molecules that can induce the dedifferentiation of C2C12 myoblasts, an assay was designed based on the notion that lineagereversed myoblasts should regain multipotency, which is the ability to differentiate into multiple mesenchymal cell lineages under conditions that typically only induce the differentiation of multipotent MSCs into adipocytes, osteoblasts, or chondrocytes. Reversine, a 2,6-disubstituted purine, was found to have the desired dedifferentiation inducing activity: it inhibits the myotube formation of C2C12 myoblasts, and only reversine-treated myoblasts can efficiently differentiate into osteoblasts and adipocytes after exposure to the appropriate differentiation conditions. Importantly, the efficient dedifferentiation effect of reversine on C2C12 cells can be shown at the clonal level, which suggests that this effect is inductive rather than selective. Furthermore, reversine appears to have similar effects on several other primary and established cell lines, which suggests that its mechanism may be general in cellular reprogramming. Affinity chromatography and other cellular studies revealed that the mechanism of reversine’s action is twofold: 1) to stage cells at a specific phase in cell cycle by interfering with a cell cycle regulator; and 2) to enhance

62

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

cytoskeletal rearrangement and growth-factor–induced reprogramming by targeting proteins involved in both pathways. This example is a proof-ofprinciple demonstration that the dedifferentiation of lineage-restricted cells to a more primitive (multipotent) state by a synthetic chemical can be achieved via a rationally designed phenotypic screen of combinatorial chemical libraries and that such concepts and technologies are readily applicable to other models. C. Regeneration Screens Regeneration screens have been carried out at both the cellular and wholeorganismal levels to identify the small molecules and genes involved in the regenerative process. Inhibitory molecules associated with myelin and the glial scar limit axon regeneration in the adult CNS, but the underlying mechanism of such regeneration inhibition is not fully understood. A small molecule screen to search for compounds that can neutralize neurite outgrowth inhibitory activity associated with the CNS myelin identified several epidermal growth factor receptor kinase inhibitors. These compounds showed a remarkable ability to counteract the effects of myelin inhibition, and they promoted the significant axon regeneration of injured optic nerve fibers, which points to a promising therapeutic avenue for enhancing axon regeneration after CNS injury (Koprivica et al., 2005). In a recent study that examined zebrafish fin regeneration, disorganized mesenchymal cells beneath the amputation plane were observed, and this was followed by cell proliferation, migration, and blastema formation, which provided evidence of dedifferentiation. The blastema is a mass of undifferentiated mesenchymal cells that have proliferated beyond the amputation plane to drive fin regrowth. In a screen of the N-ethyl-N-nitrosourea–induced mutants that fail in fin regeneration, fgf20a null mutants were identified, which indicates that fgf20a is required for regeneration-specific blastema formation and subsequent fin regeneration (Makino et al., 2005). Osteoporosis and diseases of bone loss are a major health problem associated with aging. The bisphosphonates have been widely used for the treatment of osteoporosis by inhibiting bone absorption. However, there are no agents that promote bone formation. To facilitate the identification of novel anabolic molecules, a high-throughput in vivo screen using larval zebrafish has been performed in a 6-day time period. Vitamin D3 analogs and intermittent parathyroid hormone were shown to cause dose-dependent increases in the formation of mineralized bone (Fleming et al., 2005). This fast, economical, and genetically tractable screening system provides a powerful adjunct to mammalian models for the identification of bone anabolic agents, and it offers the potential for the genetic elucidation of the pathways involved in osteoblastic activity. D. Chemical Screens of Pathways Fundamental developmental signaling pathways (e.g., Wnt, Hh, BMP, Notch; see Chapter 1), which control embryonic patterning and cell behaviors, play important roles in stem cell regulation. The deregulation of these pathways in either the embryonic or adult stage may result in diseases, such as cancer and degenerative disease.

CONCLUDING REMARKS

63

In addition to compounds that were discovered via other means (e.g., purmorphamine and cyclopamine-Hh pathway agonist and antagonist, respectively; BIO and sulindac-Wnt pathway agonist and antagonist, respectively; DAPT-Notch pathway antagonist), cell-based pathway-specific screens have also been used to identify the small-molecule regulators of these developmental pathways. For example, a series of Hh-pathway–specific agonists and antagonists have been identified through screens of synthetic compounds using 10T1/2 cells stably transfected with a plasmid containing a luciferase reporter downstream of multimerized Gli binding sites and a minimal promoter. These molecules have been used in various applications ranging from stem cell proliferation and differentiation to the induction of apoptosis in cancer cells. By carrying out Wnt pathway-specific screens of a chemical library using the TOPflash reporter assay, a 2-amino-4,6-disubstituted pyrimidine compound was identified that activates Wnt signaling in a dose-dependent manner. Interestingly, this compound does not inhibit GSK-3b, which is a major inhibitory component in the pathway; however, its activity could be blocked by a dominant negative T cell factor 4 (TCF4), which suggests that it functions upstream of the known TCF factors on the canonical Wnt signaling pathway. Importantly, this compound appears to mimic the effects of Wnt ligand in a Xenopus model, which suggests that it may be a useful tool to study physiologic processes that involve Wnt signaling. From a Wnt3a sensitized screen, a 2,6,9-trisubstituted purine compound was identified as a Wnt synergistic agonist. Consistent with its in vitro activity, this purine compound—in combination with a suboptimal dose of XWnt8— induces axis duplication in Xenopus, with high penetrance. Affinity chromatography study in conjunction with genetic confirmations has identified a GTPase activating protein of ADP-ribosylation factor (ARF-GAP) as a target of this purine compound. ARF family GTPases are best known for their function in vesicle transport. By inhibiting ARF-GAP, this purine compound activates ARF, which leads to increased levels of the cytoplasmic b-catenin; however, only in the presence of Wnt pathway activation is such translocated b-catenin stabilized to further activate downstream transcription. This mechanism may provide one explanation for the function of this purine compound as a synergistic agonist of Wnt signaling.

IV. CONCLUDING REMARKS Stem cell biology is a fast-growing field that offers new opportunities for the treatment of many devastating diseases and that provides new insights into the molecular mechanisms that control developmental processes. However, there remain significant obstacles that must be overcome before the therapeutic potential of stem cells can be realized. This requires a better understanding of the signaling pathways that control stem cell fate and an improved ability to manipulate stem cell proliferation and differentiation. Functional genomic and proteomic studies at the systems level are very likely to yield insights into the fundamental molecular mechanisms underlying stem cell fate determination. Although the small-molecule approach has been practiced in drug discovery and used in probing biology for decades, its value in the stem cell field is just now beginning to be realized (Figure 3.3). However, many challenges

64

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION O Me

HO

OH OH Me

Me H

N O

N

HO

H

F

HO

O Dexamethasone

OH

NH2

Me

Me

Me

Me

N O

OH H OH

HO

Me

5-azacytidine

OH

O

O

CO2H

All-trans retinoic acid

Ascorbic acid (Vitamin C)

O N

OCH3

O

Me

Cl

H

N

Me H

S

H

N

Me

O N H Me H

H

N

N

HO HN

O Me Me Me O OH OAc Me MeH OH HO

NH

N

N

O

CH3

Hh-Ag 1.3

Cyclopamine

Forskolin

Purmorphamine

O

Br

HO

OH NH 2 N H

S

N O CH3

OCH3

O CH3 CH3

N H

OH

H3C O

H N

N

N CH3

N

H N

F O

N

N O CH3

Pluripotin

FIGURE 3.3

O

O N CH3 H

S

O O

OEt

N

DAPT

N CH3

Neuropathiazol OH

NH2 O

N N H CH3

WHI-P131

F

OCH3

N H

N N

CH3O

IBMX

N

N H

HN CH3O

SB 203580

CH3

N

N

Reversine

O S CH3

N

PD 98059

TSA

N

N H

H N

NH2 O

O

H3C

N

F

O

N

N

OH N

NH O

N

Cardiogenol C

O

Rosiglitazone

H3C N CH3

N H

O BIO

O

NH

O

N

NH

N H

TWS119

N

NH

H3CO

N N

HO

N

N CF3

NH

N N H

OCH3 Wnt agonist

O O

N

N N H

N

N

Wnt synergistic agonist

Small compounds modulating stem cell fate.

remain, including designing better chemical libraries and screening strategies to systematically identify the small molecules that regulate the desired cellular process, developing more efficient methods to understand the underlying mechanisms, and translating in vitro discoveries into approaches for the in vivo regeneration of desired tissues and organs by small-molecule therapeutics. Nonetheless, it is clear that the identification of additional small molecules that control stem cell fate will significantly facilitate the studies of stem cell and developmental biology and contribute to the development of regenerative medicine.

GLOSSARY OF TERMS

65

SUMMARY

 Embryonic stem cells are pluripotent cells that have the potential to differ-







entiate into essentially all cell types in the organism, whereas tissue-specific adult stem cells are multipotent with the restricted potential to differentiate into certain specific cell types. It remains challenging to maintain stem cell self-renewal or to direct lineagespecific differentiation in a homogenous fashion. Methods for the precise control of stem cell fate will not only allow for the generation of desirable cells for cell-based therapy, but they will also provide excellent systems for studying the underlying mechanisms that control such processes. The realization of stem cells’ therapeutic potential will require an improved ability to control their fate and a better understanding of the precise molecular mechanisms underlying their proliferation, differentiation, migration, and survival at the systems level. The application of chemical and genomic approaches in stem cells will greatly advance our understanding of fundamental questions in stem cell and developmental biology, and it may ultimately facilitate the development of novel therapeutic strategies to treat human diseases and to stimulate tissue and organ regeneration in vivo.

GLOSSARY OF TERMS Functional genomics Functions of a genome-scale collection of genes are simultaneously evaluated in a biologic process. High-throughput screening Recent advances in automation and detection technologies allow for the evaluation of a large number of genes or small molecules at the same time in a certain biologic process. MOA Mechanism of action studies of a small molecule, including target and/or pathway identification by various methods. Proteomics Protein profiling technologies (e.g., mass spectrometry) for examining protein– protein interactions and/or protein modification and activities. Structure-and-activity-relationship studies Studies of a small molecule involving reiterated rounds of testing structurally related compounds modified via medicinal chemistry around substituent and scaffold. Self-renewal The symmetric or asymmetric division of a stem cell into two identical or different daughter cells (with one identical to the parental cell) to replicate itself and maintain the potential for differentiation. It requires combined forces of promotion of proliferation, inhibition of differentiation, and prevention of apoptosis.

66

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

SILAC Stable isotope labeling by amino acids in cell culture. It is a simple and costeffective approach to quantitative proteomics. Stem cells Cells that have the ability to self-renew for long periods of time and to differentiate into specialized cell types in response to appropriate signals, including pluripotent embryonic stem cells and multipotent adult (tissuespecific) stem cells.

REFERENCES Aza-Blanc P, Cooper CL, Wagner K, et al: Identification of modulators of TRAIL-induced apoptosis via RNAi-based phenotypic screening, Mol Cell 12(3):627–637, 2003. Beachy PA, Karhadkar SS, Berman DM: Tissue repair and stem cell renewal in carcinogenesis, Nature 432(7015):324–331, 2004. Berns K, Hijmans EM, Mullenders J, et al: A large-scale RNAi screen in human cells identifies new components of the p53 pathway, Nature 428(6981):431–437, 2004. Boyer LA, Lee TI, Cole MF, et al: Core transcriptional regulatory circuitry in human embryonic stem cells, Cell 122(6):947–956, 2005. Brandenberger R, Wei H, Zhang S, et al: Transcriptome characterization elucidates signaling networks that control human ES cell growth and differentiation, Nat Biotechnol 22(6):707–716, 2004. Burdine L, Kodadek T: Target identification in chemical genetics: the (often) missing link, Chem Biol 11(5):593–597, 2004. Burdon T, Stracey C, Chambers I, et al: Suppression of SHP-2 and ERK signalling promotes selfrenewal of mouse embryonic stem cells, Dev Biol 210(1):30–43, 1999. Carpenter AE, Sabatini DM: Systematic genome-wide screens of gene function, Nat Rev Genet 5(1):11–22, 2004. Chen S, Do JT, Zhang Q, et al: Self-renewal of embryonic stem cells by a small molecule, Proc Natl Acad Sci U S A 103:17266–17271, 2006. Chen YT, Liu P, Bradley A: Inducible gene trapping with drug-selectable markers and Cre/loxP to identify developmentally regulated genes, Mol Cell Biol 24(22):9930–9941, 2004. Dezawa M, Ishikawa H, Itokazu Y, et al: Bone marrow stromal cells generate muscle cells and repair muscle degeneration, Science 309(5732):314–317, 2005. Dezawa M, Kanno H, Hoshino M, et al: Specific induction of neuronal cells from bone marrow stromal cells and application for autologous transplantation, J Clin Invest 113(12):1701–1710, 2004. Dhara SK, Benvenisty N: Gene trap as a tool for genome annotation and analysis of X chromosome inactivation in human embryonic stem cells, Nucleic Acids Res 32(13):3995–4002, 2004. Ding S, Gray NS, Ding Q, Schultz PG: A concise and traceless linker strategy toward combinatorial libraries of 2,6,9-substituted purines, J Org Chem 66(24):8273–8276, 2001. Ding S, Gray NS, Ding Q, et al: Resin-capture and release strategy toward combinatorial libraries of 2,6,9-substituted purines, J Comb Chem 4(2):183–186, 2002a. Ding S, Gray NS, Wu X, et al: A combinatorial scaffold approach toward kinase-directed heterocycle libraries, J Am Chem Soc 124(8):1594–1596, 2002b. Ding S, Wu TY, Brinker A, et al: Synthetic small molecules that control stem cell fate, Proc Natl Acad Sci U S A 100(13):7632–7637, 2003. Elbashir SM, Harborth J, Lendeckel W, et al: Duplexes of 21-nucleotide RNAs mediate RNA interference in cultured mammalian cells, Nature 411(6836):494–498, 2001. Elliott ST, Crider DG, Garnham CP, et al: Two-dimensional gel electrophoresis database of murine R1 embryonic stem cells, Proteomics 4(12):3813–3832, 2004. Engel FB, Schebesta M, Duong MT, et al: p38 MAP kinase inhibition enables proliferation of adult mammalian cardiomyocytes, Genes Dev 19(10):1175–1187, 2005. Evans CA, Tonge R, Blinco D, et al: Comparative proteomics of primitive hematopoietic cell populations reveals differences in expression of proteins regulating motility, Blood 103 (10):3751–3759, 2004.

REFERENCES

67 Fleming A, Sato M, Goldsmith P: High-throughput in vivo screening for bone anabolic compounds with zebrafish, J Biomol Screen 10(8):823–831, 2005. Fortunel NO, Otu HH, Ng HH, et al: Comment on “‘Stemness’: transcriptional profiling of embryonic and adult stem cells” and “a stem cell molecular signature,” Science 302 (5644):393, author reply 393, 2003. Geijsen N, Horoschak M, Kim K, et al: Derivation of embryonic germ cells and male gametes from embryonic stem cells, Nature 427(6970):148–154, 2004. Guo G, Wang W, Bradley A: Mismatch repair genes identified using genetic screens in Blmdeficient embryonic stem cells, Nature 429(6994):891–895, 2004. Gygi SP, Rist B, Gerber SA, et al: Quantitative analysis of complex protein mixtures using isotopecoded affinity tags, Nat Biotechnol 17(10):994–999, 1999. Huang Q, Raya A, DeJesus P, et al: Identification of p53 regulators by genome-wide functional analysis, Proc Natl Acad Sci U S A 101(10):3456–3461, 2004. Hubner K, Fuhrmann G, Christenson LK, et al: Derivation of oocytes from mouse embryonic stem cells, Science 300(5623):1251–1256, 2003. Kehat I, Khimovich L, Caspi O, et al: Electromechanical integration of cardiomyocytes derived from human embryonic stem cells, Nat Biotechnol 22(10):1282–1289, 2004. Kiger AA, Jones DL, Schulz C, et al: Stem cell self-renewal specified by JAK-STAT activation in response to a support cell cue, Science 294(5551):2542–2545, 2001. Koprivica V, Cho KS, Park JB, et al: EGFR activation mediates inhibition of axon regeneration by myelin and chondroitin sulfate proteoglycans, Science 310(5745):106–110, 2005. Kratchmarova I, Blagoev B, Haack-Sorensen M, et al: Mechanism of divergent growth factor effects in mesenchymal stem cell differentiation, Science 308(5727):1472–1477, 2005. Makino S, Whitehead GG, Lien CL, et al: Heat-shock protein 60 is required for blastema formation and maintenance during regeneration, Proc Natl Acad Sci U S A 102(41):14599–14604, 2005. Michiels F, van Es H, van Rompaey L, et al: Arrayed adenoviral expression libraries for functional screening, Nat Biotechnol 20(11):1154–1157, 2002. Mitsui K, Tokuzawa Y, Itoh H, et al: The homeoprotein Nanog is required for maintenance of pluripotency in mouse epiblast and ES cells, Cell 113(5):631–642, 2003. Nagano K, Taoka M, Yamauchi Y, et al: Large-scale identification of proteins expressed in mouse embryonic stem cells, Proteomics 5(5):1346–1361, 2005. Ng P, Wei CL, Sung WK, et al: Gene identification signature (GIS) analysis for transcriptome characterization and genome annotation, Nat Methods 2(2):105–111, 2005. Ong SE, Blagoev B, Kratchmarova I, et al: Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics, Mol Cell Proteomics 1 (5):376–386, 2002. Paddison PJ, Silva JM, Conklin DS, et al: A resource for large-scale RNA-interference-based screens in mammals, Nature 428(6981):427–431, 2004. Pardal R, Clarke MF, Morrison SJ: Applying the principles of stem-cell biology to cancer, Nat Rev Cancer 3(12):895–902, 2003. Pittenger MF, Mackay AM, Beck SC, et al: Multilineage potential of adult human mesenchymal stem cells, Science 284(5411):143–147, 1999. Qi X, Li TG, Hao J, et al: BMP4 supports self-renewal of embryonic stem cells by inhibiting mitogen-activated protein kinase pathways, Proc Natl Acad Sci U S A 101(16):6027–6032, 2004. Sato N, Meijer L, Skaltsounis L, et al: Maintenance of pluripotency in human and mouse embryonic stem cells through activation of Wnt signaling by a pharmacological GSK-3-specific inhibitor, Nat Med 10(1):55–63, 2004. Sato N, Sanjuan IM, Heke M, et al: Molecular signature of human embryonic stem cells and its comparison with the mouse, Dev Biol 260(2):404–413, 2003. Scheel JR, Ray J, Gage FH, Barlow C: Quantitative analysis of gene expression in living adult neural stem cells by gene trapping, Nat Methods 2(5):363–370, 2005. Sinha S, Chen JK: Purmorphamine activates the Hedgehog pathway by targeting Smoothened, Nat Chem Biol 2(1):29–30, 2006. Sperger JM, Chen X, Draper JS, et al: Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors, Proc Natl Acad Sci U S A 100(23):13350–13355, 2003. Takahashi T, Lord B, Schulze PC, et al: Ascorbic acid enhances differentiation of embryonic stem cells into cardiac myocytes, Circulation 107(14):1912–1916, 2003. To C, Epp T, Reid T, et al: The Centre for Modeling Human Disease Gene Trap resource, Nucleic Acids Res 32(Database issue):D557–D559, 2004.

68

CHEMICAL AND FUNCTIONAL GENOMIC APPROACHES TO STUDY STEM CELL BIOLOGY AND REGENERATION

Toyooka Y, Tsunekawa N, Akasu R, Noce T: Embryonic stem cells can form germ cells in vitro, Proc Natl Acad Sci U S A 100(20):11457–11462, 2003. Tulina N, Matunis E: Control of stem cell self-renewal in Drosophila spermatogenesis by JAKSTAT signaling, Science 294(5551):2546–2549, 2001. Warashina M, Min KH, Kuwabara T, et al: A synthetic small molecule that induces neuronal differentiation of adult hippocampal neural progenitor cells, Angew Chem Int Ed Engl 45 (4):591–593, 2006. Wheeler DB, Carpenter AE, Sabatini DM: Cell microarrays and RNA interference chip away at gene function, Nat Genet 37(Suppl):S25–S30, 2005. Wichterle H, Lieberam I, Porter JA, Jessell TM: Directed differentiation of embryonic stem cells into motor neurons, Cell 110(3):385–397, 2002. Wu TY, Ding S, Gray NS, Schultz PG: Solid-phase synthesis of 2,3,5-trisubstituted indoles, Org Lett 3(24):3827–3830, 2001. Wu X, Ding S, Ding Q, et al: Small molecules that induce cardiomyogenesis in embryonic stem cells, J Am Chem Soc 126(6):1590–1591, 2004a. Wu X, Walker J, Zhang J, et al: Purmorphamine induces osteogenesis by activation of the hedgehog signaling pathway, Chem Biol 11(9):1229–1238, 2004b. Zheng L, Liu J, Batalov S, et al: An approach to genomewide screens of expressed small interfering RNAs in mammalian cells, Proc Natl Acad Sci U S A 101(1):135–140, 2004.

FURTHER READING http://stemcells.nih.gov/stemcell/scireport.asp Stanford Genomic Resources http://genome-www.stanford.edu/ Science, Functional Genomics Home http://www.sciencemag.org/feature/plus/sfg/ FunctionalGenomics.org.uk http://www.functionalgenomics.org.uk/ Functional and Comparative Genomics: Human Genome Research in Progress http://www.ornl.gov/sci/techresources/Human_Genome/research/function.shtml Harvard Institute of Proteomics http://www.hip.harvard.edu/

4

ASSESSING NEURAL STEM CELL PROPERTIES USING LARGE-SCALE GENOMIC ANALYSIS SOOJUNG SHIN, JONATHAN D. CHESNUT, and MAHENDRA S. RAO Invitrogen Corporation, Carlsbad, CA

INTRODUCTION The nervous system is one of the earliest organ systems to differentiate from the blastula stage embryo. Neural stem cells (NSCs) are the cells in the nervous system that give rise to all neurons and supporting glial cells by symmetric and asymmetric divisions. This can be mimicked in culture, and NSCs can be derived from human embryonic stem cell (ESC) cultures over a period of 2 to 3 weeks (Reubinoff et al., 2001; Shin et al., 2006; Zhang et al., 2001). ESCs are the in vitro counterpart of the inner cell mass of the blastula-stage embryo, which gives rise to every component of our body. In vivo, the primitive neural tube forms by approximately the fourth week of gestation, and neurogenesis has commenced by the fifth week of development in humans (Kennea et al., 2002). At the time of neurulation or around the fourth week (see Chapter 12), the neuroectoderm segregates from the ectoderm by a process called neural induction. The initially formed neural plate then undergoes a stereotypic set of morphogenetic movements to form a hollow tube, which is comprised primarily of stem cells by a process called primary neurulation (Rao, 1999). The neural crest, which will form the peripheral nervous system, segregates from the central nervous system at this stage (see Chapter 26). Stem cells that will generate the central nervous system reside in the ventricular zone (VZ) throughout the rostrocaudal axis; they appear to be regionally specified; they proliferate at different rates; and they express different positional markers. The anterior neural tube undergoes a dramatic expansion and can be delineated into three primary vesicles: the forebrain (prosencephalon), the midbrain (mesencephalon), and the hindbrain (rhombencephalon). Differential growth and further segregation lead to the additional delineation of the prosencephalon into the telencephalon and the diencephalon and of the rhombencephalon into the metencephalon and the myelencephalon. The caudal neural tube does not undergo a similar expansion, but it does increase in size to parallel the growth of the embryo, and it undergoes further differentiation to form the spinal cord. The properties of VZ stem cells have been characterized (Rao, 2004; Schubert et al., 2000), and they appear to be homogenous, despite the acquisition of rostrocaudal and dorsoventral identity. Principles of Developmental Genetics © 2007, Elsevier Inc. All rights reserved.

69

ASSESSING NEURAL STEM CELL PROPERTIES USING LARGE-SCALE GENOMIC ANALYSIS

70

As development proceeds, the VZ is much reduced in size, and additional zones of mitotically active precursors can be identified. Mitotically active cells that accumulate adjacent to the VZ have been called subventricular zone (SVZ) cells. This SVZ is later called the subependymal zone as the VZ is reduced to a single layer of ependymal cells. The SVZ is prominent in the forebrain, and it can be identified as far back as the fourth ventricle. No SVZ is detectable in the more caudal regions of the brain, and, if it exists, it is likely a very small population of cells. An additional germinal matrix that is derived from the rhombic lip of the fourth ventricle, called the external granule layer, generates the granule cells of the cerebellum. Like the VZ, the SVZ can be divided into subdomains that express different rostrocaudal markers and that generate phenotypically distinct progeny. Distinct SVZ domains identified include the cortical SVZ, the medial ganglion eminence, and the lateral ganglion eminence. The proportion of SVZ stem cells declines with development, and, in the adult, multipotent stem cells are likely present only in regions of ongoing neurogenesis (e.g., the anterior SVZ, the SVZ underlying the hippocampus). At this stage, marker expression is relatively heterogeneous (Bernier et al., 2000; Doetsch et al., 1996; Pevny et al., 2003). Stem cells do not generate differentiated progeny directly, but rather they generate dividing populations of more restricted precursors that are analogous to the blast cells or restricted progenitors described in the hematopoietic lineages (Bedi et al., 1995; Katsura et al., 2001; Mujtaba et al., 1999; see Chapter 34). These precursors can divide and self-renew, but they are located in regions that are distinct from the stem cell population, and they can be distinguished from the stem cell population by the expression of cell surface and cytoplasmic markers (Cai et al., 2004b; Kalyani et al., 1997; Liu et al., 2004). Investigators have begun efforts to analyze stem cell populations (Table 4.1) using a variety of techniques with the idea that, by understanding

TABLE 4.1 Methods That Have Been Used to Characterize Neural Stem Cell Populations Authors

Cells Characterized

Method Used

Reference

Luo et al.

Neural stem cells and progenitor cells Neuroepithelial cells

Microarray

(Luo et al., 2002)

Subtractive suppression hybridization Immunohistochemistry

(Cai et al., 2004b)

Microarray Microarray

(Wright et al., 2003) (Abramova et al., 2005) (Geschwind et al., 2001) (Cai et al., 2006)

Cai et al. Liu et al. Svendsen Abramova et al. Geschwind et al. Cai et al. Miura et al. Brandenberger et al. Richards et al.

Astrocyte-restricted precursors Neural stem cells SSEA1-positive cells Central nervous system progenitors Neurosphere forming cells Embryonic stem cells Embryonic stem cells Embryonic stem cells

Microarray Massively parallel signature sequencing Massively parallel signature sequencing Expressed sequence tag scan Serial analysis of gene expression

(Liu et al., 2004)

(Miura et al., 2004a) (Brandenberger et al., 2004) (Richards et al., 2006)

THE IMPORTANCE OF A GLOBAL ANALYSIS AND THE CAVEATS WHEN COMPARING CELL SAMPLES

71

these populations and identifying the factors that regulate self-renewal and direct differentiation, one will be able to modulate the development and response of stem cells to environmental signals. Many of these approaches depend on large-scale analytic tools that rely on the comparison of purified populations of cells that differ with regard to their stage of development or their exposure to factors or that carry specific genetic abnormalities. In this chapter, we focus on general principles that should guide such an analysis, the techniques used to perform such an analysis, and how data mining efforts have provided important insights into the properties of NSCs.

I. THE IMPORTANCE OF A GLOBAL ANALYSIS AND THE CAVEATS WHEN COMPARING CELL SAMPLES The overall disposition of a cell depends on the steady state of a complex of interacting factors. The integration of these instructions occurs in the nucleus through combinations of signal-activated and tissue-restricted transcription factors binding to and controlling related enhancers or cis-regulatory modules of coexpressed genes. Additional regulation is provided by previously unappreciated epigenetic mechanisms such as histone modulation and CpG island methylation and by small untranslated RNA (microRNA). Thus, the response of the cell to any one perturbation is dependent on context, and it explains in part the conflicting results that have been reported. For example, the effect of Sonic hedgehog on NSCs is dependent on the presence or absence of fibroblast growth factor (Wechsler-Reya et al., 1999). Likewise, the response to bone morphogenetic protein depends on the density of the culture and the presence or absence of various regulatory genes (Rajan et al., 2003; Wilson et al., 2001). This context-dependent response suggests that the overall state of a cell needs to be understood before perturbation experiments are initiated so that consistent and meaningful analyses of the results can be obtained. Several variables remain poorly understood. For example, no distinction has been made between long-term self-renewing populations and short-term self-renewing populations. Although the evidence that stem cells age is quite clear (Shen et al., 1998; Svendsen, 2000), no analysis so far has taken into account the effects of aging, the acquisition of karyotypic abnormalities, the differences as a result of the acquisition of positional identity, or the differences between types of NSCs that are present during development. For example, radial glia type stem cells, transdifferentiated stem cell populations, VZ stem cells, SVZ-derived stem cells, and neurosphere-forming stem cells fulfill the criteria of NSC such as self-renewal and the ability to differentiate into neurons and glia, but the comparisons among them have not been fully understood. Two other observations have suggested that caution needs to be exercised as stem cell populations are analyzed. Stem cells propagated in culture stochastically differentiate, and, as such, they are invariably contaminated by various amounts of differentiated cells. For example, the proportion of stem cells in a neurosphere culture can vary from 1% to 2% to up to 100%. Notably, the largest contaminating populations are astrocytes and astrocyte precursors, which are dividing and expressing Nestin in culture. A confounding point is that these cells are difficult to distinguish from stem cells using the

72

ASSESSING NEURAL STEM CELL PROPERTIES USING LARGE-SCALE GENOMIC ANALYSIS

TABLE 4.2 Advantages and Disadvantages of Various Methods of Gene Expression Analysis Massively Parallel Signature Sequencing

Serial Analysis of Gene Expression

Microarray

Minimum amount of sample required Detection capacity Data presentation

20 mg of total RNA

1 mg of total RNA

1 mg of total RNA

2,000,000 Transcripts per million

100,000 Tag number detected

Sensitivity Data calibration and standardization Technical biases

þþþþþ Not required

þþþ Not required

48,000 Hybridization intensity using fluorescence þþ Required

Failure of sequencing reaction; duplication of the sequence

Failure of sequencing reaction; duplication of the sequence

High Long (10 weeks)

High Long (6 months)

Cost Turnaround time

Background hybridization from rivaling signal; signal overlapping; detector saturation Low Short

standard battery of tests. A second important observation made was that there are species differences between stem cells, and, thus, extrapolating from mouse to human is fraught with caveats (Barker et al., 2003; Ginis et al., 2004). The genome data sets for the two species are also not identical, thereby making cross-species comparisons difficult to interpret. Each of these differences will add variability to the results and make cross-laboratory comparisons difficult unless attention is paid to the quality of the sample and detailed information is provided regarding the time of isolation, the age at which the cells are isolated, the number of passages in culture, and the degree of contaminating cells present. These differences must be documented and taken into account when comparing data sets, because the noise from such variability can mask important critical differences between cell populations. In addition, when analyzing the cells, it is desirable to use a reliable and reproducible method that is cost-effective and sensitive enough to detect with high fidelity the global differences among populations of cells as well as the subtle differences introduced as the cells are propagated in the culture or as they mature. Although several different methodologies have been proposed, none of the cross-platform comparisons is very useful unless sophisticated normalizing algorithms and consideration of the technical variables inherent in largescale analyses are carefully considered (Table 4.2). Nevertheless, as improvements are made in the ability to obtain pure populations of cells, to harvest RNA from single cells or small amounts of tissue, to construct libraries, to sort cells, and to obtain high-quality genomic information, such large-scale analysis has become increasingly possible. However, it is recommended that analysis and comparisons be limited to one stage of development, in one species, with a single platform. The samples should be carefully examined for the presence of contaminating populations, and the degree of contamination should be assessed (Figure 4.1). This initial quality control will be critical for the yielding of useful results.

THE USE OF A REFERENCE STANDARD

73

FIGURE 4.1 Flowchart of techniques used to characterize cell populations using large-scale analysis. Samples for global analysis require verification in advance to ensure dependable data production. After the quality of the sample is controlled, the method of large-scale analysis that is most appropriate for the purpose needs to be determined. After the generated data set is processed and analyzed, an independent method is selected to confirm the acquired results.

II. THE USE OF A REFERENCE STANDARD When comparing large batches of data generated in different laboratories using different techniques or slightly different cell culture protocols, one must determine how best this can be done. Several strategies have been proposed, including the idea of a reference standard (Dybkaer et al., 2004; Novoradovskaya et al., 2004). This idea, although new, appears to be underappreciated in the stem cell field. However, most researchers have found it all but impossible to mine across data sets, because there are too many variables that need to be normalized and too many assumptions that need to be made. Furthermore, there are often circumstances in which one simply lacks the data to make any appropriate assumptions. In the ESC field, several strategies have been proposed (Loring et al., 2006). These include establishing a publicly available and well-curated data set that can be used as a ready reference, a set of standards that are readily available from a commercial or not-for-profit provider, or a control sample that all investigators can use as a standard. In principle, each of these could be applied to the NSC field, but, to our knowledge, no such common database exists as yet. Immortalized or cancer stem cell lines, such as C17.2, RT-4 or more recently identified cancer stem cell lines harvested from the appropriate species of interest, have been proposed as possible standards (Imada et al., 1978; Snyder et al., 1992; Steindler, 2002). However, it is important when using such lines as a reference to carefully assess the subclone that is being used. C17 subclones, for example, have shown remarkable variability, and diametrically opposite results have been reported, depending on the subclone used. The karyotype of this line is unstable, which may account for some of the differences seen. Nevertheless, because it has been so widely used, it could serve as a reference, provided that sufficient care was taken to use the same passage sample banked at American Type Culture Collection or some other responsible cell-banking facility. Fetal tissue samples from which pure populations of stem cells can be harvested at a defined stage of development in rodents may be an alternative choice for a reference standard. Many commercial entities provide such

74

ASSESSING NEURAL STEM CELL PROPERTIES USING LARGE-SCALE GENOMIC ANALYSIS

samples, and these could therefore become a de facto standard. The equivalent stages of development are not readily accessible in humans, however, and, as such, an alternative control will need to be considered. Sorting or negativeselection strategies have been shown to enrich for stem cell populations, and markers that define the stem cell stage are available; thus, a reference standard for human stem cell analysis could be considered. Alternatively, a publicly available and well-curated data set could be provided that could serve as a digital reference standard. In the absence of any of the above, we recommend obtaining RNA, genomic DNA, and microRNA from NTERA2. This is not an NSC line, but it readily differentiates into neurons and glia. It has been carefully analyzed by several groups, and labeled and unlabeled clones are both commercially available and available from American Type Culture Collection. The line is often used as a comparator for ESC work, and, thus, significant data on several different platforms are already available. However, these cells are not optimal when detailed high-resolution comparisons are required (Schwartz et al., 2005). In our laboratory, having an internal reference standard has proven invaluable in allowing us to compare NSCs to each other and to samples run at different times and to compare our results with those of several other colleagues without the necessity of repeating all of the experiments. These standards have also allowed us to check the quality of markers, our fluorescence activated cell sorting efficiencies, and the quality of our antibodies, and they have provided a basis for comparing across different laboratories. By running a reference sample in one laboratory and sending results to another laboratory, our experiments can be easily compared, and, over time, crossplatform comparisons also become possible (Figure 4.2). After the concept of a standard is accepted widely, commercial providers can provide RNA, DNA, and genomic material from such a reference that can

FIGURE 4.2 The importance of using a reference standard. The reference standard makes it possible to compare data sets from different slides, different time points, different laboratories, and even different methods.

75

METHODS OF ANALYSIS

be used as a comparator for all types of large-scale studies. In the ESC and microarray fields, groups have been established to determine such standards, and such uniformity has yielded useful results (Brazma et al., 2001; Husser et al., 2006; Wei et al., 2005), and we fully expect that a similar effort in the NSC field will be equally useful.

III. METHODS OF ANALYSIS The past several years have seen dramatic advances in technology and equally dramatic reductions in cost. Both genomic and proteomic methods are now available at costs that allow an average small laboratory to begin performing such experiments. The amount of material required for such an analysis has also become much less than was previously necessary, thereby making these experimental approaches feasible even when the number of stem cells available is limited. For example, single-nucleotide polymorphism analysis to examine overall allelic variability using a 500-K chip set costs about $500. Likewise, a genomewide gene expression profile using an Illumina 48,000 transcript chip costs about the same amount. Mitochondrial analysis and nuclear run-on assay expenses are also in the same ballpark. Even more importantly, the requirement of material has been dramatically reduced. We estimate that one can perform the entire battery of tests (excluding proteomic analysis) with about 2 million stem cells in any species. Approximately 10 million cells are enough for a mass spectrometry–based analysis, with equivalent amounts of material required for stable isotopic labeling using amino acids in cell culture and other similar methods. The material required will be further reduced as technology advances, and the costs are likely to be driven down further. Indeed, it is expected that whole-genome sequencing will cost less than $1000 in the near future and that profiling services will require one tenth of the current material. These and other technical advances have allowed large-scale gene expression analysis to be performed by a variety of techniques (Table 4.3), although many of the results remain unpublished. In general, however, investigators have focused on gene expression profiling followed by epigenetic analysis and chromatin immunoprecipitation (ChIP)-onChIP type studies. Mitochondrial sequencing, histone and chromatin modifications, and run-on assays have not been used as frequently. In most cases, data have been limited to mouse and human cells, because the genomic databases that are required for such analysis have not been as well developed in other species. In the next section, we briefly describe some of the methods used. A. Epigenetic Modulation Over the past few years, the importance of heritable epigenetic remodeling has been highlighted in the regulation of stem cell proliferation, cell fate determination, TABLE 4.3

Various Methods Used to Profile Cell Populations

Epigenetic modulation MicroRNAs Mitochondrial sequencing Transcriptome mapping Nuclear run-on assays Proteomic analysis, glycosylation maps, and other protein-mapping strategies

ASSESSING NEURAL STEM CELL PROPERTIES USING LARGE-SCALE GENOMIC ANALYSIS

76

and carcinogenesis (Beaujean et al., 2004; Huntriss et al., 2004; Meehan, 2003; Ohgane et al., 2004; Vignon et al., 2002). However, it has been difficult to study these events on a global scale. The ability to grow large numbers of cells and to differentiate them along specific pathways, coupled with the ability to perform such studies in a high-throughput fashion, suggests that this will change in the near future. Global methylation studies can be performed using a microarray (Maitra et al., 2005). Illumina has recently described a bead-array strategy to look at methylation patterns at 1500 loci encompassing regulatory elements in almost 400 genes. These include most genes known to be regulated during early embryonic development and those altered during tumorigenesis. These arrays have been used to examine methylation profiles in cancer stem cells and ESCs (Bibikova et al., 2006), and experiments assessing NSCs are underway. Assessing the epigenetic profile of NSCs will be important before using them for transplant therapy, because the maintenance of a particular epigenetic profile is probably critical for the appropriate function of cells. B. MicroRNA MicroRNAs are small noncoding RNA genes found in most eukaryotic genomes, and they are involved in the posttranscriptional regulation of gene expression. The microRNAs are transcribed in the cell nucleus, where they are processed into pre-microRNAs. Further processing occurs in the cytoplasm, where the pre-microRNAs are cleaved into their final 22-nucleotide–long form, which appears to regulate gene expression via transcriptional, translational, or protein degradation regulation (Bartel, 2004; Szymanski et al., 2003). Recent reports have identified global strategies for identifying microRNAs, and more than 450 such untranslated RNAs have been identified in humans, mice, and other species (Houbaviy et al., 2003; Lewis et al., 2003; Rajewsky et al., 2004). These approaches include computational analysis using sophisticated algorithms that recognize potential microRNA coding sequences and potential binding sites. Other strategies have included making microRNA chips (Krichevsky et al., 2003) and sequencing protocols analogous to the massively parallel signature sequencing (MPSS) developed by Lynx therapeutics, which can be used to obtain quantitative data about microRNA made by a particular cell, which in turn predicts the overall state of the stem cell. C. Mitochondrial Sequencing Structural and functional abnormalities in mitochondria lead to functional defects in the nervous system, the muscles, and other organ systems. Somatic mitochondrial mutations are common in human cancers, aging cells, and cells maintained in culture for prolonged periods. Mitochondrial DNA is also relevant to nuclear transfer, and estimating the stability of it is important for assessing the response of cells to stress and for determining their ability to propagate in culture. Techniques to examine mitochondrial DNA mutations have been under development for some time. A recent description of a polymerase chain reaction (PCR)-based approach for sequencing vertebrate mitochondrial genomes has attracted much attention for being more rapid and economical than traditional methods, which use cloned mitochondrial DNA and primer walking. Maitra et al. (2004) have developed a mitochondrial Custom RefSeq microarray as an array-based sequencing platform for the rapid and high-throughput

METHODS OF ANALYSIS

77

analysis of mitochondrial DNA. The MitoChip contains oligonucleotide probes synthesized using standard photolithography and solid-phase synthesis, and it is able to sequence more than 29 kb of double-stranded DNA in a single assay. It is useful to note that many mutations arise in the D-loop regions and that a simple PCR amplification and sequencing process would capture a large amount of information. No published data about baseline mitochondrial sequence and its change after culture of NSC are currently available. However, several laboratories have initiated such experiments, and we expect data about the long-term viability of NSCs (insofar as mitochondrial stability) to be available soon. D. Transcriptome Mapping Efforts have begun to identify the complete DNA binding sites and corresponding genes targeted by the transcriptional factors. One approach is to use an in silico computational strategy to retrieve putative genes with such binding sites. A direct and physiologic approach is to perform the ChIP of factors cross-linked in vivo to DNA targets followed by the identification of the specific DNA binding sites. Promoter chips, which range from a focused selection of genes to complete gene sets, are now being made. The findings from such a project will be of immense biologic value in providing a description and understanding of the hierarchal relationships between groups of transcription factors and their target genes as they perform their tasks in embryonic development and the specification of lineage fates and terminal differentiation. E. Nuclear Run-On Assays One analogous approach to identifying regulatory elements is to perform the labeling of newly processed RNA to examine genes that are induced only after a specific stimulus. Such hybridizations, although they require larger amounts of material, are feasible with cell lines and with ESCs, and they can provide a global overview of the network of the transcriptional responses to a specific stimulus. More importantly, they provide an element of temporal control by allowing one to better place individual genes within a transcriptional network. Although such arrays have not been run with NSCs, experiments in other systems have yielded exciting results (Li et al., 2006), and we expect similar results from NSCs in the near future. F. Proteomic Analysis, Glycosylation Maps, and Other Protein Mapping Strategies Most of our discussion has been about methods for assessing genomic differences between cells. However, posttranslational modifications play a crucial role in modifying genomic information and increasing the complexity of information that can be processed by a cell. The very complexity of the proteome has made it difficult to study on a large scale. Recently, however, multiple technical breakthroughs have begun to allow large-scale analyses. These include advances in the sensitivity of mass spectrometry, stable isotopic labeling using amino acids in cell culture, developing variations on two-dimensional gels, labeling techniques to identify key proteins that are altered under different conditions, and the development of methods for isolating and sequencing small quantities of proteins (Elliott et al., 2004; Freeze, 2003; Ong et al., 2003). Proteomic analyses of NSCs have not yet been reported, despite the fact that cell lines have been available for several years.

78

ASSESSING NEURAL STEM CELL PROPERTIES USING LARGE-SCALE GENOMIC ANALYSIS

IV. DATA MINING: CHROMOSOME MAPPING, PATHWAY ANALYSIS, AND DATA REPRESENTATION A major problem with large-scale analysis has been knowing how to interpret the data, how to compare them, and how to extract meaningful and biologically relevant information. Biologists as a rule cannot simply look at long lists of genes to identify critical information, and examining the most abundant gene may not be of biological significance. For example, changes in notch or b-catenin signals of twofold or less may be biologically significant, whereas 10- or even 100-fold differences in the expression of some genes may be irrelevant for the biologic state of the cell. A particularly telling example of the relevancy of some differential gene expression data comes from the ESC literature, in which digital differential display identified TEX15 and DPPA5 as genes that are highly expressed in ESCs but low or absent in all other populations examined (Adjaye et al., 2005; Kim et al., 2005; Lagarkova et al., 2006). These results were verified by PCR and immunocytochemistry, although data from knockout mice showed that these genes are dispensable for all assessed functions (Amano et al., 2006). Alternatively, a similar strategy identified Nanog, a previously unknown key regulator of ESC differentiation (Chambers et al., 2003; Mitsui et al., 2003). In the field, this has led to multiple attempts to consider how one should analyze data sets that are generated. In our laboratory, we have made the following assumptions: before pooling data or subjecting them to analysis, the quality of the sample used is tested. For example, with NSCs, the harvested sample is assessed for its expression of known NSC markers and the absence of markers of differentiation (to measure contamination; Figure 4.3, A). The presence or

FIGURE 4.3 A comparison of NSC, astrocyte precursor cells, and oligodendrocyte precursor cells using global gene expression analysis. The process of profiling NSC is shown as an example. A, Astrocyte precursor cells and oligodendrocyte precursor cells are chosen as comparison groups, and the sample quality is monitored by examining marker expressions and their differentiation potential. B, Global comparison is shown in terms of gene numbers detected in each sample. C, The expressed gene numbers are subcategorized to show gene numbers in certain intensity. D, The relatedness among populations is accessed and shown by the R2 score. E, The common and uncommon genes for each population are visualized in a Venn diagram. F, The results are verified by independent methods of polymerase chain reaction and immunocytochemistry. (See color insert.)

DATA MINING: CHROMOSOME MAPPING, PATHWAY ANALYSIS, AND DATA REPRESENTATION

79

absence of these markers on the array is determined, and the ability of the array to detect such differences is assessed by running the same sample on an array. If an array hybridization or MPSS analysis fails to detect an expected result, then those data are not used, because all subsequent predictions are too uncertain. This analysis allows one to determine the expected sensitivity of the result, and it provides a rough idea of the sampling space (how much one will miss). We then examine carefully the intensity distribution of the expression levels of the genes present on the array (Figure 4.3, B and C). We have noted that, in most cell types, the distribution is quite similar, and, therefore, any alteration suggests technical errors. Although normalization algorithms can be employed in an attempt to use a particular anomalous data set, we generally red flag it, because most normalization algorithms tend to skew results. Empirically, we have determined that comparison across platforms is fraught with peril. Only positive results can be considered, and negative results are generally not interpretable. For example, ESC-derived NSC samples were taken and examined by MPSS and Illumina bead arrays, and a concordance of around 50% was shown, whereas the concordance rate for the sample run on a second Illumina array was close to 95% (unpublished observations). Even with such a comparison, only the presence or absence of an expression pattern can be considered, and no attempts to compare expression levels between different methods should be made. We have also determined that amplification tends to provide a different pattern than using unamplified RNA, even if the same sample is used. This appears to not be the result of operator error, because the amplification of different biologic replicates performed at different times yields more similar results than comparison between amplified and unamplified samples. Therefore, only samples that have been processed identically are compared as far as it is practical. After we are comfortable with the quality of each sample processed, we examine differential gene expression by examining pairwise comparisons rather than pooling the data, or we compare them with a reference standard of baseline data that have been generated. This allows one to generate larger data sets and to compare across laboratories. For example, with NSCs, we suggest using human universal RNA as a potential reference standard: it is widely available, it is standardized, it has been run across multiple platforms, and such data sets are publicly available. This method allows one to readily determine if hybridization results are within the normal range and whether the data are usable and comparable with results from other laboratories. After we have determined that we have a reasonable set of data, we then determine an appropriate cutoff for sensitivity with which we are comfortable. This ranges from array phenotype to phenotype, and it is an important criterion in any assessment. In MPSS, for example, a theoretical sensitivity when 3 million tags are sequenced is 3 transcripts per million (Miura et al., 2004a). However, we have empirically determined that testing or validating expression at such low levels is difficult, and, as such, it may or may not be useful to consider. In our hands, an expression level of 50 transcripts per million is readily verifiable, and it is a cutoff that we routinely use (Cai et al., 2006), although this does mean that we are potentially discarding useful information (which can be substantial, because a majority of genes are expressed at low levels). Likewise, with Illumina bead arrays, we use a cutoff of 50 to 100

80

ASSESSING NEURAL STEM CELL PROPERTIES USING LARGE-SCALE GENOMIC ANALYSIS

arbitrary intensity units, although the theoretical sensitivity of the arrays is much lower. However, it is important that this be made clear in any publication or, alternatively, that the raw data be made available for independent analysis. This is particularly important because the genomic databases are constantly being curated, and expressed sequence tags assigned to a particular locus are being reassigned as better data become available. This curation sometimes means that a gene tag on an array may not represent the gene that it was originally thought to identify. Various groups have estimated this frequency, and they have recognized that it is an important source of error in these types of analyses. Therefore, we strongly recommend that such curation and updating be a regular part of the data analysis. After we have determined a cutoff, established a set of phenotypes for analysis, and collected the curated data, we then can begin to consider how best to analyze the data. This analysis is dependent on the biologic questions one wishes to pose, and, although each strategy will be different, we can perhaps highlight some of the simple strategies that can be used. We currently feel that the genomic data that are most complete are those from the human and the mouse. Therefore, we have focused on large-scale analysis in these two species. We find that mapping the expressed genes onto a genome browser (chromosome mapping) provides a very useful overview of global patterns of gene expression. It allows one to study the regulation of genes on a global scale, to identify genomic hotspots, and to correlate with known chromosomal breakpoints, single-nucleotide polymorphism data sets, data developed by groups working in different disciplines. We would recommend the University of California, San Francisco golden path browser for this purpose. A second important strategy that we routinely use is what we call a “pathway analysis.” Here, rather than looking at the expression of an individual gene, we examine an entire signaling pathway by mapping the expression of all detected genes for that pathway. A visual representation of the “on” and “off” state of the pathway can be easily obtained, and observation of the pathway as cells differentiate can provide a much clearer look at whether genes in that pathway are important in the process of differentiation. For example, such a pathway analysis clearly identified the LIF/GP130 pathway as being important in the NSCs of humans but not of mice (unpublished data). It also helps identify key genes that must be tested in verification studies. Multiple commercial programs to perform such pathway analyses exist, and we recommend using any one of these. Overall, our experience has been that, if an effort is made to develop well-curated data sets, the verification of observations is generally greater than 50%, and one can glean important and reliable information. Although a hit rate of 50% may sound low, it is useful to consider that this is much better than the 1 in 30,000 chance of finding a functionally useful gene that one started off with before analysis. For example, there are currently about 10 NSC biomarkers that have been identified to date. However, using a large-scale analysis, one could readily identify perhaps 200 such markers, of which perhaps 25% (50) would be novel or unexpected genes. With a hit rate of 50%, one could identify 25 new markers during a 3-month experiment, thus more than doubling the number of known NSC markers. In the next section, we will discuss some general observations made about NSCs.

GENERAL OBSERVATIONS ABOUT THE PROPERTIES OF NEURAL STEM CELLS

81

V. GENERAL OBSERVATIONS ABOUT THE PROPERTIES OF NEURAL STEM CELLS We find that NSCs appear to be similar to other cells in that they synthesize about 10,000 to 12,000 genes of an estimated 35,000 or 40,000 genes annotated in the RefSeq database. Average total RNA per cell tends to be higher among stem cells as compared with other cells (average, 5–10 pg), but it is similar to levels seen in metabolically active cells. The distribution of transcript frequency suggests that most genes are transcribed at relatively low levels (T and 1298A>C polymorphisms result in elevated endogenous homocysteine levels, and they are associated with an increased risk of NTD. Environmental factors also play a role in the etiology of craniofacial defects, such as cleft palate and cleft lip. Orofacial clefts have been associated with maternal cigarette smoking, alcohol consumption, and a lack of folic acid supplementation. Human maternal periconceptional intake of multivitamins containing folic acid has been associated with a reduction in the risk of delivering infants with clefts; however, this reduction in risk has not been observed in all studies. The transforming growth factor alpha (TGFa) genotype has also been shown to be important for contributing to the development of cleft palate. Some genetic studies have shown a two- to five-fold increased risk of clefting among individuals with the less-common allele for TGFa. TGFa is a secretory protein that binds to the epidermal growth factor receptor, and it has been localized to palatal epithelium during mouse palatal closure. Studies using targeted knockout mouse models and spontaneous mutants have been powerful in the validation of the importance of TGF signaling in clefting (i.e., TGFb3 / mutant mice exhibit clefting; Proetzel et al., 1995). Moreover, the discovery of numerous mouse mutants with cleft palate has confirmed the complex genetics of this trait and the multiple signaling pathways that are involved with clefting. Although human and mouse cleft development is remarkably similar, mice do not develop cleft lip.

II. MODEL ORGANISMS A. Primary Model Organisms in the Study of Development and Disease 1. Unicellular Organisms a. Yeast The budding yeast Saccharomyces cerevisiae and the fission yeast Schizosaccharomyces pombe are single-celled fungi with distinct life cycles. S. cerevisiae is the first eukaryotic organism for which the entire genome sequence was completed (Goffeau et al., 1996). The yeast genome contains about 6000 genes, and about 20% of human disease genes have counterparts in yeast. Yeast’s rapid generation time and simple and inexpensive maintenance under laboratory conditions make it advantageous for both classic genetic studies and high-throughput genomic approaches. S. cerevisiae provides a

134

UNDERSTANDING HUMAN BIRTH DEFECTS THROUGH MODEL ORGANISM STUDIES

TABLE 7.1

Strengths and Weaknesses of Model Organisms

Species

Strengths

Weaknesses

Homologous recombination; powerful genetic and proteomic technologies; complete genome sequence; simple and inexpensive maintenance; basic eukaryotic cell organelles present; cell cycle control similar to animals

Unicellular; no distinct tissues

Powerful genetic and proteomic technologies; simple biologic processes similar to animals

Unicellular; no distinct tissues

Excellent genetics; effective RNA interference; powerful molecular techniques; complete genome sequence; suppressor/ enhancer screens; fully known morphology; transparent; small number of cells and all cell lineages characterized; longterm storage ( 80 C); known neuronal connectivity

No homologous recombination; difficult gene-expression analysis

Powerful genetics; molecular techniques; complete genome sequence; suppressor/enhancer screens; mosaic analysis; effective RNA interference; easily generated transgenics; well-characterized development

No embryo freezing; difficult embryologic manipulations; difficult targeted gene disruptions

Vertebrate; external fertilization; large number of eggs; transparent; accessible developmental stages; organ systems similar to higher vertebrates; morpholinos; RNA interference; mutagenesis screens

No homologous recombination; difficult to generate transgenics

Xenopus laevis

Vertebrate; external embryonic development; large size; identifiable blastomeres; easy embryo manipulations

Long time to sexual maturity; no genetics; difficult to generate transgenic animals

Gallus gallus

Descriptive embryology; ideal for embryologic manipulations (e.g., transplants of limbs, neural crest, notochord)

Limited genetics

Mus musculus

Vertebrate; development, cell types, and tissues similar to human; powerful genetics; targeted gene disruption by homologous recombination; transgenic technologies; large mutant collection; source for primary cell cultures

Development is in utero; expensive maintenance; relatively long time to sexual maturity and maturity; early embryonic lethal phenotypes difficult to study (resorption in vivo)

Unicellular organisms Saccharomyces cerevisiae

Dictyostelium discoideum

Invertebrates Caenorhabditis elegans

Drosophilia melanogaster

Vertebrates Danio rerio

MODEL ORGANISMS

135

significant advantage for experimentation because of its fast and easy means of gene cloning, gene disruptions, gene overexpressions, and single-step gene replacements (Table 7.1). Because it can grow as either haploids or diploids, recessive mutations can be identified by phenotypic changes in the haploid strain. In addition, complementation analyses can be performed in yeast. In contrast with mammalian cells, in which redundant processes are often obstacles to understanding the function of a specific gene, yeast provide a clean readout against a null background. As a result of the high degree of conservation of basic molecular and cellular mechanisms between yeast and human cells, it represents a highly useful system for the investigation of cell architecture and fundamental cellular mechanisms. The genes controlling the eukaryotic cell cycle were identified mainly through studies of the yeasts S. cerevisiae and S. pombe. Yeast has proven to be an extremely good model for cancer studies, because its mechanism of cell division and its response to DNA damage are similar to human cells. It is also used as a model in apoptosis, aging, and DNA repair studies. However, limitations exist, because not every human disease gene has an ortholog in yeast, and pathologies that affect specific tissues, organs, and physiologic functions cannot be assessed at the single-cell level. Moreover, as a result of its unicellularity, the functions of the genes that are expressed as different isoforms in different cell types cannot be analyzed in yeast. As a whole, yeast has been an excellent model for studying conserved biologic cell mechanisms that affect developmental processes in a cell-autonomous context. They have also proven useful for translating the consequences of human mutations that act in a non–cell-autonomous fashion at the tissue or organ level. b. Dictyostelium discoideum Dictyostelium discoideum amoebae thrive in moist soil. Nutritional stress drives cells to aggregate by means of chemotactic signals, and these aggregates then differentiate into multicellular fruiting bodies that contain spores. Its recently published genome encodes approximately 12,500 genes. Although this is more than twice the number of genes in yeast, it is still only about half that of humans, and the rarity of alternative splicing simplifies its proteome even further as compared with those of vertebrates (Eichinger et al., 2005). However, efficient genetic manipulations by gene targeting and replacement as well as by insertional mutagenesis, suppressor screens, and RNA interference (RNAi) make Dictyostelium a popular experimental system. Dictyostelium exemplifies many processes that are characteristic of complex eukaryotes, including cytokinesis, motility, phagocytosis, chemotaxis, signal transduction, and aspects of development such as cell sorting, pattern formation, and cell-type determination and differentiation. It has been used by researchers to study the mechanisms of action of myosin mutations that cause cardiac myopathies, the molecular basis of cisplatin (a drug used for the treatment of cancer), and the mechanism of action of lithium and VPA, which are used for the treatment of depressive disorders (Egelhoff et al., 1993; Eickholt et al., 2005; Li et al., 2000). It has also been established as a host model for the pathogenesis of infectious diseases such as malaria, Legionnaire’s disease, salmonellosis, tuberculosis, listeriosis, and pseudomoniasis. Although Dictyostelium is a very good model for the study of simple cellular behaviors, its limited cellular diversity and its absence of distinct tissues limit its use as a model for eukaryotic cell function.

136

UNDERSTANDING HUMAN BIRTH DEFECTS THROUGH MODEL ORGANISM STUDIES

2. Invertebrate Models a. Caenorhabditis elegans Caenorhabditis elegans is a small soil nematode with a short life cycle of 3.5 days at 20 C. Adult hermaphrodites of this species can give birth to more than 300 progeny by self-fertilization. The worms are very easy to grow in the laboratory on agar medium in Petri dishes, and stocks can also be frozen, thereby allowing one to create large numbers of mutant strains with limited maintenance. Its cell lineage, the complete connectivity of its nervous system, and its nerve–muscle synaptic connections are known; this has made it a powerful model for studying basic neurodevelopmental mechanisms. As a result of its transparency, it is possible to view internal structures, especially with enhancement by green fluorescent protein and differential interference contrast in live animals. The genome sequence of C. elegans was completed in 1998, and it revealed that 43% of C. elegans genes have human orthologs, including numerous disease genes (Culetto and Sattelle, 2000). A major strength of using C. elegans as a model system is that genetic manipulations are easily performed and tools are well developed. The ability to breed a mutation to homozygosity in a single generation facilitates the performance of genetic screens for recessive mutations more readily than in other organisms. The powerful gene-mapping strategies based on single nucleotide polymorphism detection can also be performed in worms. Rapid cosmid rescue-transformation of mutant animals by microinjection is also available. RNAi is effectively and easily applied to C. elegans in the knockdown of specific genes. The organism is also well suited for second-site suppressor/ enhancer screens that facilitate the determination of components of a genetic pathway after a single gene involved in that process has been identified (Jorgensen and Mango, 2002). C. elegans has been extremely powerful in the study of the apoptotic pathway in higher eukaryotes, because the key components of the apoptosis machinery appear to be conserved between humans and nematodes. Many genetic diseases involve the dysregulation of apoptotic programs, such as sclerosis, type I diabetes mellitus, Hashimoto thyroiditis, Sjo¨gren syndrome, and certain cancers (e.g., melanoma), thereby making C. elegans an important model for elucidating their pathogenesis. C. elegans has also been used as a model system for studying the mechanisms of aging, neurodegenerative diseases, muscular dystrophy, polycystic kidney disease, and other human diseases. Despite its unique advantages, C. elegans has limitations as a model organism. One of these limitations is, of course, its relative divergence from humans as compared with that of another much-studied invertebrate, the fly. Also, difficulties in the direct analysis of gene expression and the performance of embryologic manipulations are other experimental limitations. b. Drosophila melanogaster The fruit fly Drosophila melanogaster has been used by researchers for more than 100 years in the areas of gene discovery and genetic analyses. Since the completion of the genome sequence of Drosophila in 2000, 61% of its genes have been shown to have human counterparts (Adams et al., 2000). Its rapid generation time and the availability of various forward and reverse genetics approaches make Drosophila a powerful model organism. Most of

MODEL ORGANISMS

137

the studies in the fly have been performed by using forward genetics, in which the chemical mutagen EMS or P-element transposition are used in large-scale phenotype-based mutagenesis screens. The availability of single nucleotide polymorphism maps has also been useful for mapping specific mutations. P-element transposons have also been used to screen for second-site modifiers (enhancers and suppressors) of a specific sensitized background, which permits the identification of new genes involved in a given developmental pathway or process. Large collections of P-element insertion stocks have been generated, thereby allowing for the direct screening for mutant phenotypes. Misexpression and overexpression phenotypes have been generated in a spatiotemporal fashion using the GAL4-UAS system. Finally, clonal analyses have been extensively used in Drosophila to trace cell lineages, to analyze lethal mutations at later stages of development, and to distinguish cell-autonomous versus nonautonomous actions of genes after applying the FLP/FRT recombination system in somatic lineages. Loss-of-function studies of a particular gene have been analyzed by different methods, such as the imprecise excision of P-elements, targeted gene replacement, or RNAi. RNAi has been shown to effectively block gene expression in vivo in Drosophila. It has been shown that 75% of human genes known to be associated with disease have a Drosophila ortholog (Reiter et al., 2001). Moreover, Drosophila shows similarities to humans in basic biologic cell processes, including gene expression, membrane trafficking, cytoskeleton organization, extracellular matrix, determination of cellular asymmetry, epithelial organization, neuronal connectivity, synaptic function, cell–cell and intracellular signaling pathways, and apoptosis. Arguably, Drosophila has been the model organism of choice for dissecting the genetic pathways that affect neurodevelopment. Recently, Drosophila has been applied to identifying mechanisms of human neurodegenerative disease, including Alzheimer, Parkinson, and Huntington diseases. Neurodegenerative diseases share common features. They are caused by dominant mutations; they exhibit a late onset and progressive neuronal degeneration, and they are associated with the formation of highly stable protein aggregates. In the case of polyglutamine diseases, repeat length correlates with the severity of the phenotype. It has been shown that flies can mimic human pathology in several respects. Transgenic flies expressing polyglutamine repeats showed that increased repeat length causes neural degeneration, and degeneration is typically seen in late Drosophila development (Marsh et al., 2000). As in the human, condition neuropathology in the fly is progressive, and protein aggregates form upon the expression of mutant polyglutamine repeat peptides (Davies et al., 1997; Marsh et al., 2000). Studies with Drosophila have shown that flies are helpful for the sharing of mutant phenotypes that are similar to those of human diseases and also for facilitating the identification of the components of a given developmental pathway. For example, when the adenomatous polyposis (APC) gene, which is responsible for numerous intestinal polyps that predispose individuals to colon cancer, was identified by positional cloning in 1991, its function was unknown. Later, the identification of APC’s interaction with b-catenin provided the first clues about its function. However, the link between the Wnt signaling pathway and APC was established after b-catenin (Armadillo in Drosophila) and the Wnt signaling components were discovered by genetic analyses in Drosophila. It is now known that the Wnt pathway plays a critical role in the pathogenesis of colon cancer. Similarly, the identification of the

138

UNDERSTANDING HUMAN BIRTH DEFECTS THROUGH MODEL ORGANISM STUDIES

Notch, Shh, and Nodal pathways in Drosophila has shed light on many aspects of the dysregulation of these signaling mechanisms in human biology and disease. Although Drosophila has been a very powerful model organism, several limitations exist. Flies differ from humans by their much simpler circulatory systems, immune responses, skeletal systems, and cognitive abilities. Numerous subtle differences exist in the functions of individual proteins. Humans often have several copies of a gene that is present in only one copy in the fly genome. The different forms of many human genes have evolved to acquire different expression patterns and unique functions in different cell types. For example, unlike Drosophila, there are three hedgehog proteins (Shh, Indian hedgehog [Ihh], and Desert hedgehog [Dhh]) in mammals. Among these, Ihh has been shown to be important for the differentiation of prehypertrophic chondrocytes, which is a cell type that is not found in flies. In fact, skeletal development serves as a contrast with neurodevelopment when considering the relevance of the Drosophila model to human disease. 3. Vertebrate Models a. Danio rerio (zebrafish) Zebrafish (Danio rerio) has been an attractive vertebrate model system because of its simple and inexpensive maintenance at high densities in the laboratory and the large numbers of embryos that can be produced in a very short time. Zebrafish embryos develop most of the major organ systems— including the cardiovascular, nervous, and digestive systems—in less than a week. The transparent embryos develop externally, which allows for direct observation of the morphologic defects. Although targeted gene knockout technology has not yet been developed for zebrafish, specific genes can be knocked down transiently by the injection of morpholino antisense oligomers. High-resolution simple sequence length polymorphisms and radiation hybrid maps are available to help with the genetic mapping of mutations. Large-scale forward genetic screens have been generated in zebrafish by using N-ethyl-Nnitrosourea (ENU) and insertional mutagenesis. The mutants that were obtained from these screens showed embryonic patterning and organ system defects of the retina, bone, cartilage, brain, hematopoietic system, digestive system, and cardiovascular system. These mutants represent a useful tool for identifying genes that are involved in human disorders. Zebrafish embryos have long been a model for studying teratogen-induced malformations such as human FAS. The characteristic features of human FAS, such as brain defects, are also observed in zebrafish that are exposed to ethanol. It is a useful model not only for dissecting disease-associated genes and pathways but also for testing for environmental toxins in humans, for drug-target identification, and for the in vivo validation of targets before clinical trials (Blader and Strahle, 1998). b. Xenopus laevis and chick Xenopus laevis is a nonmammalian vertebrate that has been used as a model system as a result of its several advantages, including external embryonic development, large size, and easy experimental manipulations. It has been used for the study of embryonic development, the patterning of the basic body plan, the determination of cell fate, and the early patterning of major

MODEL ORGANISMS

139

organs (e.g., the digestive, circulatory, and nervous systems). Studies in Xenopus have helped elucidate the formation and function of Spemann’s organizer. Similar to zebrafish, loss of function can be generated by morpholino oligos, but gene targeting is not yet widely available. Another nonmammalian vertebrate model is the chick. The chicken has significant advantages as a model because of its low cost and availability, its fully sequenced genome, and its simple physical manipulations in ovo. Direct access to the embryo facilitates the removal of tissue, the implantation of morphogen-soaked beads, heterotopic transplantations, and both viral and nonviral mediated gene transfer to tissues. The chicken has proven to be a superb model organism for the study of limb development, and it has enabled the identification of key organizing centers, such as the zone of polarizing activity and the apical ectodermal ridge. It has also been used in the study of viruses and cancer. The first tumor virus (Rous sarcoma virus) and oncogene (src) were identified in the chicken. However, one of the disadvantages of these nonmammalian vertebrate animal models is the relative divergence of some genes and their functions from that seen in humans. Moreover, direct genetic manipulations are difficult to perform, which can be a problem during the creation of transgenic strains and targeted gene disruptions. c. Mouse The completion of the mouse genome sequence in 2002 demonstrated that 80% of mouse genes had a single human ortholog (Waterston et al., 2002). Well-developed genetic manipulations and the ability to use powerful molecular tools made the mouse an even more valuable model system (Bedell et al., 1997). ENU-induced point mutations have been successfully used to generate large-scale chemical mutagenesis screens (Justice et al., 1999). These screens enabled the production of allelic series of mutations that complemented classical loss-of-function alleles with hypomorphic and neomorphic alleles. Thus, the potential embryonic lethality of recessive null mutations could be circumvented. In vitro manipulations of embryonic stem cells (ESCs) allowed for the application of reverse genetic approaches, including of knockout technology and insertional gene trapping. Homologous recombination-based gene targeting had the greatest initial impact on the understanding of gene function in mice. Although loss-of-function studies have provided important information about many human diseases, they have been less informative in situations involving genetic redundancies and early embryonic lethality. The lethality of recessive null alleles prevented the analysis of the later functions of a gene. However, this has been ultimately circumvented by generating conditional alleles by incorporating Cre recombinase-mediated Lox P excision in the targeting vector or by isolating point mutations from ENU screens. Abundant comparative genetic studies between human and mouse mutant alleles have revealed cases of both excellent and poor phenocopies. In some cases, both the inheritance pattern and the phenotypes correlated well. In others, phenotypic differences and incomplete penetrance could be observed. For example, variable phenotypicexpressivity in addition to strain-dependent penetrance can often be observed in mouse models. Factors that affect the degree of phenotypic correlation between mice and humans include the following: 1) differences in the dosage sensitivity of the mutation and the affected pathway; 2) differences in redundant pathways and genes; 3) genetic and

140

UNDERSTANDING HUMAN BIRTH DEFECTS THROUGH MODEL ORGANISM STUDIES

epigenetic modifiers specific for inbred strains of mice versus outbred humans; 4) true divergent function of the target gene in mouse versus humans; and 5) differential effects seen early versus later in development. Although knockout mice might present with a more severe phenotype than the human phenotype or have no clinical phenotype at all, differences may be used advantageously for identifying modifiers, alternative developmental pathways, and novel gene interactions. An example is the Lesch– Nyhan syndrome mouse model. Lesch–Nyhan syndrome is an X-linked disorder characterized by hyperuricemia, choreoathetosis, spasticity, and mental retardation that results from a complete lack of hypoxanthine–guanine phosphoribosyltransferase (HPRT). HPRT knockout mice are phenotypically normal and healthy, with only subtle changes in brain dopamine levels. However, they provided a good biochemical model for studying the in vivo consequences of HPRT mutation on metabolite alterations (Jinnah et al., 1994). In contrast with the generation of loss-of-function mutations, gain-offunction mutants have been primarily generated in a targeted fashion by pronuclear injection into generate transgenic mice. Here, mice have served as valuable models of overexpression of wild-type and/or mutant proteins. The use of tissue-specific promoters to direct transgene expression has enabled the misexpression of proteins in tissues not normally expressing the target gene, whereas the development of BAC transgenesis has enabled more physiologic overexpression models. Despite its many advantages as a model system and its similarity to humans in both anatomy and physiology, the mouse is far from the perfect experimental system. Unlike lower-order vertebrates, early organogenesis, differentiation, and development are hard to observe, because these processes take place in utero. Dissecting early essential functions during embryogenesis from later tissue-specific or organ-specific functions can be challenging and may require lengthy genetic manipulations and breeding schemes. Finally, mice have a comparatively longer life cycle, and colonies are relatively expensive to maintain in laboratories. As an alternative rodent model, the laboratory rat (Rattus norvegicus) has been a long-established model for studying human disease and physiology. Unfortunately, gene-targeting studies in the rat have been limited by the lack of availability of embryonic stem cells and technical inefficiencies. However, transgenesis has been achieved by microinjection and, more recently, by retroviral/lentiviral integration. The ability to move mouse/rodent modeling of human disease into an accelerated phase will be facilitated by the widespread availability of null mutations for all genes and then by the generation of a series of allelic mutations in each gene. To achieve this latter goal, the ability to rapidly generate targeted single nucleotide substitutions in ESCs will be important. Progress in new technologies, such as in vivo RNAi using lentiviral transfer into ESCs and pronuclei, is providing novel avenues for generating hypomorphic and loss-of-function alleles more rapidly. However, off-target effects as well as potential unwanted effects on the endogenous microRNA processing machinery have yet to be evaluated. B. Mouse Models for Human Birth Defects Developmental defects arising from genetic mutations can be broadly divided into those with structural origins and those with metabolic origins.

MODEL ORGANISMS

141

1. Structural Defects Structural defects are congenital malformations that result from altered patterning, differentiation, proliferation, remodeling, and the apoptosis of affected tissues during development. Resulting phenotypes can be associated with most cellular processes, including the following: 1) morphogen and morphogen antagonists; 2) morphogen–receptor interactions; 3) signal transduction downstream of receptor–ligand interactions; 4) transcription; 5) RNA processing, posttranslational modification, and trafficking; and 6) matrix production and cell matrix interactions. Skeletal malformations exemplify a common class of structural birth defects, and developmental dysplasias of the skeleton can be used to illustrate the salient features that are characteristic of this class. a. Skeletal dysplasias as a model for comparative mouse–human genetic analyses Formation of the skeleton requires the differentiation of the mesenchymal stem cell via either the chondrogenic lineage or the osteoblastic lineage. Chondrogenesis involves differentiation first into chondroblasts, then into proliferating chondrocytes, next into prehypertrophic chondrocytes, and finally into hypertrophic chondrocytes during the process of endochondral ossification (see Chapter 39). Terminally differentiated hypertrophic chondrocytes undergo apoptosis and are replaced by mineralizing bone. By contrast, intramembranous ossification involves the direct differentiation of the stem cell into preosteoblasts, early osteoblasts, mature osteoblasts, and, finally, into terminally differentiated osteocytes. Both cell differentiation processes must be patterned in the body plan’s three-dimensional space (i.e., proximal–distal, anterior–posterior, and dorsal–ventral). Two critical transcription factors directing the specification of the mesenchymal stem cell to the respective chondrogenic or osteoblastic lineages are Sox9 and Runx2. In humans, haploinsufficiency of SOX9 (a transcription factor with an sex determining region Y [SRY]-related high-mobility group [HMG] box domain) causes campomelic dysplasia. Campomelic dysplasia is an autosomal dominant neonatal chondrodysplasia characterized by a severe dwarfism that affects all cartilage-derived structures and that is also characterized by frequent male-to-female sex reversal. During mouse embryonic development, Sox9 is expressed in all prechondrocytic mesenchymal condensations. Later, its expression is maintained at high levels in fully differentiated chondrocytes. Sox9þ/ mice have been generated by classic gene-targeting strategies. Heterozygous Sox9 mice die perinatally and phenocopy most of the skeletal abnormalities seen among patients with campomelic dysplasia, including cleft palate, hypoplasia, and the bending of many skeletal structures (Bi et al., 2001). Moreover, studies using Sox9 knockout chimeric mice showed that mesenchymal cells that did not express Sox9 were unable to differentiate into chondrocytes or to contribute to mesenchymal condensations, which suggests that Sox9 is essential for chondrocyte differentiation and cartilage formation (Bi et al., 1999). Runx2 is a runt domain transcription factor that is essential for osteoblast cell fate commitment and chondrocyte maturation. It is expressed in osteochondro progenitors, developing osteoblasts, and in a subset of chondrocytes. The haploinsufficiency of RUNX2 in humans causes dominantly inherited cleidocranial dysplasia (CCD). CCD is characterized by skeletal anomalies (including open fontanels), late closure of cranial sutures with Wormian bones, delayed eruption of permanent dentition, rudimentary clavicles, and

142

UNDERSTANDING HUMAN BIRTH DEFECTS THROUGH MODEL ORGANISM STUDIES

short stature. Several mouse models of Runx2 function have been generated. Runx2 null mice show a complete lack of bone formation as a result of a maturational arrest of osteoblasts (Komori et al., 1997; Otto et al., 1997). These mice have only cartilage anlagen of the skeleton and no evidence of osteoblasts or mineralization. Heterozygous mice phenocopy CCD with delayed ossification of the fontanels and hypoplastic clavicles. In addition, transgenic mouse models demonstrated that Runx2 is also important for postnatal bone development. The overexpression of Runx2 in osteoblasts caused osteopenia and fractures in transgenic mice (Liu et al., 2001), whereas it has been shown that the overexpression of Runx2 in cartilage induced chondrocyte hypertrophy while partially rescuing the chondrocyte maturation defects in Runx2 null mice (Takeda et al., 2001). These data demonstrate a complex spatiotemporal function of Runx2 in skeletal development. The function of Runx2 in chondrocyte maturation has been confirmed in humans. Both decreased hypertrophy and decreased expression of RUNX2 target genes were found in human CCD cartilage (Zheng et al., 2005). In general, there is an excellent correlation between mouse and human mutations for these two transcription factors. This reflects the strict dosage requirement for many transcription factors in general and the often near-complete penetrance of phenotypes associated with alterations in transcription-factor expression levels. In fact, most human transcription-factor diseases are dominantly inherited, and the majority are the result of loss-of-function mutations. Similarly, most corresponding mouse mutants are semidominant and faithfully phenocopy the human condition. The pathogenic consequences of transcription factor aberrations ultimately reflect the dysregulation of their target transcriptional network of genes. Considering skeletogenesis, key targets for Sox9 and Runx2 include matrix structural proteins. It is not surprising that mutations in the matrix structural proteins were the first identified in the osteochondrodysplasias, but, because of the complexity of matrix-cell and matrix-environment interactions, comparative mouse–human studies have been insufficiently robust to reveal all influences. For example, the tetraped versus the biped nature of rodents versus primates dictates different biomechanical forces impinging on the phenotypic expression of matrix alterations in the skeleton. In fact, the tremendous clinical variability associated with both fibrillar collagen mutations in humans points to the contribution of many modifying factors. Human mutations of type I collagen in the COL1A1 and COL1A2 genes cause osteogenesis imperfecta (brittle bone disease). Type II collagen (COL2A1) mutations cause disproportionate dwarfism or chondrodysplasias of varying severities (achondrogenesis type II, hypochondrogenesis, spondyloepiphyseal dysplasia congenita, Kniest dysplasia, and Stickler syndrome spectrum). Still, observations from mouse studies have affected our understanding of the underlying disease mechanisms. For example, subtle substitution mutations in the fibrillar collagens act in a dominant negative fashion, producing a severe phenotype. Large deletions and null mutations cause a quantitative loss-of-function effect and are associated with milder phenotypes replicated in the first COL1A1 mouse models. Retroviral insertional mutants that affected levels of COL1A1 expression caused mild osteogenesis imperfecta in mice, whereas transgenic mice harboring additional copies of the gene with engineered point mutations had severe osteogenesis imperfecta (Bonadio et al., 1990; Stacey et al., 1988). As “knock-in” technologies have improved, the replacement of the wild-type allele with a point mutation

MODEL ORGANISMS

143

replicating human alleles has become a powerful approach for modeling. However, there are still significant limitations for mouse models in the study of human disease. Of note, type X collagen (COL10A1) mutations cause Schmid metaphyseal chondrodysplasia via dominant loss-of-function mechanisms. Interestingly, mice that are null for COL1A1 appear to be phenotypically normal (Rosati et al., 1994). Morphogen signaling occurs upstream of transcriptional networks. A primary signaling pathway regulating chondrocyte proliferation and differentiation is the fibroblast growth factor signaling pathway. Recurrent FGFR3 mutations have been shown to cause most cases of achondroplasia, which is the most common form of disproportionate dwarfism. Ultimately, the mouse models for Fgfr3 mutations elucidated the pathogenesis of this disorder. Fibroblast growth factor signaling negatively regulates chondrocyte proliferation, and the achondroplasia mutation is an activating mutation. A complete loss of function of Fgfr3 in mice caused overgrowth of the growth plate (Colvin et al., 1996). By contrast, transgenic mice expressing the achondroplasia mutation as well as knock-in mutants phenocopied human achondroplasia (Naski et al., 1996; Wang et al., 1996). 2. Metabolic–Endocrine Defects Developmental defects can also be associated with classic inborn errors of metabolism. These diseases are characterized by deficiencies of enzymes and transporters that result in the dysregulation of metabolite flux. The consequent accumulation of toxic upstream precursors, the deficiency of downstream products, and/or the stimulation of alternative metabolic pathways occur and exceed the threshold for clinical disease. Major categories of disease include organic acidemias, peroxisomal disorders, lysosomal storage disorders, carbohydrate metabolism disorders, amino acidopathies, fatty acid oxidation defects, mitochondrial/respiratory chain defects, and urea cycle disorders (Lanpher et al., 2006). Some conditions can cause structural developmental defects, especially in target organs like the liver, muscle, and brain. Others may be primarily associated with metabolic decompensation and/or chronic neurologic symptoms. Modeling these diseases in the mouse has represented a considerable challenge. The divergence of metabolic pathways and the differential use of alternative disposal pathways allow for a significant discordance in phenotypes between mouse and man, as previously exemplified by the HPRT enzyme. Galactosemia and urea synthesis provide excellent examples, comparing in the one case the difficulty and the other the facility of using mouse modeling for the study of human inborn errors of metabolism. Galactosemia is an autosomal recessive disease of carbohydrate metabolism that results from a deficiency of the enzyme galactose-1-phosphate uridyltransferase (Galt). Galactose and the derived toxic products galactose-1phosphate and galactitol accumulate in the blood, leading to neonatal morbidity and mortality. Even with tight dietary control, patients suffer long-term morbidity that is likely related to a deficiency of downstream product and dysregulation of the glycosylation machinery. In affected infants, symptoms present soon after the ingestion of a lactose-based formula or breast milk. Homozygous individuals exhibit vomiting, rapid weight loss, hepatomegaly, and jaundice. Long-term complications include mental retardation, cataracts, hepatomegaly, and ovarian failure. Animal models have been used to better

144

UNDERSTANDING HUMAN BIRTH DEFECTS THROUGH MODEL ORGANISM STUDIES

understand galactosemia. Animal studies typically involve high-galactose diets overloading the galactose metabolic pathway. Additionally, Galt knockout mice have been generated and studied (Leslie et al., 1996). However, Galt null mice appear to be normal and to have no evidence of neonatal toxicity. Even after keeping the animals on a high-galactose diet for weeks, no obvious phenotype was observed. The knockout of other enzymes in this pathway, such as galactokinase, also fails to fully phenocopy the human condition. Hence, these data suggest the evolution of alternative galactose handling pathways in mice in addition to the classic Galt-mediated oxidation. Interestingly, there are more examples of such discordance in metabolism. In contrast with the galactose metabolism pathway, urea synthetic pathways have been relatively well conserved in mice and humans. Urea cycle disorders (UCDs) are caused by the deficiencies of the enzymes required for transferring nitrogen from ammonia and aspartate into urea. Initial signs of UCDs in infants include somnolence, poor feeding, vomiting, seizures, lethargy, and coma. Persistent hyperammonemia, if not treated, may cause irreversible neuronal damage. Knockout mice for most of the enzymes of the UCDs have been generated, and they faithfully replicate the neonatal hyperammonemic phenotype. However, whether they might also replicate the long-term morbidity of UCDs that is not directly attributable to hyperammonemia is unknown because of the perinatal lethality of the mutants. Animal models of inborn errors of metabolism have proven in general to be excellent models for the development of therapy. Numerous metabolic disease models have been used for the development of protein replacement and gene therapies, and substantial long-term correction has been observed in many of them. Where they have often failed is in predicting the host toxicity to therapy. Hence, there are different therapeutic indices for a given specific therapy when moving from small-animal to large-animal models. This is because toxicity to treatment—especially to biologic therapies—cannot always be predicted reliably in rodents. Ultimately, although small-animal models such as rodents are excellent from the perspective of assessing efficacy, toxicities should be evaluated in nonhuman primate models and humans in phase I settings. Examples of successes in genetic therapies developed with the help of mouse models include protein replacement therapies for several of the lysosomal storage diseases, including the mucopolysaccharidoses. Still, there are examples in which the mouse model does not replicate human pathology, and this has been a major obstacle to the development of therapy. This is most prominently reflected in the cystic fibrosis transmembrane regulator mouse model of cystic fibrosis, in which lung pathology is absent (Snouwaert et al., 1992).

III. PERSPECTIVES Integrating the study of human genetic disease with comparative analyses of model organisms has proven to be a powerful approach to elucidating basic developmental mechanisms, understanding the pathogenesis of disease, and testing novel therapeutic approaches. The choice of model organism needs to consider the nature of the studied pathway, whether the mutation is cell autonomous, and the contribution of interorgan interactions. For highly conserved cell-autonomous mechanisms (e.g., the cell cycle, genome stability cilia), the choice of lower-order model organisms may facilitate the rapid dissection of

145

PERSPECTIVES

components of the developmental and cellular pathway. For complex behavioral phenotypes or later-evolved genetic networks (e.g., the regulation of skeletal development), higher-order mammalian models may be required. Ultimately, the analysis must take into account gene–gene and gene–environment interactions and the early versus the late effects of the target gene. The availability of genomic information and molecular tool kits for gene targeting and replacement has rapidly accelerated the generation of loss- and gain-offunction models. A major goal for further facilitating the translation of information between species will be approaches for rapidly introducing single nucleotide variants into the model genomes so that an allelic series can be quickly obtained. Another important goal would be to extend the ability to introduce loss- and gain-of-function alleles into the germline of larger-animal models. Ultimately, if the pace of generating models can be further accelerated, a bottleneck will still remain in the characterization of resulting phenotypes. Doubtlessly, success with this challenge will still depend on old-fashioned hard work, intuition, insight, and some degree of luck. Ultimately, the ability to begin to integrate global methods of phenotypic analyses into mechanistic hypotheses that are testable will be critical in completing the loop of translational research between humans and model organisms (Figure 7.1). The goal of modeling human disease is to enable the testing of hypotheses involving structure and function. These may be deduced from the study of genotype–phenotype correlations in human disease. Translation from human to model organism study should result in the formulation of a genetic and biochemical pathway that both explains the phenotype and generates testable predictions. The translation of mechanistic information back to humans would then involve the testing of these predictions in humans. In closing this loop, we should gain information about a basic developmental process as well as the consequence of dysregulation of this process in humans. The goal of clinical translation would be to develop methodologies to measure and ultimately rescue this pathway.

FIGURE 7.1 insert).

The loop of translational research between human and model organisms. (See color

146

UNDERSTANDING HUMAN BIRTH DEFECTS THROUGH MODEL ORGANISM STUDIES

SUMMARY

 Developmental defects can stem from structural metabolic defects or

  

 

exposure to environmental agents. Model organisms are indispensable tools for understanding the mechanisms of complex biologic pathways and human diseases. The choice of model organism needs to consider the nature of the studied pathway, whether the mutation is cell autonomous, and the contribution of interorgan interactions. As a result of the availability of genomic information, a powerful genetic toolkit, and the relative conservation of organogenesis, the mouse model has been the most used for studying human developmental disorders. Factors that affect the degree of phenotypic correlation between mice and humans include the following: 1) differences in the dosage sensitivity of the mutation and the affected pathway; 2) differences in redundant pathways and genes; 3) genetic and epigenetic modifiers specific for inbred strains of mice versus outbred humans; 4) true divergent function of the target gene in mouse versus humans; and 5) differential effects seen early versus later in development. Comparative human and model organism studies should facilitate a translational loop that enables the elucidation of pathogenic mechanisms and testable predictions that can be validated back in the human condition. Technological advances that will accelerate comparative studies include the ability to rapidly generate single nucleotide substitutions in the germ cell, high-throughput methods, and the associated informatic tools to characterize tissue phenotypes on a molecular level.

GLOSSARY OF TERMS Choreoathetosis Sudden involuntary movements of the limbs and the facial muscles. Hyperuricemia A high level of uric acid in the blood. Organic acidemia A class of inherited metabolic disorders that lead to the accumulation of organic acids in biologic fluids (blood and urine). Pleiotropic effect A single gene that produces multiple phenotypic traits. Skeletal dysostosis The abnormal formation of bone caused by the lack of proper ossification.

REFERENCES Adams MD, Celniker SE, Holt RA, et al: The genome sequence of Drosophila melanogaster, Science 287:2185–2195, 2000. Barr MM: Super models, Physiol Genomics 13:15–24, 2003. Bearer CF: Mechanisms of brain injury: L1 cell adhesion molecule as a target for ethanol-induced prenatal brain injury, Semin Pediatr Neurol 8:100–107, 2001.

REFERENCES

147 Bedell MA, Jenkins NA, Copeland NG: Mouse models of human disease. Part I: techniques and resources for genetic analysis in mice, Genes Dev 11:1–10, 1997. Bi W, Deng JM, Zhang Z, et al: Sox9 is required for cartilage formation, Nat Genet 22:85–89, 1999. Bi W, Huang W, Whitworth DJ, et al: Haploinsufficiency of Sox9 results in defective cartilage primordia and premature skeletal mineralization, Proc Natl Acad Sci U S A 98:6698–6703, 2001. Blader P, Strahle U: Ethanol impairs migration of the prechordal plate in the zebrafish embryo, Dev Biol 201:185–201, 1998. Bonadio J, Saunders TL, Tsai E, et al: Transgenic mouse model of the mild dominant form of osteogenesis imperfecta, Proc Natl Acad Sci U S A 87:7145–7149, 1990. Colvin JS, Bohne BA, Harding GW, et al: Skeletal overgrowth and deafness in mice lacking fibroblast growth factor receptor 3, Nat Genet 12:390–397, 1996. Cooper MK, Porter JA, Young KE, Beachy PA: Teratogen-mediated inhibition of target tissue response to Shh signaling, Science 280:1603–1607, 1998. Culetto E, Sattelle DB: A role for Caenorhabditis elegans in understanding the function and interactions of human disease genes, Hum Mol Genet 9:869–877, 2000. Dahme M, Bartsch U, Martini R, et al: Disruption of the mouse L1 gene leads to malformations of the nervous system, Nat Genet 17:346–349, 1997. Davies SA, Stewart EJ, Huesmann GR, et al: Neuropeptide stimulation of the nitric oxide signaling pathway in Drosophila melanogaster Malpighian tubules, Am J Physiol 273:R823–R827, 1997. Edison RJ, Muenke M: Central nervous system and limb anomalies in case reports of first-trimester statin exposure, N Engl J Med 350:1579–1582, 2004a. Edison RJ, Muenke M: Mechanistic and epidemiologic considerations in the evaluation of adverse birth outcomes following gestational exposure to statins, Am J Med Genet 131:287–298, 2004b. Egelhoff TT, Lee RJ, Spudich JA: Dictyostelium myosin heavy chain phosphorylation sites regulate myosin filament assembly and localization in vivo, Cell 75:363–371, 1993. Ehlers K, Sturje H, Merker HJ, Nau H: Spina bifida aperta induced by valproic acid and by alltrans-retinoic acid in the mouse: distinct differences in morphology and periods of sensitivity, Teratology 46:117–130, 1992. Eichinger L, Pachebat JA, Glockner G, et al: The genome of the social amoeba, Dictyostelium discoideum, Nature 435:43–57, 2005. Eickholt BJ, Towers GJ, Ryves WJ, et al: Effects of valproic acid derivatives on inositol triphosphate depletion, teratogenicity, glycogen synthase kinase-3beta inhibition, and viral replication: a screening approach for new bipolar disorder drugs derived from the valproic acid core structure, Mol Pharmacol 67:1426–1433, 2005. Finnell RH, Waes JG, Eudy JD, Rosenquist TH: Molecular basis of environmentally induced birth defects, Annu Rev Pharmacol Toxicol 42:181–208, 2002. Fitzky BU, Witsch-Baumgartner M, Erdel M, et al: Mutations in the Delta7-sterol reductase gene in patients with the Smith-Lemli-Opitz syndrome, Proc Natl Acad Sci U S A 95:8181–8186, 1998. Fransen E, Van Camp G, D’Hooge R, et al: Genotype-phenotype correlation in L1 associated diseases, J Med Genet 35:399–404, 1998. Goffeau A, Barrell BG, Bussey H, et al: Life with 6000 genes, Science 274:546, 1996563-547. Jinnah HA, Wojcik BE, Hunt M, et al: Dopamine deficiency in a genetic mouse model of LeschNyhan disease, J Neurosci 14:1164–1175, 1994. Jorgensen EM, Mango SE: The art and design of genetic screens: Caenorhabditis elegans, Nat Rev Genet 3:356–369, 2002. Justice MJ, Noveroske JK, Weber JS, et al: Mouse ENU mutagenesis, Hum Mol Genet 8:1955–1963, 1999. Komori T, Yagi H, Nomura S, et al: Targeted disruption of Cbfa1 results in a complete lack of bone formation owing to maturational arrest of osteoblasts, Cell 89:755–764, 1997. Lammer EJ, Chen DT, Hoar RM, et al: Retinoic acid embryopathy, N Engl J Med 313:837–841, 1985. Lanpher B, Brunetti-Pierri N, Lee B: Inborn errors of metabolism: the flux from Mendelian to complex diseases, Nat Rev Genet 7:449–460, 2006. Leslie ND, Yager KL, McNamara PD, Segal S: A mouse model of galactose-1-phosphate uridyl transferase deficiency, Biochem Mol Med 59:7–12, 1996.

148

UNDERSTANDING HUMAN BIRTH DEFECTS THROUGH MODEL ORGANISM STUDIES

Li G, Alexander H, Schneider N, Alexander S: Molecular basis for resistance to the anticancer drug cisplatin in Dictyostelium, Microbiology 146 (Pt 9), 2219–2227, 2000. Liu W, Toyosawa S, Furuichi T, et al: Overexpression of Cbfa1 in osteoblasts inhibits osteoblast maturation and causes osteopenia with multiple fractures, J Cell Biol 155:157–166, 2001. Marsh JL, Walker H, Theisen H, et al: Expanded polyglutamine peptides alone are intrinsically cytotoxic and cause neurodegeneration in Drosophila, Hum Mol Genet 9:13–25, 2000. Naski MC, Wang Q, Xu J, Ornitz DM: Graded activationof fibroblast growth factor receptor 3 by mutations causing achondroplasia and thanatophoric dysplasia, Nat Genet 13:233–237, 1996. Otto F, Thornell AP, Crompton T, et al: Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development, Cell 89:765–771, 1997. Proetzel G, Pawlowski SA, Wiles MV, et al: Transforming growth factor-beta 3 is required for secondary palate fusion, Nat Genet 11:409–414, 1995. Reiter LT, Potocki L, Chien S, et al: A systematic analysis of human disease-associated gene sequences in Drosophila melanogaster, Genome Res 11:1114–1125, 2001. Roessler E, Belloni E, Gaudenz K, et al: Mutations in the human Sonic Hedgehog gene cause holoprosencephaly, Nat Genet 14:357–360, 1996. Rosati R, Horan GS, Pinero GJ, et al: Normal long bone growth and development in type X collagen-null mice, Nat Genet 8:129–135, 1994. Snouwaert JN, Brigman KK, Latour AM, et al: An animal model for cystic fibrosis made by gene targeting, Science 257:1083–1088, 1992. Stacey A, Bateman J, Choi T, et al: Perinatal lethal osteogenesis imperfecta in transgenic mice bearing an engineered mutant pro-alpha 1(I) collagen gene, Nature 332:131–136, 1988. Takeda S, Bonnamy JP, Owen MJ, et al: Continuous expression of Cbfa1 in nonhypertrophic chondrocytes uncovers its ability to induce hypertrophic chondrocyte differentiation and partially rescues Cbfa1-deficient mice, Genes Dev 15:467–481, 2001. Wang TR, Wang WP, Hwu WL, Lee ML: Fibroblast growth factor receptor 3 (FGFR3) gene G1138A mutation in Chinese patients with achondroplasia, Hum Mutat 8:178–179, 1996. Watanabe H, Yamazaki M, Miyazaki H, et al: Phospholipase D2 functions as a downstream signaling molecule of MAP kinase pathway in L1-stimulated neurite outgrowth of cerebellar granule neurons, J Neurochem 89:142–151, 2004. Waterham HR, Wijburg FA, Hennekam RC, et al: Smith-Lemli-Opitz syndrome is caused by mutations in the 7-dehydrocholesterol reductase gene, Am J Hum Genet 63:329–338, 1998. Waterston RH, Lindblad-Toh K, Birney E, et al: Initial sequencing and comparative analysis of the mouse genome, Nature 420:520–562, 2002. Zheng Q, Sebald E, Zhou G, et al: Dysregulation of chondrogenesis in human cleidocranial dysplasia, Am J Hum Genet 77:305–312, 2005.

FURTHER READING Model organisms http://www.ncbi.nlm.nih.gov/About/model/ Saccharomyces Genome Database http://www.yeastgenome.org/ Wormbase (C. elegans) http://www.wormbase.org/ Flybase (Drosophila) http://www.flybase.org/ Mouse genome informatics http://www.informatics.jax.org/

II EARLY EMBRYOLOGY, FATE DETERMINATION, AND PATTERNING

8

GERM LINE DETERMINANTS AND OOGENESIS KELLY M. HASTON and RENEE A. REIJO PERA Institute for Stem Cell Biology and Regenerative Medicine, Stanford University, Palo Alto, CA

INTRODUCTION Germ cells are the only cells in the body that are destined to pass genetic information from one generation to the next; by contrast, somatic cells give rise to cells of the body that are ultimately destined to die. Thus, the allocation (also called specification) of germ cell versus somatic cell fate is of primary importance to all species; it occurs very early in embryo development (Saffman and Lasko, 1999). Despite the critical importance of germ cell development to all species, however, two divergent methods of germ cell specification and maintenance are apparent in animals (Saffman and Lasko, 1999; Houston and King, 2000; Wylie, 2000). First, in nonmammalian species, germ cell fate is determined by the inheritance of germ plasm, which is microscopically distinct oocyte cytoplasm that is particularly rich in RNAs and RNA-binding proteins and that segregates with cells destined to become germ cells (Saffman and Lasko, 1999; Houston and King, 2000; Wylie, 2000). By contrast, in mammalian species, both male and female germ cells are specified independently of germ plasm via inductive signaling from neighboring cells (Lawson and Hage, 1994; Tam and Zhou, 1996; Lawson et al., 1999; McLaren, 1999; Ying, 2000; Ying et al., 2001; Yoshimizu, 2001). Both modes of specification are discussed in more detail below. We then move on to discuss the migration and maturation of gametes, with the later discussion particularly focusing on oogenesis.

I. GERM CELL SPECIFICATION A. Germ-Plasm–Dependent Specification of the Germ Cell Lineage Specialized cytoplasm, called germ plasm, is found in diverse nonmammalian species that include Caenorhabditis elegans (nematodes), Drosophila melano-

150

Principles of Developmental Genetics

© 2007, Elsevier Inc. All rights reserved.

GERM CELL SPECIFICATION

151

gaster (flies), Xenopus laevis (frogs), and Danio rerio (fish) (Wylie, 2000; Zhou and King, 2004). The germ plasm is observed in oocytes and cleavage-stage embryos; it is microscopically dense; and it is enriched greatly in RNAs, RNA-binding proteins, mitochondria, and ribosomes (Wylie, 2000; Zhou and King, 2004; see Figure 8.1). Germ plasm is referred to by different names in different species; it is referred to as P granules in C. elegans, pole plasm in D. melanogaster, and simply as germ plasm in X. laevis. Nonetheless, several common properties define germ plasm. First, in organisms that specify germ cell fate via germ plasm, the decision to segregate the germ cell from somatic lineages occurs before gastrulation. Second, the germ plasm or germline granules segregate at all times with cells of the germ cell lineage, in both embryonic and postembryonic development (see Figure 8.1). Indeed, germ plasm plays a determinative role: cells that inherit germ plasm develop as germ cells, whereas, in the absence of germ plasm, germ cells do not develop. Finally, as noted by several reviewers, despite the differences between the two processes of germ cell allocation, a number of factors that retain their function in germ cell development are highly conserved between organisms that specify germ cells via germ plasm and those that do not. For example, in Drosophila, the disruption of genes such as Oskar, Vasa, Tudor, Germ cell-less, and Aubergine results in the lack of a germ line (Santos and Lehmann, 2004). These genes function to assemble the germ plasm, in which the highly conserved interacting RNA-binding proteins Pumilio and Nanos are localized. These proteins in turn repress translation and thus indirectly silence gene transcription in nascent germ cells (Lin and Spradling, 1997; Forbes and Lehmann, 1998; Parisi and Lin, 1999; Santos and Lehmann, 2004). In Pumilio and Nanos mutants, nascent germ cells may divide prematurely, migrate abnormally, or subsequently die during early embryo development (Jaruzelska et al., 2003; Tsuda et al., 2003; Santos and Lehmann, 2004). Recently, the homologues of these and other germ plasm components have been identified in mammalian germ cells, and, in many cases, these are required for germ cell development, despite the fact that, in mammals, germ cells are not specified via germ plasm in the oocyte and cleavage-stage embryos. B. Germ-Plasm–Independent Germ Cell Specification (Inductive Signaling) Initially, in mice, a founder population of approximately 45 primordial germ cells (PGCs) is formed (Chiquoine, 1954; Ginsburg et al., 1990). Fate-mapping studies have been used to examine germ cell specification in mammals, and they have revealed that germ cells are specified in the proximal epiblast in mice (Tam and Zhou, 1996; see Figure 8.2) in response to signals from the neighboring extraembryonic ectoderm, particularly Bmp4 signaling (Fujiwara et al., 2001; see Figure 8.2). However, it is notable that the proximal epiblast is not predestined to a germ cell fate, because transplantation of the distal epiblast to contact the extraembryonic ectoderm also results in germ cell formation (Tam and Zhou, 1996). Furthermore, the fate of proximal epiblast cells is ultimately to form both germ cells and extraembryonic mesoderm. Thus, it is likely that the extraembryonic ectoderm provides one of the first signals for germ cell specification in the epiblast. Then, a second as yet uncharacterized signal must be required to distinguish extraembryonic mesoderm from germ cells. Germ cells are definitively recognized after gastrulation, at 7.2 days post coitum, as an extraembryonic cluster of cells at the base of the allantois that express tissue

152

GERM LINE DETERMINANTS AND OOGENESIS

nonspecific alkaline phosphatase, Oct4, and Stella (Chiquoine, 1954; Scholer et al., 1990; Scholer et al., 1990; Saitou et al., 2002; see Figure 8.2). Notably, although epiblast cells migrate through the primitive streak during gastrulation, the physical act of migration does not appear to be necessary for defining germ cell versus somatic cell fates (Ying et al., 2001; Yoshimizu, 2001; Pesce et al., 2002). The search for molecules that are required for the specification of germ cells in mammals has been intriguing. Early studies suggested that perhaps a gene called Stella may be a determinant of the germ cell lineage (Saitou et al., 2002). However, disruption of the mouse Stella gene indicated that the gene encoded a factor that was required for embryo growth; indeed, embryos derived from oocytes that are null for Stella do not develop (Payer et al., 2003; Bortvin et al., 2004). Subsequently, Ohinata and colleagues (2005) identified the Blimp1 gene as a critical determinant of the germ cell lineage. Although the Blimp1 gene encodes a transcriptional repressor that is widely expressed during development, its function was shown to be required for the establishment of the primordial germ cell population (via repression of somatic Hox genes) and for the subsequent migration and proliferation of germ cell populations (Ohinata et al., 2005; see Figure 8.2).

II. CONSERVED GENES IN ORGANISMS THAT SPECIFY GERM CELLS VIA GERM PLASM AND INDUCTIVE MECHANISMS In the description above, the specification of germ cells via germ plasm and via inductive signaling was contrasted. However, as the DNA sequences of multiple organisms have been assembled and reproductive biologists have probed gene function across species, it has become clear that many key genes that function in establishing and maintaining germ cell populations are conserved. In particular, family members of genes such as Vasa, Pumilio, Nanos, and Deleted in AZoospermia (DAZ) have homologs in diverse organisms. A. Vasa Homologs of the Vasa gene family encode RNA-binding proteins of the DEAD-box helicase family, which are specifically expressed in germ cells in all animals examined. The Vasa gene was first identified as a maternal-effect gene that is required for the proper establishment of abdominal segments and for the formation of the pole cells in Drosophila (Schu¨pbach and Wieschaus, 1986). Subsequently, the gene was also shown to function in oogenesis (Styhler et al., 2002). Studies in diverse species have verified that this gene has a function in germ cell development; however, the phenotypes may differ from species to species. For example, in mice, the disruption of mouse Vasa homolog leads to the meiotic arrest of male germ cells, whereas there is a reduction in premeiotic germ cells in female flies and an arrest in the pachytene stage of oogenesis in nematodes (Tanaka et al., 2000; Kuznicki et al., 2000; Styhler et al., 2002). B. Pumilio and Nanos The Pumilio and Nanos genes are among the most well-characterized genes in invertebrates. In Drosophila, these genes encode interacting proteins that are

CONSERVED GENES IN ORGANISMS THAT SPECIFY GERM CELLS VIA GERM PLASM

153

required for the formation of nascent germ cells and for the establishment of the anterior–posterior axis (Kobayashi et al., 1996; Wreden et al., 1997; Asaoka et al., 1998; Asaoka-Taguchi et al., 1999; Deshpande et al., 1999). Then, later in development, these genes may be required for oogenesis as well (Forbes and Lehmann, 1998). In C. elegans, there are numerous members of the Pumilio, Fbf, and Nanos gene families (Subramaniam and Seydoux, 1999; Wickens et al., 2002; Subramaniam and Seydoux, 2003). Where studied, these homologs also function in gametogenesis, with some homologs encoding proteins that interact to promote the spermatogenesis-to-oogenesis switch and others required for the incorporation of germ cells into the gonad or the progression of germ cell development through meiosis (Kraemer et al., 1999; Subramaniam and Seydoux, 1999; Crittenden et al., 2002; Subramaniam and Seydoux, 2003). Pumilio and Nanos homologs have also been identified in mammals, including humans (Rongo et al., 1997; Castrillon et al., 2000; Tanaka et al., 2000; Mochizuki et al., 2001; Jaruzelska et al., 2003; Moore et al., 2003; Tsuda et al., 2003). Indeed, a number of factors that interact in invertebrate germ cells have also been shown to interact in vertebrates, including the Pumilio and Nanos proteins (Jaruzelska et al., 2003; Moore et al., 2003; Tsuda et al., 2003). Moreover, the loss of function of some homologs results in infertility; in particular, the loss of function of Nanos2 results in the defective development of male germ cells, and the loss of Nanos3 function results in the impaired maintenance of PGCs during migration in both sexes (Tsuda et al., 2003). Other known vertebrate Nanos homologs include Xcat-2 in Xenopus and Nos1 and Nos2 in zebrafish; the products of these genes have been shown to localize to germ plasm in these species, and zebrafish Nos1 has been shown to be required for PGC migration and survival (Mosquera et al., 1993; Koprunner et al., 2001). C. Deleted in AZoospermia Human DAZ was identified in a screen for Y chromosome genes that cause azoospermia (or the production of few or no germ cells) when deleted in men (Reijo et al., 1995; Reijo et al., 1996). Subsequently, autosomal homologs called Deleted in AZoospermia-Like (DAZL) were identified in mice and humans, and they were also shown to be expressed only in germ cells (Cooke et al., 1996; Reijo et al., 1996; Yen et al., 1996; Menke et al., 1997). Finally, a third homolog called boule was identified as a meiotic regulator. It was hypothesized to be the ancestral member of this family, because this gene is conserved in both vertebrates and invertebrates (Eberhart et al., 1996; Xu et al., 2001; Xu et al., 2003). Notably, all members of the DAZ gene family encode RNA-binding proteins that contain a highly conserved RNA recognition motif, and they may bind several different RNAs during the development of male and female germ cells pre- and postmeiotically (Houston et al., 1998; Venables and Eperon, 1999; Houston and King, 2000; Tsui et al., 2000; Tsui et al., 2000; Venables et al., 2001; Jiao et al., 2002; Collier et al., 2005; Fox et al., 2005; Reynolds et al., 2005). The loss of function of members of this gene family demonstrates that they function in germ plasm, in meiosis, and postmeiotically in different organisms (Ruggiu et al., 1997; Houston et al., 1998; Houston and King, 2000; Karashima et al., 2000; Saunders et al., 2003; Dann et al., 2006). In addition, studies of DAZL

154

GERM LINE DETERMINANTS AND OOGENESIS

function in humans have demonstrated that, despite their key role in germ cell development in numerous species genes, the DAZ and DAZL genes are among the most variable in the human genome, with variants that correlate with reproductive parameters in both men and women (Teng et al., 2006; Tung et al., 2006a; 2006b).

III. GERM CELL MIGRATION After specification and just before or during the early stages of gastrulation, germ cell migration occurs. PGCs migrate out of the embryo proper and reside in extraembryonic tissues until gastrulation is complete. PGC migration in the mouse occurs between embryonic days 7.5 and 13.5, when the PGCs travel through the developing gut to become incorporated into the primitive gonad (see Figure 8.2). The migration of PGCs is a multistep process during which the cells migrate through different tissue types and environments (Anderson et al., 2000; Molyneaux et al., 2001; Molyneaux and Wylie, 2004; Santos and Lehmann, 2004). Both migratory and survival signals appear to be required during this period for successful PGC development. After embryonic day 7.5, there is a distinct population of cells that appear to have a germ-cell– specific gene-expression profile. Several factors are involved in germ cell survival as PGCs progress along their migratory route. For example, the gene Dead end, which is required for the initiation of migration in zebrafish, appears to act in the mouse as a survival factor. Mice with mutations in this gene display a decrease in PGC number by embryonic day 8.0 and have increased testicular germ cell tumors detected after birth (Youngren et al., 2005). Another gene, Tiar, encodes an RNA-binding protein that is necessary for PGC survival and that has been implicated as a regulator of apoptosis. The Tiar protein is required on embryonic day 11.5 during migration from the hindgut to the genital ridge, and mice lacking this protein fail to develop oogonia or spermatogonia (Beck et al., 1998). Furthermore, the growth factor Fgf2, its receptor Fgf2-IIIb, and other genes (e.g., the antiapoptotic gene Bax) are required for the survival of germ cells during the migration from the hindgut to the genital ridge (Sette et al., 2000; Stallock et al., 2003; Takeuchi et al., 2005).

Germ Plasm

cell atic Som ages line

Somatic Cells

Ger m linea line ge Oocyte

Two Cell Embryo

Four Cell Embryo

Egg or Sperm

The pregastrulation specification of germline fate through the inheritance of germ plasm. Oocytes from species with predetermined germ cell specification contain a microscopically dense complex that is enriched in RNAs, RNA-binding proteins, mitochondria, and ribosomes. During the earliest embryonic cell divisions, germ plasm segregates to cells that will eventually give rise to the germ cell lineage. Cells that do not contain germ plasm give rise to somatic lineages. (See color insert.)

FIGURE 8.1

155

GERM CELL MIGRATION

Specification

Migration

Colonization

BMP Pre-PGCs in proximal epiblast

E6.25

E7.5

PGCs in extraembryonic mesoderm

E8.5

PGCs in endoderm

Fetal gonad

Genital ridge

E10.5

E13.5

Blimp- 1

Fragilis Oct 4 Stella TNAP C-Kit Dead end Nanos 3 B1 Integrins CXCR4/SDF1 Dazl Vasa

FIGURE 8.2 Primordial germ cell (PGC) development in the mouse. Stages of PGC development during the postfertilization mouse embryo, from embryonic days 6.25 to 13.5. Displayed below the figure is the PGC-specific expression of genes that are either required for or diagnostic of the specified stage of PGC development. PGCs are displayed as red dots, which are specified in the proximal epiblast (blue) at embryonic day 6.25 in response to bone morphogenetic protein signaling from the extraembryonic tissues (green arrows). PGCs then migrate into the extraembryonic mesoderm by embryonic day 7.5. By embryonic day 8.5, they have entered the definitive endoderm, and they begin their migration down the developing hindgut. They colonize the primitive fetal gonad between embryonic days 11.5 and 13.5. (Modified from National Institutes of Health: Stem cells: scientific progress and future research directions [Web site]: stemcells.nih. gov/info/scireport/2001report.htm. Accessed December 11, 2006. See color insert.)

Although some survival signals required during the earliest stages of PGC migration have been identified, early migratory steps are poorly understood, and few factors have been identified that clarify what triggers these events. However, several mediators during the later stages of migration involving the hindgut-to-genital-ridge transition have been resolved. On approximately embryonic day 8.0, the PGC population migrates out of the extraembryonic mesoderm and enters the embryonic (definitive) endoderm. Although they display motile behavior both in vivo and in vitro, they do not actively migrate at this point; rather, they are carried along with the endoderm as the hindgut invaginates between embryonic days 8.5 and 9.0, and, by embryonic day 9.0, PGCs are in the hindgut epithelium (Anderson et al., 2000; Molyneaux et al., 2001; Molyneaux and Wylie, 2004). Between embryonic days 9.0 and 9.5, PGCs emerge from the hindgut and migrate into the developing genital ridge. On embryonic day 10.5, the PGCs are moving toward the genital ridges from divergent sites, and PGCs that do not reach the genital ridge undergo apoptosis. The migration of PGCs from the hindgut to the genital ridge is an essential step in germ cell development. Factors involved in this step include the RNA-binding protein Nanos3 (described previously), the adhesion molecules b1 integrins, the forkhead/winged helix transcription factor Foxc1, and the G-protein–coupled receptor CXCR4 and its ligand SDF-1 (Anderson et al.,

156

GERM LINE DETERMINANTS AND OOGENESIS

1999; Anderson et al., 2000; Ara et al., 2003; Molyneaux et al., 2003; Tsuda et al., 2003; Mattiske et al., 2006). In particular, Nanos3 is expressed in PGCs by embryonic day 9.5, and the loss of this protein leads to a decline of PGC number by embryonic day 11.5 in both sexes (Tsuda et al., 2003; see Figure 8.2). Similarly, the loss of b1 integrins leads to a decline in PGC number before they reach the genital ridges, although the mechanism for this reduction is not understood (Anderson et al., 1999). Foxc1 mutants also result in the failure of PGCs to exit the hindgut (Mattiske et al., 2006). Finally, the G-protein–coupled receptor CXCR4 is expressed on the surface of the PGCs, and both the body wall mesenchyme and the genital ridges express SDF-1; the loss of either the receptor or the ligand leads to the failure of PGCs to reach the genital ridge, and this is followed by cell death of ectopic PGCs in the hindgut (Ara et al., 2003; Molyneaux et al., 2003; see Figure 8.2). Although this is not an exhaustive list, it is clear from this brief overview of factors involved in hindgut-to-genital-ridge migration that both PGC autonomous and nonautonomous interactions are required for both the earlier and later stages of PGC survival and migration. By embryonic day 11.5, the majority of the PGCs (~25,000) are in the genital ridges, where they are becoming nonmotile and beginning to aggregate into sex-specific organization with the somatic tissue (Anderson et al., 2000; Molyneaux et al., 2001; Molyneaux and Wylie, 2004; see Figure 8.2).

IV. GERM CELL SEX DETERMINATION After the germ cells reenter the embryo and migrate into and invade the genital ridges, they will colonize, proliferate mitotically, complete the process of resetting the genomic imprints (described later) by erasure (on embryonic days 9.5 to 11.5), and differentiate as either male or female germ cells (Gomperts et al., 1994; Hajkova et al., 2002). Before entry into the gonad, the development of male and female germ cells has been indistinguishable; subsequently, however, germ cell development in the male and female gonads will diverge (Swain, 2006). Germ cells that colonize an ovary will enter the first meiotic prophase and become oogonia at approximately embryonic day 13.5, whereas those that colonize the testis will not progress to meiotic prophase and will instead mitotically divide to form a pool of spermatogonia that may replicate or differentiate throughout the life of the male (Swain, 2006). Several reports document early events in the sex determination of mammalian germ cells (Menke et al., 2003; Bowles et al., 2006; Koubova et al., 2006). The differentiation of cells to female or male germ cells occurs independently of sex chromosome composition and instead is dependent on gonadal sex (Swain, 2006). An early molecular marker of female sexual differentiation is the expression of the Stra8 gene, which is expressed in an anteriorto-posterior wave in germ cells in the ovary between embryonic days 12.5 and 16.5 (Menke et al., 2003). Coincident with or just after Stra8 expression, the downregulation of the Oct4 gene (a marker of germ cells that is most highly expressed before meiosis) and the upregulation of Dmc1 (a gene that encodes a meiotic protein) ensues (Menke et al., 2003). These observations further suggest that local signals may regulate the sexual differentiation of the germ cells in an anterior-to-posterior fashion and promote meiotic entry (for more information about meiosis is provided later in this chapter).

GENOMIC IMPRINTING

157

Early studies suggested that both male and female germ cells are inherently programmed to enter meiosis in the fetal gonad and that the inhibition of meiosis via a hypothesized “meiosis-inhibiting factor” must be a critical event in male germ cells (McLaren and Southee, 1997). Moreover, data indicated that one of the factors implicated in regulating the sex-specific timing of meiotic initiation in mice is retinoic acid, which may act by regulating the expression of Stra8 (Koubova et al., 2006). Furthermore, in the male, the identity of a factor that may act as the meiosis-inhibiting factor was suggested with the disruption of a gene that encodes the retinoid-degrading enzyme CYP26B1; testis in which this gene is disrupted possess germ cells that enter meiosis precociously, as is expected of germ cells in the ovary (Bowles et al., 2006). Together, this work provides the first molecular explanation of how germ cell sexual differentiation is regulated in mammals.

V. GENOMIC IMPRINTING Genomic imprinting is another fundamental property of mammalian germ cells. Epigenetic modifications to the genome (typically by DNA methylation) result in the expression of genes from only one of the two parental chromosomes. The “instructions” regarding gene expression are established in the parental germ cells via differential methylation of the DNA. Methylation predominantly occurs on the cytosine residue of CpG dinucleotides, which typically cluster together to form “CpG islands” (Reik and Walter, 2001). The establishment of these imprints in germ cells is essential for fetal, placental, and behavioral development (Reik and Walter, 2001). The misregulation of imprinted genes, which leads to biallelic expression, has been implicated in growth and neuronal disorders in multiple mammalian species, including humans (Allegrucci et al., 2004; Kelly and Trasler, 2004; see Chapter 5). Interestingly, genomic imprints are altered throughout the life cycle of the organism (Reik and Walter, 2001; Hajkova et al., 2002). Upon fertilization, imprinting is maintained through the replication and segregation of chromosomes during development. Although the mechanism for the maintenance of specific regions of DNA is not fully understood, one of the five mammalian DNA methyltransferases, Dnmt1, has been identified as a maintenance methyltransferase that is essential for the maintenance of methylation during DNA replication (Bestor, 2000). As germ cells develop in the new organism, genomic imprints in PGCs are erased by a wave of demethylation, which occurs as they are arriving at the primitive gonadal ridge. It is at this stage that an active erasure of methylation at the imprinted regions occurs by an unidentified demethylation agent. Imprints are then reestablished in a sex-specific manner in the PGCs through de novo methylation by Dnmt3a, Dnmt3b, and Dnmt3L (Okano, 1999; Bourc’his et al., 2001; Hata et al., 2002). The imprints are established in germ cells as they mature into sperm or eggs at different time points for female versus male imprints. Imprints in prospermatogonia are imposed before they enter meiosis, whereas imprints in oocytes are reestablished later, at different stages of oogenesis, in a gene-specific manner (Bestor and Bourc’his, 2004). The reestablishment of sex-specific imprints is an essential step in gametogenesis, which completes the cycle toward a mature germ cell that is competent to give rise to viable offspring.

158

GERM LINE DETERMINANTS AND OOGENESIS

VI. MEIOSIS A. General Properties of Meiosis Meiosis is defined by a series of stages with characteristic landmark events, and it is highly conserved across species (Kleckner, 1996). Prophase I is a defining stage of meiosis that encompasses many unique features, including the formation of the synaptonemal complex, the pairing of homologous chromosomes, and the formation of chiasmata between homologs. These features are common to almost all species, and they are in place to ensure that homologous chromosomes pair and remain together until the first meiotic division (Kleckner, 1996). Prophase can be divided into four stages: leptotene, zygotene, pachytene, and diplotene. DNA replication begins in the preleptotene to leptotene stages, and sister chromatids begin to condense. Then, during zygotene, sister chromatids synapse along their length and form lateral elements that contain synaptonemal complex proteins (SCPs), such as SCP2 and SCP3. At pachytene, synaptonemal complex formation is complete, and recombination nodules that contain proteins such as MLH1 are clearly visible. Finally, at the diplotene stage, homologous chromosomes begin to separate, and only the chiasmata (the sites of the recombination machinery) hold the chromosomes together. Meiotic division continues to progress in an orderly fashion from meiosis I to meiosis II, unless errors in the recombination or chromosomal segregation machinery trigger arrest at one of two checkpoints, either during prophase or at the metaphase–anaphase transition (Roeder, 1997; Roeder and Bailis, 2000). Then, after meiosis, germ cells continue to develop through a process called spermiogenesis in males, during which they become mature elongated spermatids by the compaction of their chromatin into the sperm head, by the production of other sperm components (e.g., the flagellar tail), and by oocyte maturation in females (Hunt and Hassold, 2002). There are many genes involved in the initiation of and the progression through meiosis, and this has been a subject of intense study for several decades in many different organisms (Baker et al., 1976; Kleckner, 1996; Smith and Nicolas, 1998; Roeder and Bailis, 2000; Hunt and Hassold, 2002). The disruption of the function of meiotic genes generally leads to meiotic arrest and subsequent apoptosis of the germ cells or aneuploidy (Lahn and Page, 1997). B. Errors in Meiotic Chromosome Segregation Aneuploidy arises during meiosis I or II by nondisjunction or the premature separation of sister chromatids, and it is a rare event in most organisms. For example, rates of aneuploidy in meiotic cells have been reported as 1 in 10,000 cells for Saccharomyces cerevisiae, 1 in 6000 cells for D. melanogaster, and approximately 1 in 100 to 1 in 200 for mice (Hassold and Hunt, 2001). Surprisingly, the rate of aneuploidy in humans may be as high as 1 in 10 to 1 in 30, depending on factors such as age and sex (Hassold and Hunt, 2001). Aneuploidy is detected in approximately 5% of clinically recognized pregnancies. However, in general, most aneuploidies are eliminated early in gestation. Among fetal deaths occurring between about 6 to 8 weeks’ and 20 weeks’ gestation, about 35% are trisomic or monosomic; this rate decreases to about 4% among stillbirths (fetal deaths occurring between about 20 weeks’ gestation and term) and to about 0.3% among newborns (Hassold

159

OOGENESIS

and Hunt, 2001). The most common abnormalities in both stillbirths and newborns are trisomy 21 and sex chromosome trisomies (47,XXX; 47,XXY; 47,XYY). The overall rate of 5% aneuploidy in all human conceptions is likely an underestimate, because it does not include an analysis of undetected pregnancies during the first few weeks of gestation (Hassold and Hunt, 2001). Aneuploidy is also common in human gametes. Approximately 2% of human sperm are aneuploid, and 20% to 25% of oocytes are aneuploid (depending on age) (Hassold and Hunt, 2001). Studies of the origin of aneuploidy over the past decade have revealed that maternal errors predominate among almost all trisomies, with paternal errors accounting for nearly 50% of 47,XXYs and trisomy 2.

VII. OOGENESIS Oogenesis is the meiotic division of a diploid oocyte into a haploid ovum (egg). The process of oogenesis is dependent on the development of an ovarian follicle (folliculogenesis), which is the formation of the somatic cells surrounding the developing oocyte that will become the functional unit of the ovary

A

Primary Oocyte

2C

Primordial Follicle Squamous granulosa cells

B

Primary PrimaryOocyte Oocyte Arrested in Prophase of Meiosis I

2C

Cuboidal granulosa cells

Zona pellucida

C

Primary Oocyte Arrested in Prophase of MeiosisI I

Primary Follicle

Secondary Follicle 2C Theca cells

D

Antral/Ovulatory Follicle

Secondary Oocyte Meiosis I complete Arrested in Metaphase of Meiosis II

2C 1C First polar body

Cumulus cells

E

(degenerates)

Ovum Meiosis II complete

Fertilization 1C 1C Second polar body (degenerates)

Oogenesis and folliculogenesis. Oogenesis and folliculogenesis are intertwined processes that form the functional unit of the ovary. A, Both processes begin when the primary oocyte, which at this point is diploid (2C), becomes surrounded by squamous granulosa cells to make up the primordial follicle. B, The primary oocyte enters meiosis and becomes arrested during the prophase of meiosis I as the granulosa cells of the follicle transition from squamous to cuboidal to form a primary follicle. During this process, the zona pellucida or egg coat is formed around the primary oocyte. C, The primary oocyte maintains its meiotic arrest, whereas the primary follicle develops into the secondary follicle through the formation of multiple layers of granulosa cells surrounded by theca cells. D, After the formation of the antral follicle, the oocyte is ovulated; it is surrounded by cumulus cells, and it completes meiosis I, extruding the first polar body, which will degenerate. It is now a secondary oocyte, and it is arrested during the metaphase of meiosis II. E, Finally, upon fertilization, meiosis II is completed with the extrusion of the second polar body, which will also degenerate. This will leave an ovum, which is haploid (1C) and fully competent to support embryonic growth. (See color insert.)

FIGURE 8.3

160

GERM LINE DETERMINANTS AND OOGENESIS

(see Figure 8.3). The interaction between the oocyte and surrounding follicular structure is essential for the correct timing of events during oogenesis (Mehlmann, 2005). These events include the maintenance or release of the two periods of meiotic arrest that the oocyte experiences during its development, which occur during the prophase of meiosis I and the metaphase arrest of meiosis II. Because of the critical nature of the oocyte/follicle interaction, folliculogenesis is highly regulated by endocrine, paracrine, and autocrine factors (Roy and Matzuk, 2006). In vertebrates, oogenesis begins during embryogenesis, after the primordial germ cells migrate into the primitive ovary. These cells now complete mitotic divisions and become primary oocytes by entering meiosis and arresting during the diplotene phase of meiotic prophase I (see Figure 8.3, B). The first step in follicular development is dependent on the expression of a basic helixloop-helix transcription factor, factor in the germ line a (Figla), by the primary oocytes, which triggers the formation of a primordial follicle that consists of a single layer of flat, squamous, pregranulosa cells around the oocyte. Figla is essential in the formation of the primordial follicle, and Figla-null mice are infertile and do not have primordial follicles (Liang et al., 1997). In mice, the majority of primary oocytes are in primordial follicles within 1 to 2 days of birth, whereas, in humans, primordial follicles form at 19 weeks’ gestation (Choi and Rajkovic, 2006). After primordial follicles are established, further follicle development involves the periodic recruitment of a subset of the follicles into a maturation cycle either postnatally (in mice) or at puberty (in humans). Although the mechanism of follicle recruitment is not well understood, the process begins with the transition of a primordial follicle to a primary follicle, at which time the granulosa cells surrounding the oocytes now undergo a squamous-tocuboidal shape change and begin to undergo proliferation (see Figure 8.3, A and B). The oocyte-specific expression of the homeobox transcription factor Nobox is required for the primordial-to-primary transition to occur, and Nobox mutants display a block in oocyte development and infertility; they also have a decreased number of somatic cells surrounding the oocytes (Rajkovic et al., 2004). The primary oocyte also continues to express Figla, which drives the expression of the zona pellucida genes during primary follicle formation (Liang et al., 1997). The zona pellucida gene products ZP-2 and ZP-3 function at this time to form the zona pellucida oocyte coat that will be needed later, during oocyte maturation and fertilization. Also important in the developing crosstalk between the oocyte and its surrounding somatic follicle is the expression by granulosa cells of the forkhead transcription factor forkhead box L2, FOXL2, which is required for the squamous-to-cubiodal transition of the granulosa cells (Choi and Rajkovic, 2006). The primary follicle then develops into a secondary follicle through the formation of two or more layers of cuboidal granulosa cells surrounding the primary oocyte, and this is followed by the development of theca cells around the granulose cell layer (see Figure 8.3, B and C). This process is dependent on the oocyte-specific expression of growth differentiation factor 9 (Gdf9) and bone morphogenetic protein 15 (BMP-15). Both are members of the TGF-b superfamily of secreted proteins, and both are required for the primary-tosecondary transition. Without the oocyte-specific expression of these genes, the development of the somatic cells surrounding the oocytes fails, which will lead to the cell death of the oocyte (Carabatsos et al., 1998). Within

GERM CELL DEVELOPMENT IN HUMANS AND INFERTILITY

161

the secondary follicles, the primary oocytes reach their maximum sizes of ~75 mm and ~100 mm in diameter in mice and humans, respectively (Mehlmann et al., 2004; Mehlmann, 2005), and they are competent to resume meiosis; however, the primary oocyte is still arrested in meiotic prophase I. The G-protein coupled receptor Gpr3 maintains the meiotic arrest by promoting high levels of cAMP in the oocyte, and Gpr3-mutant mice undergo spontaneous oocyte maturation (Mehlmann et al., 2004; Mehlmann, 2005); however, the ligand that activates the receptor has not been identified. The next step in folliculogenesis is the formation of the antral follicle, which is the formation of a fluid-filled cavity or antrum (see Figure 8.3, D). Furthermore, at this stage, the granulosa layer is divided by the antrum into two separate compartments: the outer layer of cuboidal mural granulosa cells and the cumulus cells, which surround the oocyte (Mehlmann et al., 2004; Mehlmann, 2005). The stages of antral development are dependent on endocrine signals from the pituitary gonadotropin hormone follicle stimulating hormone and luteinizing hormone. Along with triggering the growth and differentiation of the somatic cells surrounding the oocyte, these signals also induce steroidogenic enzyme expression. Growth and steroid production help bring the oocyte to a maturation point during the late antral/preovulatory stage. However, although endocrine signaling has played an important part in this process, crosstalk between the oocytes and the somatic environment is still essential, and this is exemplified by the oocyte-specific transcription factor, Taf4b. Taf4b is a TATA box-binding protein (TBP)-associated factor that is required for the antral-to-preovulatory transition, and female mice lacking this gene are infertile (Falender et al., 2005). By the late antral/preovulatory stage, the oocyte responds to a surge in luteinizing hormone, which leads to ovulation through an interaction with luteinizing-hormone receptors on the mural granulosa cells. In frogs and fish, the signal is propagated by the stimulation of steroid hormones (Haccard and Jessus, 2006), but how the mural granulosa cells transmits this signal to either the cumulus cells or the oocyte in mammals is currently unknown. Other factors intrinsic to the follicle are also important during this period. One example in the mouse is the orphan nuclear receptor steroidogenic factor 1, which is expressed by the somatic cells of the follicle. Female mice that lack this gene in their ovary have antral follicles, but they do not ovulate (Pangas and Rajkovic, 2006). After ovulation proceeds, it promotes a resumption of meiosis, with the concurrent extrusion of the first polar body followed by arrest in meiosis metaphase II (see Figure 8.3, E). Now considered a secondary oocyte and surrounded by cumulus cells, the oocyte is able to undergo fertilization by sperm. Upon fertilization, the oocyte will complete meiosis II, extrude the second polar body, and be fully functional and able to support embryonic development.

VIII. GERM CELL DEVELOPMENT IN HUMANS AND INFERTILITY Infertility is common among both men and women. Although human reproduction and fertility have been studied for many years, few genes have been identified that contribute to human germ cell production. However, several studies have demonstrated that the age at onset of menopause has a significant genetic component; this property is likely to reflect the quantity and quality

162

GERM LINE DETERMINANTS AND OOGENESIS

of female germ cells that are formed and differentiated. Family history is a significant predictor of early menopause (menopause at age Pitx2 pathway mediating cell-type-specific proliferation during development, Cell 111:673–685, 2002. Lee RY, Luo J, Evans RM, et al: Compartment-selective sensitivity of cardiovascular morphogenesis to combinations of retinoic acid receptor gene mutations, Circ Res 80:757–764, 1997. Lincoln J, Alfieri CM, Yutzey KE: BMP and FGF regulatory pathways control cell lineage diversification of heart valve precursor cells, Dev Biol 292:292–302, 2006. Marvin MJ, Di Rocco G, Gardiner A, et al: Inhibition of Wnt activity induces heart formation from posterior mesoderm, Genes Dev 15:316–327, 2001. McElhinney DB, Geiger E, Blinder J, et al: NKX2.5 mutations in patients with congenital heart disease, J Am Coll Cardiol 42:1650–1655, 2003. Megarbane A, Salem N, Stephan E, et al: X-linked transposition of the great arteries and incomplete penetrance among males with a nonsense mutation in ZIC3, Eur J Hum Genet 8: 704–708, 2000. Mjaatvedt CH, Nakaoka T, Moreno-Rodriguez R, et al: The outflow tract of the heart is recruited from a novel heart-forming field, Dev Biol 238:97–109, 2001. Nguyen-Tran VT, Kubalak SW, Minamisawa S, et al: A novel genetic pathway for sudden cardiac death via defects in the transition between ventricular and conduction system cell lineages, Cell 102:671–682, 2000. Oda T, Elkahloun AG, Pike BL, et al: Mutations in the human Jagged1 gene are responsible for Alagille syndrome, Nat Genet 16:235–242, 1997. Patient RK, McGhee JD: The GATA family (vertebrates and invertebrates), Curr Opin Genet Dev 12:416–422, 2002.

720

HEART PATTERNING AND CONGENITAL DEFECTS

Patterson KD, Cleaver O, Gerber WV, et al: Homeobox genes in cardiovascular development, Curr Top Dev Biol 40:1–44, 1998. Prall OW, Elliott DA, Harvey RP: Developmental paradigms in heart disease: insights from tinman, Ann Med 34:148–156, 2002. Pritchard JK, Cox NJ: The allelic architecture of human disease genes: common disease-common variant. . .or not? Hum Mol Genet 11:2417–2423, 2002. Roberts KE, McElroy JJ, Wong WP, et al: BMPR2 mutations in pulmonary arterial hypertension with congenital heart disease, Eur Respir J 24:371–374, 2004. Robinson SW, Morris CD, Goldmuntz E, et al: Missense mutations in CRELD1 are associated with cardiac atrioventricular septal defects, Am J Hum Genet 72:1047–1052, 2003. Satoda M, Zhao F, Diaz GA, et al: Mutations in TFAP2B cause Char syndrome, a familial form of patent ductus arteriosus, Nat Genet 25:42–46, 2000. Schott JJ, Benson DW, Basson CT, et al: Congenital heart disease caused by mutations in the transcription factor NKX2–5, Science 281:108–111, 1998. Stennard FA, Harvey RP: T-box transcription factors and their roles in regulatory hierarchies in the developing heart, Development 132:4897–4910, 2005. Svensson EC, Huggins GS, Lin H, et al: A syndrome of tricuspid atresia in mice with a targeted mutation of the gene encoding Fog-2, Nat Genet 25:353–356, 2000. Tanaka Y, Okada Y, Hirokawa N: FGF-induced vesicular release of Sonic hedgehog and retinoic acid in leftward nodal flow is critical for left-right determination, Nature 435:172–177, 2005. Tartaglia M, Gelb BD: Noonan syndrome and related disorders: genetics and pathogenesis, Annu Rev Genomics Hum Genet 6:45–68, 2005. Tevosian SG, Deconinck AE, Tanaka M, et al: FOG-2, a cofactor for GATA transcription factors, is essential for heart morphogenesis and development of coronary vessels from epicardium, Cell 101:729–739, 2000. Vissers LE, van Ravenswaaij CM, Admiraal R, et al: Mutations in a new member of the chromodomain gene family cause CHARGE syndrome, Nat Genet 36:955–957, 2004. Vitelli F, Morishima M, Taddei I, et al: Tbx1 mutation causes multiple cardiovascular defects and disrupts neural crest and cranial nerve migratory pathways, Hum Mol Genet 11:915–922, 2002. Ward C: Clinical significance of the bicuspid aortic valve, Heart 83:81–85, 2000. Ware SM, Peng J, Zhu L, et al: Identification and functional analysis of ZIC3 mutations in heterotaxy and related congenital heart defects, Am J Hum Genet 74:93–105, 2004. Ya J, Schilham MW, de Boer PA, et al: Sox4-deficiency syndrome in mice is an animal model for common trunk, Circ Res 83:986–994, 1998. Yamagishi H, Yamagishi C, Nakagawa O, et al: The combinatorial activities of Nkx2.5 and dHAND are essential for cardiac ventricle formation, Dev Biol 239:190–203, 2001. Yutzey KE, Kirby ML: Wherefore heart thou? Embryonic origins of cardiogenic mesoderm, Dev Dyn 223:307–320, 2002.

RECOMMENDED RESOURCES Online Mendelian Inheritance in Man: http://www.ncbi.nlm.nih.gov/omim Human Gene Mutation Database: http://www.hgmd.cf.ac.uk/ac/index.php Development of the Heart: http://embryology.med.unsw.edu.au/Movies/heart.htm Harvey RP, Rosenthal N, editors: Heart development, Burlington, MA, 1998, Academic Press, Inc.

33

BLOOD VESSEL FORMATION KARINA YANIV and BRANT M. WEINSTEIN Laboratory of Molecular Genetics, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD

INTRODUCTION The development of the vascular system is one of the earliest events in organogenesis. All other organs depend on a vascular supply for the delivery of nutrients, oxygen, and cellular and humoral factors and for the clearance of wastes. Serious disruptions in the formation of the vascular network are lethal early during postimplantation development in amniotes, whereas the maintenance of vessel integrity and the control of vessel physiology have important consequences throughout embryonic and adult life. Many of the processes that take place during normal vascular development in the embryo are reactivated in situations of neoangiogenesis in the adult, including tissue regeneration, wound healing, and tumor formation. A full understanding of the signaling pathways of vascular development is essential when searching for new targets for therapeutic intervention during pathologic situations.

I. EMERGENCE OF THE BLOOD VASCULAR SYSTEM A. Basic Concepts The cardiovascular system consists of the heart, the blood vascular system, and the lymphatic vascular system, which is discussed separately later in this chapter. During vertebrate embryogenesis, it is the first functional organ system to develop, because embryonic growth and differentiation are critically dependent on the transport of oxygen, nutrients, and waste products throughout the early vasculature. The heart and blood vessels form a closed circulatory loop, with blood never leaving the vessels, except through leakage or hemorrhage. The vascular system is composed of two fundamental types of blood vessels: arteries and veins. Arteries carry blood away from the heart to tissues, whereas veins return blood back to the heart (except for pulmonary veins). The circulatory systems of fish, amphibians, reptiles, birds, and mammals show various stages of evolution. In fish, the system has only one circuit, with the blood being pumped

Principles of Developmental Genetics

721

722

BLOOD VESSEL FORMATION

through the capillaries of the gills, where it is oxygenated and then sent on to the capillaries of the body tissues before being returned to the heart; this is known as single circulation. The heart of the fish is therefore only a single pump that consists of two main chambers: the atrium and the ventricle. In air-breathing vertebrates, an additional circulatory loop is incorporated to accommodate the pulmonary circulation; blood flows from the heart to the lungs for oxygenation and then back again to the heart before being sent out again for distribution to other tissues. In amphibians and reptiles, this double circulatory system is used, but the heart is not always completely divided into two separate pumps (e.g., amphibians have a three-chambered heart). Birds and mammals show complete separation of the heart into two pumps for a total of four heart chambers (two atria and two ventricles). Oxygen-saturated blood leaves the heart through the aortic arch, branching out and dispersing into ever-smaller–caliber arteries and arterioles. Eventually, the blood supply comes into contact with all of the living cells of the body in the capillaries, which are microscopic vessels through which life-sustaining substances and wastes move readily in and out. From the capillaries, blood moves into small venules and then progressively larger veins, which merge together into the vena cava before returning to the heart. Blood vessels all have a similar basic histologic structure (Figure 33.1). They are composed of two basic cell types: vascular endothelial cells (ECs)

Blood vessels come in two fundamental types: arteries and veins. Both types of vessels are composed of an inner endothelium (tunica intima) surrounded by internal elastic tissue, a smooth muscle cell layer (tunica media), external elastic tissue, and fibrous connective tissue (tunica adventitia). Larger-caliber arteries have a thicker smooth muscle cell layer, whereas larger veins possess specialized structures such as valves. The two networks of tubes are completely separate at the level of the larger vessels but are connected distally through a system of fine capillaries. (Reproduced from Cleaver and Krieg, 1999, with permission. See color insert.)

FIGURE 33.1

EMERGENCE OF THE BLOOD VASCULAR SYSTEM

723

and vascular smooth muscle cells (VSMCs). The inner epithelial lining of the blood vessels adjacent to the lumen is a thin, single-layered epithelium of ECs, whereas smooth muscle cells or pericytes surround the EC layer. In larger blood vessels, the inner endothelial lining is called the intimal layer, and it is surrounded by a medial layer that is composed of multiple layers of VSMCs embedded in an elastin-rich extracellular matrix. The medial layer is itself surrounded by an extracellular matrix-rich layer called the adventitial layer. The inner lining of the vessel is the endothelium, a single-cell–thick layer of vascular ECs that is surrounded by subendothelial connective tissue. This is surrounded by a muscular layer of VSMCs, which is highly developed in arteries. Finally, there is a further layer of connective tissue called the adventitia, which contains nerves that supply the muscular layer as well as nutrient capillaries in the larger blood vessel. The earliest primitive vessels in vertebrate embryos form by a process called vasculogenesis, during which mesodermal cells differentiate into endothelial precursor cells called angioblasts. These angioblasts then differentiate in situ into ECs and coalesce to form the earliest vessels, which in mammalian and avian embryos often appear as a relatively unstructured vascular plexus. The subsequent remodeling of vasculogenic vessels and their further growth and remodeling to form the complex and elaborate network of vessels found in the mature vasculature is called angiogenesis. In nonsprouting angiogenesis or intussusception, preexisting vessels subdivide in two by the formation of transvascular posts or pillars, whereas, during sprouting angiogenesis, new vessels grow by sprouting and growth from preexisting vessels (see Chapter 21). After the primitive endothelial tubes are formed, the endothelium secretes factors that lead to the recruitment and/or induction of primordial smooth muscle through a process called vascular myogenesis. Several recent reviews have carefully documented the current state of knowledge regarding the differentiation and growth of VSMCs to form the tunica media (Carmeliet, 2000; Hungerford and Little, 1999). The origins of VSMCs remain unclear. Different models have been proposed for the induction and differentiation of these cells (Bergwerff et al., 1998; DeRuiter et al., 1997; Rosenquist and Beall, 1990; Vrancken Peeters et al., 1999); however, it is important to note that the complex origin of VSMCs seems to be dependent on their location. This suggests that individual growth factors and their receptors will have different effects on VSMC growth and differentiation in specific vascular beds. B. Emergence and Specification of Endothelial Cells The initial phase of vascular development involves the differentiation of endothelial precursor cells, called angioblasts, from mesoderm. Angioblasts are endothelial precursors that have certain characteristics of ECs but that have not yet assembled into functional vessels (Flamme et al., 1997). Quail/chick transplantation experiments have shown that two subsets of mesoderm— somitic and splanchnopleuric—have the potential to give rise to endothelial progenitors in avians (Coffin and Poole, 1988; Pardanaud and Dieterlen– Lievre, 1999). In zebrafish embryos, angioblasts detected by the expression of vascular endothelial growth factor receptor (VEGFR-2/flk1/kdr) arise and segregate from the lateral plate mesoderm at the 7-somite stage (Fouquet et al., 1997; Liao et al., 1997; Thompson et al., 1998). Cell-lineage analysis

724

BLOOD VESSEL FORMATION

suggests that these angioblasts contribute to the primordia of the dorsal aorta, the cardinal veins, and even the intersegmental vessels in the zebrafish trunk (Childs et al., 2002; Zhong et al., 2001). In the developing mouse embryo, the formation of blood islands in the extraembryonic yolk sac marks the onset of vasculogenesis (Risau and Flamme, 1995). Blood islands develop from aggregates of mesodermal cells at approximately embryonic day 7.5 to 8 of mouse development. They consist of an inner layer of primitive hematopoietic cells and a peripheral population of angioblasts. These angioblasts differentiate into ECs, form a lumen, migrate, and interconnect to form a primary vascular plexus (Risau and Flamme, 1995). The close spatial relationship as well as the simultaneous emergence of the hematopoietic cells and ECs within the blood islands has led to the hypothesis that they arise from a common precursor, the hemangioblast (His, 1900; Murray, 1932; Williams et al., 1980; see Chapter 34). Several lines of evidence support the existence of such progenitors with dual potentials. For example, angioblasts and hematopoietic progenitors express many of the same transcription factors and surface receptors, such as CD34 (Fina et al., 1990), Flk, Flt1, Tie1, Tie2 (Dumont et al., 1995), and SCL/tal–1 (Kallianpur et al., 1994). Moreover, the development of both endothelial and hematopoietic compartments is impaired in embryos bearing a mutation or a dominant-negative form of one of the relevant genes (Shalaby et al., 1995; Shivdasani et al., 1995; Visvader et al., 1998). Similarly, the zebrafish cloche mutation results in defects in blood cells and blood vessels (Liao et al., 1997; Stainier et al., 1995). Probably the most compelling evidence for the hemangioblast, however, comes from studies involving the in vitro differentiation system of mouse embryonic stem cells, in which the development of the endothelial and hematopoietic lineages in embryoid bodies recapitulates events that take place in vivo in the yolk sac blood islands (Doetschman et al., 1985; Vittet et al., 1996; Wang et al., 1992; Wiles and Keller, 1991). Using this model system, Choi et al. (1998) and Chung et al. (2002) isolated a transient population of cells that expresses markers that are common to both cell lineages (SCL/tal– 1, CD34, and Flk–1), and they determined that single cells could give rise to clones containing both hematopoietic cells and ECs. More recently, an Flk1þ population of cells was identified in the posterior primitive streak of embryonic day 7 to 7.5 mouse embryos, thus providing further support for the existence of a common progenitor for hematopoietic cells and ECs in vivo (Huber et al., 2004). Although many studies looking at morphology, gene expression, and mutant phenotypes support the concept of a common hematopoietic cell and EC progenitor in the yolk sac in vivo, other studies have argued against its existence. In the avian embryo, a clonal differentiation assay of VEGFR-2–positive cells sorted out from the very early blastodisc failed to give rise to mixed colonies of endothelial and hematopoietic cells (Eichmann et al., 1997). In addition, the lineage tracing of cells from the primitive streak to the yolk sac has failed to identify cells with endothelial and hematopoietic potential (Kinder et al., 2001). Finally, Ferkowicz et al. (2003) showed that hematopoietic and endothelial progenitors can be distinguished by their differential expression of CD41 as soon as they exit from the primitive streak. The differences between these apparently contradictory facts still need to be resolved. Some of the discrepancies may result from differences in the

EMERGENCE OF THE BLOOD VASCULAR SYSTEM

725

timing of commitment to the different lineages in various organisms. Another explanation may be that there are distinct populations of precursor cells, some of which have multilineage “hemangioblastic” potential and others that contribute only to single lineages. EC differentiation in the embryo proper does not occur in close association with hematopoiesis, except for on the floor of the aorta (reviewed by Jaffredo et al., 2005a), where the endothelial and hematopoietic clusters are present in close proximity. Recent data have suggested that the endothelium of this region might be hemogenic (i.e., capable of giving rise to definitive hematopoietic cells through an endothelial intermediate). However, other studies suggest that hematopoietic precursor cells originate from the surrounding mesenchyme and migrate through the aorta wall before subsequently entering the circulation (reviewed by Jaffredo et al., 2005a; 2005b). Further work allowing for the direct tracing of endothelial progenitors as well as additional molecular determinants of the hemangioblast will presumably shed additional light on the molecular pathways leading to hematopoietic cell and EC lineage differentiation. C. De Novo Formation of a Primary Vascular Plexus: Vasculogenesis After endothelial progenitors have been specified, they interconnect to form a dispersed capillary plexus that supports blood cell circulation and that matures into a vascular network by extensive pruning and remodeling. This process, which involves the differentiation of ECs from the mesoderm and is followed by their coalescence into tubes, is called vasculogenesis (Risau, 1997). Intraembryonic vasculogenesis is initiated in the cranial region of embryonic day 7.5 mouse embryos with the emergence of endocardial progenitor cells. Concomitantly, the aortic primordia first become discernible (Drake and Fleming, 2000). In the chick embryo, the dorsal aorta as well as several capillaries has differentiated by the time that a heartbeat begins at the 12somite stage (reviewed by Eichmann et al., 2005). As is seen in other vertebrates, the first angioblasts in the zebrafish arise from lateral plate mesoderm. The angioblasts migrate to the trunk midline between the 10-and 15-somite stages and coalesce to form the primary axial vessels of the trunk (reviewed by Weinstein, 2002). As shown in a very detailed atlas of the blood vessels of the developing zebrafish (Isogai et al., 2001), the same primary vasculogenic vessels that establish the initial circulatory circuits within mammalian and avian embryos are present in the zebrafish, including the dorsal aorta and the posterior cardinal vein in the trunk and the internal carotid artery, the primordial hindbrain channel, the anterior cardinal vein, and the basilar artery in the head (Figure 33.2). D. Remodeling and Maturation of the Vascular Plexus: Angiogenesis After its initial establishment, the vasculogenic primary vascular plexus becomes extensively remodeled and elaborated on by subsequent angiogenesis. This angiogenic remodeling and growth are essential for tissue and organ growth and repair, and an imbalance in this process contributes to numerous malignant, inflammatory, ischemic, infectious, and immune disorders. Angiogenesis takes two main forms: sprouting and nonsprouting (reviewed by Scappaticci, 2002). Sprouting angiogenesis refers to the development of new blood vessels by budding and growth from preexisting vessels. It involves the proteolytic degradation of the extracellular matrix adjacent to an existing vessel,

726

BLOOD VESSEL FORMATION

FIGURE 33.2 Major primitive vessels form by vasculogenesis during embryonic development. A, Early avian yolk sac vascular plexus, with larger arterial and venous vessels already apparent. B, In situ hybridization for the fli1 gene (arrows) marks emerging vascular and hematopoietic progenitors in the early lateral mesoderm of the 10-somite–stage zebrafish (composite image, dorsal view, anterior up). C, Early vasculogenic vessels in confocal microangiography of a 24-hour postfertilization zebrafish embryo. In fish and amphibians, most vasculogenic vessels form as single vascular tubes without an intermediate stage of vascular plexus formation. (Panel A modified from Popoff, 1894).

migration and proliferation of ECs from the wall of the vessel, followed by lumen formation and the maturation of functional capillaries from the mobilized ECs. Nonsprouting angiogenesis occurs by intussusception, during which the subdivision of previously existing vessels takes place by the formation of transvascular pillars that split the vessel into several new capillaries. Intussusceptive and sprouting angiogenesis are both employed to remodel, elaborate, and expand on an initial vascular plexus. This occurs concomitantly with the acquisition of VSMCs or pericyte cells, which stabilize the nascent vasculature and which are essential for the maturation phase of vessel development. During organ and tissue growth, blood vessels must continuously grow and adapt to meet the needs for nutrients and oxygen. The organs and tissues signal to the vessels to promote their growth and, if necessary, their regression. They also provide cues that cause ECs to adopt functional specialties and the specific features that particular organs need to interact properly with the circulatory system (Nikolova and Lammert, 2003). It has become clear in recent years that ECs are functionally and molecularly heterogeneous. In addition to arterial–venous distinctions, ECs from different organs and tissues frequently express different genes. Conversely, increasing evidence suggests

EMERGENCE OF THE BLOOD VASCULAR SYSTEM

727

that ECs, in turn, provide instructive morphogenetic cues to the surrounding organs to help determine their location, differentiation, and morphology during development and in the adult. Blood vessels and organ-specific cells interact with each other continuously throughout development and postnatal life, and the coordination between ECs and the cells within the organs and tissues that they serve generates a functional organ with a vascular system that is adjusted to its needs. E. Molecular Regulation of Vascular Development Vascular development and pathologic blood vessel formation are promoted by signals that are received and processed by ECs or their precursors. In recent years, a number of different signaling pathways involved in the development of the vascular system have been described (reviewed by Coultas et al., 2005). The best characterized involve receptor tyrosine kinases, although other classes of signaling inputs are also important (reviewed by Folkman and D’Amore, 1996; Ilan et al., 1998; Yancopoulos et al., 1998). In the following sections, we review two pathways: the vascular endothelial growth factor (VEGF) and the angiopoietin/Tie receptor pathways, which are almost totally endothelial specific and which are critical players in vascular development. Other signaling factors and pathways that are important for the guidance and patterning of the developing vasculature are discussed in Section IV. 1. Vascular Endothelial Growth Factor Signaling VEGF has been shown to be important for the migration, proliferation, maintenance, and survival of ECs, and it is critical during both vasculogenesis and angiogenesis (reviewed by Carmeliet and Collen, 2000; Yancopoulos et al., 2000). VEGF, which is now commonly referred to as VEGF-A, was the first growth factor described to be a mitogen specifically for ECs (Carmeliet and Conway, 2001; Ferrara, 1999; Ferrara et al., 2003). It was initially defined, characterized, and purified for its ability to induce vascular permeability and to promote EC proliferation. The VEGF family includes five characterized VEGF relatives in mammals (VEGF-A through VEGF-D and placental growth factor) that display differential interaction with three related receptor tyrosine kinases (VEGFR-1/Flt-1, VEGFR-2/Flk-1, and VEGFR3/Flt-4) and a number of ancillary receptor components, such as the neuropilins (NPs; reviewed by Goishi and Klagsbrun, 2004). VEGF-A signals through binding to VEGFR-1 and VEGFR-2, which are restricted largely to vascular endothelium in their expression, and this accounts for the specificity of VEGF-A signaling. By contrast, VEGFR-3 is restricted largely to lymphatic endothelium (Kukk et al., 1996). VEGF-A is produced by different cell types, including tumor cells, macrophages, T cells, and smooth muscle cells (Klagsbrun and D’Amore, 1996). It is thought to play a major role in tumor-induced neovascularization, and, recently, a humanized monoclonal antibody directed against VEGF-A has shown efficacy for the clinical treatment of colorectal and renal tumors (Willett et al., 2004). During embryonic development, the expression of both VEGF-A and its receptor VEGFR-2 correlate closely with sites of vessel formation (Jakeman et al., 1993; Liang et al., 1998; Shweiki et al., 1993). The most conclusive evidence for the critical role of VEGF-A as a key regulator of both vasculogenesis and angiogenesis has come from the

728

BLOOD VESSEL FORMATION

knockout mice (Carmeliet et al., 1996; Ferrara, 1996; Shalaby et al., 1995). In embryos lacking either VEGF-A or VEGFR-2, blood islands, ECs, and major vessels fail to develop in appreciable numbers, resulting in embryonic lethality between embryonic days 8.5 and 9.5. Remarkably, the disruption of even one of the two VEGF-A alleles results in embryonic lethality between embryonic days 11 and 12, demonstrating a strict dose–dependent requirement for VEGF-A during embryogenesis and making this one of very few genes showing haploinsufficiency during murine development. The inactivation of VEGFR-1 also leads to embryonic lethality (Fong, 1995; Fong et al., 1995; Vajkoczy et al., 1999). Although ECs are found at embryonic and extraembryonic sites, the resulting vessels are abnormally organized, apparently as a result of an overproliferation of ECs (Fong et al., 1995; Kearney et al., 2002), which suggests that this receptor negatively regulates or restrains angiogenesis. A third VEGFR, VEGFR3/Flt-4, is essential for lymphatic development, and it is activated by binding to VEGF-C. VEGFR3/Flt-4–deficient embryos also show defects in the formation of the circulatory system (Dumont et al., 1998). VEGF-A plays an important role also during early postnatal life (Gerber et al., 1999). The partial inhibition of VEGF-A achieved by Cre-loxP– mediated gene targeting results in increased mortality, stunted body growth, and impaired organ development. More recently, a critical role for this factor has also been demonstrated during adult neovascularization (Grunewald et al., 2006). Together, these data account for the major position of the VEGF signaling system in vascular formation. 2. Angiopoietin/Tie Signaling After the discovery of VEGF-A, a second family of growth factors important for EC survival and vascular remodeling was identified, with members of this family being called the angiopoietins (Davis et al., 1996; reviewed by Gale and Yancopoulos, 1999). The angiopoietins have been shown to have important functions during angiogenesis. Like the VEGFRs, the specificity of angiopoietins for vascular endothelium results from the restricted distribution of their tyrosine kinase receptors Tie1 and Tie2/Tek on ECs (Dumont et al., 1995; Sato et al., 1995). Angiopoietin 1 (Ang1) seems to be important for the stabilization of vessel walls by promoting interactions between vascular ECs and surrounding pericytes and smooth muscle cells. Consistent with a constitutive stabilizing role, Ang1 is widely expressed in adult normal tissues (Suri et al., 1996). In murine embryos that are deficient in Ang1, early stages of VEGF-dependent vascular development appear to occur rather normally. However, the remodeling and stabilization of the primitive vascular plexus is severely perturbed, and this leads to embryonic lethality. A similar phenotype (although it was more evident in the brain capillary plexus) was reported for murine embryos lacking Tie2 receptor (Sato et al., 1995). Alternatively, the transgenic overexpression of Ang1 leads to striking hypervascularization by promoting vascular maturation and inhibiting normal vascular pruning (Suri et al., 1998). Together, these results suggest a critical role for the Ang1/Tie2 system in the normal remodeling, maturation, and stabilization of the developing vasculature. In sharp contrast with Ang1, angiopoietin 2 (Ang2) also binds to the Tie2 receptor, but it is unable to activate the Tie2 receptor, and thus it acts as a

ARTERIAL–VENOUS DIFFERENTIATION

729

natural antagonist for the Ang1/Tie2 interaction. The transgenic overexpression of Ang2 results in a lethal phenotype that is reminiscent of that seen in Ang1 or Tie2 knockout mice (Maisonpierre et al., 1997). Ang2 is highly expressed at sites of vascular remodeling, and it is hypothesized to destabilize mature vessels, thus rendering them more amenable to remodeling, regression, or additional angiogenic growth, depending on other signals that the vessels are receiving (most notably VEGF; Yancopoulos et al., 2000). Ang2 knockout mice have a complex phenotype (Gale et al., 2002) that includes some vascular defects, which supports a role for Ang2 in postnatal angiogenesis and/or vascular remodeling but which prominently includes malformations of the lymphatic vasculature. Large and small lymphatic vessels are generally able to be formed, but they have defects in their overall organization. The mice appear to be normal at birth, but soon after the start of feeding they develop severe defects as a result of lymphatic dysfunction, and they die around postnatal day 14. To help further clarify the role of Ang2 in blood and lymphatic vessel development, Gale et al. (2002) generated mice in which the Ang2 gene was replaced by Ang1. These mice showed an almost complete rescue of the lymphatic defects, but the postnatal remodeling defects of the retinal blood vasculature persisted. One interpretation of this phenotype is that Ang2 normally acts as a Tie2 agonist in the lymphatic vasculature (because Ang1 can rescue the defect), but it acts as a Tie2 antagonist in the retinal vasculature (because Ang1 cannot rescue the defect). Like Tie2, the closely related Tie1 receptor is primarily expressed in vascular ECs (Sato et al., 1993). However, until recently, ligands for this receptor had not been identified, and it remained an orphan receptor. Vascular development proceeds normally in Tie1-deficient mice up to approximately embryonic day13.0, but shortly thereafter they begin to show signs of edema, local hemorrhage, and rupturing of microvessels, and they die between embryonic days 13.5 and 18.5 (Puri et al., 1995; Sato et al., 1995). Although previous studies failed to demonstrate the binding of angiopoietins to Tie1, it has recently been demonstrated that Ang1 and Ang4 may function as activating ligands for this receptor (Saharinen et al., 2005) and that its activation is promoted by the formation of heterodimeric complexes with Tie2 (Marron et al., 2000; Saharinen et al., 2005; Tsiamis et al., 2002). The mechanisms underlying the effects of this binding are not yet known, and it still remains unclear whether Tie1 can function as an independent receptor. The generation of mice with conditional null alleles or the development of more effective and specific inhibitors may reveal the precise roles of Tie signaling and of the different ligands and receptors during later development and in mature vessels.

II. ARTERIAL–VENOUS DIFFERENTIATION The most fundamental dichotomy in the blood vascular system and one of the very first steps in the differentiation of endothelium is the specification of arterial and venous identity (reviewed by Eichmann et al., 2005; Lawson and Weinstein, 2002; Rossant and Hirashima, 2003; Torres-Vazquez et al., 2003). Although it has long been assumed that the specification of arterial versus venous endothelial fate was a late event in development defined primarily by anatomic sites and hemodynamic forces, evidence has recently begun to emerge that suggests that the identity of ECs is genetically determined before

730

BLOOD VESSEL FORMATION

the onset of circulation and before vessel assembly. Beginning with studies in the mouse showing that ephrinB2 and its receptor EphB4 are specifically expressed in arterial and venous endothelium, respectively (Wang et al., 1998), a large number of studies have now highlighted the signaling pathways involved in the differentiation of arteries and veins. A variety of evidence indicates that arterial–venous identity is acquired early during vasculogenesis, well before the onset of blood flow. In murine (Wang et al., 1998), zebrafish (Lawson et al., 2001), and avian (Herzog et al., 2005) embryos, the differential expression of ephrinB2 and EphB4 (Lawson et al., 2001; Wang et al., 1998), NP-1 and NP-2 (Herzog et al., 2001), and other markers in arteries and veins precedes the initiation of blood flow (Figure 33.3). The fluorescent labeling of single angioblasts during early somitogenesis stages in the zebrafish showed that angioblasts give rise to either arterial or venous vascular ECs but not to mixed clones, thus further supporting the idea of early specification of arterial–venous identity (Zhong et al., 2001). Other evidence suggests, however, that there is considerably plasticity in the early vasculature with respect to arterial–venous identity. Grafting experiments in avian embryos indicate that EC fate remains plastic up to embryonic day 7, which is well after circulation is initiated. Ectopic grafts from embryonic quail arteries or veins can switch their cell fate after transplantation into host chick embryos (Moyon et al., 2001; OthmanHassan et al., 2001), and time-lapse imaging experiments have shown that changes in circulatory flow patterns can regulate arterial–venous differentiation in the avian embryo yolk sac (le Noble et al., 2004). However, this plasticity is gradually lost later during development, perhaps as a result of association with VSMCs or other nonendothelial components of the vascular wall (Moyon et al., 2001). It seems likely that the overall vascular structure

FIGURE 33.3

A reduction in Notch signaling in zebrafish embryos perturbs arterial–venous identity. In situ hybridization of the trunk dorsal aorta (red arrows) and cardinal vein (blue arrows) in 25-somite–stage Notch-deficient mindbomb (mibta52b) mutant and wild-type sibling zebrafish embryos. In wild-type animals, ephrinB2a (efnb2a) expression is apparent in the dorsal aorta but not the cardinal vein (upper left). However, in notch-deficient mibta52b mutant embryos, efnb2a expression is absent (lower left). By contrast, flt4 expression is restricted to the cardinal vein in wild-type animals by the 25-somite stage (upper right), whereas, in mibta52b mutant embryos, flt4 expression persists within both the cardinal vein and the dorsal aorta (lower right). All panels show lateral views of the mid trunk, dorsal up, anterior to the left. (Figure modified from Lawson et al., 2001. See color insert.)

ARTERIAL–VENOUS DIFFERENTIATION

731

during embryogenesis is defined by both the hemodynamics of circulatory flow and by the initial genetically programmed, intrinsic determination of arterial versus venous fate. Recent work has shown a molecular cascade of events involved in the establishment of arterial–venous identity. We review some of these molecular players and their roles below. A. Eph–Ephrin Signaling The Eph-receptor tyrosine kinases constitute the largest known family of growth factor receptors, and they are activated by the equally numerous membrane-bound ephrins as their ligands (Adams et al., 1999; see Chapter 22). Although initially characterized in the nervous system, key roles for ephrinB and the EphB receptor in vascular development have been suggested in recent studies. In an initial knockout study in which the ephrinB2 locus was targeted with a tau–lacZ gene (Wang et al., 1998), the presence of this reporter was exclusively detected in the arteries even before the establishment of circulation, whereas in situ hybridization experiments on the same embryos showed the specific expression of its receptor, EphB4, in the venous endothelium. As noted previously, this was the first evidence for the genetic predetermination of arterial and venous fate. When the EphB4 locus was inactivated, vascular defects similar to those in mice lacking ephrinB2 were observed (Gerety et al., 1999). Animals carrying homozygous knock-in mutations of either the ligand or receptor display the proper differential expression of knockedin transgenes and the normal initial formation of the major intraembryonic arterial and venous trunk vessels, which implies that this signaling pathway is not required for the first steps of arterial–venous specification. In both cases, however, later defects in the remodeling of the primary vascular plexus as well as in the maintenance of artery–vein separation were observed that led to death around embryonic day 9.5 (Adams et al., 1999; Gerety et al., 1999; Wang et al., 1998). Although ephrinB2 is also expressed in VSMCs, an endothelial-specific knockout of this gene displayed a very similar phenotype to the conventional ephrinB2 null mutant mice, thus demonstrating that the gene is critically required in ECs, at least for its earliest vascular functions (Gerety and Anderson, 2002). Interestingly, the complementary expression of ephrinB2 and EphB4 in arteries and veins is also present in adults, which suggests an important role for the reciprocal expression of these genes not only during development but also for the continued maintenance of proper arterial–venous differentiation in adults (Gale et al., 2001) Molecularly, the ephrinB–EphB system can function bidirectionally. As for most other receptor tyrosine kinases, ligand binding induces “forward” signaling in EphB4, mainly through phosphotyrosine-mediated pathways. However, ephrins can also signal into their host cell (referred to as reverse signaling) via their cytoplasmic tail (reviewed by Kullander and Klein, 2002). The important role of ephrinB2 reverse signaling in angiogenesis was confirmed by recent studies of mice carrying a deletion of the cytoplasmic tail of ephrinB2 (Adams et al., 2001). In these mutant embryos, although the migration of neural crest cells (induced by Eph forward signaling) is normal, the remodeling of the vasculature is severely affected, and this suggests a critical role for ephrinB2 reverse signaling in this process. However, other researchers found that mice that were homozygous for novel knock-in alleles

732

BLOOD VESSEL FORMATION

of the ephrinB2 cytoplasmic tail targeting either the PDZ interaction site (ephrinB2 V/V) or the conserved tyrosine residues (ephrinB25F/5F) survived the initial requirement of ephrinB2 in embryonic vascular remodeling (Makinen et al., 2005). However, the ephrinB2 V/V mice exhibited major lymphatic defects, which suggests that there is an additional critical role for ephrinB2 reverse signaling via the PDZ interaction site in the postnatal remodeling of the lymphatic system. In sum, although the expression of ephrinB2 in arteries and EphB4 in veins is required for normal vascular development, the fact that mice that lack ephrinB2 continue to express a LacZ transgene inserted at the ephrinB2 locus indicates that these genes are not required for the initial fate decision that distinguishes arterial and venous endothelial progenitors. This suggests that upstream factors may regulate the proper expression of arterial- and venousspecific genes as well as the determination of arterial–venous fate. Growing evidence suggests that the Notch signaling pathway plays a key role in this process. B. Notch Signaling The Notch signaling pathway is an evolutionarily conserved intercellular signaling mechanism that controls cell fate specification in a variety of tissues in nearly all animal species that have been investigated so far (Artavanis–Tsakonas et al., 1999; Chitnis, 1995). Several lines of evidence indicate an important role for Notch signaling during vascular development in vertebrates. 1. Members of the Notch Signaling Pathway Are Expressed in the Vasculature Notch receptors and ligands are both present within the developing vasculature. Four different Notch receptors (Notch1 through Notch4) and five ligands (delta-like [Dll]1, Dll3, Dll4, Jagged-1 [Jag1], and Jag2), have been identified in mammals (Nye and Kopan, 1995). Notch1 (Del Amo et al., 1992; Reaume et al., 1992; Taichman et al., 2002), Notch2 (Del Amo et al., 1992; Zimrin et al., 1996), Notch3 (Villa et al., 2001), Notch4 (Krebs et al., 2000; Shirayoshi et al., 1997; Uyttendaele et al., 1996; Villa et al., 2001), and the Notch ligands Dll4, Jag1, and Jag2 are all expressed in vascular ECs during early embryogenesis in mice (Krebs et al., 2000; Shirayoshi et al., 1997; Villa et al., 2001). Members of the Notch family of receptors and ligands are also expressed in the vasculature in addition to various other tissues in other vertebrates such as chicken and zebrafish (Lawson et al., 2001; Vargesson et al., 1998; Zhong et al., 2000). Interestingly, the vascular expression of Notch receptors and ligands in the vasculature of different species has nearly always been reported to be restricted to arterial but not venous ECs (Lawson et al., 2001; Shutter et al., 2000; Villa et al., 2001), which suggests a role for this pathway during arterial differentiation. 2. Vascular Defects in Animals Deficient for Notch Receptors and Ligands Genetic analysis in mice and humans has revealed various types of vascular defects associated with Notch pathway mutants. The dominant genetic disorder cerebral autosomal–dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) is caused by mutations in the human Notch3 gene (reviewed by Kalimo et al., 2002). In patients with CADASIL,

ARTERIAL–VENOUS DIFFERENTIATION

733

vascular lesions occur throughout the arterial tree, including in the arteries and arterioles within muscle, skin, and peripheral nerves (Fryxell et al., 2001; Ruchoux et al., 1994; Schroder et al., 1995). Transgenic mice overexpressing murine Notch3 also exhibit vascular CADASIL–like phenotypes (Ruchoux et al., 2003). Alagille syndrome (AGS) is a genetic disorder caused by a mutation in the Notch receptor Jag1 (Xue et al., 1999). AGS is a major form of chronic liver disease in childhood, with severe morbidity and a mortality rate of 10% to 20% of affected individuals. Cardiac defects are also seen in more than 95% of patients with AGS (Gridley, 2003), and noncardiac vascular defects such as stenosis, aneurysm, and hemorrhage are also frequent in these patients, accounting for a third or more of their mortality (Emerick et al., 2005; Kamath et al., 2004). Further evidence supporting the important role of the Notch signaling pathway in the formation and/or maintenance of the vasculature has come from in vivo studies carried out in mice, rats, and zebrafish. Notch1 and Notch1/Notch4 mutant mouse embryos display severe defects in angiogenic vascular remodeling, which lead to death from vascular defects and hemorrhaging at around embryonic day 10.5 (Krebs et al., 2000; Swiatek et al., 1994; reviewed by Rossant and Howard, 2002). By contrast, the expression of an activated form of Notch4 specifically in the embryonic endothelium leads to disorganized vascular development and a reduction of the number of small vessels, thus resulting in embryonic lethality at embryonic day 10.5 (Uyttendaele et al., 2001). These results indicate that either a loss or an excess activation of Notch receptors in the vasculature causes defects in blood vessel morphogenesis. In similar fashion, mice lacking Notch ligands Jag1 and Dll1 die early during gestation as a result of severe defects in vascular remodeling (Barrantes et al., 1999; Xue et al., 1999). In addition, mutations in Dll4 (a Notch ligand expressed specifically in developing arterial ECs) lead to the defective development of the dorsal aorta and cardinal veins, the formation of arterial–venous shunts, and the downregulation of arterial markers and the upregulation of venous markers in the dorsal aorta (Duarte et al., 2004; Gale et al., 2004). 3. The Notch Signaling Pathway is Required for Arterial–Venous Cell Fate Determination The artery-restricted pattern of expression of all identified vascular Notch signaling components suggested a specific role in arterial–venous cell fate determination, but it was not until recently that this role was confirmed by functional studies in zebrafish and mice. Zebrafish embryos deficient in Notch signaling as a result of a mutation in the mindbomb (mib) gene (Jiang et al., 1996) or of microinjection with a dominant–negative form of the transcription repressor Suppressor of hairless (Su(H); the common downstream effector of Notch signaling; Wettstein et al., 1997) display a loss of arterial markers such as ephrinB2a and Notch5 (Lawson et al., 2001; Zhong et al., 2001) that is accompanied by the ectopic expression of normally vein–restricted markers in the arteries (see Figure 33.3). Conversely, the activation of the Notch pathway by the either ubiquitous or endothelial-specific expression of Notch–intracellular domain induces the ectopic expression of artery markers in veins (Lawson et al., 2002). These data support the idea that arterial fate is specified and maintained by the Notch pathway via repression of the venous fate.

734

BLOOD VESSEL FORMATION

The gridlock (grl) gene (Zhong et al., 2001) has been reported to be a target of the Notch pathway in the zebrafish vasculature. The grl gene encodes a basic helix–loop–helix protein that belongs to the Hairy and Enhancer of split family of transcriptional repressors (Nakagawa et al., 1999). The vascular expression of grl is restricted to the aorta, and zebrafish embryos with mutations in the grl gene fail to establish trunk circulation as a result of the incomplete formation of the aorta (Stainier et al., 1995; Zhong et al., 2000). A number of studies have reported different results regarding the potential role of grl in arterial differentiation downstream of Notch. In one report, the injection of grl mRNA into wild-type zebrafish embryos repressed the expression of the venous markers flt4 and EphB4 (as did the injection of activated Notch5) and enhanced the expression of the arterial gene ephrinB2 (Zhong et al., 2001). However, in another study, grl was shown to be normally expressed in the dorsal aorta of embryos that lacked Notch activity, despite the ectopic expression of other artery–vein molecular markers and clear effects on vascular morphology (Lawson et al., 2001). These observations indicate that grl might not be the functional repressor of flt4 in vivo, and suggest the existence of other Hairy-related transcription factors that might mediate the repression of flt4 downstream of the Notch pathway. The mammalian ortholog of grl, Hey2, is expressed in the developing cardiovascular system, and it has been shown to be a direct target of Notch signaling in vitro (Nakagawa et al., 2000). Although knockout of the Hey2 gene does not seem to affect artery/vein fate (Donovan et al., 2002; Fischer et al., 2002; Sakata et al., 2002), the double mutation of Hey2 and Hes1 (another Hairy-related transcription factor) produces a loss of arterial markers and vascular shunts, although dorsal aorta formation and morphology are also severely affected (Fisher and Caudy, 1998), which is not the case in Notch–deficient zebrafish. Thus, the activation of Notch signaling seems to be a conserved requirement for the specification of arterialcell fate in vertebrates, but additional research will be required to clarify the potential role of different downstream response genes. C. VEGF Signaling as a Regulator of Arterial–Venous Cell Fate In addition to its general role in establishing and maintaining ECs (described previously), recent work demonstrated a more specific role for the VEGF system in the promotion of arterial endothelial differentiation upstream of Notch signaling (Lawson et al., 2002; Mukouyama et al., 2002; Stalmans et al., 2002). In cultured murine embryonic ECs, the addition of the VEGF-A isoforms VEGF120 and VEGF164 induce ephrinB2 expression up to 50%, whereas the addition of other growth factors, such as nerve growth factor and brain-derived neurotrophic factor, can only promote ephrinB2-positive cells to about 10% (Mukouyama et al., 2002). Recent studies have shown that postnatal mice that express only the VEGF188 isoform (but not the predominant VEGF164 or VEGF120 isoforms) have reduced numbers of ephrinB2–positive arterioles in their retinas, whereas the number of venules and capillaries is unaffected (Stalmans et al., 2002). Conversely, mice that carry a VEGF164 transgene have increased numbers of ephrinB2-positive capillaries and a concomitant decrease in EphB4-positive blood vessels in the heart (Visconti et al., 2002). Taken together, these results suggest that VEGF is sufficient to promote

ARTERIAL–VENOUS DIFFERENTIATION

735

arterial endothelial differentiation in vivo and in vitro, and it seems to do so independently of its ability to induce proliferation or the survival of all ECs. The studies in mice are consistent with those in the zebrafish in demonstrating that VEGF is necessary and sufficient for arterial differentiation. A reduction of VEGF-A levels in zebrafish embryos by antisense morpholino injection prevents the expression of artery-specific markers such as ephrinB2, and it blocks the formation of arteries, whereas veins are largely unaffected. These defects can be rescued by the activation of Notch signaling (Lawson et al., 2002). However, VEGF is unable to rescue the arterial defects seen in embryos with reduced Notch activity. These results, together with additional experiments performed in zebrafish (Lawson et al., 2002) indicate that VEGF acts upstream of Notch signaling during arterial differentiation. Although analogous in vivo experiments to demonstrate VEGF functions upstream of Notch have not been performed in mammals, exogenous VEGF can induce the expression of Notch1 and Delta4 in human ECs in vitro (Liu et al., 2003). The mechanisms underlying the artery-specific effects of VEGF–A remain unclear, because the major receptors (VEGFR-1 and VEGFR-2) are expressed on all ECs (reviewed by Ferrara et al., 2003; Klagsbrun and Eichmann, 2005). The NPs, which are VEGF coreceptors, are differentially expressed on arteries and veins. NP1 is preferentially expressed in arteries, whereas NP2 expression is restricted to the venous/lymphatic endothelium (Herzog et al., 2001). However, arterial–venous restriction precedes differential NP expression, and experiments carried out in both zebrafish and mice (Lawson et al., 2002; Mukouyama et al., 2002) support the idea that differential NP expression probably reinforces an arterial–venous decision that has already been initiated. Receptor output could also be regulated by the type of ligand available in the local environment and by genetic interactions among the different VEGF receptors contributing to blood vessel diversity during development (Covassin et al., 2006). D. The Role of Sonic Hedgehog in Arterial–Venous Differentiation Hedgehogs are a class of 19-kDa proteins that interact with heparin on the cell surface through an N-terminal basic domain and that are tethered to the surface through cholesterol and fatty acyl modification. Hedgehog signaling is crucial throughout development; Sonic hedgehog (Shh) is important for the determination of cell types in structures that lie adjacent to the notochord, and increasing evidence suggests a role for this signaling molecule in angiogenesis (reviewed by Lawson and Weinstein, 2002). Studies in zebrafish have shown that Shh acts upstream of VEGF to induce arterial endothelial differentiation (Lawson et al., 2002). The expression of VEGF is upregulated by Shh in both mouse (Pola et al., 2001) and zebrafish (Lawson et al., 2002) embryos, whereas its expression is lost in the somites and hypochord of zebrafish lacking Shh function, which results in the formation of a single midline axial blood vessel that expresses venous markers only (Lawson et al., 2002). In mice, Shh administered to aged animals induces new vessel growth in ischemic hind limbs, but Shh has no effect on EC migration or proliferation in vitro, although it does induce the expression of proangiogenic VEGF and angiopoietins-1 and -2 from interstitial mesenchymal cells (Pola et al., 2001). The two studies described suggest that Shh plays an indirect role in angiogenesis as an activator of other downstream angiogenic factors (Lawson et al., 2002; Pola et al., 2001), but other recent work suggests that Hedgehog

736

BLOOD VESSEL FORMATION

A molecular pathway for arterial–venous fate determination. Studies in the zebrafish have shown that vascular endothelial growth factor acts downstream of Sonic hedgehog and upstream of the Notch pathway to determine arterial cell fate. A variety of different methods were used to either increase (left side) or decrease (right side) the levels and/or activities of each of these signaling pathways, as shown. Loss of Notch, vascular endothelial growth factor, or Sonic hedgehog signaling results in the loss of arterial identity, whereas the exogenous activation or overexpression of these factors causes the ectopic expression of arterial markers. “Molecular epistasis” experiments were performed by combining different methods to assemble these components into an ordered pathway. (For further information about the zebrafish studies used to derive this pathway, see Lawson et al., 2001 and 2002).

FIGURE 33.4

signals may also be received by ECs directly to promote their proper morphogenesis into tubular vessels (Vokes et al., 2004). Together, all of the studies described show that a molecular pathway consisting of the sequential activation of Hedgehog, VEGF, and Notch signaling regulates arterial differentiation in developing zebrafish (Figure 33.4) and probably also in mammals.

III. EMERGENCE OF THE LYMPHATIC SYSTEM In addition to the blood vascular system, vertebrates possess a completely separate and parallel network of endothelial vessels called the lymphatic vascular system. Unlike the blood circulatory system, the lymphatic system is a blindended system of vessels that protect and maintain the fluid environment of the body by filtering and draining away lymphatic fluid. Lymphatic fluid is a clear, colorless fluid that contains water, dissolved molecules, and a few blood cells. The lymphatic system is not closed, and it has no single, central pump. The mammalian and avian lymphatic systems begin with innumerable blind-ended, thin-walled capillaries and larger vessels that drain lymphatic fluid from the extracellular spaces of all organs and tissues into larger

EMERGENCE OF THE LYMPHATIC SYSTEM

737

collecting tubes (Oliver, 2004). These vessels are lined with a continuous single layer of overlapping ECs that form loose intercellular junctions, which makes them highly permeable to large macromolecules, pathogens, and migrating cells. Larger lymph vessels have one-way, semilunar valves, and the lymph moves slowly and under low pressure as a result of the action of surrounding skeletal muscles, which help to squeeze fluid through them. This fluid is transported to progressively larger lymphatic vessels that culminate in the right lymphatic duct (for lymph from the right upper body) and the thoracic duct (for the rest of the body). These ducts drain into the blood circulatory system at the right and left subclavian vein. Under normal conditions, the lymphatic vascular system is necessary for the return of extravasated interstitial fluid and macromolecules to the blood circulation, for immune defense, and for the uptake of dietary fats. It has an important role during embryonic development, and the growth and proliferation of lymphatic vessels is an essential feature of tissue repair and inflammation in most organs (Leu et al., 2000). Impaired functioning of lymphatic vessels can result in the formation of lymphoedema (Witte et al., 2001), whereas tumor-associated lymphangiogenesis may contribute to the spread of cancer cells from solid tumors. Thus far, it is unclear how tumor cells enter the lymphatic system; however, using lymphatic-specific molecular markers, many studies have shown that tumor cells activate peritumoral and intratumoral lymphangiogenesis (Mandriota et al., 2001; Skobe et al., 2001; Stacker et al., 2001). In contrast with the extensive molecular and functional characterization of blood vascular endothelium, comparatively little is known about the mechanisms that control the formation, differentiation, and function of lymphatic vessels. The lack of specific markers has made it difficult to elucidate the mechanisms that underlie the development of the lymphatic system, and its origin has remained controversial. The most widely accepted view of early lymphatic development was described by Sabin (1902). On the basis of ink injection experiments, she postulated that the two primitive jugular lymph sacs originated from ECs that bud from large veins early during development. The peripheral lymphatic vessels subsequently form by centrifugal sprouting from these primary lymph sacs. More recent studies have provided support for this model, and they have suggested molecular players that might be important for different stages of lymphatic EC (LEC) emergence and specification (Figure 33.5). Several studies with mice deficient in the homeobox transcription factor Prox-1 (Wigle and Oliver, 1999) or VEGF-C/VEGFR3 signaling (Karkkainen et al., 2004; Makinen et al., 2001) support this model. Prox-1, although broadly expressed during embryonic development, has been identified as a specific marker of a subpopulation of ECs that give rise to the lymphatic system (Oliver et al., 1993; Wigle and Oliver, 1999). As early as embryonic day 10.5, Prox-1–positive cells are detected in the wall of the cardinal vein. As development proceeds, these Prox-1–positive cells appear to bud from the cardinal vein to give rise to the lymphatic jugular sacs. Interestingly, the inactivation of Prox-1 in mice leads to a complete arrest of lymphatic system development (Wigle and Oliver, 1999). By contrast, vasculogenesis and angiogenesis are unaffected, which demonstrates that Prox–1 activity is specifically required for the normal development of the lymphatic system. These findings indicate that Prox–1 is a “master gene” in the program, specifying LEC fate.

738

BLOOD VESSEL FORMATION

FIGURE 33.5 A proposed model for lymphatic emergence from venous endothelium (modified from Oliver, 2004). Lymphatic endothelium emerges from “competent” venous endothelium expressing the marker LYVE-1. Polarized expression of Prox-1 on a subset of venous endothelial cells “biases” these cells toward producing lymphatic endothelium and induces or allows for the continued maintenance of the expression of a number of different lymphatic endothelial cell genes, including vascular endothelial growth factor receptor 3, the receptor for the lymphangiogenic factor vascular endothelial growth factor C. (See Oliver [2004] for a more comprehensive discussion of this model.)

The signals that determine which ECs in the cardinal vein will activate Prox–1 and become lymphatic are unknown (Oliver and Detmar, 2002). The migration of the LECs toward the lymph sacs was shown to be critically dependent on the presence of the growth factor VEGF-C. In mice that are deficient for this growth factor, Prox-1–expressing LECs are formed, but they fail to migrate toward the lymph sacs and subsequently die around embryonic day 17 as a result of the formation of massive lymphedema, because they do not develop any lymphatic vessels (Karkkainen et al., 2004). VEGF-C specifically binds to its high-affinity tyrosine kinase receptor VEGFR-3 (reviewed by Jussila and Alitalo, 2002). During mouse embryogenesis, the pattern of expression of VEGFR-3 (Flt-4) also coincides with Sabin’s model of lymphatic development. VEGFR-3 is first expressed in a subset of blood vascular ECs, and it subsequently becomes restricted to LECs (Kaipainen et al., 1995; Oh et al., 1997) Recent studies have demonstrated that signaling via VEGFR-3 is sufficient to induce lymphangiogenesis in transgenic mice (Jeltsch et al., 1997; Makinen et al., 2001). Moreover, the expression of a dominant–negative VEGFR-3 in the skin of transgenic mice blocks lymphangiogenesis andinduces the regression of already formed lymphatic vessels, which demonstrates that VEGFR-3 signaling is required not only for the initial formation but also for the maintenance of the lymphatic vasculature (Makinen et al., 2001). However, consistent with the earlier expression of VEGFR-3 in the blood vasculature is the fact that VEGFR-3–deficient mice show defective blood vessel development at early embryonic stages, and the embryos die on embryonic day 9.5 (Dumont, 1998; Dumont et al., 1998). Thus, it seems that VEGFR-3 has an essential function in the remodeling of the primary capillary vasculature before the formation of the lymphatic vessels. VEGF-C also binds to the Neuropilin-2 receptor, which is specifically expressed by veins and

PATTERNING OF THE DEVELOPING VASCULATURE

739

which subsequently becomes restricted to lymphatic vessels (Karkkainen et al., 2001; Yuan et al., 2002). Mice deficient for NRP-2 show selective defects in the formation of lymphatic vessels (Yuan et al., 2002). An alternative model for the emergence of the lymphatic formation was proposed by Huntington and McClure (1910), who suggested that mesenchymal lymphangioblast-like cells are the source of lymphatic vessels and that lymphatics arise in the mesenchyme independent of the veins and then subsequently establish venous connections. This model has been supported by work performed in avian embryos by quail/chick grafting experiments (Schneider et al., 1999; Wilting et al., 2001, 2006) and in Xenopus (Ny et al., 2005), in which both transdifferentiated venous cells and lymphangioblasts have been reported to contribute to newly formed lymph vessels. The debate surrounding LEC origins remains unresolved in large part because of the lack of an effective model organism that allows one to easily observe lymphatic cells in vivo and to perform defined genetic and experimental manipulation of the lymphatic system. Recently, however, the presence of a well-defined lymphatic system that shares characteristics of lymphatic vessels found in higher vertebrates was reported in the zebrafish (Yaniv et al., 2006). Using live imaging of transgenic zebrafish, researchers were able to trace the origins of LECs from their region of origin and through their incorporation into the thoracic duct (the main lymphatic vessel described in vertebrates), thus providing the first direct in vivo evidence for a venous origin for primitive lymphatic vessels, as proposed by Sabin a century ago. Further direct in vivo imaging studies in the zebrafish should similarly allow one to determine whether later-forming and/or peripheral lymphatics arise predominantly through the further proliferation of these initial LECs or via recruitment from mesenchyme as proposed for birds and frogs.

IV. PATTERNING OF THE DEVELOPING VASCULATURE The gross vascular anatomy is characterized by a reproducible pattern of blood vessels. At least for the major vessels (e.g., the aorta), characteristic features such as lumen size, branching angles, and curvature along the vascular tree are quite reproducible. There are also designated sites for secondary sprouts (e.g., intersomitic vessels, main vessels penetrating different organs), whereas microvessels and capillaries formed by intussusceptive angiogenesis are mostly nonstereotyped. The control of branch patterning includes both attractive and repulsive guidance signals, and it is regulated by both positive and negative regulators. The cellular and molecular mechanisms that govern blood vessel assembly at appropriate sites in the organism are poorly understood, yet understanding this regulation is critical to the ability to design therapeutics around vessel production in vivo. In the next few sections, we review recent advances in the understanding of how vascular patterning is established during embryonic development. A. Assembly of the Primary Axial Vessels The formation of the main axial vessels of the trunk—the dorsal aorta and the cardinal vein—occurs by vasculogenesis, which is the local aggregation

740

BLOOD VESSEL FORMATION

of angioblast progenitors arising in the mesoderm. Angioblasts specified in the lateral mesoderm migrate to the midline to form vascular cords that subsequently undergo morphogenesis into the lumenized vascular tubes of the axial vessels (the dorsal aorta and the cardinal vein). Evidence suggests that midline cues are important for the assembly of these initial vasculogenic vessels. Studies in the zebrafish have shown that the notochord is required for dorsal aorta formation. Zebrafish embryos with mutations in the floating head (encodes Xnot, a homeobox gene; Schulte-Merker et al., 1994) or no tail (encodes Brachyury; Talbot et al., 1995) genes lack a differentiated notochord, and they also specifically lack a dorsal aorta, although they still form the cardinal vein (Fouquet et al., 1997; Sumoy et al., 1997). Wild-type notochord cells transplanted back into floating head mutants can locally rescue the assembly of aortic primordia (Fouquet et al., 1997), which suggests that notochord-derived signals are required for aortic specification. The floating head and no tail mutants also fail to form the hypochord, an endodermally derived thin strip of cells that lies immediately ventral to the notochord, just above the aorta in fish and amphibian embryos. Studies in Xenopus and zebrafish have shown that the hypochord expresses a soluble short isoform of VEGF that could potentially act as a medium to a long-range graded signal for the medial migration and assembly of the angioblasts that contribute to the immediately juxtaposed dorsal aorta (Cleaver and Krieg, 1998; Lawson et al., 2002; Liang et al., 2001). However, experiments designed to directly test this idea have not yet been performed, and studies in the zebrafish suggest that VEGF expressed by the adjacent somites in response to notochord-derived Hedgehog signals may be more critical than hypochord-derived VEGF for dorsal aorta assembly (Lawson et al., 2002). The loss of VEGF activity causes a loss of dorsal aorta formation in mice, although it is not clear what tissues provide midline vascular patterning signals. Avians and mice lack a hypochord, but embryonic structures such as the somites or the primitive endoderm ventral to the aorta express VEGF and could be mediating arterial angioblast migration and aorta assembly. The role of the notochord in dorsal aorta formation in avians and mice is also not clear. Mouse Brachyury mutants lack the posterior notochord, but they still form a posterior dorsal aorta (Hogan and Bautch, 2004); however, this does not rule out the role of the notochord in the anterior region, where the smoothened phenotype is most severe (Vokes et al., 2004). In avians and mice, the dorsal aortae are initially present as a pair of vessels on either side of the midline rather than a single midline vessel, as they are in fish and amphibians. Studies using quail/chick chimeras in which axial structures were removed showed that the notochord acts as a midline barrier to impede avian angioblasts from crossing the axial midline (Klessinger and Christ, 1996), and a subsequent report showed that this was the result of notochord-expressed bone morphogenetic protein antagonists (Reese et al., 2004). Other recent studies using mouse/avian chimeras have shown that the neural tube is the source of a positive patterning signal, and they have identified this signal once again as VEGF-A (Ambler et al., 2001, 2003; Hogan and Bautch, 2004). In addition to VEGF, Hedgehog signaling also seems to be important for the assembly of the dorsal aorta in zebrafish and mice (Ingham et al., 2000; Lawson et al., 2002; Vokes et al., 2004). As noted previously, notochord-derived

PATTERNING OF THE DEVELOPING VASCULATURE

741

Hedgehog signaling has been shown to be important in the pathway leading to arterial differentiation in zebrafish because it induces VEGF expression in the somites. However, unlike deficiencies in Notch signaling, the loss of Hedgehog signaling leads to a complete loss of the dorsal aorta and not simply to its misspecification as vein, which suggests that Hedgehog signaling plays a role in dorsal aorta assembly in addition to its activity upstream of arterial differentiation (Lawson et al., 2002). Recent studies in Xenopus and mice confirm this, showing that Hedgehog signaling is important for the morphogenesis of a vascular tube (Vokes et al., 2004). B. The Role of Guidance Factors in Developmental Angiogenesis After the assembly of embryonic primary vasculogenic vessels such as the dorsal aorta and the cardinal vein, further remodeling and ramification of the vasculature are achieved by developmental angiogenesis, as described previously. Angiogenesis in adults is in most cases clearly directed and guided by local tissue requirements for oxygen and nutrients, but, during early development, it follows highly stereotypic and evolutionarily well-conserved patterns of vascular network assembly that are reminiscent of the way in which the initial assembly of the nervous system and axon tracts follows a conserved and stereotypic program of assembly (see Chapter 24). Indeed, the two systems display remarkable anatomic parallels, and vessels and nerves frequently course adjacent to one another. This has led to the idea that the assembly of the vasculature might depend on similar (or even the same) attractive and repulsive guidance cues from surrounding tissues that help to direct growing vessels along specific pathways (Weinstein, 2005). A variety of recent studies have begun to show that many of the molecular pathways used for the guidance of migrating axons are also employed in the direction of the stereotypic pathways followed by developing blood vessels. The formation of the trunk intersegmental vessels provides a good example of a stereotypic and conserved process of vascular network assembly (Figure 33.6, A and B). Intersegmental vessels are conserved features of the trunk vasculature of all vertebrates. They form bilaterally on opposite sides of the trunk along each intersomitic boundary (vertical myoseptum). Studies in zebrafish have shown that these vessels grow and elongate in a choreographed and reproducible fashion (Isogai et al., 2003). Initially, primary vascular sprouts emerge from the dorsal aorta bilaterally at each intersomitic boundary, and they grow dorsally around the notochord and neural tube. As they reach the dorsal–lateral surface of the neural tube, they branch rostrally and caudally, fusing with adjacent intersegmental sprouts to form a pair of continuous vessels along the dorsal trunk called the dorsal longitudinal anastomotic vessels. The formation of the primary sprout-derived vascular lattice is followed by a wave of secondary vascular sprouts that emerge from the posterior cardinal vein. About half of these connect with the base of primary segments, which then become intersegmental veins. The rest of the primary segments, which remain connected only to the dorsal aorta, become intersegmental arteries. Recent studies have shown that well-known neuronal guidance factors play important roles in the guidance and patterning of these and other developing vessels, although in many cases the details of their vascular activities remain unclear.

742

BLOOD VESSEL FORMATION

FIGURE 33.6 Trunk vascular network assembly and its guidance. A, The anatomy of the zebrafish trunk and its blood vessels by approximately three days postfertilization. At this stage, there is active flow through the dorsal aorta (DA), the posterior cardinal vein (PCV), and most intersegmental arteries (ISA) and intersegmental veins (ISV). The intersegmental arteries and veins are linked together dorsally via paired dorsal longitudinal anastomotic vessels (DLAV). All of these vessels are shown relative to adjacent tissues and structures in the mid trunk, including the gut (G), the myotomes (M), the notochord (N), the neural tube (NT), the left pronephric duct (P), and the yolk mass (Y). In addition to the functioning vessels noted here, parachordal vessels (PAV) run longitudinally to either side of the notochord, along the horizontal myoseptum. At 3 days postfertilization, the parachordal vessels do not yet carry flow. Anterior is to the left and above the plane of the page, and dorsal is up. B, Schematic diagram illustrating the steps that lead to the assembly of the trunk angiogenic vascular network. For clarity, the diagram shows the vessels on only one side of the trunk. B.i, Primary sprouts emerge bilaterally exclusively from the dorsal aorta (red). B.ii, Primary sprouts grow dorsally, branching cranially and caudally at the level of the dorsal–lateral roof of the neural tube. B.iii, Branches interconnect on either side of the trunk to form two dorsal longitudinal anastomotic vessels (DLAV). B.iv, Secondary sprouts begin to emerge exclusively from the posterior cardinal vein (blue). B.v, Some secondary sprouts connect to the base of primary segments, whereas others do not. B.vi, Primary segments with patent connections to secondary segments become intersegmental veins (blue), whereas primary segments that remain connected only to the dorsal aorta become intersegmental arteries (red). Most of the secondary sprouts that do not connect to primary segments serve instead as ventral roots for the parachordal vessels (PAV). C and D, Blood vessels in the mid trunk of control morpholino (C) or plexinD1 morpholino (D) injected 48 hour postfertilization into fli1–EGFP transgenic embryos. In control morpholino-injected animals, intersegmental vessels extend along the boundaries between somites, avoiding the semaphorin-rich central regions (C). In animals deficient in plexinD1, intersegmental vessels sprout, branch, and grow without regard for somitic boundaries (D). Anterior is to the left and dorsal is up in all panels. (Panels A and B are modified from Isogai et al., 2003. Panels C and D are modified from Torres–Vazquez et al., 2004. See these references for further details. See color insert.)

1. Semaphorin Signaling Some of the most conclusive evidence for neuronal guidance factors playing important roles in guiding and patterning the developing vasculature has come from studies of semaphorin signaling and blood vessels in zebrafish and mice. Semaphorins are a large family of cell-associated and secreted proteins that signal through multimeric receptors (Bagri and Tessier-Lavigne, 2002). Membrane-associated semaphorins bind directly to plexin receptors, whereas secreted semaphorins bind to NP-plexin receptor complexes. Recent

PATTERNING OF THE DEVELOPING VASCULATURE

743

work has shown that semaphorin–plexin signaling regulates the guidance and patterning of the vasculature in a manner similar to that of the repulsive guidance roles of semaphorins in the nervous system. ECs express various NP and plexin receptors, including the endothelial-specific receptor plexinD1 (Basile et al., 2004; Gitler et al., 2004; Miao et al., 1999; Soker et al., 1998; Torres-Vazquez et al., 2004). In zebrafish, type 3 semaphorins are expressed in the center of each somite, and semaphorin-plexinD1 signaling mediates the repulsive guidance of growing intersegmental vessels to restrict their paths to semaphorin-free corridors along the intersomitic boundaries (TorresVazquez et al., 2004). The loss of function of either the plexinD1 receptor or the trunk semaphorins in zebrafish causes the mispatterning of intersegmental vessels, which sprout at irregular positions and grow and branch aberrantly throughout the trunk instead of maintaining paths along the intersomitic boundaries (see Figure 33.6, C and D). In mice, the targeted inactivation of plexinD1 also causes the mispatterning of intersegmental vessels as well as increased vascularization of the somites, which normally exclude vessels (Gitler et al., 2004; Gu et al., 2005). In this case, the primary ligand is apparently Sema3E, which is normally expressed in a region of the somites that is adjacent to the intersegmental vessels (Gu et al., 2005), although the murine plexinD1 receptor is capable of responding to Sema3A in an NP-dependent manner (Gitler et al., 2004). However, a mouse mutant that is deficient in semaphorin binding to both NP1 and NP2 forms intersegmental vessels normally, which suggests that Sema3E binds plexinD1 directly rather than via an NP receptor. Further studies will be needed to work out the precise ligand–receptor interactions for semaphorin signaling in the vasculature. 2. Slit–Robo Signaling Several recent studies have implicated Slits and their receptors in angiogenesis. There are four slit receptors (“Roundabout” receptors or Robos) in mammals (Robos). Robo4, which is structurally divergent from the other Robos, shows highly endothelial-cell–specific expression both in vitro and during mouse embryogenesis (Park et al., 2003). In the adult, Robo4 is present at sites of both normal and pathologic active angiogenesis, including that of tumor vessels (Huminiecki et al., 2002). The role of Slit–Robo signaling in vascular guidance remains controversial. One study showed that Robo4 binds Slit0002 and that it is able to inhibit the migration of Robo4-expressing cells in vitro (Park et al., 2003), but other studies either failed to detect such binding (Suchting et al., 2005) or demonstrated a promigratory effect of Slit2 on ECs (Wang et al., 2003). Substantial vascular defects have not been reported in either Robo4 or Slit ligand knockout mice, although this could be explained by redundancy in the expression of both ligands and receptors (Long et al., 2004). However, defects in intersegmental vessel formation have been reported in zebrafish after the morpholino-mediated knockdown of Robo4 (Bedell et al., 2005). Again, further studies will be required to determine thenature of the in vivo role of Slit–Robo signaling during vascular development. 3. Netrin Signaling Another important set of guidance cues during nervous system patterning is provided by the netrins, which are a family of highly conserved laminin-

744

BLOOD VESSEL FORMATION

related secreted proteins (Hedgecock et al., 1990; Ishii et al., 1992; Serafini et al., 1994). Netrins can mediate the attraction of neurons by activating Deleted in colorectal cancer (DCC) receptor family members that are expressed on axons (Fazeli et al., 1997; Serafini et al., 1996), including DCC and neogenin (Chan et al., 1996; Keino-Masu et al., 1996). Conversely, netrin binding to members of the Uncoordinated-5 (UNC5) receptor family results in axon repulsion. Like Slits, the role of netrins in the guidance of the vasculature is still somewhat unclear. Several different studies carried out both in vitro and in vivo in mice and zebrafish have shown that netrin-1 is a proangiogenic factor for vascular ECs (Park et al., 2004; Wilson et al., 2006), but another study has shown that netrin-1 can act as a repellent in vesselguidance via the UNC5B receptor (Lu et al., 2004). In the zebrafish, netrin-1a is expressed along the horizontal myoseptum that divides the dorsal and ventral halves of the somites. One report suggests that the loss of netrin-1a prevents the formation of the parachordal vessels that normally run through the horizontal myoseptum, which suggests that netrin-1a provides a positive cue for the growth of these vessels (Wilson et al., 2006). However, another report suggests that the loss of netrin-1a in the zebrafish promotes increased vessel branching and growth along the somites (Lu et al., 2004). The differing conclusions of all of these studies may at least partly reflect the biology of netrins, which are capable of both repulsive and attractive signaling, depending on the cellular and environmental context. Further studies will be required to clarify the differences between these and other results regarding the activity of netrins and their receptors in the vasculature. There is clearly much to be done before our understanding of how guidance factors regulate vascular patterning begins to approach the depth of current understanding of how these factors regulate neural patterning. The expression of many of these factors adjacent to or in opposition with developing vessel tracts suggests that they play important roles in establishing the anatomic form of vascular networks (Figure 33.7), but conclusive in vivo evidence is, in most cases, still lacking. It is likely that, as is seen in the nervous system, extensively redundant guidance factors for the vasculature coordinate their activities in a complex spatial and temporal interaction to shape the stereotypic pattern of early blood vessels. The strong parallels uncovered between these two systems already imply that studies carried out in the nervous system that are aimed at dissecting this interaction will continue to have relevance to the vascular system (and, perhaps, vice versa).

V. CONCLUDING REMARKS The differentiation of ECs and the formation of the vascular system are of central importance during embryonic vascular development. Research carried out over the past years has shed light on the different steps involved in the formation of the vascular system. The emergence of endothelial progenitors, their coalescence into the primary vascular system (vasculogenesis) and the further remodeling and elaboration of the initial plexus (angiogenesis), the differentiation of vessels into arteries and veins, and the formation of the lymphatic system have been extensively studied. Genetic evidence has highlighted the roles of many molecules that affect vasculogenesis, angiogenesis, and lymphangiogenesis in a complex and tightly regulated manner. Many

CONCLUDING REMARKS

745

FIGURE 33.7 Members of various families of guidance factors are expressed in discrete patterns in the developing trunk. Netrins and Slits are expressed in the ventral neural tube. Netrin is expressed along the horizontal myoseptum, and semaphorins and ephrins are expressed within the somites. These signaling molecules are well positioned to provide potential repulsive or attractive guidance cues for blood vessels such as the intersegmental vessels (which run vertically along intersomitic boundaries), the parachordal vessels (which run longitudinally along the horizontal myosepta), and the vertebral arteries (which run longitudinally on either side of the ventral neural tube). NT, Neural tube; No, notochord; DA, dorsal aorta; PCV, posterior cardinal vein (Image from Weinstein, 2005.)

of the signaling pathways implicated in vascular development are reactivated during disease states of angiogenesis or vessel regression, thus making a full understanding of the complexities of these pathways important for identifying new targets for therapeutic intervention during pathologic situations such as severe tissue ischemia, coronary heart disease, and tumor-promoted angiogenesis. A particularly interesting aspect of recent research is the heterogeneity displayed by ECs. To date, it is clear that, although all vessels share the same endothelial basis, each cell type has acquired unique characteristics that are vital to cardiovascular system function. To what extent is the formation of different types of vessels defined solely by intrinsic programs or influenced by local cues? How does the intimate association with their cognate organs influence ECs to adopt functional specialties such as the blood–brain barrier and the fenestrated endothelium in the kidney glomeruli? The answers to these questions are of vital importance when aiming to refine therapeutic applications to specific subsets of the vasculature. Much of the past decade of research on vascular development has focused on the central role played by VEGF and its receptors. Although there is still much more to uncover about VEGF signaling, a challenge for the coming decade will be to incorporate our understanding of the role of this pathway into a larger framework of multiple and sometimes highly specific regulators. Genetic, molecular, and cell biologic tools are now available for the study of vessel formation in a diverse array of model systems, thus creating a wide and useful array of tools for vascular research. Because many of our insights into the mechanisms underlying the formation of the vascular system have come from developmental studies, it seems likely that further work involving the early stages of vascular formation will continue to shed light on the activities of these pathways during normal and pathologic adult neovascularization.

746

BLOOD VESSEL FORMATION

SUMMARY

 Primitive blood vessels in vertebrate embryos form by a process called vas-









culogenesis, during which mesodermal cells differentiate into endothelial precursor cells called angioblasts. These angioblasts then differentiate in situ into ECs, and they coalesce to form the earliest vessels. The subsequent growth and remodeling of the vasculature occur by angiogenesis, during which new blood vessels form from preexisting vessels by endothelial sprouting and splitting. The VEGF signaling pathway has been shown to be important for the migration, proliferation, maintenance, and survival of ECs and to be critical during both vasculogenesis and angiogenesis. In addition to its important function during embryonic development, VEGF is thought to play a major role in tumor-induced neovascularization. Angiopoietin–Tie signaling is important for the maturation and stabilization of the nascent vascular network. One of the most fundamental steps in the differentiation of the vasculature is the specification of arterial and venous endothelium. Recent work has highlighted the role of Notch, VEGF, and Hedgehog signaling in the specification of arterial identity. Classically, differences between arteries and veins were attributed to physiologic factors such as the direction and pressure of blood flow. However, recent work has shown that molecular distinctions between arterial and venous endothelium appear before the onset of blood flow, and it has highlighted the roles of genetic pathways including Notch, VEGF, and Hedgehog signaling in specifying arterial identity. Vertebrates possess a second, blind-ended vascular system, the lymphatic system, that is responsible for clearing and draining fluids and macromolecules that leak from blood vessels into the interstitial spaces of tissues and organs. Evidence suggests that the first LECs emerge by transdifferentiation from the ECs of primitive veins. The transcription factor Prox1 is a critical regulator of LEC specification, whereas the VEGF family members VEGF-C and VEGF-D are important for LEC migration and lymphangiogenesis. During early development, newly formed vessels follow a highly stereotypic and evolutionarily well-conserved pattern of network assembly that is reminiscent of that followed by the nervous system and axon tracts. A variety of recent studies have shown that many well-known molecular pathways used for the guidance of migrating axons play an important role in the guidance and patterning of developing blood vessels.

ACKNOWLEDGEMENT This work was supported by the intramural program of the NICHD and by an EMBO fellowship to KY.

GLOSSARY Angioblast Mesodermal-derived endothelial precursors that have certain characteristics of endothelial cells but that have not yet assembled into functional vessels.

747

REFERENCES

Angiogenesis The formation of new blood vessels from preexisting vessels by endothelial sprouting and splitting. This process of remodeling of the primary capillary network leads to the formation of mature arteries and veins. Lymphatic system A blind-ended system of vessels that protect and maintain the fluid environment of the body by filtering and draining away lymphatic fluid. Under normal conditions, the lymphatic vascular system is necessary for the return of extravasated interstitial fluid and macromolecules to the blood circulation, for immune defense, and for the uptake of dietary fats. Vasculogenesis The process of the formation of primitive vessels in vertebrate embryos during which mesodermal cells differentiate into endothelial precursor cells called angioblasts. These angioblasts then differentiate in situ into endothelial cells and coalesce to form the earliest vessels. Vascular endothelial growth factor The first growth factor described to be a specific mitogen for endothelial cells. It is important for the migration, proliferation, maintenance, and survival of endothelial cells, and it is critical during both vasculogenesis and angiogenesis. It also plays an important role in arterial differentiation.

REFERENCES Adams RH, Diella F, Hennig S, et al: The cytoplasmic domain of the ligand ephrinB2 is required for vascular morphogenesis but not cranial neural crest migration, Cell 104:57–69, 2001. Adams RH, Wilkinson GA, Weiss C, et al: Roles of ephrinB ligands and EphB receptors in cardiovascular development: demarcation of arterial/venous domains, vascular morphogenesis, and sprouting angiogenesis, Genes Dev 13:295–306, 1999. Ambler CA, Nowicki JL, Burke AC, Bautch VL: Assembly of trunk and limb blood vessels involves extensive migration and vasculogenesis of somite–derived angioblasts, Dev Biol 234:352–364, 2001. Ambler CA, Schmunk GM, Bautch VL: Stem cell–derived endothelial cells/progenitors migrate and pattern in the embryo using the VEGF signaling pathway, Dev Biol 257:205–219, 2003. Artavanis-Tsakonas S, Rand MD, Lake RJ: Notch signaling: cell fate control and signal integration in development, Science 284:770–776, 1999. Bagri A, Tessier-Lavigne M: Neuropilins as Semaphorin receptors: in vivo functions in neuronal cell migration and axon guidance, Adv Exp Med Biol 515:13–31, 2002. Barrantes IB, Elia AJ, Wunsch K, et al: Interaction between Notch signalling and Lunatic fringe during somite boundary formation in the mouse, Curr Biol 9:470–480, 1999. Basile JR, Barac A, Zhu T, et al: Class IV semaphorins promote angiogenesis by stimulating Rho-initiated pathways through plexin–B, Cancer Res 64:5212–5224, 2004. Bedell VM, Yeo SY, Park KW, et al: roundabout4 is essential for angiogenesis in vivo, Proc Natl Acad Sci U S A 102:6373–6378, 2005. Bergwerff M, Verberne ME, DeRuiter MC, et al: Neural crest cell contribution to the developing circulatory system: implications for vascular morphology? Circ Res 82:221–231, 1998. Carmeliet P: Mechanisms of angiogenesis and arteriogenesis, Nat Med 6:389–395, 2000. Carmeliet P, Collen D Molecular basis of angiogenesis. Role of VEGF and VE–cadherin, Ann N Y Acad Sci 902:249–262; discussion 263–264, 2000. Carmeliet P, Conway EM: Growing better blood vessels, Nat Biotechnol 19:1019–1020, 2001. Carmeliet P, Ferreira V, Breier G, et al: Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele, Nature 380:435–439, 1996.

748

BLOOD VESSEL FORMATION

Chan SS, Zheng H, Su MW, et al: UNC-40, a C. elegans homolog of DCC (Deleted in Colorectal Cancer), is required in motile cells responding to UNC–6 netrin cues, Cell 87:187–195, 1996. Childs S, Chen JN, Garrity DM, Fishman MC: Patterning of angiogenesis in the zebrafish embryo, Development 129:973–982, 2002. Chitnis AB: The role of Notch in lateral inhibition and cell fate specification, Mol Cell Neurosci 6:311–321, 1995. Choi K, Kennedy M, Kazarov A, et al: A common precursor for hematopoietic and endothelial cells, Development 125:725–732, 1998. Chung YS, Zhang WJ, Arentson E, et al: Lineage analysis of the hemangioblast as defined by FLK1 and SCL expression, Development 129:5511–5520, 2002. Cleaver O, Krieg PA: VEGF mediates angioblast migration during development of the dorsal aorta in Xenopus, Development 125:3905–3914, 1998. Cleaver O, Krieg PA: Molecular mechanisms of vascular development, In Harvey RP, Rosenthal N, editors: Heart development, San Diego, 1999, Academic Press, pp. 221–252. Coffin JD, Poole TJ: Embryonic vascular development: immunohistochemical identification of the origin and subsequent morphogenesis of the major vessel primordia in quail embryos. Development 102:735–748, 1988. Covassin LD, Villefranc JA, Kacergis MC, et al: Distinct genetic interactions between multiple Vegf receptors are required for development of different blood vessel types in zebrafish, Proc Natl Acad Sci U S A 103:6554–6559, 2006. Davis S, Aldrich TH, Jones PF, et al: Isolation of angiopoietin–1, a ligand for the TIE2 receptor, by secretion–trap expression cloning, Cell 87:1161–1169, 1996. Del Amo FF, Smith DE, Swiatek PJ, et al: Expression pattern of Motch, a mouse homolog of Drosophila Notch, suggests an important role in early postimplantation mouse development, Development 115:737–744, 1992. DeRuiter MC, Poelmann RE, VanMunsteren JC, et al: Embryonic endothelial cells transdifferentiate into mesenchymal cells expressing smooth muscle actins in vivo and in vitro, Circ Res 80:444–451, 1997. Doetschman TC, Eistetter H, Katz M, et al: The in vitro development of blastocyst–derived embryonic stem cell lines: formation of visceral yolk sac, blood islands and myocardium, J Embryol Exp Morphol 87:27–45, 1985. Donovan J, Kordylewska A, Jan YN, Utset MF: Tetralogy of Fallot and other congenital heart defects in Hey2 mutant mice, Curr Biol 12:1605–1610, 2002. Drake CJ, Fleming PA: Vasculogenesis in the day 6.5 to 9.5 mouse embryo, Blood 95:1671–1679, 2000. Duarte A, Hirashima M, Benedito R, et al: Dosage–sensitive requirement for mouse Dll4 in artery development, Genes Dev 18:2474–2478, 2004. Dumont DJ, Fong GH, Puri MC, et al: Vascularization of the mouse embryo: a study of flk–1, tek, tie, and vascular endothelial growth factor expression during development, Dev Dyn 203:80–92, 1995. Dumont DJ, Jussila L, Taipale J, et al: Cardiovascular failure in mouse embryos deficient in VEGF receptor–3, Science 282:946–949, 1998. Eichmann A, Corbel C, Nataf V, et al: Ligand–dependent development of the endothelial and hemopoietic lineages from embryonic mesodermal cells expressing vascular endothelial growth factor receptor 2, Proc Natl Acad Sci U S A 94:5141–5146, 1997. Eichmann A, Yuan L, Moyon D, et al: Vascular development: from precursor cells to branched arterial and venous networks, Int J Dev Biol 49:259–267, 2005. Emerick KM, Krantz ID, Kamath BM, et al: Intracranial vascular abnormalities in patients with Alagille syndrome, J Pediatr Gastroenterol Nutr 41:99–107, 2005. Fazeli A, Dickinson SL, Hermiston ML, et al: Phenotype of mice lacking functional Deleted in colorectal cancer (Dcc) gene, Nature 386:796–804, 1997. Ferkowicz MJ, Starr M, Xie X, et al: CD41 expression defines the onset of primitive and definitive hematopoiesis in the murine embryo, Development 130:4393–4403, 2003. Ferrara N: Vascular endothelial growth factor, Eur J Cancer 32A:2413–2422, 1996. Ferrara N: Molecular and biological properties of vascular endothelial growth factor, J Mol Med 77:527–543, 1999. Ferrara N, Gerber HP, LeCouter J: The biology of VEGF and its receptors, Nat Med 9:669–676, 2003. Fina L, Molgaard HV, Robertson D, et al: Expression of the CD34 gene in vascular endothelial cells, Blood 75:2417–2426, 1990.

REFERENCES

749 Fischer A, Leimeister C, Winkler C, et al: Hey bHLH factors in cardiovascular development, Cold Spring Harb Symp Quant Biol 67:63–70, 2002. Fisher A, Caudy M: The function of hairy–related bHLH repressor proteins in cell fate decisions, Bioessays 20:298–306, 1998. Flamme I, Frolich T, Risau W: Molecular mechanisms of vasculogenesis and embryonic angiogenesis, J Cell Physiol 173:206–210, 1997. Folkman J, D’Amore PA: Blood vessel formation: what is its molecular basis? Cell 87:1153–1155, 1996. Fong GH, Rossant J, Gertsenstein M, Breitman ML: Role of the Flt–1 receptor tyrosine kinase in regulating the assembly of vascular endothelium, Nature 376:66–70, 1995. Fouquet B, Weinstein BM, Serluca FC, Fishman MC: Vessel patterning in the embryo of the zebrafish: guidance by notochord, Dev Biol 183:37–48, 1997. Fryxell KJ, Soderlund M, Jordan TV: An animal model for the molecular genetics of CADASIL (cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy), Stroke 32:6–11, 2001. Gale NW, Baluk P, Pan L, et al: Ephrin–B2 selectively marks arterial vessels and neovascularization sites in the adult, with expression in both endothelial and smooth–muscle cells, Dev Biol 230:151–160, 2001. Gale NW, Dominguez MG, Noguera I, et al: Haploinsufficiency of delta–like 4 ligand results in embryonic lethality due to major defects in arterial and vascular development, Proc Natl Acad Sci U S A 101:15949–15954, 2004. Gale NW, Thurston G, Hackett SF, et al: Angiopoietin–2 is required for postnatal angiogenesis and lymphatic patterning, and only the latter role is rescued by Angiopoietin–1, Dev Cell 3:411–423, 2002. Gale NW, Yancopoulos GD: Growth factors acting via endothelial cell–specific receptor tyrosine kinases: VEGFs, angiopoietins, and ephrins in vascular development, Genes Dev 13: 1055–1066, 1999. Gerber HP, Hillan KJ, Ryan AM, et al: VEGF is required for growth and survival in neonatal mice, Development 126:1149–1159, 1999. Gerety SS, Anderson DJ: Cardiovascular ephrinB2 function is essential for embryonic angiogenesis, Development 129:1397–1410, 2002. Gerety SS, Wang HU, Chen ZF, Anderson DJ: Symmetrical mutant phenotypes of the receptor EphB4 and its specific transmembrane ligand ephrin–B2 in cardiovascular development, Mol Cell 4:403–414, 1999. Gitler AD, Lu MM, Epstein JA: PlexinD1 and semaphorin signaling are required in endothelial cells for cardiovascular development, Dev Cell 7:107–116, 2004. Goishi K, Klagsbrun M: Vascular endothelial growth factor and its receptors in embryonic zebrafish blood vessel development, Curr Top Dev Biol 62:127–152, 2004. Gridley T: Notch signaling and inherited disease syndromes, Hum Mol Genet 12:R9–13, 2003. Grunewald M, Avraham I, Dor Y, et al: VEGF–induced adult neovascularization: recruitment, retention, and role of accessory cells, Cell 124:175–189, 2006. Gu C, Yoshida Y, Livet J, et al: Semaphorin 3E and plexin–D1 control vascular pattern independently of neuropilins, Science 307:265–268, 2005. Hedgecock EM, Culotti JG, Hall DH: The unc–5, unc–6, and unc–40 genes guide circumferential migrations of pioneer axons and mesodermal cells on the epidermis in C. elegans, Neuron 4:61–85, 1990. Herzog Y, Guttmann–Raviv N, Neufeld G: Segregation of arterial and venous markers in subpopulations of blood islands before vessel formation, Dev Dyn 232:1047–1055, 2005. Herzog Y, Kalcheim C, Kahane N, et al: Differential expression of neuropilin–1 and neuropilin–2 in arteries and veins, Mech Dev 109:115–119, 2001. His W: Lecithoblast und angioblast der wirbeltiere, Abhandl Math–Phys Ges Wiss 26:171–328, 1900. Hogan KA, Bautch VL: Blood vessel patterning at the embryonic midline, Curr Top Dev Biol 62:55–85, 2004. Huber TL, Kouskoff V, Fehling HJ, et al: Haemangioblast commitment is initiated in the primitive streak of the mouse embryo, Nature 432:625–630, 2004. Huminiecki L, Gorn M, Suchting S, et al: Magic roundabout is a new member of the roundabout receptor family that is endothelial specific and expressed at sites of active angiogenesis, Genomics 79:547–552, 2002. Hungerford JE, Little CD: Developmental biology of the vascular smooth muscle cell: building a multilayered vessel wall, J Vasc Res 36:2–27, 1999.

750

BLOOD VESSEL FORMATION

Huntington G, McClure C: The anatomy and development of the jugular lymph sac in the domestic cat (Felis domestica), Am J Anat 10:177–311, 1910. Ilan N, Mahooti S, Madri JA: Distinct signal transduction pathways are utilized during the tube formation and survival phases of in vitro angiogenesis, J Cell Sci 111(Pt 24):3621–3631, 1998. Ingham PW, Nystedt S, Nakano Y, et al: Patched represses the Hedgehog signalling pathway by promoting modification of the Smoothened protein, Curr Biol 10:1315–1318, 2000. Ishii N, Wadsworth WG, Stern BD, et al: UNC–6, a laminin–related protein, guides cell and pioneer axon migrations in C. elegans, Neuron 9:873–881, 1992. Isogai S, Horiguchi M, Weinstein BM: The vascular anatomy of the developing zebrafish: an atlas of embryonic and early larval development, Dev Biol 230:278–301, 2001. Isogai S, Lawson ND, Torrealday S, et al: Angiogenic network formation in the developing vertebrate trunk, Development 130:5281–5290, 2003. Jaffredo T, Bollerot K, Sugiyama D, et al: Tracing the hemangioblast during embryogenesis: developmental relationships between endothelial and hematopoietic cells, Int J Dev Biol 49:269–277, 2005a. Jaffredo T, Nottingham W, Liddiard K, et al: From hemangioblast to hematopoietic stem cell: an endothelial connection? Exp Hematol 33:1029–1040, 2005b. Jakeman LB, Armanini M, Phillips HS, Ferrara N: Developmental expression of binding sites and messenger ribonucleic acid for vascular endothelial growth factor suggests a role for this protein in vasculogenesis and angiogenesis, Endocrinology 133:848–859, 1993. Jeltsch M, Kaipainen A, Joukov V, et al: Hyperplasia of lymphatic vessels in VEGF–C transgenic mice, Science 276:1423–1425, 1997. Jiang YJ, Brand M, Heisenberg CP, et al: Mutations affecting neurogenesis and brain morphology in the zebrafish, Danio rerio, Development 123:205–216, 1996. Jussila L, Alitalo K: Vascular growth factors and lymphangiogenesis, Physiol Rev 82:673–700, 2002. Kaipainen A, Korhonen J, Mustonen T, et al: Expression of the fms–like tyrosine kinase 4 gene becomes restricted to lymphatic endothelium during development, Proc Natl Acad Sci U S A 92:3566–3570, 1995. Kalimo H, Ruchoux MM, Viitanen M, Kalaria RN: CADASIL: a common form of hereditary arteriopathy causing brain infarcts and dementia, Brain Pathol 12:371–384, 2002. Kallianpur AR, Jordan JE, Brandt SJ: The SCL/TAL–1 gene is expressed in progenitors of both the hematopoietic and vascular systems during embryogenesis, Blood 83:1200–1208, 1994. Kamath BM, Spinner NB, Emerick KM, et al: Vascular anomalies in Alagille syndrome: a significant cause of morbidity and mortality, Circulation 109:1354–1358, 2004. Karkkainen MJ, Haiko P, Sainio K, et al: Vascular endothelial growth factor C is required for sprouting of the first lymphatic vessels from embryonic veins, Nat Immunol 5:74–80, 2004. Karkkainen MJ, Saaristo A, Jussila L, et al: A model for gene therapy of human hereditary lymphedema, Proc Natl Acad Sci U S A 98:12677–12682, 2001. Kearney JB, Ambler CA, Monaco KA, et al: Vascular endothelial growth factor receptor Flt–1 negatively regulates developmental blood vessel formation by modulating endothelial cell division, Blood 99:2397–2407, 2002. Keino–Masu K, Masu M, Hinck L, et al: Deleted in colorectal cancer (DCC) encodes a netrin receptor, Cell 87:175–185, 1996. Kinder SJ, Loebel DA, Tam PP: Allocation and early differentiation of cardiovascular progenitors in the mouse embryo, Trends Cardiovasc Med 11:177–184, 2001. Klagsbrun M, D’Amore PA: Vascular endothelial growth factor and its receptors, Cytokine Growth Factor Rev 7:259–270, 1996. Klagsbrun M, Eichmann A: A role for axon guidance receptors and ligands in blood vessel development and tumor angiogenesis, Cytokine Growth Factor Rev 16:535–548, 2005. Klessinger S, Christ B: Axial structures control laterality in the distribution pattern of endothelial cells, Anat Embryol (Berl) 193:319–330, 1996. Krebs LT, Xue Y, Norton CR, et al: Notch signaling is essential for vascular morphogenesis in mice, Genes Dev 14:1343–1352, 2000. Kukk E, Lymboussaki A, Taira S, et al: VEGF–C receptor binding and pattern of expression with VEGFR–3 suggests a role in lymphatic vascular development, Development 122:3829–3837, 1996. Kullander K, Klein R: Mechanisms and functions of Eph and ephrin signalling, Nat Rev Mol Cell Biol 3:475–486, 2002.

REFERENCES

751 Lawson ND, Scheer N, Pham VN, et al: Notch signaling is required for arterial–venous differentiation during embryonic vascular development, Development 128:3675–3683, 2001. Lawson ND, Vogel AM, Weinstein BM: sonic hedgehog and vascular endothelial growth factor act upstream of the Notch pathway during arterial endothelial differentiation, Dev Cell 3:127–136, 2002. Lawson ND, Weinstein BM: Arteries and veins: making a difference with zebrafish, Nat Rev Genet 3:674–682, 2002. le Noble F, Moyon D, Pardanaud L, et al: Flow regulates arterial–venous differentiation in the chick embryo yolk sac, Development 131:361–375, 2004. Leu AJ, Berk DA, Lymboussaki A, et al: Absence of functional lymphatics within a murine sarcoma: a molecular and functional evaluation, Cancer Res 60:4324–4327, 2000. Liang D, Chang JR, Chin AJ, et al: The role of vascular endothelial growth factor (VEGF) in vasculogenesis, angiogenesis, and hematopoiesis in zebrafish development, Mech Dev 108:29–43, 2001. Liang D, Xu X, Chin AJ, et al: Cloning and characterization of vascular endothelial growth factor (VEGF) from zebrafish. Danio rerio, Biochim Biophys Acta 1397:14–20, 1998. Liao W, Bisgrove BW, Sawyer H, et al: The zebrafish gene cloche acts upstream of a flk–1 homologue to regulate endothelial cell differentiation, Development 124:381–389, 1997. Liu ZJ, Shirakawa T, Li Y, et al: Regulation of Notch1 and Dll4 by vascular endothelial growth factor in arterial endothelial cells: implications for modulating arteriogenesis and angiogenesis, Mol Cell Biol 23:14–25, 2003. Long H, Sabatier C, Ma L, et al: Conserved roles for Slit and Robo proteins in midline commissural axon guidance, Neuron 42:213–223, 2004. Lu X, Le Noble F, Yuan L, et al: The netrin receptor UNC5B mediates guidance events controlling morphogenesis of the vascular system, Nature 432:179–186, 2004. Maisonpierre PC, Suri C, Jones PF, et al: Angiopoietin–2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis, Science 277:55–60, 1997. Makinen T, Adams RH, Bailey J, et al: PDZ interaction site in ephrinB2 is required for the remodeling of lymphatic vasculature, Genes Dev 19:397–410, 2005. Makinen T, Jussila L, Veikkola T, et al: Inhibition of lymphangiogenesis with resulting lymphedema in transgenic mice expressing soluble VEGF receptor–3, Nat Med 7:199–205, 2001. Mandriota SJ, Jussila L, Jeltsch M, et al: Vascular endothelial growth factor–C–mediated lymphangiogenesis promotes tumour metastasis, EMBO J 20:672–682, 2001. Marron MB, Hughes DP, Edge MD, et al: Evidence for heterotypic interaction between the receptor tyrosine kinases TIE–1 and TIE–2, J Biol Chem 275:39741–39746, 2000. Miao HQ, Soker S, Feiner L, et al: Neuropilin–1 mediates collapsin–1/semaphorin III inhibition of endothelial cell motility: functional competition of collapsin–1 and vascular endothelial growth factor–165, J Cell Biol 146:233–242, 1999. Moyon D, Pardanaud L, Yuan L, et al: Plasticity of endothelial cells during arterial–venous differentiation in the avian embryo, Development 128:3359–3370, 2001. Mukouyama YS, Shin D, Britsch S, et al: Sensory nerves determine the pattern of arterial differentiation and blood vessel branching in the skin, Cell 109:693–705, 2002. Murray PDF: The development in vitro of the blood of the early chick embryo, Proc Roy Soc London 11:497–521, 1932. Nakagawa O, McFadden DG, Nakagawa M, et al: Members of the HRT family of basic helix– loop–helix proteins act as transcriptional repressors downstream of Notch signaling, Proc Natl Acad Sci U S A 97:13655–13660, 2000. Nakagawa O, Nakagawa M, Richardson JA, et al: HRT1, HRT2, and HRT3: a new subclass of bHLH transcription factors marking specific cardiac, somitic, and pharyngeal arch segments, Dev Biol 216:72–84, 1999. Nikolova G, Lammert E: Interdependent development of blood vessels and organs, Cell Tissue Res 314:33–42, 2003. Ny A, Koch M, Schneider M, et al: A genetic Xenopus laevis tadpole model to study lymphangiogenesis, Nat Med 11:998–1004, 2005. Nye JS, Kopan R: Developmental signaling. Vertebrate ligands for Notch, Curr Biol 5:966–969, 1995. Oh SJ, Jeltsch MM, Birkenhager R, et al: VEGF and VEGF–C: specific induction of angiogenesis and lymphangiogenesis in the differentiated avian chorioallantoic membrane, Dev Biol 188:96–109, 1997.

752

BLOOD VESSEL FORMATION

Oliver G, Detmar M: The rediscovery of the lymphatic system: old and new insights into the development and biological function of the lymphatic vasculature, Genes Dev 16:773–783, 2002. Oliver G, Sosa–Pineda B, Geisendorf S, et al: Prox 1, a prospero–related homeobox gene expressed during mouse development, Mech Dev 44:3–16, 1993. Othman–Hassan K, Patel K, Papoutsi M, et al: Arterial identity of endothelial cells is controlled by local cues, Dev Biol 237:398–409, 2001. Pardanaud L, Dieterlen–Lievre F: Manipulation of the angiopoietic/hemangiopoietic commitment in the avian embryo, Development 126:617–627, 1999. Park KW, Crouse D, Lee M, et al: The axonal attractant Netrin–1 is an angiogenic factor, Proc Natl Acad Sci U S A 101:16210–16215, 2004. Park KW, Morrison CM, Sorensen LK, et al: Robo4 is a vascular–specific receptor that inhibits endothelial migration, Dev Biol 261:251–267, 2003. Pola R, Ling LE, Silver M, et al: The morphogen Sonic hedgehog is an indirect angiogenic agent upregulating two families of angiogenic growth factors, Nat Med 7:706–711, 2001. Popoff D: Dottersack–gafa´sse des Huhnes, Wiesbaden, 1894, Kreidl’s Verlag. Puri MC, Rossant J, Alitalo K, et al: The receptor tyrosine kinase TIE is required for integrity and survival of vascular endothelial cells, EMBO J 14:5884–5891, 1995. Reaume AG, Conlon RA, Zirngibl R, et al: Expression analysis of a Notch homologue in the mouse embryo, Dev Biol 154:377–387, 1992. Reese DE, Hall CE, Mikawa T: Negative regulation of midline vascular development by the notochord, Dev Cell 6:699–708, 2004. Risau W: Mechanisms of angiogenesis, Nature 386:671–674, 1997. Risau W, Flamme I: Vasculogenesis, Annu Rev Cell Dev Biol 11:73–91, 1995. Rosenquist TH, Beall AC: Elastogenic cells in the developing cardiovascular system. Smooth muscle, nonmuscle, and cardiac neural crest, Ann N Y Acad Sci 588:106–119, 1990. Rossant J, Hirashima M: Vascular development and patterning: making the right choices, Curr Opin Genet Dev 13:408–412, 2003. Rossant J, Howard L: Signaling pathways in vascular development, Annu Rev Cell Dev Biol 18:541–573, 2002. Ruchoux MM, Chabriat H, Bousser MG, et al: Presence of ultrastructural arterial lesions in muscle and skin vessels of patients with CADASIL, Stroke 25:2291–2292, 1994. Ruchoux MM, Domenga V, Brulin P, et al: Transgenic mice expressing mutant Notch3 develop vascular alterations characteristic of cerebral auttosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy, Am J Pathol 162:329–342, 2003. Sabin F: On the origin of the lymphatic system from the veins, and the development of the lymph hearts and thoracic duct in the pig, Am J Anat 1:367–389, 1902. Saharinen P, Kerkela K, Ekman N, et al: Multiple angiopoietin recombinant proteins activate the Tie1 receptor tyrosine kinase and promote its interaction with Tie2, J Cell Biol 169:239–243, 2005. Sakata Y, Kamei CN, Nakagami H, et al: Ventricular septal defect and cardiomyopathy in mice lacking the transcription factor CHF1/Hey2, Proc Natl Acad Sci U S A 99:16197–16202, 2002. Sato TN, Qin Y, Kozak CA, Audus KL: Tie–1 and tie–2 define another class of putative receptor tyrosine kinase genes expressed in early embryonic vascular system, Proc Natl Acad Sci U S A 90:9355–9358, 1993. Sato TN, Tozawa Y, Deutsch U, et al: Distinct roles of the receptor tyrosine kinases Tie–1 and Tie–2 in blood vessel formation, Nature 376:70–74, 1995. Scappaticci FA: Mechanisms and future directions for angiogenesis–based cancer therapies, J Clin Oncol 20:3906–3927, 2002. Schneider M, Othman–Hassan K, Christ B, Wilting J: Lymphangioblasts in the avian wing bud, Dev Dyn 216:311–319, 1999. Schroder JM, Sellhaus B, Jorg J: Identification of the characteristic vascular changes in a sural nerve biopsy of a case with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), Acta Neuropathol (Berl) 89:116–121, 1995. Schulte–Merker S, van Eeden FJ, Halpern ME, et al: no tail (ntl) is the zebrafish homologue of the mouse T (Brachyury) gene, Development 120:1009–1015, 1994. Serafini T, Colamarino SA, Leonardo ED, et al: Netrin–1 is required for commissural axon guidance in the developing vertebrate nervous system, Cell 87:1001–1014, 1996. Serafini T, Kennedy TE, Galko MJ, et al: The netrins define a family of axon outgrowth–promoting proteins homologous to C. elegans UNC–6, Cell 78:409–424, 1994.

REFERENCES

753 Shalaby F, Rossant J, Yamaguchi TP, et al: Failure of blood–island formation and vasculogenesis in Flk–1–deficient mice, Nature 376:62–66, 1995. Shirayoshi Y, Yuasa Y, Suzuki T, et al: Proto–oncogene of int–3, a mouse Notch homologue, is expressed in endothelial cells during early embryogenesis, Genes Cells 2:213–224, 1997. Shivdasani RA, Mayer EL, Orkin SH: Absence of blood formation in mice lacking the T–cell leukaemia oncoprotein tal–1/SCL, Nature 373:432–434, 1995. Shutter JR, Scully S, Fan W, et al: Dll4, a novel Notch ligand expressed in arterial endothelium, Genes Dev 14:1313–1318, 2000. Shweiki D, Itin A, Neufeld G, et al: Patterns of expression of vascular endothelial growth factor (VEGF) and VEGF receptors in mice suggest a role in hormonally regulated angiogenesis, J Clin Invest 91:2235–2243, 1993. Skobe M, Hawighorst T, Jackson DG, et al: Induction of tumor lymphangiogenesis by VEGF–C promotes breast cancer metastasis, Nat Med 7:192–198, 2001. Soker S, Takashima S, Miao HQ, et al: Neuropilin–1 is expressed by endothelial and tumor cells as an isoform–specific receptor for vascular endothelial growth factor, Cell 92:735–745, 1998. Stacker SA, Caesar C, Baldwin ME, et al: VEGF–D promotes the metastatic spread of tumor cells via the lymphatics, Nat Med 7:186–191, 2001. Stainier DY, Weinstein BM, Detrich HW 3rd, et al: Cloche, an early acting zebrafish gene, is required by both the endothelial and hematopoietic lineages, Development 121:3141–3150, 1995. Stalmans I, Ng YS, Rohan R, et al: Arteriolar and venular patterning in retinas of mice selectively expressing VEGF isoforms, J Clin Invest 109:327–336, 2002. Suchting S, Heal P, Tahtis K, et al: Soluble Robo4 receptor inhibits in vivo angiogenesis and endothelial cell migration, FASEB J 19:121–123, 2005. Sumoy L, Keasey JB, Dittman TD, Kimelman D: A role for notochord in axial vascular development revealed by analysis of phenotype and the expression of VEGR–2 in zebrafish flh and ntl mutant embryos, Mech Dev 63:15–27, 1997. Suri C, Jones PF, Patan S, et al: Requisite role of angiopoietin–1, a ligand for the TIE2 receptor, during embryonic angiogenesis, Cell 87:1171–1180, 1996. Suri C, McClain J, Thurston G, et al: Increased vascularization in mice overexpressing angiopoietin–1, Science 282:468–471, 1998. Swiatek PJ, Lindsell CE, del Amo FF, et al: Notch1 is essential for postimplantation development in mice, Genes Dev 8:707–719, 1994. Taichman DB, Loomes KM, Schachtner SK, et al: Notch1 and Jagged1 expression by the developing pulmonary vasculature, Dev Dyn 225:166–175, 2002. Talbot WS, Trevarrow B, Halpern ME, et al: A homeobox gene essential for zebrafish notochord development, Nature 378:150–157, 1995. Thompson MA, Ransom DG, Pratt SJ, et al: The cloche and spadetail genes differentially affect hematopoiesis and vasculogenesis, Dev Biol 197:248–269, 1998. Torres–Vazquez J, Gitler AD, Fraser SD, et al: Semaphorin–plexin signaling guides patterning of the developing vasculature, Dev Cell 7:117–123, 2004. Tsiamis AC, Morris PN, Marron MB, Brindle NP: Vascular endothelial growth factor modulates the Tie–2:Tie–1 receptor complex, Microvasc Res 63:149–158, 2002. Uyttendaele H, Ho J, Rossant J, Kitajewski J: Vascular patterning defects associated with expression of activated Notch4 in embryonic endothelium, Proc Natl Acad Sci U S A 98: 5643–5648, 2001. Uyttendaele H, Marazzi G, Wu G, et al: Notch4/int–3, a mammary proto–oncogene, is an endothelial cell–specific mammalian Notch gene, Development 122:2251–2259, 1996. Vajkoczy P, Menger MD, Vollmar B, et al: Inhibition of tumor growth, angiogenesis, and microcirculation by the novel Flk–1 inhibitor SU5416 as assessed by intravital multi–fluorescence videomicroscopy, Neoplasia 1:31–41, 1999. Vargesson N, Patel K, Lewis J, Tickle C: Expression patterns of Notch1, Serrate1, Serrate2 and Delta1 in tissues of the developing chick limb, Mech Dev 77:197–199, 1998. Villa N, Walker L, Lindsell CE, et al: Vascular expression of Notch pathway receptors and ligands is restricted to arterial vessels, Mech Dev 108:161–164, 2001. Visconti RP, Richardson CD, Sato TN: Orchestration of angiogenesis and arteriovenous contribution by angiopoietins and vascular endothelial growth factor (VEGF), Proc Natl Acad Sci U S A 99:8219–8224, 2002. Visvader JE, Fujiwara Y, Orkin SH: Unsuspected role for the T–cell leukemia protein SCL/tal–1 in vascular development, Genes Dev 12:473–479, 1998.

754

BLOOD VESSEL FORMATION

Vittet D, Prandini MH, Berthier R, et al: Embryonic stem cells differentiate in vitro to endothelial cells through successive maturation steps, Blood 88:3424–3431, 1996. Vokes SA, Yatskievych TA, Heimark RL, et al: Hedgehog signaling is essential for endothelial tube formation during vasculogenesis, Development 131:4371–4380, 2004. Vrancken Peeters MP, Gittenberger–de Groot AC, Mentink MM, Poelmann RE: Smooth muscle cells and fibroblasts of the coronary arteries derive from epithelial–mesenchymal transformation of the epicardium, Anat Embryol (Berl) 199:367–378, 1999. Wang B, Xiao Y, Ding BB, et al: Induction of tumor angiogenesis by Slit–Robo signaling and inhibition of cancer growth by blocking Robo activity, Cancer Cell 4:19–29, 2003. Wang HU, Chen ZF, Anderson DJ: Molecular distinction and angiogenic interaction between embryonic arteries and veins revealed by ephrin–B2 and its receptor Eph–B4, Cell 93:741–753, 1998. Wang R, Clark R, Bautch VL: Embryonic stem cell–derived cystic embryoid bodies form vascular channels: an in vitro model of blood vessel development, Development 114:303–316, 1992. Weinstein BM: Plumbing the mysteries of vascular development using the zebrafish, Semin Cell Dev Biol 13:515–522, 2002. Weinstein BM: Vessels and nerves: marching to the same tune, Cell 120:299–302, 2005. Wettstein DA, Turner DL, Kintner C: The Xenopus homolog of Drosophila Suppressor of hairless mediates Notch signaling during primary neurogenesis, Development 124:693–702, 1997. Wigle JT, Oliver G: Prox1 function is required for the development of the murine lymphatic system, Cell 98:769–778, 1999. Wiles MV, Keller G: Multiple hematopoietic lineages develop from embryonic stem (ES) cells in culture, Development 111:259–267, 1991. Willett CG, Boucher Y, di Tomaso E, et al: Direct evidence that the VEGF–specific antibody bevacizumab has antivascular effects in human rectal cancer, Nat Med 10:145–147, 2004. Williams SK, Gillis JF, Matthews MA, et al: Isolation and characterization of brain endothelial cells: morphology and enzyme activity, J Neurochem 35:374–381, 1980. Wilson BD, Ii M, Park KW, et al: Netrins promote developmental and therapeutic angiogenesis, Science 313:640–644, 2006. Wilting J, Aref Y, Huang R, et al: Dual origin of avian lymphatics, Dev Biol 292:165–173, 2006. Wilting J, Papoutsi M, Othman–Hassan K, et al: Development of the avian lymphatic system, Microsc Res Tech 55:81–91, 2001. Witte MH, Bernas MJ, Martin CP, Witte CL: Lymphangiogenesis and lymphangiodysplasia: from molecular to clinical lymphology, Microsc Res Tech 55:122–145, 2001. Xue Y, Gao X, Lindsell CE, et al: Embryonic lethality and vascular defects in mice lacking the Notch ligand Jagged1, Hum Mol Genet 8:723–730, 1999. Yancopoulos GD, Davis S, Gale NW, et al: Vascular–specific growth factors and blood vessel formation, Nature 407:242–248, 2000. Yancopoulos GD, Klagsbrun M, Folkman J: Vasculogenesis, angiogenesis, and growth factors: ephrins enter the fray at the border, Cell 93:661–664, 1998. Yaniv K, Isogai S, Castranova D, et al: Live imaging of lymphatic development in the zebrafish, Nat Med 12:711–716, 2006. Yuan L, Moyon D, Pardanaud L, et al: Abnormal lymphatic vessel development in neuropilin 2 mutant mice, Development 129:4797–4806, 2002. Zhong TP, Childs S, Leu JP, Fishman MC: Gridlock signalling pathway fashions the first embryonic artery, Nature 414:216–220, 2001. Zhong TP, Rosenberg M, Mohideen MA, et al: gridlock, an HLH gene required for assembly of the aorta in zebrafish, Science 287:1820–1824, 2000. Zimrin AB, Pepper MS, McMahon GA, et al: An antisense oligonucleotide to the notch ligand jagged enhances fibroblast growth factor–induced angiogenesis in vitro, J Biol Chem 271:32499–32502, 1996.

RECOMMENDED RESOURCES Coultas L, Chawengsaksophak K, Rossant J: Endothelial cells and VEGF in vascular development, Nature 438:937–945, 2005. Oliver G: Lymphatic vasculature development, Nat Rev Immunol 4:35–45, 2004. Torres-Vazquez J, Kamei M, Weinstein BM: Molecular distinction between arteries and veins, Cell Tissue Res 314:43–59, 2003.

34

BLOOD INDUCTION AND EMBRYONIC FORMATION XIAOYING BAI and LEONARD I. ZON Howard Hughes Medical Institute, Children’s Hospital of Boston, Boston, MA

INTRODUCTION Each day, our body produces billions of new white blood cells, red blood cells, and platelets to replace the blood cells lost during the process of cell turnover. These mature blood cells are generated by a small population of stem cells called hematopoietic stem cells (HSCs) that reside in the bone marrow. HSCs are defined by their capacity of self-renew, and their multilineage differentiation can give rise to all blood cells in our body, including erythrocytes, granulocytes (basophils, eosinophils, and neutrophils), lymphocytes (B cells, T cells, and natural killer cells), monocytes/macrophages, and platelets (Figure 34.1). When transplanted into a host that has been lethally irradiated to remove endogenous HSCs, the donor-derived HSCs can reconstitute all blood lineages throughout life span (Fleischman et al., 1982; Harrison et al., 1988; Spangrude et al., 1988; Jordan and Lemischka, 1990; Chaddah et al., 1996; Osawa et al., 1996). The emergence of HSCs and the blood system can be traced back to the early stages of embryogenesis. Pioneering studies addressing the formation of the blood system were performed in chick embryos because of the easy accessibility of embryos for observation and manipulation. Later, this work was extended into mammalian models such as the mouse. More recently, Xenopus and zebrafish models have been extensively used to decipher the molecular pathways involved in hematopoiesis. These studies have revealed a remarkably conserved developmental program of hematopoiesis. In addition, in vitro studies using hematopoietic progenitor cell culture and embryonic stem cell (ESC)-derived embryonic bodies greatly contribute to our understanding of the hematopoietic hierarchy. In this chapter, we will review the general process of hematopoiesis in these vertebrate model organisms and highlight the important molecules that regulate blood development.

Principles of Developmental Genetics © 2007, Elsevier Inc. All rights reserved.

755

756

BLOOD INDUCTION AND EMBRYONIC FORMATION

The general process of hematopoiesis in the vertebrate. Important regulatory factors are shown. The zebrafish mutants are in italics, with the mutant gene indicated in parentheses.

FIGURE 34.1

I. ORIGIN OF BLOOD CELLS DURING EMBRYOGENESIS A. Induction of Hematopoiesis in the Early Embryo In vertebrates, blood cells derive from the ventral mesoderm (see Chapter 13). During gastrulation, the mesoderm is induced and subsequently patterned to adopt a dorsal or ventral fate. Studies in Xenopus and zebrafish have uncovered mesoderm-inducing factors that include members of the transforming growth factor b (TGF-b) family and the fibroblast growth factors (MunozSanjuan and H-Brivanlou, 2001). In addition, the T-box transcription factor VegT has also been found to be important for mesoderm formation (Zhang et al., 1998). The zebrafish mutant spadetail, which is caused by the loss of function of tbx16 (the homologue of VegT), displays a severe defect in the mesodermal and endodermal derivatives in the trunk, including an absence of blood (Kimmel et al., 1989; Ho and Kane, 1990; Thompson et al., 1998). Patterning of the mesoderm is regulated by antagonistic interactions between the ventralizing signals and the dorsalizing signals (Graff, 1997; Thomsen, 1997). Considerable evidence suggests that the ventralizing signals are mediated by members of the bone morphogenetic proteins (BMPs), a subgroup in the TGF-b superfamily. The overexpression of BMP2, BMP4, or BMP7 results in the loss of dorsal derivatives (e.g., the muscle, the notochord), and it expands ventral mesoderm fates (e.g., the blood, the kidney; Dale et al., 1992b; Jones et al., 1992; Fainsod et al., 1994; Clement et al., 1995; Wang et al., 1997). Animal cap (ectodermal) explants in Xenopus revealed that the ectopic expression of BMP4 induces the molecular markers of hematopoiesis, such as Gata-2 and Scl (Maeno et al., 1996; Mead et al., 1998a). Conversely, blocking BMP pathways by the overexpression of a dominant-negative BMP receptor inhibits blood formation and causes an expansion of the dorsal derivatives (Graff et al., 1994; Maeno et al., 1994;

ORIGIN OF BLOOD CELLS DURING EMBRYOGENESIS

757

Suzuki et al., 1994). Furthermore, several zebrafish mutants with disruptions in the BMP pathway, such as swirl (BMP2 mutant), snailhouse (BMP7 mutant), and somitobun (mutant of Smad5, the downstream signal transducer of the BMP pathway), lack the ventral tissues (i.e., dorsalized mutants) and have defects in blood formation (see Figure 34.1; Mullins et al., 1996; Kishimoto et al., 1997; Nguyen et al., 1998; Hild et al., 1999; Dick et al., 2000; Schmid et al., 2000). Conversely, a defect in the chordin gene ventralizes the zebrafish embryo and expands blood formation, thus suggesting that chordin is a dorsalizing signal. Similar dorsalizing molecules are also identified in Xenopus, including chordin, noggin, and follistatin (Piccolo et al., 1996; Zimmerman et al., 1996; Iemura et al., 1998). The overexpression of these factors dorsalizes the ventral mesoderm by interacting with BMP signals to prevent receptor activation. Gene knockout studies in mice also support the role of the BMP pathways in mesoderm patterning and blood induction. Targeted disruption of mouse BMP4, BMP2, or Alk3 (a BMP receptor) genes resulted in severe mesoderm deficiency, and, in certain genetic backgrounds, primitive hematopoiesis in these mutant embryos was severely disrupted (Winnier et al., 1995; Lawson et al., 1999). Recently, members of the Hedgehog family (which are secreted by the visceral endoderm in mouse embryos) have been found to appear to be capable of inducing blood formation in mouse embryo explant cultures, perhaps through the upregulation of BMP4 (Dyer et al., 2001). In addition to its general effect on hematopoiesis through dorsal–ventral patterning, a recent study in zebrafish showed that Alk8, a BMP receptor, regulates the specification of the myeloid lineage during early embryogenesis, thus suggesting that BMP/TGF-b signaling may also regulate hematopoiesis in a lineage-specific pattern (Hogan et al., 2006). The BMP pathways are possibly mediated by the Mix family of transcription factors, which belongs to the paired class of homeobox genes (Mead et al., 1998b). Seven distinct Mix factors have been isolated in Xenopus, including Mix.1, Mix.2, Mix.3, Bix1, Bix2, Bix3, and Bix4 (Mead et al., 1996; Vize, 1996; Ecochard et al., 1998; Henry and Melton, 1998; Tada et al., 1998). Mix genes are transiently expressed in the future mesoderm and/or the endoderm during gastrulation, and they are induced by activin (another TGF-b family member) and BMP. An activin response element has been identified in the Mix.2 promoter (Chen et al., 1997). The overexpression of Mix.1 induces excessive blood formation and ventralizes the embryos in a way that is similar to that seen with the BMP overexpression phenotype (Dale et al., 1992a; Jones et al., 1992). The single Mix gene homologue in mouse, Mixl1/mMix, is expressed in the primitive streak of the gastrulating embryo, and it marks the cells that are destined to form mesoderm and endoderm (Pearce and Evans 1999; Robb et al., 2000). Mixl1 null mice display numerous mesodermal and endodermal defects that result in embryonic lethality at day 8.5 postcoitum (dpc). Using Mixl1 null ESCs, Elefanty et al. recently demonstrated that Mixl1 is required for efficient hematopoiesis and BMP4induced ventral mesoderm patterning (Ng et al., 2005). Conversely, the induction of Mixl in ESC-derived embryonic bodies results in the acceleration of mesoderm development and the expansion of hematopoietic progenitors (Willey et al., 2006). Taken together, these studies suggest that the Mix family may participate in the BMP signaling pathways in patterning the mesoderm toward a ventral and hematopoietic fate.

758

BLOOD INDUCTION AND EMBRYONIC FORMATION

B. From Hemangioblast to Hematopoietic Stem Cells When embryonic hematopoiesis initiates, hematopoietic and endothelial cells emerge simultaneously in close association with each other from the mesoderm. The developmental proximity between these two cell types has led to the hypothesis that they arise from a common progenitor called the hemangioblast. Numerous studies have suggested such a common origin for hematopoietic and endothelial lineages. For example, mice lacking Flk-1, which is an early endothelial marker, fail to generate both endothelial and hematopoietic cells (Shalaby et al., 1995). Flk-1 null ESCs also fail to contribute to either vessels or blood cells in chimeric mice (Shalaby et al., 1995). Furthermore, single Flk-1þ cells from avian embryos can develop into either hematopoietic or endothelial cells, depending on the presence of the vascular endothelial growth factor (Eichmann et al., 1997). Similarly, Scl, which is an early hematopoietic marker, is also required for the endothelial lineage, because Scl null ESCs fail to contribute to the formation of the vitelline vessels in the mouse yolk sac (YS; Visvader et al., 1998). In support of the mouse studies, a zebrafish mutant Cloche lacks both endothelial and hematopoietic cells, and Scl expression is greatly reduced (Stainier et al., 1995; Liao et al., 1997). The overexpression of Scl can partially rescue both blood and vascular defects, which suggests that the mutant gene functions upstream of Scl, perhaps at the hemangioblast level (Liao et al., 1998). Using in vitro differentiated ESCs, Choi et al. (1998) have isolated a blast colony-forming cell (BL-CFC) population. BL-CFCs behave like hemangioblasts in that they express a number of genes that are common to both endothelial and hematopoietic lineages, and they have potential to form either lineage. Recently, Huber et al. (2004) have identified BL-CFCs in the posterior region of the primitive streak of gastrulating mouse embryo, thereby providing evidence for the existence of hemangioblasts in vivo. Most interestingly, adult HSCs isolated from human bone marrow and human cord blood have been shown to have vascularizing potential, thereby suggesting that hemangioblasts are also present during the postnatal stage (Pelosi et al., 2002). C. Primitive Hematopoiesis Hematopoiesis in vertebrate embryos is characterized by two successive waves occurring at anatomically distinct sites. Primitive hematopoiesis is transient, generating cells mainly in the erythroid lineage (see Figure 34.1), although macrophages and megakaryocytes are also found in primitive hematopoiesis. The second wave, which is called definitive hematopoiesis, lasts for the life of the organism and produces HSCs that are capable of giving rise to all blood lineages (see Figure 34.1). In mouse embryos, the first hematopoietic cells appear extraembryonically, within the mesoderm-derived YS blood island (Figure 34.2; Palis et al., 1999; 2001). The expression of the hematopoietic genes Scl and Lim only domain 2 (Lmo-2) marks the initiation of YS hematopoiesis at 7 dpc (Palis et al., 2001). Subsequently, Gata-1, which is an erythroid-specific transcription factor, is detected in the YS (Pevny et al., 1995; Palis et al., 2001). Most of the primitive blood cells are red cells (erythrocytes). Different from the enucleated definitive red cells, primitive erythrocytes retain their nucleus, and they predominantly express the embryonic hemoglobins (z, bH1, and ey;

ORIGIN OF BLOOD CELLS DURING EMBRYOGENESIS

759

Steiner, 1973). Between 8.5 and 9 dpc, the primitive erythrocytes enter the circulation, and, by 9 dpc, the primitive erythropoietic potential of the YS has disappeared (Palis et al., 1999). Similar to mammals, avian primitive hematopoiesis initiates in the YS (see Figure 34.2). Differentiated primitive red blood cells are detected at day 1.5, and embryonic circulation starts by day 2 (Evans, 1997). In Xenopus, the ventral blood island (VBI) is functionally equivalent to the mammalian and avian YS, which also develops from the ventral mesoderm (Mangia et al., 1970). By 24 hours postfertilization (hpf), Scl, c-Myb, and Gata-1 can be detected in the developing VBI (see Figure 34.2; Turpen et al., 1997). Primitive erythrocytes start to express the hemoglobin at 40 hpf, and, by 50 hpf, the circulation is established (Mangia et al., 1970).

FIGURE 34.2 The site of primitive hematopoiesis in different species. Mouse: Upper panel, A mouse embryo (8.5 dpc) from an e-globin lacZ transgenic mouse line stained for the expression of b-galactosidase (dark blue) in primitive erythroid cells in the yolk sac. Middle panel, The in situ hybridization of an embryonic globin probe to a 10.5-dpc mouse embryo section showing the cellular composition of the yolk sac. The yolk sac is seen as a membrane surrounding the embryo, and it contains a series of blood islands. Two of these (rectangle) are shown at higher magnification in the lower panel. Lower panel, Visceral endoderm and mesoderm components are seen enveloping clusters of primitive erythroblasts (Eryp) surrounded by endothelial cells. Chick: Upper panel, A 2-day-old (10 somite pairs) chick embryo showing the expression of the Lmo-2 gene. The dotted pattern reveals the distribution of the blood islands, which will give rise to the first (primitive) generation of erythrocytes. SV, Sinus venosus. Lower panel, Cross-section at the level indicated previously showing the maturation steps of the blood islands (BI). Arrows indicate the hemangioblasts that will give rise to the endothelial cell (EC) and the hematopoietic cell (HC). 1, This immature blood island is full of hematopoietic and endothelial cells that are already differentiated. 2, The blood island has matured; several cells have been freed, creating a hole. 3, The blood island is fully mature. The hematopoietic cells are free and detached from one another. LP, Lateral plate; M, mesoderm. Xenopus: In situ hybridization of the Gata-1 probe to a swimming tadpole-stage embryo reveals the high expression of Gata-1 in the ventral blood island region. Zebrafish: The in situ hybridization of an embryonic globin probe to a 24 hpf embryo reveals the intermediate cell mass region. (Mouse figures adapted from Baron, 2001, with permission. Chick figures adapted from Jaffredo et al., 2003, with permission. Xenopus figures adapted from Mead et al., 2001, with permission. See color insert.)

760

BLOOD INDUCTION AND EMBRYONIC FORMATION

Unlike other vertebrate systems in which primitive hematopoiesis occurs extraembryonically, primitive hematopoiesis in zebrafish takes place in the embryo proper at the intermediate cell mass (ICM). Derived from the posterior intermediate mesoderm, the ICM precursors are first evident around the 2-somite stage by the expression of the hematopoietic marker Scl in bilateral stripes of cells flanking the paraxial mesoderm (Davidson et al., 2003). At the 4-somite stage, the expression of the erythroid-specific transcription factor Gata-1 is detected in a subset of Sclþ cells (Davidson et al., 2003), thus indicating the erythropoietic commitment of the ICM precursors. Erythroid precursors then migrate toward the trunk midline to form the ICM, and, at 15 hpf, they begin expressing embryonic globins (see Figure 34.2; Al-Adhami and Kunz, 1977; Willett et al., 1999). Between 24 and 26 hpf, the heart starts beating, and the erythroblasts enters the circulation, where they subsequently mature into primitive erythrocytes (Willett et al., 1999). D. Definitive Hematopoiesis and Hematopoietic Stem Cells Primitive hematopoiesis is transient and subsequently replaced by the definitive wave of hematopoiesis, which generates HSCs that give rise to all adult blood lineages throughout the life span. Definitive HSCs colonize the fetal liver (FL) and later the bone marrow; however, their origin has been controversial during the past few decades. Using quail–chick chimeras in which the embryo of a quail was grafted onto the YS of a chick, Dieterlen-Lievre (1975) showed that the chick-derived YS cells only contribute to primitive but not definitive hematopoiesis in the quail embryo. Further inspection showed that definitive HSCs come from intraembryonic regions that are closely associated with the aorta (Dieterlen-Lievre and Martin, 1981). Similarly, in mouse embryos, definitive hematopoietic progenitors have been detected at 9 dpc in the aorta–mesonephros–gonads (AGM) region. At 10 dpc, AGMderived cells have long-term repopulating potential in lethally irradiated adult recipients lacking the endogenous HSCs (Muller et al., 1994), thus demonstrating HSC activity in the AGM. Recently, the AGM origin of definitive HSCs was also demonstrated in zebrafish (Figure 34.3, C; Thompson et al., 1998). Morphologically, putative hematopoietic cells appear as clusters budding from the ventral wall of the dorsal aorta as well as from the endothelium of vitelline/umbilical arteries, and they express HSC markers such as Flk-1, Scl, and Runx-1 (see Figure 34.3; Garcia-Porrero et al., 1995; Marshall et al., 1999; North et al., 1999; Tavian et al., 1999). In Runx-1–deficient mouse embryos, the formation of the intra-aortic clusters is disrupted, and definitive hematopoiesis is blocked (North et al., 1999). Therefore, a hypothesis of “hemogenic endothelium” (see Figure 34.1) has been proposed in which definitive HSCs are generated through an endothelial intermediate that has the potential to give rise to hematopoietic cells (Jaffredo et al., 1998). In addition to the AGM region, definitive hematopoietic potential has been described in the mouse YS. Palis et al. (1999) discovered definitive erythroid progenitors expressing the adult globin in the YS at 8 dpc, before the onset of circulation. As circulation begins, these cells are found in the bloodstream and subsequently in the liver, which suggests that YS-derived erythroid cells can colonize the FL. YS cells taken at 9 to 10 dpc can provide long-term (up to 1 year) multilineage blood reconstitution for conditioned newborn recipients, but they cannot be engrafted into irradiated adult recipients (Yoder et al., 1997a;

ORIGIN OF BLOOD CELLS DURING EMBRYOGENESIS

761

FIGURE 34.3 The sites of definitive hematopoiesis in different species. The in situ hybridization of the Runx-1 probe in A, mouse, B, chick, and C, zebrafish embryos reveals Runx-1 expression in the ventral endothelium of the dorsal aorta. In C, the cross section of the dorsal aorta is shown on the right, with the dorsal aorta (red arrowhead) and the Runx-1 in situ signal (blue arrowhead) indicated. (Adapted from Jeffredo et al., 2005, with permission. See color insert.)

1997b); When they are cultured with AGM-derived stromal cells, however, YS cells isolated at 8 dpc provide long-term HSCs in adult recipients (Matsuoka et al., 2001), which suggests that the YS-derived cells have the potential to become definitive HSCs given the proper environmental cues. Although the contribution of YS-derived cells to definitive hematopoiesis in mice is still under debate, the VBI in Xenopus has been demonstrated to give rise to both primitive and definitive hematopoiesis. As mentioned previously, primitive hematopoiesis occurs at the VBI in Xenopus. The second site of amphibian hematopoiesis, the dorsal–lateral plate, is analogous to the mammalian AGM (Kau and Turpen, 1983; Maeno et al., 1985). Both VBI– and dorsal–lateralplate–derived cells are found to colonize the liver and the thymus, where they later give rise to all blood lineages in larvae and adults (Kau and Turpen, 1983; Smith et al., 1989; Bechtold et al., 1992; Chen and Turpen, 1995). Recent studies have suggested that there may be new anatomic sites participating in HSC development during vertebrate embryogenesis. Caprioli et al. (1998; 2001) have discovered cells with hemangioblastic potential in the avian allantois. Similarly, hematopoietic stem cell activity has been described in the murine placenta, which is an equivalent structure to the avian allantois. Alvarez-Silva et al. (2003) found that the murine placenta contains a large number of hematopoietic progenitors that can differentiate into multilineages when they are cultured in vitro. Gekas et al. (2005) showed that HSCs with long-term reconstitution activity could be found in the placenta before their appearance in the bloodstream, which is

762

BLOOD INDUCTION AND EMBRYONIC FORMATION

suggestive of the de novo hematopoietic potential of the placenta. Moreover, between 11.5 and 12.5 dpc, the placental HSC pool expands, and this results in more than 15 times the number of HSCs as compared with the number found in the AGM region(Gekas et al., 2005). This expansion of placental HSCs may result from the continuous generation of HSCs in the placenta or the homing of HSCs to the placenta from other sites. Further studies are required to verify the origin of placental HSCs. E. Colonization of Hematopoietic Organs by Embryonic Precursors After their birth in the AGM and/or the YS, HSCs migrate to the newly forming hematopoietic organs, where they reside in specific locations called stemcell niches, which provide a microenvironment for HSC self-renewal and differentiation. In the stem cell niches, HSCs undergo extensive proliferation and differentiation, and they give rise to all blood lineages to support the growth of the organism. The migration of HSCs from the AGM and/or the YS to the hematopoietic organs is suggested by the quantitative temporal measurement of HSC activity during mouse embryogenesis. In mouse embryos, HSC activity in the AGM and YS decrease after 11 dpc, and it becomes undetectable by 13 dpc (Muller et al., 1994; Sanchez et al., 1996). This decrease in HSC activity in the AGM and YS is accompanied by an exponential increase in HSC activity in the FL from 12 to 15 dpc and later the in bone marrow (Morrison et al., 1995). Although the mechanisms of HSC migration and colonization are largely unknown, several recent studies have suggested possible regulatory factors. Mice deficient in b1-integrin (an adhesion molecule) have hematopoietic progenitors that are present in the YS, the AGM, and the bloodstream; however, these progenitors are unable to properly seed the hematopoietic organs, and this results in a complete absence of fetal hematopoiesis (Potocnik et al., 2000). Chemokine–chemokine-receptor interactions have also been implicated in the migration of HSCs. Mice deficient in the chemokine stromal-cell–derived factor 1a (SDF-1a) or its receptor, CXCR4, fail to establish bone marrow hematopoiesis, although the FL hematopoiesis is normal (Zou et al., 1998; Godin et al., 1999; Ara et al., 2003), which suggests a critical role for SDF-1a–CXCR4 interaction in HSC migration to the bone marrow>. Consistent with these results, Wright et al. recently demonstrated that FL HSCs migrate in response to SDF-1a in vitro (Wright et al., 2002). Moreover, the migratory response of FL HSCs is greatly enhanced in the presence of another chemokine, stem cell factor (SCF; Christensen et al., 2004), which has well-established roles in HSC maintenance, survival, and proliferation. The mutation of the SCF-encoding gene or its receptor c-Kit gene leads to profound hematopoietic defects (Russell, 1979). Consistent with a possible function in cell migration, SCF and c-Kit are expressed along the migration pathways of the germ cells, the melanocytes, the central nervous system, and the hematopoietic cells. In mice, the FL predominates hematopoiesis from 12 dpc through birth, giving rise to all blood lineages to support the growing fetus (Delassus and Cumano, 1996; Mebius and Akashi, 2000; Mebius et al., 2001). At the end of fetal life, the spleen becomes a predominant erythropoietic organ, and it aids in the transition from FL to the bone marrow hematopoiesis (Sasaki and Matsumura, 1988; Godin et al., 1999). The bone marrow is the last

GENETIC APPROACHES TO THE STUDY OF THE TRANSCRIPTION FACTORS IN BLOOD DEVELOPMENT

763

hematopoietic organ to develop in the fetus, but it is the primary niche for HSCs in the adult. The bone marrow HSC reservoir is largely established during postnatal life, when liver hematopoiesis ceases. Other vertebrates use different hematopoietic organs during the fetal and adult stages. For example, in Xenopus, the liver is the main hematopoietic organ in both the larval and adult stages (Chen and Turpen, 1995), whereas, in the avian system, the site of hematopoiesis shifts directly from the AGM to the bone marrow (Dieterlen-Lievre and Martin, 1981). In zebrafish, the kidney maintains the larval and adult hematopoiesis (Willett et al., 1999).

II. GENETIC APPROACHES TO THE STUDY OF THE TRANSCRIPTION FACTORS IN BLOOD DEVELOPMENT The process of HSC generation, proliferation, and differentiation involves complex interactions of transcription factors that modulate the genetic switches along a particular developmental pathway. During the past few decades, many transcription factors have been identified as being essential for blood development by targeted gene disruption in the murine system. More recently, zebrafish has become a powerful model organism for the systematic genetic analysis of vertebrate development, particularly with regard to hematopoiesis. Extensive genetic screens in zebrafish have yielded at least 26 complementation groups of blood mutants (Ransom et al., 1996), thus providing a great source for the genetic dissection of vertebrate hematopoietic pathways. A. Targeted Gene Disruption in Mice The ability to disrupt specific genes by homologous recombination in murine ESCs has allowed investigators to address loss-of-function questions in a mammalian system (Capecchi, 1989). Targeted mutations in mice have identified several important transcription factors that act at distinct stages during hematopoiesis. 1. Transcription Factors Acting at the Hematopoietic–Stem-Cell/ Progenitor Level Flk-1 is a tyrosine kinase receptor for the vascular endothelial growth factor (VEGF). It is expressed in the extraembryonic mesoderm that is destined to give rise to both the vascular and hematopoietic components of the YS blood island of the mouse embryo (Yamaguchi et al., 1993). Expression is maintained in the endothelial cells as well as in the primitive hematopoietic progenitors. Mice that are homozygous for the disruption of Flk-1 die between 8 and 9 dpc, because they lack both the vascular and the primitive blood progenitors (Shalaby et al., 1995). Furthermore, Flk-1–null ESCs fail to contribute to either the vessels or the blood cells in chimeric mouse embryos (Shalaby et al., 1995). These findings suggested that Flk-1 plays essential functions in both blood and endothelial development, perhaps at the hemangioblast stage. However, another model suggests that Flk-1 does not play an instructive role in hematopoiesis. Hidaka et al. (1999) found that, by altering the culture condition, Flk-deficient ESCs can give rise to hematopoietic lineages in vitro in differentiated embryonic bodies, thereby suggesting

764

BLOOD INDUCTION AND EMBRYONIC FORMATION

that Flk-1 regulates the migration of hemangioblasts to the proper environment that is permissive to hematopoiesis (Hidaka et al., 1999). Scl/Tal-1, which was originally identified in a chromosomal translocation in T-cell acute lymphoblast leukemia, encodes a basic helix–loop–helix transcription factor (Hershfield et al., 1984; Begley et al., 1991). By heterodimerizing with E2A products (E12 and E47), Scl recognizes the E-box motifs (CANNTG; Hsu et al., 1994; Shivdasani and Orkin, 1996). It can also form complexes with Gata, Ldb-1, and Lmo-2 to regulate the erythroid differentiation (Wadman et al., 1997). Scl expression is detected in hematopoietic, vascular, and neuronal tissues (Green et al., 1992; Drake et al., 1997). Scl null mice die around 8.5 dpc, and they completely lack blood cells; Scl null ESCs fail to contribute to any definitive hematopoietic lineage, thereby demonstrating the essential role of Scl in the formation of HSCs (Shivdasani et al., 1995; Porcher et al., 1996). Although Scl is essential for the genesis of HSCs, the conditional knockout of Scl in adult mice does not disturb the HSC function, which suggests that Scl is not continually required for the identity and function of HSCs (Mikkola et al., 2003). The function of Scl in endothelial lineage has also been suggested. Although Scl-deficient embryos have some vascular endothelial cells, Scl null ESCs fail to contribute to the formation of the vitelline vessels in chimera mouse embryos, which suggests that Scl is required for the late events in endothelial development (Visvader et al., 1998). The Lmo-2 gene is also involved in a chromosomal translocation in T-cell acute lymphoblast leukemia (Rabbitts, 1998). It encodes a Lim-domain protein that acts as a bridge between DNA-binding transcription factors such as Scl and Gata-1 (Wadman et al., 1997). The disruption of Lmo-2 in mice causes lethality at 9 dpc, with a complete lack of YS hematopoiesis (Warren et al., 1994). Lmo-2 null ESCs do not contribute to any adult blood lineage or to the endothelial cells of large vessels (Yamada et al., 1998), thus demonstrating that Lmo-2 has a similar function to Scl in hematopoiesis and angiogenesis. Gata-2 belongs to the Gata family of transcription factors, which contain two homologous zinc-finger domains and which bind to the Gata consensus sequence (T/AGATAA/G; Crispino et al., 1999; Shimizu et al., 2001). It is highly expressed in the extraembryonic mesoderm, the immature blood progenitors, and the HSCs (Minegishi et al., 1999; 2003). Gata-2 null mice have markedly reduced primitive and definitive hematopoiesis, and they die around 10 to 11 dpc (Tsai et al., 1994). Hematopoietic progenitors from Gata-2 null ESCs proliferate poorly in vitro, and they undergo extensive apoptosis (Tsai and Orkin, 1997), which suggests that Gata-2 is essential for the proliferation and survival of early progenitors. Furthermore, studies involving the use of Gata-2 heterozygous mice revealed that the Gata-2þ/ bone marrow has reduced numbers of HSCs and that the cells exhibit a higher frequency of cell death; this suggests that the dose of Gata-2 is important for the maintenance of adult HSC homeostasis (Rodrigues et al., 2005). Core binding factor (CBF) is a heterodimeric transcriptional factor that consists of a DNA-binding subunit Runx-1 (also known as AML1/CBFA2/ PEBP2aB) and a non–DNA-binding subunit CBF-b (Ogawa et al., 1993a; 1993b; Wang et al., 1993). Runx-1 belongs to the Runx family of transcription factors, which bind to DNA through an evolutionarily conserved Runt domain. CBF-b associates with Runx-1 and enhances its DNA-binding affinity. Mice that are deficient in either subunit lack all definitive blood lineages, but primitive hematopoiesis is not affected (Okuda et al., 1996; Sasaki

GENETIC APPROACHES TO THE STUDY OF THE TRANSCRIPTION FACTORS IN BLOOD DEVELOPMENT

765

et al., 1996; Wang et al., 1996a; 1996b). Hematopoietic colony assays have revealed that the AGM cells from Runx-1 / embryos cannot form any blood lineage, which suggests that CBF is required at the level of stem cells (Mukouyama et al., 2000). In support of this, Runx-1 expression is found in the endothelial cells lining the ventral aspect of the dorsal aorta in the AGM region as well as in other intra-aortic sites from which hematopoietic clusters are thought to emerge (see Figure 34.3). In Runx-1–deficient mice, no hematopoietic clusters are generated, which suggests that CBF is required for the “budding” of definitive HSCs from the intra-aortic endothelium (North et al., 1999). In the adult, however, conditional knockout studies in mice have suggested that Runx-1 is not required for the maintenance of HSCs in the adult bone marrow (Ichikawa et al., 2004). c-Myb is the cellular homolog of the v-Myb oncogene. It is highly expressed in immature hematopoietic progenitors, but its expression decreases as they differentiate (Shivdasani and Orkin, 1996). Mice lacking c-Myb have normal primitive hematopoiesis but a marked loss of definitive progenitors in the FL, which results in death at 15 dpc (Mucenski et al., 1991). AGM cells from c-Myb / embryos do not generate hematopoietic cells in vitro (Mukouyama et al., 1999); this is indicative of there being an essential function of c-Myb during early definitive hematopoiesis. 2. Lineage-Specific Transcription Factors Gata-1, which is the founding member of the Gata family of zinc-finger transcription factors, serves as a central regulator for erythroid gene transcription and development. The Gata motif has been found in virtually all characterized erythroid-specific genes (Evans et al., 1988). Disrupting Gata-1 function in mice results in embryonic lethality at 11.5 dpc from fatal anemia caused by a block in erythroid differentiation at the proerythroblast stage accompanied by apoptosis (Pevny et al., 1991; 1995; Fujiwara et al., 1996), thus demonstrating its critical role in erythroid commitment and differentiation. In addition, the selective knockout of Gata-1 expression in megakaryocytes blocks megakaryocyte differentiation (Shivdasani et al., 1997). Conversely, the forced expression of Gata-1 in a myeloid progenitor cell line promotes megakaryocytic and erythroid differentiation (Visvader et al., 1992). Taken together, these studies establish an instructing role of Gata-1 in megakaryo/erythroid lineage specification and differentiation. Fog is a multitype zinc-finger protein that has been identified as a binding partner of Gata-1 in yeast two-hybrid screens. It is coexpressed with Gata-1 during hematopoietic development, and it cooperates in the mediation of erythroid and megakaryocytic differentiation. Mutant Gata-1 that is unable to interact with Fog fails to support erythroid maturation (Crispino et al., 1999), which suggests a crucial role of Gata-1–Fog-1 interaction during erythropoiesis. The targeted disruption of Fog in mice leads to blocked erythropoiesis (Tsang et al., 1998); the result is similar to the phenotype of Gata-1–knockout mice, thus further supporting Fog’s cooperative function with Gata-1. Eklf is an erythroid Kruppel factor that belongs to the Kruppel zinc-finger protein family, which binds a CACC motif that has been found in many erythroid-specific genes, including adult b-globin (Miller and Bieker, 1993; Feng et al., 1994; Crossley et al., 1996). Eklf null mice die at the FL stage as a result of severe anemia and b-globin deficiency (Nuez et al., 1995). Primitive

766

BLOOD INDUCTION AND EMBRYONIC FORMATION

erythropoiesis and embryonic globin expression are normal, and they demonstrate the pivotal role of Eklf in the activation of adult b-globin expression during the late stages of erythropoiesis (Perkins et al., 1995). B. The Zebrafish Bloodless Mutants Over the past 20 years, zebrafish has been proven to be a powerful model for large-scale genetic investigations in vertebrates. The unique advantages of zebrafish, such as external fertilization and embryogenesis, the large brood size, and the relatively short developmental period, greatly facilitate the forward genetic screens. The transparent zebrafish embryo is especially helpful for the detection and analysis of blood mutants. To date, at least 26 complementary groups with hematopoietic defects have been isolated, and they can be categorized into four phenotypic classes: bloodless, hypochromic, decreasing blood, and photosensitive mutants (Ransom et al., 1996). Here, we will focus on the bloodless class, in which no or few blood cells can be detected in the circulation that begins at 24 hpf. 1. spadetail (spt) The Spt mutation is caused by a defect in the tbx16 gene, which encodes a T-box transcription factor (Griffin et al., 1998). In spt embryos, the mesodermal cell migration is affected, and the paraxial–mesodermal cells are mislocated to the tail, which results in aberrant somite patterning (Kimmel et al., 1989). In addition, spt embryos have specific defects in primitive erythropoiesis in the trunk region, but they retain the normal development of macrophage and myeloid lineage in the head (Thompson et al., 1998; Amacher et al., 2002). Overexpression of Scl can rescue blood defects in spt embryos (Dooley et al., 2005), which suggests that spt acts upstream of Scl. Accordingly, the expression of early hematopoietic markers such as Scl, Lmo-2, and Gata2 and of the erythroid marker Gata-1 is absent in the trunk hematopoietic precursors, but it is retained in the developing endothelial cells. Although the mesoderm-patterning defect in spt embryos may lead to the disruption in the blood specification, Rohde et al. (2004) used transplantation assays to find that spt function is required both cell-autonomously for hematopoiesis and non–cell autonomously for creating the proper environment for red cell development. 2. kugelig (kgg) The kgg mutant is characterized by severe anemia, a shortened tail, and reduced yolk tube extension (Hammerschmidt et al., 1996). The defective gene is cdx4, which is a member of the caudal-related homeobox transcription factor family (Davidson et al., 2003). The cdx family in vertebrates has been implicated in the anterior–posterior patterning of the embryonic axis through the regulation of the Hox genes. Consistent with the role in Hox regulation, the expression pattern of at least nine Hox genes is altered in kgg mutants. The number of ICM precursors expressing Scl and Gata-1 is also reduced, which indicates an early defect in hematopoiesis. This is not caused by a general posterior patterning defect resulting from perturbed Hox gene expression, because the adjacent pronephric tissue is normally patterned in kgg mutants. Instead, the Hox genes have been implicated to have an integral role in hematopoiesis, because the overexpression of Hoxb7 and Hoxa9 rescues the

GENETIC APPROACHES TO THE STUDY OF THE TRANSCRIPTION FACTORS IN BLOOD DEVELOPMENT

767

blood defect without correcting the morphologic defects in kgg (Davidson et al., 2003). Recently, using muse ESC culture, Wang et al. (2005) found that ectopic cdx4 expression promotes hematopoietic mesoderm specification, increases hematopoietic progenitor formation, and, together with HoxB4, enhances the multilineage hematopoietic engraftment of lethally irradiated adult mice. Taken together, these studies demonstrate the specific function of the cdx–Hox pathway in vertebrate blood development. 3. cloche (clo) The clo mutants have virtually no blood and vascular cells, and they lack heart endocardium (Stainier et al., 1995). Gene expression analyses of HSC markers and angioblast markers including Scl, Lmo-2, Gata-2, Runx-1, Fli-1, and Flk-1 have revealed a near complete absence of hematopoietic and endothelial lineages in clo embryos (Liao et al., 1998). The overexpression of Scl but not BMP4 can at least partially rescue the clo defect, which suggests that clo acts downstream of BMP4 but upstream of Scl, perhaps at the hemangioblast level. The mutant gene in clo has not yet been cloned as a result of the telomeric location of the gene (Liao et al., 2000). Uncovering the mutation responsible for clo is expected to provide insight into the molecular events that direct the commitment of mesoderm toward the blood and/or endothelial fates. 4. bloodless (bls) The bls mutant is especially interesting in that the blood defect is restricted to primitive hematopoiesis. Mutant embryos are bloodless until 5 days postfertilization (dpf; Liao et al., 2002), after which the blood cells begin to repopulate the animal, thereby allowing the embryos to survive to adulthood. Scl and Gata-1 expression are greatly reduced during primitive hematopoiesis in the ICM region, and rare specified hematopoietic progenitors undergo apoptosis (Liao et al., 2002). Definitive hematopoiesis is delayed in bls mutants. At 36 hpf, c-Myb expression in the AGM is weak as compared with that of the wild-type embryos; however, by 48 hpf, c-Myb expression is recovered in the bls mutants, which suggests the recovery of definitive hematopoiesis (J. Galloway, unpublished result). Similarly, the expression of Rag1, which is a marker of lymphocytes, is absent in the bls embryos at 4.5 dpf but recovered by 7.5 dpf. Although the defect gene is not yet cloned, it has been speculated that the bls gene may encode a secreted signal on the basis of its non– cell-autonomous defect (Liao et al., 2002). 5. vlad tepes (vlt) The vlt mutant is caused by a nonsense mutation in the zebrafish Gata-1 gene that results in a truncated protein that is unable to bind DNA or to mediate Gata-specific transactivation (Lyons et al., 2002). The mutant embryos have a severe reduction in erythroid progenitors, and this results in few or no blood cells at the onset of the circulation. Expression analyses reveal the normal expression of early hematopoietic markers such as Scl and Lmo2 but a great reduction or even the absence of a number of erythroid markers throughout development, thus demonstrating that the fundamental role of Gata-1 in erythropoiesis is conserved in zebrafish. Recently, Galloway et al. (2005) showed that the loss of Gata-1 function transforms the erythroid

768

BLOOD INDUCTION AND EMBRYONIC FORMATION

precursors into myeloid cells, thereby demonstrating that Gata-1 is required to determine erythroid versus myeloid fate during blood development. 6. moonshine (mon) Mon mutants are characterized by severe anemia, increased apoptosis in ICM and the tail fin, and the enhanced proliferation of iridophores (Ransom et al., 1996). The expression of hematopoietic markers such as Scl and Lmo2 and of the erythroid marker Gata-1 is initiated normally, which suggests that the erythroid precursors are formed in mon embryos. However, these precursors undergo apoptosis that is concomitant with the reduction and eventual loss of hematopoietic markers by the 20-somite stage (Ransom et al., 2004). The differentiation of lymphoid and myeloid cells is not affected in mon mutants. Most of mon homozygous animals die between 10 and 14 dpf, although rare mutant animals survive to adulthood. Analyses of hematopoiesis in the kidneys of these rare survivors revealed a severe block in erythroid differentiation at the proerythroblast stage (Ransom et al., 2004). Therefore, the mon gene is required during both primitive and definitive hematopoiesis specifically for erythroid differentiation. The defect gene in mon mutants encodes the zebrafish ortholog of the mammalian transcriptional intermediary factor 1g (TIF1g), a member of the TIF1 family of transcription cofactors (Ransom et al., 2004). The mechanism of TIF1g function is largely unknown, but recent studies in Xenopus and in human cell culture have suggested an interaction between TIF1g and TGF-b/BMP pathways (Dupont et al., 2005). Consistent with the blood defects in mon mutants, it has been found that the RNAi-mediated knockdown of TIF1g in human hematopoietic progenitor cells inhibits erythroid differentiation in response to the TGF-b signal (He et al., 2006). Whether TIF1g functions in a novel pathway or cooperates with classic erythropoietic factors such as Gata-1 to regulate erythropoiesis is currently under investigation.

III. CLINICAL APPLICATIONS OF HEMATOPOIETIC STEM CELLS Understanding the basic biology of blood development and HSC formation has paved the way for the clinical usage of HSCs. The capacity of HSCs to selfrenew and ultimately give rise to all blood lineages makes them uniquely situated as a powerful tool for the treatment of a variety of blood diseases that are untreatable by traditional approaches. The best example is bone marrow transplantation, which has been used for the treatment of cancer-related hematopoietic deficiency and bone marrow failure states (Antin and Smith, 1995). The transplantation of HSCs from adult bone marrow requires, first of all, the purification of HSCs. This is achieved by sorting HSCs on the basis of their unique surface marker profiles. Using monoclonal antibodies to select bone marrow cells on the basis of surface marker expression, both mouse and human HSCs can be properly isolated. All HSC activity in adult mouse bone marrow has been found in a population marked by the composite phenotype of c-Kitþ, Thy-1.1lo, lineage markers /lo, and Sca-1þ (Spangrude et al., 1988). These cells, when transplanted at the single-cell level, give rise to the longterm reconstitution of hematopoiesis in a lethally irradiated host. In humans, the combination of CD34þ, c-Kitþ, Thy-1lo, and lineage markers /lo cells resulted in the purification of the HSC population, with 85% to 95% purity

769

SUMMARY

(Baum et al., 1992). In human bone marrow, however, HSCs are very rare. Only 0.5% to 5% of bone marrow cells are CD34þ (Civin et al., 1990), and, of these cells, only 10% to 20% express the Thy-1lo, lineage marker– phenotype (Baum et al., 1992). To recover enough HSCs for transplantation, growth factors such as the granulocyte colony-stimulating factor (GCSF) are injected to stimulate the proliferation of HSCs and to mobilize HSCs out of the bone marrow and into the peripheral blood (Murray et al., 1995). The HSC-enriched mononuclear cell fraction can then be collected from the blood and sorted on the basis of the surface markers. Isolated HSCs also provide good targets for gene therapy, which involves the introduction and expression of recombinant genes in somatic cells. For example, genetic hematopoietic disorders that are caused by mutation at a single locus can be treated by introducing a functional copy of the gene into the isolated HSCs using retrovirus-mediated gene transfer (Sutton et al., 1998; Case et al., 1999; Miyoshi et al., 1999) followed by the transplantation of these “corrected” HSCs back to the patient. Such gene therapy approaches have been performed in clinical trials on patients with severe combined immunodeficiency disease, and immune restoration has been observed after the transplantation (Bordignon et al., 1995). Studies in different vertebrate models during the past few decades have significantly contributed to our understanding of the nature of HSCs. We believe that future studies using these model organisms will continually help us to understand the mechanisms by which HSCs differentiate into mature, functional cells. This will ultimately improve the use of HSCs for clinical applications.

SUMMARY In this chapter, we reviewed the development of the vertebrate blood system. Using different model organisms, many factors controlling hematopoiesis have been identified, and a highly conserved genetic program is beginning to emerge.

 Vertebrate hematopoiesis occurs by a multistep process that begins with

   



the induction of ventral mesoderm (see Figure 34.1). The BMP signaling pathway and its antagonists play important roles in patterning the mesoderm toward a ventral and hematopoietic fate. Endothelial and hematopoietic lineages are believed to derive from a common precursor called the hemangioblast. Primitive hematopoiesis occurs at extraembryonic YS blood islands in mammals or their equivalent in other species. Definitive hematopoiesis, which mainly occurs intraembryonically in the AGM region, produces definitive hematopoietic stem cells (HSCs) that ultimately give rise to all of the blood lineages throughout the life span. HSCs subsequently migrate through the circulation to colonize the newly forming hematopoietic organs (e.g., the FL [mouse, human, and Xenopus], the bone marrow [bird, mouse, and human], the kidney [zebrafish]) for further proliferation and differentiation into mature blood cells. Hematopoiesis is highly regulated by complex interactions among growth factors, cytokines, and transcription factors. Gene knockout studies in mice and the genetic mutants generated in zebrafish are powerful tools for identifying the essential genes that control hematopoiesis.

770

BLOOD INDUCTION AND EMBRYONIC FORMATION

ACKNOWLEDGMENTS We gratefully thank Jenna L. Galloway and Teresa V. Bowman for helpful advice and critical reading of the manuscript. L.I.Z. is an investigator of Howard Hughes Medical Institute.

GLOSSARY Definitive hematopoiesis The second wave of hematopoiesis that occurs shortly after the primitive hematopoiesis and that generates all blood lineages. In mammals, definitive hematopoiesis initiates in the aorta–mesonephros–gonads region. Hemangioblast An hypothesized common precursor for both blood cells and blood vessel endothelium cells. Hematopoiesis The developmental process by which various types of blood cells are formed. Hematopoietic stem cells The precursor cells that give rise to all types of blood cells. As stem cells, they are defined by their ability to self-renew and to form multiple cells types. In human adults, these cells are located in the bone marrow. Primitive hematopoiesis A transient wave of hematopoiesis that generates the first blood cells (mainly the red cells) in embryos. In mammals, the site of primitive hematopoiesis is the yolk sac.

REFERENCES Al-Adhami MA, Kunz YW: Ontogenesis of haematopoietic sites in Brachydanio rerio, Dev Growth Differ 19:171–179, 1977. Alvarez-Silva M, Belo-Diabangouaya P, Salaun J, et al: Mouse placenta is a major hematopoietic organ, Development 130:5437–5444, 2003. Amacher SL, Draper BW, Summers BR, et al: The zebrafish T-box genes no tail and spadetail are required for development of trunk and tail mesoderm and medial floor plate, Development 129:3311–3323, 2002. Antin JH, Smith BR: Bone marrow transplantation, In Handin RI, Lux SE, Stossel TP Blood: Principles and practice of hematology, Philadelphia, PA, 1995, J. B. Lippincott Co., DD. 2055–2103. Ara T, Itoi M, Kawabata K, et al: A role of CXC chemokine ligand 12/stromal cell-derived factor1/pre-B cell growth stimulating factor and its receptor CXCR4 in fetal and adult T cell development in vivo, J Immunol 170:4649–4655, 2003. Baum CM, Weissman IL, Tsukamoto AS, et al: Isolation of a candidate human hematopoietic stem-cell population, Proc Natl Acad Sci U S A 89:2804–2808, 1992. Bechtold TE, Smith PB, Turpen JB: Differential stem cell contributions to thymocyte succession during development of Xenopus laevis, J Immunol 148:2975–2982, 1992. Begley CG, Visvader J, Green AR, et al: Molecular cloning and chromosomal localization of the murine homolog of the human helix-loop-helix gene SCL, Proc Natl Acad Sci U S A 88:869–873, 1991. Bordignon C, Notarangelo LD, Nobili N, et al: Gene therapy in peripheral blood lymphocytes and bone marrow for ADA-immunodeficient patients, Science 270:470–475, 1995.

REFERENCES

771 Capecchi MR: Altering the genome by homologous recombination, Science 244:1288–1292, 1989. Caprioli A, Jaffredo T, Gautier R, et al: Blood-borne seeding by hematopoietic and endothelial precursors from the allantois, Proc Natl Acad Sci U S A 95:1641–1646, 1998. Caprioli A, Minko K, Drevon C, et al: Hemangioblast commitment in the avian allantois: cellular and molecular aspects, Dev Biol 238:64–78, 2001. Case SS, Price MA, Jordan CT, et al: Stable transduction of quiescent CD34(þ)CD38( ) human hematopoietic cells by HIV-1-based lentiviral vectors, Proc Natl Acad Sci U S A 96: 2988–2993, 1999. Chaddah MR, Wu DD, Phillips RA: Variable self-renewal of reconstituting stem cells in long-term bone marrow cultures, Exp Hematol 24:497–508, 1996. Chen X, Weisberg E, Fridmacher V, et al: Smad4 and FAST-1 in the assembly of activin-responsive factor, Nature 389:85–89, 1997. Chen XD, Turpen JB: Intraembryonic origin of hepatic hematopoiesis in Xenopus laevis, J Immunol 154:2557–2567, 1995. Choi K, Kennedy M, Kazarov A, et al: A common precursor for hematopoietic and endothelial cells, Development 125:725–732, 1998. Christensen JL, Wright DE, Wagers AJ, et al: Circulation and chemotaxis of fetal hematopoietic stem cells, PLoS Biol 2:E75, 2004. Civin CI, Strauss LC, Fackler MJ, et al: Positive stem cell selection—basic science, Prog Clin Biol Res 333:387–401: discussion 402, 1990. Clement JH, Fettes P, Knochel S, et al: Bone morphogenetic protein 2 in the early development of Xenopus laevis, Mech Dev 52:357–370, 1995. Crispino JD, Lodish MB, MacKay JP, et al: Use of altered specificity mutants to probe a specific protein-protein interaction in differentiation: the GATA-1:FOG complex, Mol Cell 3:219–228, 1999. Crossley M, Whitelaw E, Perkins A, et al: Isolation and characterization of the cDNA encoding BKLF/TEF-2, a major CACCC-box-binding protein in erythroid cells and selected other cells, Mol Cell Biol 16:1695–1705, 1996. Davidson AJ, Ernst P, Wang Y, et al: cdx4 mutants fail to specify blood progenitors and can be rescued by multiple hox genes, Nature 425:300–306, 2003. Delassus S, Cumano A: Circulation of hematopoietic progenitors in the mouse embryo, Immunity 4:97–106, 1996. Dick A, Hild M, Bauer H, et al: Essential role of Bmp7 (snailhouse) and its prodomain in dorsoventral patterning of the zebrafish embryo, Development 127:343–354, 2000. Dieterlen-Lievre F: On the origin of haemopoietic stem cells in the avian embryo: an experimental approach, J Embryol Exp Morphol 33:607–619, 1975. Dieterlen-Lievre F, Martin C: Diffuse intraembryonic hemopoiesis in normal and chimeric avian development, Dev Biol 88:180–191, 1981. Dooley KA, Davidson AJ, Zon LI: Zebrafish scl functions independently in hematopoietic and endothelial development, Dev Biol 277:522–536, 2005. Drake CJ, Brandt SJ, Trusk TC, et al: TAL1/SCL is expressed in endothelial progenitor cells/angioblasts and defines a dorsal-to-ventral gradient of vasculogenesis, Dev Biol 192:17–30, 1997. Dupont S, Zacchigna L, Cordenonsi M, et al: Germ-layer specification and control of cell growth by Ectodermin, a Smad4 ubiquitin ligase, Cell 121:87–99, 2005. Dyer MA, Farrington SM, Mohn D, et al: Indian hedgehog activates hematopoiesis and vasculogenesis and can respecify prospective neurectodermal cell fate in the mouse embryo, Development 128:1717–1730, 2001. Ecochard V, Cayrol C, Rey S, et al: A novel Xenopus mix-like gene milk involved in the control of the endomesodermal fates, Development 125:2577–2585, 1998. Eichmann A, Corbel C, Nataf V, et al: Ligand-dependent development of the endothelial and hemopoietic lineages from embryonic mesodermal cells expressing vascular endothelial growth factor receptor 2, Proc Natl Acad Sci U S A 94:5141–5146, 1997. Evans T: Developmental biology of hematopoiesis, Hematol Oncol Clin North Am 11:1115–1147, 1997. Evans T, Reitman M, Felsenfeld G: An erythrocyte-specific DNA-binding factor recognizes a regulatory sequence common to all chicken globin genes, Proc Natl Acad Sci U S A 85:5976–5980, 1988. Fainsod A, Steinbeisser H, De Robertis EM: On the function of BMP-4 in patterning the marginal zone of the Xenopus embryo, EMBO J 13:5015–5025, 1994.

772

BLOOD INDUCTION AND EMBRYONIC FORMATION

Feng WC, Southwood CM, Bieker JJ: Analyses of beta-thalassemia mutant DNA interactions with erythroid Kruppel-like factor (EKLF), an erythroid cell-specific transcription factor, J Biol Chem 269:1493–1500, 1994. Fleischman RA, Custer RP, Mintz B: Totipotent hematopoietic stem cells: normal self-renewal and differentiation after transplantation between mouse fetuses, Cell 30:351–359, 1982. Fujiwara Y, Browne CP, Cunniff K, et al: Arrested development of embryonic red cell precursors in mouse embryos lacking transcription factor GATA-1, Proc Natl Acad Sci U S A 93: 12355–12358, 1996. Galloway JL, Wingert RA, Thisse C, et al: Loss of gata1 but not gata2 converts erythropoiesis to myelopoiesis in zebrafish embryos, Dev Cell 8:109–116, 2005. Garcia-Porrero JA, Godin IE, Dieterlen-Lievre F: Potential intraembryonic hemogenic sites at preliver stages in the mouse, Anat Embryol (Berl) 192:425–435, 1995. Gekas C, Dieterlen-Lievre F, Orkin SH, et al: The placenta is a niche for hematopoietic stem cells, Dev Cell 8:365–375, 2005. Godin I, Garcia-Porrero JA, Dieterlen-Lievre F, et al: Stem cell emergence and hemopoietic activity are incompatible in mouse intraembryonic sites, J Exp Med 190:43–52, 1999. Graff JM: Embryonic patterning: to BMP or not to BMP, that is the question, Cell 89:171–174, 1997. Graff JM, Thies RS, Song JJ, et al: Studies with a Xenopus BMP receptor suggest that ventral mesoderm-inducing signals override dorsal signals in vivo, Cell 79:169–179, 1994. Green AR, Lints T, Visvader J, et al: SCL is coexpressed with GATA-1 in hemopoietic cells but is also expressed in developing brain, Oncogene 7:653–660, 1992. Griffin KJ, Amacher SL, Kimmel CB, et al: Molecular identification of spadetail: regulation of zebrafish trunk and tail mesoderm formation by T-box genes, Development 125: 3379–3388, 1998. Hammerschmidt M, Pelegri F, Mullins MC, et al: Mutations affecting morphogenesis during gastrulation and tail formation in the zebrafish Danio rerio, Development 123:143–151, 1996. Harrison DE, Astle CM, Lerner C: Number and continuous proliferative pattern of transplanted primitive immunohematopoietic stem cells, Proc Natl Acad Sci U S A 85:822–826, 1988. Henry GL, Melton DA: Mixer, a homeobox gene required for endoderm development, Science 281:91–96, 1998. Hershfield MS, Kurtzberg J, Harden E, et al: Conversion of a stem cell leukemia from a T-lymphoid to a myeloid phenotype induced by the adenosine deaminase inhibitor 2’-deoxycoformycin, Proc Natl Acad Sci U S A 81:253–257, 1984. Hidaka M, Stanford WL, Bernstein A: Conditional requirement for the Flk-1 receptor in the in vitro generation of early hematopoietic cells, Proc Natl Acad Sci U S A 96:7370–7375, 1999. Hild M, Dick A, Rauch GJ, et al: The smad5 mutation somitabun blocks Bmp2b signaling during early dorsoventral patterning of the zebrafish embryo, Development 126:2149–2159, 1999. Ho RK, Kane DA: Cell-autonomous action of zebrafish spt-1 mutation in specific mesodermal precursors, Nature 348:728–730, 1990. Hogan BM, Layton JE, Pyati UJ, et al: Specification of the primitive myeloid precursor pool requires signaling through Alk8 in zebrafish, Curr Biol 16:506–511, 2006. Hsu HL, Wadman I, Baer R: Formation of in vivo complexes between the TAL1 and E2A polypeptides of leukemic T cells, Proc Natl Acad Sci U S A 91:3181–3185, 1994. Huber TL, Kouskoff V, Fehling HJ, et al: Haemangioblast commitment is initiated in the primitive streak of the mouse embryo, Nature 432:625–630, 2004. Ichikawa M, Asai T, Saito T, et al: AML-1 is required for megakaryocytic maturation and lymphocytic differentiation, but not for maintenance of hematopoietic stem cells in adult hematopoiesis, Nat Med 10:299–304, 2004. Iemura S, Yamamoto TS, Takagi C, et al: Direct binding of follistatin to a complex of bone-morphogenetic protein and its receptor inhibits ventral and epidermal cell fates in early Xenopus embryo, Proc Natl Acad Sci U S A 95:9337–9342, 1998. Jaffredo T, Alais S, Bollerot K, et al: Avian HSC emergence, migration, and commitment toward the T cell lineage, FEMS Immunol Med Microbiol 39:205–212, 2003. Jaffredo T, Gautier R, Eichmann A, et al: Intraaortic hemopoietic cells are derived from endothelial cells during ontogeny, Development 125:4575–4583, 1998. Jones CM, Lyons KM, Lapan PM, et al: DVR-4 (bone morphogenetic protein-4) as a posteriorventralizing factor in Xenopus mesoderm induction, Development 115:639–647, 1992. Jordan CT, Lemischka IR: Clonal and systemic analysis of long-term hematopoiesis in the mouse, Genes Dev 4:220–232, 1990.

REFERENCES

773 Kau CL, Turpen JB: Dual contribution of embryonic ventral blood island and dorsal lateral plate mesoderm during ontogeny of hemopoietic cells in Xenopus laevis, J Immunol 131:2262–2266, 1983. Kimmel CB, Kane DA, Walker C, et al: A mutation that changes cell movement and cell fate in the zebrafish embryo, Nature 337:358–362, 1989. Kishimoto Y, Lee KH, Zon L, et al: The molecular nature of zebrafish swirl: BMP2 function is essential during early dorsoventral patterning, Development 124:4457–4466, 1997. Lawson KA, Dunn NR, Roelen BA, et al: Bmp4 is required for the generation of primordial germ cells in the mouse embryo, Genes Dev 13:424–436, 1999. Liao EC, Paw BH, Oates AC, et al: SCL/Tal-1 transcription factor acts downstream of cloche to specify hematopoietic and vascular progenitors in zebrafish, Genes Dev 12:621–626, 1998. Liao EC, Paw BH, Peters LL, et al: Hereditary spherocytosis in zebrafish riesling illustrates evolution of erythroid beta-spectrin structure, and function in red cell morphogenesis and membrane stability, Development 127:5123–5132, 2000. Liao EC, Trede NS, Ransom D, et al: Non-cell autonomous requirement for the bloodless gene in primitive hematopoiesis of zebrafish, Development 129:649–659, 2002. Liao W, Bisgrove BW, Sawyer H, et al: The zebrafish gene cloche acts upstream of a flk-1 homologue to regulate endothelial cell differentiation, Development 124:381–389, 1997. Lyons SE, Lawson ND, Lei L, et al: A nonsense mutation in zebrafish gata1 causes the bloodless phenotype in vlad tepes, Proc Natl Acad Sci U S A 99:5454–5459, 2002. Maeno M, Mead PE, Kelley C, et al: The role of BMP-4 and GATA-2 in the induction and differentiation of hematopoietic mesoderm in Xenopus laevis, Blood 88:1965–1972, 1996. Maeno M, Ong RC, Suzuki A, et al: A truncated bone morphogenetic protein 4 receptor alters the fate of ventral mesoderm to dorsal mesoderm: roles of animal pole tissue in the development of ventral mesoderm, Proc Natl Acad Sci U S A 91:10260–10264, 1994. Maeno M, Tochinai S, Katagiri C: Differential participation of ventral and dorsolateral mesoderms in the hemopoiesis of Xenopus, as revealed in diploid-triploid or interspecific chimeras, Dev Biol 110:503–508, 1985. Mangia F, Procicchiami G, Manelli H: On the development of the blood island in Xenopus laevis embryos: light and electron microscope study, Acta Embryol Exp (Palermo) 2:163–184, 1970. Marshall CJ, Moore RL, Thorogood P, et al: Detailed characterization of the human aorta-gonadmesonephros region reveals morphological polarity resembling a hematopoietic stromal layer, Dev Dyn 215:139–147, 1999. Matsuoka S, Tsuji K, Hisakawa H, et al: Generation of definitive hematopoietic stem cells from murine early yolk sac and paraaortic splanchnopleures by aorta-gonad-mesonephros region-derived stromal cells, Blood 98:6–12, 2001. Mead PE, Brivanlou IH, Kelley CM, et al: BMP-4-responsive regulation of dorsal-ventral patterning by the homeobox protein Mix.1, Nature 382:357–360, 1996. Mead PE, Deconinck AE, Huber TL, et al: Primitive erythropoiesis in the Xenopus embryo: the synergistic role of LMO-2, SCL and GATA-binding proteins, Development 128:2301–2308, 2001. Mead PE, Kelley CM, Hahn PS, et al: SCL specifies hematopoietic mesoderm in Xenopus embryos, Development 125:2611–2620, 1998a. Mead PE, Zhou Y, Lustig KD, et al: Cloning of Mix-related homeodomain proteins using fast retrieval of gel shift activities, (FROGS), a technique for the isolation of DNA-binding proteins, Proc Natl Acad Sci U S A 95:11251–11256, 1998b. Mebius R, Akashi K: Precursors to neonatal lymph nodes: LT betaþCD45þCD4þCD3- cells are found in fetal liver, Curr Top Microbiol Immunol 251:197–201, 2000. Mebius RE, Miyamoto T, Christensen J, et al: The fetal liver counterpart of adult common lymphoid progenitors gives rise to all lymphoid lineages, CD45þCD4þCD3- cells, as well as macrophages, J Immunol 166:6593–6601, 2001. Mikkola HK, Klintman J, Yang H, et al: Haematopoietic stem cells retain long-term repopulating activity and multipotency in the absence of stem-cell leukaemia SCL/tal-1 gene, Nature 421:547–551, 2003. Miller IJ, Bieker JJ: A novel, erythroid cell-specific murine transcription factor that binds to the CACCC element and is related to the Kruppel family of nuclear proteins, Mol Cell Biol 13:2776–2786, 1993. Minegishi N, Ohta J, Yamagiwa H, et al: The mouse GATA-2 gene is expressed in the para-aortic splanchnopleura and aorta-gonads and mesonephros region, Blood 93:4196–4207, 1999.

774

BLOOD INDUCTION AND EMBRYONIC FORMATION

Minegishi N, Suzuki N, Yokomizo T, et al: Expression and domain-specific function of GATA2 during differentiation of the hematopoietic precursor cells in midgestation mouse embryos, Blood 102:896–905, 2003. Miyoshi H, Smith KA, Mosier DE, et al: Transduction of human CD34þ cells that mediate longterm engraftment of NOD/SCID mice by HIV vectors, Science 283:682–686, 1999. Morrison SJ, Hemmati HD, Wandycz AM, et al: The purification and characterization of fetal liver hematopoietic stem cells, Proc Natl Acad Sci U S A 92:10302–10306, 1995. Mucenski ML, McLain K, Kier AB, et al: A functional c-myb gene is required for normal murine fetal hepatic hematopoiesis, Cell 65:677–689, 1991. Mukouyama Y, Chiba N, Hara T, et al: The AML1 transcription factor functions to develop and maintain hematogenic precursor cells in the embryonic aorta-gonad-mesonephros region, Dev Biol 220:27–36, 2000. Mukouyama Y, Chiba N, Mucenski ML, et al: Hematopoietic cells in cultures of the murine embryonic aorta-gonad- mesonephros region are induced by c-Myb, Curr Biol 9:833–836, 1999. Muller AM, Medvinsky A, Strouboulis J, et al: Development of hematopoietic stem cell activity in the mouse embryo, Immunity 1:291–301, 1994. Mullins MC, Hammerschmidt M, Kane DA, et al: Genes establishing dorsoventral pattern formation in the zebrafish embryo: the ventral specifying genes, Development 123:81–93, 1996. Munoz-Sanjuan I, H-Brivanlou A: Early posterior/ventral fate specification in the vertebrate embryo, Dev Biol 237:1–17, 2001. Murray L, Chen B, Galy A, et al: Enrichment of human hematopoietic stem cell activity in the CD34þThy-1þLin- subpopulation from mobilized peripheral blood, Blood 85:368–378, 1995. Ng ES, Azzola L, Sourris K, et al: The primitive streak gene Mixl1 is required for efficient haematopoiesis and BMP4-induced ventral mesoderm patterning in differentiating ES cells, Development 132:873–884, 2005. Nguyen VH, Schmid B, Trout J, et al: Ventral and lateral regions of the zebrafish gastrula, including the neural crest progenitors, are established by a bmp2b/swirl pathway of genes, Dev Biol 199:93–110, 1998. North T, Gu TL, Stacy T, et al: Cbfa2 is required for the formation of intra-aortic hematopoietic clusters, Development 126:2563–2575, 1999. Nuez B, Michalovich D, Bygrave A, et al: Defective haematopoiesis in fetal liver resulting from inactivation of the EKLF gene, Nature 375:316–318, 1995. Ogawa E, Inuzuka M, Maruyama M, et al: Molecular cloning and characterization of PEBP2 beta, the heterodimeric partner of a novel Drosophila runt-related DNA binding protein PEBP2 alpha, Virology 194:314–331, 1993a. Ogawa E, Maruyama M, Kagoshima H, et al: PEBP2/PEA2 represents a family of transcription factors homologous to the products of the Drosophila runt gene and the human AML1 gene, Proc Natl Acad Sci U S A 90:6859–6863, 1993b. Okuda T, van Deursen J, Hiebert SW, et al: AML1, the target of multiple chromosomal translocations in human leukemia, is essential for normal fetal liver hematopoiesis, Cell 84:321–330, 1996. Osawa M, Hanada K, Hamada H, et al: Long-term lymphohematopoietic reconstitution by a single CD34- low/negative hematopoietic stem cell, Science 273:242–245, 1996. Palis J, Chan RJ, Koniski A, et al: Spatial and temporal emergence of high proliferative potential hematopoietic precursors during murine embryogenesis, Proc Natl Acad Sci U S A 98: 4528–4533, 2001. Palis J, Robertson S, Kennedy M, et al: Development of erythroid and myeloid progenitors in the yolk sac and embryo proper of the mouse, Development 126:5073–5084, 1999. Pearce JJ, Evans MJ: Mml, a mouse Mix-like gene expressed in the primitive streak, Mech Dev 87:189–192, 1999. Pelosi E, Valtieri M, Coppola S, et al: Identification of the hemangioblast in postnatal life, Blood 100:3203–3208, 2002. Perkins AC, Sharpe AH, Orkin SH: Lethal beta-thalassaemia in mice lacking the erythroid CACCC-transcription factor EKLF, Nature 375:318–322, 1995. Pevny L, Lin CS, D’Agati V, et al: Development of hematopoietic cells lacking transcription factor GATA-1, Development 121:163–172, 1995. Pevny L, Simon MC, Robertson E, et al: Erythroid differentiation in chimaeric mice blocked by a targeted mutation in the gene for transcription factor GATA-1, Nature 349:257–260, 1991. Piccolo S, Sasai Y, Lu B, et al: Dorsoventral patterning in Xenopus: inhibition of ventral signals by direct binding of chordin to BMP-4, Cell 86:589–598, 1996.

REFERENCES

775 Porcher C, Swat W, Rockwell K, et al: The T cell leukemia oncoprotein SCL/tal-1 is essential for development of all hematopoietic lineages, Cell 86:47–57, 1996. Potocnik AJ, Brakebusch C, Fassler R: Fetal and adult hematopoietic stem cells require beta1 integrin function for colonizing fetal liver, spleen, and bone marrow, Immunity 12:653–663, 2000. Rabbitts TH: LMO T-cell translocation oncogenes typify genes activated by chromosomal translocations that alter transcription and developmental processes, Genes Dev 12:2651–2657, 1998. Ransom DG, Bahary N, Niss K, et al: The zebrafish moonshine gene encodes transcriptional intermediary factor 1gamma, an essential regulator of hematopoiesis, PLoS Biol 2:E237, 2004. Ransom DG, Haffter P, Odenthal J, et al: Characterization of zebrafish mutants with defects in embryonic hematopoiesis, Development 123:311–319, 1996. Robb L, Hartley L, Begley CG, et al: Cloning, expression analysis, and chromosomal localization of murine and human homologues of a Xenopus mix gene, Dev Dyn 219:497–504, 2000. Rodrigues NP, Janzen V, Forkert R, et al: Haploinsufficiency of GATA-2 perturbs adult hematopoietic stem-cell homeostasis, Blood 106:477–484, 2005. Rohde LA, Oates AC, Ho RK: A crucial interaction between embryonic red blood cell progenitors and paraxial mesoderm revealed in spadetail embryos, Dev Cell 7:251–262, 2004. Russell ES: Hereditary anemias of the mouse: a review for geneticists, Adv Genet 20:357–459, 1979. Sanchez MJ, Holmes A, Miles C, et al: Characterization of the first definitive hematopoietic stem cells in the AGM and liver of the mouse embryo, Immunity 5:513–525, 1996. Sasaki K, Matsumura G: Spleen lymphocytes and haemopoiesis in the mouse embryo, J Anat 160:27–37, 1988. Sasaki K, Yagi H, Bronson RT, et al: Absence of fetal liver hematopoiesis in mice deficient in transcriptional coactivator core binding factor beta, Proc Natl Acad Sci U S A 93: 12359–12363, 1996. Schmid B, Furthauer M, Connors SA, et al: Equivalent genetic roles for bmp7/snailhouse and bmp2b/swirl in dorsoventral pattern formation, Development 127:957–967, 2000. Shalaby F, Rossant J, Yamaguchi TP, et al: Failure of blood-island formation and vasculogenesis in Flk-1-deficient mice, Nature 376:62–66, 1995. Shimizu R, Takahashi S, Ohneda K, et al: In vivo requirements for GATA-1 functional domains during primitive and definitive erythropoiesis, EMBO J 20:5250–5260, 2001. Shivdasani RA, Fujiwara Y, McDevitt MA, et al: A lineage-selective knockout establishes the critical role of transcription factor GATA-1 in megakaryocyte growth and platelet development, EMBO J 16:3965–3973, 1997. Shivdasani RA, Mayer EL, Orkin SH: Absence of blood formation in mice lacking the T-cell leukaemia oncoprotein tal-1/SCL, Nature 373:432–434, 1995. Shivdasani RA, Orkin SH: The transcriptional control of hematopoiesis, Blood 87:4025–4039, 1996. Smith PB, Flajnik MF, Turpen JB: Experimental analysis of ventral blood island hematopoiesis in Xenopus embryonic chimeras, Dev Biol 131:302–312, 1989. Spangrude GJ, Heimfeld S, Weissman IL: Purification and characterization of mouse hematopoietic stem cells, Science 241:58–62, 1988. Stainier DY, Weinstein BM, Detrich HW, 3rd: Cloche, an early acting zebrafish gene, is required by both the endothelial and hematopoietic lineages, Development 121:3141–3150, 1995. Steiner R: On the kinetics of erythroid cell differentiation in fetal mice. II. DNA and hemoglobin measurements of individual erythroblasts during gestation, J Cell Physiol 82:219–230, 1973. Sutton RE, Wu HT, Rigg R, et al: Human immunodeficiency virus type 1 vectors efficiently transduce human hematopoietic stem cells, J Virol 72:5781–5788, 1998. Suzuki A, Thies RS, Yamaji N, et al: A truncated bone morphogenetic protein receptor affects dorsal-ventral patterning in the early Xenopus embryo, Proc Natl Acad Sci U S A 91:10255–10259, 1994. Tada M, Casey ES, Fairclough L, et al: Bix1, a direct target of Xenopus T-box genes, causes formation of ventral mesoderm and endoderm, Development 125:3997–4006, 1998. Tavian M, Hallais MF, Peault B: Emergence of intraembryonic hematopoietic precursors in the pre-liver human embryo, Development 126:793–803, 1999. Thompson MA, Ransom DG, Pratt SJ, et al: The cloche and spadetail genes differentially affect hematopoiesis and vasculogenesis, Dev Biol 197:248–269, 1998.

776

BLOOD INDUCTION AND EMBRYONIC FORMATION

Thomsen GH: Antagonism within and around the organizer: BMP inhibitors in vertebrate body patterning, Trends Genet 13:209–211, 1997. Tsai FY, Keller G, Kuo FC, et al: An early haematopoietic defect in mice lacking the transcription factor GATA-2, Nature 371:221–226, 1994. Tsai FY, Orkin SH: Transcription factor GATA-2 is required for proliferation/survival of early hematopoietic cells and mast cell formation, but not for erythroid and myeloid terminal differentiation, Blood 89:3636–3643, 1997. Tsang AP, Fujiwara Y, Hom DB, et al: Failure of megakaryopoiesis and arrested erythropoiesis in mice lacking the GATA-1 transcriptional cofactor FOG, Genes Dev 12:1176–1188, 1998. Turpen JB, Kelley CM, Mead PE, et al: Bipotential primitive-definitive hematopoietic progenitors in the vertebrate embryo, Immunity 7:325–334, 1997. Visvader JE, Elefanty AG, Strasser A, et al: GATA-1 but not SCL induces megakaryocytic differentiation in an early myeloid line, EMBO J 11:4557–4564, 1992. Visvader JE, Fujiwara Y, Orkin SH: Unsuspected role for the T-cell leukemia protein SCL/tal-1 in vascular development, Genes Dev 12:473–479, 1998. Vize PD: DNA sequences mediating the transcriptional response of the Mix.2 homeobox gene to mesoderm induction, Dev Biol 177:226–231, 1996. Wadman IA, Osada H, Grutz GG, et al: The LIM-only protein Lmo2 is a bridging molecule assembling an erythroid, DNA-binding complex which includes the TAL1, E47, GATA-1 and Ldb1/NLI proteins, EMBO J 16:3145–3157, 1997. Wang Q, Stacy T, Binder M, et al: Disruption of the Cbfa2 gene causes necrosis and hemorrhaging in the central nervous system and blocks definitive hematopoiesis, Proc Natl Acad Sci U S A 93:3444–3449, 1996a. Wang Q, Stacy T, Miller JD, et al: The CBFbeta subunit is essential for CBFalpha2 (AML1) function in vivo, Cell 87:697–708, 1996b. Wang S, Krinks M, Kleinwaks L, et al: A novel Xenopus homologue of bone morphogenetic protein-7 (BMP-7), Genes Funct 1:259–271, 1997. Wang S, Wang Q, Crute BE, et al: Cloning and characterization of subunits of the T-cell receptor and murine leukemia virus enhancer core-binding factor, Mol Cell Biol 13:3324–3339, 1993. Wang Y, Yates F, Naveiras O, et al: Embryonic stem cell-derived hematopoietic stem cells, Proc Natl Acad Sci U S A 102:19081–19086, 2005. Warren AJ, Colledge WH, Carlton MB, et al: The oncogenic cysteine-rich LIM domain protein rbtn2 is essential for erythroid development, Cell 78:45–57, 1994. Willett CE, Cortes A, Zuasti A, et al: Early hematopoiesis and developing lymphoid organs in the zebrafish, Dev Dyn 214:323–336, 1999. Willey S, Ayuso-Sacido A, Zhang H, et al: Acceleration of mesoderm development and expansion of hematopoietic progenitors in differentiating ES cells by the mouse Mix-like homeodomain transcription factor, Blood 107:3122–3130, 2006. Winnier G, Blessing M, Labosky PA, et al: Bone morphogenetic protein-4 is required for mesoderm formation and patterning in the mouse, Genes Dev 9:2105–2116, 1995. Wright DE, Bowman EP, Wagers AJ, et al: Hematopoietic stem cells are uniquely selective in their migratory response to chemokines, J Exp Med 195:1145–1154, 2002. Yamada Y, Warren AJ, Dobson C, et al: The T cell leukemia LIM protein Lmo2 is necessary for adult mouse hematopoiesis, Proc Natl Acad Sci U S A 95:3890–3895, 1998. Yamaguchi TP, Dumont DJ, Conlon RA, et al: flk-1, an flt-related receptor tyrosine kinase is an early marker for endothelial cell precursors, Development 118:489–498, 1993. Yoder MC, Hiatt K, Dutt P, et al: Characterization of definitive lymphohematopoietic stem cells in the day 9 murine yolk sac, Immunity 7:335–344, 1997a. Yoder MC, Hiatt K, Mukherjee P: In vivo repopulating hematopoietic stem cells are present in the murine yolk sac at day 9.0 postcoitus, Proc Natl Acad Sci U S A 94:6776–6780, 1997b. Zhang J, Houston DW, King ML, et al: The role of maternal VegT in establishing the primary germ layers in Xenopus embryos, Cell 94:515–524, 1998. Zimmerman LB, De Jesus-Escobar JM, Harland RM: The Spemann organizer signal noggin binds and inactivates bone morphogenetic protein 4, Cell 86:599–606, 1996. Zou YR, Kottmann AH, Kuroda M, et al: Function of the chemokine receptor CXCR4 in haematopoiesis and in cerebellar development, Nature 393:595–599, 1998.

RECOMMENDED RESOURCES

777

FURTHER READING Dale L, Howes G, Price BM, et al: Bone morphogenetic protein 4: a ventralizing factor in early Xenopus development, Development 115:573–585, 1992. Jaffredo T, Nottingham W, Liddiard K, et al: From hemangioblast to hematopoietic stem cell: an endothelial connection? Exp Hematol 33:1029–1040, 2005.

RECOMMENDED RESOURCES Alvarez-Silva M, Belo-Diabangouaya P, Salaun J, et al: Mouse placenta is a major hematopoietic organ, Development 130:5437–5444, 2003. Baum CM, Weissman IL, Tsukamoto AS, et al: Isolation of a candidate human hematopoietic stem-cell population, Proc Natl Acad Sci U S A 89:2804–2808, 1992. Case SS, Price MA, Jordan CT, et al: Stable transduction of quiescent CD34(þ)CD38(-) human hematopoietic cells by HIV-1-based lentiviral vectors, Proc Natl Acad Sci U S A 96:2988–2993, 1999. Choi K, Kennedy M, Kazarov A, et al: Papadimitriou, J.C., and Keller, G, A common precursor for hematopoietic and endothelial cells. Development 125:725–732, 1998. Davidson AJ, Ernst P, Wang Y, et al: cdx4 mutants fail to specify blood progenitors and can be rescued by multiple hox genes, Nature 425:300–306, 2003. Dieterlen-Lievre F, Martin C: Diffuse intraembryonic hemopoiesis in normal and chimeric avian development, Dev Biol 88:180–191, 1981. Eichmann A, Corbel C, Nataf V, et al: Ligand-dependent development of the endothelial and hemopoietic lineages from embryonic mesodermal cells expressing vascular endothelial growth factor receptor 2, Proc Natl Acad Sci U S A 94:5141–5146, 1997. Fleischman RA, Custer RP, Mintz B: Totipotent hematopoietic stem cells: normal self-renewal and differentiation after transplantation between mouse fetuses, Cell 30:351–359, 1982. Graff JM, Thies RS, Song JJ, et al: Studies with a Xenopus BMP receptor suggest that ventral mesoderm-inducing signals override dorsal signals in vivo, Cell 79:169–179, 1994. Huber TL, Kouskoff V, Fehling HJ, et al: Haemangioblast commitment is initiated in the primitive streak of the mouse embryo, Nature 432:625–630, 2004. Jaffredo T, Nottingham W, Liddiard K, et al: From hemangioblast to hematopoietic stem cell: an endothelial connection? Exp Hematol 33:1029–1040, 2005. Liao EC, Paw BH, Oates AC, et al: SCL/Tal-1 transcription factor acts downstream of cloche to specify hematopoietic and vascular progenitors in zebrafish, Genes Dev 12:621–626, 1998. Miyoshi H, Smith KA, Mosier DE, et al: Transduction of human CD34þ cells that mediate longterm engraftment of NOD/SCID mice by HIV vectors, Science 283:682–686, 1999. Murray L, Chen B, Galy A, et al: Enrichment of human hematopoietic stem cell activity in the CD34þThy-1þLin- subpopulation from mobilized peripheral blood, Blood 85:368–378, 1995. Osawa M, Hanada K, Hamada H, et al: Long-term lymphohematopoietic reconstitution by a single CD34- low/negative hematopoietic stem cell, Science 273:242–245, 1996. Porcher C, Swat W, Rockwell K, et al: The T cell leukemia oncoprotein SCL/tal-1 is essential for development of all hematopoietic lineages, Cell 86:47–57, 1996. Ransom DG, Haffter P, Odenthal J, et al: Characterization of zebrafish mutants with defects in embryonic hematopoiesis, Development 123:311–319, 1996. Sanchez MJ, Holmes A, Miles C, et al: Characterization of the first definitive hematopoietic stem cells in the AGM and liver of the mouse embryo, Immunity 5:513–525, 1996. Shivdasani RA, Mayer EL, Orkin SH: Absence of blood formation in mice lacking the T-cell leukaemia oncoprotein tal-1/SCL, Nature 373:432–434, 1995.

35

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE THOMAS M. SCHULTHEISS Beth Israel Deaconess Medical Center and Harvard Medical School, Molecular and Vascular Medicine Unit, Beth Israel Deaconess Medical Center, Boston, MA

INTRODUCTION The kidney is a vital organ, and its main functions include the excretion of metabolic waste products and the maintenance of water balance. The basic functional unit of the vertebrate kidney is the nephron (Figure 35.1), which consists of three main components: a glomerulus, which filters the blood; a tubule, which reabsorbs substances from the glomerular filtrate and into which substances are secreted for excretion; and a nephric duct, which transmits the contents of the tubule to the exterior. Although all vertebrate kidneys are comprised of nephrons, there is great variety in the morphology of the nephron and the arrangement of nephrons into kidneys, both among different species of vertebrates and in different kidney tissues within the same animal. All of these different types of vertebrate kidneys are produced by modifications of the kidney developmental program. Thus, studying kidney development in different vertebrates is important not only for what it can tell us about the kidney itself but also because it provides an excellent window into the question of how the basic building blocks of a tissue can be modified to create a great variety of forms. Kidney development has been the subject of experimental embryologic study for more than 100 years, and, during the past 20 or so years, a large number of genes have been identified that play roles in kidney formation. It is not possible in this chapter to review comprehensively all of these studies; excellent classic and recent reviews and indeed whole books have been written about the topic, and the reader is referred to these for topics that are not included or covered in insufficient depth in this chapter (Dressler, 2002; 2006; Fraser, 1950; Goodrich, 1895; 1930; Lechner and Dressler, 1997;

778

Principles of Developmental Genetics

© 2007, Elsevier Inc. All rights reserved.

VERTEBRATE KIDNEY ANATOMY

779

Schematic of a single nephron, illustrating its major components. Blood is filtered from capillaries at the glomerulus (Gl) across a specialized basement membrane in the kidney capsule (Ca). Upon crossing the capsular membrane, the filtrate is now in the nephrocoele (Nc), which is a subdivision of the body cavity or coelom. In the primitive situation, the nephrocoele is connected to the general coelom via a peritoneal funnel (PF), but this connection is lost in many types of nephrons. From the nephrocoele, the filtrate passes through a nephrostome (Ns), which is typically ciliated, and into the lumen of the kidney tubule (T). In the tubule, substances are absorbed from the lumen and secreted into it by the epithelial tubular lining cells. Finally, the tubular contents drain into the nephric duct (ND), from which they exit to the exterior. Typically the tubule is much longer and more convoluted than illustrated here. (Illustration from Torrey TW: Carnegie Institution of Washington Publication 603, Contributions to Embryology 35:175–197, 1954. Reprinted with permission of the publisher.)

FIGURE 35.1

Romer, 1955; Saxen, 1987; Vize et al., 1997; 2003; Yu et al., 2004). Rather than trying to restate that which has already been ably described, the current review will discuss vertebrate kidney development with a somewhat different emphasis from that of other recent reviews, namely from an evolutionary– developmental perspective. In other words, we will review the field with an eye toward trying to understand the basic developmental mechanisms that are common to the formation of all types of kidneys, as well as the mechanisms that have evolved to modify the structure of nephrons and their organization into kidneys. It is the author’s belief that our understanding of the regulation of kidney development will be aided if we ask the following about each experimental finding: is this telling us something about kidney formation in general or about a specific type of modification in the kidney developmental program that has the effect of producing a specific type of kidney morphology? It will become clear in this chapter that we are only at the very beginning of being able fit what we currently know about the molecular regulation of vertebrate kidney development into an evolutionary–developmental context. It is hoped that by raising questions and pointing out areas in which our knowledge is lacking that this chapter may spur research and thought that broadens our understanding of kidney formation.

I. VERTEBRATE KIDNEY ANATOMY We will first briefly review some of the most important morphologies of vertebrate nephrons and kidneys. The purpose of this section is to lay out

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

780

the variation displayed by vertebrate nephrons and kidneys that our models of kidney development will have to explain. The material is presented at the histologic level and largely without molecular data, because the main concepts and problems were already elucidated using purely morphologic criteria beginning approximately 100 years ago. During the later parts of this chapter, we will revisit many of these forms and concepts and attempt to connect them to more recent molecular and experimental data. A. The Holonephros Current consensus holds that the agnathostome fishes (jawless fishes), including the hagfish and lamprey, constitute a monophyletic outgroup to the remainder of the extant vertebrates (Takezaki et al., 2003), and thus features that are shared by agnathostomes and other vertebrates are good candidates for being conserved from the common ancestor that gave rise to all extant vertebrates. The kidneys of the embryonic hagfish and the apodans (legless amphibians) have a common, simple structure, and they are thus candidates for resembling the ancestral vertebrate kidney. The kidney of the hagfish embryo and the apodans extends throughout much of the length of the organism, and it contains one kidney tubule per segment; these are connected by a nephric duct that drains to the outside at the cloaca (Dean, 1899; Fraser, 1950; Goodrich, 1930; Price, 1897; 1904–1905; Romer, 1955). Because of its relative uniformity along the anterior–posterior axis, this type of kidney has been called a holonephros. The structure of the holonephros suggests that the vertebrate kidney is primitively a segmental organ. Holonephric tubules consist of dorsal outpouchings from the embryonic coelom in a region between the somite and the lateral plate (the nephrocoele; Figures 35.2 and 35.3). At their dorsalmost aspect, each tubule bends posteriorly to join the next most posterior tubule, thereby connecting successive tubules with each other. These connecting regions comprise the nephric duct, which drains the urine produced by the tubules to the outside. In holonephric nephrons, the glomus, which is the vascular component of the kidney, is a segmental branch from the aorta that is associated with the ventral side of the nephric coelom (see Figure 35.2). As in many other anamniotes, the gloma are typically not intimately associated with the kidney tubules, because they are separated from them by the space of the nephrocoele. A connection called the peritoneal funnel also exists between the nephric and the lateral plate coela, thereby potentially allowing fluid in the lateral plate coelom to be taken up by the nephric tubules. Thus, if we consider the holonephros as being representative of the primitive vertebrate kidney, such a kidney would be expected to contain one kidney tubule per segment, with the tubules originating as outpouchings from the dorsal coelomic wall, with a nephric duct connecting the dorsal aspects of the tubules, and with a segmental vascular component associated with the ventral coelomic wall. It should be noted that our knowledge of hagfish embryonic kidney anatomy is based on just a few series of histologic sections reported during the late nineteenth and early twentieth centuries, because hagfish embryos, which develop in the deep ocean, are extraordinarily hard to come by (Ota and Kuratani, 2006). There is no reason to think that the anatomic descriptions are inaccurate. However, because of the important position of the

VERTEBRATE KIDNEY ANATOMY

781

FIGURE 35.2 Section through an apodan (Hypogeophis) embryo, illustrating features of the primitive nephron. Note that the glomerulus (gl), a branch from the aorta (A), is separated from the entrance to the kidney tubule (the nephrostome, ns) by the relatively large space of the nephrocoele (nc). my, Myotome; N, notochord; pc, peritoneal canal; sc, spinal cord; scl, sclerotome; splc, splanchnocoele. (Reproduced from Goodrich ES: Studies on the structure and development of vertebrates, London, 1930, Macmillan.)

FIGURE 35.3

Schematic illustrating the process of tubule and duct formation in an animal with one tubule per embryonic segment. Tubules (t) bud out dorsally from the nephrocoele (nc) and bend posteriorly, joining up with tubules from more posterior segments to form the nephric duct (ld). In the most primitive case (the holonephros), each segment of duct is formed by the posterior extension of the immediately anterior tubule. The current figure illustrates a more derived situation in which the duct is formed from a combination of contributions from posteriorly extending tubules as well as from the posterior extension of the duct itself. lp, Lateral plate; m, myotome. (Modified from Goodrich ES: Studies on the structure and development of vertebrates, London, 1930, Macmillan.)

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

782

hagfish in vertebrate phylogeny, it would be of great interest to have access to hagfish embryos and to be able to examine them using modern molecular and microscopic methods. B. The Pronephros The pronephros derives its name from its position in the anterior part of the organism and because of the fact that it develops first during embryonic development (Figure 35.4; Romer, 1955; Vize et al., 1997). It serves as the embryonic kidney of many fish and amphibians. As compared with the holonephros, the pronephros can be thought of as a differentiation of only the anteriormost part of the kidney developmental field. Although the anamniote pronephros is often small (consisting of only one pair of tubules in zebrafish and three pairs in Xenopus), it is functional, and zebrafish mutants that lack a pronephros die as embryos (Drummond et al., 1998). Birds and mammals also form a pronephros, but it has limited or no functionality, and it typically begins degenerating shortly after its formation (Fraser, 1950; Romer, 1955). As will be discussed later, the avian and mammalian pronephros appears to be evolutionarily conserved because pronephros formation is fundamentally linked to the formation of the nephric duct, which is essential for the development of the mesonephros and the metanephros. Thus, the avian/mammalian pronephros is a largely atavistic structure that is preserved because it serves an essential embryologic function.

FIGURE 35.4

Illustration of types of kidney tissue at different stages of a developing mammalian embryo. A, Pronephros, with a few tubules draining into a nephric duct (here called the archinephric duct). B, The degeneration of much of the pronephros and the formation of a large mesonephros that extends over much of the trunk. C, The formation of the metanephros posterior to the mesonephros. D, The further development of the metanephros and the degeneration of the mesonephros, except for where it becomes associated with the testis. Note the single attachment of the metanephros to the nephric duct, which is in contrast with the multiple connections of the mesonephros. (Illustration from Romer AS: The vertebrate body, Philadelphia, 1955, WB Saunders.)

EARLY KIDNEY DEVELOPMENT AND SPECIFICATION OF THE INTERMEDIATE MESODERM

783

Because pronephric tubules form only in the anterior end of the embryo, the nephric duct in animals with a pronephros is typically quite long, because it extends from the pronephros to the cloaca. Recent investigations have found that some molecular markers that are found in the nephric duct of other vertebrate kidneys are found only in the distalmost part of the zebrafish nephric duct (Van Campenhout et al., 2006). Thus, the functional “tubule” of at least some embryos with a functional pronephros may extend to more posterior regions of the embryo than had been previously thought, and the true “nephric duct” of such embryos may only comprise the more posterior regions of the straight portion of the urinary system. The relation of the glomus to the tubule is variable in the pronephros. For example, in Xenopus, the glomus is separated from the entrance to the tubules by the space of the nephrocoele (Vize et al., 1997), as in the holonephric nephron (see Figure 35.2), whereas zebrafish pronephroi contain true glomeruli in which the vascular and tubule elements of the kidney are intimately associated (Drummond and Majumdar, 2003). C. The Mesonephros and Metanephros The mesonephros is the adult kidney of fish and amphibians and the functional embryonic kidney of birds and mammals. It lies posterior to the pronephros, and it usually extends over a larger portion of the trunk (see Figure 35.4). It typically contains multiple nephrons per body segment, with multiple connections of the these nephrons to the nephric duct (Cebrian et al., 2004). Mesonephroi typically contain true glomeruli, with an intimate association between the vascular and tubular components (Sainio, 2003). Metanephroi are formed only in amniotes, where they serve as the adult kidney. The metanephros originates at the posterior end of the kidney morphogenetic field, although it can grow into more anterior locations, such as is seen in birds. In contrast with the mesonephros, the metanephros has only one or a few connections to the nephric duct (see Figure 35.4), which suggests that it may be derived from only one or a few embryonic segments. Despite its apparent origin from a small primordium, the metanephros can grow quite large, as would be expected for an organ that processes the metabolic and water balance needs of large adult animals. To achieve a large size despite having limited connections to the duct, the nephrons of the metanephros are highly branched, and they contain typical glomeruli (Fraser, 1950; Romer, 1955). Having briefly reviewed the anatomy of the main vertebrate kidney forms, we will now turn to a discussion of the formation of these vertebrate kidney structures. We will focus on three areas that have received a significant amount of experimental attention: (1) the specification and early development of the kidney primordium; (2) the development of the nephric duct; and (3) the formation of the mammalian metanephros. Other topics will be discussed in a comparative manner as part of the treatment of the three main topics.

II. EARLY KIDNEY DEVELOPMENT AND SPECIFICATION OF THE INTERMEDIATE MESODERM In the basic vertebrate body plan, the structure of the holonephros suggests that the region of mesoderm lateral to the somites is competent to form

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

784

nephric tissue and that this nephrogenic activity is uniform throughout a large segment of the anterior–posterior body axis. We will consider here two aspects of the establishment of this nephrogenic competence. First, we will discuss how the nephrogenic field is established along the medial–lateral (ML) (also referred to as the dorsal–ventral [DV]) axis of the mesoderm (i. e., why does kidney-forming tissue develop immediately lateral [or ventral] to the somites?). Because the location of the kidney-forming tissue along the ML axis is conserved throughout the vertebrates (an exception is in the zebrafish pronephros, which is certainly a specialized adaptation; this is discussed later), many of the factors that regulate the ML aspects of early kidney patterning are likely to be conserved among vertebrates. Second, we will consider how the nephrogenic field is established along the anterior–posterior (AP) axis. As discussed previously, kidney morphology along the AP axis is highly variable among different vertebrates and within individual animals, and alterations in the AP dimension of kidney formation are likely to play a role in the production of different types of kidneys. A. The Medial–Lateral Axis of Early Kidney formation The source of all kidney tissue is the intermediate mesoderm (IM), a region of mesoderm that lies between the somites and the lateral plate in the developing embryo (Figure 35.5). In the chicken embryo (as in mice and zebrafish), the earliest molecular marker of the IM is the putative transcription factor Odd1, the expression of which is initiated before the formation of a morphologically distinct IM (Figure 35.6; James et al., 2006; So and Danielian, 1999). Strictly speaking, Odd1 is not a specific IM marker, because its expression extends into the medial part of the lateral plate. In chicken and mouse embryos, Odd1 is expressed only in undifferentiated kidney precursor tissues; upon the differentiation of kidney tubules or duct, Odd1 is downregulated (James

Diagram of the intermediate mesoderm (IM) in the chicken embryo. A, The IM is a strip of mesoderm that is located lateral to the somites. B, C, and D, The nephric duct rudiment forms when a portion of the IM bulges dorsally and separates from the IM. Subsequently, the nephric duct rudiment extends posteriorly and becomes epithelialized to form the nephric duct. Tubules subsequently form from the nephrogenic cord (c). coe, Coelom; d, nephric duct; im, intermediate mesoderm; lp, lateral plate; n, notochord; np, neural plate; nt, neural tube; psm, presomitic mesoderm; som, somite. (Adapted from James RG, Schultheiss TM: Patterning of the avian intermediate mesoderm by lateral plate and axial tissues, Dev Biol 253:109–124, 2003.)

FIGURE 35.5

EARLY KIDNEY DEVELOPMENT AND SPECIFICATION OF THE INTERMEDIATE MESODERM

785

FIGURE 35.6 Early gene expression in the chicken embryo intermediate mesoderm (IM). A, Odd1 is expressed shortly after gastrulation in the nascent IM (arrow). B, By Hamburger and Hamilton stage 9, Odd1 is expressed robustly in the IM and medial lateral plate. Its expression extends anteriorly until the axial level of the first somite. C and D, A second wave of IM gene expression begins at approximately Hamburger and Hamilton stage 9. At this stage, Pax2 is just beginning to be expressed. Pax2 expression is limited to the most medial aspect of the Odd1 expression domain, and it extends anteriorly only as far as the sixth somite axial level. (Adapted from James RG, Kamei CN, Wang Q, et al: Odd-skipped related 1 is required for development of the metanephric kidney and regulates formation and differentiation of kidney precursor cells, Development 133:2995–3004, 2006; see color insert.)

et al., 2006). This is in contrast with other markers described later, which are maintained in differentiated kidney tissues. A second phase in IM gene expression begins in chicken embryos that are at Hamburger and Hamilton (HH) stage 8 (Hamburger and Hamilton, 1951), which occurs approximately 6 hours after the initiation of Odd1 expression, when the expression of the transcription factors Pax2 and Lim1 is detected in the most medial region (i.e., adjacent to the somites) of the Odd1 expression domain (see Figure 35.6; James and Schultheiss, 2003; Mauch et al., 2000). Unlike Odd1, Pax2 expression is confined to the IM and future kidney tissues, including both duct and tubular tissue, whereas Lim1 is initially expressed solely in the forming nephric duct (Schultheiss et al., 2003). At HH stage 10, Wt1 expression is first detectable in the IM and the most medial part of the lateral plate (James and Schultheiss, 2003). Zebrafish, Xenopus, and mouse also express Pax2 (or its paralog Pax8), Lim1, and Wt1 during early pronephros formation, which indicates that these genes are part of a shared vertebrate early kidney developmental program (Carroll et al., 1999; Dressler et al., 1990; Fujii et al., 1994; Kreidberg et al., 1993; Majumdar et al., 2000; Serluca and Fishman, 2001). In all of these species, Pax2 and Pax8 are expressed in the kidney tubules and duct, whereas Wt1 expression comes to be associated primarily with the forming glomerulus. In summary, IM and early kidney gene expression can be divided crudely into two stages. Phase 1 consists of Odd1 expression in the IM and the medial part of the lateral plate, whereas phase 2 consists of the activation of kidney-specific genes in the more medial sector of the Odd1-expressing domain. The avian embryo has been used by a number of groups to gain insight into the timing of developmental commitment to a kidney fate along the ML axis and to determine the tissues and molecules that regulate such commitment. Transplantation experiments using chick–quail chimeras have determined that, when prospective pronephros cells reside in the primitive streak

786

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

(HH stages 4, 5, and 6), they are not committed to a kidney fate, because they will change fates if they are transplanted into prospective somite or lateral plate regions (James and Schultheiss, 2003; Psychoyos and Stern, 1996). Similarly, prospective somite and lateral plate regions of the streak will adopt a kidney fate if they are transplanted into the IM region of the primitive streak. During the approximately 6 hours after exit from the primitive streak, prospective pronephros cells become progressively resistant to respecification upon transplantation into non-IM regions of the embryo. Thus, by HH stages 8 and 9 (concomitant with the initiation of Pax2 and Lim1 expression in the IM), the prospective IM maintains Pax2 and Lim1 if it is transplanted into either the somite or lateral plate regions (James and Schultheiss, 2003). The tissues and molecular factors that regulate the ML aspects of IM formation are just beginning to be uncovered. In the avian embryo, it seems clear at this point that signals from tissues that are both lateral and medial to the IM are important for specifying the IM. The first evidence that lateral signals could induce IM gene expression came from experiments in which lateral plate and somite tissue were combined in tissue culture. It was found that Pax2 was induced in the somite tissue, thereby indicating that a factor in the lateral plate could induce early kidney gene expression in paraxial mesoderm (James and Schultheiss, 2003). Subsequent experiments have found that bone morphogenetic protein (BMP) signaling is an important component of the lateral plate activity (James and Schultheiss, 2005). Purified BMP-2 or BMP-4 or constitutively active BMP receptors activate IM gene expression in the paraxial mesoderm both in vivo and in vitro, and the repression of BMP signaling using the BMP antagonist noggin represses IM gene expression in the IM. Interestingly, IM genes are activated only by specific levels of BMP signaling. At high levels of BMP signaling, lateral plate but not IM genes are activated, whereas lower levels of BMP signaling activate IM genes. This dose-sensitive effect of BMP signaling is cell autonomous (James and Schultheiss, 2005). Thus, one factor that regulates IM patterning is BMP signaling, with specific levels of BMP signaling being required to induce the expression of IM genes. One source of BMP signaling in the embryo is likely to be the lateral plate mesoderm itself, which expresses BMP-4, whereas levels of BMP signaling may be modulated by BMP antagonists expressed in midline and paraxial tissues (James and Schultheiss, 2003; Schultheiss et al., 1997). BMP signaling has also been implicated in the regulation of early kidney gene expression in zebrafish (Melby et al., 2000). In Xenopus, the ectopic expression of a combination of Lim1 and Pax8 leads to the production of ectopic pronephric tubules (Carroll and Vize, 1999). Interestingly, these ectopic tubules appear to be generated only in the region of the prospective somites, thereby suggesting that the paraxial mesoderm is more permissive for kidney formation than the more ventral regions of the embryo. This is consistent with the data from chickens, and it suggests that conditions in the lateral plate (with its high levels of BMP signaling) may be nonpermissive for kidney development (Figure 35.7; James and Schultheiss, 2005). However, BMP signaling is almost certainly not the only important signal that regulates IM gene expression. Several laboratories have found evidence for a midline or paraxial signal that promotes IM gene expression in both avian and amphibian embryos. The blockage of communication between the dorsal and intermediate regions of the mesoderm prevents the activation of IM genes in chicken embryos (Barak et al., 2005; Mauch et al., 2000), and

EARLY KIDNEY DEVELOPMENT AND SPECIFICATION OF THE INTERMEDIATE MESODERM

787

FIGURE 35.7 Model of factors regulating intermediate mesoderm (IM) formation. In the somite, IM genes are repressed by somite transcription factors such as FoxC1 and FoxC2. Bone morphogenetic protein signaling in the IM region represses somite transcription factor gene expression and derepresses the expression of IM genes such as Odd1. An unknown factor or factors that likely originate from axial or paraxial tissues is required for the robust expression of other IM genes such as Pax2 and Lim1. In the lateral plate, high levels of bone morphogenetic protein signaling repress IM gene expression through unknown molecular mechanisms. (Modified from James RG, Schultheiss TM: Bmp signaling promotes intermediate mesoderm gene expression in a dose-dependent, cell-autonomous and translation-dependent manner, Dev Biol 288:113–125, 2005).

dorsal structures have also been found to promote kidney tubule formation in Xenopus (Seufert et al., 1999). Such a dorsal signal is unlikely to be simply a BMP antagonist, because transplants of Hensen’s node (the most dorsal embryonic structure) into the lateral plate can induce IM gene expression at a stage of embryonic development at which implants of noggin-expressing cells do not have this effect (James and Schultheiss, 2003, 2005). In addition, in vivo and in vitro studies have found that BMP signaling is more efficient for the induction of the expression of Odd1 than it is of more specific kidney markers such as Pax2 and Lim1 (James and Schultheiss, 2005). Because Pax2 and Lim1 expression is activated after Odd1 and only in the region immediately adjacent to the somites, it is likely that an additional signal emanating from midline or paraxial tissues is required for the expression of the full set of IM genes. The molecular nature of this dorsal signal (or signals) is not currently known, but it is under active investigation. In vitro experiments with Xenopus animal caps have found that a combination of activin and retinoic acid (RA) signaling promotes the generation of kidney tubules (Moriya et al., 1993). Although both activin-like (Schier, 2003; Sive, 1993) and RA (Swindell et al., 1999) signals are associated with dorsal embryonic structures, it is not clear how the in vitro findings relate to the requirement for dorsal signaling or to the generation of kidney tubules in vivo. One of the effects of activin in the in vitro animal cap experiments is likely to be mesoderm induction, but that does not rule out more kidney-specific roles for activin-like signaling during later stages of development. It is of interest that activin and RA have

788

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

also been found to promote kidney gene expression in mouse embryonic stem cells (Kim and Dressler, 2005). One important area of current research is attempting to make connections between the signals that induce IM formation and the activation of IM-specific genes. Misexpression in the IM of the forkhead family members FoxC1 and FoxC2, which are normally expressed in the somites, leads to the repression of IM gene expression on a cell-autonomous basis, whereas complementary experiments have found that, in mouse FoxC1 and FoxC2 knockouts, IM gene expression is expanded into the somite region (Wilm et al., 2004). Taken together, the FoxC1, FoxC2, and BMP data have been combined into a working model in which IM gene expression is under negative regulation by somite transcription factors and in which BMP signaling (at the appropriate dose) represses somite gene expression in the IM and thereby derepresses IM gene expression (see Figure 35.7; James and Schultheiss, 2005). One major aim for future research will be to identify other signals, including those from dorsal sources, that induce a subportion of the Odd1-expressing region to express kidney-specific markers, such as Pax2, Pax8, Lim1, and Wt1. B. The Anterior–Posterior Axis of Early Kidney formation Many aspects of kidney morphology vary along the AP axis, including the following: (a) There is an anterior limit to the portion of the IM that goes on to form kidney tissue. In the chicken embryo, this limit lies at the axial level of somite 6 (Barak et al., 2005), whereas, in Xenopus, it lies at approximately somite 3 (Vize et al., 1997). (b) The pronephros, mesonephros, and metanephros are located at different locations along the AP axis, and they have very different morphologies. (c) In many vertebrates, only anterior IM is capable of giving rise to nephric duct tissue (this topic is discussed further in the next section). It is likely that interactions of the basic kidney developmental program with factors that pattern the AP axis underlie the different kidney developmental fates along the AP axis. Much less is known about the AP dimension of kidney formation than is known about the ML dimension. Hox genes are good candidates for regulating at least some of these aspects of kidney AP patterning. Members of the Hox11 paralogous family are expressed in the metanephros but not in more anterior kidney tissue, and the loss of Hox11 function results in the absence of the metanephros (Patterson et al., 2001; Wellik et al., 2002). The anterior border of the pronephros and the posterior border of the duct-forming region correlate with the expression boundaries of particular Hox genes (H. Barak, R. James, R. Reshef, and T. M. Schultheiss, unpublished data), but the functional significance of this correlation has not yet been established. Experiments in the chick embryo have investigated the time at which the AP pattern in the IM is established and the factors that regulate such patterning (Barak et al., 2005). Normally, IM adjacent to somites 1 though 5 in the chick embryo expresses Odd1, but it will never express specific kidney markers such as Pax2 or Lim1. While it is still residing in the primitive streak,prospective IM that is fated to lie adjacent to somites 1 through 5 can be

FORMATION OF THE NEPHRIC DUCT

789

induced to express kidney markers if it is transplanted into an older primitive streak that gives rise to the IM adjacent to somites 6 and those posterior to it. As with ML patterning, prospective anterior and posterior IM tissues become fixed to their respective identities during the course of their migration into the IM region. Interestingly, if prospective anterior IM residing in the primitive streak is transplanted directly into mature IM adjacent to somites 1 through 5 (thereby eliminating its normal migration path), the transplanted tissue goes on to express kidney markers ectopically. These data suggest that kidneyinducing signals are present throughout the AP axis of the embryos (and not just at the axial levels at which the kidneys form). The process of the migration of cells from the primitive streak into the IM of axial levels 1 through 5 appears to expose cells to factors that inhibit their ability to respond to kidney-inducing signals, thus restricting kidney gene expression to the axial levels of somite 6 and of those posterior to it. Before leaving the topic of early kidney patterning, it is worth noting that the relationship between the ML and AP axes of the early embryonic kidney is somewhat different in chicken and mouse embryos as compared with, for example, zebrafish embryos. In birds and mammals, the tubule, glomerulus, and duct primordia of each nephron are located at the same axial levels but at different regions of the ML axis. This is likely to be a conserved feature from the primitive vertebrate state, because it is shared with the holonephros (see Figure 35.2). In the zebrafish pronephros, on the other hand, the glomus primordia (as marked by Wt1 expression) is located more anteriorly than other regions of the pronephros (Serluca and Fishman, 2001). The Pax2 domain partly overlaps the Wt1 domain and extends more posteriorly, and it is later associated with the tubule and anterior duct, whereas a more posterior Sim1 domain is later associated with the duct. Thus, in zebrafish, cell fates within the nephron are deployed along the AP axis, and this is unlike the situation in the bird and mammal pronephros. These differences may have their origin in the modification in the fish pronephros of an ancestral kidney differentiation pattern as compared with its higher degree of conservation in birds and mammals. Regardless of how the difference evolved, it is important to be aware of these different patterns when comparing developmental mechanisms in the different species. In particular, factors that regulate glomus as compared with tubule formation might be expected to be associated with AP positioning in the zebrafish but with DV positioning in amniote embryos.

III. FORMATION OF THE NEPHRIC DUCT The development of the nephric duct is an important topic for a number of reasons. It is one of three main structural units of all kidney types (the others being the tubules and the glomus/glomerulus), and studies of nephric duct formation can yield insights into how the different components of the kidney are differentiated with respect to each other. From an evolutionary perspective, the formation of the nephric duct is quite variable across vertebrate groups, and its study offers the opportunity for insights into the developmental mechanisms that account for such variability. In many species, the nephric duct undergoes a remarkable migratory process that is still poorly understood at the molecular level. Finally, as will be discussed in the section about the

790

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

metanephros, the nephric duct plays an essential role in the induction of the metanephric (as well as the mesonephric) tubules. To gain some evolutionary context, we begin our discussion of the nephric duct by again considering the development of the holonephros. It must be stressed again that our knowledge of holonephros nephric duct formation is derived from serial histologic sections of a very limited number of specimens; thus, it is urgently in need of a revisit. That being said, classic descriptions of holonephros nephric duct formation hold that the dorsalmost part of the kidney tubule in each body segment bends posteriorly and merges with the tubule of the next most posterior segment (see Figure 35.3; Fraser, 1950; Price, 1897; 1904–1905). These connections between tubules collectively make up the nephric duct. Thus, each segment of the embryo that contains a kidney tubule also generates a segment of the nephric duct. The formation of the most posterior portion of the duct, from the most posterior tubule to the opening through which the duct drains to the outside, has not been described. When we move to the pronephros, a problem arises: if duct segments form only where there are tubules (as in the holonephros), how can a duct be generated that drains urine from the pronephros at the anterior end of the embryo to the exit at the posterior end? It appears that at least two developmental mechanisms have evolved to solve this problem. The most highly studied mechanism is seen in amphibian embryos as well as in avian and mammalian embryos. In these species, a nephric duct rudiment forms in the vicinity of the pronephros and proceeds to extend posteriorly to where it drains into the cloaca or bladder (Schultheiss et al., 2003). In the chicken, the duct rudiment forms as a dorsal outcropping from the intermediate mesoderm; this is reminiscent of the dorsal extensions that give rise to the holonephros nephric duct (see Figure 35.5; James and Schultheiss, 2003). The fact of duct migration has been convincingly demonstrated in both amphibian and avian embryos by fate mapping and microsurgical techniques (Obara-Ishihara et al., 1999; Poole and Steinberg, 1981; Schultheiss et al., 2003). There is some evidence that an alternative mechanism may contribute to the process of nephric duct formation in zebrafish and Xenopus embryos. In Xenopus, it has been reported that cells in the posterior trunk of the embryo can participate in duct formation (Cornish and Etkin, 1993), although in this setting there is also a contribution from the anterior migrating duct rudiment. This is unlike the situation in the chick, in which all duct tissue appears to originate in the anterior duct rudiment (Obara-Ishihara et al., 1999; Schultheiss et al., 2003). In one study in zebrafish, the labeling of cells in the flank of the embryo did not uncover any evidence for the migration of the duct (Serluca and Fishman, 2001). One caveat to this experiment is that it does not rule out duct migration at an earlier stage of development. In any case, these studies in Xenopus and zebrafish should be followed up, because they indicate that, in at least some species, it may be possible to generate duct tissue without any connection to tubular tissue. Despite many decades of study, the molecular mechanisms that regulate duct rudiment formation and extension are still obscure (Schultheiss et al., 2003). Studies in Xenopus have implicated Notch signaling in the allocation of the pronephric rudiment into duct and tubule primordia (McLaughlin et al., 2000). In the chicken, BMP signaling is required for the epithelialization of the duct rudiment but not for the posterior migration of the duct primordia (Obara-Ishihara et al., 1999).

FORMATION OF THE MAMMALIAN METANEPHROS

791

Despite the differences between holonephric and pronephric nephric duct formation, there may be some fundamental similarities. In both situations, there appears to be a connection between duct and tubule formation, such that the duct forms as a dorsal extension from a region that is forming (e.g., the holonephros) or that will form (e.g., the pronephros) kidney tubules. The difference between the two situations is that, in the pronephros, the connection between duct and tubule formation is limited to the pronephric region, whereas, in the holonephros, the two tissue types are linked throughout the length of the body axis. It is of interest that, in birds and mammals, in which the pronephros is essentially nonfunctional, the nephric duct forms in association with rudimentary pronephric tubules. Thus, there may be an essential link between pronephric tubule and nephric duct formation, with the nonfunctional pronephric tubules being a byproduct of the need to generate a nephric duct. It is not clear whether the migration of the nephric duct rudiment is an evolutionary innovation or rather if it is an elaboration of the basic tendency of the dorsalmost regions of the kidney tubules to extend posteriorly, as is seen in the holonephros. In amniote embryos, the IM of the mesonephric and metanephric regions does not appear to have any ability to generate nephric duct tissue. Thus, blocking the migration of the duct rudiment in chicken embryos leads to the absence of a duct posterior to the blockade (Gruenwald, 1937; Waddington, 1938). One caveat to interpreting this finding is that the nephric duct is itself required for the induction of mesonephric and metanephric tubule formation (as is discussed later). Thus, if tubule formation is required for duct formation, then the essential precursors for duct formation would not be generated in embryos in which duct migration has been blocked. The main contribution of studies in the mouse embryo to the understanding of nephric duct formation has come from genetics. Abnormalities in the formation of the nephric duct are seen in many mouse mutants, including Odd1, Pax2, Pax8, Lim1, and Gata3 (Bouchard et al., 2002; Grote et al., 2006; James et al., 2006; Shawlot and Behringer, 1995). However, because of the mutual interactions between duct and tubulogenic mesenchyme (discussed in the next section) and because many of these genes are expressed in both the duct and the prospective tubules, it is usually unclear whether these genes are required autonomously in the duct. One exception is Gata3, which is not expressed outside of the duct during early kidney formation (Grote et al., 2006); thus, it appears to play a specific role in duct formation. Gata3 mutants do form a duct, but it exhibits abnormalities in posterior extension and morphogenesis. The absence of a significant number of genes that are required specifically for duct formation and that are not also required for other aspects of kidney formation should maybe not be surprising given the common origin in the IM of both the duct and the tubules and the intimate relationship between duct and tubule formation throughout kidney development.

IV. FORMATION OF THE MAMMALIAN METANEPHROS We will now turn to the metanephros, which is the most-studied component from a molecular standpoint, not only of the mammalian nephric system, but of vertebrate nephrogenesis in general. Studies of metanephros development

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

792

are the main source of knowledge regarding the genetics of vertebrate kidney development. However, again we must bear in mind that the metanephros is a specialized structure, and thus not all of the developmental pathways that regulate metanephros formation will be found to be involved in the formation of other types of kidneys. During this discussion, we will introduce comparisons between metanephric and other forms of kidney development for which such data exist to attempt to determine which features of metanephric development appear to be conserved in other types of vertebrate kidney development and which appear to be specializations that are specific to metanephros formation. A. Overview The first morphologic evidence for metanephros formation is the appearance of a bud called the ureteric bud (UB) near the posterior end of the nephric duct. The UB invades the adjacent intermediate mesoderm, which at this point is called the metanephric mesenchyme (MM; Figure 35.8). As a result of reciprocal interactions between the UB and the MM (Dressler, 2006; Lechner and Dressler, 1997; Yu et al., 2004), the UB branches, and cells of the MM condense at the tips of the branching UB. The condensed MM patches give rise to tubules and glomeruli, whereas the branched UB gives rise to the collecting system of the kidney. After multiple rounds of branching and condensation, the result is a highly organized arrangement of thousands of nephrons (approximately 10,000 in the mouse and 1,000,000 in the human metanephros) that drain into a central ureter (Saxen, 1987). From a purely morphologic point of view, the metanephros differs from the mesonephros in its highly branched structure and single ureteral drainage as compared with the mesonephros, which is much less highly branched and which has multiple connections to the nephric duct. Although the metanephros is typically larger than the mesonephros, it appears that the region of the IM (the MM) that gives rise to the metanephros extends the length of only a few somites, whereas the mesonephros derives from a much larger extent of trunk IM. Thus, the production of a large metanephros required the evolution

FIGURE 35.8 Steps in the formation of mouse metanephric nephrons. (Figure originally published in Yu J, McMahon AP, Valerius MT: Recent genetic studies of mouse kidney development, Curr Opin Genet Dev 14:550–557, 2004. Reprinted with permission of the publisher. See color insert.)

FORMATION OF THE MAMMALIAN METANEPHROS

793

of a mechanism for greatly expanding the nephrogenic mesenchyme while generating nephrons. It is this tremendous expansion of the nephrogenic mesenchyme that underlies many of the features of kidney development that are specific to the metanephros. B. The Ureteric Bud The first morphologically recognizable step in metanephros formation is the formation of a bud from the nephric duct: the UB. Formation of the UB is dependent on the signaling of the Gdnf/cRet pathway, with Gdnf expressed in the MM and its receptor cRet expressed in the nephric duct (Moore et al., 1996; Sainio et al., 1997b; Sanchez et al., 1996; Schuchardt et al., 1995). Mutations in either Gdnf or cRet lead to the absence of UB formation. Because cRet is expressed throughout the length of the nephric duct and Gdnf is also expressed in more anterior regions of the embryo, the question arises as to why only one UB forms and why it forms in the region of the MM. The examination of the Gdnf expression pattern sheds some light on this issue. Just before the time of UB formation, Gdnf expression becomes restricted to a region that is adjacent to the future UB site. This restriction is regulated at least in part by a mechanism involving FoxC1 and FoxC2 as well as Robo2/Slit2. In mutants of FoxC1 and FoxC2 or Robo2/Slit2, Gdnf remains expressed in more anterior IM regions, and multiple UBs are formed (Grieshammer et al., 2004; Kume et al., 2000). Thus, it appears that a broad region of the nephric duct is competent to form UBs and that the restriction of the ligand Gdnf is important for the generation of a single UB. Consistent with this interpretation is the fact that ectopic Gdnf can cause multiple buds to form from a nephric duct in vitro (Sainio et al., 1997b). Although the restriction of Gdnf provides an explanation for how a single UB is generated at this particular time in development, this does not explain why ectopic UBs do not form at earlier times, before Gdnf expression has become restricted. One possibility is that other, currently unknown components of the UB induction pathway are not functional until after embryonic day 10.5, thus preventing the premature initiation of UB formation. C. The Metanephric Mesenchyme Before its Interaction with the Ureteric Bud The mouse MM forms as an expanded region of IM at the axial level of the hind limb. On embryonic day 10.5, before its interaction with the UB, the MM is not readily distinguished by morphologic criteria. However, already at this time, the MM exhibits important molecular characteristics, and it expresses a significant number of kidney-associated regulatory genes, including Odd1, Eya1, Six1, Six2, Sall1, Gdnf, Wt1, and Pax2 (Armstrong et al., 1993; Brodbeck and Englert, 2004; Dressler et al., 1990; Kalatzis et al., 1998; Moore et al., 1996; Nishinakamura et al., 2001; Ohto et al., 1998; Pichel et al., 1996; Sanchez et al., 1996; So and Danielian, 1999; Xu et al., 2003). This is an important point, because it indicates that the MM has been significantly patterned before its interaction with the UB and that it does not require interaction with the UB for the initial expression of kidney-specific genes. It is likely (although not yet demonstrated) that the mechanisms that regulate initial IM formation (including BMP signaling, as discussed previously) are also involved in the early specification of the MM. Most genes that are characteristic of the MM at this time, such as Eya1, Six1, Six2, Sall1, Gdnf,

794

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

Odd1, Wt1, and Pax2, are not confined to the MM; they are also being expressed in more anterior regions of the IM. As discussed previously, among the few exceptions are genes of the Hox11 paralogous cluster, which are expressed in the MM region but not in more anterior IM (Patterson et al., 2001; Wellik et al., 2002). Because compound mutants in the Hox11 group do not generate a metanephros, it is possible that Hox11 family members play a role in specifying the metanephric character of this region of the IM. Many of the genes expressed in the MM before its interaction with the UB have been studied in mouse mutations. From these studies, Odd1 appears to lie genetically upstream of most genes that are known to be expressed in the MM, because Odd1 mutants exhibit an absence of Gdnf, Pax2, Eya1, Sall1, and Six2 (James et al., 2006; Wang et al., 2005). One exception is Wt1, which is expressed relatively normally in Odd1 mutants (James et al., 2006). Eya1 is required for the expression of other known MM genes aside from Odd1 and Wt1 (Sajithlal et al., 2005; Xu et al., 1999), and Six1 is required for Pax2, Six2, and Sall1 expression (Xu et al., 2003). Recent data from mouse metanephros cultures indicate that Wt1-regulated vascular endothelial growth factor signaling within the MM is also important for the activation of a subset of early metanephric genes (Gao et al., 2005). In mutants of Odd1, Wt1, and Eya1 (and of course of Gdnf), the UB does not form, whereas mutants for Six1 begin to form a UB (Xu et al., 2003), and Sall1 mutants show an arrest of UB formation at a somewhat later stage (Nishinakamura et al., 2001). Thus, the beginnings of a molecular pathway for early MM gene expression are emerging, with Odd1 lying at the most upstream point and being followed by Eya1, then Six1, and then Pax2, Six2, and Sall1. The expression of Wt1, which in turns regulates vascular endothelial growth factor expression, appears to be regulated largely independently of this pathway. It must be noted, however, that certain aspects of the regulatory relationships between these genes do not fit easily into simple hierarchical patterns; the interactions among gene products are likely to be complex. Although most genetic studies of mouse kidney development have focused on the metanephros, some of these studies have also provided information about the genes required for mesonephros formation. Some genes that are expressed in both mesonephric and metanephric mesenchyme are apparently required for metanephric but not mesonephric kidney development. Thus, embryos lacking Eya1, Six1, and Gdnf have apparently normal mesonephroi (Moore et al., 1996; Nishinakamura et al., 2001; Sajithlal et al., 2005; Sanchez et al., 1996; Xu et al., 1999; 2003), whereas embryos lacking Odd1 and Wt1 have defective (although present) mesonephroi (James et al., 2006; Sainio et al., 1997a). One possible explanation for the apparent lack of requirement for some of these genes in mesonephros formation is the existence of redundant pathways in mesonephros but not metanephros development. Another possible explanation is that most genetic studies have focused on the examination of the metanephros. Thus, subtle alterations in the mesonephros may have been missed in some cases. In this context, it should be noted that the normal mesonephros in the mouse is very small, containing only a dozen or so tubules, and it is thus atypical as compared with other amniotes. Therefore, it is possible that genes that appear to be dispensable for mouse mesonephric development may actually be required for the normal development of a more substantial mesonephros. Most interesting is the

FORMATION OF THE MAMMALIAN METANEPHROS

795

possibility that genes such as Eya1, Six1, and Gdnf, although expressed in the mesonephros and metanephros, are truly required only for metanephric development. This is consistent with the possibility (discussed later) that there are distinct regulatory pathways that operate in the metanephros and that function to generate the tremendous degree of controlled growth that is characteristic of the metanephros. D. Differentiation of the Metanephric Mesenchyme At the onset of interaction with the UB, gene expression in the MM is relatively uniform. After the initiation of UB formation and branching, gene expression in the MM becomes localized to specific subdomains (see Figure 35.8). MM condenses adjacent to the branching UB to form a “cap” of cells (Cho and Dressler, 2003; Dressler, 2002). Cells of the cap express Odd1, Eya1, Six2, Gdnf, Pax2, and other genes (Dressler et al., 1990; James et al., 2006; Kalatzis et al., 1998; Sainio et al., 1997b), although it is not clear whether all of the cap cells express all of the genes simultaneously. The next phase of MM differentiation is the further condensation of the region of the cap closest to the UB tip to form pretubular aggregates, which typically form on the side of the UB closest to the ureteric stalk (see Figure 35.8). The pretubular aggregates then epithelialize to form epithelial vesicles (EVs), which are the progenitors of the nephric tubules and glomeruli. It is worth noting that the formation of the EVs is an example of mesenchymal-to-epithelial transition, which is much rarer than its opposite, epithelialto-mesenchymal transition. There are several lines of evidence that indicate that Wnt signaling is important in the progression from the condensed mesenchyme of the caps to the formation of pretubular aggregates and EVs. The first evidence came from experiments in which MM was cultured in organ culture. During the 1950s, Grobstein et al. (1955) found that, if the MM is placed into culture, it will not undergo differentiation; however, if it is cocultured with the UB, it will differentiate into epithelialized tubules. In 1994, it was reported that fibroblasts expressing Wnt1 can substitute for the UB (Herzlinger et al., 1994). However, Wnt1 is not expressed in the UB, so the translation of this in vitro finding to kidney development in vivo was not clear. Subsequently, it was found that Wnt4 is expressed in pretubular aggregates and that the loss of Wnt4 prevents the subsequent development of the aggregates into kidney tubules (Kispert et al., 1998). Recently, it was reported that Wnt9b is expressed at the tips of the branching ureter and that, in embryos carrying Wnt9b mutations, MM differentiation is arrested before the pretubular aggregate stage (Carroll et al., 2005). Thus, a picture is emerging in which Wnt signaling is required during at least two phases in MM differentiation: a Wnt9b signal from the UB tips is required for the formation of pretubular aggregates, and a Wnt4 signal originating from the aggregates themselves is required for further tubular differentiation. Although Wnt9b and Wnt4 originate from different tissues, it could be argued that they both are part of a continuing requirement for Wnt signaling to allow MM differentiation to proceed. One possible model that is currently not proven is that the condensed cells of the cap may be blocked in their differentiation and that the function of Wnt signaling would be to derepress kidney differentiation. The function of carefully deployed Wnt signaling in this model would be to allow for the selective differentiation of parts of the MM into tubules

796

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

while other regions of the MM are maintained in a replicating, undifferentiated state to provide more material for subsequent rounds of tubule formation. Some tangential supporting evidence for this type of model comes from experiments in which the overexpression of Odd1 leads to the inhibition of epithelial kidney tubule formation (James et al., 2006). Support for the idea that differentiation of the condensed mesenchyme is under active suppression comes from the recently published Six2 mutant which undergoes precocious metanephric tubule formation, implying that Six2 is involved in the maintenance of condensed mesenchyme in a non-epithelialized state (Self et al., 2006). After its formation, each EV undergoes a complex process of regionalization and growth to generate a kidney tubule and glomerulus. On a morphologic level, the EV expands to form first a “comma-shaped body” and then an “S-shaped body,” which fuses (by a poorly understood mechanism) with a branch of the UB (Cho and Dressler, 2003). On a molecular level, the EV shows some signs of regionalization shortly after its formation, with respect to Notch signaling and integrin gene expression (Cheng et al., 2003; Cho and Dressler, 2003). Fibroblast growth factor signaling is critical for proper EV growth and patterning; conditional fibroblast growth factor 8 mutants show a loss of the loop of Henle and other nephron patterning defects (Grieshammer et al., 2005). Notch signaling also appears to be important in the patterning of the EV, because loss-of-function Notch mutations result in a loss of proximal tubule and glomerulus formation (Cheng et al., 2003). The picture that is beginning to emerge is that initial patterning of the tubule takes place already at the EV stage and that an important element of tubular differentiation is the selective growth of portions of this patterned rudiment. BMP-7 signaling has also been implicated in the subsequent growth of the metanephric kidney (Dudley et al., 1995). Subsequent growth and differentiation of the metanephric mesenchyme include further differentiation of the glomerulus, specialization of the tubular tissue, the formation and differentiation of the stroma (which lies between the nephrons), and the differentiation of the vascular system of the kidney. Most of these processes are not yet well understood at the molecular level, but the interested reader can find summaries of current knowledge in recent reviews (Abrahamson and Wang, 2003; Cho and Dressler, 2003; Cullen-McEwen et al., 2005; Woolf and Yuan, 2003; Yosypiv and El-Dahr, 2005). E. Further Differentiation of the Ureteric Bud and Collecting System When we left the story of the UB, it had just branched from the nephric duct in response to Gdnf signals emanating from the MM. Subsequently, the UB undergoes a large series of branching divisions that result in the highly branched collecting system of the metanephros. Although these branching events have often been described as bifurcations, in fact many different types of branching patterns have been observed (Cebrian et al., 2004). Mosaic mice in which green fluorescent protein is expressed in scattered cells in the UB have been used to trace the movement of cells within the UB as it branches (Shakya et al., 2005). These studies have found that the green-fluorescentprotein–labeled cRet-/- cells can participate in UB formation and branching if they are accompanied by wild-type cells, but that such mutant cells cannot move into the ampullae of the collecting ducts, where future branching events will take place. Thus, cRet signaling may be required not for collecting duct

SUMMARY AND SYNTHESIS

797

formation per se but rather for the ability to undergo budding and branching events. While signals from the MM are required for the induction of a UB and for subsequent branching of the UB in vivo, the UB can undergo extensive branching in vitro in the absence of the MM (Qiao et al., 1999). The culture medium in these experiments contained conditioned medium from MM cultures. These findings suggest that the UB itself possesses an intrinsic ability to undergo branching morphogenesis in response to general environmental cues from the MM. After UB formation, Gdnf expression becomes confined to the MM adjacent to the branching points of the UB (Miyamoto et al., 1997). This localized Gdnf expression has often been interpreted as inducing branching in the UB. However, the data indicating that the UB can branch in response to a nonlocalized Gdnf signal require some revisiting of this interpretation, and they suggest that the function of Gdnf and other localized signals in the MM might be to modify an autonomous branching program in the UB. The mechanisms of renal branching morphogenesis have been the subject of a large number of studies, which have been recently reviewed (Costantini, 2006). F. Concluding Remarks Concerning Metanephros Development As discussed above, the formation of the mammalian metanephros is an iterative process that leads to the generation of multiple generations of nephrons. This requires an intricate balance between the maintenance and proliferation of precursor tissue on the one hand and the differentiation of nephrons on the other. An important but poorly understood issue concerns how this balance between proliferation and differentiation is maintained, and the reason that the kidney stops growing and generating new nephrons after it reaches a certain size (which occurs approximately 3 weeks after birth in mice) also needs to be addressed. At this time, condensed mesenchymal cells can no longer be seen at the periphery of the kidney, which is the zone in which new nephrons are generated. The assumption has been that, at this time, all kidney precursor cells have been depleted; however, this has not been conclusively demonstrated. It would be interesting to determine whether any markers that are characteristic of MM cells are expressed in older kidneys and, if so, to attempt to characterize these cells.

SUMMARY AND SYNTHESIS The preceding discussion has reviewed the formation of selected components of the nephric system in several different vertebrate species. In this concluding section, I would like to touch on a few of the many ways that alterations in developmental mechanisms may underlie the great diversity of vertebrate kidney morphologies. When thinking about the relationships among the various vertebrate kidneys, it is helpful to consider the various types of vertebrate kidney in the context of their evolution from a postulated common ancestor as represented by the holonephric kidney. One can interpret the subsequent evolution of vertebrate kidneys as modifications of the relatively uniform holonephros. Thus, in other vertebrates, one sees specialization along the body axis, with

798

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

the posterior regions in general containing more complex kidney tissue. Selection for such specializations may be driven by animal size and/or the existence of an animal in more than one environment during its life cycle. Thus, larger animals would need larger kidneys to process the additional metabolic wastes, and such larger kidneys would most likely be generated at the posterior end of the animals, which is generated later during development. The transition to a terrestrial environment, with its strong need to conserve water, would also potentially select for a new type of kidney— the metanephros—that is more compact and that contains organizational features such as loops of Henle, which facilitate water retention. Although it is reasonable to suggest that genes that control AP patterning, such as members of the Hox family, may regulate the types of kidney tissue that develop along the AP axis, molecular data supporting this hypothesis have not yet been reported. Other genes (e.g., Eya1, Six1) that are expressed throughout the IM appear to be required only for metanephros formation, and they may be involved in regulating aspects of kidney development that are specific to the metanephros, such as the extensive proliferation that characterizes formation of the metanephros. Another departure from the holonephros comes with respect to the role of the nephric duct during tubule differentiation. The formation of the holonephric or, more importantly, the pronephric tubules does not appear to require induction from a nephric duct. Indeed, there is evidence (in Xenopus, at least) that the pronephric tubules are patterned before the specification of the nephric duct (Brennan et al., 1998). By contrast, mesonephric and metanephric tubules require activity from the nephric duct to induce and/or promote kidney tubule differentiation. It is possible that a role for the duct in regulating tubule formation emerged after the evolution of pronephroi. Embryos with functional pronephroi typically have, for a period of development, a nephric duct that runs through a region of trunk mesoderm that does not form tubules. This situation may have led to a role for the duct in regulating the differentiation of the mesonephros and the metanephros. In this respect, it would be interesting to know whether known inducers of kidney differentiation (e.g., members of the Wnt family) are expressed in regions of the nephric duct that pass through these regions of the amphibian embryo or whether such trunk mesoderm is competent to respond to Wnt signaling before the time at which it normally differentiates into mesonephric tubules. Comparisons among the various vertebrate kidneys also gives some insight into the development of the glomerulus. Although in the metanephroi of mammals and birds the glomerulus and the kidney tubule are intimately associated, this is almost certainly not true in the primitive vertebrate nephron. In the hagfish and many other fish and amphibian embryos, the bloodfiltering site (glomerulus) and the filtrate absorption and secretion site (tubule) are separated from each other by the coelomic space. The development of more efficient nephrons, with closely associated glomeruli and tubules, had to involve a modification of developmental programs so that the tubule and the glomerulus developed from the same basic primordia. One way of thinking about this issue may be to consider the individual renal vesicles that are generated during metanephros formation as equivalent to nephrocoeles, with each renal vesicle containing tubule and glomerulus precursors. Thus, one of the modifications of the kidney developmental program that may have

ACKNOWLEDGMENTS

799

occurred during vertebrate IM evolution is a movement from the formation of one large coelom in the IM to the formation of multiple coela, with each giving rise to a complete nephron unit. The molecular mechanisms by which the modification of this developmental program may have come about are not currently known. These are but a few of the many ways in which the kidney developmental program may have been modified to produce a variety of kidney types, both within a single species and among different species. Major challenges for the future are to identify additional developmental mechanisms that regulate kidney formation and to attempt to understand how the modification of such mechanisms can generate such a wide variety of kidney forms. As reviewed previously, genetics has already contributed in a major way to the identification of the molecular pathways that regulate kidney formation, particularly in the mouse metanephros and the zebrafish pronephros. Promising areas for future research include the development of in vitro systems in which kidney development can be induced in a controlled manner and the use of genomics-based approaches to identify genes with expression that is regulated under these conditions.

SUMMARY

 The basic building block of the kidney is the nephron, which consists of a 

    

glomerulus that filters the blood, a tubule that processes the glomerular filtrate, and a duct that drains the filtrate to the outside. The primitive vertebrate kidney is a holonephros, which contains one nephron per body segment. Other vertebrate kidneys include the pronephros, the mesonephros, and the metanephros, all of which are modifications of the holonephric kidney. All kidney tissue is derived from the intermediate mesoderm (IM). IM formation is regulated by signals from the lateral plate (one of which is BMP) and by unknown signals from axial or paraxial tissues. Modifications in the kidney developmental program underlie the generation of different kidney types from the IM. The metanephros is the adult kidney of amniote vertebrates, and it uses specialized developmental processes to generate a highly branched organ with a large number of nephrons. Some developmental mechanisms that regulate metanephros formation are specialized for the formation of the metanephros, whereas others are also used in the generation of other kidney types.

ACKNOWLEDGMENTS Many thanks to good friends and colleagues with whom it has been a pleasure to discuss some of the ideas contained in this article, including Iain Drummond, Doris Herzlinger, Richard James, and Rami Reshef. This work was supported by grants from the National Institutes of Health (National Institute of Diabetes and Digestive and Kidney Diseases) R01 DK59980 and R01 DK71041, and the Fogerty International Center R03 TW006864.

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

800

GLOSSARY Agnathostome A group of fish species characterized by lack of a jaw. Amniote The class of vertebrates that have embryos that contain amniotic membranes; this class includes mammals, birds, and reptiles. (Its opposite, the anamniote class, includes the vertebrates that lack amniotic membranes, such as fish and amphibians.) Apodan A class of legless amphibians that have embryos that possess holonephric-type kidneys. Axial Pertaining to the body axis, including the notochord and the neural tube. Coelom (plural: coela) The embryonic body cavity, which develops as a space within the embryonic mesoderm. Intermediate mesoderm A strip of mesoderm that is lateral to the somites and that gives rise to kidney tissue. Lateral plate Mesoderm that is lateral to the intermediate mesoderm and that gives rise to the limbs and tissues of the flank. Nephric Pertaining to the kidney. Paraxial Adjacent to the body axis; mainly pertaining to the somites.

REFERENCES Abrahamson DR, Wang R: Development of the glomerular capillary and its basement membrane. In P Vize P, AS Woolf AS, JBL Bard JBL, editors: The kidney: from normal development to congenital disease, San Diego, 2003, Academic Press, pp. 221–249. Armstrong JF, Pritchard-Jones K, Bickmore WA, et al: The expression of the Wilms’ tumour gene, WT1, in the developing mammalian embryo, Mech Dev 40:85–97, 1993. Barak H, Rosenfelder L, Schultheiss TM, Reshef R: Cell fate specification along the anterior-posterior axis of the intermediate mesoderm, Dev Dyn 232:901–914, 2005. Bouchard M, Souabni A, Mandler M, et al: Nephric lineage specification by Pax2 and Pax8, Genes Dev 16:2958–2970, 2002. Brennan HC, Nijjar S, Jones EA: The specification of the pronephric tubules and duct in Xenopus laevis, Mech Dev 75:127–137, 1998. Brodbeck S, Englert C: Genetic determination of nephrogenesis: the Pax/Eya/Six gene network, Pediatr Nephrol 19:249–255, 2004. Carroll TJ, Park JS, Hayashi S, et al: Wnt9b plays a central role in the regulation of mesenchymal to epithelial transitions underlying organogenesis of the Mammalian urogenital system, Dev Cell 9:283–292, 2005. Carroll TJ, Vize PD: Synergism between Pax-8 and lim-1 in embryonic kidney development, Dev Biol 214:46–59, 1999.

REFERENCES

801 Cebrian C, Borodo K, Charles N, Herzlinger DA: Morphometric index of the developing murine kidney, Dev Dyn 231:601–608, 2004. Cheng HT, Miner JH, Lin M, et al: Gamma-secretase activity is dispensable for mesenchyme-toepithelium transition but required for podocyte and proximal tubule formation in developing mouse kidney, Development 130:5031–5042, 2003. Cho EA, Dressler GR: Formation and development of nephrons. In Vize P, Woolf AS, Bard JBL, editors: The kidney: from normal development to congenital disease, San Diego, 2003, Academic Press, pp. 195–210. Cornish JA, Etkin LD: The formation of the pronephric duct in Xenopus involves recruitment of posterior cells by migrating pronephric duct cells, Dev Biol 159:338–345, 1993. Costantini F: Renal branching morphogenesis: concepts, questions, and recent advances, Differentiation 74:402–421, 2006. Cullen-McEwen LA, Caruana G, Bertram JF: The where, what and why of the developing renal stroma, Nephron Exp Nephrol 99:e1–e8, 2005. Dean B: On the embryology of Bdellostoma stouti, Verlag von Gustav Fischer in Jena 221–276, 1899. Dressler G: Tubulogenesis in the developing mammalian kidney, Trends Cell Biol 12:390–395, 2002. Dressler GR: The cellular basis of kidney development, Annu Rev Cell Dev Biol 22:509–529, 2006. Dressler GR, Deutsch U, Chowdhury K, et al: Pax2, a new murine paired-box-containing gene and its expression in the developing excretory system, Development 109:787–795, 1990. Drummond I, Majumdar A: The pronephric glomus and vasculature, In Vize P, Woolf AS, Bard JBL, editors: The kidney: from normal development to congenital disease, San Diego, 2003, Academic Press, pp. 61–73. Drummond IA, Majumdar A, Hentschel H, et al: Early development of the zebrafish pronephros and analysis of mutations affecting pronephric function, Development 125:4655–4667, 1998. Dudley AT, Lyons KM, Robertson EJ: A requirement for bone morphogenetic protein-7 during development of the mammalian kidney and eye, Genes Dev 9:2795–2807, 1995. Fraser EA: The development of the vertebrate excretory system, Biol Rev 25:159–187, 1950. Fujii T, Pichel JG, Taira M, et al: Expression patterns of the murine LIM class homeobox gene lim1 in the developing brain and excretory system, Dev Dyn 199:73–83, 1994. Gao X, Chen X, Taglienti M, et al: Angioblast-mesenchyme induction of early kidney development is mediated by Wt1 and Vegfa, Development 132:5437–5449, 2005. Goodrich ES: Coelom, genital ducts, and nephridia, Quart J Micro Sci 37, 1895. Goodrich ES: Studies on the structure and development of vertebrates, London, 1930, Macmillan. Grieshammer U, Cebrian C, Ilagan R, et al: FGF8 is required for cell survival at distinct stages of nephrogenesis and for regulation of gene expression in nascent nephrons, Development 132:3847–3857, 2005. Grieshammer U, Le M, Plump AS, et al: SLIT2-mediated ROBO2 signaling restricts kidney induction to a single site, Dev Cell 6:709–717, 2004. Grobstein C: Inductive interactions in the development of the mouse metanephros, J Exp Zool 130:319–340, 1955. Grote D, Souabni A, Busslinger M, Bouchard M: Pax 2/8-regulated Gata 3 expression is necessary for morphogenesis and guidance of the nephric duct in the developing kidney, Development 133:53–61, 2006. Gruenwald P: Zur entwicklungsmechanick der Urogenital-systems beim Huhn, Wilhelm Roux Arch Entw Mech 136:786–813, 1937. Hamburger V, Hamilton HL: A series of normal stages in the development of the chick embryo, J Morphol 88:49–92, 1951. Herzlinger D, Qiao J, Cohen D, et al: Induction of kidney epithelial morphogenesis by cells expressing Wnt-1, Dev Biol 166:815–818, 1994. James RG, Kamei CN, Wang Q, et al: Odd-skipped related 1 is required for development of the metanephric kidney and regulates formation and differentiation of kidney precursor cells, Development 133:2995–3004, 2006. James RG, Schultheiss TM: Patterning of the avian intermediate mesoderm by lateral plate and axial tissues, Dev Biol 253:109–124, 2003. James RG, Schultheiss TM: Bmp signaling promotes intermediate mesoderm gene expression in a dose-dependent, cell-autonomous and translation-dependent manner, Dev Biol 288:113–125, 2005.

802

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

Kalatzis V, Sahly I, El-Amraoui A, Petit C: Eya1 expression in the developing ear and kidney: towards the understanding of the pathogenesis of branchio-oto-renal (BOR) syndrome, Dev Dyn 213:486–499, 1998. Kim D, Dressler GR: Nephrogenic factors promote differentiation of mouse embryonic stem cells into renal epithelia, J Am Soc Nephrol 16:3527–3534, 2005. Kispert A, Vainio S, McMahon AP: Wnt-4 is a mesenchymal signal for epithelial transformation of metanephric mesenchyme in the developing kidney, Development 125:4225–4234, 1998. Kreidberg JA, Sariola H, Loring JM, et al: WT-1 is required for early kidney development, Cell 74:679–691, 1993. Kume T, Deng K, Hogan BL: Murine forkhead/winged helix genes Foxc1 (Mf1) and Foxc2 (Mfh1) are required for the early organogenesis of the kidney and urinary tract, Development 127:1387–1395, 2000. Lechner MS, Dressler GR: The molecular basis of embryonic kidney development, Mech Dev 62:105–120, 1997. Majumdar A, Lun K, Brand M, Drummond IA: Zebrafish no isthmus reveals a role for pax2.1 in tubule differentiation and patterning events in the pronephric primordia, Development 127:2089–2098, 2000. Mauch TJ, Yang G, Wright M, et al: Signals from trunk paraxial mesoderm induce pronephros formation in chick intermediate mesoderm, Dev Biol 220:62–75, 2000. McLaughlin KA, Rones MS, Mercola M: Notch regulates cell fate in the developing pronephros, Dev Biol 227:567–580, 2000. Melby AE, Beach C, Mullins M, Kimelman D: Patterning the early zebrafish by the opposing actions of bozozok and vox/vent, Dev Biol 224:275–285, 2000. Miyamoto N, Yoshida M, Kuratani S, et al: Defects of urogenital development in mice lacking Emx2, Development 124:1653–1664, 1997. Moore MW, Klein RD, Farinas I, et al: Renal and neuronal abnormalities in mice lacking GDNF, Nature 382:76–79, 1996. Moriya N, Uchiyama H, Asashima M: Induction of pronephric tubules by activin and retinoic acid in presumptive ectoderm of Xenopus laevis, Dev Growth Differ 35:123–128, 1993. Nishinakamura R, Matsumoto Y, Nakao K, et al: Murine homolog of SALL1 is essential for ureteric bud invasion in kidney development, Development 128:3105–3115, 2001. Obara-Ishihara T, Kuhlman J, Niswander L, Herzlinger D: The surface ectoderm is essential for nephric duct formation in intermediate mesoderm, Development 126:1103–1108, 1999. Ohto H, Takizawa T, Saito T, et al: Tissue and developmental distribution of Six family gene products, Int J Dev Biol 42:141–148, 1998. Ota KG, Kuratani S: The history of scientific endeavors towards understanding hagfish embryology, Zoolog Sci 23:403–418, 2006. Patterson LT, Pembaur M, Potter SS: Hoxa11 and Hoxd11 regulate branching morphogenesis of the ureteric bud in the developing kidney, Development 128:2153–2161, 2001. Pichel JG, Shen L, Sheng HZ, et al: Defects in enteric innervation and kidney development in mice lacking GDNF, Nature 382:73–76, 1996. Poole TJ, Steinberg MS: Amphibian pronephric duct morphogenesis: segregation, cell rearrangement and directed migration of the Ambystoma duct rudiment, J Embryol Exp Morphol 63:1–16, 1981. Price GC: Development of the excretory organs of a myxinoid Bdellostoma stouti Lockington, Zoolog Jahrb 10:205–226, 1897. Price GC: A further study of the development of the excretory organs in Bdellostoma stouti, Amer J Anat 4:117–138, 1904–1905. Psychoyos D, Stern CD: Fates and migratory routes of primitive streak cells in the chick embryo, Development 122:1523–1534, 1996. Qiao J, Sakurai H, Nigam SK: Branching morphogenesis independent of mesenchymal-epithelial contact in the developing kidney, Proc Natl Acad Sci U S A 96:7330–7335, 1999. Romer AS: The vertebrate body, Philadelphia, 1955, WB Saunders. Sainio K: Development of the mesonephric kidney, In Vize P, Woolf AS, Bard JBL, editors: The kidney: from normal development to congenital disease, San Diego, 2003, Academic Press, pp. 75–86. Sainio K, Hellstedt P, Kreidberg JA, et al: Differential regulation of two sets of mesonephric tubules by WT-1, Development 124:1293–1299, 1997a. Sainio K, Suvanto P, Davies J, et al: Glial-cell-line-derived neurotrophic factor is required for bud initiation from ureteric epithelium, Development 124:4077–4087, 1997b.

REFERENCES

803 Sajithlal G, Zou D, Silvius D, Xu PX: Eya1 acts as a critical regulator for specifying the metanephric mesenchyme, Dev Biol 284:323–336, 2005. Sanchez MP, Silos-Santiago I, Frisen J, et al: Renal agenesis and the absence of enteric neurons in mice lacking GDNF, Nature 382:70–73, 1996. Saxen L: Organogenesis of the kidney, London, 1987, Cambridge University Press. Schier AF: Nodal signaling in vertebrate development, Annu Rev Cell Dev Biol 19:589–621, 2003. Schuchardt A, Srinivas S, Pachnis V, Costantini F: Isolation and characterization of a chicken homolog of the c-ret proto-oncogene, Oncogene 10:641–649, 1995. Schultheiss TM, Burch JBE, Lassar AB: A role for bone morphogenetic proteins in the induction of cardiac myogenesis, Genes Dev 11:451–462, 1997. Schultheiss TM, James RG, Listopadova A, Herzlinger D: Formation of the nephric duct, In Vize P, Woolf AS, Bard JBL, editors: The kidney, Amsterdam, 2003, Academic Press, pp. 464–477. Self M, Lagutin OV, Bowling B, Hendrix J, Cai Y, Dressler GR, Oliver G Six2 is required for suppression of nephrogenesis and progenitor renewal in the developing kidney, EMBO J 25: 5214–5228, 2006. Serluca FC, Fishman MC: Pre-pattern in the pronephric kidney field of zebrafish, Development 128:2233–2241, 2001. Seufert DW, Brennan HC, DeGuire J, et al: Developmental basis of pronephric defects in Xenopus body plan phenotypes, Dev Biol 215:233–242, 1999. Shakya R, Watanabe T, Costantini F: The role of GDNF/Ret signaling in ureteric bud cell fate and branching morphogenesis, Dev Cell 8:65–74, 2005. Shawlot W, Behringer RR: Requirement for Lim1 in head-organizer function, Nature 374: 425–430, 1995. Sive HL: The frog princess: a molecular formula for dorsoventral patterning in Xenopus, Genes Dev 7:1–12, 1993. So PL, Danielian PS: Cloning and expression analysis of a mouse gene related to Drosophila oddskipped, Mech Dev 84:157–160, 1999. Swindell EC, Thaller C, Sockanathan S, et al: Complementary domains of retinoic acid production and degradation in the early chick embryo, Dev Biol 216:282–296, 1999. Takezaki N, Figueroa F, Zaleska-Rutczynska Z, Klein J: Molecular phylogeny of early vertebrates: monophyly of the agnathans as revealed by sequences of 35 genes, Mol Biol Evol 20: 287–292, 2003. Van Campenhout C, Nichane M, Antoniou A, et al: Evi1 is specifically expressed in the distal tubule and duct of the Xenopus pronephros and plays a role in its formation, Dev Biol 294:203–219, 2006. Vize PD, Seufert DW, Carroll TJ, Wallingford JB: Model systems for the study of kidney development: use of the pronephros in the analysis of organ induction and patterning, Dev Biol 188:189–204, 1997. Vize P, Woolf AS, Bard JBL, editors: The kidney: from normal development to congenital disease, San Diego, 2003, Academic Press. Waddington CH: The morphogenetic function of a vestigal organ in the chick, J Exp Biol 15:271–377, 1938. Wang Q, Lan Y, Cho ES, et al: Odd-skipped related 1 (Odd 1) is an essential regulator of heart and urogenital development, Dev Biol 288:582–594, 2005. Wellik DM, Hawkes PJ, Capecchi MR: Hox11 paralogous genes are essential for metanephric kidney induction, Genes Dev 16:1423–1432, 2002. Wilm B, James RG, Schultheiss TM, Hogan BL: The forkhead genes, Foxc1 and Foxc2, regulate paraxial versus intermediate mesoderm cell fate, Dev Biol 271:176–189, 2004. Woolf AS, Yuan HT: Development of kidney blood vessels, In Vize P, Woolf AS, Bard JBL, editors: The kidney: from normal development to congenital disease, San Diego, 2003, Academic Press, pp. 251–266. Xu PX, Adams J, Peters H, et al: Eya1-deficient mice lack ears and kidneys and show abnormal apoptosis of organ primordia, Nat Genet 23:113–117, 1999. Xu PX, Zheng W, Huang L, et al: Six1 is required for the early organogenesis of mammalian kidney, Development 130:3085–3094, 2003. Yosypiv IV, El-Dahr SS: Role of the renin-angiotensin system in the development of the ureteric bud and renal collecting system, Pediatr Nephrol 20:1219–1229, 2005. Yu J, McMahon AP, Valerius MT: Recent genetic studies of mouse kidney development, Curr Opin Genet Dev 14:550–557, 2004.

TOPICS IN VERTEBRATE KIDNEY FORMATION: A COMPARATIVE PERSPECTIVE

804

FURTHER READING An excellent review of comparative kidney morphology and embryology can be found in Elizabeth Fraser’s review from 1950 (Fraser, 1950). A more exhaustive treatment of the subject can be found in Goodrich’s classic (Goodrich, 1930). Saxen’s book-length treatment of kidney development remains the standard synthesis of the classical literature regarding kidney embryology, with an emphasis on mammalian metanephros formation (Saxen, 1987). For more contemporary treatments of kidney development, the book edited by Peter Vize (Vize et al., 2003) contains many outstanding chapters. Vize has also published an excellent review of the pronephros as a model system for studying kidney development (Vize et al., 1997). Finally, for readers interested in a compact, up-to-date review of the molecular aspects of kidney development, Dressler has recently published an excellent overview (Dressler, 2006).

36

DEVELOPMENT OF THE GENITAL SYSTEM HONGLING DU and HUGH S. TAYLOR Division of Reproductive Endocrinology and Infertility, Department of Obstetrics, Gynecology and Reproductive Sciences, Yale University School of Medicine, New Haven, CT

INTRODUCTION In mammals, sex development is a genetically and hormonally controlled process, which involves three main sequential processes. It begins with the establishment of chromosomal or genetic sex at fertilization, when a sperm (with a Y or an X chromosome) fertilizes an ovum (with an X chromosome). This initial phase of genital development represents the genetic sex determination, without any morphologic indication of sex. The early embryo is phenotypically identical in both sexes. Therefore, this is referred to as the indifferent stage of sexual development. Next, gonadal differentiation is initiated in accordance with the expression of sex differentiation genes. During this phase, the gonads begin to acquire sexual characteristics and differentiate into either testes or ovaries. Subsequently, the development of internal or external sexual duct systems takes place. The morphologic differentiation of sex is considered to begin with gonadal differentiation and to progress with sexual duct system development, which is influenced by the gonads. In this chapter, we will list the genetic factors that lead to sexual and gonadal differentiation and compare them among different species. Genetic and hormonal factors, which play a role in the formation of the internal and external genitals of both sexes, will be discussed. Malformations of the genital system and its mechanisms will be described.

I. GENETIC SEX DETERMINATION A. Sex Chromosomes In mammals, sex determination is accomplished by a chromosomal mechanism. Whether a mammalian embryo develops into a male or a female is Principles of Developmental Genetics © 2007, Elsevier Inc. All rights reserved.

805

806

DEVELOPMENT OF THE GENITAL SYSTEM

determined by its complement: XX embryos become females, and XY embryos become males. In humans, as in other mammals, females have two copies of a large X chromosome, and males have a single X and a much smaller and heterochromatic Y chromosome. The Y chromosome has a strong testis-determining effect on the indifferent stage, thus determining male gonad development. The hypothetical factor on the Y chromosome required for male differentiation is called testis-determining factor (TDF). The male-specific Y chromosome in mammals, in addition to playing a vital role in sex determination, also harbors the genes that are required for spermatogenesis (Tiepolo et al., 1976; Welshons et al., 1959). The number of X chromosomes appears to be unimportant to sex determination, as indicated by the loss of an X chromosome in patients with Turner syndrome (45,X or 45,XO). These patients present gonadal dysgenesis, but they are phenotypically females, with the ovaries represented by gonadal streaks. To summarize, the presence of a Y chromosome results in the differentiation of the embryonic somatic cells of the gonad into testes rather than ovaries. The absence of a Y chromosome results in female gonad development and the formation of the ovaries. The gonads then determine the type of sexual differentiation in the internal genital ducts and the external genitalia. Nonmammalian vertebrate species have a variety of sex chromosomal systems, such as ZZ–WZ, WX–XX–WY, and XY (Schartl, 2004). For example, birds have differentiated Z and W chromosomes in which the W is usually small and heterochromatic. As opposed to mammals, male birds are homogametic, which means that the sex produces one type of gamete with respect to sex chromosome content. Those with two copies of the Z chromosome (ZZ) are male, whereas those with a one each of the Z and W chromosomes (WZ) are female. In some reptiles, such as crocodilians and marine turtles, the development of the embryonic gonad into testis or ovary is dependent on temperature. By contrast, snakes have a fixed genetic sex determination system. The chromosomal system is the WZ type, and the incubation temperature of the eggs cannot influence the development of the embryonic gonads. In amphibians, most species have homomorphic sex chromosomes, which are morphologically identical members of an homologous pair of chromosomes. However, there are also some species with heteromorphic sex chromosomes, which are a chromosome pair with some homology but that differ with regard to size, shape, and staining properties. In the Japanese frog Rana rugosa, populations with heteromorphic XY, homomorphic XY, and heteromorphic WZ sex chromosomes have been identified, even within the same species. In fish, the chromosomal mechanisms show enormous variation. XY and WZ systems are the most common. In addition, XO, ZO, X1X2Y, XY1Y2, and Y autosome fusion have been described. Systems exist in which multiple sex chromosomes are present in a population. For example, three types of sex chromosomes (X, W, and Y) coexist in a population of platyfish (Xiphophorus maculatus). WX, XX, and WY become females, whereas XY and YY fish become males. However, in fish and amphibians, sex can be reverted by age, social factors, and temperature or hormone treatment, or this can even occur spontaneously. In nonmammalian vertebrates, the variety of the chromosomal system is the consequence of the dynamic process of the evolution of sex determination mechanisms.

GENETIC SEX DETERMINATION

807

B. The Sex Determination Gene 1. Evolution of Sex Chromosomes In mammals, male sex determination is controlled by genes on the Y chromosome. Evolutionary comparisons show that X and Y chromosomes were originally homologous. Although the X chromosome is consistent in size and gene content, the Y chromosome is much more variable among mammalian species. A comparison of the gene content of sex chromosomes from the three major groups of extant mammals (placentals, marsupials, and monotremes) shows that part of the X chromosome and a corresponding region of the Y chromosome is shared by all mammals and thus must be very ancient. In humans, the X and Y chromosomes share two short regions at either end (the pseudoautosomal regions [PARs]) throughout which they are homologous and thus may recombine (Burgoyne, 1998). The PAR on the short arms of the X and Y chromosomes (PAR1) undergoes pairing and recombination at meiosis. By contrast, the PARs on the long arms of the X and Y chromosomes (PAR2) pair only infrequently, and the homology observed in this region is probably maintained by gene conversion. In addition, many genes and pseudogenes on the Y chromosomes have homologues on the X chromosomes. The evolution of the mammalian Y took place in several cycles of addition and attrition as autosomal regions were added to the pseudoautosomal region of one sex chromosome, recombined onto the other, and degraded on the Y. This indicates that no matter how different in size and gene content today, the X and Y chromosomes were once homologues. The Y chromosome seems to have degraded progressively over the last 200 million years, perhaps as a consequence of keeping the sex-determining gene together with allied malespecific genes. Similar degradation occurs in single chromosomes that do not undergo recombination, irrespective of sex. Snakes have a ZZ male and a ZW female system. Like the mammalian X chromosome, the Z chromosome of snakes is large, containing about 6% of the genome in all snake families. Birds that are as distantly related as the chicken and the emu have Z chromosomes that are nearly identical genetically (Shetty, 1999). The W chromosome is considered more variable, and it has different sizes in various families of birds and snakes. The W chromosome is largely homologous to the Z chromosome in the emu, but it has become small and heterochromatic in the chicken. The process of W chromosome degradation has therefore taken place to different extents and independently in different bird and snake lineages. The differentiation of the X and Y chromosomes is thought to have been initiated in an ancestral mammal when an allele at a single locus on the protoY took over a male-determining function from an ancestral genetic or environmental sex-determining system. The comparison of sequences of X-Y shared genes across species shows that the Y copy changes far more rapidly than the X copy. Comparative mapping studies show no homology between bird Z and mammalian X sex chromosomes, which indicates that the two sex chromosomes systems evolved independently. Other vertebrate classes show a wide variety of genetic and environmental sex determination mechanisms, so it is not possible to infer the sex-determining system of the common reptilian ancestor. 2. Sex-Determining Genes on the Y Chromosome in Mammals In the mammalian XY chromosomal sex-determining system, sex differentiation depends on Y-chromosome–specific genes that trigger male development. Several testis-determining candidate genes have been identified: a minor

808

DEVELOPMENT OF THE GENITAL SYSTEM

male-specific antigen (HYA); the zinc finger Y chromosome (ZFY); and the sex-determining region of the Y chromosome (SRY). a. HYA HYA is a minor male-specific antigen. It was originally discovered during the mid 1950s as a transplantation or histocompatibility antigen that caused female mice of a certain inbred strain to reject male skin of their own strain. The HYA/hya gene encoding the HYA has been mapped to the long arm of the Y chromosome in humans and the short arm of the Y chromosome in mice. During the 1970s, HYA was proposed to be the TDF as a result of two traits: HYA is uniquely expressed on male cells, and the presence of HYA is associated with testis determination as indicated by XX sex-reversed male mice. These mice have two X chromosomes, but they are clearly phenotypically male. As a result of a chromosome rearrangement, the HYA gene is expressed in these XX mice, which provides evidence that is consistent with HYA being the testis-determining Y-encoded gene (Wachtel, 1975). However, during the 1980s, it was found that HYA is absent from certain mice that develop testes and that are of indisputably male phenotype; this finding disputed the function of the HYA as the TDF (McLaren et al., 1984). b. ZFY Another candidate gene, ZFY, which is located on the Y chromosome, encodes a zinc finger protein. It lies close to the pseudoautosomal boundary on the short arm of the human Y chromosome. In the mouse, Zfy was found to consist of two duplicated genes, Zfy-1 and Zfy-2, which are both present on the normal human Y chromosome. ZFY was initially considered as a TDF candidate (Page, 1987), but its lack of conservation on the Y chromosome in marsupials made it an unlikely candidate for a universal mammalian TDF. More refined mapping of the human Y chromosome defined a new minimum sex-determining region that lacked ZFY, and this ultimately excluded ZFY from a role in sex determination (Palmer et al., 1989). c. SRY In 1991, the SRY gene was cloned from a region closely linked to ZFY, and it has been confirmed as the TDF needed from the Y chromosome to establish male development (Figure 36.1; Koopman et al., 1990; Sinclair et al.,1990). SRY/sry is a small intronless gene that encodes a protein with a conserved DNA-binding high mobility group (HMG) box. SRY is a member of a large family of SRY-like HMG-box containing genes. The presence of an SRY mutation in about 15% of human XY females supported the proposition that this gene represented the TDF. The identity of sry as the TDF was determined in the mouse. The sry gene is absent in a strain of XY mice that are phenotypically female. In the transgenic mouse, sry can cause XX mice to undergo sex reversal and develop as males. This occurs despite the fact that they lack all other genes of the Y chromosome. SRY is transcribed in the genital ridges of embryos just before testis differentiation, but it is not expressed in the gonads of female mice embryos. In addition, sry was cloned from marsupials and shown to map to the Y chromosome, which indicates that it represents the common ancestral mammalian TDF. The SRY gene encodes a transcription factor that regulates the genes that are responsible for testicular development.

809

GENETIC SEX DETERMINATION

Yp

PseudoAutosomal region

1A1 (140kb)

60kb SRY

Centromere 1

1A2 (140kb)

35kb

ZFY

2

Yq

3 1B 4A 1C

1959

1966

1986

1987

1989

1990

FIGURE 36.1 A history of the sex-determination gene (Sry) localized on the Y chromosome. The central role of the Y chromosome in male sex determination has been recognized for many years. In 1987, ZFY was isolated and initially equated with the TDF. In 1989, another area located near ZFY was found to play role in male sex determination. In 1991, SRY gene was isolated in this area and identified as the TDF. (Adapted from Sultan et al., 1991.)

3. Sex-Determining Genes on the Z or Y Chromosome in Nonmammalian Vertebrates Sex-determination systems are diverse in vertebrates. As described previously, in mammals, sex differentiation depends on sex-determining genes. However, in nonmammalian vertebrates, sex is also determined by heredity, the environment, or both. Interestingly, SRY-like sex-determining genes have also been identified in nonmammalian vertebrates, and they are considered to be one of the sex-determining factors. a. DMRTI Encoding Doublesex and mab-3–related transcription factor 1 (Dmrt1) is a Z-linked candidate for the male sex-determining gene. In birds, there is no copy of Dmrt1 on the W chromosome, and males have two copies of the gene. Dmrt1 has been implicated in male sexual development in many vertebrate species, and it is considered to be the “Sry gene” of nonmammalian vertebrates (Zarkower, 2001). Recent advances in the characterization of the human Dmrt0001 gene shows that multiple transcripts are expressed in the human testis. Thus, Dmrt1 is likely to have an important role in the evolution of the sexual development mechanisms of many species (Cheng et al., 2006). b. DMY The DM domain gene on the Y chromosome (DMY) has been found in the sex-determining region of the Y chromosome of the teleost medaka fish Oryzias latipes. Mutations of the DMY cause a simple sex reversal in medaka, and this is confirmed by naturally occurring mutant females in several wild populations. DMY appears to be closely related to Dmrt1 with regard to both nucleotide

810

DEVELOPMENT OF THE GENITAL SYSTEM

sequence (93% identity) and function (Matsuda, 2003). So far, the evolutionary comparison of sex-determination genes indicates that molecular similarities among phyla are only present in the fly doublesex, the worm mab-3, and the vertebrate Dmrt1 (dsx- and mab3-related transcription factor 1)/Dmy genes.

II. GONADAL DIFFERENTIATION A. Primordial Germ Cell Migration Primordial germ cells (PGCs) are the embryonic precursors of the gametes. In all systems, PGCs form far from the site of the developing gonads and migrate to the sites of developing ovaries or testes. This isolation may be important in the maintenance of their unique characters. It is accepted that PGC migration occurs in three phases: separation, migration, and colonization. Several mechanisms have been hypothesized to explain PGC migration, including self-movement, attraction by chemotactic factors, PGC–PGC interactions, substrate guidance, and interaction with extracellular matrix molecules. In human embryos, PGCs are visible early during the fourth week of gestation among the endodermal cells from the posterior wall of the yolk sac, near the origin of the allantois. During the folding of the embryo, part of the yolk sac is incorporated into the embryo, and PGCs migrate along the wall of the hindgut and through the dorsal mesentery, until they approach the newly appearing genital ridges late in the fifth week. During the sixth week, PGCs migrate into the underlying mesenchyme, and they become incorporated into the primary sex cords. The migration of PGCs is considered to be accomplished by the active ameboid movement of the cells in response to a permissive extracellular matrix substrate. In the mouse, PGCs are induced to form in the proximal epiblast, where their formation is dependent on the expression of bone morphogenetic proteins 4 and 8b in extraembryonic tissue (Lawson et al., 1999; Ying et al. 2001). During gastrulation, they move through the primitive streak and invade the definitive endoderm, the parietal endoderm, and the allantois. In both XX and XY mice, PGCs can first be detected at 7.5 days post coitum (dpc) at the base of the allantois. By 9.0 dpc, the PGCs in the definitive endoderm become incorporated into the hindgut. Between 9.0 dpc to 9.5 dpc, PGCs migrate through the dorsal side of the hindgut to colonize the developing genital ridge. At 10.5 dpc, PGCs begin to cluster, forming a network of migrating cells. By 11.5 dpc, most PGCs have colonized the genital ridge. The entire migration takes approximately 4 days. In mouse embryos, two genes required for the migration of PGCs have been identified: c-kit and steel. c-kit encodes a protein receptor that is located on the surface of the migrating PGC; steel encodes a growth factor that is expressed in the somatic cells that are placed in the pathway of the migrating PGC and that functions as the ligand of c-kit. The steel/c-kit interaction forms a ligand–receptor pathway that is required for PGC colonization, survival, and migration (Motro et al. 1991). fragillis/mil-1 was recently identified as the gene that initiates PGC motility. It is a member of an interferon-inducible family of genes that is implicated in homotypic cell–cell adhesion and cell-cycle control. At 7.25 dpc, fragilis is

GONADAL DIFFERENTIATION

811

expressed in the posterior epiblast, with the highest level of expression overlapping the region where PGCs are formed. Twenty-four hours later, its expression is downregulated as PGCs scatter and move into the endoderm (Saitou et al., 2002). The expression of two other members of the fragilis gene family (fragilis 2 and fragilis 3) was found in nascent PGCs that play a role in maintaining PGC migration. PGC migration in flies can be divided into four stages: (1) internalization of the pole cells; (2) emigration of PGCs from the gut; (3) the lateral migration of PGCs; and (4) gonad coalescence. PGCs arise from the posterior pole of the developing embryo, where localized maternal components become segregated into pole cells. Next, PGCs migrate out of the ventral side of the gut along the basal surface and into the lateral mesoderm, where they coalesce with the somatic cells of the gonad. torso is a maternally inherited transcript encoding a tyrosine kinase. Its activity is required for the efficient incorporation of pole cells into the hindgut pocket. torso signaling in flies initiates pole cell multiplication, and it may be analogous to the role of c-kit in PGC development in mice. The loosening of cell–cell contacts between the cells of midgut epithelium allows PGCs to emigrate from the gut, forcing them away from the gut and toward the overlying mesoderm by a repulsive signal mediated by wunen/wunen2. Thus, wunen gene expression appears to be responsible for directing the migration of the PGC in flies. Similarly, in zebrafish, PGCs are specified by maternal components that become segregated into four clusters within the cleaving embryo. During gastrulation, these PGCs cluster more dorsally and align at the border between the head and trunk mesoderm, or they align within the lateral mesoderm. Next, PGCs migrate posteriorly to colonize the developing gonad. dead end is a gene that has been identified in zebrafish that plays a specific role in the initiation of PGC motility. dead end homologues have been identified in PGCs in Xenopus, chicken, and mouse. However, in the mouse, dead end is expressed in PGCs after the migratory stages; hence, its function in PGC development may not be entirely conserved. Because there are enormous ethical, technical, and logistic problems with regard to in vivo studies of the movement of PGCs in humans, many experiments have been carried out in other animal models. The initiation of PGC motility is currently poorly understood, and it may be controlled by speciesspecific mechanisms. Despite their different origins, the early development of PGCs in flies and mice is quite similar, and the survival and early migration of PGCs in these systems require signaling via tyrosine kinase receptors (torso and c-kit, respectively). A tyrosine kinase receptor with a role similar to that of torso/c-kit has yet to be identified in zebrafish. The initiation of PGC motility in zebrafish is controlled by the mRNA binding protein dead end. PGC guidance mechanisms have been well studied in all three species, and they require chemoattractants that signal via G-protein–coupled receptors, cell–cell adhesion, and, probably, specific interactions between PGCs and the extracellular matrix. Despite all this, evidence now suggests that there is no active migration of PGCs in the human embryo and that the displacement of germ cells can be explained by the global growth and movement of the embryo (Freeman, 2003). The analysis of recent data suggests that human PGCs do not actively migrate at any stage (either up or down) but rather that they are embedded in

812

DEVELOPMENT OF THE GENITAL SYSTEM

the caudal tissues of the embryo (e.g., allantois, hindgut, coelomic serosa) and carried along passively during the curvilinear unrolling of this caudal region. The similarities between mouse and human embryos are so impressive that it has been doubted whether PGCs are moving independently in the mouse embryo, either. It has been proposed that, in the mammalian embryo, there is a passive carriage of multiplying cells in a caudal direction rather than an active ascent of PGCs. Such carriage of cells accompanies the normal development of the caudal part of the embryo (the so-called embryonic unrolling). This new hypothesis for a passive migration of PGCs in the human embryo encourages a reexamination of evidence for the previously widely accepted active migration of PGCs in other species. B. Origin of the Gonads The gonads are derived from three sources: the coelomic epithelium, the underlying mesenchyme, and the PGCs. Gonadal organogenesis begins with the appearance of the genital ridge. Initially, a thickened area of coelomic epithelium develops on the medial aspect of the mesonephros. After PGCs begin to colonize the ventral region of the urogenital ridge, cells of the coelomic epithelium and the underlying mesenchymal cells undergo active proliferation. Branches of blood vessels from the dorsal aorta and the cardinal veins send endothelial cells into the genital ridges. Proliferating coelomic and mesenchymal cells, together with arriving PGCs and endothelial cells, form a cluster of condensed cells that gradually become the undifferentiated gonad. In the mouse, the urogenital ridge develops beginning around 9.5 dpc. The morphologic establishment of the undifferentiated gonad takes approximately 48 hours (10 to 12 dpc). During this stage, the initially amorphous cluster of condensed cells becomes segregated into two compartments: (1) an epithelial compartment formed by PGCs and epithelial-like cells surrounded by a basal membrane; and (2) a stromal compartment formed by mesenchymal cells, fibroblasts, and blood vessels. In the human, gonadal development is first identified during the fifth week of gestation. The gonads arise from an elongated region of mesoderm along the ventromedial border of the mesonephros. The indifferent gonad develops in close association with the mesonephros, an embryonic kidney that contributes to both the male and female reproductive tracts. Germ cells induce cells of the mesonephros to form the genital ridge. Cells in the cranial part of this region condense to form the adrenocortical primordia, and those of the caudal part become the genital ridges. Soon, finger-like epithelial cords called primary sex cords grow into the underlying mesenchyme. C. The Bipotential Gonad and the Sex-Determining Switch In mammals, both testis and ovary share a common origin—the bipotential gonad—that possesses neither distinctly male nor female characteristics. The indifferent gonad can differentiate into either testes or ovaries, depending on the presence or absence of a Y chromosome, respectively. The fate of the gonad is specified by the SRY gene product. In all other reproductive organs (discussed later), sexually dimorphic development does not depend directly on chromosomal compliment but rather on the presence of either the male or female gonad.

GONADAL DIFFERENTIATION

813

There are essentially three different cell lineages present in the gonads in addition to the germ cells. Each lineage has a bipotential fate and is capable of differentiating along either the male or female pathway. The supporting cell lineage will give rise to Sertoli cells in the testis and follicle cells in the ovary. These cells surround the germ cells and provide an appropriate growth environment. The steroidogenic cell lineage produces the sex steroid hormones that will contribute to the development of the secondary sexual characteristics of the embryo. In the male, these correspond with the Leydig cells; in the female, these correspond with the theca cells. Finally, a connective cell lineage contributes to the formation of the organ as a whole in both the testes and the ovaries. Early testis development is characterized by the formation of testicular cords that contain Sertoli and germ cells, with the Leydig cells excluded to the interstitium. The connective cell lineage is a major contributor to cord formation as the peritubular myoid cells surround the Sertoli cells; together, they lay down the basal lamina. The testis is also characterized by rapid and prominent vascularization. Organization of the ovary takes place later in development and is less structured, with the connective tissue lineage giving rise to stromal cells and with no myoid cell equivalent. There are three genes that have been identified as necessary for the development of the undifferentiated or bipotential gonad in the mouse: wt-1, sf1, and lim-1 (Luo et al., 1994; Pritchard-Jones et al., 1990; Shawlot et al., 1995). The wt-1 gene was isolated from humans who developed the Wilms’ kidney tumor, and it is specifically expressed within the developing genital ridge and the kidney in mouse embryos. wt-1 knockout mouse embryos lack kidneys and a genital ridge. The orphan nuclear receptor steroidogenic factor 1 gene (sf1) is a regulator of the tissue-specific expression of cytochrome P-450 steroid hydroxylases, and it is a member of the nuclear hormone receptor family. It is present in the mouse urogenital ridge at the earliest stage of organogenesis. The targeted disruption of Steroidogenic factor 1 (sf-1) in mice prevents the establishment of undifferentiated gonads, and all sf-1 null mutants develop a female phenotype. lim-1, which is part of the LIM-homeodomain subclass of LIM proteins, is also important for urogenital development. lim-1 null mutant mice typically die approximately 10 dpc, and the few that develop to term lack kidneys and gonads. D. Differentiation of Testes In mammals, testes differentiation begins at an earlier stage of development than does the differentiation of ovaries. Testis determination is normally initiated in males by the expression of the SRY gene on the Y chromosome in the bipotential gonad that is common to both males and females (Figure 36.2). In the human, sry expression can first be detected at 41 days. In the mouse, gonads, sry expression can be detected by 10.5 dpc, and testis differentiation occurs approximately 36 hours later, between 12.0 dpc and 12.5 dpc (Hacker et al., 1995). sry expression in gonadal somatic cells initiates the differentiation of Sertoli cells, which are known as the supporting cell lineage of the testis that is essential for its subsequent differentiation. Sertoli cells polarize and aggregate around germ cells to support the growth and maturation of germ cells. After sry expression, sry-related HMG box-9 (sox9),

814

DEVELOPMENT OF THE GENITAL SYSTEM

WT1, SF1 and LIM-1 Male (XY)

a. Indifferent Gonad

Female (XX)

SRY

No SRY

Testis

Testosterone HOX

MIS

Wolffian ducts Differentiation

Ovary

b. Sex determination No Testosterone

Müllerian ducts Degeneration

No MIS

Wolffian ducts Degeneration

Müllerian ducts Differentiation

HOX WNT

No androgen Dihydrotestosterone c. Sex differentiation Male genitalia Development

Female genitalia Development

FIGURE 36.2 Genetic and hormonal factors in the mammalian sex determination and differentiation. a, There are three genes identified as necessary for development of the undifferentiated or bipotential gonad in the mouse: wt1, sf1, and lim-1. b, Testis determination is initiated in males by expression of the sry gene on the Y chromosome; in the absence of SRY expression, the bipotential gonad develops as an ovary. c, Male hormones promote both Wolffian duct and male genitalia development, while Mu¨llerian ducts degenerate in male. In the absence of male hormones, both Mu¨llerian ducts and female genitalia differentiate, while the Wolffian ducts degenerate in female. hox genes are involved in the development genital ducts both in male and female. The wnt gene is involved in female genital duct development.

which is another definitive testis-differentiation gene, is activated (Morais et al., 1996). sox9 upregulation is the earliest marker of pre-Sertoli cells. sry and sox9 expression overlap in Sertoli cells, colocalizing to the nucleus of pre-Sertoli cells as early as 11.5 dpc. As the sox9-expressing population expands between 11.5 dpc and 12.5 dpc, the number of cells that coexpress sry decreases until 12.5 dpc, when sry expression is extinguished and sox9 expression is confined to Sertoli cells inside of the cords. Several male-specific cellular events (e.g., glycogenesis, coelomic epithelium proliferation, mesonephric migration, vasculogenesis) are induced in XY gonads after the onset of sry and sox9 expression. Although the sry gene was discovered more than 10 years ago, the mechanism of how sry functions as a testis-determining factor remains unknown. Heterozygous human sox9 mutations cause campomelic dysplasia, a severe skeletal disorder that involves defective cartilage development. Many of these male patients also have gonadal dysgenesis. Heterozygous mice that are haploinsufficient for sox9 die perinatally as a result of skeletal malformation. Dax1 is an X-linked orphan nuclear receptor that is expressed in the ventromedial hypothalamus, the pituitary gonadotropes, the adrenal cortex, the testis, and the ovary. The duplication of the X-chromosomal-region–spanning Dax1 results in dosage-sensitive, male-to-female sex reversal. Alternatively, Dax1 deficiency influences testis cord formation. It appears that appropriate Dax1 levels are critical for normal testis development: too little or too much of the factor can have an antitestis effect. Recently, Dax1 was reported to function as an early mediator of testis development downstream of sry. The analysis of Dax1 / mice implies that Dax1 functions at an early step downstream of sry or even possibly in a parallel path-

GONADAL DIFFERENTIATION

815

way of sry to establish Sertoli cell differentiation. The data also show that Dax1 may play an important functional role in Leydig cells (Meeks et al., 2003). The lifespan of fetal Leydig cells can be divided into three stages: differentiation, fetal maturity, and regression. Leydig cells differentiate after Sertoli cells during fetal development under their paracrine action. There are two signaling systems that are essential for fetal Leydig cell differentiation: the Desert hedgehog (Dhh)–Patched system and the platelet-derived growth factor (PDGF)-receptor a system (Brennan et al., 2003; Clark et al., 2000). The ligands Dhh and PDGF-A are produced by fetal Sertoli cells, whereas fetal Leydig cells express their cognate receptors (Patched for Dhh and PDGF-Ra for PDGF-A). The recognition of the involvement of these signaling systems in the process of fetal Leydig cell differentiation came from gene inactivation experiments in mice. Genetic analysis has placed PDGF-Ra upstream of Dhh. Mutants of both Dhh and PDGF-Ra have early defects in the partitioning of the testis cord and the interstitial compartment and severely impaired differentiation of fetal Leydig cells. A similar human phenotype was described for a 46,XY patient with a Dhh mutation. In both the mouse knockout and the human mutation, female external genitalia with a blind vagina were observed. Moreover, the Wolffian duct derivatives and prostate were decreased in size, indicating insufficient androgen production by fetal Leydig cells. Anti-Mu¨llerian hormone (AMH), which is also known as Mu¨llerian inhibitor substance, is a key peptide hormone produced by Sertoli cells, and it belongs to the transforming growth factor b family. It is the earliest marker of testis formation. In the human, AMH is secreted by fetal Sertoli cells starting around 8 weeks. In the mouse, AMH transcripts are first present in the pre-Sertoli cells at 12.5 dpc, when these cells are forming cords. This factor acts as an important molecular switch to turn on a network of downstream factors that are involved in testicular development as well as male sex differentiation (Munsterberg et al., 1991). E. Differentiation of Ovaries In mammalian embryos, the testis pathway is the active pathway in gonad development. In mice lacking a Y chromosome (more specifically the sry gene), gonadal development occurs slowly. In the absence of SRY expression, the bipotential gonad develops as an ovary. In the female mouse, the gonads retain an undifferentiated stage longer than in the male. Strings of oocytes form indistinct cords that are observed at 14 dpc. By 15 dpc, these are separated by stromal cell partitions and blood vessels. In humans, the ovary is not positively identifiable until about the tenth week of gestation. Ovarian differentiation depends on the presence of germ cells, so if primordial germ cells fail to reach the genital ridges, streak ovaries are formed. In the absence of oocytes, somatic cells transdifferentiate toward testicular tissue, and this includes the appearance of XX Sertoli cells (Nilsson et al., 2002; Nilsson et al., 2004). Similarly to the testis, the ovary contains primitive sex cords in the medullary region, but these are not as well developed as they are in the testis. In the female, the initial sex cords degenerate, but they are replaced by new epithelial proliferation that results in a new set of sex cords. These cortical sex cells superficially penetrate the mesenchyme, remaining near the outer cortex of the ovary, which is the location of the

816

DEVELOPMENT OF THE GENITAL SYSTEM

female germ cells. Rather than forming an interconnected network, these cords form distinct clusters surrounding germ cells. This is the initial origin of the ovarian follicles. The epithelial sex cords differentiate into granulosa cells, whereas the mesenchymal cells form the thecal cells. Relatively few genes have been shown to be specific to the early female gonadal development. Dax1 was cloned from an X-chromosomal region in humans that was responsible for dosage-sensitive sex reversal. It is specifically expressed in XX gonads after 11.5 dpc, and it was initially suggested as a proovarian or antitestis candidate gene. However, Dax1 loss of function on the XX background does not prevent ovary development. Subsequent studies have shown an expected role for Dax1 in testis development, which indicates that its actions are highly dependent on the timing and level of expression (Bardoni et al., 1994). The existence of sex reversal XX individuals that develop as males in the absence of the sry gene led to the proposal that sry normally represses a factor (Z) that functions at the top of a genetic cascade as a repressor of male development. Thus, according to this theory, the Z factor would be repressed by sry in the male, and it would be independent of sry on an XX genetic background. Loss of a Z factor would be sex-reversing on XX (female-to-male). One candidate for this Z factor is Wingless-related integration site 4 (Wnt-4), which acts as a partial antitestis gene by repressing aspects of male development in the female gonad. In the mouse, Wnt genes are expressed in the mesonephric mesenchyme between 9.5 dpc and 10.5 dpc and in the gonadal mesenchyme of both sexes at 11 dpc. By 11.5 dpc, gonadal Wnt-4 expression is downregulated in the male but maintained in the female. Wnt-4 signaling is required to maintain the female germ line and to suppress the differentiation of Leydig cell precursors (McElreavey et al., 1993). Testicular differentiation can occur in the absence of germ cells, but these cells are essential for the formation and maintenance of follicles in the ovary. In their absence, follicles degenerate into cord-like structures, and XX cells express male markers such as sox9 and AMH (Brennan et al., 2004). In humans, a loss-of-function mutation in Wnt0004 caused Mayer–Rokitansky–Kuster–Hauser syndrome, which is characterized by the defective development of Mu¨llerian derivatives and the duplication of a chromosomal region containing Wnt4; this condition was associated with a case of human XY sex reversal.

III. DEVELOPMENT OF THE GENITAL DUCTS A. The Indifferent Stage Like the gonads, the sexual ducts pass through an early indifferent stage. Unlike the gonads, in which a single tissue is bipotential, the indifferent genital ducts involve two options: the mesonephric (Wolffian) ducts and the paramesonephric (Mu¨llerian ) ducts. Both male and female embryos have two pairs of genital or sexual ducts. These ducts can differentiate into male or female reproductive organs according to the hormonal status of the fetus. In mammalian embryos, the testis secretes several hormones that promote Wolffian duct differentiation into the male reproductive tract. They form the epididymides, the ductus, and the ejaculatory system when the Mu¨llerian ducts

DEVELOPMENT OF THE GENITAL DUCTS

817

degenerate. In the absence of male hormones, the Wolffian ducts degenerate, whereas the Mu¨llerian ducts persist and differentiate into the female internal reproductive tract, which is made up of fallopian tubes, the uterus, and the superior portion of the vagina. The fate of the indifferent genital ducts depends on the gender of the gonad (Figure 36.3). B. Development of the Male Genital Duct Owing to the expression of SRY, the bipotential gonad of males becomes the testis. Hormones play an essential role in regulating male sexual development after the testis has formed. In mammals, this regulation depends on three key hormones produced by the fetal testis: AMH, testosterone, and insulin-like factor 3 (INSL3). In the absence of these critical testicular hormones, female sex differentiation occurs. In males, the Mu¨llerian duct system forms early on but subsequently regresses. The elimination of the Mu¨llerian ducts in the male fetus is driven by AMH, which is a transforming growth factor b superfamily member (Viger et al., 2005). AMH is secreted by Sertoli cells. The expression of AMH starts around 12.5 dpc in the mouse and at around 8 weeks in the human. It is main-

Sexual differentiation of the reproductive system. Before sexual differentiation, both male and female embryos have bipotential gonads, as they possess both Wolffian and Mu¨llerian ducts (a). These ducts can differentiate into male or female reproductive organs according to the hormonal status of the fetus. Owing to the expression of Sry, the bipotential gonad of males becomes the testis, which secretes several hormones including testosterone, MIS, or AMH and Insl3 (b). Testosterone promotes Wolffian duct differentiation into the male reproductive tract, and MIS eliminates the Mu¨llerian ducts. In females, the bipotential gonad becomes the ovary (c). In the absence of male hormones, the Wolffian ducts regenerate, whereas the Mu¨llerian ducts persist and differentiate into the female reproductive tract. (Adapted from Kobayashi and Behringer, 2003.)

FIGURE 36.3

818

DEVELOPMENT OF THE GENITAL SYSTEM

tained throughout fetal development, and it then declines markedly after birth. In XY embryos, AMH regulates male sex differentiation by triggering the regression of the paramesonephric (Mu¨llerian) duct. In the human, the primary role for AMH in sex development is to cause a gradient of cranial-tocaudal regression of the Mu¨llerian ducts during a short period from 8 to 10 weeks of gestation. This is achieved by the protein binding to a similarly expressed gradient of AMH type II receptor in mesenchymal cells, which induces apoptosis of the epithelial cells of the Mu¨llerian ducts. In the human, the absence of AMH expression or an inactivating mutation of the AMH type II receptor gene causes persistent Mu¨llerian duct syndrome in males. In mice, the elimination of the Mu¨llerian duct system in male fetuses is essentially complete by 16.5 dpc. Several transcription factors are involved in the regulation of AMH expression. The first factor shown to be crucial for AMH expression is the orphan nuclear receptor sf-1, which was identified as an essential regulator of the sex steroid synthesis in the adrenal glands and gonads (Shen et al., 1994). In the mouse, the targeted disruption of the sf-1 gene prevents Mu¨llerian duct regression (Luo et al., 1994). In transgenic mice, an intact sf-1 binding site is required for the sex-specific expression of AMH (Giuili et al., 1997). The ability of sf-1 to activate AMH transcription has been demonstrated. This regulation is modulated through direct protein interactions with the factors Wnt1, sox9, and Dax1. 1. Androgens Are Also Essential for Normal Male Sex Differentiation Testosterone is secreted by the Leydig cells of the testes. During fetal life, testosterone promotes virilization of the urogenital tract in two ways. First, it stimulates the mesonephric (Wolffian) ducts to develop and differentiate into the epididymides, the seminal vesicles, and the vasa deferens. Second, in the early urogenital sinus and external genitalia, testosterone is rapidly transformed into dihydrotestosterone by the enzyme steroid 5a-reductase to induce the development of the male urethra, the prostate, the penis, and the scrotum. In humans, 5a-reductase II deficiency is a cause of pseudohermaphroditism. Affected individuals are 46,XY males who have an autosomal-recessive disorder that is characterized by an external female phenotype at birth, bilateral testes, and normally virilized Wolffian structures that terminate in the vagina. Testosterone is also essential for testis descent into the scrotum during fetal development. Testis descent constitutes an essential step in the male sex differentiation process. In mammals, this process follows two distinct and sequential stages: the intra-abdominal stage and the inguinoscrotal stage. The INSL3 or relaxin-like factor, which is a member of the insulin-like hormone superfamily, seems to be required to regulate the intra-abdominal stage of testicular descent, although the mechanism remains poorly understood. INSL3 is expressed early in fetal mouse Leydig cells, and INSL3 knockout male mice are bilaterally cryptorchid; the gubernacular bulbs fail to develop, and they resemble normal female gubernacular structures. Evidence shows that the steroid hormones estradiol and diethylstilbestrol could downregulate INSL3 expression in the fetal Leydig cells (Nef et al., 2000). Hox genes encode homeodomain proteins that act as transcriptional regulators. Hox genes have a well-characterized role in embryonic development, during which they determine identity along the anterior–posterior body axis (see also the chapter by Kenyon in this book). In mice and humans, Hox genes are clustered in four unlinked genomic loci (designated Hoxa-d or HOXA-D),

DEVELOPMENT OF THE GENITAL DUCTS

819

which contain a subset of nine to thirteen genes each. The role of mammalian Hox genes in regulating segmental patterning of axial structures and the limb is well established. Hox genes also play a similar role in the specification of developmental fate in the individual regions of the male reproductive tract. Hoxa-10 is expressed along the mesonephric duct from the caudal epididymis to the point at which the ductus deferens inserts into the urethra. Mutants of both Hoxa-10 and Hoxa-11 exhibit a homeotic transformation that results in the partial transformation of the ductus deferens to the epididymis (Bomgardner et al., 2003; Podlasek et al., 1999). Hoxa-13 and Hoxd-13 are the 50 -most members of the Hox A and D clusters. Hoxa-13 is expressed in the terminal part of the digestive and urogenital tracts during embryogenesis, specifically in the genital tubercle from which the male accessory sex organs derive (Dolle et al., 1991; Warot et al., 1997). A Hoxd-13 loss-of-function mutant demonstrated that Hoxd-13 is essential to external genital development. Hoxd-13 is the most caudally expressed Hox gene in the genitourinary tract. It is expressed in both the mesenchyme and the epithelium of the Wolffian duct and the urogenital sinus (Fromental-Ramain et al., 1996; Oefelein et al., 1996). Another homeobox gene, emx2, which is a mammalian homologue of the Drosophila empty spiracles, is also expressed in the epithelial component of the intermediate mesoderm. emx2-/- mutants completely lack reproductive tracts and gonads. In the mutant embryos, the Wolffian duct forms on embryonic day 10.5, but it subsequently degenerates on embryonic day 11.5. emx2 expression is only detected during the early stages of reproductive duct formation. It suggests that this gene is only required for a specific time period during the development of the intermediate mesoderm, possibly providing a survival signal. C. Development of the Female Genital Duct In contrast with the Wolffian duct, the Mu¨llerian duct persists in the absence of external signals, and it must be directed to degenerate in males. In females, the absence of sry results in ovarian organization. In the absence of male hormones, the Wolffian ducts degenerate, whereas the Mu¨llerian ducts persist and differentiate into the female reproductive tract, including the oviduct (fallopian tube), the uterus, and the upper portion of the vagina. Several genes are involved in the development of female genital duct development. Although few genes have been identified in humans, mouse knockout studies have shed light on a set of genes that are essential for the regulation of Mu¨llerian duct formation. The Wnt gene family, which is homologous to the Drosophila Segment polarity gene Wingless, encodes secreted glycoproteins. They are involved in sex determination and the development of several female reproductive organs. In mice, a subset of Wnt gene family members has been identified to regulate Mu¨llerian duct development. For example, Wnt0004 expression is crucial for the formation of the Mu¨llerian ducts, because it is required for tubule formation. Both Wnt4-deficient male and female mice completely lack Mu¨llerian ducts, and Wnt4-mutant females even differentiate a normal male reproductive tract. Although Wnt0004 initiates Mu¨llerian duct formation, Wnt7a may regulate its further outcome. It is expressed in the Mu¨llerian duct epithelium in both sexes from 12.5 dpc to 14.5 dpc. After the Mu¨llerian duct regression, expression is lost in males. In females, it persists in the epithelium of the Mu¨l-

820

DEVELOPMENT OF THE GENITAL SYSTEM

lerian duct derivatives throughout life (Heikkila et al., 2001). From Wnt7a mutant female mice studies, it has been learned that Wnt7a is required for the proper differentiation of the oviduct and the uterus. In Wnt7a mutant male mice, the Mu¨llerian ducts do not regress. This suggests that Wnt genes have a function not only in Mu¨llerian duct formation but also in the regression of these ducts in males. Another member of this family, Wnt5a, is also critical for female reproductive tract development; it is needed for the formation of the genital tubercle. Wnt5a-deficient female have a coiled and shortened uterus and a poorly defined cervix and vagina. The Hox gene family has also been identified to regulate female reproductive tract development. As described previously, Hox genes play a role in male genital duct development. This chapter will now address how Hox gene expression also directs Mu¨llerian duct differentiation and pattern formation. In the developing Mu¨llerian duct, a number of posterior Abdominal B (Abd B) homeobox genes were found to be expressed in partially overlapping patterns along the anterior–posterior axis. Abd B genes are expressed according to their 30 -50 order in the Hox clusters: Hoxa-9 is expressed in the oviduct, Hoxa-10 is expressed in the developing of the uterus, Hoxa-11 is found in the primordia lower uterus and cervix, and Hoxa-13 is seen in the upper vagina (Figure 36.4; Taylor et al., 1997). The targeted mutagenesis of these genes results in region-specific defects along the female reproductive tract. Hoxa10 deficiency causes the homeotic transformation of the anterior part of the uterus into an oviduct-like structure, and it also causes reduced fertility in females. Hoxa-13 null embryos show agenesis of the posterior portion of the Mu¨llerian duct. When the Hoxa-11 gene is replaced by the Hoxa-13 gene, posterior homeotic transformation occurs in the female reproductive tract: the uterus, in which Hoxa-11 but not Hoxa-13 is normally expressed, becomes similar to the more posterior cervix and vagina, in which Hoxa-13 is normally expressed. It is also revealed that Wnt7a is required for the maintenance of

FIGURE 36.4 Expression of HOX gene is arranged in the linear fashion along the paramesonephric duct. Hox9 is expressed in areas destines to become fallopian tube. Hoxa10 is expressed in the developing uterus. Hoxa11 is expressed in the primordia of the lower uterine segment and cervix and upper vagina in the developing uterus and cervix. Hoxa13 is expressed in the upper vagina. (Adapted from Taylor, 2000. See color insert.)

DEVELOPMENT OF THE EXTERNAL GENITALIA

821

Hoxa-10 and Hoxa-11 expression in the uterus in mice. Wnt5a resides in the same genetic pathway as Wnt7a and Hoxa genes during female reproductive tract development, and emx2 also plays a role in female reproductive tract development. In contrast with the Wolffian duct, the Mu¨llerian duct never forms in emx2 mutants. Genes such as lim1 and Paired-box gene 2 (pax2) are also indispensable for the early steps of Mu¨llerian duct development. lim1 plays an essential role in mouse head and urogenital system development. It has a dynamic expression pattern in the Mu¨llerian duct as early as embryonic day 11.5, which suggests that lim1 function is crucial for the initial formation of the Mu¨llerian duct. The analysis of lim1 null mice revealed that Mu¨llerian duct derivatives were completely absent. Furthermore, no Wolffian duct derivatives were ever observed in lim1 null neonatal males, thus demonstrating that lim1 is required for the formation of both sexual ducts. pax2, which is a member of the Pax gene family, encodes a homeodomain transcription factor that is homologous to the Drosophila pair-rule gene Paired. pax2 null mice also lack a reproductive tract. Unlike the lim1 null embryos, pax2 null mutants present both Wolffian and Mu¨llerian duct formation, but these subsequently degenerate. This phenotype correlates with pax2 expression at this stage in both reproductive tracts on embryonic day 13.5, thus indicating a cell-autonomous role for pax2 in the developing Wolffian and Mu¨llerian ducts. The development of the female reproductive tract also depends on estrogenic hormones by the fetal ovaries. Estrogen action is mediated by estrogen receptors (ERs), which belong to the nuclear receptor superfamily of ligandinducible transcription factors. In fetal mice, an ERa signal was detected in the nuclei of the surrounding cells of the Mu¨llerian ducts as early as 11.5 dpc. No expression of ERb was detected in the mouse Mu¨llerian duct. However, ERbs were detected in rat Mu¨llerian ducts from 15.5 dpc to 21.5 dpc. The expression levels of ERbs are much lower than those of ERas. The studies show that ERa is likely a dominant ER subtype that is used in Mu¨llerian duct development (Greco et al., 1991). ERa knockout mice have a small but normally patterned reproductive tract.

IV. DEVELOPMENT OF THE EXTERNAL GENITALIA A. The Indifferent Stage The external genitalia also pass through an undifferentiated state before distinguishing sexual characteristics appear. The external genitalia are derived from a complex of mesodermal tissue located around the cloaca. In the human, a very early midline elevation called the genital eminence is situated just cephalic to the proctodeal depression. This structure soon develops into a prominent genital tubercle at the cranial end of the cloacal membrane early during the fourth week. Genitalia swellings and urogenital folds soon develop on each side of the cloacal membrane. The development of the genital tubercle is initially regulated by Hox gene expression. Located at the terminal part of the urogenital system, the genital tubercle expresses the 50 -most genes from the Hox gene clusters, specifically Hoxa-13 and Hoxd-13. The early phase of outgrowth of the genital tubercle also depends on the interacting signals of Sonic hedgehog and the fibroblast growth factors 8 and 10 (Carlson, 2004).

822

DEVELOPMENT OF THE GENITAL SYSTEM

B. External Genitalia The masculinization of the indifferent external genitalia is caused by androgens produced by the testis. Unlike the internal genitalia in the male, the external genitalia do not respond directly to testosterone. 5a-Reductase converts testosterone to dihydrotestosterone in the external genitalia. Under the influence of dihydrotestosterone, the genital tubercle in the male undergoes a second phase of elongation to form the penis, and the genital swellings enlarge to form the scrotal pouches. 5a-Reductase plays an important role in external genitalia development. As described previously, 5a-reductase II deficiency will cause male patients to have an external female phenotype at birth and to exhibit normally virilized Wolffian structures that terminate in the vagina. In the absence of androgen signaling, the feminization of the indifferent external genitalia occurs. The genital tubercle in the female becomes the clitoris; the genital folds become the labia minora, and the genital swellings develop into the labia majora.

V. MALFORMATIONS OF THE GENITAL SYSTEM Malformations of the genital system are intrinsic defects in the developing human embryo that result in localized abnormalities during the development of the reproductive duct system. Genetic events can result in congenital genital malformations during very early development stages. In the human, numerous factors can also affect the development of the reproductive tract, such as infectious agents, drugs or pharmaceutical products, environmental chemicals, physical agents, and maternal diseases. Despite their different origins, all of these factors cause some type of genetic abnormalities that ultimately induce genital malformation. As described previously, genetic abnormalities occurring in animal models, such as chromosomal anomalies and gene mutations, influence the development of the genital system, thereby causing congenital malformations. A. Abnormalities of Sexual Differentiation 1. Turner Syndrome (Gonadal Dysgenesis) Turner syndrome is characterized by defective gonadal development in women with a karyotypic sex chromosome abnormality (45,X or 45,XO). These individuals are phenotypic females. Individuals with this syndrome possess primordial germ cells that degenerate shortly after they reach the gonads. Affected individuals generally are of short stature, and they present with undifferentiated (streak) gonads. As expected, the internal and external reproductive structure develops as female as a result of the absence of AMH and testosterone. 2. Swyer Syndrome Swyer syndrome, which is also known as XY gonadal dysgenesis, is a heterogenous condition with variant forms that are caused, in most cases, by a structural abnormality on the Y chromosome that leads to sry loss of function. Swyer syndrome has also been associated with autosomal mutations such as chromosome 9p deletions. Patients with Swyer syndrome are born

MALFORMATIONS OF THE GENITAL SYSTEM

823

without functional gonads; instead, they present simply with gonadal streaks. Affected individuals are phenotypically females at birth. However, because the streak gonads are incapable of producing the sex hormones that are essential for puberty, these patients do not develop most secondary sex characteristics without hormone replacement. 3. True Hermaphroditism Hermaphroditism is a rare condition in which ovarian and testicular tissues exist in the same person. The testicular tissue contains seminiferous tubules and spermatozoa, and the ovarian tissue contains follicles or corpora albicantia. Hermaphrodite patients typically show a chromosomal male– female mosaicism in which both the male XY and the female XX chromosome pairs are present. External genitalia may show traits from both sexes. 4. Female Pseudohermaphroditism Female pseudohermaphroditism is characterized by male or ambiguous genitalia coupled with a female karyotype (46,XX). The XX male syndrome is a heterogeneous disorder. The presence of the sry gene transposed with the X chromosome leads to male differentiation. About 80% of XX males express sry. Some cases of sry-negative XX males have been reported. This may be the result of an unrecognized XX/XY chimerism or an XX/XXY mosaicism, although an autosomal-recessive disorder has also been proposed as the intrinsic cause of these less common cases. 5. Male Pseudohermaphroditism Male pseudohermaphroditism refers to a condition that affects 46,XY individuals with differentiated testes who exhibit varying degrees of feminization. In cases of male pseudohermaphroditism, there is a spectrum of external genitalia; some individuals are completely phenotypically female, whereas others appear to be normal males with varying spermatogenesis and/or pubertal virilization. Between these two extremes is a wide area of ambiguity. Deficiency of the enzyme 5a reductase can result in external genitalia that appear to be female in an individual with an XY karyotype. High testosterone production at puberty can drive male external differentiation and an apparent “sex change” during adolescence. 6. Testicular Feminization (Androgen Insensitivity) Syndrome Individuals with testicular feminization syndrome have a normal XY chromosomal complement; however, they are resistant to androgens (testosterone). This usually results from a mutation in the androgen receptor, which consequently leads to some extent of—or even total—external genitalia feminization. Complete testicular feminization results in an individual who looks outwardly female. Because these individuals produce AMH, the Mu¨llerian ducts degenerate; however, the Wolffian ducts lack the ability to respond to testosterone, and they therefore regress. Patients with androgen insensitivity have no internal genitalia. The vagina in these individuals is a short structure that lacks communication with any internal organ. In these patients, breast development proceeds normally as a result of the lack of androgens; however, they still have the ability to convert testosterone to estrogen.

824

DEVELOPMENT OF THE GENITAL SYSTEM

VI. Vestigial Structures from the Embryonic Genital Ducts A. Mesonephric Duct Remnants Mesonephric duct remnants are also known as Gartner’s ducts. In females, the remains of the cranial parts of the mesonephros may persist as the epoophoron or the paroophoron. The caudal parts of the mesonephric ducts are often seen in histologic sections along the uterus or the upper vagina as Gartner’s ducts. Portions of these duct remnants sometimes enlarge to form cysts. B. Paramesonephric Duct Remnants Remnants of the paramesonephric (Mu¨llerian) ducts can be found in the male as a uterus-like structure. Historically, this has been called a masculine uterus. The remnants appear as one or two thin, uterus-like tubes that are medial to the ducti deferentes, with or without a medial corpus lying between the ampullae. The paramesonephric remnants resemble a normal female uterus, but the endometrium consists primarily of amorphous extracellular matrix.

VII. OTHER ABNORMALITIES OF THE GENITAL DUCT SYSTEM A. Failed Mu¨llerian Duct Fusion or Cannulation with Clinical Correlation Normal uterine development depends on both the fusion of the two paramesonephric ducts and the absorption of the fused walls to create a single cavity. Failed Mu¨llerian duct fusion occurs in females when the ducts do not meet or fuse. Fusion abnormalities can occur at any point along the Mu¨llerian ducts; they may involve an isolated junction or the entire duct. The condition can have a range of results, from a small branch of the apex of the duct being affected to a complete duplication of the uterus into separate structures, each with a single fallopian tube. Failure to absorb the intervening wall between the two fused paramesonephric ducts results in a septum that separates the canals of a fused uterus. Both fusion defects and a septum can result in pregnancy complications such as miscarriage and premature delivery. B. Congenital Absence of the Vas Deferens The congenital bilateral absence of vas deferens occurs in males when the tubes that carry sperm from the testes (the vas deferens) fail to develop normally. This condition can occur alone or as a sign of cystic fibrosis. The testes usually develop and function normally, but sperm cannot be transported out of the epididymis. This condition occurs in men with a cystic fibrosis gene mutation; however, they often do not have any of the other health problems associated with that disease (e.g., progressive lung damage, chronic digestive system problems).

VIII. CONCLUSION Reproductive tract development follows from an ordered set of divergent signals that begins with chromosomal complement. It is one of the few instances of such a clear bimodal heterogeneity in development. The adaptation

825

SUMMARY

of conservation of these mechanisms in multiple organisms attests to their evolutionary value. The whole development includes three stages: (1) sex determination; (2) the differentiation of the internal genital ducts; and (3) the differentiation of the external genitalia. Interestingly, there are indifferent stages before the differentiation of distinguishing sexual characteristics in which numerous genetic factors are involved. The reasons that development requires this indifferent stage and the ways in which the required genes are initiated are still unclear. There are some genes, such as Hox, lim1, and pax2, that are involved in the development of the genital ducts in both males and females. How do they direct these two different structures? How is the expression of the sry gene initiated in the XY embryo? Finally, although the traditional dogma considers that the migration of PGCs occurs via self-movement, recent new theories question this mode of cell movement. Many questions in this field remain to be answered, and the exploitation of new genetic and genomic information will be critical in the answering of these important questions.

SUMMARY

 In mammals, sex determination is accomplished by a chromosomal mech-

 







anism. It begins at the time of fertilization through the coupling of two gametes: either two X chromosomes (XX in females) or an X and a Y chromosome (XY in males). The sry gene is the sex determination gene of the Y chromosome. Gonadal differentiation begins after the migration of the PGCs into the indifferent gonad. Testis determination is normally initiated in males by the expression of the sry gene. In the absence of sry expression, the bipotential gonad develops as an ovary. Ovarian differentiation is dependent on the presence of germ cells. However, germ cells are not necessary for testicular differentiation. The indifferent genital ducts consist of the mesonephric (Wolffian) ducts and the paramesonephric (Mu¨llerian) ducts. In mammalian embryos, the testis secretes several hormones, such AMH, testosterone, and INSL3, that promote Wolffian duct differentiation into the male reproductive tract, whereas the Mu¨llerian ducts degenerate in males. In the absence of the male hormones, the Wolffian ducts degenerate, whereas the Mu¨llerian ducts persist and differentiate into the female internal reproductive tract. The fate of the indifferent genital ducts depends on the gender of the gonad. Genital duct development is a hormonally and genetically, controlled process. Hox gene family members are involved in the development of genital ducts in both males and females. Along the anterior–posterior axis of the genital duct, Hox genes are expressed according to their 30 -50 order in the Hox clusters. Wnt gene family members are also involved in both female gonadal differentiation and female genital duct development. The external genitalia pass through an undifferentiated state before distinguishing sexual characteristics appear. The development of the genital tubercle is initially regulated by Hox gene expression. Because the genital tubercle is located at the terminal part of the urogenital system, it expresses the 50 -most genes from the Hox gene clusters, specifically Hoxa-13 and Hoxd-13.

826

DEVELOPMENT OF THE GENITAL SYSTEM

 Malformations of the genital system are intrinsic defects in the developing human that result in localized abnormalities during the development of the reproductive duct system. The genetic abnormalities, such as chromosomal anomalies and gene mutations, influence the development of the genital system and cause malformations.

GLOSSARY Mu¨llerian ducts Two embryonic tubes that extend along the mesonephros that become the uterine tubes, the uterus, and part of the vagina in the female and that form the prostatic utricle in the male. Also known as paramesonephric ducts. Sex chromosome Either of a pair of chromosomes (usually designated as X or Y) in the germ cells of most animals and some plants that combine to determine the sex and sex-linked characteristics of an individual. In mammals, XX results in a female and XY results in a male. Sex determination The process by which the sex of an organism is determined. In many species, the sex of an individual is dictated by the two sex chromosomes (X and Y) that it receives from its parents. In mammals, some plants, and a few insects, males are XY and females are XX; in birds, reptiles, some amphibians, and butterflies, the reverse is true. In bees and wasps, males are produced from unfertilized eggs, and females are produced from fertilized eggs. In 1991, sry was identified as the sex-determination gene of the Y chromosome. Environmental factors can also affect sex determination in some fish and reptiles. In turtles, for example, sex is influenced by the temperature at which the eggs develop. Primordial germ cell An embryonic cell that gives rise to a germ cell from which a gamete (i.e., an egg or a sperm) develops. Wolffian duct The duct in the embryo that drains the mesonephric tubules. It becomes the vas deferens in the male, and it forms vestigial structures in the female. Also known as the mesonephric duct.

REFERENCES Bardoni B, Zanaria E, Guioli S, et al: A dosage sensitive locus at chromosome Xp21 is involved in male to female sex reversal, Nat Genet 7:497–501, 1994. Bomgardner D, Hinton BT, Turner TT: 50 hox genes and meis 1, a hox-DNA binding cofactor, are expressed in the adult mouse epididymis, Biol Reprod 68:644–650, 2003. Brennan J, Capel B: One tissue, two fates: molecular genetic events that underlie testis versus ovary development, Nat Rev Genet 5:509–521, 2004. Brennan J, Tilmann C, Capel B: Pdgfr-alpha mediates testis cord organization and fetal Leydig cell development in the XY gonad, Genes Dev 17:800–810, 2003. Burgoyne PS: The mammalian Y chromosome: a new perspective, Bioessays 2:363–366, 1998.

REFERENCES

827 Carlson BM: Urogenital system, In Ozols I, Milnes J, editors: Human embryology and developmental biology, Philadelphia, 2004, Elsevier 393–419. Cheng HH, Ying M, Tian YH, et al: Transcriptional diversity of DMRT1 (dsx- and mab3-related transcription factor 1) in human testis, Cell Res 16:389–393, 2006. Clark AM, Garland KK, Russell LD: Desert hedgehog (Dhh) gene is required in the mouse testis for formation of adult-type Leydig cells and normal development of peritubular cells and seminiferous tubules, Biol Reprod 63:1825–1838, 2000. Dolle P, Izpisua-Belmonte JC, Brown JM, et al: HOX-4 genes and the morphogenesis of mammalian genitalia, Genes Dev 5:1767–1776, 1991. Freeman B: The active migration of germ cells in the embryos of mice and men is a myth, Reproduction 125:635–643, 2003. Fromental-Ramain C, Warot X, Messadecq N, et al: Hoxa-13 and Hoxd-13 play a crucial role in the patterning of the limb autopod, Development 122:2997–3011, 1996. Giuili G, Shen WH, Ingraham HA: The nuclear receptor SF-1 mediates sexually dimorphic expression of Mu¨llerian inhibiting substance, in vivo, Development 124:1799–1807, 1997. Greco TL, Furlow JD, Duello TM, et al: Immunodetection of estrogen receptors in fetal and neonatal female mouse reproductive tracts, Endocrinology 129:1326–1332, 1991. Hacker A, Capel B, Goodfellow P, et al: Expression of Sry, the mouse sex determining gene, Development 121:1603–1614, 1995. Heikkila M, Peltoketo H, Vainio S: Wnts and the female reproductive system, J Exp Zool 290:616–623, 2001. Kobayashi A, Behringer RR: Developmental genetics of the female reproductive tract in animals, Nat Rev Genet 4:969–980, 2003. Koopman P, Munsterberg A, Capel B, et al: Expression of a candidate sex-determining gene during mouse testis differentiation, Nature 348:450–452, 1990. Lawson KA, Dunn NR, Roelen BA, et al: Bmp4 is required for the generation of primordial germ cells in the mouse embryo, Genes Dev 13:424–436, 1999. Luo X, Ikeda Y, Parker KL: A cell-specific nuclear receptor is essential for adrenal and gonadal development and sexual differentiation, Cell 77:481–490, 1994. Matsuda M: Sex determination in fish: lessons from the sex-determining gene of the teleost medaka, Oryzias latipes, Dev Growth Differ 45:397–403, 2003. McElreavey K, Vilain E, Abbas N, et al: A regulatory cascade hypothesis for mammalian sex determination: SRY represses a negative regulator of male development, Proc Natl Acad Sci U S A 90:3368–3372, 1993. McLaren A, Simpson E, Tomonari K, et al: Male sexual differentiation in mice lacking H-Y antigen, Nature 312:552–555, 1984. Meeks JJ, Russell TA, Jeffs B, et al: Leydig cell-specific expression of DAX1 improves fertility of the Dax1-deficient mouse, Biol Reprod 69:154–160, 2003. Morais da Silva S, Hacker A, Harley V, et al: Sox9 expression during gonadal development implies a conserved role for the gene in testis differentiation in mammals and birds, Nat Genet 14:62–68, 1996. Motro B, van der Kooy D, Rossant J, et al: Contiguous patterns of c-kit and steel expression: analysis of mutations at the W and Sl loci, Development 113:1207–1221, 1991. Munsterberg A, Lovell-Badge R: Expression of the mouse anti-Mu¨llerian hormone gene suggests a role in both male and female sexual differentiation, Development 113:613–624, 1991. Nef S, Shipman T, Parada LF: A molecular basis for estrogen-induced cryptorchidism, Dev Biol 224:354–361, 2000. Nilsson EE, Kezele P, Skinner MK: Leukemia inhibitory factor (LIF) promotes the primordial to primary follicle transition in rat ovaries, Mol Cell Endocrinol 188:65–73, 2002. Nilsson EE, Skinner MK: Kit ligand and basic fibroblast growth factor interactions in the induction of ovarian primordial to primary follicle transition, Mol Cell Endocrinol 214:19–25, 2004. Oefelein M, Chin-Chance C, Bushman W: Expression of the homeotic gene Hox-d13 in the developing and adult mouse prostate, J Urol 155:342–346, 1996. Page DC, Mosher R, Simpson EM, et al: The sex-determining region of the human Y chromosome encodes a finger protein, Cell 51:1091–1104, 1987. Palmer MS, Sinclair AH, Berta P, et al: Genetic evidence that ZFY is not the testis-determining factor, Nature 342:937–939, 1989. Podlasek CA, Seo RM, Clemens JQ, et al: Hoxa-10 deficient male mice exhibit abnormal development of the accessory sex organs, Dev Dyn 214:1–12, 1999.

828

DEVELOPMENT OF THE GENITAL SYSTEM

Pritchard-Jones K, Fleming S, Davidson D, et al: The candidate Wilms0 tumour gene is involved in genitourinary development, Nature 346:194–197, 1990. Saito T, Kuang JQ, Bittira B, et al: Xenotransplant cardiac chimera: immune tolerance of adult stem cells, Ann Thorac Surg 74:19–24, 2002. Schartl M: Sex chromosome evolution in non-mammalian vertebrates, Curr Opin Genet Dev 14:634–641, 2004. Shawlot W, Behringer RR: Requirement for Lim1 in head-organizer function, Nature 374:425–430, 1995. Shen WH, Moore CC, Ikeda Y, et al: Nuclear receptor steroidogenic factor 1 regulates the Mu¨llerian inhibiting substance gene: a link to the sex determination cascade, Cell 77:651–661, 1994. Shetty S, Griffin DK, Graves JA: Comparative painting reveals strong chromosome homology over 80 million years of bird evolution, Chromosome Res 7:289–295, 1999. Sinclair AH, Berta P, Palmer MS, et al: A gene from the human sex-determining region encodes a protein with homology to a conserved DNA-binding motif, Nature 346:240–244, 1990. Sinclair AH, Berta P, Palmer MS, et al: A gene from the human sex-determining region encodes a protein with homology to a conserved DNA-binding motif, Nature 346:240–244, 1990. Sultan C, Lobaccaro JM, Medlej K, et al: SRY and male sex determination, Horm Res 36:1–3, 1991. Taylor HS: The role of Hox genes in human implantation, Hum Reprod Update 6:75–79, 2000. Taylor HS, Vandenheuvel GB, Igarashi P: A conserved Hox axis in the mouse and human female reproductive system - late establishment and persistent adult expression of the Hoxa cluster genes, Biol Reprod 57:1338–1345, 1997. Tiepolo L, Zuffardi O: Localization of factors controlling spermatogenesis in the nonfluorescent portion of the human Y chromosome long arm, Hum Genet 34:119–124, 1976. Viger RS, Silversides DW, Tremblay JJ: New insights into the regulation of mammalian sex determination and male sex differentiation, Vitam Horm 70:387–413, 2005. Wachtel SS, Ono S, Koo GC, et al: Possible role for H–Y antigen in the primary determination of sex, Nature 257:235–236, 1975. Warot X, Fromental-Ramain C, Fraulob V, et al: Gene dosage-dependent effects of the Hoxa-13 and Hoxd-13 mutations on morphogenesis of the terminal parts of the digestive and urogenital tracts, Development 124:4781–4791, 1997. Welshons WJ, Russell LB: The Y-chromosome as the bearer of male determining factors in the mouse, Proc Natl Acad Sci U S A 45:560–566, 1959. Ying Y, Qi X, Zhao GQ: Induction of primordial germ cells from murine epiblasts by synergistic action of BMP4 and BMP8B signaling pathways, Proc Natl Acad Sci U S A 98:7858–7862, 2001. Zarkower D: Establishing sexual dimorphism: conservation amidst diversity? Nat Rev Genet 2:175–185, 2001.

FURTHER READING Nanda I, Haaf T, Schartl M, et al: Comparative mapping of Z = Orthologous genes in vertebrates: implications for the evolution of avian Sex chromosomes, Cytogenet Genome Res 99:178– 184, 2002. Schartl M: Sex Chromosome evolution in non-mammalian vertebrates, Curr Opin Genet Dev 14:634–641, 2004.

RECOMMENDED RESOURCES The Urogenital System: http://sprojects.mmi.mcgill.ca/embryology/ug/ Human Embryology, Genital System: http://www.embryology.ch/anglais/ugenital/planmodgenital.html Developmental Biology: 6th ed.: http://www.ncbi.nlm.nih.gov/books/bv.fcgi?rid=dbio Recent Progress in Hormone Research: http://rphr.endojournals.org/cgi/reprint/57/1/1 Molecular Regulation of Mu¨llerian Development by Hox Genes: http://www.annalsnyas.org/cgi/ content/full/1034/1/152 Genes in Genital Malformations and Male Reproductive Health: http://www.cbra.org.br/pages/ publicacoes/animalreproduction/issues/download/AR012.pdf

37

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS KATE G. ACKERMAN1–3 and DAVID R. BEIER1,2 1

Division of Genetics, Brigham and Women’s Hospital; Harvard Medical School; 3 Department of Medicine, Division of Emergency Medicine, Children’s Hospital, Boston, MA 2

INTRODUCTION A normal diaphragm is required for breathing and for normal pulmonary development in humans. Diaphragmatic defects are relatively common congenital disorders, and they have a significant impact on families and the health care system, because they frequently result in perinatal lethality or long-term chronic disease (West and Wilson, 2005). The most severe diaphragmatic defects result from the abnormal development of the diaphragm during early gestation. These defects, which are most commonly referred to as congenital diaphragmatic hernias (CDHs), are associated with abnormal pulmonary development, and they occur as frequently as 1 in every 2500 live births (Skari et al., 2000). Death or long-term complications are usually the consequences of severe respiratory insufficiency from abnormal pulmonary development and pulmonary hypoplasia. The reported mortality rate is variable, because tertiary care centers are unable to account for neonates who die before transport. Despite advances in medical therapies, population-based studies measuring outcome from the time of antenatal diagnosis report mortality rates of 58% to 79% (Beresford and Shaw, 2000; Skari et al., 2000; Stege et al., 2003). Survival is improved with an increased antenatal termination rate and with diaphragmatic defects that are not associated with other congenital anomalies (Stege et al., 2003). The normal embryogenesis of the diaphragm is not well understood, and many different abnormal diaphragmatic phenotypes have been observed in Principles of Developmental Genetics © 2007, Elsevier Inc. All rights reserved.

829

830

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

humans. Although the defects associated with severe pulmonary hypoplasia usually occur in the posterior regions of the diaphragm or involve the entire hemidiaphragm, abnormalities of the anterior diaphragm are often asymptomatic and present as incidental findings in the older child or adult. Posterior diaphragmatic hernias are isolated defects in 60% of cases, whereas other affected individuals have complex CDH (i.e., they have additional malformations) or syndromic CDH (i.e., they have a constellation of anomalies that matches a syndromic pattern). Diaphragmatic defects are considered to be complex genetic disorders, and they likely occur as a result of a variety of mechanisms, including cytogenetic aberrations and de novo point mutations in important developmental pathway genes (Slavotinek, 2005). Heritability measured by sibling precurrence or twin studies is low, and transmission rates from affected to offspring are unknown, because low survival rates have precluded reproduction (Pober et al., 2005). This chapter will review our current understanding of the normal and abnormal embryogenesis of the diaphragm. Until recently, the interpretation of diaphragmatic development has come from detailed anatomic evaluation in humans. During the 1980s, a teratogenic model of congenital diaphragmatic defects in rodents became widely available, which provided a valuable research tool (Costlow and Manson, 1981). Over the past 15 years, mouse mutants have provided us with more sophisticated methods of studying diaphragmatic development, and advances in genetic research in humans has allowed for candidate locus analysis and gene discovery in affected patients (Slavotinek, 2005). These complementary research strategies have begun to provide insight into the fundamental mechanisms of diaphragmatic development and its perturbation in human congenital disease.

I. DIAPHRAGMATIC ANATOMY A. Anatomy of the Diaphragm The diaphragm separates the thoracic cavity from the abdominal cavity, and it serves as our major respiratory muscle. Like other muscular structures, it is comprised of connective tissue, skeletal muscle, nerves, and blood vessels. The basic anatomy of the mature diaphragm is shown in Figure 37.1, A. Structures that pass through the diaphragm include the esophagus and the inferior vena cava. The esophageal orifice is surrounded by muscle, which helps to maintain the gastroesophageal junction. The muscle of the diaphragm attaches to the body wall. Anterior midline muscle is present in two distinct bundles that attach to the xiphoid process. Areas of decreased muscularization of the peripheral diaphragm occur between and lateral to these anterior muscle bundles and in the posterolateral regions, the latter regions are often called the lumbocostal triangles. Diaphragmatic defects may occur in any region of the diaphragm (see Figure 37.1). B. Human Developmental Diaphragmatic Defects Human diaphragmatic defects should be characterized by type and location. Historically, however, they have been grouped under names that comprise a variety of defects that may or may not have similar embryologic mechanisms.

DIAPHRAGMATIC ANATOMY

831

FIGURE 37.1 Mouse diaphragms showing normal diaphragmatic structure and the location of diaphragmatic defects. (See color insert.)

1. Type The terminology used to describe the types of diaphragmatic defects has not been standardized. For example, the term eventration technically means “the upward displacement of organs into the thorax.” This could arise from a lack of normal muscularization in a region, from a weakness of the connective tissue component of the diaphragm, or from a diaphragm lacking function as a result of the loss of phrenic nerve innervation. Defects may occur as a hole in the diaphragm without a membrane (i.e., hernia); they may have a thin but highly attenuated membrane (i.e., sac hernia), or they may occur as muscularization defects that cause mild herniation (i.e., eventration). The most common severe diaphragmatic defects occur in the posterior region and usually do not have a membrane or sac. Diaphragmatic defects occurring in the anterior regions almost always have a sac. The embryologic differences between these types are not understood. 2. Location a. Posterior Posterior or posterolateral hernias are often called CDHs or Bochdalek hernias. The name Bochdalek hernia was adopted after Dr. Victor Bochdalek described the posterolateral hernias in 1848 (Irish et al., 1996). Dr. Bochdalek hypothesized that these hernias occurred through the vulnerable lumbocostal triangle; this may be a mechanism for some (but not all) posterior hernias. Most human posterolateral defects are thought to arise at the site of fusion of the pleuroperitoneal fold tissue with the septum transversum tissue, causing a failure to obliterate the pleural canals or from a deficiency in pleuroperitoneal tissue,

832

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

which then fails to fuse to the septum transversum (Sweeney, 1998; Babiuk et al., 2003). Although there is limited evidence to support this presumptive mechanism, it is consistent with the examination of some defects in humans that retain posterolateral diaphragmatic tissue and that have intact lumbocostal triangles. The mechanisms of hernias in the posteromedial location are unknown. The multiple phenotypes of diaphragmatic defects in the posterior regions require careful evaluation so that a more accurate nomenclature system may be developed. b. Anterior Hernias occurring in the anterior regions of the diaphragm are often classified as Morgagni hernias. Giovanni Battista Morgagni described various types of anterior hernias during the 1700s (Irish et al., 1996). Classically, the Morgagni hernia occurs through the Morgagni foramina, which are just lateral to the anterior muscle bundles that attach to the xiphoid process. Some use the term Morgagni hernia to describe herniation through the right-sided foramen and the term Larrey hernia to refer to herniation on the left side. Others classify hernias on either side or across the midline as Morgagni– Larrey or just Morgagni hernias (Salman et al., 1999; Loong and Kocher, 2005). c. Central or other Hernias involving the anterior central tendon are often called septum transversum hernias (Paci et al., 2005), as are midline hernias that extend into the ventral wall, which are usually associated with Pentalogy of Cantrell (Wesselhoeft and DeLuca, 1984). Hernias may also occur laterally or in the anterolateral diaphragm. It is unknown how these defects should be classified, although many are large and have been associated with pulmonary hypoplasia. Defects involving the entire bilateral or hemidiaphragm usually do not involve the crural tissue. These are traditionally thought to be severe variants of the Bochdalek posterior hernia. 3. Hiatal Hernias Hiatal hernias are generally not considered under the broad definition of CDHs, because they tend to be mild and present in older age groups, although at least some of these cases are clearly congenital. Congenital intrathoracic stomach (the most extreme case of hiatal hernia) has been reported in neonates (Hendrickson et al., 2003; Petersons et al., 2003). These defects result in the upward deviation of the stomach (often with a shortened esophagus and, in the most extreme case, an intrathoracic stomach) or in the herniation of a portion of the stomach or other abdominal contents alongside the esophagus (paraesophageal hernia). The pathogenesis of these hernias is unknown but probably heterogeneous. 4. Bilateral Defects Bilateral diaphragmatic defects are common, especially among populations of patients who have complex CDHs with other anomalies. Bilateral defects are usually symmetric in location, but they may vary by type. For example, a patient may have a posteromedial hernia with no sac on one side and a posteromedial muscularization defect causing a mild eventration on the other. Patients with hemidiaphragmatic aplasia might have diffuse muscularization

DIAPHRAGMATIC DEVELOPMENT

833

defects on the contralateral side. Findings of different types of defects within the same patients or within members of the same family suggest that they are developmentally related (Thomas et al., 1976; Rodgers and Hawks, 1986; Elberg et al., 1989; Akel and Nasr, 2001; St. Peter et al., 2005).

II. DIAPHRAGMATIC DEVELOPMENT A. Muscularization of the Diaphragm Muscle must be present in the diaphragm for the normal contraction that is needed to create a negative intrathoracic pressure to inflate the lungs. Muscle is always present in a characteristic pattern that forms the main portion of the diaphragm, the diaphragmatic crus in the posterior or dorsal region, and the paraesophageal muscle (see Figure 37.1). Although it was previously believed that diaphragmatic muscle formed from an ingrowth of body wall muscle, it is now evident that this muscle is derived from separate migratory populations of muscle precursor cells. The diaphragm is a hypaxial muscular structure, just like the tongue, the limbs, the shoulder muscles, the intercostal muscles, and the abdominal muscles. Although intercostal and abdominal muscles are nonmigratory hypaxial structures, the formation of muscle in the diaphragm, tongue, shoulder, and limbs requires a complex series of signaling events to allow muscle precursor cells to delaminate from the ventrolateral somite, to maintain motility, to reach target organs, and to differentiate at the appropriate time in development (Dietrich, 1999; Birchmeier and Brohmann, 2000; Bailey et al., 2001). Most studies of this process have concentrated on limb muscle, and some of the genes required for limb muscularization are also important for other hypaxial structure muscularization (e.g., c-Met); however, it is clear that these processes are unique and that distinct subpopulations of muscle progenitor cells encounter and respond to different guidance cues for each structure. For example, Lbx1 and CXCR4/SDF-1 are important for limb and tongue muscularization, but they are not required for normal diaphragmatic muscularization (Brohmann et al., 2000; van der Weyden et al., 2002). Genes required for normal diaphragmatic muscularization have been identified through the analysis of mutant mouse models, and they are listed in Table 37.1. Pax3 (paired box gene 3) is required for the establishment of the muscle progenitor pool in the ventral dermomyotomes. Mutant Pax3 mice (Splotch mice) have impaired development of the hypaxial muscles, and they have amuscular diaphragms (Li et al., 1999). The loss of this muscularization has been partially attributed to deficient expression of the c-Met receptor that is transcriptionally controlled by Pax3 (Epstein et al., 1996). The tyrosine kinase receptor, scatter factor/hepatocyte growth factor, and its ligand, c-Met, control the delamination and migration of migrating muscle precursor cells (Dietrich et al., 1999). Mice deficient in c-Met have amuscular diaphragms (Babiuk et al., 2003). Mice with a hypomorphic mutation in the Gata transcription cofactor Fog2 (Zfpm2) have an abnormal pattern of diaphragmatic muscularization with the apparent overgrowth of crural muscle and a lack of posterolateral muscle. A decreased and abnormal pattern of hepatocyte growth factor expression in the region of migration onto the diaphragm was found in mutant mice (Ackerman et al., 2005). The mechanisms that guide the development of the muscle pattern in the diaphragm are unknown. The development of

834

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

TABLE 37.1 Major Mouse Models of Abnormal Diaphragmatic Muscle Migration Gene

Phenotype

c-Met Fog2 (Zfpm2)

Amuscular diaphragm1 Muscle patterning defect (hepatocyte growth factor patterning abnormal)2 Amuscular diaphragm (signals c-met)3 Thin diaphragm, not functional4 Amuscular diaphragm (signals c-met)5

Gab1 MyoD Pax3 (Splotch) 1

Babiuk RP, Zhang W, Clugston R, et al: Embryological origins and development of the rat diaphragm, J Comp Neurol 455:477–487, 2003. 2 Ackerman KG, Herron BJ, Vargas SO, et al: Fog2 is required for normal diaphragm and lung development in mice and man, PLoS Genet 1:58–65, 2005. 3 Sachs M, Brohmann H, Zechner D, et al: Essential role of Gab1 for signaling by the c-Met receptor in vivo, J Cell Biol 150:1375–1384, 2000. 4 Kablar B, Krastel K, Ying C, et al: MyoD and Myf-5 differentially regulate the development of limb versus trunk skeletal muscle, Development 124:4729–4738, 1997. 5 Li J, Liu KC, Jin F, et al: Transgenic rescue of congenital heart disease and spina bifida in Splotch mice, Development 126:2495–2503, 1999.

this characteristic patterning has been described in embryonic rats (Babiuk et al., 2003) and in embryonic muscle cell reporter mice carrying a reporter LacZ gene controlled by the transcription factor myocyte enhancer factor 2 (Figure 37.2; Naya et al., 1999). These studies show that the muscle is first present in the crural regions and in the central regions, where the primordial diaphragmatic tissue from the pleuroperitoneal fold start to form the diaphragm. The migratory muscle (myotubes) extend toward the lateral body walls before advancing to the anterior regions (Babiuk et al., 2003). The phrenic nerves innervate the diaphragm. During embryogenesis, the phrenic axons initially target primordial diaphragmatic tissue in the pleuroperitoneal fold, which descends to form the posterior and lateral regions of the diaphragm. There is a close correlation between myotube formation in the diaphragm and the location and extension of phrenic intramuscular axons

A

Ventral

abd wall

Dorsal

C

B

*

CT

Dorsal

FIGURE 37.2 Early diaphragmatic muscularization in wild-type mouse embryos carrying the desMEF2-lacZ reporter gene, which defines embryonic muscle cells (blue) after X-gal staining. A and B, On mouse embryonic day 13.5, muscularization is present, but it has not reached the ventral regions, and it has not completely reached the abdominal wall (*). C, On embryonic day 14.5, the muscularization is complete. Muscle never covers the central tendon region (CT). (See color insert.)

DIAPHRAGMATIC DEVELOPMENT

835

with axonal growth after myotube formation (Allan and Greer, 1997; Babiuk et al., 2003). A lack of proper diaphragmatic innervation or a lack of use may result in decreased muscle cell mass or the atrophy of diaphragmatic muscle (Wolpowitz et al., 2000), and increased phrenic nerve use (through exercise such as opera singing) will cause muscle hypertrophy (Woodring and Bognar, 1998). There is no evidence that abnormal innervation of the diaphragm affects muscle cell patterning or migration. B. Development of the Central Tendon and Connective Tissue Components The diaphragm consists of both connective tissue and muscle, and the connective tissue portion is able to form completely if muscularization does not take place (Babiuk et al., 2003). The membranous diaphragm is thought to be derived from cells of both the septum transversum and the pleuroperitoneal folds (Greer et al., 2000; Yuan et al., 2003). The central tendon is the connective tissue portion of the diaphragm that is not populated by muscle cells (see Figure 37.1). The anterior portion of the central tendon sits under the heart, and it is attached to pericardial tissue. Severe disruptions of this region occur in humans with Pentalogy of Cantrell (Cantrell et al., 1958; Wesselhoeft and DeLuca, 1984; Carmi and Boughman, 1992). Milder disruptions are evident in mice that are deficient in the Slit3 gene. The Slit family of secreted proteins is highly conserved and critical for normal embryogenesis. Slit3 is expressed in the mesothelium of the developing diaphragm, and deficiency in two independent knockout models is associated with the herniation of thinned central tendon tissue, which fails to separate from the liver (Liu et al., 2003; Yuan et al., 2003). This has been attributed to a decrease in the proliferation of mesenchymal cells in the central tendon and to the hindered migration of mesothelial buds, which separate the liver from the diaphragm (Yuan et al., 2003). The defects observed are rare in humans, and there is no known association yet between Slit deficiency and abnormal human diaphragmatic development. The majority of the membranous diaphragm is believed to be derived from cells that migrate from the pleuroperitoneal folds. The pleuroperitoneal folds are wedge-shaped structures that arise from the lateral cervical walls and fuse with the septum transversum ventrally during embryogenesis (Figure 37.3). Both migratory muscle precursor cells (expressing Pax3 and MyoD) and phrenic nerves are present in the pleuroperitoneal folds (Babiuk et al., 2003). Further evidence that this tissue contributes to the diaphragm comes from the nitrofen diaphragmatic hernia model. Nitrofen (2,4-dichlorophenyl-p-nitrophenylether) is a teratogen that induces CDHs in rodent embryos after exposure in utero (Costlow and Manson, 1981). Nitrofen-induced defects occur in the posterior diaphragm and mimic the severe Bochdalek hernias seen in neonates. In rodent embryos that have been exposed to nitrofen, the structure of the pleuroperitoneal folds is abnormal, and the region of disruption corresponds anatomically with the region of absent diaphragm (Greer et al., 2000). Nitrofen also induces hernias in diaphragms that have no muscular contribution and in mice that do not grow lungs, which suggests that there is a direct effect on the integrity of the connective tissue component of the diaphragm that is not related to muscularization or lung growth (Babiuk and Greer, 2002). The mechanism of action of nitrofen is unknown, although it is at least partially dependent on retinoic acid (Greer et al., 2003).

836

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

CV

A

CV

E L

L

PN

H

Transverse section of mouse embryo on embryonic day 11.5. The diaphragm is not yet formed at this stage. Pleuroperitoneal fold cells that are thought to contribute to the diaphragm are evident at this stage (boxes). Phrenic motor axons are seen within this tissue (PN). A, Aorta; CV, cardinal vein; E, esophagus; L, lung; H, heart.

FIGURE 37.3

Mice with a loss of chicken ovalbumin upstream promoter transcription factor II (COUP-TFII, which is also called NrfF2) or Wilms tumor suppressor gene (Wt1), also have posterior diaphragmatic membrane defects. COUPTFII is a nuclear orphan receptor of the steroid/thyroid hormone receptor family (Park et al., 2003). COUP-TFII knockout mice die during embryogenesis before diaphragm formation, but investigators have used Cre-lox technology to create mice with a conditional loss of COUP-TFII in the abdominal mesenchyme, and these mice have posterior diaphragmatic defects (You et al., 2005). Wt1 is necessary both for tumor suppression and for the normal development of many organs, including the urogenital, splenic, and cardiac systems. Wt1 knockout mice have incomplete formation of the diaphragm (Kreidberg et al., 1993), and the PPF in Wt1 null mice is structurally abnormal (Clugston et al., 2006). Mutations in WT1 in humans cause a variety of defects and syndromes, some of which have been associated with congenital diaphragmatic defects (Royer-Pokora et al., 2004; Scott et al., 2005). There have been no human cases of isolated diaphragmatic defects associated with mutations in the WT1 gene (Nordenskjold et al., 1996). Mice with a compound loss of retinoic acid receptors also have diaphragmatic defects (Mendelsohn et al., 1994), and it is likely that the formation of the posterior diaphragm in both rodents and humans is dependent on retinoic acid. Vitamin A (retinol) and the retinoid signaling pathway are known to be crucial for normal embryogenesis (Ross et al., 2000). Retinol is transported to the cytoplasm by retinol binding proteins, where it is converted to retinoic acid before entering the nucleus to bind the retinoic acid receptors (RAR and RXR families), which leads to the regulation of many target genes (Greer et al., 2003). Diaphragmatic defects occur in rodents that are fed a diet that is deficient in vitamin A (Anderson, 1941; Wilson et al., 1953; Greer et al., 2003), and nitrofen affects embryonic retinoic acid production (Chen et al., 2003; Mey et al., 2003; Babiuk et al., 2004). In human newborns with CDH, preliminary studies have found reduced levels of both plasma retinol

LUNG DEVELOPMENT AND THE DIAPHRAGM

837

and retinol-binding protein (Major et al., 1998). The CDH candidate genes COUP-TFII, Fog2, and Gata4 are all suspected of playing roles in retinoic acid signaling (Malpel et al., 2000; Pereira et al., 2000; Huggins et al., 2001; Clabby et al., 2003).

III. LUNG DEVELOPMENT AND THE DIAPHRAGM A. Pulmonary Hypoplasia Associated with Diaphragmatic Defects Posterior diaphragmatic defects in humans are a major health problem, because they are associated with fatal or debilitating pulmonary hypoplasia. Defects in other regions of the diaphragm are usually not associated with pulmonary hypoplasia, although there are exceptions to this rule. The lungs of children with CDH have the inadequate ability to exchange gasses (likely as a result of inadequate branching and parenchymal development), and they also have pulmonary hypertension that can lead to vascular compromise. Often, these patients develop right heart failure from severe pulmonary hypertension during the newborn period. Unlike the physiologic abnormality persistent pulmonary hypertension of the newborn, the pulmonary hypertension in CDH is often only partially responsive to medical therapy. This may be the result of the severely reduced surface area of the pulmonary vascular system and of inappropriate remodeling or adaptation to elevated pulmonary pressures (Taira et al., 1998; Shehata et al., 1999; Kinsella et al., 2005; Jesudason, 2006). The mechanism of development of pulmonary hypoplasia has been a topic of considerable debate; however, it is becoming more evident that it is multifactorial and heterogeneous. Pulmonary hypoplasia may occur as a secondary defect when the lungs are compressed by herniated abdominal contents, but it also may occur as a primary defect when genes are affected that are necessary for both primary lung and primary diaphragmatic development. Those with primary lung defects are most likely those with the most severe pulmonary disease and clinical course. B. Evidence for Secondary Pulmonary Hypoplasia It is known that lungs develop hypoplasia and branching deficiency as a result of a lack of normal diaphragmatic excursion resulting from nervous system dysfunction (Goldstein and Reid, 1980; Fewell et al., 1981; Harding et al., 1993; Hill et al., 1994; Harding and Hooper, 1996; Harding, 1997) or abnormal muscularization (Tseng et al., 2000). When diaphragmatic hernias are created surgically in the in utero lamb, the animal develops pulmonary hypoplasia as a result of a compressive phenomenon (Lipsett et al., 1997; Ting et al., 1998; Lipsett et al., 2000). The contribution of secondary forces to pulmonary hypoplasia is supported by the findings of more severe abnormalities in the lung ipsilateral to the diaphragmatic defect (Areechon and Reid, 1963; Boyden, 1972; Goldstein and Reid, 1980). It has been suggested that pulmonary hypoplasia is not as severe when there are small diaphragmatic defects, because there is less herniation of abdominal contents. In general, lung size measured in utero correlates with respiratory outcome (Lipshutz et al., 1997; Laudy et al., 2003), although this is a subject of current controversy. Exceptions to this rule may occur as a result of technical issues or potential abnormalities in primary lung development (Heling et al., 2005).

838

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

C. Evidence for Primary Pulmonary Hypoplasia Teratogenic (nitrofen) and genetic rodent models have provided evidence that there are common developmental mechanisms that control lung and diaphragm development. Pulmonary hypoplasia develops as a primary defect in both the nitrofen and the Fog2 mouse models. When these lungs are removed and cultured in vitro (away from the influence of the developing diaphragm), they show growth or structural abnormalities. Lungs from nitrofen-exposed embryos have delayed branching, disorganization, and a reduction of late pulmonary epithelial developmental markers (Guilbert et al., 2000). Nitrofen-exposed embryos also have pulmonary smooth muscle functional abnormalities in vitro (Belik et al., 2003). Lungs from Fog2 mutant mice show growth delay and a specific structural branching defect (Ackerman et al., 2005). The CDH candidate genes Fog2, Wt1, COUP-TFII, and Gata4 (Table 37.2) are all expressed in the lung during development. Fog2 and Gata4 are both necessary for normal structural development and growth, and they are expressed in

TABLE 37.2

Mouse Genetic Models of Congenital Diaphragmatic Hernias

Gene

Mouse Model

Human Evidence

COUP-TFII (NR2F2)

Conditional knockouts (using Nkx3.2 Cre) have posterior hernias1 Hypomorphic allele homozygotes have loss of posterolateral musculature and abnormal muscular patterning2 Null allele carriers have anterior hernias, normal muscle Some null allele carriers have central hernias3 Null has central diaphragmatic rupture4 Compound nulls have posterior diaphragmatic hernias5 Null has central midline hernia6,7 Null has posterior hernias8

Multiple gene deletions (syndromic)

Fog2 (Zfpm2)

Gata4 Lox Retinoic acid receptors Slit3 Wt1

De novo mutation (not syndromic)

None Multiple gene deletions None None None Syndromic

1 You LR, Takamoto N, Yu CT, et al: Mouse lacking COUP-TFII as an animal model of Bochdalek-type congenital diaphragmatic hernia, Proc Natl Acad Sci U S A 102:16351–16356, 2005. 2 Ackerman KG, Herron BJ, Vargas SO, et al: Fog2 is required for normal diaphragm and lung development in mice and man, PLoS Genet 1:58–65, 2005. 3 Jay PY, Bielinska M, Erlich JM, et al: Impaired mesenchymal cell function in Gata4 mutant mice leads to diaphragmatic hernias and primary lung defects. Dev Biol 301:602–614, 2007. 4 Hornstra IK, Birge S, Starcher B, et al: Lysyl oxidase is required for vascular and diaphragmatic development in mice, J Biol Chem 278:14387–14393, 2003. 5 Mendelsohn C, Lohnes D, Decimo D, et al: Function of the retinoic acid receptors (RARs) during development (II). Multiple abnormalities at various stages of organogenesis in RAR double mutants, Development 120:2749–2771, 1994. 6 Liu J, Zhang L, Wang D, et al: Congenital diaphragmatic hernia, kidney agenesis and cardiac defects associated with Slit3-deficiency in mice, Mech Dev 120:1059–1070, 2003. 7 Yuan W, Rao Y, Babiuk RP, et al: A genetic model for a central (septum transversum) congenital diaphragmatic hernia in mice lacking Slit3, Proc Natl Acad Sci U S A 100:5217–5222, 2003. 8 Kreidberg JA, Sariola H, Loring JM, et al: WT-1 is required for early kidney development, Cell 74:679–691, 1993.

GENETICS OF HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS/CONGENITAL DIAPHRAGMATIC HERNIA

839

the early developing pulmonary mesenchyme (Ackerman et al., 2006; Jay et al., 2007). COUP-TFII is expressed in the developing lung mesenchyme, and it probably plays a role in retinoic-acid–mediated lung development (Malpel et al., 2000). Experimental design prohibits the evaluation of lung development in the COUP-TFII diaphragmatic hernia mouse model, because this gene was not deleted in the lung (You et al., 2005). Wt1 is expressed in the pulmonary mesothelium, and Wt1 knockout mice have abnormal lungs; however, these defects have not been further characterized. A neonate reported to have a mutation in Fog2 had severe bilateral pulmonary hypoplasia that was out of proportion with what would have been suspected on the basis of the severity of the diaphragmatic defect (Ackerman et al., 2005). Because Fog2 in the mouse is critical for both primary lung and diaphragm development, it is likely that this patient also had primary developmental abnormalities of both organs, which caused a severe phenotype. Hopefully, the discovery of similar associations with other candidate genes will improve our ability to better understand and predict risk on the basis of embryologic mechanisms.

IV. GENETICS OF HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS/CONGENITAL DIAPHRAGMATIC HERNIA A. Human Defects: Isolated Versus Syndromic or Complex Human diaphragmatic defects are considered to be complex birth defects. Unlike other complex genetic diseases (e.g., asthma, diabetes) that require the inheritance and interaction of multiple genetic predisposing factors (e.g., single-nucleotide polymorphisms, mutations), diaphragmatic defects may occur (in mice) with a single identified genetic change. Because the phenotypes of these disorders are heterogeneous and the pattern of inheritance is inconsistent, it is likely that mutations or other genetic variations in many different developmental pathway genes could cause abnormal diaphragmatic development. Diaphragmatic defects usually occur in isolation, although 30% of patients may have other anomalies (Enns et al., 1998). Nonsyndromic defects are considered to be sporadic, because the recurrence risk within a family is low (Huggins et al., 2001); however, families with multiple affected nonsyndromic diaphragmatic defects have been reported (Hitch et al., 1989; Farag et al., 1994; Kufeji and Crabbe, 1999). Because mortality and morbidity are high, good heritability data for the risk of transmission from an affected parent to an offspring are not yet available. Patients with multiple birth defects may fit the characteristic of a syndrome associated with CDH (i.e., syndromic CDH), or they may be identified as having complex CDH (i.e., not fitting a syndrome). Syndromic CDH occurs in 10% of all patients with a diaphragmatic defect, although this percentage varies widely on the basis of the population evaluated. Syndromes associated with CDH include Fryns syndrome (Slavotinek et al., 2005), Wt1 syndromes (e.g., Denys–Drash), craniofrontonasal syndrome, and Simpson–Golabi–Behmel syndrome (Slavotinek, 2005). (For a comprehensive review of syndromes associated with diaphragmatic defects, see the Congenital Diaphragmatic Hernia Overview at http://www. genetests.org.)

840

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

B. Genetic Abnormalities Cytogenetic abnormalities determined by high-resolution G-banded karyotype probably occur in 10% of all individuals with CDH, but higher percentages have been reported for populations that have syndromic or complex cases. The most common large cytogenetic abnormalities are trisomy 18 and isochromosome 12p (Pallister–Killian syndrome). Recurrent small cytogenetic abnormalities have been reported and are a resource for the discovery of CDH candidate genes (Lopez et al., 2006). Cytogenetic CDH hot spots that are sufficiently small or common are being used for candidate gene discovery, and they are listed in Table 37.3. The effort to define small cytogenetic aberrations that TABLE 37.3 Major Cytogenetic Hot Spots for Candidate Gene Discovery in Human Congenital Diaphragmatic Defects Location

Evidence

Candidate Genes

1q41–42.12

Multiple deletions, one translocation (syndromic, Fryns syndrome)1–3 Multiple deletions (syndromic)4–7 Multiple balanced translocations, deletion8–10 Multiple cases with trisomy (syndromic)12 Multiple deletions (syndromic)4,7,13

Too many to speculate

8p23.1 8q22.3(23.1) 11q23.3-qter 15q26.2 1

Gata4 most likely Fog211 Too many to speculate 5 Mb region, COUP-TFII most likely14

Kantarci S, Casavant D, Prada C, et al: Findings from aCGH in patients with congenital diaphragmatic hernia (CDH): a possible locus for Fryns syndrome, Am J Med Genet A 140:17–23, 2006. 2 Youssoufian H, Chance P, Tuck-Muller CM, Jabs EW: Association of a new chromosomal deletion [del(1)(q32q42)] with diaphragmatic hernia: assignment of a human ferritin gene, Hum Genet 78:267–270, 1988. 3 Smith SA, Martin KE, Dodd KL, Young ID: Severe microphthalmia, diaphragmatic hernia and Fallot’s tetralogy associated with a chromosome 1;15 translocation, Clin Dysmorphol 3:287–291, 1994. 4 Slavotinek A, Lee SS, Davis R, et al: Fryns syndrome phenotype caused by chromosome microdeletions at 15q26.2 and 8p23.1, J Med Genet 42:730–736, 2005. 5 Slavotinek AM, Moshrefi A, Davis R, et al: Array comparative genomic hybridization in patients with congenital diaphragmatic hernia: mapping of four CDH-critical regions and sequencing of candidate genes at 15q26.1–15q26.2, Eur J Hum Genet 14:999–1008,2006. 6 Shimokawa O, Miyake N, Yoshimura T, et al: Molecular characterization of del(8)(p23.1p23.1) in a case of congenital diaphragmatic hernia, Am J Med Genet A 136:49–51, 2005. 7 Lopez I, Bafalliu JA, Bernabe MC, et al: Prenatal diagnosis of de novo deletions of 8p23.1 or 15q26.1 in two fetuses with diaphragmatic hernia and congenital heart defects, Prenat Diagn 26:577–580, 2006. 8 Temple IK, Barber JC, James RS, Burge D: Diaphragmatic herniae and translocations involving 8q22 in two patients, J Med Genet 31:735–737, 1994. 9 Howe DT, Kilby MD, Sirry H, et al: Structural chromosome anomalies in congenital diaphragmatic hernia, Prenat Diagn 16:1003–1009, 1996. 10 Cappellini A, Sala E, Colombo D, et al: Monosomy 8q and features of Fryns’ syndrome (abstract), Eur J Hum Genet 4(Suppl 1):29, 1996. 11 Ackerman KG, Herron BJ, Vargas SO, et al: Fog2 is required for normal diaphragm and lung development in mice and man, PLoS Genet 1:58–65, 2005. 12 Klaassens M, Scott DA, van Dooren M, et al: Congenital diaphragmatic hernia associated with duplication of 11q23-qter, Am J Med Genet A 140:1580–1586, 2006. 13 Klaassens M, van Dooren M, Eussen HJ, et al: Congenital diaphragmatic hernia and chromosome 15q26: determination of a candidate region by use of fluorescent in situ hybridization and array-based comparative genomic hybridization, Am J Hum Genet 76:877–882, 2005.

841

SUMMARY

are not detectable by routine karyotypes has been advanced by array comparative genomic hybridization technology. Deletions at human chromosomes 15q26.1–26.2 have been extensively mapped to narrow the region of interest to four major candidate genes (Klaassens et al., 2005; Slavotinek et al., 2006). COUP-TFII is the most likely causal gene in the region based on mouse models of diaphragmatic defects (You et al., 2005). Deletions in the 8p23.1 region include the transcription factor Gata4 (Shimokawa et al., 2005; Slavotinek et al., 2006), which interacts with Fog2 for normal heart, lung, and probably diaphragm development (Crispino et al., 2001). Translocations in the 8q22.3– 23.1 region have been reported in unrelated individuals with nonsyndromic CDH (Temple et al., 1994; Howe et al., 1996). Deletions in this region have not been found in array comparative genomic hybridization experiments, which suggests that the loss of genes in this region results in low viability. In a cohort of 30 deceased patients with diaphragmatic defects evaluated for Fog2 mutations, one patient had a de novo heterozygous nonsense mutation that was predicted to result in one functional copy of FOG2. This child died at 5 hours of life as a result of severe respiratory distress, and was found to have a posterior diaphragmatic eventration with severe bilateral pulmonary hypoplasia. The sequencing of the DNA of patients with milder phenotypes is ongoing, but it has not revealed additional mutations. It is interesting that both COUP-TFII and Gata4 interact with Fog2 (Crispino et al., 2001; Huggins et al., 2001); this makes a developmental pathway for diaphragm development requiring these three genes likely.

VI. CONCLUSIONS In this chapter, we have reviewed normal diaphragmatic anatomy, common diaphragmatic structural abnormalities, and the genes that are currently known to be necessary for normal diaphragmatic development. The integration of developmental genetics in animal models with human genetics and development has enhanced our current understanding of the development of the diaphragm, and the continued integration of these fields will hopefully lead us to the genetic and mechanistic classification of all human congenital diaphragmatic defects. This will be important for providing prognostic information to parents and heritability risk information to affected individuals.

SUMMARY

 Posterior hernias are associated with defects in the pleuroperitoneal folds during development.

 Defects of specific genes, including Fog2, COUP-TF II, Gata4, Slit3, and Wt1, result in defective diaphragmatic development in animal models.

 In humans, Fog2 is the only gene that has so far been shown to be asso-



ciated with isolated diaphragmatic defect. WT1 mutations can be associated with syndromic diaphragmatic defects. COUP-TFII is located in a cytogenetic hot spot for CDH, and it is deleted in some humans with syndromic or complex CDH. Defects in pulmonary development may occur concordantly with defects in diaphragmatic development, and they may account for the poor outcomes of many affected children.

842

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

ACKNOWLEDGMENTS The authors have no conflicts of interest. The desMEF2 lacZ reporter mice were provided by Dr. Eric Olson of The University of Texas Southwestern Medical Center at Dallas.

GLOSSARY Bochdalek hernia A congenital diaphragmatic hernia of the posterolateral region of the diaphragm. This term is commonly used to describe any diaphragmatic hernia in the lateral or posterior region. Central tendon (of the diaphragm) The unmuscularized portion of the diaphragm, most of which sits under the heart and which is attached to the liver inferiorly by the falciform ligament Congenital diaphragmatic hernia This term is most often used to describe the Bochdalek hernia or other posterolateral diaphragmatic hernias, but it technically refers to any type of diaphragmatic hernia of congenital origin. Eventration (of the diaphragm) The protrusion of abdominal contents into the thoracic cavity. This term is usually used to describe a mild protrusion versus a hernia, which involves massive protrusion. Others use the term to describe a defect including the entire hemidiaphragm. An eventration may occur with or without muscularization defect. Morgagni hernia A diaphragmatic hernia in the anterior region of the diaphragm. This term is commonly used to describe a variety of types of anterior hernias. Muscularization defect A complete lack of muscle in a region of the diaphragm or abnormally patterned diaphragmatic musculature. Depending on the location and size of the muscularization defect, the result might be an eventration, a sac hernia, or no apparent protrusion of abdominal contents. Pentalogy of Cantrell A syndrome of congenital defects involving the abdominal wall, the sternum, the diaphragm, the pericardium, and the heart. These are considered to be ventral developmental field defects. Pleuroperitoneal fold Tissue in the developing embryonic thorax that will form a major part of the diaphragm and that will separate the pleural and peritoneal cavities.

REFERENCES Ackerman KG, Herron BJ, Vargas SO, et al: Fog2 is required for normal diaphragm and lung development in mice and man, PLoS Genetics 1:58–65, 2005.

REFERENCES

843 Ackerman KG, Wang JN, Fujiwara Y, et al: Gata4 is necessary for normal pulmonary lobar development, Am J Respir Cell Mol Biol 36:391–397, 2007. Akel S, Nasr W: Multiple ipsilateral congenital diaphragmatic pathologies: rarities to consider, Eur J Pediatr Surg 11:200–203, 2001. Allan DW, Greer JJ: Embryogenesis of the phrenic nerve and diaphragm in the fetal rat, J Comp Neurol 382:459–468, 1997. Anderson DH: Incidence of congenital diaphragmatic hernia in the young of rats bred on a diet deficient in vitamin A, Am J Dis Child 62:888–889, 1941. Areechon W, Reid L: Hypoplasia of the lung with congenital diaphragmatic hernia, Br Med J 1:230–233, 1963. Babiuk RP, Greer JJ: Diaphragm defects occur in a congenital diaphragmatic hernia model independent of myogenesis and lung formation, Am J Physiol Lung Cell Mol Physiol 283: L1310–L1314, 2002. Babiuk RP, Thebaud B, Greer JJ, et al: Reductions in the incidence of nitrofen-induced diaphragmatic hernia by vitamin A and retinoic acid, Am J Physiol Lung Cell Mol Physiol 286: L970–L973, 2004. Babiuk RP, Zhang W, Clugston R, et al: Embryological origins and development of the rat diaphragm, J Comp Neurol 455:477–487, 2003. Bailey P, Holowacz T, Lassar AB, et al: The origin of skeletal muscle stem cells in the embryo and the adult, Curr Opin Cell Biol 13:679–689, 2001. Belik J, Davidge ST, Zhang W, et al: Airway smooth muscle changes in the nitrofen-induced congenital diaphragmatic hernia rat model, Pediatr Res 53:737–743, 2003. Beresford MW, Shaw NJ: Outcome of congenital diaphragmatic hernia, Pediatr Pulmonol 30:249–256, 2000. Birchmeier C, Brohmann H: Genes that control the development of migrating muscle precursor cells, Curr Opin Cell Biol 12:725–730, 2000. Boyden EA: The structure of compressed lungs in congenital diaphragmatic hernia, Am J Anat 134:497–508, 1972. Brohmann H, Jagla K, Birchmeier C: The role of Lbx1 in migration of muscle precursor cells, Development 127:437–445, 2000. Cantrell JR, Haller JA, Ravitch MM: A syndrome of congenital defects involving the abdominal wall, sternum, diaphragm, pericardium, and heart, Surg Gynecol Obstet 107:602–614, 1958. Carmi R, Boughman JA: Pentalogy of Cantrell and associated midline anomalies: a possible ventral midline developmental field, Am J Med Genet 42:90–95, 1992. Chen MH, MacGowan A, Ward S, et al: The activation of the retinoic acid response element is inhibited in an animal model of congenital diaphragmatic hernia, Biol Neonate 83:157–161, 2003. Clabby ML, Robison TA, Quigley HF, et al: Retinoid X receptor alpha represses GATA-4mediated transcription via a retinoid-dependent interaction with the cardiac-enriched repressor FOG-2, J Biol Chem 278:5760–5767, 2003. Clugston RD, Klattig J, Englert C, et al: Teratogen-induced, dietary and genetic models of congenital diaphragmatic hernia share a common mechanism of pathogenesis, Am J Pathol 169:1541–1549, 2006. Costlow RD, Manson JM: The heart and diaphragm: target organs in the neonatal death induced by nitrofen (2,4-dichlorophenyl-p-nitrophenyl ether), Toxicology 20:209–227, 1981. Crispino JD, Lodish MB, Thurberg BL, et al: Proper coronary vascular development and heart morphogenesis depend on interaction of GATA-4 with FOG cofactors, Genes Dev 15:839–844, 2001. Dietrich S: Regulation of hypaxial muscle development, Cell Tissue Res 296:175–182, 1999. Dietrich S, Abou-Rebyeh F, Brohmann H, et al: The role of SF/HGF and c-Met in the development of skeletal muscle, Development 126:1621–1629, 1999. Elberg JJ, Brok KE, Pederson SA, Kock KE: Congenital bilateral eventration of the diaphragm in a pair of male twins, J Pediatr Surg 24:1140–1141, 1989. Enns GM, Cox VA, Goldstein RB, et al: Congenital diaphragmatic defects and associated syndromes, malformations, and chromosome anomalies: a retrospective study of 60 patients and literature review, Am J Med Genet 79:215–225, 1998. Epstein JA, Shapiro DN, Cheng J, et al: Pax3 modulates expression of the c-Met receptor during limb muscle development, Proc Natl Acad Sci U S A 93:4213–4218, 1996. Farag TI, Bastaki L, Marafie M, et al: Autosomal recessive congenital diaphragmatic defects in the Arabs, Am J Med Genet 50:300–301, 1994. Fewell JE, Lee CC, Killerman JA: Effects of phrenic nerve section on the respiratory system of fetal lambs, J Appl Physiol 51:293–297, 1981.

844

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

Goldstein JD, Reid LM: Pulmonary hypoplasia resulting from phrenic nerve agenesis and diaphragmatic amyoplasia, J Pediatr 97:282–287, 1980. Greer JJ, Allan DW, Babink RP, Lemke RP: Recent advances in understanding the pathogenesis of nitrofen-induced congenital diaphragmatic hernia, Pediatr Pulmonol 29:394–399, 2000. Greer JJ, Babiuk RP, Thebaud B: Etiology of congenital diaphragmatic hernia: the retinoid hypothesis, Pediatr Res 53:726–730, 2003. Greer JJ, Cote D, Allan DW: Structure of the primordial diaphragm and defects associated with nitrofen-induced CDH, J Appl Physiol 89:2123–2129, 2000. Guilbert TW, Gebb SA, Shannon JM: Lung hypoplasia in the nitrofen model of congenital diaphragmatic hernia occurs early in development, Am J Physiol Lung Cell Mol Physiol 279: L1159–L1171, 2000. Harding R: Fetal pulmonary development: the role of respiratory movements, Equine Vet J Suppl Jun(24):32–39, 1997. Harding R, Hooper SB: Regulation of lung expansion and lung growth before birth, J Appl Physiol 81:209–224, 1996. Harding R, Hooper SB, Han VK: Abolition of fetal breathing movements by spinal cord transection leads to reductions in fetal lung liquid volume, lung growth, and IGF-II gene expression, Pediatr Res 34:148–153, 1993. Heling KS, Wauer RR, Hammer H: Reliability of the lung-to-head ratio in predicting outcome and neonatal ventilation parameters in fetuses with congenital diaphragmatic hernia, Ultrasound Obstet Gynecol 25:112–118, 2005. Hendrickson RJ, Fenton L, Hall D, Karrer FM: Congenital paraesophageal hiatal hernia, J Am Coll Surg 196:483, 2003. Hill AC, Adzick NS, Stevens MB: Fetal lamb pulmonary hypoplasia: pulmonary vascular and myocardial abnormalities, Ann Thorac Surg 57:946–951, 1994. Hitch DC, Carson JA, Smith EI: Familial congenital diaphragmatic hernia is an autosomal recessive variant, J Pediatr Surg 24:860–864, 1989. Howe DT, Kilby MD, Sirry H: Structural chromosome anomalies in congenital diaphragmatic hernia, Prenat Diagn 16:1003–1009, 1996. Huggins GS, Bacani CJ, Boltax J: Friend of GATA 2 physically interacts with chicken ovalbumin upstream promoter-TF2 (COUP-TF2) and COUP-TF3 and represses COUP-TF2-dependent activation of the atrial natriuretic factor promoter, J Biol Chem 276:28029–28036, 2001. Irish MS, Holm BA, Glick PL: Congenital diaphragmatic hernia. A historical review, Clin Perinatol 23:625–653, 1996. Jay PY, Bielinska M, Erlich JM, et al: Impaired mesenchymal cell function in Gata4 mutant mice leads to diaphragmatic hernias and primary lung defects, Dev Biol 301:602–614, 2007. Jesudason EC: Small lungs and suspect smooth muscle: congenital diaphragmatic hernia and the smooth muscle hypothesis, J Pediatr Surg 41:431–435, 2006. Kinsella JP, Ivy DD, Abman SH: Pulmonary vasodilator therapy in congenital diaphragmatic hernia: acute, late, and chronic pulmonary hypertension, Semin Perinatol 29:123–128, 2005. Klaassens M, van Dooren M, Eussen HJ, et al: Congenital diaphragmatic hernia and chromosome 15q26: determination of a candidate region by use of fluorescent in situ hybridization and array-based comparative genomic hybridization, Am J Hum Genet 76:877–882, 2005. Kreidberg JA, Sariola H, Loring JM, et al: WT-1 is required for early kidney development, Cell 74:679–691, 1993. Kufeji DI, Crabbe DC: Familial bilateral congenital diaphragmatic hernia, Pediatr Surg Int 15:58–60, 1999. Laudy JA, Van Gucht M, Van Dooren MF, et al: Congenital diaphragmatic hernia: an evaluation of the prognostic value of the lung-to-head ratio and other prenatal parameters, Prenat Diagn 23:634–639, 2003. Li J, Liu KC, Jin F, et al: Transgenic rescue of congenital heart disease and spina bifida in Splotch mice, Development 126:2495–2503, 1999. Lipsett J, Cool JC, Runciman SC, et al: Morphometric analysis of pulmonary development in the sheep following creation of fetal diaphragmatic hernia, Pediatr Pathol Lab Med 17:789–807, 1997. Lipsett J, Cool JC, Runciman SI, et al: Morphometric analysis of preterm fetal pulmonary development in the sheep model of congenital diaphragmatic hernia, Pediatr Dev Pathol 3:17–28, 2000. Lipshutz GS, Albanese CT, Feldstein VA, et al: Prospective analysis of lung-to-head ratio predicts survival for patients with prenatally diagnosed congenital diaphragmatic hernia, J Pediatr Surg 32:1634–1636, 1997. Liu J, Zhang L, Wang D, et al: Congenital diaphragmatic hernia, kidney agenesis and cardiac defects associated with Slit3-deficiency in mice, Mech Dev 120:1059–1070, 2003.

REFERENCES

845 Loong TP, Kocher HM: Clinical presentation and operative repair of hernia of Morgagni, Postgrad Med J 81:41–44, 2005. Lopez I, Bafalliu JA, Bernabe MC, et al: Prenatal diagnosis of de novo deletions of 8p23.1 or 15q26.1 in two fetuses with diaphragmatic hernia and congenital heart defects, Prenat Diagn 26:577–580, 2006. Major D, Cadenas M, Fournier L, et al: Retinol status of newborn infants with congenital diaphragmatic hernia, Pediatr Surg Int 13:547–549, 1998. Malpel S, Mendelsohn C, Cardoso WV: Regulation of retinoic acid signaling during lung morphogenesis, Development 127:3057–3067, 2000. Mendelsohn C, Lohnes D, Decimo D, et al: Function of the retinoic acid receptors (RARs) during development (II). Multiple abnormalities at various stages of organogenesis in RAR double mutants, Development 120:2749–2771, 1994. Mey J, Babiuk RP, Clugston R, et al: Retinal dehydrogenase-2 is inhibited by compounds that induce congenital diaphragmatic hernias in rodents, Am J Pathol 162:673–679, 2003. Naya FJ, Wu C, Richardson JA, et al: Transcriptional activity of MEF2 during mouse embryogenesis monitored with a MEF2-dependent transgene, Development 126:2045–2052, 1999. Nordenskjold A, Tapper-Persson M, Anvret M: No evidence of WT1 gene mutations in children with congenital diaphragmatic hernia, J Pediatr Surg 31:925–927, 1996. Paci M, de Franco S, Della Valle E, et al: Septum transversum diaphragmatic hernia in an adult, J Thorac Cardiovasc Surg 129:444–445, 2005. Park JI, Tsai SY, Tsai MJ: Molecular mechanism of chicken ovalbumin upstream promoter-transcription factor (COUP-TF) actions, Keio J Med 52:174–181, 2003. Pereira FA, Tsai MJ, Tsai SY: COUP-TF orphan nuclear receptors in development and differentiation, Cell Mol Life Sci 57:1388–1398, 2000. Petersons A, Liepina M, Spitz L: Neonatal intrathoracic stomach in Marfan’s syndrome: report of two cases, J Pediatr Surg 38:1663–1664, 2003. Pober BR, Lin A, Russell M, et al: Infants with Bochdalek diaphragmatic hernia: sibling precurrence and monozygotic twin discordance in a hospital-based malformation surveillance program, Am J Med Genet A 138:81–88, 2005. Rodgers BM, Hawks P: Bilateral congenital eventration of the diaphragms: successful surgical management, J Pediatr Surg 21:858–864, 1986. Ross SA, McCaffery PJ, Drager UC, De Luca LM: Retinoids in embryonal development, Physiol Rev 80:1021–1054, 2000. Royer-Pokora B, Beier M, Henzler M, et al: Twenty-four new cases of WT1 germline mutations and review of the literature: genotype/phenotype correlations for Wilms tumor development, Am J Med Genet A 127:249–257, 2004. Salman AB, Tanyel FC, Senocak ME, Buyukpamukcu N: Four different hernias are encountered in the anterior part of the diaphragm, Turk J Pediatr 41:483–488, 1999. Scott DA, Cooper ML, Stankiewicz P, et al: Congenital diaphragmatic hernia in WAGR syndrome, Am J Med Genet A 134:430–433, 2005. Shehata SM, Tibboel D, Sharma HS, Mooi WJ: Impaired structural remodelling of pulmonary arteries in newborns with congenital diaphragmatic hernia: a histological study of 29 cases, J Pathol 189:112–118, 1999. Shimokawa O, Miyake N, Yoshimura T, et al: Molecular characterization of del(8)(p23.1p23.1) in a case of congenital diaphragmatic hernia, Am J Med Genet A 136:49–51, 2005. Skari H, Bjornland K, Haugen G, et al: Congenital diaphragmatic hernia: a meta-analysis of mortality factors, J Pediatr Surg 35:1187–1197, 2000. Slavotinek A, Lee SS, Davis R, et al: Fryns syndrome phenotype caused by chromosome microdeletions at 15q26.2 and 8p23.1, J Med Genet 42:730–736, 2005. Slavotinek AM: The genetics of congenital diaphragmatic hernia, Semin Perinatol 29:77–85, 2005. Slavotinek AM, Moshrefi A, Davis R, et al: Array comparative genomic hybridization in patients with congenital diaphragmatic hernia: mapping of four CDH-critical regions and sequencing of candidate genes at 15q26.1–15q26.2, Eur J Hum Genet 14:999–1008, 2006. St. Peter SD, Shah SR, Little DC, et al: Bilateral congenital diaphragmatic hernia with absent pleura and pericardium, Birth Defects Res A Clin Mol Teratol 73:624–627, 2005. Stege G, Fenton A, Jaffray B: Nihilism in the 1990s: the true mortality of congenital diaphragmatic hernia, Pediatrics 112(3 Pt 1):532–535, 2003. Sweeney LJ: Basic concepts in embryology, New York, 1998, McGraw-Hill. Taira Y, Yamataka T, et al: Adventitial changes in pulmonary vasculature in congenital diaphragmatic hernia complicated by pulmonary hypertension, J Pediatr Surg 33:382–387, 1998.

846

DIAPHRAGMATIC EMBRYOGENESIS AND HUMAN CONGENITAL DIAPHRAGMATIC DEFECTS

Temple IK, Barber JC, James RS, Burge D: Diaphragmatic herniae and translocations involving 8q22 in two patients, J Med Genet 31:735–737, 1994. Thomas MP, Stern LM, Morris LL: Bilateral congenital diaphragmatic defects in two siblings, J Pediatr Surg 11:465–467, 1976. Ting A, Glick PL, Wilcox DT, et al: Alveolar vascularization of the lung in a lamb model of congenital diaphragmatic hernia, Am J Respir Crit Care Med 157:31–34, 1998. Tseng BS, Cavin ST, Booth FW, et al: Pulmonary hypoplasia in the myogenin null mouse embryo, Am J Respir Cell Mol Biol 22:304–315, 2000. van der Weyden L, Adams DJ, Bradley A: Tools for targeted manipulation of the mouse genome, Physiol Genomics 11:133–164, 2002. Wesselhoeft CW Jr, DeLuca FG: Neonatal septum transversum diaphragmatic defects, Am J Surg 147:481–485, 1984. West SD, Wilson JM: Follow up of infants with congenital diaphragmatic hernia, Semin Perinatol 29:129–133, 2005. Wilson JG, Roth CB, Warkany J: An analysis of the syndrome of malformations induced by maternal vitamin A deficiency. Effects of restoration of vitamin A at various times during gestation, Am J Anat 92:189–217, 1953. Wolpowitz D, Mason TB, Dictrich P, et al: Cysteine-rich domain isoforms of the neuregulin-1 gene are required for maintenance of peripheral synapses, Neuron 25:79–91, 2000. Woodring JH, Bognar B: Muscular hypertrophy of the left diaphragmatic crus: an unusual cause of a paraspinal “mass,” J Thorac Imaging 13:144–145, 1998. You LR, Takamoto N, Yu CT, et al: Mouse lacking COUP-TFII as an animal model of Bochdalektype congenital diaphragmatic hernia, Proc Natl Acad Sci U S A 102:16351–16356, 2005. Yuan W, Rao Y, Babiuk RP, et al: A genetic model for a central (septum transversum) congenital diaphragmatic hernia in mice lacking Slit3, Proc Natl Acad Sci U S A 100:5217–5222, 2003.

RECOMMENDED RESOURCE Gene Tests: www.genetests.org—A publicly funded (National Institutes of Health) medical genetics information resource that includes GeneReviews of congenital disorders, including congenital diaphragmatic hernia.

38

FORMATION OF VERTEBRATE LIMBS YINGZI YANG Genetic Disease Research Branch, National Human Genome Research Institute, National Institutes of Health, Bethesda, MD

INTRODUCTION There are considerable morphologic and functional differences among the limbs of different vertebrate species. Most species use their limbs to support their body weight, to walk, and to run. However, birds and bats use their forelimbs to fly, and whales use their limbs to swim. Human beings use their arms, legs, hands, and feet to perform more complicated skills and artistic tasks. However, a closer examination of all vertebrate limbs reveals that their structures are, in fact, remarkably similar. The origin of these similarities is believed to stem from the possession of a common ancestor. Most vertebrates have four limbs. No matter what purposes the limbs are serving, the skeletal elements constructing the limbs, the muscles operating them, and the nerves controlling the muscles always bear basic similarities. For example, both the mouse forelimb and the chick wing have a shoulder, a girdle, a humerus, a radius, an ulna and digits, although the number of digits varies. More importantly, the structural and functional similarities of vertebrate limbs are determined by similar developmental processes when the limbs form in the embryo. Vertebrate limbs develop from the embryonic structure called limb bud. The limb bud forms by localized proliferation of the lateral plate mesoderm at certain axial levels at the dorsal–ventral (DV) boundary. In the developing mouse embryo, a visible fore limb bud appears at 9.5 days postcoitum, and the hindlimb bud forms slightly later. The limb mesenchyme gives rise to patterned limb structures that form later during development that include cartilage, bone, tendon, and ligaments. Muscle, nerves, and blood vessels are derived from cells that migrate into the limb bud during development. According to the molecular regulation and morphogenetic events, limb development can be divided into three stages: limb initiation, limb patterning, and late limb morphogenesis. This chapter will focus on the first two stages. Principles of Developmental Genetics Dr. Yang contributed to this article in her personal capacity. The views expressed are her own and do not necessarily represent the views of the National Institutes of Health or the United States Government.

847

848

FORMATION OF VERTEBRATE LIMBS

The investigation of limb development is a very active field of developmental biology. Before the tools of molecular developmental biology were developed, classic embryological studies using mainly chick limb buds contributed significantly to the current understanding of limb development by identifying the signaling centers that control limb pattern formation. The chick limb bud has been one of the main systems of choice for embryological studies of limb development because the chick limb bud is accessible and large in size, and it can be easily manipulated in ovo without affecting other developmental processes. For example, signaling centers that control the threedimensional limb patterning were identified through tissue recombination experiments in the chick limb bud, which could not have been performed in mouse embryos, which develop in utero. More recently, the availability of genetic tools in mice in combination with embryologic studies in chick have led to the identification of the signaling molecules and pathways that mediate that function of the previously identified signaling centers, which will be discussed further in this chapter. Because of the pleiotrophic roles of the signaling pathways identified in limb development, molecular human and mouse genetic studies have identified genetic variants in the limb signaling pathways leading to both human birth defects and diseases, such as polydactyly, brachydactyly, Greig cephalopolysyndactyly syndrome, Gorlin syndrome, and Bardet–Biedl syndrome.

I. LIMB INITIATION The positions of limb bud formation along the anterior–posterior (AP) axis are determined genetically in embryonic development. The area in the lateral plate mesoderm that is competent to form a limb is called the limb field. The limb field was identified by tissue graft experiments in classic embryological studies (Kieny, 1960; 1968). In the chick embryo, when it is grafted to an ectopic location, grafted limb field tissue directs the development of an entire limb. It is interesting to note that the size of limb field is bigger than the actual size of the lateral plate that forms a limb bud during normal development. It has been suggested that Hox genes expressed at different levels along the AP axis determine where a limb bud will form (Cohn et al., 1997; Davidson et al., 1991; Izpisua-Belmonte et al., 1991). There is a narrow time window during which the grafted lateral plate mesoderm can induce the formation of an ectopic limb. Younger mesoderm requires associated somite tissue to induce limb formation, whereas older mesoderm from the already formed endogenous limb bud loses such inductive ability (Dhouailly and Kieny, 1972). Because the early limb bud is simply a mesoderm core covered by the surface ectoderm, limb initiation requires extensive epithelial–mesenchymal interactions. However, classic embryological work has demonstrated that it is the lateral mesenchymal cells that provide the signals to initiate the process of limb development and that determine the specific identities of the limb type (arm/leg or wing/leg) and the axial levels of limb bud formation (Detwiler, 1933; Hamburger, 1938). In the pre-limb bud mesoderm, T-box transcription factors Tbx5 and Tbx4 are expressed in the future forelimb and hindlimb areas, respectively (Agarwal et al., 2003; Naiche and Papaioannou, 2003; Takeuchi et al., 2003; see Chapter 16). However, despite the highly conserved

LIMB INITIATION

849

DNA sequences and functions between Tbx5 and Tbx4, only Tbx5 is required for limb bud initiation. No forelimb bud forms in Tbx5 / mutant embryos, but, in Tbx4 / embryos, hindlimb bud forms and then degenerates. The function of Tbx5 is cell-autonomous as a transcription factor. However, limb initiation requires signaling from the lateral mesoderm to the overlying ectoderm. Fibroblast growth factor 10 (Fgf10), which has been found to play a critical role in limb initiation, appears to mediate the role of Tbx5 in limb bud initiation by signaling to the overlying ectoderm. First, it was found that beads soaked in recombinant FGF1, FGF2, and FGF4 can induce the formation of a complete and morphologically normal limb when implanted to the lateral plate mesoderm (Cohn et al., 1995). Second, it was found that Fgf10 is expressed in the limb field and that Fgf10 / mice develop without limbs. Similar to what has been observed in the Tbx4 / mouse embryos, a tiny limb bud does form, but it quickly degenerates in the Fgf10 / mice (Min et al., 1998; Sekine et al., 1999). Third, in the Tbx5 / and Tbx4 / mouse embryos, Fgf10 expression is either never induced or weakly expressed and then quickly lost after initial expression (Naiche and Papaioannou, 2003). Fourth, FGF receptor 2 (Fgfr2) appears to mediate Fgf10 signaling in controlling limb initiation. Fgfr2 / mutant embryos also form very small limb buds, which degenerate quickly. The expression of Tbx5 is intact in the Fgfr2 / embryos (Agarwal et al., 2003). Therefore, Tbx5 and Tbx4 act upstream of Fgf10 during limb initiation. Another signaling pathway that plays a critical role in limb initiation is the Wnt/b-catenin signaling pathway. Central to this pathway is the stabilization of b-catenin, which is phosphorylated and degraded in the absence of Wnt signaling. When Wnt signaling is active, b-catenin phosphorylation is inhibited, and it then translocates to the nucleus where it binds lymphoid enhancer-binding factor/T-cell factors (Lef/ Tcfs) to activate downstream gene expression. Lef1 and Tcf1 are coexpressed with Tbx5 or Tbx4 in the prospective and early limb mesoderm. Expression of Lef1/Tcf1 is also lost in the prospective limb field of the Tbx5 / embryos. However, the expression of Tbx5 is intact in Lef1 / /Tcf1 / double mutant embryos (Agarwal et al., 2003). The Lef1 / /Tcf1 / double mutant embryos also form much smaller limb buds that degenerate soon after (Galceran et al., 1999). The epithelial–mesenchymal interaction during limb initiation is likely to be mediated by the interaction between Fgf and the Wnt/b-catenin signaling. After Tbx5 is expressed, Wnt signaling acts in concert with Tbx5 to activate Fgf10 expression fully in the developing limb bud. Tbx5 alone can activate Fgf10 expression, whereas the canonical Wnt signaling sustains high levels of Fgf10 expression during limb initiation (Figure 38.1; Agarwal et al., 2003). Tbx5, Fgf10, and Wnt3 are initially expressed in a broader area than just the presumptive limb bud. It is likely that Tbx5 establishes a limb field by activating the expression of Fgf10 in the mesenchyme of the limb field. Because Wnt3 is expressed in the ectoderm overlying the presumptive limb bud mesoderm, Wnt3 and Fgf10 may mediate the extensive ectoderm– mesenchymal interaction during early limb initiation, and they may maintain the expression of each other by forming a positive feedback loop (see Figure 38.1). This is supported by the finding that Wnt3 is required for limb initiation (Barrow et al., 2003). It appears that Wnt3 signals to both the limb mesenchyme and the ectoderm to regulate limb initiation and elongation. It has been demonstrated that Wnt3 signals through b-catenin within the

850

FORMATION OF VERTEBRATE LIMBS

FIGURE 38.1 Signaling pathways regulating limb initiation and apical ectoderm ridge (AER) formation. Tbx4 and Tbx5 are expressed in the lateral plate mesoderm (LPM) before the expression of Fgf10, Lef1, and Tcf1. Tbx4 and Tbx5 directly activate Fgf10 expression. Fgf10 and Wnt3 expressed in the surface ectoderm (SE) may form a positive feedback loop to achieve the high expression level required for limb bud outgrowth and AER induction. At this stage, Wnt3 signaling for the regulation of Fgf10 expression is mediated by Lef1 and Tcf1. Wnt3 signals in the surface ectoderm itself through b-catenin, which leads to AER induction and Fgf8 expression. After Fgf8 is induced, it joins the Wnt3/Fgf positive feedback loop by forming a positive feedback loop with Fgf10. This new Fgf8–Fgf10 positive feedback loop is required for AER maintenance and proximal–distal limb outgrowth.

surface ectoderm to control the formation of the apical ectoderm ridge (AER; see Figure 38.1). AER is a thickened epithelial structure that forms at the DV boundary of the early limb bud, and it is required for limb bud outgrowth. AER formation is a critical developmental event during limb initiation. The removal of either Wnt3 or b-catenin genetically from the early limb ectoderm blocks AER formation (Barrow et al., 2003), and the limb development in these mutant embryos phenocopies that seen in the Fgf10 / and the Lef1 / / Tcf1 / mouse embryos. In the Lef1 / /Tcf1 / mouse embryos, all Wnt/bcatenin activity in the presumptive limb region may be blocked. Because early limb mesoderm as well as Fgf10 induces AER formation that can also be induced by Wnt/b-catenin signaling within the ectoderm without Fgf10, it appears that Wnt3 expression is induced by the early presumptive limb mesoderm, possibly through Fgf10.

II. LIMB BUD OUTGROWTH AND PATTERNING The limb is a three-dimensional structure with three axes: AP (thumb to little finger), proximal–distal (PD; shoulder to finger tip), and DV (back of hand to palm). Proper limb bud development requires patterning along the three axes, which is controlled by three signaling centers. The three signaling centers are established after limb initiation, and then the limb develops autonomously by the coordination among these three signaling centers. AER is the signaling center that directs PD limb outgrowth. The function of AER in PD limb development was identified by the classic experiments performed by John Saunders in 1948 (Saunders, 1948; Summerbell, 1974). In these series of experiments, it was found that AER removal in the chick limb bud led to limb truncation along the PD axis. Earlier AER removal leads to limb truncation at more proximal limb levels. Fgf family members that are expressed in the AER (mainly Fgf8 and Fgf4) have been identified to mediate the function of AER by molecular genetic studies in both the mouse and the

LIMB BUD OUTGROWTH AND PATTERNING

851

chick. In the chick limb bud, heparin beads coated with either FGF4 or FGF8 can rescue limb truncation caused by AER removal when implanted to the distal edge of the AER-stripped limb bud (Crossley et al., 1996; Niswander et al., 1993). In the mouse, the removal of Fgf4 and Fgf8 function in the AER blocks limb outgrowth (Sun et al., 2002). Apart from the substantial growth of the limb bud along the PD axis, the limb bud is also patterned along the PD axis. The humerus forms at the most proximal part of the limb bud (also called the styropod), whereas the radius and ulna form from the middle segment of the limb bud (also called the zeugopod). The distal limb bud (also called the autopod) forms the hand plate, which includes the metatarsal bones, the tarsal bones, and the digits (Figure 38.2). For a long time, patterning along the PD axis was thought to be regulated by the progress zone (i.e., the progress zone model). In this model (Summerbell and Lewis, 1975), cells in the progress zone are progenitors of cartilage and connective tissues (see Chapter 39). Parts of the limb are specified in PD sequence by an autonomous timing mechanism operating in a “progress zone” of undifferentiated growing mesenchyme under the influence of the AER, which serves to keep the cells in the progress zone actively dividing and relatively undifferentiated (Globus and Vethamany-Globus, 1976). If the AER is surgically removed, the limb bud will stop growing, and truncation occurs along the PD axis with earlier AER removal causing more proximal limb truncation. This progress model is consistent with the order of skeletal formation in the limb, which occurs first in the more proximal limb region. A central tenet to the progress zone model is that the cells in the progress zone undergo a progressive change in positional information such that their specification depends on when they leave the progress zone and undergo differentiation. It is thought that cells that leave the progress earlier will adopt more proximal limb fates. Recent fate mapping studies have challenged the progress zone model (Dudley et al., 2002). It is proposed in the new studies that, instead of being specified progressively as the limb bud grows out, PD specification of the limb occurs early during limb development, and patterning along the PD axis is determined at the same time. Cell-labeling experiments performed in early chick limb bud indicate that cells in the limb bud are specified as proximal or distal early on, because early-labeled limb cells were rarely found in two

FIGURE 38.2 Skeletal preparation of a developing forelimb in a mouse embryo (arm, A) and a chick embryo (wing, B). The mouse and chick limbs have similar structures. The forelimbs of the mouse and chick both contain styropod (humerus, H), zeugopod (radius [R] and ulna [U]), and autopod (digits). The mouse limb has five digits (1 through 5), whereas the chick wing has three digits (2 through 4; see color insert).

852

FORMATION OF VERTEBRATE LIMBS

adjacent territories. Then, controlled by the AER, the regions of different cells expand at different times before differentiation to form the complete limb. The fate mapping study started by reassessing cell behavior in the distal chick limb bud immediately after the AER was removed. From these experiments, it is confirmed that there is extensive cell death after AER removal. Interestingly, the domain of cell loss is always the same size, regardless of when the AER is removed. Therefore, later AER removal allows a bigger part of the limb to grow. Further fate mapping studies of the presumed progress zone cells under the AER revealed that these cells died after AER removal instead of taking a more proximal fate predicted by the progress zone model. All of these studies suggest that AER prevents distal limb bud cell death, and that it is required for the expansion of prespecified limb territories along the PD axis. This view is supported by the genetic studies of the crucial Fgf factors expressed in the AER, Fgf4 and Fgf8 (Sun et al., 2002). What was surprising is that a burst of Fgf8 in the nascent AER activated the Fgf4 gene, which was enough to allow some skeletal elements to develop, particularly the humerus. In most cases, more distal elements also form, although both genes were inactivated completely soon after limb bud formation. Fgf signaling from the AER has two important functions. First, it is required at the earliest stages of limb development to establish a limb bud of normal size. Second, it is needed for limb bud cells to survive. The important observation in the study is that, when Fgf4 and Fgf8 are removed at later stages of limb development, there are not enough cells in the limb bud to produce the required skeletal elements. The Fgf4/Fgf8 mutant limbs do not simply lose the distal limb structures. Fgf signaling is not just required for limb outgrowth; it also plays a critical role in the PD patterning of the limb. Patterning along the PD axis of the limb is determined by the interaction of regional specific transcription factors expressed in the limb mesoderm and the signaling pathways in the AER and the limb mesoderm. Restriction of the expression of evolutionally conserved homeobox genes Meis1 and Meis2 to proximal regions of the limb bud is essential for limb development (Capdevila et al., 1999; Mercader et al., 2000). Ectopic Meis2 expression in the distal limb bud severely disrupts limb outgrowth by repressing distal genes, whereas bone morphogenetic proteins (Bmps) and Hoxd genes restrict Meis2 expression to the proximal limb bud. Combinations of Bmps and AER factors are sufficient to distalize proximal limb cells. Retinoic acid (RA) is an upstream activator of the proximal determinant genes Meis1 and Meis2. RA promotes the proximalization of limb cells, and endogenous RA signaling is required to maintain the proximal Meis domain in the limb. RA synthesis and signaling range are restricted to proximal limb domains after limb initiation by the AER activity, which is mainly mediated by Fgfs. Fgfs have a specific function in promoting distalization through the inhibition of RA production and signaling. Although the AER serves to provide the growth signal along the PD axis, the limb bud type is controlled by the mesoderm. In other words, the AER signal is permissive but not instructive for limb development. When the chick wing bud mesoderm is recombined with the leg bud ectoderm, a wing develops (Zwilling, 1955). It was also found that the development of the cross-species recombination of limbs (e.g., chick/rat) was typical of the species contributing the mesoderm (Jorquera and Pugin, 1971). Moreover, if the AP axis of the AER is reversed with reference to the mesoderm, the pattern of mesoderm differentiation is unchanged (Zwilling, 1956a). Indeed, it has been

LIMB BUD OUTGROWTH AND PATTERNING

853

found that Tbx5 and Tbx4 are expressed in the forelimb and hindlimb mesoderms, respectively, but not in the ectoderm. In addition, the hindlimb identity is determined by the mesoderm-specific transcription factor Pitx1 (Lanctot et al., 1999; Logan and Tabin, 1999; Szeto et al., 1999). The interaction between the AER and the limb mesoderm is not one-way. Equally important is the maintenance of AER structure and function by the underlying limb mesoderm. The presumed mesoderm-derived maintenance factor is called the apical ectoderm maintenance factor (AEMF). In the tissue recombination experiments, it was found that an older AER would change its morphology to resemble that of a young AER if it is recombined with a young mesoderm. The normal configuration of AER is also controlled by the limb mesoderm (Zwilling, 1956b). If the limb ectoderm is recombined with nonlimb mesoderm (flank lateral plate or posterior somites), then the AER degenerates within 2 days of the operation. However, if a small piece of limb mesoderm is added beneath a part of the AER, that part of the AER survives (Zwilling, 1972). It has been suggested that the AEMF is not distributed evenly through the mesoderm. Rather, it is more concentrated in the posterior half of the limb bud, which is covered by a thicker and longer AER. Because Fgf10 forms a positive feedback loop with Fgf8 during the limb elongation stage and this feedback loop is required for limb outgrowth and AER maintenance (see Figure 38.1), Fgf10 is obviously qualified to be an AEMF. However, Fgf10 may not be the only AEMF. A classic example of the AEMF deficiency in limb development is found in limb deformity (ld) mutant mice, in which AER degenerates prematurely as a result of a mesodermal defect (Kuhlman and Niswander, 1997). The molecular nature of the defective AEMF in the ld limb has been identified to be a secreted signaling molecule, Gremlin (Khokha et al., 2003; Zuniga et al., 1999). Gremlin is an antagonist of Bmp signaling that is expressed in the distal limb mesoderm. Because Bmp signaling promotes AER degeneration, Gremlin is required for the maintenance of AER integrity by antagonizing Bmp signaling. The second signaling center is the zone of polarizing activity (ZPA), which is a group of mesenchymal cells located at the posterior limb margin and immediately adjacent to the AER. The polarizing activity was also discovered by Saunders (Saunders and Gasseling, 1968), who found that ZPA tissue grafted to the anterior limb bud leads to digit duplication in a way that is a mirror image of the endogenous digits. When limb mesoderm cells are dissociated and packed into the ectoderm jacket, the resulting limb has digits, but their specific identities cannot be determined. However, when a polarizing posterior mesenchyme graft is added, this results in a much more normal skeleton with recognizable digits, with the most posterior digit forming closest to the polarizing posterior mesenchyme graft (MacCabe et al., 1973). MacCabe has also demonstrated that there is a gradient of the polarizing posterior mesenchymal activity in the chick limb bud, and it has been found that such activity appears to be mediated by a diffusible factor (Calandra and MacCabe, 1978; MacCabe and Parker, 1976). It was proposed by Summerbell (1979) that the polarizing posterior mesenchyme emits a diffusible signal that forms a gradient and that specifies the AP axis in a dose-dependent manner. Molecular genetic studies have identified that a vertebrate Hedgehog family member called Sonic hedgehog (Shh) mediates the activity of ZPA. Ectopically expressed Shh in the anterior mouse limb bud, Shh-expressing cells, and Shh-protein–coated beads implanted in the anterior chick limb

854

FORMATION OF VERTEBRATE LIMBS

bud all lead to mirror-image digit duplication (Figure 38.3; Chan et al., 1995; Riddle et al., 1993). The active Shh signal corresponds to a 19-kD N-terminal peptide generated by autoproteolytic cleavage that is modified by the covalent addition of cholesterol and palmitate (Chamoun et al., 2001; Porter et al., 1995). Many studies have underscored the general long-range signaling capacity of Shh, and loss-of-function genetics has established essential Shh functions during embryogenesis, the maintenance of stem cells and disease in vertebrates. Shh has been demonstrated to act as a morphogen that patterns the DV axis of the developing neural tube by forming a morphogen gradient (Ericson et al., 1996). However, in the limb, although there is an absolute requirement of Shh signaling for AP patterning of the distal limb skeleton (because a lack of Shh in various vertebrate species results in the loss of the posterior zeugopod [ulna] and digits 5 to 2 [Chiang et al., 2001]), it is still not a settled issue whether Shh produced by the ZPA also patterns the limb bud by acting as a diffusible morphogen. Shh protein can diffuse from the ZPA to elicit a response at a distance in the limb bud mesenchyme (Lewis et al., 2001). Cells responding to Shh signaling activate the Gli1 transcription factor, but genetic analysis in the mouse shows that Gli1 is not essential for limb bud development (Park et al., 2000). The related Gli2 protein also functions as a positive downstream mediator of Shh signaling, but, again, it is not essential for limb bud development (Mo et al., 1997). Rather, Shh signaling seems to enable the distal progression of limb bud morphogenesis and the formation of the digit arch by inhibiting the proteolytic production of the repressor form of another Gli family member, the Gli3 protein (te Welscher et al., 2002b; Wang et al., 2000), which is expressed primarily in cells that do not express Shh. However, when the direct contribution of ZPA cells to digit primordial was analyzed by genetic fate mapping (Harfe et al., 2004), it was found that ZPA cells give rise to descendants that either remain ZPA cells or that join a distal–anteriorly expanding population of descendants that no longer express Shh. Together, these two populations of Shh descendant cells give rise to the most posterior digits (5 and 4), parts of digit 3, and the ulna. Therefore, digit 2 is the only skeletal element that is not derived from cells that have expressed Shh at some stage. These results challenge the relevance of a spatial morphogen gradient, because digit 2 may be the only one that depends on long-range Shh signaling. Indeed, the reduction of Shh mobility across the limb bud affects digit 2, whereas digits 3 to 5 form normally. In addition, when kinetic studies were performed to address how the identities of digits 3 to 5 are specified, it was revealed that the fates of Shh descendants

FIGURE 38.3 The implantation of Shh-expressing cells in the anterior limb bud leads to mirror image digit duplication. Shh-expressing cell pellets are implanted in the anterior chick limb bud under the apical ectoderm ridge. The ectopic Shh activity causes mirror image digit duplication (from 2–3-4 to 4–3-2–3-4; see color insert).

LIMB BUD OUTGROWTH AND PATTERNING

855

are progressively restricted posteriorly: descendant cells that do not express Shh and that are born early contribute to all three digits, whereas the ZPA cells expressing Shh for the longest time contribute exclusively to digit 5. It appears that limb bud cells somehow acquire a kinetic memory of the signal received. Thus, the length of time that the Shh-expressing cells and their nonexpressing descendants that are exposed to Shh signaling earlier determines the identities of the three posterior digits. This seems to be inconsistent with the spatial morphogen gradient model, according to which long-range Shh signaling specifies digits 4, 3, and 2 (not just digit 2). However, during the dynamic process of AP limb patterning, the cells that remain in ZPA or the non-Shh expressing descendants closer to ZPA obviously receive higher doses of Shh. Therefore, their more posterior identity can also be explained by a spatial Shh gradient model. It is very likely that the digit identity is patterned by both the spatial gradient and the temporal duration of the Shh morphogen. Indeed, when limb bud cells responding to Shh signaling were marked by analyzing the transcriptional activation of Gli1 (a direct transcriptional target of Shh signaling) at specific time points of limb development, it was found that the mesenchymal cells giving rise to digits 5 to 2 and to the ulna responded to Shh signaling (Ahn and Joyner, 2004). However, although the cumulative Shh response was the highest in the posterior mesenchyme (digit 5) and progressively lower toward the anterior, no specific thresholds of response were found as predicted by the morphogen gradient model. By contrast, the Shh responsiveness of the most posterior cells (fated to form digit 5), which are exposed to Shh the longest, was reduced with time. In addition, the number of marked Gli1-expressing cells is reduced, and their distribution is altered in the limb buds of mouse embryos lacking the Gli2 gene, despite the fact that Gli2deficient limbs develop completely normally. These results, together with the analysis of Gli3-deficient limb buds, indicate that it is not the positive response to Shh as mediated by Gli1 and Gli2 but rather its inhibitory effects on Gli3 repressor (Gli3R) formation that determines digit identities. In particular, the most anterior digit 1 is specified in the absence of Shh by high levels of Gli3R, whereas cumulative high levels of Shh response effectively repress Gli3R formation and thereby specify the most posterior digit 5. Taken together, these studies indicate that the vertebrate limb bud is patterned by a kinetic memory integrating the cumulative length and strength of Shh signaling that cells receive. These temporal and spatial gradients pattern digits 2 to 5 and the ulna. These processes are regulated by controlling the inhibition of Gli3R formation at different locations along the limb AP axis and over time. Although Shh plays a critical role in the AP patterning of the limb, the AP axis of the limb appears to be established before Shh signaling. The expression of Shh in the posterior limb bud is controlled by the basic helix–loop–helix transcription factor dhand. Gli3 restricts dhand expression to posterior mesenchyme before the activation of Shh signaling in mouse limb buds. In turn, dhand excludes anterior genes such as Gli3 and Alx4 from posterior mesenchyme. These interactions polarize the AP axis of the newly formed limb bud mesenchyme before Shh signaling (te Welscher et al., 2002a). Twist1 is also a basic helix–loop–helix transcription factor that is expressed in the developing limb bud and that is required for the maintenance of the AER. Autosomal-dominant mutations in the twist1 gene are associated with limb and craniofacial defects in humans with Saethre–Chotzen syndrome (Jabs,

856

FORMATION OF VERTEBRATE LIMBS

2004). The ectopic expression of dhand phenocopies Twist1 loss of function in the limb, and the two factors have a gene dosage-dependent antagonistic interaction. The dimerization of Twist1 and dhand can be modulated by protein kinase A- and protein phosphatase 2A-regulated phosphorylation. The third signaling center is the non-AER limb ectoderm that covers the limb bud. It sets up the DV polarity of not only the ectoderm but also the underlying mesoderm of the limb (reviewed by Niswander, 2002; Tickle, 2003). In the early embryo, before the limb bud forms, the DV polarity of the future limb is determined by the mesoderm as shown by ectoderm–mesoderm recombination experiments (Geduspan and MacCabe, 1989). However, before limb bud initiation, the ability to determine the DV polarity is transferred from the mesoderm to the ectoderm. If the DV polarity of the limb ectoderm is reversed relative to the mesenchyme in the early limb bud, the DV polarity of the mesenchyme changes in accordance with that of the ectoderm. At much later stages, when skeletal morphogenesis starts, the capacity of the mesoderm to respond to ectodermal control is lost. Wnt and Bmp signaling are required to control DV limb polarity in both the limb ectoderm and the mesoderm. Wnt7a is expressed specifically in the dorsal limb ectoderm, and it activates the expression of Lmx1b, which encodes a dorsal-specific transcription factor that determines the dorsal identity (Figure 38.4; Parr et al., 1993; Riddle et al., 1995). Mutant mice with a targeted Wnt7a null mutation develop ventralized limbs, and the expression of Lmx1b is lost (Cygan et al., 1997; Parr and McMahon, 1995). On the other hand, the ectopic expression of Wnt7a in the ventral limb dorsalizes the limb and activates Lmx1b expression. During normal limb development, Wnt7a expression in the ventral ectoderm is suppressed by En-1, which encodes a transcription factor that is expressed specifically in the ventral ectoderm (see Figure 38.4; Loomis et al., 1996). In the En-1 loss-of-function mutant limb, Wnt7a is ectopically expressed in the ventral ectoderm, and the limb is dorsalized. In the En-1 and Wnt7a double mutant mouse, the limb is similarly ventralized like it is in the Wnt7a single mutant (Cygan et al., 1997; Loomis et al., 1998). These studies indicate that either ventral cell fates are default (i.e., independent of active gene regulation by Wnt7a and En-1) or that the function of Wnt7a is to actively suppress another ventralizing regulatory pathway.

FIGURE 38.4

A schematic section of the limb bud along the dorsal–ventral axis. Wnt7a expressed in the dorsal ectoderm signals to the distal limb mesoderm to determine the dorsal limb identity by activating the expression of Lmx1b. En-1 is expressed in the ventral ectoderm, and it is required for ventral limb development by inhibiting Wnt7a expression. DE, Dorsal ectoderm; VE, ventral ectoderm.

LIMB BUD OUTGROWTH AND PATTERNING

857

It appears that this ventralizing regulatory pathway is mediated by Bmp signaling. During early limb development in the chick, Bmp2, Bmp4, and Bmp7 are expressed in the mesenchyme in an unrestricted manner along the DV axis and in the AER (Francis et al., 1994; Francis-West et al., 1995). However, in the ectoderm, these Bmps are preferentially expressed in the ventral ectoderm in a way that is coincident with En-1 and that is in a complementary pattern to that of Wnt7a in the dorsal ectoderm at the time the ectoderm provides DV information to the underlying mesenchyme. In the mouse limb bud, Bmp2 is also found to be expressed in the early ventral limb ectoderm (Lyons et al., 1990). Therefore, the ventral limb mesoderm and the ectoderm receive more Bmp signaling. These Bmps signal through Bmp receptor IA (BMPR-IA) in the limb ectoderm to establish normal DV limb pattern by activating En-1 expression in the ventral ectoderm (Ahn et al., 2001; Pizette et al., 2001). In both the mouse and the chick embryonic limb bud, the loss of Bmp signaling results in the loss of En-1 expression and dorsalized limb, whereas activated Bmp signaling leads to ventralized limb with ectopic En-1 expression in the dorsal ectoderm. It appears that the effects of Bmp signaling are mediated by Msx1 and Msx2, which are two transcription factors that are also themselves transcriptionally regulated by Bmp signaling. The loss of both Msx1 and Msx2 function in the mouse embryo also leads to the loss of En-1 expression in the ventral ectoderm, which results in the expansion of Wnt7a expression, which in turn causes an expansion of Lmx1b expression into the ventral mesenchyme, thus leading to limb dorsalization (Lallemand et al., 2005). Therefore, the function of Bmp signaling in the early limb ectoderm is upstream of En-1 for controlling DV limb polarity. It is interesting to note that, like Bmp signaling, the Wnt/b-catenin signaling pathways also act directly in the limb ectoderm to control DV patterning by controlling the expression of En-1 (Barrow et al., 2003). In contrast with Wnt7a, which is a dorsalizing factor, it appears that Wnt3, which is expressed throughout the early limb ectoderm, signals through b-catenin to ventralize the limb. Wnt3/b-catenin signaling is required in the ventral ectoderm for the expression of En-1. A loss of either Wnt3 or b-catenin in the ventral limb ectoderm resulted in a dorsalized limb that can be attributed to loss of En-1 expression. However, it appears that Bmps and Wnt/b-catenin also signal to the limb mesoderm directly to control DV patterning. When BmpR-IA is specifically inactivated only in the mouse limb bud mesoderm, the distal limb is also dorsalized without altering the expression of Wnt7a and En-1 in the limb ectoderm (Ovchinnikov et al., 2006). Likewise, in the mouse limb bud mesoderm, the loss of b-catenin (like the loss of Wnt7a in the dorsal ectoderm) leads to limb ventralization, whereas activated b-catenin, like Wnt7a ectopic expression, causes limb dorsalization. Because Wnt7a signals to the limb mesoderm directly to control Lmx1b expression and DV limb patterning, it is possible that Wnt7a signals through b-catenin in the limb mesoderm and the Wnt/b-catenin and Bmp signaling pathways antagonize each other in controlling the expression of Lmx1b directly in the mesoderm. The three signaling centers that control growth and patterning along the three axes are not functionally independent. The interaction of these three signaling centers ensures that three-dimensional growth and patterning are coordinated during vertebrate limb development (Figure 38.5). It has been shown before that polarizing posterior mesenchyme must be grafted under the AER

858

FORMATION OF VERTEBRATE LIMBS

FIGURE 38.5

Patterning along the three limb axes is coordinately regulated by interactions among the signaling pathways. Both dorsal–ventral (Wnt7a) and proximal–distal (Fgfs) signals regulate anterior–posterior limb patterning by controlling Shh expression. The anterior–posterior signal Shh is also required for proximal–distal limb outgrowth by regulating the expression of Fgfs in the apical ectoderm ridge through Gremlin. (See color insert.)

in the anterior limb to cause mirror image digit duplication (Tickle et al., 1975); this highlights an interaction between the AP and PD signaling centers. It has also been demonstrated that the disruption of the dorsal ectoderm, which controls DV patterning, results in the shortening of limb skeletal elements along the PD axis (Martin and Lewis, 1986). This suggests that the DV signaling center also interacts with the PD signaling center. At the molecular level, it is now clear that the three signaling centers indeed interact with each other through interactions of the mediating signaling molecules (see Figure 38.5). First, there is a positive feedback loop between Shh and the Fgfs expressed in the AER, which connects AP limb patterning with PD limb outgrowth. Fgf signaling from the AER is required for Shh expression, and Shh signals through Gremlin to maintain the normal expression of Fgfs in the AER (Khokha et al., 2003; Laufer et al., 1994; Niswander et al., 1994). Second, the dorsalizing signal Wnt7a is required for maintaining the expression of Shh that patterns the AP axis (Parr and McMahon, 1995; Yang and Niswander, 1995).

III. LIMB DEVELOPMENT AND DISEASES The three-dimensional limb development is orchestrated by the proper regulation of cell signaling and transcriptional networks. Molecular genetic studies of limb development in model organisms (mainly the mouse and the chick) have provided significant insight in both mutation identification and pathological mechanisms of human congenital limb deformities. In many cases, because limb deformities are not associated with other more detrimental defects, there has accumulated a wealth of clinical descriptions of different limb abnormalities. Because most of the signaling pathways controlling limb development also play roles in other developing processes, identifying the molecular nature of limb deformities has also enhanced our understanding of congenital diseases affecting other systems, such as the kidney and the central nervous system.

LIMB DEVELOPMENT AND DISEASES

859

Similar to what has been shown in mouse genetic studies, mutations in the WNT, FGF, and SHH signaling pathways in humans also affect limb formation, outgrowth, and patterning along the three axes. Wnt3 is required for limb initiation in mouse. In humans, a nonsense mutation that truncates WNT3 protein at its amino terminus has been found to be a very likely cause of the rare human genetic disorder tetra-amelia (Online Mendelian Inheritance in Man [OMIM] #273395) in four affected fetuses of a consanguineous Turkish family (Niemann et al., 2004). Tetra-amelia is characterized by the complete absence of all four limbs and other anomalies in craniofacial and urogenital development. FGF signaling plays multiple roles in limb development. The significance of FGF signaling in the human developing limb is highlighted by the findings that mutations in FGF receptors lead to limb and skeletal deformities in Apert syndrome (AS; OMIM# 101200), Jackson–Weiss syndrome (JWS; OMIM# 123150), and Pfeiffer syndrome (PS; OMIM#101600; reviewed by Naski and Ornitz, 1998; Ornitz and Marie, 2002). AS and JWS result from mutations in FGFR2. These syndromes are characterized by syndactyly of the hands and feet and the premature fusion of the cranial sutures. PS results from a single mutation in FGFR1 or one of several mutations in FGFR2. This syndrome is characterized by broad great toes and thumbs in addition to the premature fusion of cranial sutures. Mechanistic studies of the FGFR mutations in these syndromes further indicate that some and perhaps all of the mutations causing JWS and PS are the result of activating mutations in an FGFR. In addition, it has been found that the S252W mutation of AS alters the ligand-binding affinity of FGFR2 (Yu et al., 2000). The increased affinity for FGF ligands caused by this mutation may result in the enhanced activation of the receptor in the mesenchymal cells that are destined to form the digits. The SHH signaling pathway controls AP limb patterning. One of the most frequently observed human limb malformations is preaxial polydactyly (PPD), which involves disrupted AP limb patterning. Patients with PPD (particularly triphalangeal thumb polysyndactyly; OMIM#190605) and preaxial polydactyly type II (OMIM#174500) have extra digits on the sides of their thumbs or great toes, just like what has been observed in mouse mutants with ectopic Shh expression in the anterior limb bud. From the insight gained from the molecular genetic studies of the Sasquatch (Sharpe et al., 1999) and Hemimelic extratoe (Clark et al., 2000) mouse mutants with preaxial supernumerary digits, it is found that point mutations in the long-range, limb-specific regulatory element of the human SHH gene are responsible for the human limb abnormality of PPD (Lettice et al., 2002). GLI3 transduces the Hedgehog signal, and mutations in GLI3 result in Greig cephalopolysyndactyly syndrome (OMIM# 175700) or Pallister–Hall syndrome (OMIM# 146510). Mutations such as deletions or translocations resulting in the haploinsufficiency of GLI3 are found in association with Greig cephalopolysyndactyly syndrome (Brueton et al., 1988; Pettigrew et al., 1991; Vortkamp et al., 1991), whereas mutations resulting in dominantnegative GLI3 are found in Pallister–Hall syndrome (Johnston et al., 2005; Kang et al., 1997). In addition, a mutation at codon 764 of the GLI3 gene, which is three-prime to the conserved domain called post zinc finger-1, is found to cause postaxial polydactyly type A1 (OMIM# 174200; Radhakrishna et al., 1997). This condition is the result of an autosomal trait that is characterized by an extra digit in the ulnar and/or fibular side of the upper

860

FORMATION OF VERTEBRATE LIMBS

and/or lower extremities. The extra digit is usually well formed, and it articulates with the fifth or extrametacarpal/metatarsal bone. Thus, different truncated GLI3 proteins are associated with different clinical syndromes, thereby highlighting the different functions of specific GLI3 variant proteins during limb development. Because GLI2 is also required to transduce Hedgehog signaling, mutations in the human GLI2 gene have also been found to cause postaxial polydactyly apart from developmental defects affecting the pituitary gland and the brain (Roessler et al., 2003). DV limb patterning is controlled by Wnt7a/Lmx1b signaling in mice. In humans, there is evidence that the Fuhrmann syndrome (OMIM# 228930) and the Al-Awadi/Raas–Rothschild/Schinzel phocomelia syndrome (OMIM # 276820) are caused by mutations in the WNT7A gene (Woods et al., 2006). These two syndromes have been considered to have distinct limb malformations characterized by various degrees of limb aplasia/hypoplasia and joint dysplasia. It has been suggested that mutations causing a partial loss of WNT7A function lead to Fuhrmann syndrome, whereas null mutations lead to the more severe limb truncation phenotypes observed in Al-Awadi/Raas– Rothschild/Schinzel phocomelia syndrome. Again, the findings in human limb malformations provide insight into the role of WNT7A in multiple aspects of vertebrate limb development. Because LMX1B is required for determining dorsal limb identity in mice, mutations resulting in the haploinsufficiency of the human LMX1B gene have been found to cause the nail–patella syndrome (OMIM# 161200), which affects DV limb development and which is characterized by dysplasia of the nails and absent or hypoplastic patellae. Other limb features of the nail–patella syndrome include elbow stiffness with limitations involving pronation and supination. Consistent with the findings in the Lmx1b / mutant mice (Chen et al., 1998; Pressman et al., 2000) is the fact that mutations in the LMX1B gene also cause renal abnormalities and open-angle glaucoma (OMIM# 137760).

IV. CONCLUSIONS AND PERSPECTIVES Studies of classic embryology and molecular genetics in model organisms have led to the identification of molecular pathways controlling coordinated threedimensional limb development. These studies have also provided insights into the understanding of the molecular and pathologic mechanisms of congenital human limb defects. The general principles will be relevant to devising new approaches for tissue repair. In addition, because of the pleiotrophic effects of the limb development pathways in other developmental processes as what has been shown in mouse and human genetic studies, the developing limb has provided an excellent model system for understanding how different signaling pathways operate and interact with each other in embryonic development. Although many genes and pathways with roles in limb development have been identified, we are still far way from getting a full picture of how limb development is controlled at the molecular level. There are considerable gaps in our current understanding. To gain clarity with regard to the cellular basis of limb development, we also need to find creative new ways to visualize the behavior of extracellular signaling molecules and to measure their concentrations and the cellular responses they trigger. Furthermore, it is not yet clear how the three-dimensional positional information, which is controlled by

861

GLOSSARY

distinct signaling pathways, is integrated into the cells of the developing limb. Both the mouse and human genomes have been sequenced, and recent developments in functional genomic studies in mice (e.g., chemical and insertional mutagenesis and large-scale gene targeting) will provide a tremendous amount of new information for the understanding of limb development. These studies—in combination with the ever-improving and powerful techniques of the genetic mapping of human disease—have allowed for the advancement of our understanding of limb development at an unprecedented speed.

SUMMARY Vertebrate limb development has been a fertile field for the understanding of the functional mechanisms of cell–cell signaling in the control of embryonic development. The quick advancement of the molecular genetic studies of vertebrate limb development has benefited tremendously from classic embryological experiments in the chick. These studies have revealed the molecular networks that control limb development in the following areas:

 Limb bud formation at specific position along the AP body axis may be determined by the expression of Hox genes.

 Limb initiation is controlled by the transcription factors Tbx4 (hindlimb) and Tbx5 (forelimb) and by the Wnt/Fgf signaling pathways.

 Limb outgrowth and patterning along the three axes (PD, AP, and DV) are

 

controlled by three signaling centers (AER, ZPA, and non-AER limb ectoderm, respectively). Coordination among the three signaling centers is made through interactions among the signaling pathways that mediate the function of the respective signaling centers. Signaling pathways controlling limb development also often play critical roles in the regulation of other important developmental processes. Genetic variations in humans that disrupt limb development lead to congenital limb malformations and other defects.

GLOSSARY Bardet–Biedl syndrome A condition that is characterized mainly by obesity, pigmentary retinopathy, polydactyly, mental retardation, hypogonadism, and renal failure, in fatal cases. Brachydactyly A term which literally means “shortness of the fingers and toes (digits).” The shortness is relative to the length of other long bones and other parts of the body. Gorlin syndrome An autosomal-dominant cancer syndrome. Patients with this rare syndrome often have anomalies of multiple organs, many of which are subtle. Patients with Gorlin syndrome have the propensity to develop multiple neoplasms, including basal cell carcinomas and medulloblastoma, and they often demonstrate extreme sensitivity to ionizing radiation, including sunlight.

862

FORMATION OF VERTEBRATE LIMBS

Extra digits on the hands or feet also occur among patients with this syndrome. Mutations in the Patched gene have been found to cause this syndrome. Polydactyly The anatomic variant that involves more than the usual number of digits on the hands and/or feet; also known as hyperdactyly. When each hand or foot has six digits, it is sometimes called hexadactyly. Syndactyly A condition in which the fingers fail to separate into individual appendages. This separation occurs during embryological development.

REFERENCES Agarwal P, Wylie JN, Galceran J, et al: Tbx5 is essential for forelimb bud initiation following patterning of the limb field in the mouse embryo, Development 130:623–633, 2003. Ahn K, Mishina Y, Hanks MC, et al: BMPR-IA signaling is required for the formation of the apical ectodermal ridge and dorsal-ventral patterning of the limb, Development 128:4449–4461, 2001. Ahn S, Joyner AL: Dynamic changes in the response of cells to positive hedgehog signaling during mouse limb patterning, Cell 118:505–516, 2004. Barrow JR, Thomas KR, Boussadia-Zahui O, et al: Ectodermal Wnt3/beta-catenin signaling is required for the establishment and maintenance of the apical ectodermal ridge, Genes Dev 17:394–409, 2003. Brueton L, Huson SM, Winter RM, Williamson R: Chromosomal localisation of a developmental gene in man: direct DNA analysis demonstrates that Greig cephalopolysyndactyly maps to 7p13, Am J Med Genet 31:799–804, 1988. Calandra AJ, MacCabe JA: The in vitro maintenance of the limb-bud apical ridge by cell-free preparations, Dev Biol 62:258–269, 1978. Capdevila J, Tsukui T, Rodriquez Esteban C, et al: Control of vertebrate limb outgrowth by the proximal factor Meis0002 and distal antagonism of BMPs by Gremlin, Mol Cell 4:839–849, 1999. Chamoun Z, Mann RK, Nellen D, et al: Skinny hedgehog, an acyltransferase required for palmitoylation and activity of the hedgehog signal, Science 293:2080–2084, 2001. Chan DC, Laufer E, Tabin C, Leder P: Polydactylous limbs in Strong’s Luxoid mice result from ectopic polarizing activity, Development 121:1971–1978, 1995. Chen H, Lun Y, Ovchinnikov D, et al: Limb and kidney defects in Lmx1b mutant mice suggest an involvement of LMX1B in human nail patella syndrome, Nat Genet 19:51–55, 1998. Chiang C, Litingtung Y, Harris MP, et al: Manifestation of the limb prepattern: limb development in the absence of sonic hedgehog function, Dev Biol 236:421–435, 2001. Clark RM, Marker PC, Kingsley DM: A novel candidate gene for mouse and human preaxial polydactyly with altered expression in limbs of Hemimelic extra-toes mutant mice, Genomics 67:19–27, 2000. Cohn MJ, Izpisua-Belmonte JC, Abud H, et al: Fibroblast growth factors induce additional limb development from the flank of chick embryos, Cell 80:739–746, 1995. Cohn MJ, Patel K, Krumlauf R, et al: Hox9 genes and vertebrate limb specification, Nature 387:97–101, 1997. Crossley PH, Minowada G, MacArthur CA, Martin GR: Roles for FGF8 in the induction, initiation, and maintenance of chick limb development, Cell 84:127–136, 1996. Cygan JA, Johnson RL, McMahon AP: Novel regulatory interactions revealed by studies of murine limb pattern in Wnt-7a and En-1 mutants, Development 124:5021–5032, 1997. Davidson DR, Crawley A, Hill RE, Tickle C: Position-dependent expression of two related homeobox genes in developing vertebrate limbs, Nature 352:429–431, 1991. Detwiler SR: On the time of determination of the antero-posterior axis of the forelimb in Amblystoma, J Exp Zool 64:405–414, 1933. Dhouailly D, Kieny M: The capacity of the flank somatic mesoderm of early bird embryos to participate in limb development, Dev Biol 28:162–175, 1972.

REFERENCES

863 Dudley AT, Ros MA, Tabin CJ: A re-examination of proximodistal patterning during vertebrate limb development, Nature 418:539–544, 2002. Ericson J, Morton S, Kawakami A, et al: Two critical periods of Sonic Hedgehog signaling required for the specification of motor neuron identity, Cell 87:661–673, 1996. Francis PH, Richardson MK, Brickell PM, Tickle C: Bone morphogenetic proteins and a signalling pathway that controls patterning in the developing chick limb, Development 120:209–218, 1994. Francis-West PH, Robertson KE, Ede DA, et al: Expression of genes encoding bone morphogenetic proteins and sonic hedgehog in talpid (ta3) limb buds: their relationships in the signalling cascade involved in limb patterning, Dev Dyn 203:187–197, 1995. Galceran J, Farinas I, Depew MJ, et al: Wnt3a-/–like phenotype and limb deficiency in Lef1( / ) Tcf1( / ) mice, Genes Dev 13:709–717, 1999. Geduspan JS, MacCabe JA: Transfer of dorsoventral information from mesoderm to ectoderm at the onset of limb development, Anat Rec 224:79–87, 1989. Globus M, Vethamany-Globus S: An in vitro analogue of early chick limb bud outgrowth, Differentiation 6:91–96, 1976. Hamburger V: Morphogenetic and axial self-differentiation of transplanted limb primordia of 2-day chick embryos, J Exp Zool 77:337–489, 1938. Harfe BD, Scherz PJ, Nissim S, et al: Evidence for an expansion-based temporal Shh gradient in specifying vertebrate digit identities, Cell 118:517–528, 2004. Izpisua-Belmonte JC, Tickle C, Dolle P, et al: Expression of the homeobox Hox-4 genes and the specification of position in chick wing development, Nature 350:585–589, 1991. Jabs EW: TWIST and Saethre-Chotzen syndrome,, Oxford, UK, 2004, Oxford University Press. Johnston JJ, Olivos-Glander I, Killoran C, et al: Molecular and clinical analyses of Greig cephalopolysyndactyly and Pallister-Hall syndromes: robust phenotype prediction from the type and position of GLI3 mutations, Am J Hum Genet 76:609–622, 2005. Jorquera B, Pugin E: Behavior of the mesoderm and ectoderm of the limb bud in the exchanges between chicken and rat, C R Acad Sci Hebd Seances Acad Sci D 272:1522–1525, 1971. Kang S, Graham JMJr, Olney AH, Biesecker LG: GLI3 frameshift mutations cause autosomal dominant Pallister-Hall syndrome, Nat Genet 15:266–268, 1997. Khokha MK, Hsu D, Brunet LJ, et al: Gremlin is the BMP antagonist required for maintenance of Shh and Fgf signals during limb patterning, Nat Genet 34:303–307, 2003. Kieny M: Inductive role of the mesoderm in the early differentiation of the limb bud in the chick embryo, J Embryol Exp Morphol 8:457–467, 1960. Kieny M: Variation in the inductive capacity of mesoderm and the competence of ectoderm during primary induction in the chick embryo limb bud, Arch Anat Microsc Morphol Exp 57:401–418, 1968. Kuhlman J, Niswander L: Limb deformity proteins: role in mesodermal induction of the apical ectodermal ridge, Development 124:133–139, 1997. Lallemand Y, Nicola MA, Ramos C, et al: Analysis of Msx1; Msx2 double mutants reveals multiple roles for Msx genes in limb development, Development 132:3003–3014, 2005. Lanctot C, Moreau A, Chamberland M, et al: Hindlimb patterning and mandible development require the Ptx1 gene, Development 126:1805–1810, 1999. Laufer E, Nelson CE, Johnson RL, et al: Sonic hedgehog and Fgf-4 act through a signaling cascade and feedback loop to integrate growth and patterning of the developing limb bud, Cell 79:993–1003, 1994. Lettice LA, Horikoshi T, Heaney SJ, et al: Disruption of a long-range cis-acting regulator for Shh causes preaxial polydactyly, Proc Natl Acad Sci U S A 99:7548–7553, 2002. Lewis PM, Dunn MP, McMahon JA, et al: Cholesterol modification of sonic hedgehog is required for long-range signaling activity and effective modulation of signaling by Ptc1, Cell 105:599–612, 2001. Logan M, Tabin CJ: Role of Pitx1 upstream of Tbx4 in specification of hindlimb identity, Science 283:1736–1739, 1999. Loomis CA, Harris E, Michaud J, et al: The mouse Engrailed-1 gene and ventral limb patterning, Nature 382:360–363, 1996. Loomis CA, Kimmel RA, Tong CX, et al: Analysis of the genetic pathway leading to formation of ectopic apical ectodermal ridges in mouse Engrailed-1 mutant limbs, Development 125:1137–1148, 1998. Lyons KM, Pelton RW, Hogan BL: Organogenesis and pattern formation in the mouse: RNA distribution patterns suggest a role for bone morphogenetic protein-2A (BMP-2A), Development 109:833–844, 1990.

864

FORMATION OF VERTEBRATE LIMBS

MacCabe JA, Parker BW: Polarizing activity in the developing limb of the Syrian hamster, J Exp Zool 195:311–317, 1976. MacCabe JA, Saunders JWJr, Pickett M: The control of the anteroposterior and dorsoventral axes in embryonic chick limbs constructed of dissociated and reaggregated limb-bud mesoderm, Dev Biol 31:323–335, 1973. Martin P, Lewis J: Normal development of the skeleton in chick limb buds devoid of dorsal ectoderm, Dev Biol 118:233–246, 1986. Mercader N, Leonardo E, Piedra ME, et al: Opposing RA and FGF signals control proximodistal vertebrate limb development through regulation of Meis genes, Development 127:3961–3970, 2000. Min H, Danilenko DM, Scully SA, et al: Fgf-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless, Genes Dev 12:3156–3161, 1998. Mo R, Freer AM, Zinyk DL, et al: Specific and redundant functions of Gli2 and Gli3 zinc finger genes in skeletal patterning and development, Development 124:113–123, 1997. Naiche LA, Papaioannou VE: Loss of Tbx4 blocks hindlimb development and affects vascularization and fusion of the allantois, Development 130:2681–2693, 2003. Naski MC, Ornitz DM: FGF signaling in skeletal development, Front Biosci 3:d781–d794, 1998. Niemann S, Zhao C, Pascu F, et al: Homozygous WNT3 mutation causes tetra-amelia in a large consanguineous family, Am J Hum Genet 74:558–563, 2004. Niswander L: Interplay between the molecular signals that control vertebrate limb development, Int J Dev Biol 46:877–881, 2002. Niswander L, Jeffrey S, Martin GR, Tickle C: A positive feedback loop coordinates growth and patterning in the vertebrate limb, Nature 371:609–612, 1994. Niswander L, Tickle C, Vogel A, et al: FGF-4 replaces the apical ectodermal ridge and directs outgrowth and patterning of the limb, Cell 75:579–587, 1993. Ornitz DM, Marie PJ: FGF signaling pathways in endochondral and intramembranous bone development and human genetic disease, Genes Dev 16:1446–1465, 2002. Ovchinnikov DA, Selever J, Wang Y, et al: BMP receptor type IA in limb bud mesenchyme regulates distal outgrowth and patterning, Dev Biol 295:103–115, 2006. Park HL, Bai C, Platt KA, et al: Mouse Gli1 mutants are viable but have defects in SHH signaling in combination with a Gli2 mutation, Development 127:1593–1605, 2000. Parr BA, McMahon AP: Dorsalizing signal Wnt-7a required for normal polarity of D-V and A-P axes of mouse limb, Nature 374:350–353, 1995. Parr BA, Shea MJ, Vassileva G, McMahon AP: Mouse Wnt genes exhibit discrete domains of expression in the early embryonic CNS and limb buds, Development 119:247–261, 1993. Pettigrew AL, Greenberg F, Caskey CT, Ledbetter DH: Greig syndrome associated with an interstitial deletion of 7p: confirmation of the localization of Greig syndrome to 7p13, Hum Genet 87:452–456, 1991. Pizette S, Abate-Shen C, Niswander L: BMP controls proximodistal outgrowth, via induction of the apical ectodermal ridge, and dorsoventral patterning in the vertebrate limb, Development 128:4463–4474, 2001. Porter JA, von Kessler DP, Ekker SC, et al: The product of hedgehog autoproteolytic cleavage active in local and long-range signalling, Nature 374:363–366, 1995. Pressman CL, Chen H, Johnson RL: LMX1B, a LIM homeodomain class transcription factor, is necessary for normal development of multiple tissues in the anterior segment of the murine eye, Genesis 26:15–25, 2000. Radhakrishna U, Wild A, Grzeschik KH, Antonarakis SE: Mutation in GLI3 in postaxial polydactyly type A, Nat Genet 17:269–271, 1997. Riddle RD, Ensini M, Nelson C, et al: Induction of the LIM homeobox gene Lmx1 by WNT7a establishes dorsoventral pattern in the vertebrate limb, Cell 83:631–640, 1995. Riddle RD, Johnson RL, Laufer E, Tabin C: Sonic hedgehog mediates the polarizing activity of the ZPA, Cell 75:1401–1416, 1993. Roessler E, Du YZ, Mullor JL, et al: Loss-of-function mutations in the human GLI2 gene are associated with pituitary anomalies and holoprosencephaly-like features, Proc Natl Acad Sci U S A 100:13424–13429, 2003. Saunders JWJ: The proximo-distal sequence of origin of the parts of the chick wing and the role of the ectoderm, J Exp Zool 108:363–403, 1948. Saunders JWJ, Gasseling MT: Ectoderm-mesenchymal interaction in the origin of wing symmetry, In R Fleischmajer R, RE Billingham RE, , editors: Epithelia-mesenchymal interactions, Baltimore, 1986, Williams and Wilkins78–97.

RECOMMENDED RESOURCES

865

Sekine K, Ohuchi H, Fujiwara M, et al: Fgf10 is essential for limb and lung formation, Nat Genet 21:138–141, 1999. Sharpe J, Lettice L, Hecksher-Sorensen J, et al: Identification of sonic hedgehog as a candidate gene responsible for the polydactylous mouse mutant Sasquatch, Curr Biol 9:97–100, 1999. Summerbell D: A quantitative analysis of the effect of excision of the AER from the chick limb bud, J Embryol Exp Morphol 32:651–660, 1974. Summerbell D: The zone of polarizing activity: evidence for a role in normal chick limb morphogenesis, J Embryol Exp Morphol 50:217–233, 1979. Summerbell D, Lewis JH: Time, place and positional value in the chick limb-bud, J Embryol Exp Morphol 33:621–643, 1975. Sun X, Mariani FV, Martin GR: Functions of FGF signalling from the apical ectodermal ridge in limb development, Nature 418:501–508, 2002. Szeto DP, Rodriguez-Esteban C, Ryan AK, et al: Role of the Bicoid-related homeodomain factor Pitx1 in specifying hindlimb morphogenesis and pituitary development, Genes Dev 13:484–494, 1999. Takeuchi JK, Koshiba-Takeuchi K, Suzuki T, et al: Tbx5 and Tbx4 trigger limb initiation through activation of the Wnt/Fgf signaling cascade, Development 130:2729–2739, 2003. te Welscher P, Fernandez-Teran M, Ros MA, Zeller R: Mutual genetic antagonism involving GLI3 and dHAND prepatterns the vertebrate limb bud mesenchyme prior to SHH signaling, Genes Dev 16:421–426, 2002a. te Welscher P, Zuniga A, Kuijper S, et al: Progression of vertebrate limb development through SHH-mediated counteraction of GLI3, Science 298:827–830, 2002b. Tickle C: Patterning systems—from one end of the limb to the other, Dev Cell 4:449–458, 2003. Tickle C, Summerbell D, Wolpert L: Positional signalling and specification of digits in chick limb morphogenesis, Nature 254:199–202, 1975. Vortkamp A, Gessler M, Grzeschik KH: GLI3 zinc-finger gene interrupted by translocations in Greig syndrome families, Nature 352:539–540, 1991. Wang B, Fallon JF, Beachy PA: Hedgehog-regulated processing of Gli3 produces an anterior/posterior repressor gradient in the developing vertebrate limb, Cell 100:423–434, 2000. Woods CG, Stricker S, Seemann P, et al: Mutations in WNT7A cause a range of limb malformations, including Fuhrmann syndrome and Al-Awadi/Raas-Rothschild/Schinzel phocomelia syndrome, Am J Hum Genet 79:402–408, 2006. Yang Y, Niswander L: Interaction between the signaling molecules WNT7a and SHH during vertebrate limb development: dorsal signals regulate anteroposterior patterning, Cell 80:939–947, 1995. Yu K, Herr AB, Waksman G, Ornitz DM: Loss of fibroblast growth factor receptor 2 ligand-binding specificity in Apert syndrome, Proc Natl Acad Sci U S A 97:14536–14541, 2000. Zuniga A, Haramis AP, McMahon AP, Zeller R: Signal relay by BMP antagonism controls the SHH/FGF4 feedback loop in vertebrate limb buds, Nature 401:598–602, 1999. Zwilling E: Ectoderm-mesoderm relationship in the development of the chick embryo limb bud, J Exp Zool 128:423–441, 1955. Zwilling E: Interaction between limb bud ectoderm and mesoderm in the chick embryo.I. Axis establishment, J Exp Zool 132:157–171, 1956a. Zwilling E: Interaction between limb bud ectoderm and mesoderm in the chick embryo.II. Experimental limb duplication, J Exp Zool 132:173–188, 1956b. Zwilling E: Limb morphogenesis, Dev Biol 28:12–17, 1972.

RECOMMENDED RESOURCES Mariani FV, Martin GR: Deciphering skeletal patterning: clues from the limb, Nature 423:319–325, 2003. Niswander L: Interplay between the molecular signals that control vertebrate limb development, Int J Dev Biol 46:877–881, 2002. Tickle C: Making digit patterns in the vertebrate limb, Nat Rev Mol Cell Biol 7:45–53, 2006.

39

SKELETAL DEVELOPMENT PETER G. ALEXANDER,1 AMANDA T. BOYCE,2 and ROCKY S. TUAN1 1 Cartilage Biology and Orthopaedics Branch, National Institute of Arthritis, and Musculoskeletal and Skin Diseases, National Institutes of Health, United States Department of Health and Human Services, Bethesda, MD 2 Musculoskeletal Diseases Branch, National Institute of Arthritis, and Musculoskeletal and Skin Diseases, National Institutes of Health, United States Department of Health and Human Services, Bethesda, MD

INTRODUCTION Because the skeleton is present only in vertebrates, the use of alternative model systems to study skeletal development is limited. Although useful paradigms and conserved cellular and molecular pathways in morphogenesis and development have been gained from studies of invertebrates, this chapter will focus on the genetics of skeletal formation in humans, mice, and chickens. The nonlethal nature of mutations that affect the skeleton results in a multitude of human syndromes as well as mouse knockout models that include skeletal dysmorphogenesis, which is vital to our understanding of skeletogenesis. The development of the Cre-lox conditional knockout system in mice has overcome lethal null and dominant-negative mutations in genes that affect multiple organ systems to reveal their function in skeletogenesis (Gu, 1993). The mouse genome can be manipulated to overexpress the proteins that are involved in skeletogenesis in a general or a tissue-specific manner as well. Experimentally, the chicken model is of great use, because it is easily manipulated in ovo, particularly during very early stages of limb bud outgrowth. Finally, zebrafish is a convenient model, because morpholino antisense technology conveniently allows scientists to alter gene expression (Nasevicius, 2000).

I. THE APPENDICULAR SKELETON A. Origins of the Appendicular Skeleton The formation of the lateral plate mesoderm and the initial outgrowth of the limb bud have been covered in Chapter 38, and a brief summary is provided

866

Principles of Developmental Genetics

THE APPENDICULAR SKELETON

867

here (Capdevila, 2001). At a specified location along the axis of the body, cells from the lateral plate mesoderm migrate to the ectodermal surface of the embryo and begin proliferating, thereby forming the limb bud. The zone of polarizing activity and the apical ectodermal ridge (AER) determine the anterior–posterior and proximal–distal polarities of the developing skeletal elements, and they also influence cell proliferation and migration. As the limb bud grows larger, the mesenchymal cells most distant from the AER stop proliferating and condense, thus beginning the process of endochondral ossification that ultimately results in an adult bone. Endochondral ossification can be summed up as follows: mesenchymal cells condense, undergo chondrogenesis, hypertrophy, calcify, and apoptose. Blood vessels invade the space that is left behind by chondrocyte apoptosis, and osteoblasts and osteoclasts take up residence. Additionally, joints are created as single cartilage anlage segments, and they form cartilage where the two new skeletal elements articulate. B. Chondrogenesis 1. Condensation of Limb Bud Mesenchyme As the size of the limb bud increases, the AER will eventually be positioned such that the proximal mesenchymal cells will no longer be under the influence of the fibroblast growth factor (FGF) secreted by the AER. This change in FGF signaling, together with the Wnt and bone morphogenetic protein (BMP) gradients established during limb patterning, will signal the mesenchymal cells of the limb bud to begin the process of condensation. Mesenchymal cells are surrounded by an extracellular matrix (ECM) that is rich in hyaluronan, collagen type I, and an alternatively spliced form of collagen type II (IIA; Dessau, 1980; Maleski, 1996). Before condensation, cells begin to express hyaluronidase, an enzyme that digests the hyaluronan in the ECM, thereby denuding the mesenchymal cells and allowing them to communicate with each other via their cell surface proteins. At the same time, BMP-2, which is a member of the transforming growth factor-beta (TGFb) superfamily that is present throughout the mesenchyme, turns on the expression of neural cadherin (N-cadherin; Oberlander and Tuan, 1994). N-cadherin on one cell binds with N-cadherin on another cell, thereby initiating a signaling cascade that is one of several that initiates condensation (Delise et al., 2002a; 2002b). A common in vitro system used to study the condensation event is the high-density micromass culture of limb bud mesenchyme. Limb buds from day 4 chick embryos or day 11.5 mouse embryos are removed, digested, and the cells plated as high-density cell cultures. Over the course of a few days, these undifferentiated mesenchymal cells will condense and start differentiating into cartilage (Ahrens et al., 1977). The condensation event can be monitored with peanut agglutinin, which is a lectin that recognizes a galactosyl glycoprotein moiety on the surface of prechondrogenic cells (Aulthouse et al., 1987; Stringa and Tuan, 1996). Micromass cultures can be perturbed with neutralizing antibodies or infected with viruses carrying wild-type or mutated versions of proteins involved in early limb bud processes, thus elucidating the roles of these proteins. For example, treatment with anti–N-cadherin antibodies (Oberlander and Tuan, 1994) or infection with mutated N-cadherin prevents condensation, which ultimately inhibits chondrogenesis, thus demonstrating the importance of the condensation event (Delise et al., 2002a; 2002b). The importance of BMPs during early condensation has also been studied in

868

SKELETAL DEVELOPMENT

micromass cultures, but the more important studies have involved the retroviral infection of chick embryos in ovo or in mouse transgenic models. For example, the injection of retrovirus expressing the BMP inhibitor Noggin prevents condensation, thus demonstrating the necessity of BMP signaling during early skeletogenesis (Pizette et al., 2000). Similarly, mice overexpressing Noggin in their limbs fail to produce most of their skeletal elements (Tsumaki et al., 2002). Other cell–cell interaction proteins, such as neural cell adhesion molecule (N-CAM), and multiple ECM proteins and their cell surface receptors are implicated in the condensation process. TGF-b initiates the expression of fibronectin (FN), which is an extracellular protein that is vital to condensation. There are multiple splice variants of FN, but two are particularly interesting with regard to chondrogenesis (White, 2003). Before condensation, FN expresses exon IIIA. Cells spread less in the presence of the IIIA exon, which allows the cells to round up; this process allows better packing, which positively influences condensation. After condensation is complete, exon IIIA is spliced out. Versican is another ECM protein expressed in mesenchyme that, like FN IIIA, prevents cell spreading (Williams et al., 2005). In addition to fibronectin and versican, the thrombospondins (including cartilage oligomeric matrix protein [Kipnes et al., 2003]), the tenascins, and many other ECM proteins are necessary for proper condensation. At this time of development, another member of the TGF-b superfamily, growth/differentiation factor 5 (GDF-5), is expressed (Chang et al., 1994, Storm et al., 1994). In micromass models, GDF-5 regulates condensation by increasing the ability of cells to communicate via gap junctions. Connexin 43 is expressed in developing limbs in an overlapping pattern with GDF-5, which suggests that connexin 43 may be the dominant gap-junction protein (Coleman et al., 2003a; 2003b). The overexpression of GDF-5 by infecting the chick limb in ovo with GDF-5 constructs leads to larger cartilage elements with increased cell numbers. It is proposed that the increased number of cells is the result of increased cell adhesiveness as opposed to an increase in proliferation (Francis-West et al., 1999). In mice that are null for GDF-5, chondrocyte condensation size is significantly reduced. Because the size of the initial condensations regulates the size of the future skeletal elements, these mice are severely dwarfed. Human mutations in the GDF-5 gene lead to a spectrum of dwarfisms, the most severe of which leads to a near absence of fingers. The Sry-related HMG box-containing transcription factor Sox9 is critical for every step of chondrogenesis, and it is widely used as the marker to detect chondrocytes. Using the Cre-lox system, Sox9 can be eliminated specifically in early mouse limbs before condensation (Akiyama et al., 2002). In these animals, no condensations form in the limb bud, and ultimately these animals are born with a complete absence of limbs. In micromass cultures created using Sox9–/– chimeric limb buds, SOX9–/– cells cannot be found in condensing regions. In SOX9–/––wild-type chimeric animals, SOX9–/– cells are excluded from the cartilage primordium (Bi et al., 1999). The introduction of a bead soaked in BMP-2 into the developing chick limb leads to an upregulation of Sox9 expression, which suggests that the BMPs are important for the initiation of chondrogenesis (Healy et al., 1999). Although the exact mechanism of the Sox9 induction of condensation is unclear, its role in the initiation of overt chondrogenesis is well understood, and it will be covered in detail in the next section of this chapter.

THE APPENDICULAR SKELETON

869

The size of the condensation is critical during skeletogenesis. A condensation that is too small may fail to undergo chondrogenesis; if it is too large, the final bone will also be too large. The early cartilaginous skeleton determines the size and shape of the future bones, and thus its development must be very well controlled. Growth factors and the proteins responsible for cell–cell and cell–matrix interactions are critical for determining the size of the condensations by altering cell adhesion and migration. However, the cells in the condensations also proliferate, and the proteins that guide proliferation are thus noteworthy. Homeobox genes have been implicated in the control of the proliferation of the early mesenchyme (Boulet et al., 2003, Goff et al., 1997). HoxA11/HoxD11 double mutant mice have shortened forelimbs as a result of decreased FGF expression in the AER (Boulet et al., 2003). Because FGFs in the AER direct outward growth, any changes in FGF expression will alter limb bud growth. Interestingly, Hox genes are activated by BMPs, which further demonstrates the importance of FGFs in skeletogenesis (Duprez et al., 1996). At the stage of mesenchymal condensation, the growing limb bud contains mesenchymal cells that have begun to adhere to one another and to communicate, thereby readying themselves for differentiation into chondrocytes. Before that happens, the edges of the condensation must be defined (Hall et al., 2000). Multiple cell–matrix interactions are thought to be involved in this process, but we will focus on just one interaction: the one that occurs between syndecan and FN (Figure 39.1). Syndecan is an integral membrane proteoglycan that is found in the mesenchymal cells that surround the initial condensation. Syndecan binds to FN, which is found in the ECM that surrounds cells during the early condensation. The interaction of syndecan and FN leads to an intracellular signaling cascade that instructs the mesenchymal cell to downregulate N-CAM. The point at which cells expressing N-CAM meet cells that are not expressing it becomes the boundary between the future cartilage anlage and the surrounding mesenchyme. Along this boundary, mesenchymal cells will flatten and become the perichondrium, which is the thin layer of cells that surrounds the cartilage skeletal elements that will later become the periosteum.

FIGURE 39.1

Establishment of the mesenchymal condensation boundary in the developing limb bud. A, Syndecan on the mesenchymal cells binds to fibronectin in the extracellular matrix of cells during early condensation. B, The interaction of syndecan and fibronectin leads to a downregulation of cell adhesion molecules such as neural cell adhesion molecules. C, Cells that no longer express neural cell adhesion molecules flatten and become the perichondrium, which is the thin layer of cells that surrounds cartilage.

870

SKELETAL DEVELOPMENT

2. Cartilage Differentiation As stated previously, the primary gene responsible for driving the conversion of mesenchymal cells to chondrocytes is Sox9, and it does this by turning on two other Sox genes: Sox5 and Sox6 (Akiyama et al., 2002). In animals lacking Sox9, Sox5 and Sox6 are never expressed; Sox5 and Sox6 are known to have Sox9 binding sites in their enhancers, thereby explaining this finding. Sox5 and Sox6 are redundant, and thus mutations in either gene lead to no overt phenotype. However, animals lacking both Sox5 and Sox6 have severely reduced long bones (Smits et al., 2001). Sox9 also binds the enhancer regions of collagen types II and XI, and it turns on their expression early during chondrogenesis (Bi et al., 1999). Later, together with Sox5 and Sox6, Sox9 increases the expression of collagen type IX, aggrecan, and link protein (Lefebvre et al., 1998). The expression of these ECM proteins is an indicator of mature cartilage. Collagen type II is the dominant fibril form of collagen in cartilage, and it is crosslinked by collagen type IX. Mutations in either collagen type II or IX lead to disorganized growth plates and early onset osteoarthritis (Cremer et al., 1998). Aggrecan is a sulfated proteoglycan that is important in the maintenance of the high water content of cartilage (Dudhia et al., 2005). These structural components are absolutely vital to the maintenance of the mechanical properties of cartilage. This is particularly important in the articular cartilage at the ends of bones in joints. In addition, growth factors can be sequestered in the ECM, and cell surface receptors have important interactions with ECM proteins. 3. Growth Plate Regulation and Long Bone Development The growth plate is an orderly arrangement of chondrocytes as they proliferate, differentiate, and apoptose, and they leave behind a mineralized matrix for osteoblasts to invade (Figure 39.2, A). A well-functioning growth plate is necessary for the lengthening of the long bones, and any perturbations

Regulation of cartilage growth plate maturation. A, The growth plate is an orderly arrangement of chondrocytes as they proliferate, differentiate, and apoptose, leaving behind a mineralized matrix for osteoblasts to invade. B, The Indian hedgehog/parathyroidhormone related protein negative feedback loop and bone morphogenetic protein/fibroblast growth factor antagonism: Indian hedgehog and bone morphogenetic proteins increase proliferation and prevent hypertrophy, whereas fibroblast growth factors prevent proliferation and promote hypertrophy. (Adapted from Goldring et al., 2005, and Kornak et al., 2003. See color insert.)

FIGURE 39.2

THE APPENDICULAR SKELETON

871

of the system will lead to gross deformities. The growth plate is separated into resting, proliferating, prehypertrophic, and hypertrophic cells. Each bone initially has two growth plates, with their resting chondrocytes toward the epiphyses and their hypertrophic zones meeting in the diaphysis. As the bone becomes larger, secondary growth plates called secondary ossification centers form in the ends of the long bones. Growth plates are present before birth, and they stay active until the end of puberty, when growth is complete. Sox5 and Sox6 are thought to initiate the formation of the growth plate by inducing proliferation between the epiphysis and diaphysis in the areas in which the columns of proliferating cells will appear (Smits et al., 2004). In animals that are null for Sox5 and Sox6, there are no columnar cells present, and growth plates fail to form. After the proliferating columnar cells are established, the cells at the medial edge must mature into prehypertrophy. The rest of this section will illustrate the carefully balanced act of proliferation versus hypertrophy. Sox5 and Sox6 prevent prehypertrophy at least in part by inhibiting the expression of FGF receptor 3 (FGFR-3). The FGFs prevent proliferation, and, in many tissues, cells either proliferate or differentiate. Thus, to prevent proliferation is to promote differentiation. How the medial edge differentiating cells escape the Sox5 and Sox6 proliferation cues is unclear, but it is likely that there are soluble signals that mediate this conversion. Upon the initiation of hypertrophy, the cells are committed and will ultimately undergo apoptosis. After the first prehypertrophic cells are established, they begin secreting signals that will interact with proliferating cells, thus setting up the cross talk that balances proliferation and hypertrophy. The Indian hedgehog (Ihh)/parathyroid hormone related protein (PTHrP) negative feedback loop and BMP/FGF antagonism will be covered in more detail to illustrate this point (see Figure 39.2, B). Ihh is a secreted morphogen that is important in both limb bud patterning and growth plate regulation. Ihh null animals show decreased proliferation of the growth plate chondrocytes as well as an increase in hypertrophy. The misexpression of Ihh leads to a decrease in hypertrophy (St-Jacques et al., 1999). PTHrP is a secreted protein that shares homology and a receptor with parathyroid hormone. Like the Ihh knockout, the PTHrP null animal exhibits increased hypertrophy (Amizuka et al., 1994). The misexpression of PTHrP gives rise to a completely cartilaginous skeleton, which indicates a complete absence of hypertrophy. The expression patterns of the proteins involved in the Ihh and PTHrP signaling pathways have been established. Postmitotic, prehypertrophic cells express Ihh. The Ihh receptor Patched is found in the perichondrium surrounding the growth plate as well as in the proliferating zone. The Ihh activation of Patched in the perichondrium turns on the expression of TGF-b, which in turn activates PTHrP expression, also in the perichondrium (Alvarez et al., 2001; 2002). The PTHrP receptor is found only in proliferating cells on the edge of prehypertrophy. PTHrP acts on the proliferating cells to prevent them from entering into hypertrophy. In addition, Ihh can act directly on the cells of the proliferative zone to stimulate proliferation. Taken together, Ihh, in concert with PTHrP, acts in a negative feedback loop: cells committed to hypertrophy secrete a protein that prevents hypertrophy in proliferating cells. The story becomes more complicated with the addition of BMP/FGF antagonism and its control over the Ihh/PTHrP feedback loop (Minina et al., 2001; 2002). FGF-2, -8, -9, -17, and -18 are all found in cartilage,

872

SKELETAL DEVELOPMENT

although their actions are dependent on which receptors they bind and where those receptors are located. FGFR-3 is found in proliferating and prehypertrophic cells, whereas FGFR-1 is found only in hypertrophic cells. Of the FGF knockout animals, only the FGF-18 null animal has an abnormal skeletal phenotype, and it closely resembles the FGFR-3 knockout. Both have expanded proliferative and hypertrophic zones, although the FGF-18 null animal also has delayed bone formation, which indicates that it may also signal through FGFR-1. FGFs have been shown to decrease proliferation by direct action as well as by decreasing Ihh expression. Decreased Ihh expression increases differentiation. FGFs not only increase commitment to hypertrophy, they also accelerate hypertrophy by directly acting on hypertrophic cells. Presumably the effects on Ihh expression and on proliferation are mediated by FGFR-3, although the latter effect is also mediated by FGFR-1. The BMPs are present throughout the cartilage and the perichondrium. They are known to increase proliferation directly and by the upregulation of Ihh. In addition, they inhibit hypertrophic differentiation independently of the Ihh/PTHrP signaling cascade. Although they upregulate Ihh, Ihh in turn upregulates them. Quite obviously, the effects of FGF and BMP are directly opposite one another, and they operate by regulating the expression of Ihh as well as by independent means. In limb culture systems, adding FGF-2 and BMP-2 together induces no change to the bones, whereas, independently, FGF-2 increases hypertrophy, and BMP-2 increases proliferation. Interestingly, the addition of BMP-2 to limb cultures expressing constitutively active FGFR-3 rescues the phenotype (Minina et al., 2001; 2002). This is particularly important because FGFR-3–activating mutations are responsible for achondroplasia, which is the most common form of human dwarfism (L’Hote et al., 2005). Another protein that is critical to the differentiation of chondrocytes to hypertrophy is core-binding factor 1, which is also known as runt-related transcription factor 2 (RUNX; Otto et al., 1997; Fugita et al., 2004). RUNX2 is found in prehypertrophic and hypertrophic chondrocytes as well as in osteoblasts. The primary phenotype of mice that are null for RUNX2 is a failure to form bone as a result of an absence of osteoblasts. However, infecting chick or mouse chondrocytes with antisense constructs for RUNX2 prevents chondrocyte hypertrophy, thus confirming its role in endochondral ossification. In addition, the infection of immature chondrocytes with RUNX2 accelerates the rate of hypertrophy. This acceleration of hypertrophy is the result in part of the upregulation of collagen type X and metal metalloproteinase 13 (MMP-13), which will be discussed later (Enomoto et al., 2000; Zheng et al., 2003). In addition, RUNX2 upregulates vascular endothelial growth factor (VEGF), which is necessary for the vascular invasion of the hypertrophic zone. Growth hormone (GH) and insulin-like growth factors (IGFs) are important for controlling the length of long bones, although their targets, molecular effects, and expression patterns are as yet unclear (Robson et al., 2002). GH is responsible for gigantism and dwarfism, because of over- or underexpression, respectively. It is known that GH is released from the pituitary gland and that it stimulates the liver to release circulating IGF-1. Eliminating liver-derived circulating IGF-1 stunts growth, which suggests that the liver is the source of chondrocyte-activating IGF. Alternatively, directly injecting GH into the tibia increases length both by directly influencing resting chondrocytes to proliferate and by increasing the local expression of IGF-1. Interestingly, the

THE APPENDICULAR SKELETON

873

growth plates of IGF-1 null mice have no alteration in cell proliferation, but they show a decrease in hypertrophic cell height. This finding suggests that perhaps IGF-2 is important in the stimulation of proliferation. Complicating matters further is the fact that multiple IGF binding proteins, which inhibit the activities of the IGFs, as well as the two IGF receptors are found in overlapping patterns throughout the growth plate. Although it is clear that the GH/IGF system is important in long bone length, more work needs to be done to clarify the process. The term hypertrophy has been mentioned multiple times, but it has not been defined. Hypertrophic cells are easily identified histologically, because they are the largest cells of the growth plate as a result of an increase in intracellular organelles. In fact, the increased height of hypertrophic cells is responsible for more than 50% of bone lengthening. In molecular terms, gene expression of collagen type X and alkaline phosphatase defines hypertrophy. Both of these proteins are important in mineralization. Collagen type X is thought to influence mineralization by stimulating calcium accumulation in matrix vesicles and by acting as a docking site for the matrix vesicles (Kwan et al., 1997). Alkaline phosphatase is found in matrix vesicles, and it leads to an accumulation of pyrophosphate, which is a mineral that is important in the formation of calcium hydroxyapatite crystals. Together, collagen type X and alkaline phosphatase act to initiate the mineralization of cartilage, which prevents any further growth in the hypertrophic zone of cartilage and sets up a scaffold for osteoblast invasion. After cells undergo hypertrophy, their matrix is remodeled, they undergo apoptosis, and the cartilage model is invaded by capillaries, although not necessarily in that order. It is believed that matrix remodeling by MMPs is the initiator of apoptosis and vascular invasion. Mouse knockout models for both MMP-9 and MMP-13 show a reduction in angiogenesis, ECM remodeling, and apoptosis (Stickens et al., 2004; Vu et al., 1998). MMP-13, which is turned on by cbfa1, cleaves collagen type II, which is still highly expressed in the hypertrophic zone. After the degradation of collagen type II, VEGF previously expressed under the influence of RUNX2 and sequestered in the ECM is released. VEGF acts as an activator of vasculogenesis as well as a recruiter of osteoclasts (Zelzer et al., 2002). Osteoclasts will be covered in detail later, but for now it is important to know that their primary function is ECM degradation. Osteoclasts express MMP-9, which further processes the collagen type II initially cleaved by MMP-13 and which, together with MMP-13, acts to degrade aggrecan. There are two requirements for capillary invasion: proangiogenic growth factors and space. The actions of MMPs provide both. In addition, MMPs are responsible for promoting apoptosis. Either directly or indirectly, MMP action either releases a proapoptotic factor from the ECM or the ECM protein fragments stimulate the cells to apoptose. C. Osteogenesis 1. Differentiation After a hypertrophic zone is established in the growth plate, some cells in the perichondrium surrounding the newly hypertrophic cells begin to differentiate into osteoblasts (bone-forming cells), and the bone collar is formed. There are at least three genes that are known to be important in this conversion: Ihh, RUNX2, and osterix. In addition, the BMP signaling pathway is

874

SKELETAL DEVELOPMENT

critical to bone formation. Ihh is secreted by prehypertrophic cells, but the Ihh receptor Patched is expressed throughout the perichondrium. In Ihh null mice, there are no osteoblasts, which highlights how vital this protein is in bone production (St-Jacques et al., 1999). Ihh affects differentiation by working upstream of the transcription factor RUNX2 as well as by increasing expression of BMPs. Animals that are null for most BMPs are early embryonic lethal, although BMP-6 and BMP-7 knockouts have mild skeletal phenotypes (Jena et al., 1997; Solloway et al., 1998). The ectopic expression of the BMP inhibitor Noggin or dominant-negative BMP receptors leads to decreased osteoblast function and differentiation. The ectopic expression of BMP-2 and BMP-4 leads to heterotopic ossification. In fact, fibrodysplasia ossificans progressiva, which is a human disease caused by the ectopic activation of the BMP pathway via mutations in an activin receptor (ACVR-1), leads to horrible disfigurations as ossification occurs throughout the soft tissues of patients’ bodies (Shore et al., 2006). BMPs are also capable of upregulating RUNX2, possibly acting as the effector of the Ihh upregulation of RUNX2. RUNX2 is the master switch that differentiates perichondrial precursor cells into preosteoblasts (Otto et al., 1997; Fugita et al., 2004). Preosteoblasts do not form bone, but they do express osteoblast specific proteins. They also express chondrocyte-specific genes, which suggests that they are not yet irreversibly committed to the osteoblast lineage. RUNX2 acts upstream of osterix, which is another transcription factor that is involved in osteoblast differentiation. Although RUNX2 null mice do not express osterix, the effect is not thought to be direct. Alternatively, BMP-2 is known to directly activate osterix expression. Osterix is responsible for converting preosteoblasts into bone-forming osteoblasts (Nakashima et al., 2002). Mice lacking osterix have completely normal growth plates that vascularize appropriately. However, they fail to form bone. Finally, communications between the perichondrial cells and the invading capillaries are required for the differentiation of osteoblasts. In one study, investigators placed cartilaginous long bones, complete with perichondrium, into the kidney capsules of host animals (Colnot et al., 2004). When a filter was placed around the anlage such that the vasculature of the kidney could not invade, maturation proceeded normally until osteoblasts were expected to differentiate and form bone. Although endothelial cells exist in the perichondrium, they can only form vessels and invade the hypertrophic zone upon contact with preexisting blood vessels. In other words, the cells of the perichondrium are not sufficient for osteogenesis. Communications between preexisting capillaries and the perichondrium are essential for osteoblast differentiation. Ultimately, VEGF is responsible for this finding, because it directs vasculogenesis. 2. Osteoclasts and Bone Remodeling Osteoclasts were previously mentioned briefly, and they were described as being responsible for ECM digestion. They are important early in bone development, when the hypertrophic zone is remodeled for capillary and osteoblast invasion. For the rest of an organism’s life, osteoclasts will work hand in hand with osteoblasts to digest and rebuild the skeleton. In fact, every year, 10% of the skeleton is digested and remodeled, thus resulting in a new skeleton every 10 years. This skeletal digestion must be carefully controlled, because too

THE APPENDICULAR SKELETON

875

much can lead to osteopenia, osteoporosis, or Paget’s disease, and too little can lead to osteopetrosis. Bone remodeling is not just meant to repair and refresh the skeletal structure; it also regulates serum calcium and phosphate homeostasis, because bone is the body’s storage bin for both elements. This additional level of complexity relies on communication with the gut, kidney, thyroid, and parathyroid (Boyle, 2003). Osteoclasts derive from circulating monocytes that commit to the macrophage lineage. A small percentage of these circulating cells find their way to the bone and adhere. Interaction with bone stromal cells (preosteoblasts) is necessary for them to differentiate into osteoclasts. Two cell surface proteins produced by the osteoblasts, colony stimulating factor 1 (CSF-1) and RANK ligand (RANKL), and an unidentified serum factor are required for osteoclastogenesis. CSF-1 alone promotes proliferation, whereas CSF-1 and RANKL together lead to differentiation, survival, and fusion. Osteoclasts are multinucleated, and fusion is a prerequisite for function. Continued exposure to RANKL leads to osteoclast activation. Mice deficient in RANKL or RANK (Dougall et al., 1999), which is the cell surface protein on osteoclasts that binds RANKL, have a complete absence of osteoclasts, whereas other monocyte-derived cells are normal (Boyle et al., 2003). Activated osteoclasts express a host of proteins that are involved in the degradation of bone and mineralized cartilage. To degrade bone, the osteoclast adheres to the bone surface, leaving a space into which it will secrete an acidic cocktail of digestive juices. To create this tight seal, the cell must express avb3 integrin, which attaches to the collagen in bone. A vacuolar ATPase then pumps protons into the space between the osteoclast and bone, decreasing the pH to around 4.5 and allowing for the dissolution of the inorganic crystals in bone. Tartrate-resistant acid phosphatase and cathepsin K are also secreted into the space, and they are responsible for digesting the organic components of bone (Boyle et al., 2003). Although RANK and RANKL are essential for osteoclast differentiation and activation, other signaling molecules influence osteoclastogenesis. Osteoprotegerin is a decoy receptor for RANKL, and it functionally blocks the interaction of RANK and RANKL. Osteoblasts can be stimulated to produce osteoprotegerin by estrogen and BMPs. Estrogen also downregulates the expression of RANKL in osteoblasts as well as the expression of the proinflammatory cytokines interleukin 1, interleukin 6, and tumor necrosis factor a, all of which positively influence osteoclast activity. Calcitonin, which is a protein that is secreted by the thyroid when the serum calcium level is elevated, prevents osteoclasts from attaching to bone. Alternatively, when the serum calcium level is low, the parathyroid gland releases parathyroid hormone, which stimulates osteoclast activity. Many additional factors, including vitamin D, corticosteroids, and PTHrP, positively regulate osteoclasts as well (Boyle et al., 2003) Although fewer signaling pathways are known to negatively regulate osteoclast activity, one prominent example is estrogen. The impact of estrogen is particularly prominent in postmenopausal women, who frequently suffer from osteoporosis; this condition is characterized by a loss of bone mass, and it is caused by increased osteoclast activity without compensatory bone growth. Although osteoporosis is much more common in women, men are also susceptible. As men age, there is a reduction of testosterone. Testosterone is the precursor of estrogen, and it directly promotes osteoblast activity; thus,

876

SKELETAL DEVELOPMENT

both an osteoanabolic factor and an inhibitor of osteocatabolism are lost. Drug companies have developed pharmacologic agents that target osteoclast activity at multiple steps. Bisphosphonate, which stimulates osteoclast death, and estrogen treatments are most commonly prescribed, but drugs targeting osteoclast adhesion and digestion are also in use or in development (Boyle et al., 2003). D. Joint Development 1. Interzone Development The initiation of the synovial joint formation is still a mystery in developmental biology. The morphologic changes are well known, but the molecular controls are still being elucidated. It is known that, at specific points along the length of the cartilage anlagen, cells begin to flatten and form what is known as the interzone (Figure 39.3, A). These cells stop producing chondrocytespecific proteins, particularly collagen type II and aggrecan, and they start producing collagen type I and hyaluronan. The cells at the center of the interzone begin to secrete an ECM that separates the cells, and this ultimately leads to cavitation. The remaining flanking interzone cells are thought to become articular chondrocytes. A mature joint consists of two opposing articular surfaces wrapped in a joint capsule and stabilized by ligament and tendon. Synovial cells line the inside of the joint, and they are responsible for producing joint fluid, which nourishes the avascular articular cartilage and lubricates the structure. Any perturbations in the joint structure, particularly as a function of age, lead to painful and often irreversible joint diseases, such as osteoarthritis and rheumatoid arthritis (Archer et al., 2003). The location of the interzone is established quite early. If the presumptive elbow region is removed before any morphologic changes occur, then the joint fails to form. The specification of the location of the joint interzone is similar to the specification of the location of limb bud outgrowth in the early embryo. Not surprisingly, mutations in homeobox genes, which are known to be important during embryonic patterning, lead to fusions in the joints of the wrists, thus implicating them in this specification (Archer et al., 2003). Two novel homeobox proteins, Cux1 (Lizarraga et al., 2002) and Barx1 (Meech

Joint formation. A, Morphologic description: cells of the interzone begin to flatten and stop producing chondrocyte-specific proteins (Archer, 2003). B, Molecular regulation: WNT14 appears before the interzone is established, and it is required for interzone formation. Proteins expressed by the interzone, such as growth/differentiation factor 5 and chordin, prevent new joint initiation within a certain distance of an established joint. (Adapted from Archer et al., 2003, and Hartmann et al., 2001. See color insert.)

FIGURE 39.3

THE APPENDICULAR SKELETON

877

et al., 2005), have recently been found in early joints. Interestingly, the exogenous expression of Cux1 in the limb bud micromass system leads to the downregulation of chondrogenic markers, which is one of the first steps in interzone formation. Concomitant with the establishment of the interzone, several growth factors and inhibitors appear. GDF-5, GDF-6, BMP-2, and chordin are all expressed in the interzone, whereas Noggin is specifically excluded. Upon cavitation, expression shifts to BMP-2, -4, -7, and chordin. The WNT proteins are candidate master genes that regulate gene expression in the interzone. WNT-14 is of particular interest (see Figure 39.3, B; Guo et al., 2004; Hartmann and Tabin, 2001). Although WNT-4, -5a, -14, and -16 are all found in joints, only WNT-14 is confirmed to have a direct effect on the presumptive joint. How WNT-14 signals is controversial, but its effects are agreed upon: it upregulates the expression of growth and differentiation factor 5 (GDF-5), chordin, and CD-44, and it downregulates Sox9 and Noggin. If exogenous WNT-14 is retrovirally expressed by virus, the region adjacent to the infection fails to form a joint where a joint would normally be. This implies that WNT-14–stimulated proteins act to prevent new joint initiation within a certain distance of an established joint. In other words, WNT signaling may be responsible for joint specification. However, at least one “initiator” joint would have to be established to begin this cascade of events. The roles of the secondary signaling molecules are still vague, but some hypotheses for their functions have been suggested. The downregulation of Sox9 is responsible for the phenotypic switch of cells from chondrocytes to interzone cells. BMPs are thought to increase the expression of hyaluronan (HA), which is an extracellular glucosaminoglycan. HA binds to CD-44, which is a receptor on the surface of interzone cells. HA can act to condense cells or separate them, depending on the quantity of HA that is produced. Interzone cells produce large quantities of HA, and this leads to cell separation, which sets the joint up for cavitation. BMPs are also thought to cause apoptosis in the interzone, thus further demonstrating their role in cavitation. HA and CD-44 upregulation is also dependent on movement. The mechanical stimulation of cells increases CD-44 and HA expression, and paralyzed animals form interzones, but they fail to cavitate (Archer et al., 2003). Interestingly, GDF-5 decreases HA production. The overexpression of GDF-5 leads to joint fusion, which is thought to be the result of greatly decreased HA production and the failure to cavitate. How GDF-5 positively regulates joint formation is unclear, although it may be that it is important in a step between interzone formation and cavitation. GDF-6 is closely related to GDF-5, and, like GDF-5, it is expressed in a specific subset of joints; however, its effects are unknown. Chordin and Noggin are known to bind to and inhibit BMP2, -4, and -7 and possibly GDF-5. Curiously, chordin null mice have no joint phenotype, whereas Noggin-deficient mice fail to form joints (Brunet et al., 1998). This failure is thought to be the result of either a failure of GDF-5 to be produced in joints or of an increase in BMP bioavailability, which could drive the recruitment of cells into the cartilage model. That the interzone expresses agonists and their antagonists in the same spatiotemporal pattern is curious. Because cells rarely produce substances in a wasteful manner, there are likely additional controls on BMP/GDF or Noggin/chordin action. For example, perhaps an additional protein that cleaves the inhibitors exists. The diffusion rate of the proteins may also be quite different, such that

878

SKELETAL DEVELOPMENT

BMPs may be able to signal quite distantly while the inhibitors are confined locally (or vice versa). Finally, how the proteins bind to the ECM may differ, with some being sequestered while others are left free to signal.

II. AXIAL SKELETON A. Origins of the Vertebrate Skeleton The vertebrate skeleton is the product of cells from three distinct embryonic lineages. The craniofacial skeleton is derived from the cephalic paraxial mesoderm and invading cranial neural crest cells; the axial skeleton is derived from paraxial mesoderm in the vicinity of the notochord and neural tube; and the appendicular skeleton is the product of lateral plate mesodermal cells. The primordia of these tissues are specified early during ontogeny. In the chick embryo, three germ layers are formed 12 hours after fertilization with the onset of gastrulation. In the Hamburger Hamilton stage 5 head process embryo (20 hours; Figure 39.4, A; Hamburger and Hamilton, 1951), several different mesodermal compartments can already be identified: the axial chordamesoderm that will give rise to the notochord and the paraxial, intermediate, and lateral plate mesoderms. Within the paraxial mesoderm, two major regions are recognized: the segmental plate, which forms along the length of the primitive streak on either side of the notochord, and cephalic mesoderm on either side of the chordamesoderm or head process (see Figure 39.4, A). The paraxial mesoderm of the trunk and the cephalic mesoderm, in combination with invading cranial neural crest, give rise to many tissues, including bone, muscle, tendon, and dermis. Although the axial skeleton is completely derived from the cells of the paraxial mesoderm, cranial neural crest cells contribute heavily to the craniofacial skeleton. These cells are formed during the specification of the cephalic neural plate from ectoderm, and they begin deepithelializing from the dorsal lip of the cephalic neural folds as they fuse to form the neural tube (see Chapter 26). B. Somitogenesis: Axial Skeletal Patterning Somites are transient embryonic structures that are critical in the realization of the metameric pattern characteristic of the vertebral column and its associated tissues. They are segmental units that bud from the segmental plate as the node passes rostrocaudally down the primitive streak (see Figure 39.4, B). In the node’s wake, the notochord forms from the chordamesoderm, and the neural plate differentiates from the ectoderm, thickening, folding, and fusing to form the neural tube. After these events, somites form from the rostral end of the paraxial mesoderm through a mesenchymal–epithelial transition in a species-specific, time-dependent fashion. In the chick, a new somite is formed every 90 minutes. Because the development of the embryo proceeds rostrocaudally, several stages of somitogenesis can be observed in one embryo (see Figure 39.4, C). The mesenchymal–epithelial transition involves an initially loosely associated ball of cells (the condensed somite) that reorganizes into a spherical epithelial structure (the epithelial somite) and that ultimately differentiates (see Figure 39.4, D). There is a small population of cells inside the presomitic mesoderm that retain their mesenchymal organization to form the cells of the somitic core, which is called the somitocoele.

AXIAL SKELETON

879

A brief summary of somitogenesis in the chick embryo. A, A head-process stage embryo after 20 hours of development. Two regions of paraxial mesoderm are identified: the segmental plate (SP) lateral to the primitive streak (PS) and the cephalic mesoderm (CM) lateral to the head process (HP), which is a ridge that is formed by the underlying chordamesoderm (ChM) that also runs along the length of the embryo under the primitive streak. The chick node called Hensen’s node (HN) is currently located at the rostral end of the primitive streak, and it defines the border of head and trunk structures. The thickening of the rostral segmental plate emanating posterolaterally from Hensen’s node indicates the cellular condensation characteristic of somitic mesoderm (SM) that here will give rise to the first somite. B, A stage-6, 4-somite embryo in which morphogenesis of the neural folds (NF) has narrowed and elongated the embryo. The arrow indicates the rostrocaudal migration of Hensen’s node (HN) down the primitive streak (PS) from its rostral extreme (*). In the node’s wake, neural folds have formed, and the segmental plate (SP) has lengthened, whereas somites (So) have formed from its rostral end at regular intervals. The cephalic mesoderm (CPM) has thickened with the invasion of neural crest cells, but it remains overtly unsegmented. C, A stage-13 embryo with defined brain vesicles, optic and otic cups, a folded heart tube (HT), and approximately 20 somites. Hensen’s node (HN) has almost reached the caudal end of the primitive streak (PS). Because the development of the embryo proceeds rostrocaudally, somites at different levels of development can be visualized: condensed somites (CSo, 1), epithelial somites (ESo, 2), and differentiated somites (DSo, 3). D, Transverse sections through the stage-13 embryo shown in C at different rostrocaudal levels as indicated in C and revealing D1, the condensed somite (CSo); D2, the epithelial somite (ESo) with specified dorsolateral dermomyotome (DM) and ventromedial sclerotome (SC) induced by the notochord (NC); and D3, the differentiated somite (DSo) in which the sclerotome has de-epithelialized, thus leaving the epithelial dermomyotome (DM) behind in the context of surrounding inductive tissues: notochord (NC), neural tube (NT), ectoderm (Ec), endoderm (En), intermediate mesoderm (IM), lateral plate mesoderm (LM), and the dorsal aortae (DA). (Adapted from Alexander PG: The role of paraxis in somitogenesis and carbon monoxide-induced axial skeletal teratogenesis, PhD thesis, Thomas Jefferson University, Philadelphia, Penn, 2001.)

FIGURE 39.4

1. Somite Patterning The process of somitogenesis is divided into several distinct phases: patterning, morphogenesis, differentiation, and the maturation of somite-derived tissues. Although the morphogenesis of the somite was described early in the history of embryology, the exact mechanism of the patterning process remains elusive. The regularity of somite architecture and formation during embryogenesis suggests that it must be controlled by a clock or a somitic oscillator within the embryo. The earliest experiments attempted to alter somite formation through surgical manipulations; these included removing individual tissues

880

SKELETAL DEVELOPMENT

from the embryo or observing individual tissues in isolation. Among the most important observations was that the segmental plate could undergo somitogenesis when separated from both the last previously formed somite and the axial structures (Packard, 1976). This finding demonstrated that segmentation does not require a continuous flow of information rostrocaudally and that the positional information or pattern for somitogenesis is established at gastrulation, possibly as the cells pass through the node to populate the paraxial mesoderm. Another enlightening observation was that implanted quail nodes could induce a secondary axis with a somite pattern dependent on the mediolateral position of the implant, thereby suggesting a possible morphogen emanating from Hensen’s node (Horbruch et al., 1979). With the application of scanning electron microscopy, investigators identified segmentation of the entire paraxial mesoderm in a number and regularity that prepatterned the somite (Packard and Meier, 1983). These segments were called somitomeres. Another important observation was that heat shock could induce somite abnormalities at regular intervals, thereby suggesting a timing-mechanism– based regulation of somitogenesis (Primmet et al., 1988). Similar segmentation defects were also observed when cell-cycle disruptors were applied to the embryo, which indicated a link between the cell cycle and the somite oscillator (Primmet et al., 1989). This led to the clock and wavefront model of somitogenesis, which invoked a molecular clock within the cells of the paraxial mesoderm that coordinated events in response to a signal wavefront moving rostrocaudally through it (Cooke and Zeeman, 1976). This quickly became the favored model of somitogenesis, and subsequent genetic studies have proven the central components of this theory to be true. A major breakthrough in understanding the somitic oscillator was made with the cloning of the chick Hairy gene, c-Hairy1 (Palmeirim et al., 1997). Whole-mount in situ hybridization for cHairy1 mRNA revealed a dynamic caudorostral expression pattern in the segmental plate (Figure 39.5). During the 90-minute cycle of chick somitogenesis, c-Hairy1 expression begins within a broad domain in the caudal third of the segmental plate (see Figure 39.5, A), which slowly moves rostrally, becoming restricted to the posterior boundary of the next somite to be formed (see Figure 39.5, C). This was the first molecular evidence of a somitic oscillator within the segmental plate. These experiments established that the cycles of c-Hairy1 expression are autonomous to the paraxial mesoderm, occurring independently of surrounding tissues. These cycles are also independent of protein expression, which indicates that this particular gene responds to the putative somitic oscillator. Although its sequence suggests that it is a transcription factor, the function of c-Hairy1 remains elusive, as does its connection to the somitic oscillator. 2. Somite Border Formation In the formation of the somite that follows, we recognize an initial condensed somite that is composed of unorganized mesodermal cells and its subsequent transformation into a spherical epithelial structure. These two processes are linked; however, different adhesion systems may be involved in each event. The boundaries of the somite are established through the Notch–Delta juxtacrine signaling system. Many components of this system exhibit cyclic expression that is autonomous to the segmental plate in phase with somite formation (Saga and Takeda, 2001; Kuan et al., 2004). The

AXIAL SKELETON

881

Dynamic c-Hairy1 expression in the chick segmental plate. A, The expression of c-Hairy in the caudal half of the segmental plate (SP), in the posterior border of the last formed somite (So), and in a region approximating the posterior half of the next somite at the beginning of the proposed somite oscillator cycle (time 0). Chick somites bud from the rostral segmental plate every 90 minutes; this is the temporal definition of the somite oscillator. B, Within 30 minutes, the caudal expression zone of c-Hairy1 expression is moving rostrally within the segmental plate, whereas the anterior zone is consolidating into a band prefiguring the posterior border of the next somite (somitomere). C, At 60 minutes, the posterior border of the next somite has formed, whereas the caudal c-Hairy1 expression zone has moved further rostral. D, The somite oscillator cycle is complete when c-Hairy1 expression returns to the caudal half of the segmental plate and an anterior zone prefiguring the posterior half of the next somite (as in A). NC, Notocord; So, somite; HN, Hensen’s node. (Somite numbering system adopted from Christ et al., 1998, in which the last formed somite is always numbered with Roman numeral I. Adapted from Palmieram et al., 1997.)

FIGURE 39.5

importance of somite border formation in axial skeletal development is reflected in mutations of the Notch–Delta signaling system, which cause somite dysmorphogenesis and dramatic forms of scoliosis. For example, the pudgy (pu) mouse is characterized by a highly disorganized axial skeleton with many hemi and fused vertebrae and fused rib elements (Kalter, 1980; Kusumi et al., 1998). These abnormalities are prefigured by highly irregular somites in 9.5 to 10.0 post coitum pu embryos. The mutation was shown to be in DLL3, a heterotypic binding partner of Notch (Dunwoodie et al., 2002). Mutations in the human homolog of DLL3 are known to cause spondylocostal dysplasia, a severe congenital scoliosis with a disorganized vertebral column characterized by multiple hemivertebrae, fused vertebrae, and fused ribs. Disturbances in the formation and maintenance of the axial skeleton lead to abnormal spinal curvatures in the neonate or adult. Abnormal axial skeletal curvature may be in the mediolateral or coronal plane (i.e., scoliosis) or in the dorsolateral or sagittal plane (i.e., kyphosis). However, most abnormal spinal curvature is mixed, with one direction of curvature being more dominant than the other. Scoliosis is the more common of these two abnormal curvatures, and it will be used as a model in this discussion. There are two major classes of scoliosis. Idiopathic scoliosis is a lateral curvature of the spine with no known cause. Depending on the definition of curvature, the incidence of idiopathic scoliosis may be as high as 2 to 3 cases per 1000 live births. It becomes evident postnatally, frequently during adolescence, and mostly in

882

SKELETAL DEVELOPMENT

females. Although the molecular genetics of idiopathic scoliosis remains unknown, there is a strong inheritable component. Many cases occur in the context of other musculoskeletal syndromes, including osteogenesis imperfecta, Marfan syndrome, Ehlers–Danlos syndrome, muscular dystrophies, cerebral palsy, and spinal trauma in which variations in the muscles and connective tissues that surround the axial skeleton cause imbalances in tension that lead to abnormal curvatures (Giampietro et al., 2003). Congenital scoliosis is evident at birth, and it results from an underlying axial skeletal dysmorphogenesis. The incidence of congenital scoliosis is 0.5 to 1 case per 1000 live births, with a slightly greater incidence seen among males. Abnormalities of the vertebrae include hemivertebrae, wedge vertebrae, block vertebrae, vertebral bars, and butterfly vertebrae, among others. Hemivertebrae may be caused by ectopic or additional segments, whereas block vertebrae and vertebral bars are likely caused by a failure to produce or maintain segmentation (Erol et al., 2004). Although congenital scoliosis can be associated with genetic syndromes such as spondylocostal dysplasia, spondylocostal dysostosis, Alagille syndrome, Klippel–Feil syndrome, and Jarcho–Levin syndrome, up to 60% of congenital scoliosis cases result from an unknown cause (Giampietro et al., 2003). The involvement of other organ systems is common in these cases, which include spinal and neural tissues; urogenital, gastrointestinal, and cardiovascular tissues; and the craniofacial and appendicular skeletons. These syndromes include Klippel–Feil syndrome, Goldenhar’s syndrome/OAV (oculoauriculovertebral dysplasia), VATER (vertebral anomalies, anal atresia, tracheo-esophageal fistula with esophageal atresia, and renal dysplasia), VACTERL (vertebral anomalies, anal atresia, cardiac malformations, tracheo-esophageal fistula, renal dysplasia and limb deformities) and the MURCS association (Mullerian duct aplasia-renal aplasia-cervicothoracic somite dysplasia; Table 39.1). In contrast with idiopathic scoliosis, these cases usually arise spontaneously, without a strong genetic component. It has been hypothesized that the production and degradation of inhibitors or activators of Notch may reflect or constitute the somitogenic oscillator itself. Lunatic fringe is an excellent candidate, because its expression oscillates at the posterior border of the next somite to be formed (Figure 39.6; Evrard et al., 1998). Cells expressing Lunatic fringe can be transplanted into the caudal segmental plate and induce ectopic border formation. The transgenic overexpression of Lunatic fringe in the mouse embryo overwhelms endogenous Lunatic fringe oscillations in the segmental plate, and it inhibits all segmentation, thereby suggesting that the degradation of this component may constitute the somitic oscillator itself (Dale et al., 2003). The concept of a wavefront in the clock and wavefront model implies an anterior–posterior gradient of a signal through the paraxial mesoderm. Two molecules with gradients of protein concentration in the paraxial mesoderm have been shown to be especially important in somitogenesis (Aulehla and Herrmann, 2004): FGF-8 (Sawada et al., 2001) and WNT-3a (Liu et al., 1999b). The message and protein of the first candidate wavefront molecule, FGF8, exhibits a caudorostral gradient in the segmental plate (see Figure 39.6; Dubrulle et al., 2001) that may be produced through an RNA decay mechanism (Dubrulle and Pourquie, 2004). Increasing the local concentration of FGF8 results in smaller somites, whereas the inhibition of FGF-8 signaling produces larger somites, thereby suggesting that this signal determines the

883

AXIAL SKELETON

TABLE 39.1

Selected Human Syndromes with Scoliosis

Syndrome

Gene

Spondylocostal dysotosis and Jarcho–Levin syndrome

Dll3Mesp2Lnfg

Alagille syndrome

Jgd1

Klippel–Feil syndrome MURCS

VATER

VACTERL

Goldenhar syndrome/OAV

Vertebral Abnormalities Multiple hemivertebrae, fused vertebrae along the length of the spine, and fused ribs Butterfly vertebrae and shortened interpedicular space Fused cervical vertebrae Shortened and fused vertebrae of the cervical and thoracic spines Hemivertebrae and fused vertebrae, usually thoracic and lumbar and includes cases of lumbar–sacral agenesis Hemivertebrae and fused vertebrae, usually thoracic and lumbar and includes cases of lumbar–sacral agenesis Fused vertebrae, usually cervical or thoracic

Other Anomalies

OMIM #

Cranial neural tube defects, genitourinary defects

227300

Craniofacial, cardiac, and ocular defects; liver disease

118450

Deafness, ear malformation, asymmetric facies Mu¨llerian duct aplasia, unilateral renal aplasia

118110

Anal atresia, tracheoesophageal fistula with esophageal atresia, renal dysplasia

192350

Anal atresia, cardiac malformations, tracheoesophageal fistula, renal dysplasia, limb deformities

192350

First and second branchial arch derivative hypoplasia, including facial clefts; esophageal atresia; cardiac, central nervous system, eye, and ear defects

164210

601076

position of the somite boundary in response to a spatially progressive signal. The second candidate wavefront molecule, WNT-3a, is only expressed in the mesoderm as it migrates from the node (see Figure 39.6; Aulehla et al., 2003). A caudorostral gradient is possibly created through the extracellular degradation of this protein. In the spontaneous WNT-3a mutant mouse vestigial tail (vt), FGF8 is downregulated in the tail buds and the segmental plate, thereby suggesting that FGF-8 acts downstream of WNT-3a (Greco et al., 1996). In either case, as the concentration of the gradient molecule is reduced

884

SKELETAL DEVELOPMENT

FIGURE 39.6 Expression patterns of several important genes during somitogenesis. A schematic representation of a chick embryo with various stages of somitogenesis, including the segmental plate with somitomeres, condensed somites (I-V), epithelial somites (VI-X), notochord (NC), neural tube (NT), neural folds (NF), Hensen’s node (HN), and primitive streak (PS). The asterisk indicates the posterior border of the next somite to be formed. The spatial gene expression pattern is indicated by shaded areas. (Adapted from Kuan et al., 2004.)

in more rostral positions, a threshold is reached in which border formation and epithelialization are permitted. 3. Somite Epithelialization The epithelialization of the somite is dependent on a different signaling and adhesion system. Although the cyclic expression of somitogenic genes is observed in isolated segmental plates, overt somite morphogenesis is absent, because signals emanating from the overlying ectoderm are required (Figure 39.7). The removal of the ectoderm is known to inhibit somite formation, but it does not alter the cycling expression of the Notch–Delta system. The signal responsible for this activity is WNT-6, which is produced by the ectoderm overlying the segmental plate (Rodriguez-Niedenfuehr et al., 2003; Schmidt et al., 2004). An important transducer of WNT signaling is b-catenin, which acts both as a transcriptional regulator in the canonic WNT pathway and a component of focal adhesion complexes. During somitogenesis, b-catenin translocates to the plasma membrane, and it associates with N-cadherin as somitic cells increase their affinity for each other and form a spheric epithelium.

AXIAL SKELETON

885

FIGURE 39.7

Tissue interactions involved in the specification of the sclerotome in the ventromedial portion of the epithelial somite. WNT-6 produced by the overlying ectoderm (ECT) induces the epithelialization of the somite. Sonic hedgehog emitted from the notochord (NC) and the floor plate of the neural tube (NT) induces Pax-1 in the sclerotome (shaded region). Pax-1 expression is inhibited dorsally and ventrally by bone morphogenetic protein 4. MYF5, MYOD, and Pax-3 are maintained in the dorsolateral portion of the epithelial somite. IM, Intermediate mesoderm; LM, lateral plate mesoderm. (Adapted from Alexander PG: The role of paraxis in somitogenesis and carbon monoxide-induced axial skeletal teratogenesis, PhD thesis, Thomas Jefferson University, Philadelphia, Penn, 2001.)

The disruption of N-cadherin function inhibits somite epithelialization (Linask et al., 1998). This disruption may reflect the inhibition of b-catenin’s transcriptional function and/or the inhibition of cadherin-based cell–cell adhesion via focal adhesion complexes. The interplay between the two functions of b-catenin is not clearly understood. Somite epithelialization requires increased cell–cell and cell–matrix interactions in the segmental plate. Cadherins responsible for homotypic cell–cell adhesion play a vital role in this; this group includes several members, such as N-cadherin (Hatta et al., 1987), cadherin 11 (Kimura et al., 1995), and protocadherin (Kim et al., 2003). ECM molecules that are important in epithelial structures, such as FN and laminin, are also upregulated at this time (Solursh et al., 1979). These molecules are expressed in the rostral half of the segmental plate, and this coincides with the expression of several transcription factors, such as paraxis (see Figure 39.6); the expression of paraxis prefigures somite formation and epithelialization (Burgess et al., 1995; Barnes et al., 1997) and regulates b-catenin activity during somitogenesis (Linker et al., 2005). The knockdown and knockout of paraxis in various animal models have revealed both the importance of paraxis in epithelialization and the importance of the somite in organizing the axial tissues (Barnes et al., 1997; Burgess et al., 1996; Sosic et al., 1997). Although evidence of segmentation is still observed, somite differentiation in paraxis knockout mice is delayed and less precise, and this results in disorganized tissues. Homozygotes lacking paraxis die at birth as a result of an inability to breathe that is caused by the disorganization of the intercostal musculature (Burgess et al., 1996). The expression pattern and function of axin2 suggest a link between the somitic gradients of WNT-3a and FGF-8 and the somitic oscillator that acts on the Notch–Delta signaling pathway (Aulehla and Herrmann, 2004). Axin2, which is a cytoplasmic component of canonic WNT pathways, is expressed in a caudorostral gradient under the control of WNT-3a. It is also expressed in a cyclic manner in the segmental plate at the posterior margin of the next somite to be formed, like Lunatic fringe. However, axin2 expression alternates with

886

SKELETAL DEVELOPMENT

Lunatic fringe (Aulehla et al., 2003). In mice that are null for the Notch ligand DLL1, Notch signaling is impaired, but axin2 oscillatory expression is still detectable. By contrast, the lack of axin2 in vt mice stops the oscillatory expression of Lunatic fringe but not its spatial expression pattern, which suggests a mechanistic link between the gradient and the clock. Axin2 downregulation during somite border formation may permit the influence of the canonic WNT-6 pathway on the segmental plate, thereby suggesting a mechanistic link with border formation and epithelialization as well. The results of microsurgical and gene knockdown techniques support this model; however, axin2 knockout mice show no axial skeletal defect. Alternatively, mutations in axin1 in the fused (fu) mouse produce profound axial skeletal dysmorphogenesis (Zeng et al., 1997). This discrepancy may reflect fundamental problems with previous experimental methods, the misinterpretation of phenomena, or a species-specific difference between mice and other developmental models. D. Differentiation of the Sclerotome After the patterning of the somite is accomplished, overt differentiation begins. The sclerotome that gives rise to the axial skeleton is derived from the ventromedial compartment of the somite (see Figure 39.4). During the differentiation of the sclerotome, the dorsolateral component of the epithelial somite, the dermomyotome, retains its epithelial structure, whereas the ventromedial sclerotome undergoes an epithelial–mesenchymal transition. This de-epithelialization is preceded by the expression of Pax-1 (see Figure 39.6; Barnes et al., 1996), a paired-box transcription factor that is disrupted in the undulated mouse (Balling et al., 1988), the phenotype of which includes scoliosis and missing vertebral bodies. The induction of Pax-1 expression is mediated by Sonic hedgehog (Shh), which is produced by the notochord and floor plate of the neural tube (see Figure 39.7; Johnson et al., 1994). Ablation of the notochord will prevent the expression of Pax-1 in the somite and the differentiation of the sclerotome. When a notochord or an Shh-soaked bead is implanted dorsal to the somite, Pax-1 is induced in the dorsal somite at the expense of myotomal markers (Brand-Saberi et al., 1993; Ebensberger et al., 1995). BMP-4 produced by the dorsal neural tube and intermediate mesoderm limits or modifies the effect of Shh to constrain sclerotome differentiation in the somite both dorsally and laterally (see Figure 39.4; McMahon et al., 1998; Tonegawa and Takahashi, 1998). After the sclerotome delaminates, it migrates or relocates to the region around the notochord. The function of Pax-1 and other early markers of sclerotome, such as paired-box gene 9 (Pax-9) and mesenchyme forkhead 1 (MFH1), is to maintain the sclerotome by promoting proliferation and preventing cell death (Christ et al., 2004). Functionally, this maintains the sclerotome in a mesenchymal state, and it presumably prevents differentiation. As the sclerotome develops, the posterior half becomes increasingly dense, whereas the anterior half remains loosely populated; this reflects an important anterior– posterior polarity. The differentiation of the sclerotome begins in the posterior half in close association with the notochord. This is coincident with a downregulation of Pax-1 and -9 and an upregulation of Sox9, which is a master regulator of chondrogenic differentiation that controls the expression of major cartilage constituents such as collagen types II and X and aggrecan in a manner that is similar to that of long-bone development in the limb. The

AXIAL SKELETON

887

wave of differentiation proceeds radially from the notochord dorsally to surround the neural tube and ventrally toward the ventral body wall. The sclerotomal cells that inhabit the area around the notochord will give rise to the centrum of the vertebral body. Other structures of the axial skeleton are also derived from the sclerotome, but they require additional regulatory mechanisms. E. Neural Arch and Rib Development An example of this alternative sclerotomal development is that of the vertebral neural arches that surround the spinal cord laterally and dorsally (Figure 39.8; Watanabe et al., 1998). The dorsal–medial sclerotome remains in a mesenchymal state, proliferating and migrating toward the dorsal ectoderm to surround the neural tube. These cells express MSX2, which maintains the mesenchymal phenotype of the sclerotome and which is required for subsequent chondrogenic differentiation (Monsoro-Burq et al., 1996; Monsoro-Burq and Le Dourain, 2000). This differentiation is dependent on BMP-4 signaling from the dorsal neural tube and the overlying ectoderm. The experimental dorsoventral rotation of the neural tube that removes the dorsal influence of BMP-4 on the dorsal sclerotome results in a failure of neural arch development (Nifuji et al., 1997).

FIGURE 39.8 Anatomy of the vertebral body. A, Anterior view of a thoracic vertebral body. Important anatomic features include the centrum (C), the neural arch (NA), the transverse process (TP), the spinous process (SP), the spinal canal (SC), the superior articular process or pedicle (SAP), and the superior costal fovea (SCF). B, Lateral view of a thoracic vertebral body. The gray area in A indicates a Pax-1–expressing sclerotomal origin. Additional anatomic features of a vertebral body include the inferior articular process or pedicle (IAP), the costal fovea of the transverse process (CFTP), and the inferior costal fovea (ICF).

888

SKELETAL DEVELOPMENT

The transverse processes and ribs are also derived from the Pax-1–expressing sclerotome. The body wall in which the ribs reside is derived from the lateral plate that grows ventrally to encase the trunk and abdominal organs. During the time of rib mesenchyme outgrowth, Pax-3–expressing myoblasts and sclerotomal cells both migrate toward the expanding lateral plate mesoderm (Henderson et al., 1999). Quail–chick chimeras, in which the sclerotome or dermomyotome of a chick was replaced by homotypic quail compartments, showed that all parts of the transverse processes and ribs are derived from the sclerotome (Oliviera-Martinez et al., 2000; Evans, 2003). When the overlying ectoderm was removed from the body wall, rib development was severely impaired, which suggests an additional and uncharacterized inductive mechanism that regulates muscle and cartilage development (Sudo et al., 2001). Much less is known about pectoral or pelvic girdle development, although errors in this cause disorders such as congenital hip dysplasia. Recent work looking at scapular development has revealed a dual origin: the blade of the scapula is derived from somitic mesoderm, whereas the head and neck are derived from the lateral plate mesoderm (Huang et al., 2000b; Ehehalt et al., 2004). The somitic cells that form the scapular blade come from the chondrogenic differentiation of dermomyotomal cells that are induced to express Pax-l. In contrast with the vertebral body, Pax-1 is induced (rather than inhibited) by BMP signals from the somatopleure (Wang et al., 2005). The identity of other factors important to the coordination of appendicular and axial skeletal articulation remains unknown. F. Tendon Development The bony elements of the axial skeleton develop in close association with the overlying musculature. The interaction is mutually dependent, and it is critical for the development of other connective tissues: the tendons and the ligaments. The characterization of a marker of ligament and tendon matrix, tenascin-C, has suggested that interactions between muscle and bone are important to the development of these connective tissues (Edom-Vovard and Duprez, 2004). Recent studies reported the discovery of a cellular marker for tendons and ligaments, scleraxis (a basic helix–loop–helix protein closely related to myogenic differentiation factor 1 [MYOD1] and paraxis), expressed throughout their development (Schweitzer et al., 2001; Brent et al., 2003). In the axial skeleton, scleraxis is first expressed at the cranial and caudal borders of the dorsalmost sclerotome in close proximity with the overlying myotome. This compartment has been named the syndetome (Brent and Tabin, 2002), and it is formed as a result of sclerotomal differentiation in a manner that is dependent on FGF8 being emitted by the myotome. The removal of the myotome prevents scleraxis induction and tendon/ligament development, whereas an implanted FGF-8–soaked bead can rescue this deficit in vivo (Brent et al., 2004). G. Resegmentation A rostrocaudal polarity of the prevertebral sclerotome was first shown in 1855 (Remak, 1855 as cited in Stern and Keynes, 1987). Visually, one can see that the posterior half of the mesenchymal sclerotome is more dense than the anterior half. In addition, early embryologists identified a group of elongated and transversely oriented sclerotomal cells separating the rostral and caudal halves of the somite (von Ebner’s fissures). The existence of these features confirms the

AXIAL SKELETON

889

observation that the two halves of the epithelial somite have different adhesive qualities and that they have an influence on migrating neural crest cell and axons (Stern et al., 1986; Norris et al., 1989; Tosney, 1991). These cells migrate exclusively over the cranial portion (Oakley and Tosney, 1991). Somite anterior– posterior polarity is acquired from the very beginning with the differential localization of the Notch ligands DLL1 and DLL3, and it is reiterated by other molecules (see Figure 39.6). The polarity of the somites may serve to direct the migration of neural crest and neuronal axons and to organize other tissues associated with the somites in preparation for the resegmentation of the sclerotomes. Resegmentation is the process by which the bony elements are realigned out of phase with respect to other somite-derived tissues (i.e., muscles, tendons, ligaments, blood vessels) and the neural ganglia (Figure 39.9) to form a functional vertebral joint. Quail–chick chimeras have shown that, to form a new

FIGURE 39.9 Resegmentation of the somites in relation to the adult axial skeleton. A, Schematic of epithelial somites with specified sclerotome (SCL) and dermomyotome (DM, black) and the caudal, more densely packed portion of the sclerotome (medium grey). B, Schematic of resegmented, de-epithelialized sclerotomes with the densely packed sclerotome (medium grey), which contributes to the annulus fibrosus of the intervertebral disc and possibly the anterior portions of the vertebral body. The loosely packed regions of the sclerotomes undergo chondrogenesis and endochondral ossification to form the majority of the vertebral body, including the centrum, the processes, and the ribs. Regional fates of the notochord (NC) are specified with light grey for the degenerating notochord neighboring the cartilaginous regions of the sclerotome and dark grey for the regions that will give rise to the nucleus pulposus of the intervertebral disc. C, Four thoracic vertebrae with the contribution of the posterior somites in A and corresponding densely packed anterior sclerotome to these vertebrae and annulus fibrosus (AF; medium grey). The contribution of the notochord to the nucleus pulposus (NP) within the intervertebral disc (IVD) is highlighted in dark grey. The relationship of the vertebral bodies to their musculature derived from the dermomyotome is indicated in black.

890

SKELETAL DEVELOPMENT

segment, the posterior half of the mesenchymal sclerotome derived from one somite recombines with the anterior half of the mesenchymal sclerotome derived from the anterior somite (Huang et al., 2000a). In this way, the softtissue elements, including the muscles, now span the space between two vertebral bodies to produce an articulating joint (Bagnall et al., 1988; Bagnall, 1992; Aoyama and Asamoto, 2000). The loosely populated half of the new sclerotomal segment gives rise to the vertebral body (see Figure 39.9). The contribution of the more dense anterior half is not as clear. Part of this half does contribute to the bony vertebral body that protects the nerves of the dorsal and ventral ganglia and the blood vessels that supply a particular axial segment passing over and through it. The other portion contributes to the soft tissues and the annulus fibrosus of the intervertebral disc, although how this is accomplished remains unknown. H. Intervertebral Joint During the resegmentation process, the mesenchymal somitocoele cells once located within the epithelial somite are relocated to intervertebral space. Vital dye staining revealed that these cells or their progeny are involved in joint development. The removal of the original somitocoelic cells and their replacement with an acellular spacer resulted in fused vertebral centra and pedicles, thereby providing more evidence that these cells either give rise to the tissues of the vertebral joint or direct its development (Mittapalli et al., 2005). The origin of the intervertebral disc, which is the specialized joint between the centra of two vertebrae (see Figure 39.9), is still controversial (Hunter et al., 2003). There are three components to the intervertebral disc: (1) the articular surfaces of the anterior and posterior vertebral bodies; (2) the annulus fibrosus, which is a specialized ligament that consists of concentric layers of collagen fibers that surround the third tissue; and (3) the nucleus pulposus, which is a more highly hydrated ECM-rich center of the intervertebral disc. As mentioned previously, the annulus fibrosis is derived from the densely packed sclerotomes after resegmentation. The origin of the nucleus pulposus, which absorbs the compressive forces along the axis, is elusive. In some vertebrates (e.g., chicks), the notochordal cells persist throughout life, whereas these cells disappear via apoptosis, terminal chondrocytic differentiation, or another phenotypic change in larger species (e.g., mice, humans). The original notochord-derived cells may be replaced by mesenchymal cells from the annulus fibrosis or other sources. Clinically, it has been conjectured that the loss of the notochordal cells is closely associated with the development of intervertebral disc degeneration and associated back pain. I. Rostrocaudal Vertebral Identity The axial skeleton is characterized by its metamerism, which is a pattern of similar—but not identical—repeating units. A comparison of a cervical, thoracic, lumbar, and sacral vertebra demonstrates the rostrocaudal variances. The regionalization of the somites and the constituent sclerotomes is determined by specific combinations of Hox genes in the paraxial mesoderm. In humans, there are 13 paralogous groups that are aligned in four clusters. These clusters are expressed in a colinear fashion rostrocaudally (30 to 50 along the genome), which suggests an epigenetic mechanism of gene regulation (Gruss and Kessel, 1991; Christ et al., 2000; see Chapter 9). Refinement of the

CRANIOFACIAL SKELETON DEVELOPMENT

891

anterior–posterior expression of these genes may occur through the action of retinoic acid (RA), a physiologic metabolite of vitamin A that is expressed throughout development in regions undergoing patterning and morphogenesis and that are known to affect skeletal development. In particular, the exogenous administration of RA can alter Hox gene expression during concomitant homeotic transformations of axial segments (Kessel, 1992; Kessel and Gruss, 1991). Such transformations are manifested as mandibular hypoplasia, cervical ribs, or thirteenth ribs. In extreme cases, RA can produce caudal agenesis (Padmanabhan, 1998). The genetic knockout of particular RA receptors or dehydrogenases can lead to regional skeletal abnormalities (Cui et al., 2003; Niederreither et al., 2002), whereas the knockout of RA-metabolizing enzymes such as CYP26A1 causes more global skeletal dysmorphogenesis (Sakai et al., 2001). Sclerotomal regional identity is determined very early during development (Fomenou et al., 2005). Heterotopic segmental plate grafts between chick and quail have shown that, when cervical regions are transplanted into thoracic regions, ribs and scapula do not develop (Christ, 1978). RA is produced by cells of Hensen’s node, and recent work has begun to link RA with the somitic oscillator and segment identity (Diez et al., 2003).

III. CRANIOFACIAL SKELETON DEVELOPMENT Skeletogenesis of the skull is highly variable among species, it involves the exact coordinated interaction of several tissue types, and it is dominated by the differentiation of cranial neural crest cells, which is in contrast with the embryonic origins of the axial and appendicular skeleton. The study of craniofacial development thus presents new challenges. Nevertheless, the face is a defining feature of an individual’s identity, and craniofacial abnormalities can have profound sociologic and physiologic consequences. Therefore, the skeletogenesis of the head and face is of intense research interest. Several major subdivisions of the bones that comprise the skull are recognized: (1) the viscerocranium, which is the skeleton of the face and pharynx; (2) the cartilaginous neurocranium, which forms the base of the skull; and (3) the membranous neurocranium, which includes the bones that form the cranial vault. Elements of the viscerocranium can be further subdivided into cartilaginous and membranous categories. The subdivisions of viscerocranium and neurocranium are based on the process of bone formation that each undergoes. A. Neural Crest The principal cellular contributors to the craniofacial skeleton are the neural crest cells. Neural crest cells are a heterogeneous, multipotential group of cells that form during the differentiation of the neural plate from the ectoderm (see Chapter 26). They arise from the border of these two tissues at the dorsal edge of the neural folds along the entire length of the neural tube. These cells can contribute to a wide variety of tissues, including cranial nerves and meninges, dermis, muscle, connective tissues, and glands. Their fate varies on the basis of intrinsic gene expression and the extrinsic inductive signals that they receive during migration and at the site of destination (Le Dourain et al., 2004).

892

SKELETAL DEVELOPMENT

The contribution of neural crest of a certain origin to particular bones is still under investigation. In vivo studies involving techniques of laser ablation and the generation of chicken–quail chimeras have revealed the migratory pattern and regional existence of the neural crest cells in the bones of the head and face. However, the contribution of neural crest to cranial facial bones was only recently clarified using the Cre-lox transgenic technique that labeled neural crest cells with lacZ using the WNT-1 promotor, a factor that has been shown to be necessary and sufficient for neural crest induction (Jiang et al., 2000). With the exception of the parietal bones, all head and neck bones were strongly labeled with lacZ, thus indicating a strong or dominant contribution by the neural crest. In summary, neural crest cells contribute to the craniofacial skeleton to varying degrees and through different skeletogenic processes (Wilkie and Moriss-Kay, 2001). The neural crest cells that populate the pharyngeal arches and the cephalic paraxial mesoderm in the region of the prechordal plate and occipital somites contribute to the cartilaginous viscerocranium and the cartilaginous neurocranium via endochondral ossification, which was described previously. Neural crest cells that populate the mesenchyme that surrounds the dorsal half of the brain form the membranous neurocranium, and those that invade the first branchial arch form the membranous viscerocranium or facial bones. Both of these cranial bone sets are formed by intramembranous ossification, which is a process that is distinct from endochondral ossification. B. Intramembranous Ossification The membranous neurocranium consists of bones that form the cranial vault. These include the frontal and parietal bones (Figure 39.10, A and B), which form from the frontal and parietal eminences (the embryonic thickenings of the cephalic mesoderm in the prosencephalic and mesencephalic regions).

FIGURE 39.10 Bones and sutures of the neurocranium. Bones of the neonatal skull, A, laterally and B, dorsally. The neurocranium is composed of the frontal (F) and parietal (P) bones separated by the coronal (C), sagittal (S), and metopial (M) sutures that join at the anterior fontanelle (AF). The neurocranium abuts the maxillary (Mx), sphenoid (not shown), and temporal (T) bones of the viscerocranium, separated by the sphenoidal fontanelle (SF) and the occipital bone (O) of the chondrocranium, separated by the lambdoid suture (L) and the mastoid fontanelle (MF). Md, Mandible. Illustrations of the two kinds of sutures, C, overlapping and D, abutting, showing the basic relationship between two lamellar bones (LB) with osteogenic fronts (Ost) separated by undifferentiated mesenchyme (Mes) in close relationship with the underlying dura mater (DM) and the overlying periosteum (PO). (Adapted from Sadler TW: Langman’s medical embryology, 7th ed, Baltimore, Md, 1995, Williams and Wilkins.)

CRANIOFACIAL SKELETON DEVELOPMENT

893

These two types of eminences are actually of differing tissue origins: the frontal is of neural crest origin, and the parietal is of cephalic paraxial mesoderm. Despite their differences in origin, they both produce bone by an alternative skeletogenic process called intramembranous ossification. Although endochondral ossification requires a cartilaginous model before bone deposition, intramembranous ossification is characterized by mesenchymal cells that condense and directly form osteogenic nodules (Ornitz and Marie, 2002). These nodules form in close association with blood vessels of the dura mater (in cranial vault development) or of other membranes, such as the perichondrial sheath of Meckel’s cartilage (in mandibular development). In fact, the dura mater and the overlying perichondrium are essential for calvarial morphogenesis, postnatal suture fusion, and osseous repair of calvarial defects. In these nodules, proteoglycan-rich chondroid is produced, and it is quickly calcified to osteoid (Lengele et al., 1996). Cells that are encased in this matrix become osteoblasts, whereas those that surround the nodules become the periosteum. The inner layer of the periosteum gives rise to a progenitor population that undergoes osteogenesis and that forms layers of bone or lamellae that characterize membranous bones (Figure 39.11). Macroscopically, this process advances from the initial nodules through the formation of long spicules of bone that fuse with one another to form plates. The bones of the cranial vault expand to abut one another during development, but they do not fuse. The juncture between cranial vault bones becomes a functional structure called a suture, which coordinates the growth of the skull with the underlying brain. The suture is an anatomically simple structure that is composed of two apposing plates in juxtaposing or overlapping configuration that are separated by a narrow space that is analogous to a growth plate (see Figure 39.10, C). The suture contains regions of

FIGURE 39.11 Microanatomy of the suture and molecular pathways regulating suture growth and osteogenesis. A schematic diagram of suture differentiation zones showing selected transcription factors expressed in the cells of the suture and signaling molecules emitted by the dura mater influencing the overlying mesenchyme. (Adapted from Ornitz and Marie, 2002.)

894

SKELETAL DEVELOPMENT

undifferentiated mesenchyme, rapidly dividing osteogenic stem cells, differentiated osteoblasts producing osteoid, and mature osteoblasts that are encased in a bony matrix (see Figure 39.11). The growth and timing of suture closure is carefully controlled by its interaction with the underlying dura mater in a manner that is analogous to the perichondrium/periosteum of endochondral ossification. The molecular mechanisms of intramembranous ossification and suture growth are not as well characterized as endochondral ossification and growth plate growth. The direct ossification (or the lack of chondrogenic intermediate) during the intramembranous process has been clearly shown in RUNX2 knockout mouse. RUNX2 is a master regulator of osteoblast-differentiation– regulating genes such as osteocalcin, osteopontin, bone sialoprotein, collagens, alkaline phosphatase, vitamin D receptor, osteoprotegerin, TGF-b receptors, and others (Stein et al., 2004). In accordance with this, investigators found that RUNX2 knockout homozygotes produce a cartilaginous skeleton without any indication of osteogenesis (Otto et al., 1997). In addition, the membranous neurocranium and the medial clavicle are completely absent. The heterozygotic phenotype is reminiscent of the human condition called cleidocranial dysplasia, which is caused by a mutation of the human RUNX2 gene (Mundlos et al., 1997). This is one of the few mutations that reveal a genuine error in the intramembranous mechanism itself. Craniosynostosis occurs in 1 in every 2500 live births, and it is caused by the premature fusion of one or more sutures. Errors in suture closure are attributed to imbalances in the proliferative and differentiation phases of suture growth. Studies of the molecular genetics behind craniosynostosis have provided information about important players during the various phases of suture development, with activating mutations in FGFR-2 being the most frequently implicated (Table 39.2; Marie et al., 2005). However, other FGF members are expressed during intramembranous ossification in the suture, and recent work has shown the involvement of other molecules that begin to build a pathway of osteogenesis in the suture (Nie et al., 2006). Loose, resting mesenchymal cells within the suture are poorly characterized, but it is known that these cells express FGFR-1 and that they are under the influence of several FGFs, including FGF2, -4, -9 and -18 (see Figure 39.11). Markers of these cells in particular are not known, but it is known that the forced expression of Hox genes (e.g., Hoxa-2, which is required for visceral cranial neural crest osteogenesis) actually inhibits suture development in the neurocranium (Creuzet et al., 2002; Couly et al., 2002). Proliferative mesenchymal cells in the suture express Twist and muscle segment homeobox 1 (MSX1), which are transcription factors that maintain cells in an undifferentiated state in several developmental contexts (Takahashi et al., 2003; Soo et al., 2002; Ishii et al., 2003). FGF-2 plays a dominant role at this stage. A switch to FGFR-1 is important for initiating differentiation and the expression of RUNX2 and aristaless-like 4 (ALX4) by downregulating the expression of twist and MSX2 (Ishii et al., 2005). Both MSX2 and Twist mutant mice reveal interesting features in intramembranous ossification. Dominant-negative mutations in MSX2 accelerate the differentiation of the sutural mesenchyme, and this results in Boston-type craniosynostosis. Alternatively, activating mutations of MSX2 prevent differentiation that leads to parietal foramina (an open suture; Winograd et al., 1997). This illustrates the delicate balance between mesenchymal proliferation and the differentiation

895

CRANIOFACIAL SKELETON DEVELOPMENT

TABLE 39.2 Selected Fibroblast Growth Factor Receptor, Msx2, and Twist Mutations and Their Associated Human Syndromes Gene

Mutation

Syndrome

Phenotype

FGFR2

C342R

Pfeiffer Crouzon Jackson–Weiss

P253R

Apert

Sagittal synostosis or cloverleaf skull, duplication of thumbs and big toe Variable cranial synostosis, hypertelorism, beaked nose, hypoplastic maxilla, and mandibular prognathism with no limb abnormalities Variable cranial synostosis, midfacial hypoplasia, and foot anomalies Variable cranial synostosis, midfacial hypoplasia, hand and foot syndactyly and variable fusion of hand, foot, and vertebral fusions

S252W or F S252L FGF3

P250R

Apert Normal (mild Crouzon) Muenke

FGFR1

P252R

Pfeiffer

Twist

S123stop

Saethre–Chotzen

Saethre–Chotzen

Msx2

416ins21 (insKIIPTLP) P148H

Boston craniosynostosis

)




No cranial synostosis with variable syndactyly Coronal synostosis, variable syndactyly and brachydactyly of the hand Sagittal synostosis or cloverleaf skull, duplication of thumbs and big toe Asymmetric cranial synostosis, widely spaced eyes, beaked nose, syndactyly and brachydactyly of the hand

Cloverleaf skull, supraorbital recession

that governs suture development (Liu et al., 1999). The importance of Twist is seen in cases of Saethre–Chotzen syndrome, in which an inactivating mutation in Twist leads to early suture fusion. Recently, the switch to mesenchymal differentiation in the suture was shown to be enabled by the activation of the BMP pathway (Kanzler et al., 2000), and this resulted in changed Twist expression. Noggin, which is an endogenous inhibitor of BMPs, is expressed in open sutures, and its downregulation is associated with suture closure that is concomitant with a decrease in Twist expression. Importantly, activating mutations in FGFR-2 that are known to cause early suture osteogenesis and closure has been shown to decrease Noggin expression. Having described a generalized sequence of suture growth regulation, the incidence of particular craniosynostotic patterns with specific genetic mutations argues that there are differences among sutures. Two specific knockout mice illustrate the point. EphrinB1 knockout mice have premature metopic suture closure (Twigg et al., 2004), whereas axin2 knockout mice are characterized by premature coronal suture closure (Yu et al., 2005). These two sutures regulate growth between the neurocranial bones of cranial paraxial mesodermal or neural crest origin, respectively. Thus, the specific molecular regulatory mechanisms of osteogenesis in different sutures can vary on the basis of the origin of the constituent cells. Osteogenic differentiation is thought to occur by similar means in both endochondral and intramembranous ossification, because RUNX2 is required for both processes. In fact, the observation that limb abnormalities are associated with many craniosynostoses suggests that craniofacial and limb

896

SKELETAL DEVELOPMENT

development share some mechanistic processes, such as the requirement for RUNX2 in osteogenesis and the involvement of the FGF signaling pathways in bone growth. Another similarity is the activity of the parathyroid hormone/IGF-1 axis involved in osteogenesis and calcium homeostasis. However, the activation of this pathway is clearly different between the intramembranous and endochondral ossification mechanisms (Suda et al., 2001). In endochondral ossification, Ihh regulates the transition from proliferative to hypertrophic chondrocytes and induces the expression of PTHrP signaling components. Without a cartilaginous intermediate, the PTHrP pathway is expressed during intramembranous ossification in a manner that is independent of Ihh. The means of this induction is still unknown, thus revealing subtle differences in the mechanisms of osteogenesis in the long bones and the calvaria that are yet to be elucidated. Techniques of experimental embryology, such as tissue transplants (e.g., chick–quail chimeras) and organ explant cultures, play an important role in the formation of the principles of development from a morphologic perspective, thereby suggesting cellular and molecular behaviors and interactions that are required to produce the changes observed. Genetic analyses have revealed the link between phenotype and genotype, and, when these analyses are applied to developmental problems, they shed light on the specific cellular and molecular pathways that characterize early ontogeny. These applications include characterizations of spontaneous mutants through positional cloning (e.g., Pudgy), knockouts (RUNX2), and transgenesis (Wnt-1–lacZ). A detailed understanding of embryology is required for the proper interpretation of these experiments. As tissue-specific expression approaches (e.g., Cre-Lox and other inducible systems) become more refined and widespread, it will be possible to manipulate in utero specific cell and tissue interactions that are otherwise inaccessible or previously unidentified. As discussed previously, these techniques are beginning to clarify fundamental mechanisms of skeletogenesis. The advent of genomics expands the experimental approaches to skeletal development, thus enabling the identification and analysis of candidate genes and their function(s) and establishing the different “transcriptosomes” responsible for the developmental processes that are critical to skeletogenesis. Information obtained from these studies should provide insights into disease causes, phenotype variability, wound healing, and the mechanisms of oncology.

SUMMARY

 Bones and connective tissues of the body are derived from two embryolo-





gic sources: the mesoderm and the neural crest. The appendicular skeleton arises from the lateral plate mesoderm, the axial skeleton arises from the paraxial mesoderm, and the craniofacial skeleton arises from a combination of the cephalic paraxial mesoderm and neural crest cells. Bones are formed through one of two processes: —Endochondral ossification involves the formation of a cartilaginous model, which is subsequently replaced by bone. —Intramembranous ossification is characterized by the direct differentiation of mesenchymal cells to osteogenic tissue. Bone growth is regulated by specific, analogous growth zones: the growth plate in the long bones and the sutures between cranial bones. The FGF

897

GLOSSARY



  



  

signaling system plays a critical role in the regulation of these growth zones. Mutations in different FGF receptors cause achondroplasias and craniosynostoses. The replacement of cartilage by osteoblasts is dependent on the prior mineralization of the cartilage matrix and vascular invasion. The activities of the hypertrophic chondrocytes and of the osteoblasts are regulated by RUNX2. Osteoclasts are of hemopoietic origin (i.e., monocytes), and, with osteoblasts, they remodel bone throughout an individual’s lifetime to adapt to mechanical demands. Joints are formed via the bifurcation or apposition of chondrogenic models. WNT-14 and the activity of b-catenin are central regulators of joint formation. The cellular origin of articular cartilage is unknown. Somites are formed from the segmental plate through the interaction of a cell-autonomous oscillator and a morphogenetic wave that coordinates somite boundary formation via the Notch–Delta juxtacrine system and epithelialization via the WNT paracrine system and cadherin activity. The dysregulation of this process leads to congenital scoliosis. The axial skeleton is derived from the ventromedial sclerotome of somites. The differentiation of the somite is regulated by paracrine factors from surrounding tissues, and the sclerotome in particular is induced by Shh. Tendons and ligaments develop from the interaction of mesodermal mesenchyme with myoblasts. The nucleus pulposus of the intervertebral disc is initially derived from the notochord. The gradual loss of “embryonic cells” from the disc may be related to disc degeneration and back pain. The craniofacial skeleton is derived from cells of different origins, and it forms through either intramembranous or endochondral ossification. The constituent cells’ origin and local environment determine the specific mechanisms of osteogenesis.

ACKNOWLEDGMENTS This work is supported by the Intramural Research Program, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health (Z01-AR-41106).

GLOSSARY Endochondral ossification A mechanism of bone formation in which a cartilage model is replaced with ossified tissue that is produced by invading osteoblasts. Growth plate The area of developing tissue between the diaphysis and the epiphysis and that is responsible for the longitudinal growth of bones. The suture, which is an analogous structure between the bones of the neurocranium, coordinates skull growth with brain development.

898

SKELETAL DEVELOPMENT

Hypertrophy An increase in bulk without a concomitant multiplication of parts. Interzone A region of nonchondrogenic cells that forms at the site of a future joint. Intramembranous ossification A mechanism of bone formation in which ossified tissue is produced directly from condensed mesenchyme lying in close association with a membranous structure. Somite An embryonic tissue that patterns and maintains the metamerism of axial tissues. Resegmentation The process by which the sclerotomes of neighboring somites reorganize to form vertebral bodies that span the space between the dermomyotomal derivatives of the same somites.

REFERENCES Ahrens PB, Solursh M, Reiter RS: Stage-related capacity for limb chondrogenesis in cell culture, Dev Biol 60:69–82, 1977. Akiyama H, Chaboissier MC, Martin JF, et al: The transcription factor Sox9 has essential roles in successive steps of the chondrocyte differentiation pathway and is required for expression of Sox5 and Sox6, Genes Dev 16:2813–2828, 2002. Alvarez J, Horton J, Sohn P, Serra R: The perichondrium plays an important role in mediating the effects of TGF-beta1 on endochondral bone formation, Dev Dyn 221:311–321, 2001. Alvarez J, Sohn P, Zeng X, et al: TGFbeta2 mediates the effects of hedgehog on hypertrophic differentiation and PTHrP expression, Development 129:1913–1924, 2002. Amizuka N, Warshawsky H, Henderson JE, et al: Parathyroid hormone-related peptide-depleted mice show abnormal epiphyseal cartilage development and altered endochondral bone formation, J Cell Biol 126:1611–1623, 1994. Aoyama H, Asamoto K: The developmental fate of the rostral-caudal half of a somite for vertebra and rib formation: experimental confirmation of the resegmentation theory using quail-chick chimeras, Mech Dev 99:71–82, 2000. Archer CW, Dowthwaite GP, Francis-West P: Development of synovial joints, Birth Defects Res C Embryo Today 69:144–155, 2003. Auhlehla A, Wehrle C, Brand-Saberi B, et al: Wnt3a plays a major role in the segmentation clock controlling somitogenesis, Dev Cell 4:395–406, 2003. Aulehla A, Herrmann BG: Segmentation in vertebrates: clock and gradient finally joined, Genes Dev 18:2060–2067, 2004. Aulthouse AL, Solursh M: The detection of a precartilage, blastema-specific marker, Dev Biol 120:377–384, 1987. Bagnall KM: The migration and distribution of somite cells after labelling with the carbocyanine dye, Dil: the relationship of this distribution to segmentation in the vertebrate body, Anat Embryol 185:317–324, 1992. Bagnall KM, Higgins SJ, Sanders EJ: The contribution made by a single somite to the vertebral column: experimental evidence in support of resegmentation using the chick-quail chimera model, Development 103:69–85, 1988. Balling R, Deutsch U, Gruss P: Undulated, a mutation affecting the development of the mouse skeleton, has a point mutation in the paired box of Pax1, Cell 55:531–535, 1988. Barnes GL, Hsu CW, Mariani BD, Tuan RS: Chicken Pax-1 gene: structure and expression during embryonic somite development, Differentiation 61:13–23, 1996. Barnes GL, Alexander PG, Hsu CW, et al: Cloning and characterization of chicken Paraxis: a regulator of paraxial mesoderm development and somite formation, Dev Biol 189:95–111, 1997.

REFERENCES

899 Bi W, Deng JM, Zhang Z, et al: Sox9 is required for cartilage formation, Nat Genet 22:85–89, 1999. Boulet AM, Capecchi MR: Multiple roles of Hoxa11 and Hoxd11 in the formation of the mammalian forelimb zeugopod, Development 131:299–309, 2004. Boyle WJ, Simonet WS, Lacey DL: Osteoclast differentiation and activation, Nature 423:337–342, 2003. Brand-Saberi B, Ebensberger C, Wilting J, et al: The ventralizing effect of the notochord on somite differentiation in chick embryos, Anat Embryol 188:239–245, 1993. Brent AE, Tabin CJ: Developmental regulation of somite derivatives: muscles, cartilage and tendon, Curr Op Genes Dev 12:548–557, 2002. Brent AE, Schweitzer R, Tabin CJ: A somitic compartment of tendon progenitors, Cell 113:235–248, 2003. Brent AE, Tabin CJ: FGF acts directly on the somitic tendon progenitors through the Ets transcription factors Pea3 and Erm to regulate Scleraxis expression, Development 131:3885–3896, 2004. Brunet LJ, McMahon JA, McMahon AP, Harland RM: Noggin, cartilage morphogenesis, and joint formation in the mammalian skeleton, Science 280:1455–1457, 1998. Burgess R, Cserjesi P, Ligon KL, Olson EN: Paraxis: a basic helix-loop-helix protein expressed in paraxial mesoderm and developing somites, Dev Biol 168:296–306, 1995. Burgess R, Rawls A, Brown D, et al: Requirement of the Paraxis gene for somite formation and musculoskeletal patterning, Nature 384:570–573, 1996. Capdevila J, Izpisua Belmonte JC: Patterning mechanisms controlling vertebrate limb development, Ann Rev Cell Dev Biol 17:87–132, 2001. Chang SC, Hoang B, Thomas JT, et al: Cartilage-derived morphogenetic proteins. New members of the transforming growth factor-beta superfamily predominantly expressed in long bones during human embryonic development, J Biol Chem 269:28227–28234, 1994. Christ B, Huang R, Wilting J: The development of the avian vertebral column, Anat Embryol 202:179–194, 2000. Christ B, Huang R, Scaal M: Formation and differentiation of the avian sclerotome, Anat Embryol 208:333–350, 2004. Christ B, Schmidt C, Huang R, et al: Segmentation of the vertebrate body, Anat Embryol (Berl) 197:1–8, 1998. Christ B, Jacob HJ, Jacob M: Regional determination of early embryonic muscle primordium. Experimental studies on quail and chick embryos (demonstration), Verh Anat Ges 72:353–357, 1978. Coleman CM, Loredo GA, Lo CW, Tuan RS: Correlation of GDF5 and connexin 43 mRNA expression during embryonic development, Anat Rec A Discov Mol Cell Evol Biol 275:1117–1121, 2003. Coleman CM, Tuan RS: Functional role of growth/differentiation factor 5 in chondrogenesis of limb mesenchymal cells, Mech Dev 120:823–836, 2003. Colnot C, Lu C, Hu D, Helms JA: Distinguishing the contributions of the perichondrium, cartilage, and vascular endothelium to skeletal development, Dev Biol 269:55–69, 2004. Cooke J, Zeeman EC: A clock and wavefront model for control of the number of repeated structures during animal morphogenesis, J Theoret Biol 58:455–476, 1976. Couly G, Creuzet S, Bennaceur S, et al: Interactions between Hox-negative cephalic neural crest cells and the foregut endoderm in patterning the facial skeleton in the vertebrate head, Development 129:1061–1073, 2002. Cremer MA, Rosloniec EF, Kang AH: The cartilage collagens: a review of their structure, organization, and role in the pathogenesis of experimental arthritis in animals and in human rheumatic disease, J Mol Med 76:275–288, 1998. Creuzet S, Couly G, Vincent C, Le Douarin NM: Negative effect of Hox gene expression on the development of the neural crest-derived facial skeleton, Development 129:4301–4313, 2002. Cui J, Michaille JJ, Jiang W, Zile MH: Retinoid receptors and vitamin A deficiency: differential patterns of transcription during early avian development and the rapid induction of RARs by retinoic acid, Dev Biol 260:496–511, 2003. Dale JK, Maroto M, Dequeant ML, et al: Periodic Notch inhibition by lunatic fringe underlies the chick segmentation clock, Nature 421:275–278, 2003. Delise AM, Tuan RS: Alterations in the spatiotemporal expression pattern and function of N-cadherin inhibit cellular condensation and chondrogenesis of limb mesenchymal cells in vitro, J Cell Biochem 87:342–359, 2002a. Delise AM, Tuan RS: Analysis of N-cadherin function in limb mesenchymal chondrogenesis in vitro, Dev Dyn 225:195–204, 2002b.

900

SKELETAL DEVELOPMENT

Dessau W, von der Mark H, von der Mark K, Fischer S: Changes in the patterns of collagens and fibronectin during limb-bud chondrogenesis, J Embryol Exp Morphol 57:51–60, 1980. Diez del Corral R, Olivera-Martinez I, Goriely A, et al: Opposing FGF and retinoid pathways control ventral neural pattern, neuronal differentiation, and segmentation during body axis extension, Neuron 40:65–79, 2003. Dougall WC, Glaccum M, Charrier K, et al: RANK is essential for osteoclast and lymph node development, Genes Dev 13:2412–2424, 1999. Dubrulle J, McGrew MJ, Pourquie O: FGF signaling controls somite boundary position and regulates segmentation clock control of spatiotemporal Hox gene activation, Cell 106:219–232, 2001. Dubrulle J, McGrew MJ, Pourquie O: Fgf0008 mRNA decay establishes a gradient that couples axial elongation to patterning in the vertebrate embryo, Nature 427:419–422, 2004. Duprez DM, Kostakopoulou K, Francis-West PH, et al: Activation of Fgf-4 and HoxD gene expression by BMP-2 expressing cells in the developing chick limb, Development 122:1821–1828, 1996. Dudhia J: Aggrecan, aging and assembly in articular cartilage, Cell Mol Life Sci 62:2241–2256, 2005. Dunwoodie SL, Clements M, Sparrow DB, et al: Axial skeletal defects caused by mutation in the spondylocostal dysplasia/pudgy gene Dll3 are associated with disruption of the segmentation clock within the presomitic mesoderm, Development 129:1795–1806, 2002. Ebensperger C, Wilting J, Brand-Saberi B, et al: Pax-1, a regulator of sclerotome development is induced by notochord and floor plate signals in avian embryos, Anat Embryol 191:297–310, 1995. Edom-Vovard F, Duprez D: Signals regulating tendon formation during chick embryonic development, Dev Dyn 229:449–457, 2004. Ehehalt F, Wang B, Christ B, et al: Intrinsic cartilage-forming potential of dermomyotomal cells requires ectodermal signals for the development of the scapula blade, Anat Embryol 208:431–437, 2004. Enomoto H, Enomoto-Iwamoto M, Iwamoto M, et al: Cbfa1 is a positive regulatory factor in chondrocyte maturation, J Biol Chem 275:8695–8702, 2000. Erol B, Tracy MR, Dormans JP, et al: Congenital scoliosis and vertebral malformations: characterization of segmental defects for genetic analysis, Pediatr Orthop 24:674–682, 2004. Evans DJ: Contribution of somitic cells to the avian ribs, Dev Biol 256:114–126, 2003. Evrard YA, Lun Y, Aulehla A, et al: Lunatic fringe is an essential mediator of somite segmentation and patterning, Nature 394:377–381, 1998. Fomenou MD, Scaal M, Stockdale FE, et al: Cells of all somitic compartments are determined with respect to segmental identity, Dev Dyn 233:1386–1393, 2005. Francis-West PH, Abdelfattah A, Chen P, et al: Mechanisms of GDF-5 action during skeletal development, Development 126:1305–1315, 1999. Fujita T, Azuma Y, Fukuyama R, et al: Runx2 induces osteoblast and chondrocyte differentiation and enhances their migration by coupling with PI3K-Akt signaling, J Cell Biol 166:85–95, 2004. Giampietro PF, Blank RD, Raggio CL, et al: Congenital and idiopathic scoliosis: clinical and genetic aspects, Clin Med Res 1:125–136, 2003. Goldring MB, Tsuchimochi K, Ijiri K: The control of chondrogenesis, J Cell Biochem 97:33–44, 2006. Goff DJ, Tabin CJ: Analysis of Hoxd-13 and Hoxd-11 misexpression in chick limb buds reveals that Hox genes affect both bone condensation and growth, Development 124:627–636, 1997. Greco TL, Takada S, Newhouse MM, et al: Analysis of the vestigial tail mutation demonstrates that Wnt3a gene dosage regulates mouse axial development, Genes Dev 10:313–324, 1996. Gruss P, Kessel M: Axial specification in higher vertebrates, Curr Biol 1:204–210, 1991. Gu H, Zou YR, Rajewsky K: Independent control of immunoglobulin switch recombination at individual switch regions evidenced through Cre-loxP-mediated gene targeting, Cell 73:1155–1164, 1993. Guo X, Day TF, Jiang X, et al: Wnt/beta-catenin signaling is sufficient and necessary for synovial joint formation, Genes Dev 18:2404–2417, 2004. Hall BK, Miyake T: All for one and one for all: condensations and the initiation of skeletal development, Bioessays 22:138–147, 2000. Hamburger V, Hamilton HL: A series of normal stages in the development of the chick embryo, J Morphol 88:49–92, 1951.

REFERENCES

901 Hartmann C, Tabin CJ: Wnt-14 plays a pivotal role in inducing synovial joint formation in the developing appendicular skeleton, Cell 104:341–351, 2001. Hatta K, Takagi S, Hajime F, Takeichi M: Spatial and temporal expression pattern of N-cadherin cell adhesion molecule correlates with morphogenetic processes of chick embryos, Dev Biol 120:215–227, 1987. Healy C, Uwanogho D, Sharpe PT: Regulation and role of Sox9 in cartilage formation, Dev Dyn 215:69–78, 1999. Henderson DJ, Conway SJ, Copp AJ: Rib truncations and fusions in the Sp2H mouse reveal a role for Pax3 in specification of the ventro-lateral and posterior parts of the somite, Dev Biol 209:143–158, 1999. Hornbruch A, Summerbell D, Wolpert L: Somite formation in the early chick embryo following grafts of Hensen’s node, J Embryol Exp Morphol 51:51–62, 1979. Huang R, Zhi Q, Brand-Saberi B, Christ B: New experimental evidence for somite resegmentation, Anat Embryol 202:195–200, 2000a. Huang R, Zhi Q, Patel K, et al: Dual origin and segmental organization of the avian scapula, Development 127:3789–3794, 2000b. Hunter CJ, Matyas JR, Duncan NA: The notochordal cell in the nucleus pulposus: a review in the context of tissue engineering, Tissue Eng 9:667–677, 2003. Ishii M, Han J, Yen HY, et al: Combined deficiencies of Msx1 and Msx2 cause impaired patterning and survival of the cranial neural crest, Development 132:4937–4950, 2005. Ishii M, Merrill AE, Chan YS, et al: Msx2 and Twist cooperatively control the development of the neural crest-derived skeletogenic mesenchyme of the murine skull vault, Development 130:6131–6142, 2003. Jena N, Martin-Seisdedos C, McCue P, Croce CM: BMP7 null mutation in mice: developmental defects in skeleton, kidney, and eye, Exp Cell Res 230:28–37, 1997. Jiang X, Rowitch DH, Soriano P, et al: Fate of the mammalian cardiac neural crest, Development 127:1607–1616, 2000. Johnson RL, Laeufer E, Riddle RD, Tabin CJ: Ectopic expression of Sonic hedgehog alters dorsalventral patterning of somites, Cell 79:1165–1173, 1994. Kalter H: A compendium of the genetically induced congenital malformations of the house mouse, Teratology 21:397–429, 1980. Kanzler B, Foreman RK, Labosky PA, Mallo M: BMP signaling is essential for development of skeletogenic and neurogenic cranial neural crest, Development 127:1095–1104, 2000. Kessel M: Respecification of vertebral identities by retinoic acid, Development 115:487–501, 1992. Kessel M, Gruss P: Homeotic transformations of murine vertebrae and concomitant alteration of Hox codes induced by retinoic acid, Cell 67:89–104, 1991. Kim SH, Jen WC, De Robertis EM, Kintner C: The protocadherin PAPC establishes segmental boundaries during somitogenesis in xenopus embryos, Curr Biol 10:821–830, 2000. Kimura Y, Matsunami H, Inoue T, et al: Cadherin-11 expressed in association with mesenchymal morphogenesis in the head, somite, and limb bud of early mouse embryos, Dev Biol 169:347–358, 1995. Kipnes J, Carlberg AL, Loredo GA, et al: Effect of cartilage oligomeric matrix protein on mesenchymal chondrogenesis in vitro, Osteoarthrit Cart 11:442–454, 2003. Kornak U, Mundlos S: Genetic disorders of the skeleton: a developmental approach, Am J Hum Genet 73:447–474, 2003. Kuan CYK, Tannahill D, Cook GMW, Keynes RJ: Somite polarity and segmental patterning of the peripheral nervous system, Mech Dev 121:1055–1068, 2004. Kusumi K, Sun ES, Kerrebrock AW, et al: The pudgy mutation disrupts Delta-homolog Dll3 and initiation of early somite boundaries, Nat Genet 19:271–278, 1998. Kwan KM, Pang MK, Zhou S, et al: Abnormal compartmentalization of cartilage matrix components in mice lacking collagen X: implications for function, J Cell Biol 136:459–471, 1997. Le Douarin NM, Creuzet S, Couly G, Dupin E: Neural crest cell plasticity and its limits, Development 131:4637–4650, 2004. Lefebvre V, Li P, de Crombrugghe B: A new long form of Sox5 (L-Sox5), Sox6 and Sox9 are coexpressed in chondrogenesis and cooperatively activate the type II collagen gene, EMBO J 17:5718–5733, 1998. Lengele B, Schowing J, Dhem A: Embryonic origin and fate of chondroid tissue and secondary cartilages in the avian skull, Anat Rec 246:377–393, 1996. L’Hote CG, Knowles MA: Cell responses to FGFR3 signalling: growth, differentiation and apoptosis, Exp Cell Res 304:417–431, 2005.

902

SKELETAL DEVELOPMENT

Linask KK, Ludwig C, Han MD, et al: N-cadherin/catenin-mediated morphoregulation of somite formation, Dev Biol 202:85–102, 1998. Linker C, Lesbros C, Gros J, et al: beta-catenin-dependent Wnt signalling controls the epithelial organisation of somites through the activation of Paraxis, Development 132:3895–3905, 2005. Liu YH, Tang Z, Kundu RK, et al: Msx2 gene dosage influences the number of proliferative osteogenic cells in growth centers of the developing murine skull: a possible mechanism for MSX2-mediated craniosynostosis in humans, Dev Biol 205:260–274, 1999a. Liu P, Wakamiya M, Shea MJ, et al: Requirement for Wnt3a in vertebrate axis formation, Nat Genet 22:361–365, 1999b. Lizarraga G, Lichtler A, Upholt WB, Kosher RA: Studies on the role of Cux1 in regulation of the onset of joint formation in the developing limb, Dev Biol 243:44–54, 2002. Maleski MP, Knudson CB: Hyaluronan-mediated aggregation of limb bud mesenchyme and mesenchymal condensation during chondrogenesis, Exp Cell Res 225:55–66, 1996. Marie PJ, Coffin JD, Hurley MM: FGF and FGFR signaling in chondroplasia and craniosynostosis, J Cell Biochem 95:888–896, 2005. McMahon JA, Takada S, Zimmermann LB, et al: Noggin-mediated antagonism of BMP signaling is required for growth and patterning of the neural tube and somite, Genes Dev 12:1438–1452, 1998. Meech R, Edelman DB, Jones FS, Makarenkova HP: The homeobox transcription factor Barx2 regulates chondrogenesis during limb development, Development 132:2135–2146, 2005. Minina E, Kreschel C, Naski MC, et al: Interaction of FGF, Ihh/Pthlh, and BMP signaling integrates chondrocyte proliferation and hypertrophic differentiation, Dev Cell 3:439–449, 2002. Minina E, Wenzel HM, Kreschel C, et al: BMP and Ihh/PTHrP signaling interact to coordinate chondrocyte proliferation and differentiation, Development 128:4523–4534, 2001. Mittapalli VR, Huang R, Patel K, et al: Arthrotome: a specific joint forming compartment in the avian somite, Dev Dyn 234:48–53, 2005. Monsoro-Burq AH, Duprez D, Watanabe Y, et al: The role of bone morphogenetic proteins in vertebral development, Development 122:3607–3616, 1996. Monsoro-Burq AH, Le Douarin N: Duality of molecular signaling involved in vertebral chondrogenesis, Curr Top Dev Biol 48:43–75, 2000. Mundlos S, Otto F, Mundlos C, et al: Mutations involving the transcription factor CBFA1 cause cleidocranial dysplasia, Cell 89:773–779, 1997. Nakashima K, Zhou X, Kunkel G, et al: The novel zinc finger-containing transcription factor osterix is required for osteoblast differentiation and bone formation, Cell 108:17–29, 2002. Nasevicius A, Ekker SC: Effective targeted gene ‘knockdown’ in zebrafish, Nat Genet 26:216–220, 2000. Nie X, Luukko K, Kettunen P: FGF signaling in craniofacial development and developmental disorders, Oral Dis 12:102–111, 2006. Niederreither K, Fraulob V, Garnier JM, et al: Differential expression of retinoic acid-synthesizing (RALDH) enzymes during fetal development and organ differentiation in the mouse, Mech Dev 110:165–171, 2002. Nifuji A, Kellermann O, Kuboki Y, et al: Perturbation of BMP signaling in somitogenesis resulted in vertebral and rib malformations in the axial skeletal formation, J Bone Min Res 12:332–342, 1997. Norris WE, Stern CD, Keynes RJ: Molecular differences between the rostral and caudal halves of the sclerotome in the chick embryo, Development 105:541–548, 1989. Oakley RA, Tosney KW: Peanut agglutinin and chondroitin-6 sulfate are molecular markers for tissues that act as barriers to axon advance in the avian embryo, Dev Biol 147:187–206, 1991. Oberlender SA, Tuan RS: Expression and functional involvement of N-cadherin in embryonic limb chondrogenesis, Development 120:177–187, 1994. Oliviera-Martinez I, Coltey M, Dohouailly D, Pourquie O: Mediolateral somatic origin of ribs and dermis determined by quail-chick chimeras, Development 127:4611–4617, 2000. Orntiz DM, Marie PJ: FGF signaling pathways in endochondral and intramembranous bone development and human genetic disease, Genes Dev 16:1446–1465, 2002. Otto F, Thornell AP, Crompton T, et al: Cbfa1, a candidate gene for cleidocranial dysplasia syndrome, is essential for osteoblast differentiation and bone development, Cell 89:765–771, 1997. Packard DS Jr: The influence of axial structures on chick somite formation, Dev Biol 53 (12):36–48, 1976.

REFERENCES

903 Packard DS Jr, Meier S: An experimental study of the somitomeric organization of the avian segmental plate, Dev Biol 97:191–202, 1983. Padmanabhan R: Retinoic acid-induced caudal regression syndrome in the mouse fetus, Reprod Toxicol 12:139–151, 1998. Primmett DR, Norris WE, Carlson GJ, et al: Periodic segmental anomalies induced by heat shock in the chick embryo are associated with the cell cycle, Development 105:119–130, 1989. Primmett DR, Stern CD, Keynes RJ: Heat shock causes repeated segmental anomalies in the chick embryo, Development 104:331–339, 1988. Pizette S, Niswander L: BMPs are required at two steps of limb chondrogenesis: formation of prechondrogenic condensations and their differentiation into chondrocytes, Dev Biol 219:237–249, 2000. Robson H, Siebler T, Shalet SM, Williams GR: Interactions between GH, IGF-I, glucocorticoids, and thyroid hormones during skeletal growth, Pediatr Res 52:137–147, 2002. Rodriguez-Niedenfuehr M, Dathe V, Jacob HJ, et al: Spatial and temporal pattern of Wnt-6 expression during chick development, Anat Embryol 206:447–451, 2003. Sakai Y, Meno C, Fujii H, et al: The retinoic acid-inactivating enzyme CYP26 is essential for establishing an uneven distribution of retinoic acid along the anterio-posterior axis within the mouse embryo, Genes Dev 15:213–225, 2001. Saga Y, Takeda H: The making of the somite: molecular events in vertebrate segmentation, Nat Rev Genet 2:835–845, 2001. Sawada A, Shinya M, Jiang YJ, et al: Fgf/MAPK signalling is a crucial positional cue in somite boundary formation, Development 128:4873–4880, 2001. Schmidt C, Stoeckelhuber M, McKinnel I, et al: Wnt0006 regulates the epithelialization process of the segmental plate mesoderm leading to somite formation, Dev Biol 271:198–209, 2004. Schweitzer R, Chyung JH, Murtaugh LC, et al: Analysis of the tendon cell fate using Scleraxis, a specific marker for tendons and ligaments, Development 128:3855–3866, 2001. Shore EM, Xu M, Feldman GJ, et al: A recurrent mutation in the BMP type I receptor ACVR1 causes inheritied and sporadic fibrodysplasia ossificans progessiva, Nat Genet 38:525–527, 2006. Smits P, Dy P, Mitra S, Lefebvre V: Sox5 and Sox6 are needed to develop and maintain source, columnar, and hypertrophic chondrocytes in the cartilage growth plate, J Cell Biol 164:747–758, 2004. Smits P, Li P, Mandel J, et al: The transcription factors L-Sox5 and Sox6 are essential for cartilage formation, Dev Cell 1:277–290, 2001. Solloway MJ, Dudley AT, Bikoff EK, et al: Mice lacking Bmp6 function, Dev Genet 22:321–339, 1998. Solursh M, Fischer M, Meier S, Singley CT: The role of extracellular matrix in the formation of the sclerotome, J Embryol Exp Morphol 54:75–98, 1979. Soo K, O’Rourke MP, Khoo PL, et al: Twist function is required for the morphogenesis of the cephalic neural tube and the differentiation of the cranial neural crest cells in the mouse embryo, Dev Biol 247:251–270, 2002. Sosic D, Brand-Saberi B, Schmidt C, et al: Regulation of Paraxis expression and somite formation by ectoderm- and neural tube-derived signals, Dev Biol 185:229–243, 1997. Stern C, Sisodaya S, Keynes R: Interactions between neuritis and somite cells: inhibition and stimulation of nerve growth in the chick embryo, J Embryol Exp Morphol 91:209–226, 1986. Stern, C, Keynes, R: Interactions between somite cells: the formation and maintenance of segment boundaries in the chick embryo. Development 99:261–272, 1987. St-Jacques B, Hammerschmidt M, McMahon AP: Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation, Genes Dev 13:2072–2086, 1999. Stickens D, Behonick DJ, Ortega N, et al: Altered endochondral bone development in matrix metalloproteinase 13-deficient mice, Development 131:5883–5895, 2004. Stringa E, Tuan RS: Chondrogenic cell subpopulation of chick embryonic calvarium: isolation by peanut agglutinin affinity chromatography and in vitro characterization, Anat Embryol 194:427–437, 1996. Stein GS, Lian JB, van Wijnen AJ, et al: Runx2 control of organization, assembly and activity of the regulatory machinery for skeletal gene expression, Oncogene 23:4315–4329, 2004. Storm EE, Huynh TV, Copeland NG, et al: Limb alterations in brachypodism mice due to mutations in a new member of the TGF beta-superfamily, Nature 368:639–643, 1994. Suda N, Baba O, Udagawa N, et al: Parathyroid hormone-related protein is required for normal intramembranous bone development, J Bone Miner Res 16:2182–2191, 2001.

904

SKELETAL DEVELOPMENT

Sudo H, Takahashi Y, Tonegawa A, et al: Inductive signals from the somatopleure mediated by bone morphogenetic proteins are essential for the formation of the sternal component of the avian ribs, Dev Biol 232:284–300, 2001. Takahashi Y, Inoue T, Gossler A, Saga Y: Feedback loops comprising Dll1, Dll3, and Mesp2, and differential involvement of Psen1 are essential for rostrocaudal patterning of somites, Development 130:4259–4268, 2003. Tonegawa A, Takahashi Y: Somitogenesis controlled by Noggin, Dev Biol 202:172–182, 1998. Tosney KW: Cell and cell-interactions that guide motor axons in the developing embryo, Bioessays 13:17–24, 1991. Tsumaki N, Nakase T, Miyaji T, et al: Bone morphogenetic protein signals are required for cartilage formation and differently regulate joint development during skeletogenesis, J Bone Miner Res 17:898–906, 2002. Twigg SR, Kan R, Babbs C, et al: Mutations of ephrin-B1 (EFNB1), a marker of tissue boundary formation, cause craniofrontonasal syndrome, Proc Natl Acad Sci U S A 101:8652–8657, 2004. Vu TH, Shipley JM, Bergers G, et al: MMP-9/gelatinase B is a key regulator of growth plate angiogenesis and apoptosis of hypertrophic chondrocytes, Cell 93:411–422, 1998. Wang B, He L, Ehehalt F, et al: The formation of the avian scapula blade takes place in the hypaxial domain of the somites and requires somatopleure-derived BMP signals, Dev Biol 287:11–18, 2005. Watanabe Y, Duprez D, Monsoro-Burq AH, et al: Two domains in vertebral development: antagonistic regulation by SHH and BMP4 proteins, Development 125:2631–2639, 1998. White DG, Hershey HP, Moss JJ, et al: Functional analysis of fibronectin isoforms in chondrogenesis: full-length recombinant mesenchymal fibronectin reduces spreading and promotes condensation and chondrogenesis of limb mesenchymal cells, Differentiation 71:251–261, 2003. Wilkie AOM, Moriss-Kay GM: Genetics of craniofacial development and malformation, Nat Genet 2:458–468, 2001. Williams DRJr, Presar AR, Richmond AT, et al: Limb chondrogenesis is compromised in the versican deficient hdf mouse, Biochem Biophys Res Comm 334:960–966, 2005. Yu HM, Jerchow B, Sheu TJ, et al: The role of Axin2 in calvarial morphogenesis and craniosynostosis, Development 132:1995–2005, 2005. Zelzer E, McLean W, Ng YS, et al: Skeletal defects in VEGF(120/120) mice reveal multiple roles for VEGF in skeletogenesis, Development 129:1893–1904, 2002. Zeng L, Fagotto F, Zhang T, et al: The mouse fused locus encodes Axin, an inhibitor of the Wnt signaling pathway that regulates embryonic axis formation, Cell 90:181–192, 1997. Zheng Q, Zhou G, Morello R, et al: Type X collagen gene regulation by Runx2 contributes directly to its hypertrophic chondrocyte-specific expression in vivo, J Cell Biol 162: 833–842, 2003.

FURTHER READING Barrantes IB, Elia AJ, Wunsch K, et al: Interaction between Notch signalling and Lunatic fringe during somite boundary formation in the mouse, Curr Biol 6:470–480, 1999. Conlon RA, Reaume AG, Rossant J: Notch1 required for the coordinate segmentation of somites, Development 121:1533–1545, 1995. Grapin-Botton A, Bonnin MA, McNaughton LA, et al: Plasticity of transposed rhombomeres: Hox gene induction is correlated with phenotypic modifications, Development 121:2707–2721, 1995. Hrabe de Angelis M, McIntyre JN, Gossler A: Maintenance of somite borders in mice requires the Delta homologue Dll1, Nature 386:717–721, 1997. Johnson J, Rhee J, Parsons SM, et al: The anterior/posterior polarity of somites is disrupted in Paraxis-deficient mice, Dev Biol 229:176–187, 2001. McCright B, Lozier J, Gridley T: A mouse model of Alagille syndrome: Notch2 as a genetic modifier of Jag1 haploinsufficiency, Development 129:1075–1082, 2002. Oka C, Nakano T, Wakeham A, et al: Disruption of the mouse RBP-J kappa gene results in early embryonic death, Development 121:3291–3301, 1995. Palmeirim I, Henrique D, Ish-Horowicz D, Pourquie O: Avian Hairy gene expression identifies a molecular clock linked to vertebrate segmentation and somitogenesis, Cell 91:639–648, 1997.

CRANIOFACIAL DEVELOPMENT/INTRAMEMBRANOUS OSSIFICATION

905

Palmeirim I, Dubrulle J, Henrique D, et al: Uncoupling segmentation and somitogenesis in the chick presomitic mesoderm, Dev Genet 23:77–85, 1998. Saga Y, Hata N, Koseki H, Taketo MM: Mesp2: a novel mouse gene expressed in the presegmented mesoderm and essential for segmentation initiation, Genes Dev 11:1827–1839, 1997. Takada S, Stark KL, Shea MJ, et al: Wnt3a regulates somite and tailbud formation in the mouse embryo, Genes Dev 8:174–189, 1994. Takahashi K, Nuckolls GH, Takahashi I, et al: Msx2 is a repressor of chondrogenic differentiation in migratory cranial neural crest cells, Dev Dyn 222:252–262, 2001. Wong PC, Zheng H, Chen H, et al: Presenilin 1 is required for Notch1 and DII1 expression in the paraxial mesoderm, Nature 387:288–292, 1997. Watanabe Y, Le Douarin NM: A role for BMP-4 in the development of subcutaneous cartilage, Mech Dev 57:69–78, 1996. Winograd J, Reilly MP, Roe R, et al: Perinatal lethality and multiple craniofacial malformations in MSX2 transgenic mice, Human Mol Genet 6:369–379, 1997. Zhang N, Gridley T: Defects in somite formation in lunatic fringe-deficient mice, Nature 394:374–377, 1998.

RECOMMENDED RESOURCES Online Mendelian Inheritance in Man: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db¼OMIM

CHONDROGENESIS/ENDOCHONDRAL OSSIFICATION Capdevila J, Izpisua Belmonte JC: Patterning mechanisms controlling vertebrate limb development, Ann Rev Cell Dev Biol 17:87–132, 2001. Hall BK, Miyake T: All for one and one for all: condensations and the initiation of skeletal development, Bioessays 22:138–147, 2000. Kornak U, Mundlos S: Genetic disorders of the skeleton: a developmental approach, Am J Hum Genet 73:447–474, 2003.

SOMITOGENESIS Aulehla A, Herrmann BG: Segmentation in vertebrates: clock and gradient finally joined, Genes Dev 18:2060–2067, 2004. Christ B, Huang R, Scaal M: Formation and differentiation of the avian sclerotome, Anat Embryol 208:333–350, 2004. Giampietro PF, Blank RD, Raggio CL, et al: Congenital and idiopathic scoliosis: clinical and genetic aspects, Clin Med Res 1:125–136, 2003. Kalter H: A compendium of the genetically induced congenital malformations of the house mouse, Teratology 21:397–429, 1980. Saga Y, Takeda H: The making of the somite: molecular events in vertebrate segmentation, Nat Rev Genet 2:835–845, 2001.

CRANIOFACIAL DEVELOPMENT/INTRAMEMBRANOUS OSSIFICATION Marie PJ, Coffin JD, Hurley MM: FGF and FGFR signaling in chondroplasia and craniosynostosis, J Cell Biochem 95:888–896, 2005. Wilkie AOM, Moriss-Kay GM: Genetics of craniofacial development and malformation, Nat Genet 2:458–468, 2001.

VI ENDODERMAL ORGANS

40

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS BILLIE A. MOORE-SCOTT and JAMES M. WELLS Division of Developmental Biology, Cincinnati Children’s Hospital Research Foundation and University of Cincinnati College of Medicine, Cincinnati, OH

INTRODUCTION The process of gastrulation subdivides the embryonic epiblast into the three primary germ layers: the ectoderm, the mesoderm, and the endoderm. During organogenesis, the endoderm germ layer contributes to many vital organs, including the liver, the pancreas, the thymus, and the thyroid, and it forms the epithelial lining of the gastrointestinal and respiratory tracts (Figures 40.1 and 40.4; reviewed by Wells and Melton, 1999). The endoderm cells of the postgastrulation vertebrate embryo are largely unspecified, and they become regionalized along the anterior–posterior (AP) and dorsal–ventral (DV) axes during the process of the formation of a primitive gut tube. The process during which endoderm cells obtain positional identity along the AP and DV axes is called patterning, which is a fundamental event that is necessary for directing endoderm cells into specific organ lineages. In addition, the endoderm itself acts as a source of signals that regulate the proper development of mesoderm and ectoderm organs such as the anterior central nervous system and the heart. Therefore, failure to appropriately pattern the endoderm can have a broad impact on the developmental outcome of the entire embryo. Several vertebrate animal model systems have greatly contributed to our understanding of early endoderm organogenesis, including Xenopus, zebrafish, chicken, quail, and mouse. It is remarkable that many aspects of early endoderm organogenesis are conserved across vertebrate species. For example, within a few days after gastrulation in amphibians, fish, birds, and mammals, unspecified endoderm cells form a primitive gut tube with distinct AP and DV domains that predict the formation of organ primordia. Furthermore, there is increasing evidence that signals and responding target genes

908

Principles of Developmental Genetics

© 2007, Elsevier Inc. All rights reserved.

909

INTRODUCTION

A

P

P

A

Ectoderm Mesoderm

E7.5

Endoderm

Cdx 1 Cdx 2 Cdx 4

Foxa2 Hex 1

A

A Foxa2

E8.5

Foregut

Nkx2.1 Hex1 Sox2

P

Midgut

Pdx1 Albumin NeuroD

Hindgut

SS

Cdx’s

A

A Tbx1 Nkx2.1 Pax8

E9.5

P

Pax1 Hoxa3 Pax9

Hoxd Hex1 Cdx’s

P

Pdx1

P

FIGURE 40.1 Early stages of endoderm organogenesis in the mouse. The left panels show a schematic of mouse embryos, and the right panels highlight the endoderm and genes that are regionally expressed along the anterior–posterior and dorsal–ventral axes. After gastrulation, which is complete 7.5 days after fertilization in the mouse, the endoderm is the outermost layer of cells (yellow) that surround the embryo. The middle layer is the mesoderm (red), and the inner layer is the ectoderm (blue). After one day of embryonic development (i.e., on embryonic day 8.5), morphogenetic movements at the anterior and posterior initiate tube formation and start the process of internalizing the endoderm. These processes result in the formation of a primitive gut tube with foregut and hindgut tubes. The middle section of the primitive gut tube (midgut) remains open at this stage. By 10.5 days of embryonic development, the formation and internalization of the primitive gut tube is complete, and organ primordia for the lungs, liver, and pancreas are morphologically evident. (See color insert.)

that direct the formation of a patterned gut tube are largely conserved across vertebrate species. In this chapter, we will discuss the morphogenetic and molecular processes that are involved in subdividing the endoderm into functional presumptive organ domains along the AP axis, with a focus on studies in the chick and mouse. These early stages of endoderm organogenesis are critical for the proper development of the gastrointestinal and respiratory tracts and for the ontogeny of the associated vital organs.

910

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

The formation and patterning of the primitive gut tube is a dynamic morphogenetic process that involves cell migration and tissue remodeling. For simplicity, we have subdivided the gut tube into the foregut, the midgut, and the hindgut (Figure 40.1). The anterior segment of the gut tube—the foregut—forms first and comes from the endoderm extending from the most anterior portion of the embryo (headfold in the postgastrula embryo, embryonic day 7.5 to 8 in the mouse) to just below the cardiac mesoderm. The foregut endoderm contributes to a remarkable number and diversity of organs, including the taste buds, the inner ear, the thyroid, the thymus, the esophagus, the trachea, the lungs, the liver, and the ventral pancreas. The hindgut forms shortly after the foregut, and it is derived from the most posterior region of the gastrula stage endoderm overlying the primitive streak. The hindgut contributes to the small and large intestines as well as to the bladder and the urogenital tract. For the purposes of this chapter, we will refer to the open portion of the gut tube during these early stages of gut tube morphogenesis as the midgut. This is the last region of the endoderm to form a tube, and it is comparatively larger than the traditionally defined midgut, which is based on anatomic descriptions of the gut of later-stage embryos. The midgut endoderm as defined at this stage contributes to part of the stomach, to the dorsal pancreas, and to the anterior portions of the small intestine, such as the duodenum.

I. FATE MAP OF THE EMBRYONIC ENDODERM The endoderm organs that develop from the foregut, midgut, and hindgut become morphologically apparent within 2 days of gut tube formation (embryonic day 10.5 in the mouse). To better understand the early stages of endoderm organogenesis, we will discuss in this section the embryonic origins of the different AP and DV domains of the gut tube. Numerous studies have begun to identify the cellular origins of endoderm organ domains using a method called lineage tracing or fate mapping. Lineage tracing is a method during which cells are labeled and followed from an early progenitor stage to a more mature stage of development (Figure 40.2). Thus, an endoderm cell labeled at the late gastrula stage can be traced to specific organ domains as embryonic development progresses. Cell lineage

D

A

D

P

V

V

Lateral view

Frontal view

~E8.0 embryo

Whole embryo culture

Embryo cultured to ~E10.5

Labeling endoderm cells for lineage analysis and generating fate maps. Cells can be labeled by injecting vital dyes (striped; DiI) or by expressing fluorescent proteins (gray). Labeled endoderm cells can be tracked through subsequent stages of development. The two left panels show a lateral and frontal view of an early somite stage embryo (~3 somites, e8 mouse). Embryos can be cultured in vitro, and the positions of the labeled cells can be determined after 1 to 3 days of development. This approach has been used to identify which regions of the early endoderm give rise to specific gut tube and organ domains and to generate fate maps.

FIGURE 40.2

FATE MAP OF THE EMBRYONIC ENDODERM

911

analyses performed in the mouse, chick, and frog have mapped the movement of gastrulation-stage endoderm cells to broad domains of the developing gut tube and early organ primordia (Chalmers and Slack, 2000; Kimura et al., 2006; Lawson and Schoenwolf, 2003; Lawson et al., 1986; Rosenquist, 1971; Tam et al., 2004). These studies suggest that many aspects of early endoderm patterning and gut tube formation are remarkably conserved across vertebrate species. Later in this chapter, we will compare the fate-mapping studies performed in two model organisms: the chick and the mouse. A. Mapping Endoderm Cell Lineage During Gut Tube Morphogenesis: From Gastrula Stages to the Early Somite Stages in the Chick and the Mouse 1. Endoderm Fate Maps at the Gastrula Stages In the mouse and the chick, endoderm cells form during gastrulation and migrate out of the primitive streak to integrate into the outer layer of cells, thereby forming the endoderm germ layer (see Chapter 14). Gastrulation in the mouse occurs between embryonic days 6 and 7.5; in the chick, this occurs between 7 and 19 hours of development (Hamburger and Hamilton [HH] stages 2–5). Cell lineage analyses of early to mid-gastrula embryos of both species suggest that the time at which endoderm cells exit the primitive streak influences where they end up along the AP axis at the end of gastrulation. For example, during the early mouse gastrula (embryonic day 6–6.5), the first endoderm cells to exit the primitive streak preferentially end up in the anterior endoderm overlying the anterior neural plate and trunk at the end of gastrulation (embryonic day 7.5; Lawson and Pedersen, 1987). During the mid- to late gastrula stages (embryonic day 7–7.5), endoderm cells exiting the primitive streak often colonize more posterior domains and remain adjacent to the node and the primitive streak. These fate-mapping studies suggest that the anterior axial endoderm cells are the first endoderm cells to form during gastrulation in the mouse, whereas posterior endoderm forms slightly later. Studies in the chick have added to our understanding of when different domains of endoderm are formed, and they have raised some interesting questions (Kimura et al., 2006; Lawson and Schoenwolf, 2003). One experiment that traced the lineage of axial/midline endoderm cells reached a similar conclusion to the studies in the mouse: that anterior endoderm forms first and is followed by more posterior trunk endoderm (Lawson and Schoenwolf, 2003). However, another study suggested that posterior/lateral endoderm is the first to incorporate into the hypoblast layer during the early gastrula stage (Kimura et al., 2006). These different studies could suggest that midline and lateral endoderm arise from different populations of endoderm precursors, and they emphasize how lineage-tracing experiments continue to reveal new and surprising aspects of endoderm development. For example, one study revealed a population of endoderm cells that exits the epiblast during gastrulation and migrate for several hours before inserting into the endoderm germ layer during the late gastrula stage (Kimura et al., 2006). B. Mapping Late Gastrula Stage Endoderm to the Developing Gut Tube The studies described previously indicate that the late gastrula endoderm (embryonic day 7.5 in the mouse; HH stages 4 and 5 in the chick) may have distinct AP and lateral domains. This section will describe how these domains

912

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

within a two-dimensional sheet of endoderm will give rise to the AP and DV domains of a three-dimensional gut tube (Wells and Melton, 1999). Figure 40.3 compares the chick and mouse endoderm fate maps and illustrates the relationships among the domains of the gastrula stage endoderm (embryonic day 7.5 in the mouse; HH stages 4 and 5 in the chick), the developing gut tube at early somite stages (embryonic day 8.5 in the mouse; HH stage 10 in the chick), and the gut tube at a stage when organ primordia start to become morphologically apparent (embryonic day 9.5 in the mouse; HH stage 17 in the chick). Figure 40.3 is a composite of several cell lineage studies that have begun to reveal the complex Mouse

Late gastrula e7.5

Anterior

Late gastrlua HH4-5

V VII VI

I II

Chick

Posterior

IV

III

Foregut

Foregut HH10 (8-10 somites)

e8.5 (8-10 somites)

Dorsal

Ventral e9.5 (~30 somites) Position at gastrulation stage

HH17 (~30 somites)

Position in the gut tube

Anterior/axial

Anterior/lateral

Dorsal foregut

Ventral forgut

Medial/axial

Medial/lateral

Dorsal forgut/midgut

Ventral forgut/midgut

Posterior/axial

Posterior/lateral

Dorsal hindgut

Ventral hindgut

Fate map of the endoderm and gut tube in the mouse and the chick. In the top panels, late gastrula stage mouse (lateral view) and chick (ventral view, endoderm facing up) embryos are shown. The dotted line in the chick shows the developing notochord. The roman numerals (I–VII) indicate the domains identified in the mouse fate-mapping studies (Lawson et al., 1986; Tam et al., 2004). In the middle panels, a ventral view of 8- to 10-somite stage embryos (embryonic day 8.5 in the mouse; Hamburger and Hamilton stage 10 in the chick) is shown in which the foregut has started to form and to be internalized. The arrows indicate the folding/ migration direction of the foregut and the hindgut. The dotted circles in the middle panels are somites, and the dotted line along the midline is the developing notochord. The dotted line in the anterior outlines the developing foregut. In the bottom panels, a lateral view of mouse (embryonic day 9.5; blue-staining cells are Pdx1-LacZ expression) and chick (Hamburger and Hamilton stage 17) embryos at the approximately 30-somite stage is shown. At this stage, the endoderm has been internalized and folded into a gut tube, with the exception of the hindgut in the chick, which has not fully formed at this stage. The colored domains of the gastrula-stage embryos roughly correspond with the equivalent-colored domains of the later-stage embryos. The axial endoderm of the gastrula embryo (domains I, II, III, and IV) contributes to both the dorsal and the ventral gut tubes, although only domains I and IV contribute ventrally. The lateral domains (V, VI, and VII) fold over to principally contribute to the ventral gut tube. For all images, the anterior is left and the posterior is right. In the mouse, the fate map of the foregut between embryonic days 8.5 and 10 has been well studied (Tremblay and Zaret, 2005). The posterior domains at these stages have not been well studied, and the domains in this figure are extrapolated from the earlier-stage fate map. (The Pdx1-LacZ mice were from Chris Wright of Vanderbilt University. See color insert.)

FIGURE 40.3

FATE MAP OF THE EMBRYONIC ENDODERM

913

morphogenetic processes involved in the formation of the gut tube (Kimura et al., 2006; Lawson et al., 1986; Rosenquist, 1971; Tam et al., 2004; Tremblay and Zaret, 2005). In the mouse, cell lineage experiments performed more than 20 years ago and more recent studies have resulted in a detailed fate map of the late gastrula stage (embryonic day 7.5) of the mouse endoderm (Lawson et al., 1986; Tam et al., 2004). These studies identified regions of endoderm on embryonic day 7.5 that will contribute to the distinct AP and DV domains of the developing gut tube (see Figure 40.3). One recent study distinguished between endoderm cells located along the midline versus those located more laterally. Midline endoderm cells tend to remain along the midline in both the ventral and dorsal regions of the gut tube. By contrast, endoderm cells that are found laterally in the embryonic day 7.5 endoderm preferentially end up in the ventral portions of the gut. These fate-mapping studies concluded that the most anterior midline endoderm cells overlying the neural plate (see Figure 40.3; region I) as well as the anterior lateral endoderm (region V) both contribute to the ventral foregut, which gives rise to the liver, the ventral pancreas, and the lungs (Lawson et al., 1986). Axial endoderm cells just anterior to the node (region II) contribute to the dorsal foregut, which forms the dorsal component of the stomach, the pancreas, and the duodenum, whereas lateral endoderm from region VI contributes to the ventral midgut. Axial endoderm overlying the node and the anterior primitive streak (region III) contributes to the dorsal midgut and hindgut, eventually contributing to the small intestines. The most posterior endoderm overlying the primitive streak (see Figure 40.3; regions IV and VII) contributes to both the dorsal and ventral portions of the hindgut, which contribute to the posterior intestinal derivatives and the urogenital tract. Again, studies suggest that the chick fate map roughly corresponds with that of the mouse (Kimura et al., 2006; Lawson and Schoenwolf, 2003; Rosenquist, 1971): the anterior axial and lateral domains of HH stage 5 endoderm contribute to the ventral foregut at HH stage 10, the endoderm overlying the middle primitive streak contributes to dorsal midgut and hindgut, and the most posterior endoderm over the primitive streak contributes to the ventral portions of the midgut and hindgut. The combined evidence suggests that the presumptive AP and DV regions of the gut tube are established as early as the gastrula stage, before gut tube morphogenesis (see Figure 40.3). Furthermore, the lineage tracing experiments demonstrate that gut tube morphogenesis involves the folding and migration of the anterior/lateral endoderm to make the foregut and of the posterior/lateral endoderm to make the hindgut. The folding/migration of the foregut continues toward the posterior, and similar expansion of the hindgut continues toward the anterior, thereby resulting in the closure of the primitive gut tube. C. Regions of the Foregut That Give Rise to Developing Organ Primordia Less is known about the domains of the early gut tube that give rise to developing the mid-gestation–stage organ primordia. However, a study describing lineage tracing experiments performed on early somite-stage mouse embryos identified domains in the developing foregut that contribute to specific foregut derivatives, including the liver and the ventral pancreas (Tremblay and Zaret, 2005). In these studies, foregut endoderm cells were labeled on embryonic day

914

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

8 to 8.5 (1–7 somites), and their descendants were analyzed at the early liver bud stage (embryonic day 9.5). One medial and two lateral domains of foregut endoderm were found to contribute to the liver bud. In addition, cells of the medial domain at the lip of the developing foregut gave rise to descendants all along the ventral midline, whereas the lateral domains contribute to specific ventral regions such as the liver. This study supports the idea that medial and lateral endoderm cells continue to migrate independently and then later converge during gut tube morphogenesis during the formation of the liver and other ventral foregut derivatives. The experiments described previously as well as studies in other organisms, including Xenopus and zebrafish, highlight the complex morphogenetic processes that occur during the formation and patterning of the primitive gut tube. The later sections discuss specific signaling pathways and target genes that functionally regulate these early stages of endoderm organogenesis.

II. GENE EXPRESSION DOMAINS PREDICT AND DETERMINE ENDODERM ORGAN PRIMORDIA Genes have been identified that are expressed in discrete AP and DV domains during the development of the gut tube. These markers are invaluable for the understanding of how endoderm is patterned at the molecular level. Many of these marker genes both predict where the endoderm organ primordia will form and play a functional role in establishing organ boundaries (Figure 40.4). In this section, we will discuss molecular studies that describe how patterns are established in the early endoderm and refined in the primitive gut tube, and we will also address the role of several of these genes during endoderm organogenesis. At the end of gastrulation in Xenopus, chick, and mouse, most reported gene expression patterns are in broad overlapping domains along the AP axis. For example, in all three species, Hematopoietically expressed homeobox1 (Hex1), orthodenticle homologue 2 (Otx2), and forkhead box A2 (Foxa2) are expressed along the anterior half of the embryo, whereas the Caudal type homeobox genes Cdx1, Cdx2, and Cdx4 (CdxA and CdxC in the chick) are broadly expressed in the posterior half (see Figure 40.1; Ang et al., 1993; Chapman et al., 2002; Frumkin et al., 1993; Gamer and Wright, 1993; Jones et al., 1999; Northrop and Kimelman, 1994). The expression domains of these factors suggest that the endoderm is broadly patterned along the AP axis at this stage. Moreover, gene targeting experiments have demonstrated that Hex1 is vital for anterior patterning as well as for subsequent thyroid and liver formation (Martinez Barbera et al., 2000), and the loss of Cdx2 function results in posterior truncations around the time of gut tube formation (Chawengsaksophak et al., 2004). Foxa2 is functionally important for the formation of the anterior primitive streak, and an endoderm-specific knockout of Foxa2 and of the closely related factor Foxa1 results in the failure to initiate liver development (Lee et al., 2005). By the time that gut tube formation occurs (embryonic day 8.5 in the mouse; HH stage 10 in the chick), a complex pattern of gene expression emerges in the foregut (Nkx2.1, Hex1, and Sox2), the posterior foregut (albumin), the midgut (Pdx1), and the hindgut (Cdx2 and Cdx4; CdxA and CdxB in the chick;

915

GENE EXPRESSION DOMAINS PREDICT AND DETERMINE ENDODERM ORGAN PRIMORDIA

Figures 40.1 and 40.6; Chalmers et al., 2000; Gamer and Wright, 1993; Gittes and Rutter, 1992; Ohlsson et al., 1993; Zeng et al., 1998). The extensive patterning of the primitive gut tube suggests a high degree of gene regulation. Two well-studied examples of highly regulated genes are Pdx1 and Cdx, which are part of the ParaHox cluster that arose from an ancestral ProtoHox cluster. The ParaHox cluster exhibits conserved spatial expression along the AP axis in divergent animal phyla, where Cdx expression defines the domain posterior to Cdx (the small and large intestines) and Pdx1 expression defines the domain anterior to Cdx (the duodenum, the pancreas, and the caudal stomach). These transcription factors may play a role in maintaining their mutually exclusive expression domains. For example, in the chick, the misexpression of Pdx1 in the posterior results in the repression of the Cdx genes, which suggests that Pdx1 directly or indirectly restricts the anterior expression limit of Cdx genes (Grapin-Botton et al., 2001). In addition, the genetic ablation of Pdx1 in the mouse results in arrested pancreatic development and abnormal epithelial development in the rostral duodenum (Grapin-Botton, 2005). D

A

P

Ultimobranchial bodies Thymus Stomach

V

Bladder, urogenital tract

Pancreas

E11.5

Liver Intestine

Lung Esophagus

Thyroid Parathyroid

Hex1 Sox2

Cdx Pdx1 Hoxa2 Hoxa3 Hoxc5 Hoxc6/Hoxb6 Hoxb8 Hoxc9/Hoxb9 Hoxa13/Hoxd13

FIGURE 40.4 Overlapping expression domains of transcription factors predict the emergence of organ primordia along the anterior–posterior axis of the fetal gut tube. The upper schematic shows a ventral view of a fetal gut tube around the time of organ bud formation (embryonic day 10.5–11 in the mouse). Anterior is left and posterior is right. The lower panel indicates the relative anterior–posterior expression boundaries of several homeobox-containing transcription factors. Hex1 and Cdx factors (Cdx1, Cdx2,and Cdx4 in the mouse; CdxA and CdxB in the chick) are expressed in the gastrula-stage endoderm (embryonic day 7.5 in the mouse), whereas other genes, such as Pdx1, are first expressed later in the gut tube (embryonic day 8.5 in the mouse). The anterior and posterior expression limits of some of these factors are important for establishing organ domains. (The lower panel was adapted from Grapin-Botton, 2005. See color insert.)

916

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

By embryonic day 9.5, the gut tube has expanded both anteriorly and posteriorly. At this stage, the pattern in the anterior foregut becomes further refined as indicated by the restricted expression of several genes (see Figure 40.1). Nkx2.1 is expressed in two domains, and this marks the presumptive thyroid and lung. Pax1, Pax9, Tbx1, Hoxa3, and Pax8 are expressed in the presumptive pharyngeal domains, and the Hoxd genes mark the posterior pharyngeal endoderm (discussed later). By embryonic day 10.5 in the mouse, organ primordia for the thyroid, the lungs, the pancreas (both ventral and dorsal), and the liver begin to bud from these patterned domains. Subsequent stages of endoderm organ development are discussed in more detail in other chapters of this book. A. Anterior–Posterior Patterning by Hox Genes One mechanism by which ParaHox transcription factors are believed to control patterning is through the regulation of the Hox genes. Hox genes are key transcriptional regulators of embryonic patterning in chordates, and they are well known for their role in patterning mesoderm and ectoderm. Hox genes are also expressed in defined domains along the AP axis of the gut (Figure 40.4), and emerging evidence suggests that they are involved in the development of the primitive gut and its derivatives (Grapin-Botton, 2005). In the mouse, the Hox cluster consists of 39 genes in four linkage groups (A, B, C, and D), with 13 paralogous families (see Chapter 9). In both the mouse and the chick, the expression domains of Hox genes have been identified along the AP axis of the developing fetal gut (Grapin-Botton, 2005). In the pharyngeal region, Hoxa2, Hoxa3, and Hoxa4 have overlapping patterns of expression with different anterior expression limits, thus defining the identity of the pharyngeal arches (Graham and Smith, 2001). Hoxc5 is expressed in a broad domain between the caudal stomach and the anterior intestine whereas Hoxc9, Hoxb8, and Hoxb9 are expressed in the most posterior portion of the fetal gut. Hoxb6 and Hoxc6 are expressed in restricted domains of the small and large intestines. In several cases, the AP limits of Hox gene expression precisely correlate with endoderm organ boundaries (e.g., between the presumptive stomach and the duodenum and between the small and large intestines; see Figure 40.4). It is clear that the Hox genes have the ability to regulate the AP patterning of the gut. For example, the transgenic misexpression of the Hox3.1 (Hoxc8) gene more anteriorly results in profound gastrointestinal tissue malformations and the loss of positional identity (Grapin-Botton, 2005). In the pharyngeal region, the deletion of Hoxa3 results in the loss of the thymus and of the parathyroid and hypoplasia of the thyroid (Manley and Capecchi, 1995, 1998). Similar gene targeting studies have demonstrated that Hoxc4 mutants have esophageal defects, that Hoxa5 is necessary for the development of the stomach, and that Hoxa13/Hoxd13 compound mutants have hindgut defects (reviewed by Grapin-Botton, 2005). Of these gut-related defects linked to Hox genes, only Hoxa13 has been shown to specifically function in the endoderm. Novel Hoxa13 mutations have been associated with hand–foot– genital syndrome, a rare dominantly inherited condition in humans. Consistent with this finding is the fact that the expression of this mutant form of Hoxa13 in the endoderm of chick embryos results in hindgut and genitourinary patterning defects (de Santa Barbara and Roberts, 2002). Other Hox

SIGNALING PROCESSES INVOLVED IN THE PATTERNING AND FORMATION OF THE FETAL GUT

917

mutations that affect gut development (Hoxa3, Hoxc4, and Hoxa5) are the result of primary defects in gut mesenchyme or neural crest cells that contribute to the developing gut. This highlights the importance of signaling between endoderm and neighboring mesoderm and ectoderm for the proper patterning and development of the gut.

III. SIGNALING PROCESSES INVOLVED IN THE PATTERNING AND FORMATION OF THE FETAL GUT Despite the cell lineage and molecular evidence suggesting that endoderm cells have positional identity by the end of gastrulation, these cells are not yet committed to specific lineages along the AP axis. In fact, there is increasing evidence that a continuum of signals after the gastrulation stage progressively restrict endoderm cell fate into specific organ lineages. Efforts to understand what regulates these important events have led to the identification of several candidate factors. These include soluble growth factors, such as fibroblast growth factors (FGFs), Wnts, Hedgehogs, and retinoic acid (RA). These signaling pathways regulate the expression of transcription factors, including the ParaHox and Hox genes, which have been discussed previously and which are important mediators of cell fate. The gastrula stage endoderm is in close proximity with both the mesoderm and the ectoderm, and it has been demonstrated in several species that soluble signals from these adjacent germ layers can influence the AP fate of the endoderm (Horb and Slack, 2001; Le Douarin, 1968; 1988; Wells and Melton, 2000). Experiments in the mouse and the frog demonstrated that endoderm cocultured with different AP populations of mesoderm will adopt the AP character of the cocultured mesoderm (Figure 40.5; Horb and Slack, 2001; Wells and Melton, 2000). Consistent with this is the fact that grafting posterior mouse endoderm, which would normally become hindgut, into the presumptive pancreatic domain of a late gastrula chick embryo causes that endoderm to be respecified to express Pdx1 (J.M. Wells, unpublished data). These data suggest that endoderm cells at the gastrula stage have not acquired an AP identity but rather that they retain a high degree of plasticity. A. Signals That Pattern the Gastrula Stage Endoderm The temporal and spatial expression pattern of several peptide growth factors is consistent with a role in AP patterning. In particular, the primitive streak, which is a structure that defines the posterior of chick and mouse embryos, expresses several FGF and Wnt ligands, including FGF2, FGF3, FGF4, FGF5, FGF8, Wnt3, Wnt5, and Wnt11. All of these are expressed in the posterior and thus could function in the AP patterning of endoderm. Although the involvement of most of these factors in endoderm patterning has not been established, FGF signaling has been shown to influence the early stages of endoderm patterning as well as to help maintain AP domains in the developing gut tube (Dessimoz et al., 2006; Wells and Melton, 2000). The FGF family of growth factors is comprised of 23 genes in mammals, and it is known to regulate cellular growth and differentiation processes throughout embryonic development and in adult tissues (Dailey et al., 2005). Of the FGF ligands that are expressed in the late gastrula embryo, only

918

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

Dissection

Endoderm gene expression

Recombination

IFABP SS β tubulin

Aen

IFABP SS β tubulin Pen

Aen

IFABP

Mesoderm/ ectoderm

Endoderm

Pen

SS β tubulin Aen

FIGURE 40.5 Using embryonic explant assays to investigate soluble signals that pattern the endoderm. Embryonic day 7.5 mouse embryos can be isolated and the germ layers dissected and cultured. The left panel schematically shows the dissection of embryos and the separation of the endoderm, mesoderm, and ectoderm germ layers. These layers can be further dissected into anterior or posterior endoderm. The middle panel illustrates the recombination of anterior endoderm or posterior endoderm with mesoderm/ectoderm. The right panel shows a reverse transcriptase–polymerase chain reaction analysis of several posterior endoderm markers, including intestinal fatty acid binding protein and somatostatin. Anterior endoderm does not normally express posterior markers when it is cultured with anterior mesoderm/ectoderm. However, when it is cultured with posterior mesoderm/ectoderm, it adopts a posterior fate as indicated by the de novo expression of intestinal fatty acid binding protein and somatostatin. Posterior endoderm will also adopt an anterior fate when it is cultured with anterior mesoderm/ectoderm (not shown). (Data adapted from Wells and Melton, 2000.)

FGF4 (eFGF in frog) has been implicated in endoderm patterning (Dessimoz et al., 2006; Wells and Melton, 2000). FGF4 is expressed in the posterior mesoderm and ectoderm adjacent to the presumptive midgut and hindgut endoderm. At this stage, receptors for FGF4 and the FGF target genes Sprouty1 and Sprouty2 are expressed in the endoderm, which suggests that the endoderm is receiving an FGF signal (Dessimoz et al., 2006). In the mouse, embryos lacking either the FGF4 or the FGF receptor 1 (FGFR1) gene have a similar phenotype and arrest at gastrulation (Ciruna et al., 1997; Yamaguchi et al., 1994). These findings suggest that FGF4 acts via FGFR1 during gastrulation and that FGF4 has unique activity that is not compensated for by other FGF ligands. As a result of early embryonic lethality, the roles of FGF4 and FGFR1 in endoderm patterning have not been identified using mouse genetics. However, using mouse endoderm explants and chick embryology, several studies have demonstrated the involvement of FGF4-mediated signaling in endoderm patterning (Dessimoz et al., 2006; Wells and Melton, 2000). In culture and in embryos, FGF4 protein (but not other FGF ligands) represses anterior cell fate and promotes posterior midgut and hindgut cell fates. The FGF4-mediated inhibition of anterior cell identity corresponds with disrupted foregut

919

SIGNALING PROCESSES INVOLVED IN THE PATTERNING AND FORMATION OF THE FETAL GUT

morphogenesis, which suggests that a direct link exists between patterning and morphogenesis. Inhibiting FGF signaling in vivo results in the loss of Pdx1 and Cdx expression and in the posterior expansion of Hex1 expression. This suggests that FGF signaling is necessary for promoting posterior cell fate and inhibiting anterior cell fate (see Figure 40.6). Consistent with this is the fact that a Cdx homolog in Xenopus, xCad3, is a posterior determinant and a direct FGF target (Haremaki et al., 2003). It is also known that different levels of FGF signaling can promote different cell fates along the AP axis (i. e., high doses of FGF4 protein induce hindgut gene expression, whereas lower doses induce midgut gene expression; Wells and Melton, 2000). Taken together, these studies suggest that FGF signaling represses foregut and promotes midgut and hindgut cell fate in a dose-dependent manner. FGF8 regulates posterior expression of FGF4 in the gastrula embryo and FGF4 was shown to be a direct target of the Wnt signaling pathway in the developing tooth (Kratochwil et al., 2002). Several Wnt ligands are expressed in the primitive streak, suggesting that Wnt signaling could regulate FGF ligand expression to promote posterior fate directly via transcription factors such as Cdx (see Figure 40.6).

Mesoderm

Late gastrula endoderm

Primitive gut tube

? Wnt FGF4 Anterior

Hex1

Posterior

Cdxs

Sox2 Nkx2.1 Hex1

IFABP Pdx1

Cdxs SS

FIGURE 40.6 Signals from the mesoderm at the gastrula stage are necessary for establishing anterior–posterior patterning during early gut tube development. The upper panel schematically shows the endoderm (light gray) and the adjacent mesoderm (drak gray). Several fibroblast growth factor (FGF) and Wnt ligands are expressed in the posterior (primitive streak) in the late-gastrula embryo. FGF4 has been shown to promote posterior gene expression in a dosedependent manner, with the highest levels promoting the most posterior fates. FGF4 also represses anterior gene (Hex1) expression (at the gastrula stages only). Although it is not shown in the context of endoderm patterning, it is known in other contexts that FGF4 is a direct target of the canonical Wnt signaling pathway (Kratochwil et al., 2002). The lower panel shows several of the genes that are expressed along the anterior–posterior axis of the primitive gut tube. Many of these genes, including Nkx2.1, Hex1, Pdx1, and Cdx, are expressed in restricted domains that predict where organ primordia will form. Alterations in FGF4-mediated signaling disrupt early endoderm patterning, and they cause morphologic and gene expression changes in the developing foregut, midgut, and hindgut domains (Dessimoz et al., 2006).

920

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

B. Establishing Presumptive Organ Domains in the Primitive Gut Tube 1. Foregut Patterning The foregut endoderm contributes to several organs, including the thyroid, the lungs, the liver, and the pancreas. The dynamic nature of foregut morphogenesis brings the endoderm into proximity with several mesodermally derived tissues that are now known to pattern the foregut. During the early stages of foregut morphogenesis (embryonic day 8 in the mouse; HH stage 6 in the chick), the ventral foregut endoderm is adjacent to the cardiac mesoderm. During the later stages (embryonic day 8.5 in the mouse; HH stage 9 in the chick), the endoderm comes in contact with the septum transversum mesenchyme and the endothelial cells of the ventral veins. The dorsal foregut endoderm is brought into contact with the notochord and the dorsal aorta. Using a combination of tissue recombination, embryology, and genetics, specific signaling molecules emanating from these tissues have been shown to pattern the foregut. The liver derives exclusively from the ventral foregut, and it is the first foregut derivative to form. Studies in the chick and the mouse suggest that the initial specification of the liver depends on signals from adjacent cardiac mesoderm (reviewed by Grapin-Botton, 2005). In mouse ventral foregut explant cultures, it was shown that the cardiac mesoderm signals to the ventral foregut to induce liver and that this occurs between the three- and eight-somite stages. Interestingly, in more posterior regions of the gut, negative signals were necessary to prevent precocious liver albumin gene expression in nonhepatic trunk endoderm. When trunk endoderm was removed from these negative signals, it expressed the liver marker albumin. Although the direct targets of the cardiac signal have not been identified, there are several transcription factors that are expressed at high levels in prehepatic ventral endoderm, including Hex1, Gata factors, Foxa1, and Foxa2, all of which are known to play roles in liver development (McLin and Zorn, 2006). Foregut explant cultures have been used to identify specific signaling molecules produced by cardiac mesoderm, particularly FGF1 (aFGF) and FGF2 (bFGF) proteins, which pattern the ventral foregut into liver, pancreas, and lung domains (Deutsch et al., 2001; Gualdi et al., 1996; Serls et al., 2005). From these studies, a model has been proposed whereby different concentration thresholds of FGF ligands pattern the ventral foregut into these different lineages, with high levels promoting liver, moderate levels promoting lung, and low levels promoting ventral pancreas. It is also possible that the length of time that an endoderm cell is exposed to the cardiac mesoderm influences its fate. For example, cell lineage experiments show that ventral foregut endoderm cells adjacent to cardiac mesoderm end up distributed along the AP length of the foregut at later stages (Tremblay and Zaret, 2005). Therefore, it is possible that endoderm cells that remain in proximity to cardiac mesoderm longer adopt a liver fate, that cells that migrate posteriorly away from the cardiac mesoderm become ventral pancreas, and that cells that migrate anteriorly become lung (Deutsch et al., 2001; Serls et al., 2005). RA is a key regulator of embryonic patterning and differentiation. Several reports in frogs, fish, chicks, and mice have implicated RA signaling in the development of foregut derivatives including in the lung, the liver, and the pancreas. RA signaling is shown to be important for the global AP patterning of the foregut endoderm and thus that it affects the development of

SIGNALING PROCESSES INVOLVED IN THE PATTERNING AND FORMATION OF THE FETAL GUT

921

multiple organs. For example, it has been shown that RA treatment is sufficient to cause an anterior shift of Pdx1 expression in embryonic chick explants of lateral endoderm plus mesoderm (Kumar et al., 2003). However, in this example, the presence of mesenchyme was required, which suggests that the effect on endoderm was indirect. Other growth factors had similar posteriorizing activity, including bone morphogenetic proteins (BMPs) and activin. Thus, it is possible that RA regulates the mesenchymal expression of these factors, which then signal to endoderm. A global foregut patterning role of RA was also suggested in the mouse, because embryos deficient in retinaldehyde dehydrogenase 2 (Raldh2) and lacking active RA signaling in the foregut region fail to develop the lungs, the stomach, and the dorsal pancreas, and they have impaired liver growth (Wang et al., 2006). RA signaling in the foregut was shown to act upstream of Pdx1, FGF10, and Hox genes, which suggests that it is globally affecting the early patterning of the foregut. There are reports in zebrafish, Xenopus, and the mouse suggesting that RA signaling plays patterning-independent roles in early pancreas development. Studies in Xenopus suggest that activation of the RA pathway does not effect the AP expression of Pdx1 (Xlhbox8) as it does in the chick (Zeynali and Dixon, 1998); however, it may affect later aspects of pancreas development (Chen et al., 2004). The pancreas develops from the dorsal and ventral foregut endoderm, and, interestingly, mouse embryos deficient for the RA-synthesizing enzyme Raldh2 fail to initiate dorsal pancreas development, but ventral pancreas development initiates normally (Martin et al., 2005). The development of the pancreatic mesenchyme is also deficient in these animals, which suggests that pancreatic defects may be the result of a loss of a mesenchymal signal. In zebrafish, however, the disruption of RA receptor function, specifically in the endoderm, perturbs normal pancreas development; this suggests a cell-autonomous role for RA signaling in the endoderm (Stafford et al., 2006). The apparent discrepancies involving the role of RA signaling in the development of various foregut derivatives may be the result of species differences, or they could suggest that RA signaling has multiple roles during the early stages of foregut patterning and the early stages of organ development. The role of Hedgehog signaling in pancreas development is more extensively discussed elsewhere in this book (see Chapter 42). However, there is evidence to suggest that a primary role of Hedgehog signaling is patterning the gut tube. During the early stages of gut tube development (embryonic days 8.5–9 in the mouse), Sonic hedgehog (Shh) is broadly expressed by the endoderm along the AP axis. As the thyroid, thymus, parathyroid, and pancreas organ rudiments begin to form, Shh expression is repressed in these budding organs. In the case of the pancreatic bud, the repression of Shh depends on signals from the notochord, which include FGF2 and activin (reviewed by Kim and Hebrok, 2001). Surprisingly, in transgenic mice in which Shh was misexpressed in the Pdx1 domain, the initiation of pancreas development and the formation of the endocrine and exocrine lineages were relatively normal. However, the adjacent mesenchyme underwent a homeotic transformation toward an intestinal fate, forming intestinal smooth muscle and interstitial cells of Cajal. Gene targeting studies in mice also show that Shh mutants display intestinal transformations of the stomach epithelium as evidenced by the expression of duodenal markers in the stomach (Ramalho-Santos et al., 2000).

922

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

Studies in the chick have also suggested that Shh is involved in the patterning of the foregut and in the establishment of the gizzard/stomach organ boundaries. This study suggests that Shh from the endoderm regulates the expression of Nkx2.5 in the gizzard and that ectopic Shh expression causes inappropriate Nkx2.5 expression in the pancreas. Shh did not directly induce Nkx2.5; rather, it induced the expression of BMP4 in adjacent mesenchyme. The cell-autonomous activation of BMP signaling in the mesenchyme through the expression of activated BMP receptors was sufficient to cause the upregulation of Nkx2.5 in the gizzard, and it resulted in perturbed smooth muscle differentiation.

C. Patterning the Anterior Foregut/Pharyngeal Domain The anterior portion of the foregut is commonly referred to as the branchial or pharyngeal region. This region of the vertebrate embryo is anatomically hallmarked by a series of bilateral bulges called arches, which decrease in size from anterior to posterior (Figure 40.7). The pharyngeal endoderm lines the interior of the pharyngeal arches, and it will transiently fuse with the surface ectoderm to form pouches (interior endoderm) and clefts (exterior ectoderm) that mark the anterior and posterior boundaries of the arches. Pharyngeal pouches form from anterior to posterior, and they exhibit regionally restricted identity as marked by the differential expression of FGF8 and paired box 1 (Pax1). During normal development, the pharyngeal endoderm will contribute to the trachea, the esophagus, the thymus, the parathyroid, the thyroid, the ultimobranchial bodies (which form the calcitonin-producing parafollicular cells of the thyroid), and the taste buds. However, multiple human disorders are associated with abnormal pharyngeal development, including esophageal atresia, tracheoesophageal fistula, VACTERL, and the DiGeorge and velocardiofacial syndromes, and this emphasizes the clinical importance of understanding anterior foregut patterning. Furthermore, from a teratologic standpoint, pharyngeal region morphology appears to be particularly sensitive to alcohol exposure, possibly as a result of the competitive interactions of alcohol with RA-metabolizing enzymes present in pharyngeal tissues (Wang, 2005). One key role of pharyngeal endoderm is as a regional source of patterning signals for the adjacent mesoderm and ectoderm (reviewed by Graham and Smith, 2001). Signaling molecules expressed by endoderm include Shh, multiple members of the FGF family, BMPs, the BMP antagonists chordin and noggin, and several RA receptors. Although it is not known exactly how these pathways intersect to pattern the pharyngeal region, we will later discuss a potential patterning model that is based on the analysis of mouse mutants and gene expression studies (see Figure 40.7). Both FGF3 and FGF8 are expressed in the pharyngeal pouch endoderm, and both are required for proper pouch formation (Crump et al., 2004). Mutants lacking FGF8 and FGFR1 have severe phenotypes that exhibit deletions of both the third and fourth arches and that result in the loss of the thymus and the parathyroid (Abu-Issa et al., 2002; Trokovic et al., 2003). FGF8 expression in the pharyngeal endoderm is lost in embryos that lack T-box 1 (Tbx1; see Chapter 16). Which suggests that Tbx1 is upstream of FGF8. Furthermore, Tbx1 mutations have been shown to contribute to

SIGNALING PROCESSES INVOLVED IN THE PATTERNING AND FORMATION OF THE FETAL GUT

A

B

Wild type E10.5 pharyngeal region

1

Arch

C

Fgf8, Tbx1, FgfR1, chordin, RARα; RARβ & Raldh2 mutants

D Shh-/-

923 Potential pharyngeal patterning model

(Ptc1) Shh signaling

1

Pouch Cleft

1st and 2nd 2

2

2

3 Aortic arches

3

4 6

4 E10.5

3rd and 4th RA signaling (RARE-LacZ)

FIGURE 40.7 Patterning the anterior foregut/pharyngeal domain. A, Ventral/coronal view of normal pharyngeal arches. The outline of an embryonic day 10.5 embryo shows the corresponding plane of section (dotted line) of the larger image. The endoderm is in yellow, the surface ectoderm is blue, and the mesenchyme of the arches is red. Arches are numbered 1 through 4 and then 6, from anterior to posterior. The pouches and clefts are numbered by that of the preceding arch. B, The embryonic day 10.5 pharyngeal phenotype of the FGF8, Tbx1, Fgfr1, chordin, RARa/RARb, and RALDH2 mutant mice in which the caudal third and fourth arches are missing. C, The embryonic day 10.5 pharyngeal phenotype of the Shh null mutant mice in which the first arch is missing and the second is severely hypomorphic. (Note: The Shh mutant embryos initiate the development of the first and second arches on embryonic day 9.5, but they lose the first arch by embryonic day 10.5, and the second arch continues to atrophy.) D, A putative model for the anterior–posterior patterning of the pharyngeal region, with Shh- and RA-mediated signals regulating the anterior and posterior regions of the pharyngeal region.

DiGeorge syndrome in humans (Jerome and Papaioannou, 2001). Patients with DiGeorge syndrome often have numerous defects in craniofacial and cardiovascular development and defects in organs that are derived from the pharyngeal pouch endoderm, including the thymus and the parathyroid. Considering the importance of Tbx1 in human disease, there is much interest in the signaling pathways that regulate Tbx1 expression, and several reports indicate that Tbx1 may be regulated by the morphogen Shh (Riccomagno et al., 2002; Yamagishi et al., 2003). Tbx1 expression in the pharyngeal region is lower in Shh-/- mutant mouse embryos, a Tbx1-promoter–driven LacZ transgene showed reduced expression in an Shh-/- background, and exposing pharyngeal mesenchyme to exogenous Shh protein resulted in the ectopic expression of Tbx1. However, several phenotypic differences between Tbx1 and Shh mutants suggest that the Hedgehog pathway is more critical for the development of the anterior pharyngeal region, whereas the Tbx1 phenotypes are predominantly in the caudal pharyngeal arches. For example, Shh mutants form a thymus (unlike the Tbx1 null embryos), whereas parathyroid development is perturbed in both mutants (Ivins et al., 2005; Moore-Scott and Manley, 2005). This phenotypic overlap suggests that Shh may act as a modifier of Tbx1 expression in the pharyngeal region. In addition to its role in global endoderm patterning, the RA pathway, which is mediated by the retinoid receptors (RARs) RARa, RARb, and RARg, is essential for the patterning of the pharyngeal endoderm. RA has been shown to regulate Shh and Tbx1 expression as well as the expression of several Hox genes, all of which are essential to the AP patterning of the pharyngeal endoderm (reviewed by Graham and Smith, 2001). In the case of Hoxb1 gene, an RA-inducible enhancer has been identified (RAIDR5; Graham and Smith, 2001; Grapin-Botton, 2005). This element drives broad

924

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

expression of Hoxb1 throughout the foregut on embryonic day 8.5, but it becomes restricted to the pharyngeal endoderm of the anterior foregut on embryonic day 9.5. In addition, exposing embryos to exogenous RA causes an anterior expansion of Hoxb1, suggesting that RA signaling regulates foregut pattern by directly regulating the transcription of Hoxb1. In Tbx1 mutants, Raldh2 expression expands anteriorly, which suggests that RA signaling is regulated by Tbx1 (Ivins et al., 2005). Embryos exposed to a pan-RAR antagonist, a pan-RAR agonist, vitamin A, or RA lose the caudal third and fourth pharyngeal arches. These treatments phenocopy those observed in the compound RARa and RARb knockouts and the hypomorphic Raldh2 mutant, and they illustrate the importance of proper RA signaling in pharyngeal region patterning (Niederreither et al., 2003; Vermot et al., 2003). BMPs are members of the transforming growth factor b superfamily of growth factors, and they have also been implicated in patterning and pharyngeal organogenesis. BMP2, BMP4, and BMP7 as well as the BMP antagonists noggin and chordin are expressed in the pharyngeal region in all three cell types, including the endoderm. The chordin null, like the FGF8 and Tbx1 null and the compound RARa/RARb mutants, fails to form the posterior arches, which implies that the inhibition of BMP signaling is part of a complex pathway that is necessary for the posterior patterning of the pharyngeal endoderm (Bachiller et al., 2003). These data establish that Tbx1, FGF8, RA, and chordin are all required for the formation of the posterior pharyngeal region, whereas the anterior mutant phenotype in the Shh mutants suggest an oppositional role for Hedgehog signaling in the AP patterning of the pharyngeal endoderm (see Figure 40.7, B and C; Graham and Smith, 2001). Moreover, Patched1 (Ptc1) is an Shh transcriptional target that is expressed in the first and second arch, whereas an RA-responsive-element–driven LacZ transgene is expressed in the third and fourth arches (see Figure 40.7, D), which supports the idea of Hedgehog signaling in the anterior arches and RA signaling in the posterior arches. It is interesting that none of the specific components of the Hedgehog, FGF, or RA signaling pathways are expressed in restricted pharyngeal domains along the AP axis. In fact, many are expressed in discrete domains within the endoderm or throughout the entirety of the pharyngeal endoderm, mesoderm, and ectoderm. It is not known how the activation of the Hedgehog and RA pathways is restricted to specific AP arches. Clearly, a complex interaction of the pathways occurs that has yet to be fully defined and thus requires further study. D. Midgut and Hindgut Patterning Early during the development of the gut tube, the posterior endoderm folds ventrally generating the hindgut. The midgut and hindgut form the small and large intestines and the hindgut also contributes to the cloaca, which forms the urogenital tract. Studies have begun to identify the molecular mechanisms that regulate posterior identity in the hindgut. As discussed previously, FGF signaling before gut tube morphogenesis is involved in the establishment of posterior endoderm identity (Dessimoz et al., 2006). These studies also found that FGF signaling is required to maintain the AP expression boundaries of Pdx1 and Cdx genes at later stages of gut tube development. Shifting FGF4-mediated signaling to the anterior caused an anterior shift in

IMPACT OF GENOMICS ON OUR UNDERSTANDING OF EARLY ENDODERM ORGANOGENESIS

925

the expression of Pdx1 and Cdx genes in the primitive gut tube, which suggests that anterior regions of the gut tube were transformed to a posterior fate. Conversely, inhibiting FGF signaling caused a loss of posterior identity. Moreover, these studies demonstrated that FGF signaling was acting directly on the endoderm. Several other signaling molecules, including BMP, activin, and RA, also have the ability to posteriorize the lateral endoderm that contributes to the ventral gut, but only in the presence of lateral plate mesoderm; this suggests that these factors may act indirectly (Kumar et al., 2003). These findings illustrate the fact that gut patterning involves reciprocal signaling between endodermally derived epithelium and mesodermally derived tissues, including the notochord and the gut mesenchyme. Additional reciprocal interactions have been identified between the hindgut endoderm and the mesoderm that result in spatially restricted Hox gene expression in the hindgut and the establishment of posterior identity in the endoderm (reviewed by Grapin-Botton, 2005; Wells and Melton, 1999). Specifically, hindgut endoderm expresses Shh, which is sufficient to induce BMP4 and Hoxd13 expression in adjacent posterior mesoderm but not in the more anterior mesoderm adjacent to the stomach. If Hoxd13 is misexpressed in the more anterior mesoderm, the adjacent stomach endoderm is transformed into an intestinal type of endoderm as assayed by morphology and marker expression. The molecules that transmit the signal from Hoxd13-expressing mesenchyme to endoderm are unidentified. The role of canonical Wnt signaling in regulating the homeostasis of intestinal epithelium is well established. Recently, several reports have implicated this pathway in the embryonic development and patterning of the gastrointestinal tract. Numerous Wnt signaling components are expressed along the AP axis of the gastrointestinal tract throughout its development (Theodosiou and Tabin, 2003). Functional evidence implicating Wnt signaling in gut development has come from studies of T cell factor/Lymphoid enhancer factor 1 (TCF/LEF) family of transcription factors, which are downstream effectors of the canonical Wnt pathway. In the chick, the expression of dominant-negative TCF4 in the mesenchyme caused secondary defects in the differentiation of gizzard epithelium. In the mouse, TCF1/TCF4-/- embryos have defects in hindgut expansion and an anterior transformation of the duodenum (Gregorieff et al., 2004). Although these studies implicate canonical Wnt signaling in gut tube patterning, it is not known whether TCF factors act in the endoderm, the mesoderm, or both.

IV. IMPACT OF GENOMICS ON OUR UNDERSTANDING OF EARLY ENDODERM ORGANOGENESIS Current technology and database organization make it possible to generate a quantitative catalog of all of the genes expressed at each step of organ development, from the specification of embryonic endoderm to the formation of functioning adult cell types such as insulin-producing b cells (see also the chapter by Gannon in this book). These databases allow researchers to analyze the cell-type–specific expression of a particular gene or to look for entire functional classes of genes, such as DNA-binding factors expressed in the developing or adult cell types. More importantly, these types of analyses

926

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

should provide a molecular foundation for future studies of organ development and information about how to direct the differentiation of stem cells into therapeutically important cell types, such as pancreatic b cells. Recently, Affymetrix microarrays were used to perform a quantitative gene-expression analysis of highly purified cells isolated from four key stages of endocrine pancreas development (Gu et al., 2004). The stages and cells analyzed in this experiment were as follows: (1) embryonic day 7.5 prepancreatic endoderm; (2) embryonic day 10.5 pancreatic progenitor cells as defined by Pdx1 expression; (3) embryonic day 13.5 endocrine progenitor cells as defined by Ngn3 expression; and (4) adult islets of Langerhans. This approach was highly successful for generating a gene expression database of the developing endocrine pancreas. More importantly, this study identified many novel genes expressed at each stage of endocrine pancreas development, and the functional analysis of one of these genes demonstrated its involvement in endocrine cell development. For the purpose of this chapter, we will focus on the stages of gut development during which the patterning of endoderm results in the establishment of the Pdx1 domain. The expression of more than 12,000 genes was measured simultaneously to identify the genes induced as endoderm is specified toward the pancreatic lineage. Genes that showed a greater than three-fold expression change during this endoderm patterning stage were grouped and analyzed. For simplicity, we will focus on two classes of regulatory molecules that control cell fate: transcription factors and growth factors/receptors. FGF, notch, and activin signaling pathways have been implicated in various aspects of early pancreas specification (reviewed by Kim and Hebrok, 2001), and DNA microarray expression data are consistent with these findings showing that Pdx1 cells express FGFR1, FGFR2, and FGFR4 as well as activin receptor IIb (data not shown), notch 3, and deltalike 1. These studies also identified a number of components of the Wnt signaling pathway that were expressed in the early pancreas, including Wnt receptors, frizzled2, frizzled4, and secreted forms of frizzled that antagonize Wnt signaling. The identification of Wnt signaling components in the early pancreas prompted several groups to perform functional studies of this pathway that have demonstrated its importance in pancreas development (Dessimoz et al., 2005; Murtaugh et al., 2005; Wells et al., 2007). These signals culminate in the expression of several transcription factors in the pancreatic endoderm, including Pdx1, Hlxb1, and Ptf1a/p48, all of which are necessary for proper pancreas development (reviewed by Grapin-Botton, 2005). Our analysis detected these genes in early pancreatic cells (data not shown), and we found numerous other transcription factors, most of which were not previously described in the developing pancreas. These transcription factors include the basic helix–loop–helix type (Id, Foxa2, and Foxa3) and the homeobox type (Barx1 and Hoxa5) as well as other classes. The function of most of these factors in pancreas development has not been determined.

V. CLINICAL APPLICATIONS FOR STUDIES IN ENDODERM ORGANOGENESIS A significant number of diseases affect endodermally derived organs, particularly the lungs, the liver, and the pancreas. Moreover, childhood diseases such as asthma and diabetes, which affect the respiratory and gastrointestinal tracts, are increasing at an alarming rate. The study of the molecular mechanisms

ACKNOWLEDGMENTS

927

underlying endoderm organogenesis is a relatively new field. Nonetheless, experiments in model organisms have already led to the identification of genes involved in diseases of the endoderm. The exciting new fields of regenerative medicine and tissue engineering have greatly benefited from research in developmental biology. For example, it has become increasingly apparent that some adult organs reactivate embryonic pathways during the process of regeneration. Injury to the lung and tracheal epithelium causes a regenerative response that is accompanied by a dramatic increase in the expression of embryonic endoderm regulatory genes, including Sox17, Foxa1, Foxa2, and others, and some of these genes may aid in regeneration (Park et al., 2006a, 2006b). Another means of translating embryonic studies into clinical applications is through the directed differentiation of embryonic stem cells into endoderm derivatives. It has recently been shown that the embryonic growth factor activin can direct the differentiation of human embryonic stem cells into endoderm in culture (D’Amour et al., 2005). This exciting advance will now allow researchers to study how human embryonic stem cell–derived endoderm is patterned and ultimately directed into specific organ lineages. A renewable source of endodermally derived tissues would have a profound impact on therapeutic approaches involving transplantation-based therapies.

SUMMARY

 The endoderm germ layer in the late-gastrula embryo (embryonic day 7.5

 

   

in the mouse; HH stages 4 and 5 in the chick) is a single layer of cells that are unspecified. Within 2 days of development (embryonic day 9.5 in the mouse; HH stage 18 in the chick), the endoderm has formed a primitive gut tube with budding organ primordia. The late-gastrula endoderm is regionalized into axial and lateral domains along the AP axis that give rise to the foregut, the midgut, and the hindgut. By the early somite stage (embryonic day 8.5 in the mouse; HH stage 10 in the chick), the developing gut tube has become highly patterned at the molecular level. Genes are expressed in overlapping and distinct domains that predict where organ buds will form. Several signaling pathways are involved in these early patterning mechanisms, including the FGF, Hedgehog, BMP, RA, and Wnt pathways. The disruption of these early patterning mechanisms can directly lead to defects in the development of endoderm organs, including the thymus, the parathyroid, the lungs, the liver, the pancreas, and the intestines. Genomics strategies are being used to identify new molecular pathways involved in endoderm organogenesis. Information from endoderm organ development is being used to differentiate human embryonic stem cells into endoderm organ cell types that ultimately could be used to treat degenerative diseases such as diabetes.

ACKNOWLEDGMENTS BMS is supported by a postdoctoral fellowship from the Juvenile Diabetes Research Foundation (JDRF HD42572). JMW receives support from the National Institutes of Health (GM072915), the Juvenile Diabetes Research Foundation (2–2003–530), and the Beta Cell Biology Consortium (BCBC 31148-R).

928

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

GLOSSARY Endoderm The inner layer of cells in embryonic development that gives rise to digestive and respiratory organs. Note that, early in the development of many vertebrate embryos, the endoderm is initially the outermost layer of cells. Foregut The anteriormost of the three divisions of the digestive tract (the foregut, the midgut, and the hindgut). Hindgut The posteriormost region of the digestive tract. Liver The largest gland in the human body, with roles in digestion, glucose regulation and storage, blood clotting, and the removal of wastes from the blood. Lung An organ that contains sac-like structures in which blood and air exchange oxygen and carbon dioxide. Organogenesis The development of organs during embryonic development. Pancreas A gland in the abdominal cavity with both exocrine and endocrine function that secretes digestive enzymes into the duodenum and that also secretes the hormones insulin and glucagon into the blood. Parathyroid A calcium homeostatic organ that secretes parathyroid hormone. Patterning The act of subdividing embryos or tissues into distinct domains along the embryonic axes. Primordia Cells, tissues, or organs at the earliest stage of development. Thyroid An endocrine organ with the functions of regulating metabolism, growth, and development. Ultimobranchial bodies Small glands that develop separately from the thyroid that will fuse with the thyroid and form the parafollicular cells of the thyroid.

REFERENCES Abu-Issa R, Smyth G, Smoak I, et al: Fgf8 is required for pharyngeal arch and cardiovascular development in the mouse, Development 129:4613–4625, 2002. Ang SL, Wierda A, Wong D, et al: The formation and maintenance of the definitive endoderm lineage in the mouse: involvement of HNF3/forkhead proteins, Development 119:1301–1315, 1993.

REFERENCES

929 Bachiller D, Klingensmith J, Shneyder N, et al: The role of chordin/Bmp signals in mammalian pharyngeal development and DiGeorge syndrome, Development 130:3567–3578, 2003. Chalmers AD, Slack JM: The Xenopus tadpole gut: fate maps and morphogenetic movements, Development 127:381–392, 2000. Chalmers AD, Slack JM, Beck CW: Regional gene expression in the epithelia of the Xenopus tadpole gut, Mech Dev 96:125–128, 2000. Chapman SC, Schubert FR, Schoenwolf GC, Lumsden A: Analysis of spatial and temporal gene expression patterns in blastula and gastrula stage chick embryos, Dev Biol 245:187–199, 2002. Chawengsaksophak K, de Graaff W, Rossant J, et al: Cdx2 is essential for axial elongation in mouse development, Proc Natl Acad Sci U S A 101:7641–7645, 2004. Chen Y, Pan FC, Brandes N, et al: Retinoic acid signaling is essential for pancreas development and promotes endocrine at the expense of exocrine cell differentiation in Xenopus, Dev Biol 271:144–160, 2004. Ciruna BG, Schwartz L, Harpal K, et al: Chimeric analysis of fibroblast growth factor receptor-1 (Fgfr1) function: a role for FGFR1 in morphogenetic movement through the primitive streak, Development 124:2829–2841, 1997. Crump JG, Maves L, Lawson ND, et al: An essential role for Fgfs in endodermal pouch formation influences later craniofacial skeletal patterning, Development 131:5703–5716, 2004. D’Amour KA, Agulnick AD, Eliazer S, et al: Efficient differentiation of human embryonic stem cells to definitive endoderm, Nat Biotechnol 23:1534–1541, 2005. Dailey L, Ambrosetti D, Mansukhani A, Basilico C: Mechanisms underlying differential responses to FGF signaling, Cytokine Growth Factor Rev 16:233–247, 2005. de Santa Barbara P, Roberts DJ: Tail gut endoderm and gut/genitourinary/tail development: a new tissue-specific role for Hoxa13, Development 129:551–561, 2002. Dessimoz J, Bonnard C, Huelsken J, Grapin-Botton A: Pancreas-specific deletion of beta-catenin reveals Wnt-dependent and Wnt-independent functions during development, Curr Biol 15:1677–1683, 2005. Dessimoz J, Opoka R, Kordich JJ, et al: FGF signaling is necessary for establishing gut tube domains along the anterior-posterior axis in vivo, Mech Dev 123:42–55, 2006. Deutsch G, Jung J, Zheng M, et al: A bipotential precursor population for pancreas and liver within the embryonic endoderm, Development 128:871–881, 2001. Frumkin A, Haffner R, Shapira E, et al: The chicken CdxA homeobox gene and axial positioning during gastrulation, Development 118:553–562, 1993. Gamer LW, Wright CV: Murine Cdx-4 bears striking similarities to the Drosophila caudal gene in its homeodomain sequence and early expression pattern, Mech Dev 43:71–81, 1993. Gittes GK, Rutter WJ: Onset of cell-specific gene expression in the developing mouse pancreas, Proc Natl Acad Sci U S A 89:1128–1132, 1992. Graham A, Smith A: Patterning the pharyngeal arches, Bioessays 23:54–61, 2001. Grapin-Botton A: Antero-posterior patterning of the vertebrate digestive tract: 40 years after Nicole Le Douarin’s PhD thesis, Int J Dev Biol 49:335–347, 2005. Grapin-Botton A, Majithia AR, Melton DA: Key events of pancreas formation are triggered in gut endoderm by ectopic expression of pancreatic regulatory genes, Genes Dev 15:444–454, 2001. Gregorieff A, Grosschedl R, Clevers H: Hindgut defects and transformation of the gastrointestinal tract in Tcf4(-/-)/Tcf1(-/-) embryos, EMBO J 23:1825–1833, 2004. Gu G, Wells JM, Dombkowski D, et al: Global expression analysis of gene regulatory pathways during endocrine pancreatic development, Development 131:165–179, 2004. Gualdi R, Bossard P, Zheng M, et al: Hepatic specification of the gut endoderm in vitro: cell signaling and transcriptional control, Genes Dev 10:1670–1682, 1996. Haremaki T, Tanaka Y, Hongo I, Okamoto H: Integration of multiple signal transducing pathways on FGF response elements of the Xenopus caudal homologue Xcad3, Development 130:4907–4917, 2003. Horb ME, Slack JM: Endoderm specification and differentiation in Xenopus embryos, Dev Biol 236:330–343, 2001. Ivins S, Lammerts van Beuren K, Roberts C, et al: Microarray analysis detects differentially expressed genes in the pharyngeal region of mice lacking Tbx1, Dev Biol 285:554–569, 2005. Jones CM, Broadbent J, Thomas PQ, et al: An anterior signalling centre in Xenopus revealed by the homeobox gene XHex, Curr Biol 9:946–954, 1999. Jerome LA, Papaioannou VE: DiGeorge syndrome phenotype in mice mutant for the T-box gene, Tbx1, Nat Genet 27:286–291, 2001.

930

PATTERNING THE EMBRYONIC ENDODERM INTO PRESUMPTIVE ORGAN DOMAINS

Kim SK, Hebrok M: Intercellular signals regulating pancreas development and function, Genes Dev 15:111–127, 2001. Kimura W, Yasugi S, Stern CD, Fukuda K: Fate and plasticity of the endoderm in the early chick embryo, Dev Biol 289:283–295, 2006. Kratochwil K, Galceran J, Tontsch S, et al: FGF4, a direct target of LEF1 and Wnt signaling, can rescue the arrest of tooth organogenesis in Lef1(-/-) mice, Genes Dev 16:3173–3185, 2002. Kumar M, Jordan N, Melton D, Grapin-Botton A: Signals from lateral plate mesoderm instruct endoderm toward a pancreatic fate, Dev Biol 259:109–122, 2003. Lawson A, Schoenwolf GC: Epiblast and primitive-streak origins of the endoderm in the gastrulating chick embryo, Development 130:3491–3501, 2003. Lawson KA, Meneses JJ, Pedersen RA: Cell fate and cell lineage in the endoderm of the presomite mouse embryo, studied with an intracellular tracer, Dev Biol 115:325–339, 1986. Lawson KA, Pedersen RA: Cell fate, morphogenetic movement and population kinetics of embryonic endoderm at the time of germ layer formation in the mouse, Development 101:627–652, 1987. Le Douarin N: Synthesis of glycogen in hepatocytes undergoing differentiation: role of homologous and heterologous mesenchyma, Dev Biol 17:101–114, 1968. Le Douarin NM: On the origin of pancreatic endocrine cells, Cell 53:169–171, 1988. Lee CS, Friedman JR, Fulmer JT, Kaestner KH: The initiation of liver development is dependent on Foxa transcription factors, Nature 445:944–947, 2005. Manley NR, Capecchi MR: The role of Hoxa-3 in mouse thymus and thyroid development, Development 121:1989–2003, 1995. Manley NR, Capecchi MR: Hox group 3 paralogs regulate the development and migration of the thymus, thyroid, and parathyroid glands, Dev Biol 195:1–15, 1998. Martin M, Gallego-Llamas J, Ribes V, et al: Dorsal pancreas agenesis in retinoic acid-deficient Raldh2 mutant mice, Dev Biol 284:399–411, 2005. Martinez Barbera JP, Clements M, Thomas P, et al: The homeobox gene Hex is required in definitive endodermal tissues for normal forebrain, liver and thyroid formation, Development 127:2433–2445, 2000. McLin VA, Zorn AM: Molecular control of liver development, Clin Liver Dis 10:1–25, 2006. Moore-Scott BA, Manley NR: Differential expression of Sonic hedgehog along the anteriorposterior axis regulates patterning of pharyngeal pouch endoderm and pharyngeal endodermderived organs, Dev Biol 278:323–335, 2005. Murtaugh LC, Law AC, Dor Y, Melton DA: Beta-catenin is essential for pancreatic acinar but not islet development, Development 132:4663–4674, 2005. Niederreither K, Vermot J, Le Roux I, et al: The regional pattern of retinoic acid synthesis by RALDH2 is essential for the development of posterior pharyngeal arches and the enteric nervous system, Development 130:2525–2534, 2003. Northrop JL, Kimelman D: Dorsal-ventral differences in Xcad-3 expression in response to FGFmediated induction in Xenopus, Dev Biol 161:490–503, 1994. Ohlsson H, Karlsson K, Edlund T: IPF1, a homeodomain-containing transactivator of the insulin gene, EMBO J 12:4251–4259, 1993. Park KS, Wells JM, Zorn AM, et al: Transdifferentiation of ciliated cells during repair of the respiratory epithelium, Am J Respir Cell Mol Biol 34:151–157, 2006a. Park KS, Wells JM, Zorn AM, et al: Sox17 influences the differentiation of respiratory epithelial cells, Dev Biol 294:192–202, 2006b. Ramalho-Santos M, Melton DA, McMahon AP: Hedgehog signals regulate multiple aspects of gastrointestinal development, Development 127:2763–2772, 2000. Riccomagno MM, Martinu L, Mulheisen M, et al: Specification of the mammalian cochlea is dependent on Sonic hedgehog, Genes Dev 16:2365–2378, 2002. Rosenquist GC: The location of the pregut endoderm in the chick embryo at the primitive streak stage as determined by radioautographic mapping, Dev Biol 26:323–335, 1971. Serls AE, Doherty S, Parvatiyar P, et al: Different thresholds of fibroblast growth factors pattern the ventral foregut into liver and lung, Development 132:35–47, 2005. Stafford D, White RJ, Kinkel MD, et al: Retinoids signal directly to zebrafish endoderm to specify insulin-expressing beta-cells, Development 133:949–956, 2006. Tam PP, Khoo PL, Wong N, et al: Regionalization of cell fates and cell movement in the endoderm of the mouse gastrula and the impact of loss of Lhx1(Lim1) function, Dev Biol 274:171–187, 2004. Theodosiou NA, Tabin CJ: Wnt signaling during development of the gastrointestinal tract, Dev Biol 259:258–271, 2003.

RECOMMENDED RESOURCES

931

Tremblay KD, Zaret KS: Distinct populations of endoderm cells converge to generate the embryonic liver bud and ventral foregut tissues, Dev Biol 280:87–99, 2005. Trokovic N, Trokovic R, Mai P, Partanen J: Fgfr1 regulates patterning of the pharyngeal region, Genes Dev 17:141–153, 2003. Vermot J, Niederreither K, Garnier JM, et al: Decreased embryonic retinoic acid synthesis results in a DiGeorge syndrome phenotype in newborn mice, Proc Natl Acad Sci U S A 100:1763–1768, 2003. Wang XD: Alcohol, vitamin A, and cancer, Alcohol 35:251–258, 2005. Wang Z, Dolle P, Cardoso W, Niederreither K: Retinoic acid regulates morphogenesis and patterning of posterior foregut derivatives, Dev Biol 297:433–445, 2006. Wells JM, Melton DA: Vertebrate endoderm development, Annu Rev Cell Dev Biol 15:393–410, 1999. Wells JM, Melton DA: Early mouse endoderm is patterned by soluble factors from adjacent germ layers, Development 127:1563–1572, 2000. Wells JM, Esni F, Boivin GP, Aronow BJ, Stuart W, Combs C, Sklenka A, Leach SD, Lowy AM. Wnt/beta-catenin signaling is required for development of the exocrine pancreas. BMC Dev Biol. 7:4, 2007. Yamagishi H, Maeda J, Hu T, et al: Tbx1 is regulated by tissue-specific forkhead proteins through a common Sonic hedgehog-responsive enhancer, Genes Dev 17:269–281, 2003. Yamaguchi TP, Harpal K, Henkemeyer M, Rossant J: fgfr-1 is required for embryonic growth and mesodermal patterning during mouse gastrulation, Genes Dev 8:3032–3044, 1994. Zeng X, Yutzey KE, Whitsett JA: Thyroid transcription factor-1, hepatocyte nuclear factor-3beta and surfactant protein A and B in the developing chick lung, J Anat 193:399–408, 1998. Zeynali B, Dixon KE: Effects of retinoic acid on the endoderm in Xenopus embryos, Dev Genes Evol 208:318–326, 1998.

RECOMMENDED RESOURCES Developmental Biology Online: http://www.uoguelph.ca/zoology/devobio/dbindex.htm Chicken Developmental Stages: http://embryology.med.unsw.edu.au/OtherEmb/chick1.htm Embryo Images: http://www.med.unc.edu/embryo_images/ Jackson Laboratory Gene Expression Database: http://www.informatics.jax.org/menus/expression_menu.shtml International Society for Stem Cell Research: http://www.isscr.org The Wnt Homepage: http://www.stanford.edu/%7Ernusse/wntwindow.html

41

DEVELOPMENTAL GENETICS OF THE PULMONARY SYSTEM THOMAS J. MARIANI Department of Medicine, Brigham and Women’s Hospital and Pulmonary Bioinformatics, The Lung Biology Center, Harvard Medical School, Boston, MA

INTRODUCTION The mammalian lung develops as a lateral bud from the ventral foregut endoderm between the developing liver and the thymus. Lung development initiates at 5 weeks of gestation in the human and at 9 days of gestation in the mouse, and it is understood as proceeding through four discrete, subsequent stages: pseudoglandular, canalicular, saccular, and alveolar (Figure 41.1). Each of these stages is defined morphologically, and each encompasses distinct structural, cellular, and regulatory features. During the first 30 weeks after initiation in the human, the bud grows into a branched tubular structure that is reminiscent of other glandular organs to comprise the conducting airways. Until 36 weeks of gestation in the human or postnatal day 4 in the mouse, these tubes end in sacs that are incapable of efficient gas exchange. Birth occurring before or at this time (if untreated) is associated with increased morbidity and mortality as a result of lung immaturity. During the last few weeks of gestation and the first few years of life, these primitive sacs undergo a morphologic process that results in the development of mature alveoli. This process involves a dramatic increase in the surface area of the lung as a result of the formation and elongation of buds or secondary crests off of the walls of the primary sacs. Numerous adult lung diseases involve the destruction of the alveolar space. Therefore, a thorough understanding of the regulation of lung development, including alveolar formation and maintenance, could identify means to promote the maturation, limit the destruction, and support the regeneration of lung function. In the rodent, lung development initiates at mid-gestation and proceeds in a delayed fashion as compared with that seen in humans throughout gestation. In fact, a considerable amount of lung maturation, including the entire process of alveogenesis, occurs postnatally in the rodent. Parallels between

932

Principles of Developmental Genetics

© 2007, Elsevier Inc. All rights reserved.

LUNG SPECIFICATION AND SYMMETRY

933

FIGURE 41.1 A timeline for human lung development. Mammalian lung development occurs in discrete morphologic stages. The name and timing of the individual stages are shown, along with a general description of the cellular and molecular processes that occur during the different stages.

the distinct morphologic events and regulatory mechanisms that occur during mammalian lung development have been appropriately drawn with vertebrate limb development and Drosophila melanogaster larval tracheal development. The regulation of these morphogenic processes has been more fully defined (Johnson et al., 1994; Tickle, 1999; Capdevila and Izpisua Belmonte, 2001; Ghabrial et al., 2003; Cabernard et al., 2004), and this has served as a basis for the identification of many of the mechanisms that contribute to mammalian lung development. This chapter will focus on the principal regulatory mechanisms governing the initiation of lung formation, airway branching morphogenesis, sacculation, respiratory epithelial differentiation, and alveolar formation. Data derived from mouse genetics will be reviewed, and information relevant for human disease will be highlighted. Other recent review articles can be referred to for additional information (Warburton et al., 2005; Cardoso and Lu, 2006).

I. LUNG SPECIFICATION AND SYMMETRY The lung is specified within the developing foregut endoderm by a process that is incompletely defined, but it requires HNF3b, Gli, Shh, and retinoic acid (RA) signaling (Ang and Rossant, 1994; Motoyama et al., 1998; Litingtung et al., 1998; Mendelsohn et al., 1994; Desai et al., 2004). The early commitment of distinct tracheal and respiratory lineages occurs. This is exemplified by tracheal formation, but lung agenesis in cases of deficiencies of either fibroblast growth factor (FGF)-10 or its cognate receptor FGF receptor (FGFR)-2 (Peters et al., 1994; Min et al., 1998; Sekine et al., 1999; De Moerlooze et al., 2000). Although HNF3b specifies completion of the foregut, these other molecules that regulate lung bud initiation appear to comprise a regulatory module, with Gli acting as a regulator of the Shh pathway and RA promoting FGF signaling. Although tracheal malformations are not uncommon in humans, the genetic bases for these defects are presently poorly defined (see Chapter 40). Mammalian lungs are obviously asymmetric, differing in the number of lobes in the left (three in humans, one in mice) and right lungs (two in humans, four in mice). The specification of lung symmetry occurs through the same mechanisms used to define the left–right axis of the whole organism;

934

DEVELOPMENTAL GENETICS OF THE PULMONARY SYSTEM

principally the transforming growth factor-b (TGF-b)–related molecules Nodal, ACVR2, Lefty-1, and Lefty-2, along with the homeobox gene Pitx2. Lefty1 appears to initiate the promotion of “left sidedness,” whereas Pitx2 promotes lobar simplification (Kitamura et al., 1999). Recent data strongly suggest that Fog2/GATA signaling is necessary for proper lobar septation (Ackerman et al., 2005); the same study also implicates Fog2 as a potential genetic determinant of lung hypoplasia and congenital diaphragmatic hernia in humans (see Chapter 37).

II. BRANCHING MORPHOGENESIS During the past decade, genetic dissection has provided tremendous insight into the regulation of airway branching morphogenesis. Numerous regulatory pathways contributing to this process, including epidermal growth factor (EGF), TGF-b/bone morphogenetic protein (BMP), and FGF signaling have been identified. Although EGF receptor mutants display severely diminished branching and blunted lung development (Warburton et al., 1992; Miettinen et al., 1997), this is likely the result of reduced cellular proliferation (Goldin and Opperman, 1980). BMP-4 and FGF-10 play essential roles in the coordination of branching morphogenesis by specifying the timing and location of bud/tubule growth and elongation or arrest and branching. Sutherland et al. (1996) were among the first to appreciate the critical role played by FGF signaling during branching morphogenesis. Using larval tracheal development in Drosophila melanogaster as a model system, they identified dynamic waves of expression for both the Drosophila FGF ligand breathless and its receptor, branchless (see Chapter 21). The activation of this system is essential to drive tubule elongation and branch integration. Detailed studies in mice have shown evolutionary conservation of this regulatory process, with FGF-10 and FGFR-2 serving as the relevant ligand and receptor, respectively (Bellusci et al., 1997). In fact, FGF-10 can promote the direction of elongation of epithelial tubules in rodent lung explant cultures (Park et al., 1998). The fine-tuning of FGF signaling is provided by Sprouty-2 and -4 as well as the Shh pathway. The Sproutys act to limit the effects of FGF signaling by inhibiting receptor tyrosine kinase activity (Mason et al., 2006), thereby limiting bud growth and/or restricting sites of bud branching. Their potential to contribute specifically to lung branching morphogenesis was initially recognized during studies of Drosophila (Hacohen et al., 1998). In mammals, the inhibition of Sprouty activity leads to increased branching in vitro (Tefft et al., 1999), whereas overexpression results in decreased branching in vivo (Mailleux et al., 2001; Perl et al., 2003). Shh signals through patched/smoothened to regulate Gli activity. Like Sprouty, Shh acts to limit bud elongation by repressing FGF signaling at the bud tip. In the absence of Shh, the pattern of branching is severely disrupted (Pepicelli et al., 1998). Likewise, ectopic overexpression of Shh results in abnormal lung development (Bellusci et al., 1997). A deficiency in Hip1, which is a protein that interacts with and inhibits the function of Shh, results in increased Shh activity and decreased branching (Chuang et al., 2003). BMP-4 appears to be involved in the propagation of FGF-10–related branching morphogenesis signaling centers that specify branch initiation and

SACCULATION AND EPITHELIAL CELL DIFFERENTIATION

935

outgrowth sites. BMP-4 is expressed in the lung bud epithelium at the tip of the growing bud, which is juxtaposed to the FGF-10 expression in the surrounding mesenchyme (Bellusci et al., 1996). BMP-4 appears to signal the inhibition of elongation and to promote branching (Weaver et al., 2000). This may function (at least in part) by promoting the accumulation of extracellular matrix molecules, such as fibronectin, which serves as a physical barrier to form branching clefts (Sakai et al., 2003). Recent data indicate that BMP-4 can additionally or alternatively have a positive effect on the promotion of branching (Bragg et al., 2001). Wnt signaling is becoming more appreciated as a component part of the regulation of lung development. Numerous Wnts and their receptors are expressed in the developing lung. As will be discussed later, Wnts appear to have a major role in the regulation of pulmonary vascular development. Within the context of branching morphogenesis, the treatment of lung bud cultures with Dkk1 (a Wnt antagonist) has been shown to inhibit branching, apparently by reducing fibronectin-dependent cleft formation (De Langhe et al., 2005). Although branching morphogenesis is an essential component of the developmental process of the lung, deficiencies in this process are largely irrelevant to human disease, because they are completely incompatible with life. The notable exceptions are those processes that contribute to tracheal development, which underlie conditions such as tracheal–esophageal fistula, and those that contribute to lobar septation and diaphragm development, which may serve as genetic determinants for some cases of congenital diaphragmatic hernia and lung hypoplasia.

III. SACCULATION AND EPITHELIAL CELL DIFFERENTIATION At 26 weeks of human gestation or embryonic day 17.5 in the mouse, the lung undergoes the process of sacculation, which dramatically changes the distal architecture of the lung. A predominant feature of sacculation is the flattening of distal airway epithelial cells; this process is regulated by numerous factors, such as GATA-6, Nkx2.1, HNF3b, C/EBPa, glucocorticoid hormones, and FGFs (Cardoso, 2000). These and other morphologic changes are accompanied by the initial expansion of the eventual dozens of distinct cell types that will occupy the pulmonary system. A proximal–distal axis for the patterning of epithelial cell fate is established before sacculation, with BMP-4 playing an important regulatory role (Weaver et al., 1999). BMP-4 may participate in the signaling that is necessary to maintain developing lung epithelium in either an undifferentiated or distally (airspace) committed state as defined by surfactant protein C (SPC) expression. Additionally, the inhibition of BMP-4 promotes a proximal airway phenotype as defined by Clara cell secretory protein (CCSP; uteroglobin) expression. Wnt/b-catenin signaling also appears to be crucial to establishing this discrete proximal versus distal epithelial fate (Mucenski et al., 2003; Shu et al., 2005). In both cases, the disruption of the signaling leads to the “proximalization” of the distal lung epithelium, thereby suggesting that these pathways function to maintain cells in a distal phenotype. Conversely, Nkx2.1, which is the earliest marker of lung specification, appears to be necessary for the establishment of either proximal or distal fate. Deficiency in Nkx2.1 leads to severe abnormalities in lung development, with the complete absence

936

DEVELOPMENTAL GENETICS OF THE PULMONARY SYSTEM

of both SPC and CCSP. Many other molecules and pathways contribute to the specification of individual cell fates, of which we currently know very little. Two examples are the identification of the necessity of Foxj1 for the differentiation of ciliated cells (Chen et al., 1998; Brody et al., 2000) and of Foxa2 for the specification of goblet cells (Wan et al., 2004).

IV. MESENCHYMAL DIFFERENTIATION AND VASCULAR DEVELOPMENT Lung parenchymal fibroblasts, airway and vascular smooth muscle, pleural mesothelial, endothelial, and vascular support cells (pericytes) are all of mesenchymal origin, arising primarily from the lateral plate mesoderm. A thorough understanding of the regulation of the expansion and specification of cells of mesenchymal origin has lagged behind that available for epithelial cells. However, recent data have identified numerous pathways that contribute to the process of mesenchymal cell proliferation and differentiation. Platelet-derived growth factor (PDGF) signaling is understood to play a dominant role in the promotion of lung mesenchymal cell development, and PDGF receptor expression patterns can distinguish smooth muscle cell lineages (Lindahl et al., 1997). For example, the distal airway mesenchyme expresses high levels of PDGF-Ra, whereas proximal airway mesenchyme expresses relatively little. By contrast, both large vessel and capillary endothelium express PDGF-B, whereas both large vessel smooth muscle cells and capillary pericytes express high levels of PDGF-Rb. In addition, PDGF-A is essential for the differentiation of a-smooth muscle actin-expressing progenitor cells, which are ultimately responsible for parenchymal elastogenesis (Bostrom et al., 1996; Lindahl et al., 1997). As with the regulation of epithelial cell differentiation, Wnt signaling contributes to mesenchymal development in the lung. A deficiency of Wnt7b results in severe abnormalities in lung mesenchymal tissue, particularly defects in the vascular smooth muscle (Shu et al., 2002). Additionally, Wnt5a can contribute to the regulation of lung Shh and FGF-10 expression, and it is necessary for the proper proliferation and/or differentiation of the distal mesenchyme (Li et al., 2002; 2005). The pleura can also act as a source of factors that contribute to the regulation of lung mesenchymal cell proliferation and differentiation. FGF-9 is predominantly expressed in the lung pleura, and its absence leads to a severe reduction in lung mesenchyme development (Colvin et al., 2001). Likewise, RA is synthesized in the pleura (via Raldh2), and it appears to be capable of affecting distal lung morphogenesis during the canalicular stage of lung development (Malpel et al., 2000). It is intuitive that the appropriate vascularization of the airspace is essential for efficient gas exchange. However, only recently has there been an appreciation for the relative importance of pulmonary vascular development and an understanding of the necessity of coordinated parenchymal capillary formation and secondary crest elongation during alveogenesis. The pulmonary vascular network arises through a combination of vasculogenesis, occurring proximal to conducting airways, and angiogenesis, arising from the aortic arches and the heart. There are two primary waves of vascularization during lung development: the first occurs at the saccular phase of development, and the second occurs during the alveolar phase.

SECONDARY CREST ELONGATION AND ALVEOGENESIS

937

Vascular endothelial growth factor (VEGF) is a master regulator of vascular development, and it appears to play a principal role in the regulation of the formation and maintenance of the lung vasculature and, thus, the alveoli. VEGF is expressed in both the mesenchyme and the epithelium during lung development (Acarregui et al., 1999) and it can promote the proliferation of endothelial as well as epithelial cells (Brown et al., 2001). The inhibition of VEGF function in adult rodents leads to a failure to maintain normal alveolar architecture and an emphysema-like phenotype (Kasahara et al., 2000; Petrache et al., 2005). Another molecule that seems to regulate vascular development in the lung is Scye1. Scye1, or EmapII, is a cytokine that is induced by apoptosis and that is shown to have tumor vasculature regulatory properties. Excess Scye1 inhibits vessel formation in an ectopic model of lung vascular development and leads to insufficient epithelial maturation (Schwarz et al., 2000). Unfortunately, the further clarification of distinct mesenchymal cell lineages and the exploration of the mechanisms controlling their establishment are limited by the lack of specific markers, such as those that are available for distinct epithelial lineages (e.g., SPC, CCSP). However, optimism that significant progress will be made on this front in the near future seems warranted. In particular, the combination of genomics technologies currently available (including the power of genome-wide expression profiling to identify cell-type–specific markers) in combination with the power of mouse genetics to target specific cell populations should facilitate this endeavor.

V. SECONDARY CREST ELONGATION AND ALVEOGENESIS Through the end of the saccular phase of lung development, the organ is woefully inefficient at performing its essential function: gas exchange. This is the result of two primary problems: a limited respiratory surface area and a poorly organized parenchymal capillary bed. Alveogenesis, which is the final stage of lung development, is the process that transforms the airspaces, primarily through the initiation and elongation of secondary septae (or crests) at points along the walls of the primary saccules. This is a highly coordinated and distinct morphogenic event, and its regulation is poorly understood. However, work during the past decade has identified some of the key pathways that contribute to the process. It is particularly important to point out that the molecules that regulate terminal lung development (both those related to insufficient maturation and those related to an inability to maintain and/or repair the lung) are likely to be more relevant to human disease. For example, mutations in the essential elastin fiber component fibrillin1 are associated with Marfan’s-associated emphysema (Neptune et al., 2003). It is clear that PDGF, FGF, and RA signaling play distinct but complementary and essential roles in the coordination of alveogenesis. As mentioned previously, PDGF-A is essential for the specification of a mesenchymal progenitor cell population that will play a prominent role in secondary crest elongation (Bostrom et al., 1996). These a-smooth muscle actin-expressing cells migrate to sites of potential secondary crest initiation during the saccular stage of development, and they contribute to secondary crest elongation at least in part by producing parenchymal elastic fibers. Coordinating appropriate parenchymal elastogenesis is a common feature of most pathways that are

938

DEVELOPMENTAL GENETICS OF THE PULMONARY SYSTEM

known to play a primary role in alveogenesis. The requirement for proper elastogenesis in the establishment of appropriate airspace structure is highlighted by the failure of the lung to undergo alveogenesis in mice that are deficient in any one of many elastin fiber components (Wendel et al., 2000; Loeys et al., 2002; Neptune et al., 2003; Liu et al., 2004; Maki et al., 2005). The development of the mammalian lung involves the repeated use of a limited number of regulatory modules (e.g., BMP-4, FGF-10) that reiteratively contribute to multiple developmental processes. The RA and FGF signaling pathways represent such regulatory modules, contributing to initial lung bud outgrowth, branching morphogenesis, and secondary crest elongation. Numerous lines of evidence reveal an important role for RA signaling in the establishment of proper alveolar structure. Deficiencies of numerous RA receptors, alone or in combination, lead to insufficient alveolar formation (Luo et al., 1996; Kastner et al., 1997; Massaro et al., 2000; McGowan et al., 2000; Massaro et al., 2003), and these deficiencies are almost always accompanied by decreased alveolar elastin production. In addition, supplementation with dietary RA can accelerate alveolar formation in neonatal rodents and promote alveolar regeneration and repair in adult rats (Massaro and Massaro, 1996; 1997; 2000). This is completely consistent with deficient alveogenesis in animals lacking one of numerous elastin fiber components, as described previously. FGFR-3 and -4 coordinately promote alveogenesis as defined by the failure of terminal lung development in mice that are lacking the expression of both receptors (Weinstein et al., 1998; Hokuto et al., 2003). It is unclear at this time what regulatory mechanisms necessary for secondary crest elongation are affected in the absence of FGF signaling. Although it may seem counterintuitive, failed alveogenesis as a result of deficiencies in FGFR-3 and -4 is accompanied by hyperactive elastogenesis. This can be rectified with the numerous observations that insufficient elastin fiber formation is associated with incomplete alveolar formation if one models secondary crest elongation as a proximal–distal process in which RA and FGF act as competing morphogens to extend the secondary septum and occupy defined proximal and distal portions, respectively (Mariani, 2004; Mariani and Kaminski, 2004). This is referred to as the “balloon” model of alveogenesis (Figure 41.2), because it envisions the alveolus extending in a distal fashion away from the airways, whereas the proximal portion remains fixed in space, much like the inflation of a balloon. In this model, the tip of the alveolar/secondary septum occupies a proximal position and remains tethered to the conducting airways through extracellular matrix, whereas the base of the septum (which initially occupies a preproximal/predistal position) extends to eventually occupy a distal position. RA signaling, which typically specifies a proximal fate, occurs at the septal tip and promotes elastogenesis in a mesenchymal cell with a discrete phenotype. This is consistent with the normal accumulation of elastin fibers at the tips of alveolar septae, which essentially defines the “neck” or opening of the alveolus. Alternatively, FGF/FGFR signaling, which typically specifies a distal fate, occurs at the distalizing “base” of the septum and suppresses mesenchymal cells from assuming the proximal, elastogenic phenotype. Although this model accurately predicts the major observations regarding the regulation of alveogenesis, many aspects await experimental validation.

TRANSITION TO AIR BREATHING

939

The “balloon” model of alveogenesis. Top left, The “sprouting and extension” model for secondary crest elongation depicts alveogenesis as secondary bud sprouting from the walls of primary septae and elongating into the airspace (Pierce and Shipley, 2000). Top right, The “balloon” model for the regulation of alveogenesis depicts secondary crest elongation as a distalization of the alveolar sac, with the tips of secondary crests being fixed in space by extracellular matrix tethering to the airways (Mariani, 2004). Bottom, This morphogenic process can be envisioned as occurring along a proximal–distal axis where alveolar saccules expand in a distal direction, like a balloon does when it is filled with air. A preproximal/predistal “stem cell” is differentiated into distinct proximal and distal cell populations. Proximalization is driven by retinoic acid signaling and characterized by elastin fiber production. Distalization is driven by fibroblast growth factor signaling, which functions to suppress the proximal phenotype. (See color insert.)

FIGURE 41.2

VI. TRANSITION TO AIR BREATHING In addition to the tissue morphogenesis and cellular differentiation that occur throughout lung development, at birth, the organ must transition from being filled with fluid to filled with air. Although a significant body of work has identified discrete mechanisms that contribute to this transition (including the removal of water from the lumen and the secretion of surfactant), recent work has highlighted some regulatory control mechanisms for this transition and revealed that these are common genetic determinants of neonatal morbidity and mortality. Pulmonary surfactant is composed of a complex mixture of lipids (both phospholipids and neutral lipids) and hydrophobic proteins that forms a physical barrier between airspace gas and surface liquid in the alveoli and that contributes a reduction in airspace surface tension. Surfactant is synthesized by type II pneumocytes, is stored in lamellar bodies and secreted during the antenatal period. Dynamic changes in the secretion, structure, and recycling of this lipoprotein complex during development, homeostasis, and lung disease are evident. For example, granulocyte macrophage colony stimulating factor is a critical regulator of surfactant recycling by alveolar macrophages, and a reduction in this activity leads to idiopathic alveolar proteinosis in humans (Trapnell et al., 2003). During the past few years, Whitsett and colleagues have defined a regulatory network that controls the production of pulmonary surfactant, which is necessary for the transition to air breathing. It was initially discovered that one of the hydrophobic protein components of surfactant, Sftpb, was essential

940

DEVELOPMENTAL GENETICS OF THE PULMONARY SYSTEM

for surfactant secretion. This was simultaneously appreciated in animal models of Sftpb deficiency and as genetic mutations in Sftpb were identified as a cause of respiratory distress in humans (Nogee et al., 1994; Clark et al., 1995). Similarly, a deficiency in the lamellar-body–associated membrane transporter Abca3 was identified as a basis for neonatal respiratory distress (Shulenin et al., 2004). Although the exact function of this transporter is not known, its association with lamellar bodies within type II pneumocytes suggests that it may participate in regulating surfactant organization and/or secretion. Subsequently, the forkhead box transcription factor, Foxa2, was identified as a master regulator of surfactant protein and Abca3 gene expression (Wan et al., 2004). Deficiency in Foxa2 in mice recapitulated many of the pathologic features of respiratory distress in humans. Recently, the transcription factor Cebpa was shown to contribute to this regulatory module. The deletion of Cebpa in mice leads to respiratory distress, structural deficiency in lung maturation, and alterations in surfactant and Abca3 protein expression (Martis et al., 2006). Cebpa is a direct target of both Foxa2 and Titf1, and it requires both of these transcription factors for appropriate expression. A regulatory pathway involving Hif2a and VEGF has also been implicated in the regulation of surfactant production and secretion and the transition to air breathing (Compernolle et al., 2002). A deficiency of Hif2a in mice leads to respiratory distress, insufficient surfactant production, and reduced type II pneumocyte VEGF production (a known target of Hif2a). Blocking VEGF function resulted in respiratory distress in mice, and VEGF supplementation promoted surfactant production and was capable of promoting the survival of prematurely delivered mice. The significance of this pathway to lung maturation and surfactant production, how it may integrate with the one described previously, and its potential role in human disease are not known.

VII. CONCLUSION Mammalian lung development is a complex process involving multiple morphogenic events that reliably result in an intricate, delicate, yet durable organ that is essential for life. Genetic studies of development in model organisms (particularly the mouse and the fruit fly) have clearly contributed to the global understanding of the mechanisms involved. Studies of these model organisms have further provided insight into the genetic and mechanistic nature of human lung diseases, such as neonatal respiratory distress syndrome. Although intensive research during the past decade has provided enlightenment concerning many of the regulatory mechanisms involved in creating this organ, much remains to be learned. We currently have a good understanding of the processes of early lung development (particularly branching morphogenesis), but we lack a thorough appreciation of the mechanisms that contribute to lung maturation, vascularization, and cell-type specification (particularly for nonepithelial cells). The recent advent of genomics-based technologies (including genome-wide expression profiling, conditional gene targeting, and whole-genome sequencing) promises to promote the rapid advancement of our current understanding. Given the potential to harness the knowledge of the regulation of lung development to both promote maturation in premature infants and to facilitate lung regeneration within the context of chronic lung disease, these future discoveries should prove to be most exciting.

941

GLOSSARY

SUMMARY

 Mammalian lung development is a highly complex and exquisitely regu-

  



lated process involving morphogenic and regulatory processes that resemble vertebrate limb development and Drosophila melanogaster larval tracheal development. Lung development proceeds through histologically defined stages, each of which involves discrete morphologic characteristics, cellular alterations, and regulatory processes. Many of the signaling pathways that contribute to the individual stages and morphogenic processes of lung development also contribute to other processes at other times. We currently have a more thorough understanding of the regulation of early lung development, including branching morphogenesis, than we have of either lung maturation or the pathways that specify individual cell types (particularly with respect to mesenchymal lineages). Processes that contribute to terminal lung development, maturation, maintenance, and regeneration are more likely to be genetic determinants of human disease at least in part because failures early during the developmental process may be compounded over time.

ACKNOWLEDGMENTS Supported in part by National Institutes of Health grant HL071885.

GLOSSARY Airspace The respiratory portion of the lung, where gas exchange occurs. Sometimes referred to as alveoli (singular: alveolus). Airspace is distinct from the airways, which conduct air into and out of the lungs. Alveogenesis/alveolization The stage of lung development that encompasses the formation of the functional, respiratory portion of the organ. This process is initiated in humans during late fetal development, and it continues through the first few years of life. In rodents, this process occurs entirely postnatally, beginning at approximately 1 week of age. Branching morphogenesis A process of repetitive tube elongation and branching that occurs throughout early and mid lung development and that gives rise to the conducting airways. Canalicular The stage of lung development that is subsequent to the pseudoglandular stage and that involves the completion of the establishment of the conducting airways. During this stage, the lung is histologically rich in airways with clearly defined lumen. This stage occurs from weeks 16 to 26 of gestation in human development and from embryonic days 16.5 to 17.5 in mouse lung development.

942

DEVELOPMENTAL GENETICS OF THE PULMONARY SYSTEM

Pseudoglandular The initial stage of lung development, subsequent to organ budding from the ventral foregut, during which the lung histologically resembles a solid organ. A predominant feature of this stage is branching morphogenesis. This stage occurs from weeks 6 to 16 of gestation in human development and from embryonic days 10.5 to 16.5 in mouse lung development. Saccular The stage of lung development subsequent to the canalicular stage that involves the initiation of respiratory cell differentiation and airspace vascular development. It is histologically defined by the flattening of the epithelium at the distal end of the airways. This stage occurs from weeks 26 to 36 of gestation in human development and from embryonic day 17.5 to postnatal day 4 in mouse lung development. Secondary crest elongation One of the major morphologic changes that occurs during alveogenesis. It involves the formation and elongation of parenchymal tissue (secondary septae or crests) from locations along the walls of existing saccules (primary septae) in the distal respiratory region of the lung. Vasculogenesis The de novo formation of blood vessels distant from existing ones. It often involves either the reorganization of a disorganized vascular plexus and/or the transdifferentiation of cells/tissues into blood vessels.

REFERENCES Acarregui MJ, Penisten ST, Goss KL, et al: Vascular endothelial growth factor gene expression in human fetal lung in vitro, Am J Respir Cell Mol Biol 20:14–23, 1999. Ackerman KG, Herron BJ, Vargas SO, et al: Fog2 is required for normal diaphragm and lung development in mice and humans, PLoS Genet 1:58–65, 2005. Ang SL, Rossant J: HNF-3 beta is essential for node and notochord formation in mouse development, Cell 78:561–574, 1994. Bellusci S, Furuta Y, Rush MG, et al: Involvement of Sonic hedgehog (Shh) in mouse embryonic lung growth and morphogenesis, Development 124:53–63, 1997. Bellusci S, Grindley J, Emoto H, et al: Fibroblast growth factor 10 (FGF10) and branching morphogenesis in the embryonic mouse lung, Development 124:4867–4878, 1997. Bellusci S, Henderson R, Winner G, et al: Evidence from normal expression and targeted misexpression that bone morphogenetic protein (BMP-4) plays a role in mouse embryonic lung morphogenesis, Development 122:1693–1702, 1996. Bostrom H, Willetts K, Pekny M, et al: PDGF-A signaling is a critical event in lung alveolar myofibroblast development and alveogenesis, Cell 85:863–873, 1996. Bragg AD, Moses HL, Serra R: Signaling to the epithelium is not sufficient to mediate all of the effects of transforming growth factor beta and bone morphogenetic protein 4 on murine embryonic lung development, Mech Dev 109:13–26, 2001. Brody SL, Yan XH, Wuerffel MK, et al: Ciliogenesis and left-right axis defects in forkhead factor HFH-4-null mice, Am J Respir Cell Mol Biol 23:45–51, 2000. Brown KR, England KM, Goss KL, et al: VEGF induces airway epithelial cell proliferation in human fetal lung in vitro, Am J Physiol Lung Cell Mol Physiol 281:L1001–L1010, 2001. Cabernard C, Neumann M, Affolter M: Cellular and molecular mechanisms involved in branching morphogenesis of the Drosophila tracheal system, J Appl Physiol 97:2347–2353, 2004. Capdevila J, Izpisua Belmonte JC: Patterning mechanisms controlling vertebrate limb development, Annu Rev Cell Dev Biol 17:87–132, 2001. Cardoso W: Lung morphogenesis revisited: old facts, current ideas, Dev Dyn 219:121–130, 2000.

REFERENCES

943 Cardoso WV, Lu J: Regulation of early lung morphogenesis: questions, facts and controversies, Development 133:1611–1624, 2006. Chen J, Knowles HJ, Hebert JL, Hackett BP: Mutation of the mouse hepatocyte nuclear factor/ forkhead homologue 4 gene results in an absence of cilia and random left-right asymmetry, J Clin Invest 102:1077–1082, 1998. Chuang PT, Kawcak T, McMahon AP: Feedback control of mammalian Hedgehog signaling by the Hedgehog-binding protein, Hip1, modulates Fgf signaling during branching morphogenesis of the lung, Genes Dev 17:342–347, 2003. Clark JC, Wert SE, Bachurski CJ, et al: Targeted disruption of the surfactant protein B gene disrupts surfactant homeostasis, causing respiratory failure in newborn mice, Proc Natl Acad Sci U S A 92:7794–7798, 1995. Colvin JS, White AC, Pratt SJ, Ornitz DM: Lung hypoplasia and neonatal death in Fgf9-null mice identify this gene as an essential regulator of lung mesenchyme, Development 128:2095–2106, 2001. Compernolle V, Brusselmans K, Acker T, et al: Loss of HIF-2alpha and inhibition of VEGF impair fetal lung maturation, whereas treatment with VEGF prevents fatal respiratory distress in premature mice, Nat Med 8:702–710, 2002. De Langhe SP, Sala FG, Del Moral PM, et al: Dickkopf-1 (DKK1) reveals that fibronectin is a major target of Wnt signaling in branching morphogenesis of the mouse embryonic lung, Dev Biol 277:316–331, 2005. De Moerlooze L, Spencer-Dene B, Revest J, et al: An important role for the IIIb isoform of fibroblast growth factor receptor 2 (FGFR2) in mesenchymal-epithelial signalling during mouse organogenesis, Development 127:483–492, 2000. Desai TJ, Malpel S, Flentke GR, et al: Retinoic acid selectively regulates Fgf10 expression and maintains cell identity in the prospective lung field of the developing foregut, Dev Biol 273:402–415, 2004. Ghabrial A, Luschnig S, Metzstein MM, Krasnow MA: Branching morphogenesis of the Drosophila tracheal system, Annu Rev Cell Dev Biol 19:623–647, 2003. Goldin GV, Opperman LA: Induction of supernumerary tracheal buds and the stimulation of DNA synthesis in the embryonic chick lung and trachea by epidermal growth factor, J Embryol Exp Morphol 60:235–243, 1980. Hacohen N, Kramer S, Sutherland D, et al: Sprouty encodes a novel antagonist of FGF signaling that patterns apical branching of the Drosophila airways, Cell 92:253–263, 1998. Hokuto I, Perl AK, Whitsett JA: Prenatal, but not postnatal, inhibition of fibroblast growth factor receptor signaling causes emphysema, J Biol Chem 278:415–421, 2003. Johnson RL, Riddle RD, Tabin CJ: Mechanisms of limb patterning, Curr Opin Genet Dev 4:535–542, 1994. Kasahara Y, Tuder RM, Taraseviciene-Stewart L, et al: Inhibition of VEGF receptors causes lung cell apoptosis and emphysema, J Clin Invest 106:1311–1319, 2000. Kastner P, Mark M, Ghyselinck N, et al: Genetic evidence that the retinoid signal is transduced by heterodimeric RXR/RAR functional units during mouse development, Development 124:313–326, 1997. Kitamura K, Miura H, Miyagawa-Tomita S, et al: Mouse Pitx2 deficiency leads to anomalies of the ventral body wall, heart, extra- and periocular mesoderm and right pulmonary isomerism, Development 126:5749–5758, 1999. Li C, Hu L, Xiao J, et al: Wnt5a regulates Shh and Fgf10 signaling during lung development, Dev Biol 287:86–97, 2005. Li C, Xiao J, Hormi K, et al: Wnt5a participates in distal lung morphogenesis, Dev Biol 248:68–81, 2002. Lindahl P, Karlsson L, Hellstrom M, et al: Alveogenesis failure in PDGF-A-deficient mice is coupled to lack of distal spreading of alveolar smooth muscle cell progenitors during lung development, Development 124:3943–3953, 1997. Litingtung Y, Lei L, Westphal H, Chiang C: Sonic Hedgehog is essential to foregut development, Nat Genet 20:58–61, 1998. Liu X, Zhao Y, Gao J, et al: Elastic fiber homeostasis requires lysyl oxidase-like 1 protein, Nat Genet 36:178–182, 2004. Loeys B, Van Maldergem L, Mortier G, et al: Homozygosity for a missense mutation in fibulin-5 (FBLN5) results in a severe form of cutis laxa, Hum Mol Genet 11:2113–2118, 2002. Luo J, Sucov HM, Bader JA, et al: Compound mutants for retinoic acid receptor (RAR) beta and RAR alpha 1 reveal developmental functions for multiple RAR beta isoforms, Mech Dev 55:33–44, 1996.

944

DEVELOPMENTAL GENETICS OF THE PULMONARY SYSTEM

Mailleux AA, Tefft D, Ndiaye D, et al: Evidence that SPROUTY2 functions as an inhibitor of mouse embryonic lung growth and morphogenesis, Mech Dev 102:81–94, 2001. Maki JM, Sormunen R, Lippo S, et al: Lysyl oxidase is essential for normal development and function of the respiratory system and for the integrity of elastic and collagen fibers in various tissues, Am J Pathol 167:927–936, 2005. Malpel S, Mendelsohn C, Cardoso WV: Regulation of retinoic acid signaling during lung morphogenesis, Development 127:3057–3067, 2000. Mariani T: Regulation of alveogenesis by reciprocal proximodistal FGF and retinoic acid signaling, Am J Respir Cell Mol Biol 31:S52–S57, 2004. Mariani TJ, Kaminski N: Gene expression studies in lung development and lung stem cell biology, Curr Top Dev Biol 64:57–71, 2004. Martis PC, Whitsett JA, Xu Y, et al: C/EBPalpha is required for lung maturation at birth, Development 133:1155–1164, 2006. Mason JM, Morrison DJ, Basson MA, Licht JD: Sprouty proteins: multifaceted negative-feedback regulators of receptor tyrosine kinase signaling, Trends Cell Biol 16:45–54, 2006. Massaro G, Massaro D: Postnatal treatment with retinoic acid increases the number of pulmonary alveoli in rats, Am J Physiol 270:L305–L310, 1996. Massaro G, Massaro D: Retinoic acid treatment abrogates elastase-induced pulmonary emphysema in rats, Nat Med 3:675–677, 1997. Massaro G, Massaro D: Retinoic acid treatment partially rescues failed septation in rats and in mice, Am J Physiol Lung Cell Mol Physiol 278:L955–L960, 2000. Massaro GD, Massaro D, Chambon P: Retinoic acid receptor-alpha regulates pulmonary alveolus formation in mice after, but not during, perinatal period, Am J Physiol Lung Cell Mol Physiol 284:L431–L433, 2003. Massaro GD, Massaro D, Chan WY, et al: Retinoic acid receptor-beta: an endogenous inhibitor of the perinatal formation of pulmonary alveoli, Physiol Genomics 4:51–57, 2000. McGowan S, Jackson SK, Jenkins-Moore M, et al: Mice bearing deletions of retinoic acid receptors demonstrate reduced lung elastin and alveolar numbers, Am J Respir Cell Mol Biol 23:162–167, 2000. Mendelsohn C, Lohnes D, Decimo D, et al: Function of the retinoic acid receptors (RARs) during development (II). Multiple abnormalities at various stages of organogenesis in RAR double mutants, Development 120:2749–2771, 1994. Miettinen PJ, Warburton D, Bu D, et al: Impaired lung branching morphogenesis in the absence of functional EGF receptor, Dev Biol 186:224–236, 1997. Min H, Danilenko D, Scully S, et al: Fgf-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless, Genes Dev 12:3156–3161, 1998. Motoyama J, Liu J, Mo R, et al: Essential function of Gli2 and Gli3 in the formation of lung, trachea and esophagus, Nat Genet 20:54–57, 1998. Mucenski ML, Wert SE, Nation JM, et al: beta-Catenin is required for specification of proximal/ distal cell fate during lung morphogenesis, J Biol Chem 278:40231–40238, 2003. Neptune ER, Frischmeyer PA, Arking DE, et al: Dysregulation of TGF-beta activation contributes to pathogenesis in Marfan syndrome, Nat Genet 33:407–411, 2003. Nogee LM, Garnier G, Dietz HC, et al: A mutation in the surfactant protein B gene responsible for fatal neonatal respiratory disease in multiple kindreds, J Clin Invest 93:1860–1863, 1994. Park W, Miranda B, Lebeche D, et al: FGF-10 is a chemotactic factor for distal epithelial buds during lung development, Dev Biol 201:125–134, 1998. Pepicelli C, Lewis P, McMahon A: Sonic Hedgehog regulates branching morphogenesis in the mammalian lung, Curr Biol 8:1083–1086, 1998. Perl AK, Hokuto I, Impagnatiello MA, et al: Temporal effects of Sprouty on lung morphogenesis, Dev Biol 258:154–168, 2003. Peters K, Werner S, Liao X, et al: Targeted expression of a dominant negative FGF receptor blocks branching morphogenesis and epithelial differentiation of the mouse lung, EMBO J 13:3296–3301, 1994. Petrache I, Natarajan V, Zhen L, et al: Ceramide upregulation causes pulmonary cell apoptosis and emphysema-like disease in mice, Nat Med 11:491–498, 2005. Pierce R, Shipley J: Retinoid-enhanced alveolization: Identifying relevant downstream targets, Am J Respir Cell Mol Biol 23:137–141, 2000. Sakai T, Larsen M, Yamada KM: Fibronectin requirement in branching morphogenesis, Nature 423:876–881, 2003.

RECOMMENDED RESOURCES

945

Schwarz MA, Zhang F, Gebb S, et al: Endothelial monocyte activating polypeptide II inhibits lung neovascularization and airway epithelial morphogenesis, Mech Dev 95:123–132, 2000. Sekine K, Ohuchi H, Fujiwara M, et al: Fgf10 is essential for limb and lung formation, Nat Genet 21:138–141, 1999. Shu W, Guttentag S, Wang Z, et al: Wnt/beta-catenin signaling acts upstream of N-myc, BMP4, and FGF signaling to regulate proximal-distal patterning in the lung, Dev Biol 283:226–239, 2005. Shu W, Jiang YQ, Lu MM, Morrisey EE: Wnt7b regulates mesenchymal proliferation and vascular development in the lung, Development 129:4831–4842, 2002. Shulenin S, Nogee LM, Annilo T, et al: ABCA3 gene mutations in newborns with fatal surfactant deficiency, N Engl J Med 350:1296–1303, 2004. Sutherland D, Samakovlis C, Krasnow M: Breathless encodes a Drosophila FGF homolog that control tracheal cell migration and the pattern of branching, Cell 87:1091–1101, 1996. Tefft D, Lee M, Smith S, et al: Conserved function of mSpry-2, a murine homolog of sprouty, which negatively modulates respiratory organogenesis, Curr Biol 9:219–222, 1999. Tickle C: Morphogen gradients in vertebrate limb development, Semin Cell Dev Biol 10:345–351, 1999. Trapnell BC, Whitsett JA, Nakata K: Pulmonary alveolar proteinosis, N Engl J Med 349:2527–2539, 2003. Wan H, Kaestner KH, Ang SL, et al: Foxa2 regulates alveolarization and goblet cell hyperplasia, Development 131:953–964, 2004. Wan H, Xu Y, Ikegami M, et al: Foxa2 is required for transition to air breathing at birth, Proc Natl Acad Sci U S A 101:14449–14454, 2004. Warburton D, Bellusci S, De Langhe S, et al: Molecular mechanisms of early lung specification and branching morphogenesis, Pediatr Res 57(5 Pt 2):26R–37R, 2005. Warburton D, Seth R, Shum L, et al: Epigenetic role of epidermal growth factor expression and signalling in embryonic mouse lung morphogenesis, Dev Biol 149:123–133, 1992. Weaver M, Dunn N, Hogan B: Bmp-4 and FGF-10 play opposing roles during lung bud morphogenesis, Development 127:2695–2704, 2000. Weaver M, Yingling J, Dunn N, et al: BMP signaling regulates proximo-distal differentiation of endoderm in mouse lung development, Development 126:4005–4015, 1999. Weinstein M, Xu X, Ohyama K, Deng C-X: FGFR-3 and FGFR-4 function cooperatively to direct alveogenesis in the murine lung, Development 125:3615–3623, 1998. Wendel D, Taylor D, Albertine K, et al: Impaired distal airway development in mice lacking elastin, Am J Respir Cell Mol Biol 23:320–326, 2000.

RECOMMENDED RESOURCES http://www.embryology.ch/anglais/rrespiratory/phasen01.html Aliotta JM, Passero M, Meharg J, et al: Stem cells and pulmonary metamorphosis: new concepts in repair and regeneration, J Cell Physiol 204:725–741, 2005. Bourbon J, Boucherat O, Chailley-Heu B, Delacourt C: Control mechanisms of lung alveolar development and their disorders in bronchopulmonary dysplasia, Pediatr Res 57(5 Pt 2): 38R–46R, 2005. Cardoso WV, Lu J: Regulation of early lung morphogenesis: questions, facts and controversies, Development 133:1611–1624, 2006. Kim N, Vu TH: Parabronchial smooth muscle cells and alveolar myofibroblasts in lung development, Birth Defects Res C Embryo Today 78:80–89, 2006. Roth-Kleiner M, Post M: Similarities and dissimilarities of branching and septation during lung development, Pediatr Pulmonol 40:113–134, 2005. Warburton D, Bellusci S, De Langhe S, et al: Molecular mechanisms of early lung specification and branching morphogenesis, Pediatr Res 57(5 Pt 2):26R–37R, 2005.

42

PANCREAS DEVELOPMENT AND STEM CELLS MAUREEN GANNON Departments of Medicine and Molecular Physiology and Biophysics, and Program in Developmental Biology, Vanderbilt University Medical Center, Nashville, TN

INTRODUCTION Approximately 18 million Americans have diabetes, which is a heterogeneous group of disorders characterized by the decreased function of insulin-producing b cells and insufficient insulin output. Diabetes results from an absolute (type 1) or relative (type 2) inadequate functional b-cell mass. Whereas type 1 diabetes is characterized by the selective autoimmune destruction of b cells (Gale, 2001), type 2 diabetes occurs when the b-cell population fails to compensate for the increased peripheral insulin resistance associated with obesity (Kahn, 1998). Thus, both forms of the disease would greatly benefit from treatment strategies that could enhance b-cell regeneration and/or proliferation. Although there have been some encouraging results from islet transplantation in achieving the remission of type 1 diabetes (Shapiro et al., 2000; Ryan et al., 2001), the limited amount of donor tissue obtainable makes this potential treatment unavailable to most patients. The ability to induce b cells or whole islets from pancreatic stem cells in vivo or in vitro or embryonic stems in vitro would provide an alternative source of transplantable tissue (Lumelsky et al., 2001; Odorico et al., 2001; Bonner-Weir and Sharma, 2002). Additionally, studies addressing the proliferation, regeneration, and neogenesis of b cells in the adult pancreas could lead to the restoration of b-cell mass in individuals with type 1 diabetes and enhanced b-cell compensation in patients with type 2 diabetes. Successful and reliable methods of generating islet endocrine cells in vivo or in vitro will benefit greatly from a thorough understanding of the normal developmental processes that occur during pancreatic organogenesis (e.g., transcription factors, cell-signaling molecules, and cell–cell interactions that regulate endocrine proliferation and differentiation from the embryonic pancreatic epithelium).

946

Principles of Developmental Genetics

© 2007, Elsevier Inc. All rights reserved.

THE INITIAL STAGES OF PANCREATIC BUD FORMATION

947

Recently, much progress has been made in the identification of the factors involved in the normal development and differentiation of the various pancreatic cell types. Interestingly, mutations of many of these developmentally important factors have been identified in individuals with diabetes. This chapter will summarize what is known about the regulation of pancreas development and about the factors that control mature islet function. We will also discuss potential pancreatic and endocrine stem/progenitor cell sources and the recent progress in generating insulin-producing cells in culture.

I. THE INITIAL STAGES OF PANCREATIC BUD FORMATION The mature pancreas is composed of two distinct functional units: the exocrine component of the pancreas, which consists of clusters of acinar cells that produce and secrete digestive enzymes such as amylase and elastase, and the ductal network, which transports the acinar secretions into the rostral duodenum. The endocrine compartment is composed of spherical clusters of at least five hormone-producing cell types: insulin (b cells), glucagon (a cells), somatostatin (d cells), ghrelin (e cells), and pancreatic polypeptide (PP cells; Figures 42.1 and 42.2). These endocrine clusters comprise microorgans known as the islets of Langerhans. Together, these islet hormones regulate glucose homeostasis by facilitating the uptake of ingested glucose into cells and stimulating glucose production by the liver during times of fasting. Acinar, ductal, and endocrine cells are all derived from the endoderm during embryonic development (Percival and Slack, 1999).

FIGURE 42.1 Schematic of pancreas and islet development. A, The pancreas arises as dorsal and ventral evaginations from the posterior foregut endoderm on embryonic day 9.5, which is marked by the expression of the Pdx1 transcription factor (yellow). Markers of the early pancreatic buds include the transcription factors Ptf1a and Hb9 (blue). Within the developing buds, a subset of cells expresses markers of the endocrine lineage, including Ngn3, Isl-1, and Pax6 (red). B, As development proceeds, the pancreatic epithelium (yellow) becomes a highly branched ductal network. Endocrine cells (green nuclei) and exocrine cells (blue nuclei) arise from the ducts. Endocrine progenitors are scattered throughout the epithelium, and they express Ngn3 (red nuclei). These cells maintain Ngn3 expression as they delaminate from the epithelium (tan-colored cells). Ngn3 is downregulated as hormone expression begins (green, red, and orange cells), and more general endocrine markers such as Isl-1 and Pax6 are expressed (green nuclei). C, Mature islets begin to form during late gestation. In the mouse, insulin-producing b cells (green) are found at the islet core, and all other hormone-producing cells are located at the periphery. (See color insert.)

948

PANCREAS DEVELOPMENT AND STEM CELLS

The epithelial component of the pancreas arises from dorsal and ventral outgrowths of the posterior foregut endoderm just caudal to the liver diverticulum in all vertebrates examined, including humans (see Figure 42.1; Wessels and Cohen, 1967; Kim et al., 1997; Kelly and Melton, 2000; Field et al., 2003; Ober et al., 2003). The dorsal and ventral buds later fuse to form a single organ (this occurs on embryonic day 12.5 in the mouse). Pancreatic bud formation can be observed as early as day 25 of gestation in humans (Ashraf et al., 2005), on embryonic day 9.5 in the mouse (Slack, 1995), and 24 hours postfertilization in zebrafish (see Figure 42.2; Field et al., 2003). In mammals, frogs, and chickens, both pancreatic buds generate exocrine and endocrine cells. By contrast, in zebrafish, the posterior (ventral) bud generates the endocrine tissue, which usually consists of a single islet, whereas the anterior (dorsal) bud gives rise primarily to the pancreatic duct and the acinar cells, although endocrine cells do arise from dorsal bud derivatives later during development (Field et al., 2003). The morphogenesis of the pancreatic epithelium yields a highly branched ductal network within which multipotent progenitors for both exocrine and endocrine cells are thought to reside (see Figure 42.1; Bouwens and De Blay, 1996; Fishman and Melton, 2002). It is unclear whether there ever exists a common progenitor cell that is capable of giving rise to all of the different cell types within the pancreas or whether specific types of progenitors with a more limited potential (i.e., endocrine, exocrine, duct) already exist at the earliest stages of pancreas development. There is currently a lack of markers that

FIGURE 42.2 Timeline of pancreas development in the mouse. Key events in mouse pancreas development are shown. A, Pancreatic bud evagination can first be detected on embryonic day 9.5. In this image, the Pdx1 expression domain is marked by the brown nuclei, and it includes the antral stomach (as) and the dorsal pancreatic bud (db), which is also marked by a pancreasspecific lacZ transgene (blue). Endocrine differentiation occurs in two waves. The first begins on embryonic day 10.5 and extends to embryonic day 13.5. The second wave begins on embryonic day 13.5 and continues until neonatal stages. B, Acinar gene expression (a; amylase in brown) begins on embryonic day 14.5. Both acinar cells and endocrine cells bud off of the ductal epithelium (d; blue). C, During late gestation, endocrine cells cluster, migrate away from the ductal epithelium, and organize into islets with b cells at the core and other hormone cell types at the periphery. Green, Insulin, red, glucagon. (See color insert.)

THE INITIAL STAGES OF PANCREATIC BUD FORMATION

949

identify progenitor cells within the pancreas akin to those that are known to be in the hematopoietic field, although several laboratories are using genome expression profiling at different stages of pancreas development in an attempt to identify such markers (Chiang and Melton, 2003; Wells, 2003; Gu et al., 2004). Lineage-tracing studies have determined, however, that all pancreatic cell types arise from a cell that expresses the pancreatic duodenal homeobox 1 (pdx1) and Pft1a/p48 transcription factors (discussed later; Gu et al., 2002; Kawaguchi et al., 2002). The expression of endocrine hormones such as glucagon and insulin is detected even at early pancreatic bud stages (embryonic day 10.5); exocrinespecific gene transcription does not commence until embryonic day 14.5 (see Figure 42.2; Gittes and Rutter, 1992). Pancreatic endocrine differentiation actually occurs in two waves during embryogenesis (see Figure 42.2; Pictet et al., 1972; Pang et al., 1994; Prasadan et al., 2002). The first wave occurs between embryonic days 9.5 and 13.5. Unlike second-wave endocrine cells, these early differentiating hormone-producing cells can develop in the absence of the critical pancreatic transcription factor pdx1 (Ahlgren et al., 1996; Offield et al., 1996); they lack other genetic markers of mature islet endocrine cells (Pang et al., 1994; Lee et al., 1999; Wilson et al., 2002), and they have been shown by lineage-tracing analyses to not contribute to mature islets (Herrera et al., 1994; Herrera, 2000). During the second wave of endocrine differentiation, which commences on embryonic day 13.5 in the mouse, the numbers of endocrine cells greatly increase. These endocrine cells go on to populate the mature islets. The mechanism for the increase in endocrine cells at these stages is unknown. The formation of mature, optimally functional islets requires the generation of appropriate numbers of each endocrine cell type, and this process is likely regulated by positive and negative factors that influence cell proliferation and differentiation. In a process that is very reminiscent of neurogenesis in Drosophila and other organisms, islet organogenesis involves the delamination of specified endocrine cells from the ductal epithelium, detachment from the ducts, and the formation of adherent islet clusters (see Figure 42.1; reviewed by Edlund, 2001). As in Drosophila, the specification of endocrine progenitors within the ductal epithelium is dependent on cell–cell communication and lateral inhibition using the Notch signaling pathway (Apelqvist et al., 1999; Jensen et al., 2000; Murtaugh et al., 2003). In the developing pancreatic ductal epithelium, cell–cell interactions involving Notch–Delta signaling determine which cells will initiate the endocrine genetic program by activating Ngn3 expression (Apelqvist et al., 1999; Jensen et al., 2000). Cells expressing higher levels of Notch signaling remain within the epithelium and actively repress Ngn3 expression, whereas those cells in which Delta levels become elevated activate Ngn3, exit from the epithelium, and ultimately give rise to the endocrine population (see Figure 42.1; Gu et al., 2002). In the developing human pancreas, cell surface proteins, including a cell adhesion molecule and certain integrins, have been identified that may mark endocrine progenitor cells within and delaminating from the ductal epithelium (Cirulli et al., 1998, 2000). After delamination, the endocrine cells begin to organize into clusters that are initially still located close to ducts (see Figure 42.1). On embryonic day 18.5, these clusters begin to lose their proximity to the ductal epithelium, they become surrounded by exocrine tissue, and they form mature islets (see Figures 42.1 and 42.2). As islets form, the endocrine cells segregate such that, in mice,

950

PANCREAS DEVELOPMENT AND STEM CELLS

the insulin-producing b cells are located at the core, and glucagon-, somatostatin-, ghrelin- and PP-producing cells are located at the periphery or mantle (see Figures 42.1 and 42.2). Little is known about how the different endocrine cell types and their precursors interact with one another to form functional islets. The processes of endocrine delamination and islet formation likely include changes in the expression of lineage-specific transcription factors, cell adhesion molecules, and extracellular matrix components.

II. INDUCTIVE INTERACTIONS DURING PANCREAS DEVELOPMENT Pancreas development is dependent on an interaction between epithelial (endodermal) and mesodermal components (Edlund, 2002; Kim and MacDonald, 2002; Kumar and Melton, 2003; Wilson et al., 2003). Signals from the notochord have been implicated in pancreas specification and outgrowth (Kim et al., 1997; Hebrok et al., 1998, 2000; Kim and Melton, 1998), whereas pancreatic mesenchyme stimulates the growth of the endodermal epithelium (Wessels and Cohen, 1967; Ahlgren et al., 1997). In turn, the endoderm influences the character of the overlying mesoderm (Slack, 1995; Apelqvist et al., 1997). A. Early Inductive Events in Pancreatic Endocrine Differentiation: The Role of the Notochord Wessels and Cohen (1967) suggested that signals derived from dorsal axial tissue such as the notochord might be involved in inducing the outgrowth of the dorsal pancreatic bud. The notochord transiently contacts the endodermal epithelium directly in the region from which the dorsal pancreatic bud will form during stages that occur before pancreatic bud outgrowth. Experimental manipulations in chick embryos revealed that, in the absence of the notochord, the dorsal pancreatic bud undergoes only limited outgrowth and branching, and it fails to activate the expression of pancreatic transcription factors (e.g., pdx1, Isl-1, Pax6) and of markers of differentiated endocrine or exocrine cells (Kim et al., 1997). By contrast, the outgrowth and differentiation of the ventral pancreatic bud occurs normally in the absence of the notochord. It is currently unclear what tissue interactions promote ventral pancreas development, although genes involved in ventral bud development have been identified (discussed later). Activin bB and fibroblast growth factor (FGF)-2 are likely to be the endogenous notochord-derived signals that induce dorsal pancreas bud outgrowth and differentiation (Hebrok et al., 1998). One of the main functions of notochord-derived factors seems to be the repression of endodermal Sonic hedgehog (Shh) expression in the region that is destined to give rise to the pancreas (Hebrok et al., 1998). Shh is expressed throughout the embryonic gut endoderm with the exception of the dorsal and ventral pancreatic bud endoderm. The transplantation of an ectopic notochord to nonpancreatic regions of the developing gut tube results in decreased Shh expression in the region that is adjacent to the transplant (Hebrok et al., 1998). The maintenance of Shh in the pancreatic field using a transgenic approach results in the impaired development of the pancreatic epithelium and the altered character of the overlying mesoderm such that it expresses markers that are consistent with small intestine smooth muscle (Apelqvist et al., 1997).

INDUCTIVE INTERACTIONS DURING PANCREAS DEVELOPMENT

951

B. Early Inductive Events in Pancreatic Endocrine Differentiation: The Role of Endothelial Cells The vasculature of the pancreas is derived from the mesodermal germ layer. Although islets represent only approximately 2% of the total mass of the pancreas in an adult, they receive up to 15% of the blood flow (Lifson et al., 1980, 1985), likely as a result of their role as endocrine organs secreting hormones directly into the bloodstream. The morphology and architecture of endothelial cells differs among the different capillary beds (LeCouter et al., 2002). Vessels are classified as continuous, fenestrated, or discontinuous. Capillaries in skeletal muscle, heart, lung, and brain have a continuous endothelium, whereas capillaries in endocrine glands such as the islets are fenestrated (LeCouter et al., 2002; Cleaver and Melton, 2003). Thus, in addition to producing angiogenic stimuli for inducing the ingrowth of new vessels, tissues provide factor(s) to direct the differentiated phenotype of the endothelium. For example, early differentiating pancreatic endocrine cells (both first- and second-wave cells) produce angiogenic factors including vascular endothelial growth factor (VEGF) and angiopoietin 1 (Brissova et al., 2006); the expression of these factors is maintained in adult islets (Christofori et al., 1995; Rooman et al., 1997; Lammert et al., 2001). The continued expression of these factors in adult islets suggests that the maintenance of a fenestrated endothelium is critical for mature islet function. In addition to islet endocrine cells communicating with vascular endothelial cells via secreted growth factors, endothelial cells have also been shown to signal to the pancreatic epithelium, thus influencing the differentiation of endocrine cells (Lammert et al., 2003a, 2003b; Yoshitomi and Zaret, 2004). Pancreatic bud outgrowth is initiated at sites in the posterior foregut endoderm, where it contacts the endothelium of major blood vessels; endocrine differentiation initially occurs in cells that have direct contact with endothelial cells (Lammert et al., 2001). On embryonic day 10.5, insulin expression is detected at sites at which the dorsal pancreatic bud contacts portal vein endothelium. The importance of vascular endothelial cells in pancreatic endocrine differentiation has been demonstrated both in tissue recombination experiments and in genetically modified mice. For example, embryonic day 8.5 endoderm cultured in the absence of endothelial cells failed to activate either Pdx1 or insulin protein expression, whereas, when undifferentiated endoderm was cultured in combination with dorsal aorta, both Pdx1 and insulin were induced (Lammert et al., 2001). The examination of VEGF receptor type 2 (VEGFR-2/flk-1) null mutant mice, which die before the second wave of endocrine differentiation, revealed that early insulin- and glucagon-positive cells fail to develop (Lammert et al., 2001; Yoshitomi and Zaret, 2004). These mice express most pancreatic/endocrine transcription factors (pdx1, hnf6, Ngn3, NeuroD, Prox1, and Hb9), with the exception of the early pancreatic bud marker, ptf1a (Yoshitomi and Zaret, 2004). Taken together, these data provide strong support for an endothelial-derived endocrine-inducing factor (or factors). Although the identity of this factor is currently unknown, it follows that, if endothelial cell numbers were to increase, the amount of the inducer (and thus the amount of endocrine cells) would also increase. To this end, the Melton laboratory generated transgenic mice expressing VEGF164 throughout the entire pancreatic bud early during development (Lammert et al., 2001). These transgenic embryos showed greatly increased

952

PANCREAS DEVELOPMENT AND STEM CELLS

vasculature in the pancreas and a corresponding threefold increase in islet number and islet area. Ectopic insulin expression was found adjacent to the ectopically induced endothelial cells. The evidence is increasing that, during both liver and pancreas development, endothelial cells produce an instructive signal that induces differentiation and morphogenesis (Lammert et al., 2001; Matsumoto et al., 2001; Yoshitomi and Zaret, 2004). What might this factor be? Vascular endothelial cells are known to produce several different secreted growth factors, including FGF, transforming growth factor (TGF)-b, Wnt, and hepatocyte growth factor (HGF; Lammert et al., 2003a). In the liver, a VEGFR-1/flt-1 receptor-specific agonist causes endothelial cells to express HGF in a paracrine fashion. HGF has been shown to be mitogenic to b cells (Otonkoski et al., 1994; Hayek et al., 1995; Garcia-Ocana et al., 2000). Thus, in pancreatic endothelial cells, VEGF signaling through VEGFR-1 may induce the expression of HGF, which in turn promotes endocrine proliferation. These studies highlight the reciprocal communication between pancreatic endocrine cells and endothelial cells.

III. GENES THAT AFFECT PANCREATIC BUD DEVELOPMENT Pancreatic development requires factors that act autonomously within the endodermally derived epithelium as well as factors that function within the adjacent associated mesenchyme. Gene inactivation in mice has identified transcription factors that affect the differentiation of all or of a subset of the pancreatic cell types. Factors that regulate the differentiation and function of islet b cells are candidates for susceptibility genes in type 2 diabetes. Indeed, most of the genetic lesions that result in a dominant form of the disease called maturity onset diabetes of the young (MODY) are associated with mutations in transcription factors that are expressed in adult b cells: HNF4a (MODY1), HNF1a (MODY3), pdx1 (MODY4), HNF1b (MODY5), and Beta2/NeuroD (MODY6; Hattersley et al., 2000). Some of these factors play critical roles in distinct stages of pancreatic/endocrine development, including pancreatic bud outgrowth and branching, endocrine differentiation, and/or mature islet function (Gannon and Wright, 1999; Edlund, 2001; Wilding and Gannon, 2004). Promoter analysis of islet-specific genes such as insulin has also helped to identify trans-acting factors that are critical for normal pancreas and/or endocrine development (Madsen et al., 1997; Sander and German, 1997; Edlund, 1998). By understanding how these different transcription factors function during the normal pathway of islet differentiation, one may be able to manipulate this pathway to induce the differentiation of pancreatic stem cells in vivo or to influence the production of functional b cells and/or islets from pancreatic or embryonic stem cells in vitro (Edlund, 2002). This section will summarize some of the results from gene inactivation and/or over-expression studies to determine the role of these factors in pancreas development, endocrine differentiation, and mature b-cell function. A. Endodermally Expressed Genes That Affect Pancreatic Bud Formation 1. Pdx1 The homeodomain transcription factor pdx1 (Ipf1) is one of the earliest known markers of the developing pancreas (see Figures 42.1 and 42.2; Gannon

GENES THAT AFFECT PANCREATIC BUD DEVELOPMENT

953

and Wright, 1999), and lineage-tracing analysis reveals that all cells within the endodermal component of the pancreas are derived from a pdx1þ cell (Gu et al., 2002). pdx1 is expressed early throughout the endoderm of the pancreatic buds as well as by the antral stomach, the rostral duodenum, and the common bile duct. However, it becomes highly enriched in insulin-producing b cells during late gestation, with lower levels of expression found in some acinar cels (see Figure 42.1; Guz et al., 1995; Offield et al., 1996; Wu et al., 1997). Pdx1 binds to and activates the insulin promoter and other b-cell–specific genes such as GLUT2, glucokinase, and islet amyloid polypeptide (Figure 42.3). In both mice and humans, the homozygous inactivation of pdx1 results in pancreatic agenesis, whereas heterozygotes have impaired b-cell function and are prone to diabetes (Jonsson et al., 1994; Offield et al., 1996; Stoffers et al., 1997, 1998; Ahlgren et al., 1998; Dutta et al., 1998). Mutations in pdx1 have also been identified in some patients with type 2 diabetes (Stoffers et al., 1997). In pdx1 null mouse mutants, the dorsal bud undergoes limited proliferation and outgrowth to form a small, irregularly branched ductule (Offield et al., 1996). Transient insulinþ cells and longer-lived glucagonþ clusters are found in the mutant epithelium (Ahlgren et al., 1996; Offield et al., 1996), and these may represent first-wave (immature) endocrine cells. Thus, pdx1 is not absolutely essential for insulin or glucagon expression. pdx1 inactivation specifically in adult b cells results in type 2 diabetes, thus demonstrating a role for pdx1 in the maintenance of b-cell function (Ahlgren et al., 1998). 2. Ptf1a Ptf1a (p48) is the pancreas-specific component of the heterotrimeric basic helix–loop–helix (bHLH) complex pancreas transcription factor 1 (PTF1; Cockell et al., 1989; Krapp et al., 1996). The two other components of the PTF1 complex are the constitutively expressed HeLa E-box binding factor (Cockell et al., 1989), and the mammalian Suppressor of Hairless (RBP-J; Obata et al., 2001) or its paralog RBP-L (Beres et al., 2006). Although it was first identified as a regulator of exocrine-specific genes (Krapp et al., 1996; Rose et al., 2001), PTF1 was subsequently shown to be essential early

Simplified pancreas transcription factor network. Some of the factors that are important in the specification of the endocrine lineage and subsequent differentiated b cells are shown. In particular, the interactions between different Maturity onset diabetes of the young genes, transcriptional targets of Hnf6 and Pdx1, and transcription factors that transactivate the insulin promoter are highlighted. Arrows indicate direct transcriptional targets.

FIGURE 42.3

954

PANCREAS DEVELOPMENT AND STEM CELLS

during development for normal endocrine and exocrine pancreas formation in both mice and humans (Krapp et al., 1998; Obata et al., 2001; Kawaguchi et al., 2002; Sellick et al., 2004). Mice lacking Ptf1a have no detectable ventral bud outgrowth; the dorsal bud initiates but then arrests as a duct-like structure that lacks differentiated acinar cells (Krapp et al., 1998; Kawaguchi et al., 2002). Small endocrine clusters that contain insulin and glucagon cells are present within this structure, and a few isolated hormoneþ endocrine cells can be detected within the spleen. These may represent first-wave endocrine cells. The results of Ptf1a gene inactivation were supported by analyses showing that Ptf1a expression is detected as early as embryonic day 9.5 throughout the developing pancreatic buds (see Figure 42.1; Kawaguchi et al., 2002). Lineage-tracing studies further revealed that acinar, ductal, and endocrine cells are all derived from a Ptf1a-expressing cell (Kawaguchi et al., 2002). In zebrafish, Ptf1a is required for the endocrine and exocrine tissue that arises from the anterior (dorsal) bud, but it is not required for posterior (ventral) bud formation and outgrowth, which differentiates into endocrine tissue (Lin et al., 2004). 3. Hlxb9 Hlxb9 is the gene that encodes the Hb9 homeodomain protein, which is prominently expressed in adult islet b cells. Hb9 is expressed transiently throughout the prepancreatic epithelium during the early stages of mouse pancreatic development (embryonic days 8 to 10.5) in both the dorsal and the ventral anlagen coincident with Pdx1 and Ptf1a expression (see Figure 42.1), and it is reexpressed later (on embryonic day 17.5) in differentiated b cells (Li et al., 1999). In the absence of Hlxb9 expression, the dorsal bud of the pancreas fails to develop (Harrison et al., 1999; Li et al., 1999). By contrast, the ventral pancreatic endoderm develops, and it forms both endocrine and exocrine tissues; however, the endocrine cells within the islets are disorganized, and they have reduced numbers of insulin-producing cells that demonstrate the reduced expression of some markers of terminal b-cell differentiation. 4. Hex The Hex homeobox gene is expressed in the ventral posterior foregut in the region that will give rise to both the liver bud and the ventral (but not the dorsal) pancreatic bud. The examination of Hex null embryos reveals a requirement for Hex in the proliferation and outgrowth of the liver bud and in the specification of the ventral pancreas (Bort et al., 2004). In the absence of Hex, the ventral pancreatic bud does not form, and it fails to express Pdx1, Ptf1a, and Hlxb9. Dorsal pancreas specification appears to be normal in these mutants. Interestingly, explants of presumptive ventral pancreas endoderm differentiate normally in the absence of Hex, expressing Pdx1 and endocrine lineage markers such as Isl-1, Ngn3, and Beta2. These results suggest that Hex null endoderm is fully competent to differentiate according to its normal fate but that it is susceptible to influences from surrounding tissues within the intact embryonic environment that prevent ventral pancreas specification and differentiation. Indeed, when cocultured with cardiogenic mesoderm, Hex / ventral endoderm no longer expresses any pancreatic markers (Bort et al., 2004).

GENES THAT AFFECT THE DIFFERENTIATION OF PARTICULAR PANCREATIC CELL TYPES

955

B. Mesodermally Expressed Genes that Affect Pancreatic Bud Formation 1. Isl-1 The LIM homeodomain transcription factor Islet-1 (Isl-1) was identified as a factor that binds to and transactivates the insulin gene promoter. It is expressed in all islet endocrine cell types in the embryo (see Figure 42.1) and the adult and in the mesodermally derived mesenchyme surrounding the dorsal (but not the ventral) pancreatic bud (Ahlgren et al., 1997). Similar to the phenotype found in Hlxb9 mutant pancreata, the dorsal pancreatic bud does not develop and differentiate in Isl-1 null embryos (Ahlgren et al., 1997). The ventral bud grows normally, but it fails to produce any endocrine hormoneþ cell types; acinar cell differentiation appears to be normal. Unlike Hlxb9 mutant pancreata, in which the dorsal mesenchyme remains intact, the absence of dorsal bud derivatives in Isl-1 null embryos is specifically the result of the loss of dorsal mesenchyme, because mesenchyme from wild-type embryos can rescue the differentiation of exocrine cells from Isl-1 mutant dorsal pancreatic endoderm in culture. Importantly, hormoneþ cells are not rescued by wild-type mesenchyme, which suggests that Isl-1 has a non–cell-autonomous role in the mesenchyme for dorsal bud outgrowth and a cell-autonomous role in the endoderm for endocrine differentiation. These results underscore the requirement for the pancreatic mesenchyme in pancreas development. 2. N-cadherin During the early stages of pancreas development (embryonic days 9.0 to 12.5), the cell adhesion molecule N-cadherin is expressed in both the pancreatic epithelium and the surrounding mesenchyme. After embryonic day 12.5, expression becomes restricted to the formation of clusters of endocrine cells. In embryos that lack N-cadherin, the dorsal pancreatic bud fails to form, although genes such as Hlxb9 and Isl-1 are expressed normally, and Shh expression is repressed in the dorsal endoderm (Esni et al., 2001). N-cadherin function is not required within the pancreatic endoderm, because coculture with wild-type mesenchyme can rescue branching morphogenesis, exocrine, and endocrine differentiation. In addition, restoring cardiac and circulatory function in N-cadherin null mice by the cardiac-specific transgenic expression of N-cadherin rescues the formation of the dorsal pancreas (Edsbagge et al., 2005). On the basis of this observation, it was proposed that soluble factors present in plasma are critical for the formation and/or maintenance of the dorsal pancreatic mesenchyme. It was found that sphingosine 1-phosphate present in plasma promotes the budding of the pancreatic endoderm by stimulating pancreatic mesenchymal cell proliferation and that sphingosine 1-phosphate receptors are located within the mesenchyme (Edsbagge et al., 2005).

IV. GENES THAT AFFECT THE DIFFERENTIATION OF PARTICULAR PANCREATIC CELL TYPES A. Genes Involved in Exocrine Differentiation There is actually very little known about factors that act to specify the exocrine cell population from progenitors within the ductal epithelium. It was previously thought that cells that fail to activate Notch became endocrine

956

PANCREAS DEVELOPMENT AND STEM CELLS

cells, whereas those that activate the Notch pathway became exocrine cells. It is now clear that Notch activation maintains cells in a proliferative, progenitor state and that other signals are required to generate acinar cells or endocrine cells from this pool. The activation of Notch has recently been shown to inhibit exocrine differentiation in both the mouse and the zebrafish (Hald et al., 2003; Murtaugh et al., 2003; Esni et al., 2004). 1. Pdx1 and Ptf1a Despite the severe defects in pancreas development that occur in the absence of either Pdx1 or Ptf1a, there are some differentiated endocrine cells detected in both mutants. By contrast, there is a complete absence of exocrine tissue in both the Pdx1 and Ptf1a null mutant animals, which highlights the fact that these genes are essential for exocrine development. As mentioned previously, Ptf1a is highly enriched in acinar cells after embryonic day 13.5, and the PTF1 heterotrimeric transcription factor complex has been shown to bind to several exocrine-specific gene promoters (Krapp et al., 1996; Rose et al., 2001). The loss of RBP-J, which is another component of the PTF1 complex, results in the accelerated differentiation of endocrine cells (Apelqvist et al., 1999), thus supporting the idea that a functional PTF1 is required for acinar differentiation. Pdx1, although it is expressed at much lower levels in mature acinar cells as compared with b cells, has also been found to regulate acinar gene expression in cooperation with Pbx1 and Meis2 (Swift et al., 1998). The maintenance of pdx1 expression in the exocrine lineage during embryonic development is required for acinar differentiation (Hale et al., 2005). 2. Mist1 The Mist1 bHLH transcription factor is expressed in acinar cells, and it activates the acinar genes that are involved in gap junction communication and coordinated exocytosis. Mist1 null mutant mice exhibit the extensive disorganization of exocrine tissue and defects in the regulated exocytosis pathway, which results in inappropriate intracellular enzyme activation. These changes mimic those observed in pancreatic injury, such as those seen with chronic pancreatitis. Thus, it has been proposed that Mist1 is a key regulator of acinar cell function, stability, and identity (Pin et al., 2001; Rukstalis et al., 2003; Johnson et al., 2004). B. Genes Involved in General Endocrine Specification and the Differentiation of Particular Lineages Although all islet endocrine cells express some common factors that promote endocrine specification and differentiation, it is likely that each particular islet endocrine cell type is specified by a different combination of transcription factors. Many of the lineage-restricted or lineage-specific transcription factors are actually expressed more broadly in the pancreatic epithelium or in the endocrine population early during development; they become gradually restricted as development proceeds to refine the pattern of gene expression to what is observed in the adult islet. Gene-expression and mutational analyses in mice have proven to strongly correlate with gene function in humans, because mutations in many of the genes that will be discussed later have been identified in individuals with type 2 diabetes, including Pdx1, Pax6, and Beta2.

GENES THAT AFFECT THE DIFFERENTIATION OF PARTICULAR PANCREATIC CELL TYPES

957

1. Prox1 The homeobox gene Prox1 is expressed in the posterior foregut endoderm in the presumptive pancreas region before bud outgrowth (Burke and Oliver, 2002), and it has been shown to be essential for normal liver bud outgrowth (Sosa-Pineda et al., 2000). On embryonic day 13.5, Prox1 is expressed in most cells throughout the pancreatic epithelium. As the second wave of endocrine differentiation commences after embryonic day 13.5, Prox1 becomes more highly expressed in NGN3þ (a marker of committed endocrine progenitor) and Isl-1þ (a marker of all islet endocrine) cells, and it is downregulated in differentiating acinar cells. After birth, Prox1 expression is maintained at high levels in the ductal epithelium and in peripheral islet cell types, with lower levels found in b cells. On embryonic day 15.5 (the time at which Prox1deficient embryos die as a result of complications in other organ systems), the Prox1 mutant pancreatic epithelium is less branched than that of wild type, and it contains many fewer endocrine cells (Wang et al., 2005). By contrast, the number of differentiated acinar cells is relatively increased, and the pancreas has increased levels of Ptf1a with decreased levels of markers of endocrine lineage (e.g., Ngn3). Thus, Prox1 may be required within a bipotential acinar/endocrine pancreatic progenitor to promote differentiation down the endocrine lineage. 2. HNF6 Hepatic nuclear factor 6 (HNF6/Onecut-1), as the name implies, was first identified in the liver, but it is actually more broadly expressed in the developing endoderm. Target genes for HNF6 include Foxa2, Pdx1 (MODY4), and HNF4 (MODY1), which are critical endodermal regulators: Foxa2 is involved in the b-cell–specific expression of Pdx1, and HNF4 activates HNF1a (MODY3). In turn, HNF1a activates Pdx1, and HNF4 regulates HNF6 (see Figure 42.3; reviewed by Jensen, 2004). Thus, alterations in the expression of a single HNF or MODY gene can affect the expression of multiple genes in this hierarchy. One of the interesting things about HNF6 is its dynamic expression pattern with the pancreas. Similar to Pdx1 and Ptf1a, HNF6 is initially expressed throughout the early pancreatic epithelium. As development proceeds, HNF6 is maintained in the ductal epithelium and in acinar cells, but it becomes downregulated—specifically in endocrine cells by embryonic day 18.5 (Landry et al., 1997; Rausa et al., 1997). The decreased expression of HNF6 in islet endocrine cells coincides with islet morphogenesis and b-cell maturity in preparation for birth (see Figure 42.2). This downregulation is critical for normal islet ontogeny and function: continued HNF6 expression in islets impairs the separation of endocrine cells from the ductal epithelium, disrupts the organization of endocrine cell types within the islet (core vs. mantle), and severely compromises b-cell physiology, thus leading to overt diabetes (Gannon et al., 2000; Tweedie et al., 2006). Despite its early broad expression pattern in the pancreatic buds, HNF6 function is not required to generate a pancreas. In the absence of HNF6, Pdx1 gene activation and pancreas bud outgrowth are delayed, and this results in a slightly hypoplastic pancreas at birth (Jacquemin et al., 2003). HNF6 is an upstream activator of Ngn3 (see Figure 42.3; Jacquemin et al., 2000), which is a transcription factor that is expressed in endocrine precursors, and it is a transcriptional target of the Notch–Delta signaling pathway.

958

PANCREAS DEVELOPMENT AND STEM CELLS

Ngn3 / mice lack endocrine cells in both the pancreas and the small intestine (Gradwohl et al., 2000). HNF6 / mice have decreased numbers of Ngn3þ cells during embryogenesis, and they lack islets at birth (Jacquemin et al., 2000). Islets do develop later, but they are abnormal, and the mice are glucose intolerant. The presence of islets in the absence of HNF6 suggests that other factors partially compensate for HNF6, and, indeed, other closely related factors are also expressed in the developing pancreas (Jacquemin et al., 2003). HNF6 function is therefore required to generate endocrine progenitors in the appropriate numbers. 3. Ngn3/NeuroD There are two closely related bHLH transcription factors that are involved in the early stages of pancreatic endocrine development: Ngn3 and Beta2/NeuroD (MODY6). Ngn3-expressing cells are first detected scattered throughout the pancreatic epithelium on embryonic day 9.5. Their numbers reach a peak on embryonic day 15.5 and then decrease to nearly undetectable levels by birth (Gradwohl et al., 2000). Ngn3þ cells are found within or adjacent to the ductal epithelium, and they do not coexpress any of the islet endocrine hormones (see Figure 42.1). In the absence of Ngn3, endocrine markers within the pancreas (including both broad and lineage-restricted transcription factors and all hormones) are missing (Gradwohl et al., 2000). Thus, Ngn3 represents the only known gene that specifically marks the endocrine progenitor population. Unfortunately, early attempts to increase endocrine differentiation and total islet mass by overexpressing Ngn3 in vivo have not proved fruitful. In transgenic animals, the expression of Ngn3 throughout the pancreatic epithelium results in a large increase in the number of glucagon-producing cells, with little to no b cells being formed (Apelqvist et al., 1999; Schwitzgebel et al., 2000). Similarly, the forced expression of Ngn3 in chick foregut endoderm yields only glucagonþ cells (Grapin-Botton et al., 2001). Interestingly, the endocrine phenotype observed after the over-expression of Ngn3 is identical to that seen in Hes1-deficient animals. In the absence of Hes-1, which normally represses Ngn3 transcription in response to Notch signaling, Ngn3 is over-expressed, thereby leading to an excess of glucagon-producing cells and the depletion of endocrine progenitors (Jensen et al., 2000). Beta2 was isolated as a transactivator of the insulin gene (see Figure 42.3), but it is actually expressed in all islet endocrine cells types during development and in the adult. It is a direct transcriptional target of Ngn3 (see Figure 42.3). The loss of Beta2 results in a dramatic decrease in all islet endocrine cell types, which suggests that Beta2 functions in the expansion of the endocrine population or in endocrine cell survival (Naya et al., 1997). The remaining endocrine cells fail to organize into normal spherical islet structures, which suggests that Beta2 also functions in islet morphogenesis. 4. Nkx2.2/6.1 Members of the NKX class of homeodomain proteins also have roles in the pancreatic endocrine lineage. Both Nkx2.2 and Nkx6.1 are expressed in most pancreatic epithelial cells during early stages of development; however, by embryonic day 15.5, Nkx2.2 becomes restricted to the endocrine cell population, and Nkx6.1 is found exclusively in insulin-producing cells and scattered cells within the ductal epithelium (Sussel et al., 1998; Sander et al.,

GENES THAT AFFECT THE DIFFERENTIATION OF PARTICULAR PANCREATIC CELL TYPES

959

2000). During late gestation, Nkx2.2 can be detected in nearly all hormoneþ cells, except for the somatostatin-producing cells. After birth, both genes are found to be expressed in the b-cell population. Mice lacking the Nkx2.2 gene have no detectable insulinþ cells at any stage that has so far been examined, and they also have a dramatic reduction in the number of glucagonexpressing cells and a more modest reduction in the number of PPþ cells (Sussel et al., 1998). The a and PP cells, although fewer in number, express other known markers of these lineages, which suggests that they are terminally differentiated. The expression of Isl-1 and synaptophysin, which are general markers of islet endocrine cells, is normal in the Nkx2.2 mutants; this suggests that the loss of Nkx2.2 does not result in a dramatic loss of endocrine cells in general. Subsequent analysis has revealed that these “extra” endocrine cells in the Nkx2.2 knockout pancreas were increased numbers of the ghrelinproducing e-cell population (Prado et al., 2004). Thus, Nkx2.2 is required for the generation of b cells and for the maintenance and expansion of a and PP cells. Nkx6.1 gene inactivation results in a highly specific profound loss of second-wave insulinþ cells (after embryonic day 13.5), with no alteration in the numbers of other islet endocrine cell types (Sander et al., 2000). Thus, in the absence of Nkx6.1, putative b cells do not adopt an alternate islet endocrine cell fate. Genetic epistasis experiments have demonstrated that Nkx6.1 functions downstream of Nkx2.2 in the expansion and terminal differentiation of the b-cell lineage (Sander et al., 2000). 5. GDF11 Growth and differentiation factor 11 (GDF11) is a member of the TGF-b family of secreted growth factors that is expressed in the embryonic pancreas throughout the epithelial component between embryonic days 11.5 and 13.5 (Harmon et al., 2004). During late gestation, GDF11 becomes restricted to acinar cells, and a loss of GDF11 results in a slight decrease in acinar cell mass. This was accompanied by a dramatic increase in the total number of endocrine cells in mutant animals as assessed by the general endocrine marker synaptophysin, which suggests an increased allocation of pancreatic progenitors to the endocrine lineage. Further analysis revealed an increase in the number of Ngn3þ cells earlier during development (embryonic day 11.5) and continuing until late gestation, possibly reflecting both precocious and increased endocrine specification. Similar to the results of the transgenic over-expression of Ngn3, increased Ngn3 expression in GDF11 mutants causes an increase in allocation to the a-cell lineage at the expense of b-cell differentiation. 6. Pax4/6 Two members of the paired class of homeodomain-containing transcription factors have been shown to function in pancreatic endocrine differentiation: Pax4 and Pax6. In the pancreas, Pax4 is specifically expressed in both first- and second-wave insulin-producing cells during embryonic development, and it is maintained in adult b cells (Sosa-Pineda et al., 1997). In the absence of Pax4, b and d cells fail to differentiate (lacking Pdx1, Hb9, and insulin expression), and there are increased numbers of glucagon- and ghrelin-producing cells (Sosa-Pineda et al., 1997; Prado et al., 2004). These data suggest

960

PANCREAS DEVELOPMENT AND STEM CELLS

that either Pax4 is separately required in the b- and d-cell lineages or that these two islet cell types arise from a common progenitor that is dependent on Pax4. In addition, the increased numbers of a and e cells suggest that cells that would have become b or d cells have instead adopted one of these two cell fates or, alternatively, that b and d cells normally produce something that inhibits the expansion of the a- and e-cell populations. In contrast with Pax4, Pax6 is expressed in all endocrine cell types within the pancreas both during embryonic development and in adults; however, the global loss of Pax6 has a specific affect on the a-cell lineage. In the absence of Pax6, there is a complete loss of glucagon-producing cells; the other endocrine cell types are present in reduced numbers, and they fail to organize into normal islet structures (Sander et al., 1997; St-Onge et al., 1997). Mice that lack both Pax4 and Pax6 have a complete loss of all pancreatic endocrine cell types (St-Onge et al., 1997). After birth, Pax6 also functions in the maintenance of the differentiated b-cell phenotype; gene inactivation in the mature endocrine population has no effect on cell number, but it results in diabetes and the decreased expression of some b-cell–specific genes (Ashery-Padan et al., 2004). 7. MafA/MafB The large Maf proteins are basic leucine zipper transcription factors that were first identified in an avian retrovirus. MafA was identified by several independent groups as an activator of insulin gene transcription (see Figure 42.3; Olbrot et al., 2002; Kajihara et al., 2003; Matsuoka et al., 2003; Kataoka et al., 2004; Zhao et al., 2005). MafA is specifically expressed in second-wave insulinþ cells beginning on embryonic day 13.5 and continuing into adulthood, thus making it a marker of more mature b cells (Matsuoka et al., 2004; Nishimura et al., 2006; Tsuchiya et al., 2006). Despite its indication as a critical b-cell factor, the global inactivation of MafA had no effect on the number of insulin-producing cells during embryonic development. Instead, the loss of MafA causes defects in b-cell gene expression and postnatal b-cell function, thus leading to diabetes (Zhang et al., 2005). The lack of a developmental islet phenotype in the MafA knockout animals may be the result of compensation by another closely related Maf family member that is also expressed in developing endocrine cells, MafB (Artner et al., 2006). MafB is also capable of activating insulin reporter gene transcription in tissue culture cells, although, in adult islets, it is expressed only in a cells, where it regulates the expression of the glucagon gene (Artner et al., 2006). During embryonic development, MafB is expressed in some Ngn3þ cells and in both first and second wave insulin- and glucagon-producing cells, becoming restricted to a cells soon after birth. Loss of MafB results in a dramatic decrease in mature a and b cells (Artner et al., 2007). 8. Brn4 Brain-4 (Brn4) is a POU-homeodomain–containing protein that is expressed in the developing pancreas, specifically in glucagon-producing cells beginning on embryonic day 10 and continuing into adulthood. However, no defect in a-cell specification or differentiation has been observed in Brn4 mutant animals (Heller et al., 2004). The over-expression of a Brn4 transgene in b cells in vivo results in the coexpression of glucagon in insulinþ cells (Hussain et al., 2002).

PANCREAS/b-CELL REGENERATION AND NEOGENESIS

961

9. Arx The Arx homeodomain–containing protein is expressed in scattered cells throughout the pancreatic buds between embryonic days 10.5 and 12.5. The inactivation of Arx causes a complete loss of the second wave of glucagonproducing a cells, and this results in severe postnatal hypoglycemia and death, because glucagon stimulates the liver to deposit glucose into the bloodstream during times of fasting (Collombat et al., 2003). The decrease in a cell number is accompanied by an increase in both somatostatin-producing d cells and b cells. Subsequent analyses strongly suggest that Arx / cells are diverted from an a-cell fate toward a b- or/d-cell fate instead. Indeed, the b-cell transcription factor Pax4 is upregulated in Arx mutants, whereas Arx is upregulated in Pax4 mutants. Thus, these two genes have opposing actions within the endocrine lineage to establish b or d cells (Pax4) and a cells (Arx).

V. PANCREAS/b-CELL REGENERATION AND NEOGENESIS Diabetes results from an absolute (type 1) or relative (type 2) inadequate functional b-cell mass. Thus, the genes and pathways involved in maintaining or altering b-cell mass are candidates for being affected in diabetic individuals. The functional analysis of these genes may lead to new therapeutic strategies for increasing existing b-cell mass in diabetic patients, and it may facilitate the production of b cells in vitro from embryonic or stem cells. A. Factors That Affect b-Cell Mass The mass of the b-cell has been shown in animal models to remain stable for the first few weeks after birth and then to gradually increase throughout the life of the organism (Scaglia et al., 1997; Bonner-Weir, 2000a, 2000c), and this is also thought to be true for humans (Butler et al., 2003). The endocrine pancreas undergoes substantial remodeling during the neonatal period, including increased apoptosis and neogenesis with progressive decreases in b-cell replication (Scaglia et al., 1997). The b-cell mass is dynamic, and there is much experimental evidence to show that the b-cell population has the potential to adapt to changing physiologic needs and increased functional demands (Bonner-Weir, 2000a, 2000b). In most situations, b-cell mass increases or decreases in accordance with metabolic demands; for example, b-cell mass increases during pregnancy and with the insulin resistance associated with obesity, whereas it decreases after parturition and after insulinoma implantation (Bernard-Kargar and Ktorza, 2001). Two types of compensation occur after increased demand on the b cells: improved function of individual cells and increased b-cell mass. Functional adaptations include a reduced threshold for glucose-stimulated insulin secretion and increased glucokinase activity, both of which lead to enhanced insulin secretion (Liu et al., 2000). Changes in b-cell mass are achieved by both hyperplasia (an increased number of cells) and hypertrophy (an increased individual cell size). The adult b-cell population has a slow turnover. At any given time, the number of b cells is determined by the balance of newly forming b cells (via the replication of existing cells and neogenesis from undifferentiated progenitor cells) and b-cell loss through apoptosis. It is estimated that there are 1% to 4% new b cells per day (Finegood et al., 1995). Thus, in the absence of

962

PANCREAS DEVELOPMENT AND STEM CELLS

apoptosis, the b-cell number would double in about a month. Indeed, between months 1 and 3 in the rat, the b-cell number doubles each month. After 3 months, the normal turnover of b cells approaches the replication rate, and, thus, the doubling of b-cell mass does not continue. Although it is generally assumed that pancreatic duct cells retain the ability to generate endocrine cells and form new islets, even late in life, the most recent data available indicate that most if not all new adult b cells arise from preexisting b cells with little to no contribution from stem cells or undifferentiated progenitors (Dor et al., 2004). Regardless of the source of new cells, the endocrine pancreas should be considered a slowly renewing tissue, although the ability of endocrine cells to undergo cell division decreases with age. In the normal adult pancreas, the presence of local factors such as TGF-b maintain the ducts in a quiescent state (Alvarez and Bass, 1999; Crisera et al., 2000); however, under certain conditions (e.g., pancreatic injury), the proliferation of the common duct is induced, and new lobes of pancreas, including endocrine cells, are formed. Glucose is one of the best stimuli for b-cell replication both in vitro and in vivo (Swenne and Andersson, 1984; Path et al., 2004). Thus, in a normally functioning pancreas, sustained elevations in blood glucose levels should lead to increased b-cell mass, thereby providing compensation for the increased glucose load. Autopsies of human patients reveal a 40% increase in b-cell mass in obese individuals, which suggests that b-cell compensation does indeed occur with increasing insulin resistance (Butler et al., 2003). Defects in b-cell mass compensation in all probability contribute to type 2 diabetes (Bernard-Kargar and Ktorza, 2001), but trying to identify the complex array of genes that affect this process is likely to be difficult. Any gene product that affects the renewal, proliferation, or turnover of b cells would be a candidate for genes involved in the cause of diabetes. Several secreted factors and their receptors have been shown to play a role in b-cell mass dynamics. For example, gut hormones such as glucagon-like peptide-1 (GLP-1), gastrin, and cholecystokinin have been shown to be important during b-cell neogenesis, regeneration after injury, and differentiation. The long-acting GLP-1 analog exendin 4 stimulates both b-cell neogenesis from ductal progenitors and the proliferation of existing b cells (De Leon et al., 2003). GLP-1–like compounds are currently being used for the treatment of type 2 diabetes in an attempt to enhance b-cell mass and function (Briones and Bajaj, 2006). In addition, several studies have demonstrated a role for the epidermal-growth-factor–related ligand betacellulin (Huotari et al., 1998) in the differentiation and proliferation of native b cells in vivo after partial pancreatectomy and in organ culture (Demeterco et al., 2000; Li et al., 2001, 2004; Huotari et al., 2002). Betacellulin in combination with the TGF-b family member activin A has been shown to convert ductal, acinar, and a-cell lines to insulin-producing b cells capable of secreting glucose in a regulated manner (Mashima et al., 1996; Watada et al., 1996; Ogihara et al., 2003; Ogata et al., 2004). Activin likely plays a role in b-cell specification and/or proliferation in vivo as well; defects in activin-receptor signaling during embryogenesis result in islet hypoplasia (Shiozaki et al., 1999; Kim et al., 2000). HGF is a mesenchyme-derived growth factor that stimulates the proliferation of both fetal and adult islets in culture (Otonkoski et al., 1994; Hayek et al., 1995), and, in combination with activin A, it is capable of converting the acinar cell line AR42J into insulin-producing cells (Mashima et al., 1996). It acts on epithelial cells through a membrane-spanning tyrosine

PANCREAS/b-CELL REGENERATION AND NEOGENESIS

963

kinase receptor, the product of the protooncogene c-met (Sonnenberg et al., 1993), which is highly expressed in b cells (Watanabe et al., 2003). In transgenic mice, the overexpression of HGF (specifically in insulin-producing cells using the rat insulin II promoter) leads to increased b-cell proliferation, and it protects b cells from apoptosis (Garcia-Ocana et al., 2000). It has also been shown that HGF can enhance islet graft survival and function in both the liver and the kidney (Garcia-Ocana et al., 2003; Lopez-Talavera et al., 2004). The adenoviral delivery of HGF to isolated mouse islets in culture markedly improved b-cell survival and proliferation. B. Evidence for Pancreas/b-Cell Regeneration Data regarding the transcription factors involved in the promotion of b-cell proliferation and/or neogenesis are lacking. In several models of pancreas regeneration and ductal metaplasia, the upregulation of Pdx1 is associated with increased ductal proliferation and increased islet neogenesis (Sharma et al., 1999; Song et al., 1999); however, a requirement for Pdx1 in these processes has not yet been demonstrated. By contrast, the winged helix transcription factor FoxO1 functions normally to inhibit b-cell proliferation and Pdx1 expression (Kitamura et al., 2002). Thus, the inhibition of FoxO1 activity in b cells is required for proliferation. The mammalian pancreas (including that of humans) has significant regenerative potential after insult or injury, although not to the same extent as the liver (Tsiotos et al., 1999; Risbud and Bhonde, 2002). Neogenesis from pancreatic stem cells has been reported to occur in several models of pancreas regeneration, including after b-cell destruction using chemical toxins such as alloxan or streptozotocin (McEvoy and Hegre, 1977; Guz et al., 2001; Risbud and Bhonde, 2002), after the induction of pancreatitis (Gress et al., 1994; Risbud and Bhonde, 2002; Taguchi et al., 2002), after cellophane wrapping (Wang et al., 1995), after partial pancreatectomy (Bonner-Weir et al., 1993; Liu et al., 2000), and after the targeting of inflammatory cytokines to the b cell (Gu and Sarvetnick, 1993). The replication of preexisting endocrine and exocrine cells is increased three- to fourfold after partial pancreatectomy. Although some of the restoration of b-cell mass is the result of the hypertrophy and hyperplasia of the remaining b cells, most of the new b-cell mass has been proposed to arise from ductal cells in a similar manner to what occurs during pancreas development in the embryo (Bonner-Weir et al., 1993). After partial pancreatectomy, there is a proliferation of ductal epithelium (in which progenitors are thought to reside) and the formation of new ductal complexes. Small ductules differentiate into new pancreatic islets and exocrine tissue, thereby forming new lobes of pancreatic tissue that resemble unoperated pancreata. These studies provide support for the existence of a stem-cell–like population in the adult pancreas (Holland et al., 2004), although the presence of dormant stem cells similar to satellite cells in muscle or of facultative stem cells that are activated in response to certain stimuli in the pancreas has not been proven. The regeneration seen with partial pancreatectomy seems to recapitulate the pathway of embryonic pancreas development, including increased translation of the critical pancreatic factor Pdx1 in the ductal epithelium after a wave of increased proliferation (Sharma et al., 1999). Ductal proliferation during pancreas regeneration is also accompanied by the increased expression of

964

PANCREAS DEVELOPMENT AND STEM CELLS

the GLUT-2 glucose transporter, which is normally only expressed in fetal ducts and mature b cells (Wang et al., 1995). C. Genes That Affect b-Cell Proliferation Progression through the cell cycle requires the activity of the heterodimeric cyclin/cyclin-dependent kinase (CDK) complexes. Progression from G1 phase to S phase is mediated by the D class of cyclins and their partners, CDK4 and CDK6. This particular complex is responsible for the phosphorylation of the retinoblastoma protein, which renders it inactive and frees up transcription factors that allow for cell cycle progression. The b cells seem to be particularly sensitive to the loss of certain cell cycle genes. For example, several recent publications suggest a selective role for CDK4 and cyclin D2 in postnatal bcell proliferation (Rane et al., 1999; Martin et al., 2003; Mettus and Rane, 2003; Georgia and Bhushan, 2004). CDK4 null mutant mice have a 50% reduction in body and organ size, but they develop surprisingly normally considering the fact that this gene regulates the passage of cells from G1 phase to S phase. The main defects in these mice are infertility as a result of the loss of pituitary lactotrophs and diabetes (when they are  2 months old) as a result of decreased postnatal b-cell proliferation and a gradual loss of b-cell mass with age (Rane et al., 1999; Mettus and Rane, 2003). CDK4 mutants are born with the appropriate number of b cells, which demonstrates that this gene is dispensable for the formation of the endocrine pancreas during embryonic development. Likewise, cyclin D2 mutant animals are born with b-cell mass that is equivalent to that of their control littermates, but they show a decline in b-cell mass beginning when they are 2 weeks old as a result of decreased postnatal b-cell proliferation (Georgia and Bhushan, 2004; Kushner et al., 2005). These mice become diabetic by the time that they are 3 months old. It has recently been shown that the Foxm1 winged helix transcription factor, which regulates a number of cell cycle genes including several cyclins and cdc25B (Leung et al., 2001; Wang et al., 2002; Costa et al., 2003), is also dispensable for embryonic stages of pancreas and islet development, but it is essential for postnatal b-cell replication and the maintenance of b-cell mass (Zhang et al., 2006). Pancreas-specific inactivation of Foxm1 using a Crelox strategy results in diabetes by 9 weeks of age as a result of a gradual loss of b cells postnatally. Taken together, these results lead to the provocative suggestion that Foxm1 may be involved in tissue regeneration in general and that maintaining Foxm1 in any cell type would prevent the decline in cell proliferation that occurs with age. Thus, redundant or parallel pathways likely exist during embryogenesis that ensure the generation of appropriate numbers of b cells, whereas mature b cells are highly susceptible to perturbations in cell cycle gene expression. A possible explanation for this is the fact that, although many cell types express both CDK4 and CDK6, pancreatic b cells express only CDK4 (Martin et al., 2003). The ability to activate these cell cycle genes or to prevent the age-dependent decline in their expression may facilitate the expansion of bcell mass in vivo or in vitro, or it may increase the proliferation of b cells in isolated islets before or after transplantation. Indeed, the expression of an activated form of CDK4 in islet b cells using the rat insulin promoter results in b cell hyperplasia and improved insulin secretion without hypoglycemia and without the formation of insulinomas (Hino et al., 2004; Marzo et al.,

PANCREAS/b-CELL REGENERATION AND NEOGENESIS

965

2004). Similarly, human islets transduced with a lentivirus expressing activated CDK4 show increased b-cell proliferation (Marzo et al., 2004). D. Generating Islets/b Cells from Stem or Progenitor Cells: The Challenges There are several potential sources for the large number of insulin-producing cells that are needed to make “islet transplantation” an option for a greater number of individuals with diabetes (Figure 42.4), including the following: (1) the proliferation and expansion of existing b cells in vivo or in vitro; (2) the proliferation and expansion of cadaver-derived islets; (3) the induction of b-cell differentiation from endogenous progenitors (embryonic ductal cells) or from adult ductal epithelium; (4) the induction of b-cell differentiation from embryonic stem cells (ESCs); and (5) the transdifferentiation of closely related cell types such as acinar, liver, and intestinal enteroendocrine cells (this concept is not discussed further in this chapter). All of these avenues are experimentally feasible, and they are currently being examined in animal models (Cirulli et al., 1998, 2000; Bonner-Weir et al., 2000; Cheung et al., 2000; Ramiya et al., 2000; Rooman et al., 2000; Lumelsky et al., 2001; Trivedi et al., 2001; Hori et al., 2002; Blyszczuk et al., 2003; Horb et al., 2003; Kahan et al., 2003; Miyatsuka et al., 2003; Blyszczuk et al., 2004; Cao et al., 2004; Koizumi et al., 2004; Lardon et al., 2004; Nakajima-Nagata et al., 2004; Imai et al., 2005). It may be that more than one of these methods will ultimately be used to derive a steady supply of insulin-producing cells. All of these strategies, with the exception of the directed differentiation of ESCs, require access to or the procurement of particular tissues or cell types that may or may not be sustainable in long-term culture. By contrast, ESCs are known to be stable in culture under the correct conditions, and they can be frozen for long-term storage, with excellent viability after thawing (Hogan et al., 1994). As discussed previously, reliable methods of generating a plentiful supply of islet endocrine cells in vivo or in vitro

FIGURE 42.4 Potential sources of transplantable b cells/islets. There are several potential avenues being explored to generate and expand mature b cells or functional islets in vitro as a replenishable supply for use in transplantation: (1) the proliferation and expansion of existing b cells; (2) the proliferation and expansion of cadaver-derived islets; (3) the induction of b-cell differentiation from endogenous progenitors (embryonic ductal cells) or from adult ductal epithelium; and (4) the induction of b-cell differentiation from embryonic stem cells. These strategies are more fully discussed in the text. (See color insert.)

966

PANCREAS DEVELOPMENT AND STEM CELLS

will benefit greatly from the identification and careful manipulation of the factors that promote the proliferation, regeneration, and neogenesis of b cells. Regardless of the source, experimentally derived insulin-producing cells must be rigorously tested for their survival, their engraftment, their physiologic function in vivo, their ability to reverse diabetes and maintain euglycemia, and their lack of tumorigenicity in a relevant animal model. The ability of in vitro derived insulin-producing cells to function as mature b cells in an animal model in vivo is critical to translating this eventually to human patients. One significant obstacle to the successful transplantation of experimentally derived insulin-producing cells is that current in vivo models for evaluating the functionality of such cells are inadequate. Another major limitation of current models is the inability to noninvasively monitor transplanted insulin-producing cells. The desired characteristics of an animal model to evaluate experimentally derived insulin-producing cells would include the following:

 The ability to accept and integrate grafts of human-derived cells or tissues (xenografts)

 The ability to reliably and safely induce diabetes before and after the transplantation of insulin-producing cells

 The ability to alter the expression of factors in the host or graft that may improve graft survival and function

 The ability to monitor b-cell mass noninvasively in clinically relevant sites like the liver

 The ability to retrieve grafted tissue and to assess differentiation and morphologic criteria

E. THE GENERATION OF INSULIN-PRODUCING CELLS FROM ADULT CELL SOURCES Several models of pancreatic or b-cell injury suggest that the adult ductal epithelium retains some capacity for the production of new b cells (neogenesis). The isolation and culture of ductal epithelium has been shown to yield isletlike clusters that contain functional insulin-producing cells (Bonner-Weir et al., 2000; Ramiya et al., 2000; Trivedi et al., 2001; Ogata et al., 2004). These studies suggest that at least a facultative (if not a genuine) endocrine stem cell exists in the adult pancreatic ducts that, when properly activated, is capable of giving rise to new, functional b cells. In addition, some studies suggest that acinar cells are capable of transdifferentiating directly to insulin-producing cells or of dedifferentiating into a ductal intermediate that can then go on to produce new endocrine cells (Rooman et al., 2000; Lardon et al., 2004; Means et al., 2005). For example, using cultured porcine fetal pancreas, acinar cells were observed to lose exocrine marker gene expression and to dedifferentiate to a multipotent progenitor cell. These immature cells have the capacity to differentiate as insulin-producing cells after transplantation (Humphrey et al., 2001). Islets themselves may also contain stem-like cells that are capable of generating new b cells (Guz et al., 2001; Li et al., 2003), although some researchers claim that the bone marrow harbors organ-specific stem cell populations that are capable of participating in regeneration after injury and possibly of generating new b cells (Ianus et al., 2003). This section will review the evidence for some of the different potential

THE GENERATION OF INSULIN-PRODUCING CELLS FROM EMBRYONIC STEM CELLS

967

avenues for deriving therapeutically useful islets or b cells for use in individuals with diabetes. Because the ductal epithelium is the source of new islet endocrine cells during embryonic development (see Figure 42.1), many investigations have concentrated on trying to generate new islets from fetal and adult ductal tissue. It is not clear whether there exists a quiescent, undifferentiated stem or progenitor cell within the pancreatic ducts and/or islets or whether differentiated duct cells can, with proper stimulation, act as progenitor cells for new acini and endocrine cells (Bouwens, 1998). A recent report suggests that the adult mouse pancreas does contain multipotent precursor cells that represent only 0.02% of the total cell population, which suggests that a true pancreatic stem cell may indeed exist in the adult (Seaberg et al., 2004). The markers and characteristics of this cell to allow for its identification in vivo are not known. There is also substantial evidence supporting the concept that differentiated duct cells can be activated to produce new endocrine cells in culture. In one study, ducts isolated from adult mouse pancreas were cultured to confluency, at which point islet-producing stem cells budded into the culture medium and formed spherical islet-like structures that expressed and secreted insulin in response to glucose (Cornelius et al., 1997; Ramiya et al., 2000). In addition, these structures rescued an experimentally induced form of diabetes, which suggests that they are functional in vivo. These cultures have been maintained for more than 3 years. New ductules that emerge in adult mice after partial pancreatectomy have also been cultured and shown to give rise to small clusters of endocrine cells that contain insulinþ cells (Kim et al., 2004). Similar results have been reported with cultured adult human duct tissue, which suggests that the ductal epithelium that remains after the islet isolation procedure for transplantation may be expanded and directed to differentiate, thus yielding a source of additional islets for transplantation (Bonner-Weir et al., 2000). Several studies have shown that the over-expression of Ngn3 in duct cultures greatly increases their capacity to give rise to endocrine cells (Heremans et al., 2002; Gasa et al., 2004). Unlike the results from Ngn3 over-expression in vivo, however, these endocrine cells include cells that express insulin. There is some evidence that genetically marked bone marrow cells can incorporate into pancreatic islets at a low frequency after transplantation into irradiated recipients (Ianus et al., 2003). These cells were reported to activate b-cell–specific genes, including insulin and Pdx1. Other studies provide evidence that bone-marrow–derived cells contribute to islet vasculature after injury and not to the formation of new b cells (Mathews et al., 2004). Overall, the ability of bone-marrow–derived cells to act as organ-specific stem cells remains controversial. More work needs to be done to determine whether cells from bone marrow actually differentiate to form cell types from other lineages or whether they provide factors that support the differentiation of new cells from within the host tissue in which they find themselves.

F. THE GENERATION OF INSULIN-PRODUCING CELLS FROM EMBRYONIC STEM CELLS Several investigators have described methods for the production of insulinexpressing cells from murine ESCs in vitro (Lumelsky et al., 2001; Hori et al., 2002; Blyszczuk et al., 2003, 2004; Kahan et al., 2003; Kim et al., 2003; Ku et al., 2004; Miyazaki et al., 2004). Although these protocols are

968

PANCREAS DEVELOPMENT AND STEM CELLS

all similar in that they go through an embryoid body stage in which pluripotent ESCs have already begun to differentiate, they differ with regard to the type of culture medium used, the selection criteria, and the presence or absence of active transgenes to promote b-cell differentiation (reviewed by Kania et al., 2004). In addition, there has been controversy with regard to the validity of the culture methods in which nestin positivity is used as an early selection criteria (Lumelsky et al., 2001; Hori et al., 2002; Miyazaki et al., 2004) and with regard to whether the insulinþ cells observed are actually functional b cells or if they may instead represent an artifact of insulin uptake from the culture medium (Rajagopal et al., 2003; Hansson et al., 2004; Kania et al., 2004; Sipione et al., 2004). As described previously, pancreatic islet cells have many characteristics in common with neurons, both in their gene expression and in their embryonic development. When nestin was identified as a marker of neuronal progenitor cells and selection for nestinþ cells facilitated neuronal differentiation from ESCs in culture (Andressen et al., 2001; Carpenter et al., 2001; Kawaguchi et al., 2001), nestin became a popular candidate as a marker for potential endocrine progenitor cells (Huang and Tang, 2003; Kim et al., 2004). However, more recent studies suggest that nestin expression within the pancreas is restricted to the exocrine lineage (Huang and Tang, 2003; Delacour et al., 2004) or to mesodermally derived cell types such as endothelial cells (Lardon et al., 2002; Selander and Edlund, 2002; Treutelaar et al., 2003). Thus, although much is known about the factors involved in pancreatic endocrine differentiation during murine embryonic development (reviewed by Jensen, 2004), there is currently a lack of consensus with regard to the best and most reliable method for consistently generating insulin-producing cells from ESCs. Ultimately, however, these strategies must be applied to human ESCs to generate human pancreatic islets or b cells, thereby providing a virtually limitless source of insulin-secreting cells for transplantation therapies. Recently, human ESCs have been shown to be capable of generating neurons (Carpenter et al., 2001) as well as other cell types from all three germ layers (reviewed by Odorico et al., 2001). A recent article reports the directed differentiation of hepatocytes in culture directly from murine ESCs without an embryoid body intermediate (Teratani et al., 2005). These hepatocytes were functional by several criteria, including the rescue of an animal model of liver cirrhosis. Because hepatocytes and pancreatic cells share a common embryologic origin and express many of the same genes, it may be that a similar approach would yield functional b cells in the future. Several groups have attempted the directed differentiation of murine ESCs into islet endocrine cells using a stepwise approach involving the adding of exogenous factors to the culture medium (Lumelsky et al., 2001; Kim et al., 2003; Ku et al., 2004). Many of the stepwise strategies were not hypothesis driven and were not based on a particular temporal sequence of secreted factors expressed in the pancreas developmentally, because this is in fact not currently known. Thus, in general, these methods have proven difficult to replicate across laboratories. That said, however, some generalities have come from these types of studies, in which insulin-producing cells are generated to a limited extent: (1) all culture strategies go through an embryoid body stage; (2) factors such as FGF, activin, betacellulin, exendin-4, and nicotinamide promote the production of insulinþ cells in the cultures; (3) the percentage of hormoneþ cells in the cultures are extremely low; (4) insulin-producing cells

969

SUMMARY

seem to be immature with regard to their gene-expression patterns and their ability to regulate insulin secretion in response to glucose; and (5) many other cell types, including neurons, are also present in the cultures. Even at their best, these methods yield only 3% insulinþ cells in the entire culture. More recently, Gordon Keller’s group has developed the concept that, to generate optimally functional b cells, ESCs must follow the same developmental pathway that endogenous endocrine progenitors follow: definitive endoderm ! foregut endoderm ! pancreas progenitor ! endocrine progenitor ! b cell (see Figure 42.4; Kubo et al., 2004; D’Amour et al., 2006). As mentioned previously, much research has been done regarding the transcription factors involved in pancreas development and islet differentiation, and some of this research is being used to try to force ESCs down a pancreatic/b cell differentiation pathway (Blyszczuk et al., 2003, 2004; Miyazaki et al., 2004). For example, the constitutive expression of either Pax4 or Pdx1 promotes the development of insulin-producing cells from murine ESCs (Blyszczuk et al., 2003). In Pax4 over-expressing ESCs, Isl-1, Ngn3, insulin, islet amyloid polypeptide, and GLUT2 mRNA levels increase significantly. These cells release insulin in response to glucose, and they are able to restore normal blood glucose levels after transplantation into a mouse model of type 1 diabetes.

VI. CONCLUSIONS The factors that control the specification and differentiation of the different pancreatic lineages continue to be elaborated. A complete understanding of normal pancreas development and endocrine differentiation will undoubtedly facilitate the generation of functional islets in vitro for use in the treatment of diabetes. Clearly, it is not sufficient to merely observe insulin reactivity in cultured cells derived from ESCs or any other source and conclude that these cells are b cells. The potential application of these cells to human disease demands a stringency whereby these cells are shown to endogenously produce insulin and to secrete it in a regulatable manner in response to alterations in extracellular glucose concentrations. Although these types of studies can initially be performed in vitro, the ultimate test is to determine whether derived insulin-producing cells can fully rescue (i.e., reverse diabetes and maintain glucose homeostasis) an animal whose endogenous b-cell population has been destroyed.

SUMMARY

 The pancreas is composed of two main cell types: exocrine cells and endocrine cells, both of which are derived from the endodermal germ layer.

 It is currently unclear if there exists a multipotent progenitor cell within  

the embryonic pancreas that is capable of giving rise to both exocrine and endocrine cell lineages. Progenitor cells for both the exocrine and endocrine cells are thought to reside within the embryonic pancreatic ductal epithelium. Many transcription factors have been identified that are critical for pancreas formation and the differentiation of the endocrine cell populations.

970

PANCREAS DEVELOPMENT AND STEM CELLS

 

 

However, less is know about the genes that control exocrine cell fate and differentiation. The b-cell mass has the capacity to expand and contract throughout the life of the organism in response to changing metabolic demands. In adults, the majority of new insulin-producing b cells come from the proliferation of preexisting b cells, although, under some conditions (e.g., pancreatic injury, obesity), b-cell neogenesis may occur from cells that are located within the ductal epithelium. There are currently no good protocols for consistently generating mature, glucose-responsive insulin-producing b cells from adult progenitor cell types or ESCs. Attempts to direct the differentiation of facultative or embryonic stem cells toward the b-cell fate will need to rely on a better understanding of the normal developmental processes that take a pluripotent cell from definitive endoderm to foregut endoderm to pancreas progenitor to endocrine progenitor to b cell.

ACKNOWLEDGMENTS I am grateful to Peter Wiebe and Laura Crawford in my laboratory for helpful discussions concerning the content and figures for this chapter. Space restrictions prevented me from including all areas of pancreas development and pancreas stem cell research. There are many laboratories around the world that are contributing greatly to these areas, and I wish I could refer to them all. MG was supported by a Career Development Award from the Juvenile Diabetes Research Foundation International (2–2002–583) and RO1s from the National Institutes of Health/National Institute of Diabetes and Digestive and Kidney Diseases (DK65131–01 and DK071052–01).

GLOSSARY b-cell The predominant endocrine cell type in the pancreatic islets; it is the only cell in the body that is capable of producing and secreting the hormone insulin. Diabetes A heterogeneous group of diseases in which the pancreas fails to produce insulin in sufficient amounts to maintain euglycemia, thus causing a dramatic rise in blood glucose levels. Endocrine cells Within the pancreas, the cells that produce hormones such as insulin and glucagon that are secreted directly into the bloodstream, where they travel to their target tissues. Euglycemia The state of maintaining blood glucose levels within the normal range (70– 120 mg/dL).

971

REFERENCES

Exocrine cells Within the pancreas, the cells that produce the digestive enzymes that are released into the pancreatic duct and eventually into the duodenum. Glucose intolerance An impaired ability to clear glucose efficiently from the bloodstream after a meal; a prediabetic state. Glucose-stimulated insulin secretion The increase in insulin released by pancreatic islets is proportional to the elevation in blood glucose; glucose must be metabolized by the b-cell for insulin to be released. Insulin resistance The physiologic state in which insulin target tissues in the periphery (mainly liver and muscle) become insensitive to insulin, thereby requiring elevated plasma insulin levels to elicit the same biologic effect. Islets Spherical clusters of pancreatic endocrine cells that are comprised mainly of insulin-producing b cells. Islet transplantation The experimental surgical procedure being used to treat some patients with type 1 diabetes; cadaver donor islets are transplanted into the liver of patients via the portal vein, where they then begin to secrete insulin and reverse diabetes. Mature onset diabetes of the young A dominant monogenic form of type 2 diabetes. Neogenesis The process of generating new b cells from stem or progenitor cells. Neurogenin 3 A basic helix–loop–helix transcription factor that is essential for the generation of all pancreatic islet cell types. Partial pancreatectomy A surgical procedure used in experimental animal models to stimulate pancreas and b-cell regeneration; a portion of the pancreas is removed, and new pancreas tissue is generated from the remnant organ. Type 1 diabetes An autoimmune disease in which the insulin-producing b cells are specifically destroyed. Type 2 diabetes A disease that is usually associated with obesity in which the b cells fail to produce enough insulin to overcome peripheral insulin resistance.

REFERENCES Ahlgren U, Jonsson J, Edlund H: The morphogenesis of the pancreatic mesenchyme is uncoupled from that of the pancreatic epithelium in IPF1/PDX1-deficient mice, Development 122: 1409–1416, 1996.

972

PANCREAS DEVELOPMENT AND STEM CELLS

Ahlgren U, Jonsson J, Jonsson L, et al: beta-cell-specific inactivation of the mouse Ipf1/Pdx1 gene results in loss of the beta-cell phenotype and maturity onset diabetes, Genes Dev 12: 1763–1768, 1998. Ahlgren U, Pfaff SL, Jessel TM, et al: Independent requirement for ISL1 in formation of pancreatic mesenchyme and islet cells, Nature 385:257–260, 1997. Alvarez C, Bass BL: Role of transforming growth factor-beta in growth and injury response of the pancreatic duct epithelium in vitro, J Gastrointest Surg 3:178–184, 1999. Andressen C, Stocker E, Klinz FL, et al: Nestin-specific green fluorescent protein expression in embryonic stem cell-derived neural precursor cells used for transplantation, Stem Cells 19: 419–424, 2001. Apelqvist A, Ahlgren U, Edlund H: Sonic hedgehog directs specialised mesoderm differentiation in the intestine and pancreas, Curr Biol 7:801–804, 1997. Apelqvist A, Li H, Sommer L, et al: Notch signalling controls pancreatic cell differentiation, Nature 400:877–881, 1999. Artner I, Blanchi B, Raum JC, et al: MafB is required for islet beta cell maturation, PNAS 104:3853–3858, 2007. Artner I, Le Lay J, Hang Y, et al: MafB: an activator of the glucagon gene expressed in developing islet alpha- and beta-cells, Diabetes 55:297–304, 2006. Ashery-Padan R, Zhou X, Marquardt T, et al: Conditional inactivation of Pax6 in the pancreas causes early onset of diabetes, Dev Biol 269:479–488, 2004. Ashraf A, Abdullatif H, Hardin W, Moates JM: Unusual case of neonatal diabetes mellitus due to congenital pancreas agenesis, Pediatr Diabetes 6:239–243, 2005. Beres TM, Masui T, Swift GH, et al: PTF1 is an organ-specific and Notch-independent basic helix–loop-helix complex containing the mammalian Suppressor of Hairless (RBP-J) or its paralogue, RBP-L, Mol Cell Biol 26:117–130, 2006. Bernard-Kargar C, Ktorza A: Endocrine pancreas plasticity under physiological and pathological conditions, Diabetes 50(Suppl 1):S30–S35, 2001. Blyszczuk P, Asbrand C, Rozzo A, et al: Embryonic stem cells differentiate into insulin-producing cells without selection of nestin-expressing cells, Int J Dev Biol 48:1095–1104, 2004. Blyszczuk P, Czyz J, Kania G, et al: Expression of Pax4 in embryonic stem cells promotes differentiation of nestin-positive progenitor and insulin-producing cells, Proc Natl Acad Sci U S A 100:998–1003, 2003. Bonner-Weir S: Islet growth and development in the adult, J Mol Endocrinol 24:297–302, 2000a. Bonner-Weir S: Life and death of the pancreatic beta cells, Trends Endocrinol Metab 11: 375–378, 2000b. Bonner-Weir S: Perspective: postnatal pancreatic beta cell growth, Endocrinology 141: 1926–1929, 2000c. Bonner-Weir S, Baxter LA, Schuppin GT, Smith FE: A second pathway for regeneration of adult exocrine and endocrine pancreas. A possible recapitulation of embryonic development, Diabetes 42:1715–1720, 1993. Bonner-Weir S, Sharma A: Pancreatic stem cells, J Pathol 197:519–526, 2002. Bonner-Weir S, Taneja M, Weir GC, et al: In vitro cultivation of human islets from expanded ductal tissue, Proc Natl Acad Sci U S A 97:7999–8004, 2000. Bort R, Martinez-Barbera JP, Beddington RS, Zaret KS: Hex homeobox gene-dependent tissue positioning is required for organogenesis of the ventral pancreas, Development 131:797–806, 2004. Bouwens L: Transdifferentiation versus stem cell hypothesis for the regeneration of islet beta-cells in the pancreas, Microsc Res Tech 43:332–336, 1998. Bouwens L, De Blay E: Islet morphogenesis and stem cell markers in rat pancreas, J Histochem Cytochem 44:947–951, 1996. Briones M, Bajaj M: Exenatide: a GLP-1 receptor agonist as novel therapy for type 2 diabetes mellitus, Expert Opin Pharmacother 7:1055–1064, 2006. Burke Z, Oliver G: Prox1 is an early specific marker for the developing liver and pancreas in the mammalian foregut endoderm, Mech Dev 118:147–155, 2002. Butler AE, Janson J, Bonner-Weir S, et al: Beta-cell deficit and increased beta-cell apoptosis in humans with type 2 diabetes, Diabetes 52:102–110, 2003. Cao LZ, Tang DQ, Horb ME, et al: High glucose is necessary for complete maturation of Pdx1VP16-expressing hepatic cells into functional insulin-producing cells, Diabetes 53: 3168–3178, 2004. Carpenter MK, Inokuma MS, Denham J, et al: Enrichment of neurons and neural precursors from human embryonic stem cells, Exp Neurol 172:383–397, 2001.

REFERENCES

973 Cheung AT, Dayanandan B, Lewis JT, et al: Glucose-dependent insulin release from genetically engineered K cells, Science 290:1959–1962, 2000. Chiang MK, Melton DA: Single-cell transcript analysis of pancreas development, Dev Cell 4:383–393, 2003. Christofori G, Naik P, Hanahan D: Vascular endothelial growth factor and its receptors, flt-1 and flk-1, are expressed in normal pancreatic islets and throughout islet cell tumorigenesis, Mol Endocrinol 9:1760–1770, 1995. Cirulli V, Beattie GM, Klier G, et al: Expression and function of alpha(v)beta(3) and alpha(v)beta (5) integrins in the developing pancreas: roles in the adhesion and migration of putative endocrine progenitor cells, J Cell Biol 150:1445–1460, 2000. Cirulli V, Crisa L, Beattie GM, et al: KSA antigen Ep-CAM mediates cell-cell adhesion of pancreatic epithelial cells: morphoregulatory roles in pancreatic islet development, J Cell Biol 140:1519–1534, 1998. Cleaver O, Melton DA: Endothelial signaling during development, Nat Med 9:661–668, 2003. Cockell M, Stevenson BJ, Strubin M, et al: Identification of a cell-specific DNA-binding activity that interacts with a transcriptional activator of genes expressed in the acinar pancreas, Mol Cell Biol 9:2464–2476, 1989. Collombat P, Mansouri A, Hecksher-Sorensen J, et al: Opposing actions of Arx and Pax4 in endocrine pancreas development, Genes Dev 17:2591–2603, 2003. Cornelius JG, Tchernev V, Kao KJ, Peck AB: In vitro-generation of islets in long-term cultures of pluripotent stem cells from adult mouse pancreas, Horm Metab Res 29:271–277, 1997. Costa RH, Kalinichenko VV, Holterman AX, Wang X: Transcription factors in liver development, differentiation, and regeneration, Hepatology 38:1331–1347, 2003. Crisera CA, Maldonado TS, Kadison AS, et al: Transforming growth factor-beta 1 in the developing mouse pancreas: a potential regulator of exocrine differentiation, Differentiation 65:255–259, 2000. D’Amour KA, Bang AG, Eliazer S, et al: Production of pancreatic hormone-expressing endocrine cells from human embryonic stem cells, Nature Biotech 24:1392–1401, 2006. De Leon DD, Deng S, Madani R, et al: Role of endogenous glucagon-like peptide-1 in islet regeneration after partial pancreatectomy, Diabetes 52:365–371, 2003. Delacour A, Nepote V, Trumpp A, Herrera PL: Nestin expression in pancreatic exocrine cell lineages, Mech Dev 121:3–14, 2004. Demeterco C, Beattie GM, Dib SA, et al: A role for activin A and betacellulin in human fetal pancreatic cell differentiation and growth, J Clin Endocrinol Metab 85:3892–3897, 2000. Dor Y, Brown J, Martinez OI, Melton DA: Adult pancreatic beta-cells are formed by selfduplication rather than stem-cell differentiation, Nature 429:41–46, 2004. Dutta S, Bonner-Weir S, Montminy M, Wright C: Regulatory factor linked to late-onset diabetes?, Nature 392:560, 1998. Edlund H: Transcribing pancreas, Diabetes 47:1817–1823, 1998. Edlund H: Developmental biology of the pancreas, Diabetes 50(Suppl 1):S5–S9, 2001. Edlund H: Pancreatic organogenesis—developmental mechanisms and implications for therapy, Nat Rev Genet 3:524–532, 2002. Edsbagge J, Johansson JK, Esni F, et al: Vascular function and sphingosine-1-phosphate regulate development of the dorsal pancreatic mesenchyme, Development 132:1085–1092, 2005. Esni F, Ghosh B, Biankin AV, et al: Notch inhibits Ptf1 function and acinar cell differentiation in developing mouse and zebrafish pancreas, Development 131:4213–4224, 2004. Esni F, Johansson BR, Radice GL, Semb H: Dorsal pancreas agenesis in N-cadherin-deficient mice, Dev Biol 238:202–212, 2001. Field HA, Ober EA, Roeser T, Stainier DY: Formation of the digestive system in zebrafish. I. Liver morphogenesis, Dev Biol 253:279–290, 2003. Finegood DT, Scaglia L, Bonner-Weir S: Dynamics of beta-cell mass in the growing rat pancreas. Estimation with a simple mathematical model, Diabetes 44:249–256, 1995. Fishman MP, Melton DA: Pancreatic lineage analysis using a retroviral vector in embryonic mice demonstrates a common progenitor for endocrine and exocrine cells, Int J Dev Biol 46: 201–207, 2002. Gale EA: The discovery of type 1 diabetes, Diabetes 50:217–226, 2001. Gannon M, Ray MK, Van Zee K, et al: Persistent expression of HNF6 in islet endocrine cells causes disrupted islet architecture and loss of beta cell function, Development 127:2883–2895, 2000. Gannon M, Wright CVE: Endodermal patterning and organogenesis, In Moody SA, editor: Cell lineage and fate determination, San Diego, 1999, Academic Press, pp. 583–615.

974

PANCREAS DEVELOPMENT AND STEM CELLS

Garcia-Ocana A, Takane KK, Reddy VT, et al: Adenovirus-mediated hepatocyte growth factor expression in mouse islets improves pancreatic islet transplant performance and reduces beta cell death, J Biol Chem 278:343–351, 2003. Garcia-Ocana A, Takane KK, Syed MA, et al: Hepatocyte growth factor overexpression in the islet of transgenic mice increases beta cell proliferation, enhances islet mass, and induces mild hypoglycemia, J Biol Chem 275:1226–1232, 2000. Gasa R, Mrejen C, Leachman N, et al: Proendocrine genes coordinate the pancreatic islet differentiation program in vitro, Proc Natl Acad Sci U S A 101:13245–13250, 2004. Georgia S, Bhushan A: beta cell replication is the primary mechanism for maintaining postnatal beta cell mass, J Clin Invest 114:963–968, 2004. Gittes GK, Rutter WJ: Onset of cell-specific gene expression in the developing mouse pancreas, Proc Natl Acad Sci U S A 89:1128–1132, 1992. Gradwohl G, Dierich A, LeMeur M, Guillemot F: neurogenin3 is required for the development of the four endocrine cell lineages of the pancreas, Proc Natl Acad Sci U S A 97:1607–1611, 2000. Grapin-Botton A, Majithia AR, Melton DA: Key events of pancreas formation are triggered in gut endoderm by ectopic expression of pancreatic regulatory genes, Genes Dev 15:444–454, 2001. Gress T, Muller-Pillasch F, Elsasser HP, et al: Enhancement of transforming growth factor beta 1 expression in the rat pancreas during regeneration from caerulein-induced pancreatitis, Eur J Clin Invest 24:679–685, 1994. Gu D, Sarvetnick N: Epithelial cell proliferation and islet neogenesis in IFN-g transgenic mice, Development 118:33–46, 1993. Gu G, Dubauskaite J, Melton DA: Direct evidence for the pancreatic lineage: NGN3þ cells are islet progenitors and are distinct from duct progenitors, Development 129:2447–2457, 2002. Gu G, Wells JM, Dombkowski D, et al: Global expression analysis of gene regulatory pathways during endocrine pancreatic development, Development 131:165–179, 2004. Guz Y, Montminy MR, Stein R, et al: Expression of murine STF-1, a putative insulin gene transcription factor, in beta cells of pancreas, duodenal epithelium and pancreatic exocrine and endocrine progenitors during ontogeny, Development 121:11–18, 1995. Guz Y, Nasir I, Teitelman G: Regeneration of pancreatic beta cells from intra-islet precursor cells in an experimental model of diabetes, Endocrinology 142:4956–4968, 2001. Hald J, Hjorth JP, German MS, et al: Activated Notch1 prevents differentiation of pancreatic acinar cells and attenuate endocrine development, Dev Biol 260:426–437, 2003. Hale MA, Kagami H, Shi L, et al: The homeodomain protein PDX1 is required at mid-pancreatic development for the formation of the exocrine pancreas, Dev Biol 286:225–237, 2005. Hansson M, Tonning A, Frandsen U, et al: Artifactual insulin release from differentiated embryonic stem cells, Diabetes 53:2603–2609, 2004. Harmon EB, Apelqvist AA, Smart NG, et al: GDF11 modulates NGN3þ islet progenitor cell number and promotes beta-cell differentiation in pancreas development, Development 131:6163–6174, 2004. Harrison KA, Thaler J, Pfaff SL, et al: Pancreas dorsal lobe agenesis and abnormal islets of Langerhans in Hlxb9-deficient mice, Nat Genet 23:71–75, 1999. Hattersley AT, Ellard S, Shepard M, et al: Phenotype-genotype relationships in maturity-onset diabetes of the young, In Matchinsky MA, editor: Molecular pathogenesis of MODYS, Basel, 2000, Karger, 15:16–34. Hayek A, Beattie GM, Cirulli V, et al: Growth factor/matrix-induced proliferation of human adult beta-cells, Diabetes 44:1458–1460, 1995. Hebrok M, Kim SK, Melton DA: Notochord repression of endodermal Sonic hedgehog permits pancreas development, Genes Dev 12:1705–1713, 1998. Hebrok M, Kim SK, St.Jacques B, et al: Regulation of pancreas development by hedgehog signaling, Development 127:4905–4913, 2000. Heller RS, Stoffers DA, Liu A, et al: The role of Brn4/Pou3f4 and Pax6 in forming the pancreatic glucagon cell identity, Dev Biol 268:123–134, 2004. Heremans Y, Van De Casteele M, in’t Veld P, et al: Recapitulation of embryonic neuroendocrine differentiation in adult human pancreatic duct cells expressing neurogenin 3, J Cell Biol 159:303–312, 2002. Herrera PL: Adult insulin- and glucagon-producing cells differentiate from two independent cell lineages, Development 127:2317–2322, 2000. Herrera PL, Huarte J, Zufferey R, et al: Ablation of islet endocrine cells by targeted expression of hormone-promoter-driven toxigenes, Proc Natl Acad Sci U S A 91:12999–13003, 1994.

REFERENCES

975 Hino S, Yamaoka T, Yamashita Y, et al: In vivo proliferation of differentiated pancreatic islet beta cells in transgenic mice expressing mutated cyclin-dependent kinase 4, Diabetologia 47:1819–1830, 2004. Hogan B, Beddington R, Costantini F, Lacy E: Manipulating the mouse embryo, Cold Spring Harbor, NY, 1994, Cold Spring Harbor Laboratory. Holland AM, Gonez LJ, Harrison LC: Progenitor cells in the adult pancreas, Diabetes Metab Res Rev 20:13–27, 2004. Horb ME, Shen CN, Tosh D, Slack JM: Experimental conversion of liver to pancreas, Curr Biol 13:105–115, 2003. Hori Y, Rulifson IC, Tsai BC, et al: Growth inhibitors promote differentiation of insulin-producing tissue from embryonic stem cells, Proc Natl Acad Sci U S A 99:16105–16110, 2002. Huang H, Tang X: Phenotypic determination and characterization of nestin-positive precursors derived from human fetal pancreas, Lab Invest 83:539–547, 2003. Humphrey RK, Smith MS, Kwok J, et al: In vitro dedifferentiation of fetal porcine pancreatic tissue prior to transplantation as islet-like cell clusters, Cells Tissues Organs 168:158–169, 2001. Huotari MA, Miettinen PJ, Palgi J, et al: ErbB signaling regulates lineage determination of developing pancreatic islet cells in embryonic organ culture, Endocrinology 143:4437–4446, 2002. Huotari MA, Palgi J, Otonoski T: Growth factor-mediated proliferation and differentiation of insulin-producing INS-1 and RINm5F cells: identification of betacellulin as a novel beta-cell mitogen, Endocrinology 139:1494–1499, 1998. Hussain MA, Miller CP, Habener JF: Brn-4 transcription factor expression targeted to the early developing mouse pancreas induces ectopic glucagon gene expression in insulin-producing beta cells, J Biol Chem 277:16028–16032, 2002. Ianus A, Holz GG, Theise ND, Hussain MA: In vivo derivation of glucose-competent pancreatic endocrine cells from bone marrow without evidence of cell fusion, J Clin Invest 111: 843–850, 2003. Imai J, Katagiri H, Yamada T, et al: Constitutively active PDX1 induced efficient insulin production in adult murine liver, Biochem Biophys Res Commun 326:402–409, 2005. Jacquemin P, Durviaux SM, Jensen J, et al: Transcription factor hepatocyte nuclear factor 6 regulates pancreatic endocrine cell differentiation and controls expression of the proendocrine gene ngn3, Mol Cell Biol 20:4445–4454, 2000. Jacquemin P, Lemaigre FP, Rousseau GG: The Onecut transcription factor HNF-6 (OC-1) is required for timely specification of the pancreas and acts upstream of Pdx-1 in the specification cascade, Dev Biol 258:105–116, 2003. Jensen J: Gene regulatory factors in pancreatic development, Dev Dyn 229:176–200, 2004. Jensen J, Heller RS, Funder-Nielsen T, et al: Independent development of pancreatic alpha- and beta-cells from neurogenin3-expressing precursors: a role for the notch pathway in repression of premature differentiation, Diabetes 49:163–176, 2000. Johnson CL, Kowalik AS, Rajakumar N, Pin CL: Mist1 is necessary for the establishment of granule organization in serous exocrine cells of the gastrointestinal tract, Mech Dev 121:261–272, 2004. Jonsson J, Carlsson L, Edlund T, Edlund H: Insulin-promoter-factor 1 is required for pancreas development in mice, Nature 371:606–609, 1994. Kahan BW, Jacobson LM, Hullet DA, et al: Pancreatic precursors and differentiated islet cell types from murine embryonic stem cells: an in vitro model to study islet differentiation, Diabetes 52:2016–2024, 2003. Kahn BB: Type 2 diabetes: when insulin secretion fails to compensate for insulin resistance, Cell 92:593–596, 1998. Kajihara M, Sone H, Amemiya M, et al: Mouse MafA, homologue of zebrafish somite Maf 1, contributes to the specific transcriptional activity through the insulin promoter, Biochem Biophys Res Commun 312:831–842, 2003. Kania G, Blyszczuk P, Wobus AM: The generation of insulin-producing cells from embryonic stem cells—a discussion of controversial findings, Int J Dev Biol 48:1061–1064, 2004. Kataoka K, Shioda S, Ando K, et al: Differentially expressed Maf family transcription factors, c-Maf and MafA, activate glucagon and insulin gene expression in pancreatic islet alphaand beta-cells, J Mol Endocrinol 32:9–20, 2004. Kawaguchi A, Miyata T, Sawamoto K, et al: Nestin-EGFP transgenic mice: visualization of the self-renewal and multipotency of CNS stem cells, Mol Cell Neurosci 17:259–273, 2001.

976

PANCREAS DEVELOPMENT AND STEM CELLS

Kawaguchi Y, Cooper B, Gannon M, et al: The role of the transcriptional regulator Ptf1a in converting intestinal to pancreatic progenitors, Nat Genet 32:128–134, 2002. Kelly OG, Melton DA: Development of the pancreas in Xenopus laevis, Dev Dyn 218:615–627, 2000. Kim D, Gu Y, Ishii M, et al: In vivo functioning and transplantable mature pancreatic islet-like cell clusters differentiated from embryonic stem cell, Pancreas 27:34–41, 2003. Kim SK, Hebrok M, Li E, et al: Activin receptor patterning of foregut organogenesis, Genes Dev 14:1866–1871, 2000. Kim SK, Hebrok M, Melton DA: Notochord to endoderm signaling is required for pancreas development, Development 124:4243–4252, 1997. Kim SK, MacDonald RJ: Signaling and transcriptional control of pancreatic organogenesis, Curr Opin Genet Dev 12:540–547, 2002. Kim SK, Melton DA: Pancreas development is promoted by cyclopamine, a hedgehog signaling inhibitor, Proc Natl Acad Sci U S A 95:13036–13041, 1998. Kim SY, Lee SH, Kim BM, et al: Activation of nestin-positive duct stem (NPDS) cells in pancreas upon neogenic motivation and possible cytodifferentiation into insulin-secreting cells from NPDS cells, Dev Dyn 230:1–11, 2004. Kitamura T, Nakae J, Kitamura Y, et al: The forkhead transcription factor Foxo1 links insulin signaling to Pdx1 regulation of pancreatic beta cell growth, J Clin Invest 110:1839–1847, 2002. Koizumi M, Doi R, Toyoda E, et al: Hepatic regeneration and enforced PDX-1 expression accelerate transdifferentiation in liver, Surgery 136:449–457, 2004. Krapp A, Knofler M, Frutiger S, et al: The p48 DNA-binding subunit of transcription factor PTF1 is a new exocrine pancreas-specific basic helix-loop-helix protein, EMBO J 15:4317–4329, 1996. Krapp A, Knofler M, Lederman B, et al: The bHLH protein PTF1-p48 is essential for the formation of the exocrine and the correct spatial organization of the endocrine pancreas, Genes Dev 12:3752–3763, 1998. Ku HT, Zhang N, Kubo A, et al: Committing embryonic stem cells to early endocrine pancreas in vitro, Stem Cells 22:1205–1217, 2004. Kubo A, Shinozaki K, Shannon JM, et al: Development of definitive endoderm from embryonic stem cells in culture, Development 131:1651–1662, 2004. Kumar M, Melton D: Pancreas specification: a budding question, Curr Opin Genet Dev 13: 401–407, 2003. Kushner JA, Ciemerych MA, Sicinska E, et al: Cyclins D2 and D1 are essential for postnatal pancreatic beta-cell growth, Mol Cell Biol 25:3752–3762, 2005. Lammert E, Cleaver O, Melton DA: Induction of pancreatic differentiation by signals from blood vessels, Science 294:564–567, 2001. Lammert E, Cleaver O, Melton DA: Role of endothelial cells in early pancreas and liver development, Mech Dev 120:59–64, 2003a. Landry C, Clotman F, Hioki T, et al: HNF-6 is expressed in endoderm derivatives and nervous system of the mouse embryo and participates to the cross-regulatory network of liver-enriched transcription factors, Dev Biol 192:247–257, 1997. Lardon J, Huyens N, Rooman I, Bouwens L: Exocrine cell transdifferentiation in dexamethasonetreated rat pancreas, Virchows Arch 444:61–65, 2004. Lardon J, Rooman I, Bouwens L: Nestin expression in pancreatic stellate cells and angiogenic endothelial cells, Histochem Cell Biol 117:535–540, 2002. LeCouter J, Lin R, Ferrara N: Endocrine gland-derived VEGF and the emerging hypothesis of organ-specific regulation of angiogenesis, Nat Med 8:913–917, 2002. Lee YC, Damholt AB, Billestrup N, et al: Developmental expression of proprotein convertase 1/3 in the rat, Mol Cell Endocrinol 155:27–35, 1999. Leung TW, Lin SS, Tsang AC, et al: Over-expression of FoxM1 stimulates cyclin B1 expression, FEBS Lett 507:59–66, 2001. Li H, Arber S, Jessell TM, Edlund H: Selective agenesis of the dorsal pancreas in mice lacking homeobox gene Hlxb9, Nat Genet 23:67–70, 1999. Li L, Seno M, Yamada H, Kojima I: Promotion of beta-cell regeneration by betacellulin in ninety percent-pancreatectomized rats, Endocrinology 142:5379–5385, 2001. Li L, Seno M, Yamada H, Kojima I: Betacellulin improves glucose metabolism by promoting conversion of intraislet precursor cells to beta-cells in streptozotocin-treated mice, Am J Physiol Endocrinol Metab 285:E577–E583, 2003. Li L, Yi Z, Seno M, Kojima I: Activin A and betacellulin: effect on regeneration of pancreatic beta-cells in neonatal streptozotocin-treated rats, Diabetes 53:608–615, 2004.

REFERENCES

977 Lifson N, Kramlinger KG, Mayrand RR, Lender EJ: Blood flow to the rabbit pancreas with special reference to the islets of Langerhans, Gastroenterology 79:466–473, 1980. Lifson N, Lassa CV, Dixit PK: Relation between blood flow and morphology in islet organ of rat pancreas, Am J of Physiol 249:E43–E48, 1985. Lin JW, Biankin AV, Horb ME, et al: Differential requirement for ptf1a in endocrine and exocrine lineages of developing zebrafish pancreas, Dev Biol 270:474–486, 2004. Liu YQ, Nevin PW, Leahy JL: Beta-cell adaptation in 60% pancreatectomy rats that preserves normoinsulinemia and normoglycemia, Am J Physiol Endocrinol Metab 279:E68–E73, 2000. Lopez-Talavera JC, Garcia-Ocana A, Sipula I, et al: Hepatocyte growth factor gene therapy for pancreatic islets in diabetes: reducing the minimal islet transplant mass required in a glucocorticoid-free rat model of allogeneic portal vein islet transplantation, Endocrinology 145:467–474, 2004. Lumelsky N, Blondel O, Laeng P, et al: Differentiation of embryonic stem cells to insulin-secreting structures similar to pancreatic islets, Science 292:1389–1394, 2001. Madsen OD, Jensen J, Petersen HV, et al: Transcription factors contributing to the pancreatic beta-cell phenotype, Horm Metab Res 29:265–270, 1997. Martin J, Hunt SL, Dubus P, et al: Genetic rescue of Cdk4 null mice restores pancreatic beta-cell proliferation but not homeostatic cell number, Oncogene 22:5261–5269, 2003. Marzo N, Mora C, Fabregat ME, et al: Pancreatic islets from cyclin-dependent kinase 4/R24C (Cdk4) knockin mice have significantly increased beta cell mass and are physiologically functional, indicating that Cdk4 is a potential target for pancreatic beta cell mass regeneration in Type 1 diabetes, Diabetologia 47:686–694, 2004. Mashima H, Shibata H, Mine T, Kojima I: Formation of insulin-producing cells from pancreatic acinar AR42J cells by hepatocyte growth factor, Endocrinology 137:3969–3976, 1996. Mathews V, Hanson PT, Ford E, et al: Recruitment of bone marrow-derived endothelial cells to sites of pancreatic beta-cell injury, Diabetes 53:91–98, 2004. Matsumoto K, Yoshitomi H, Rossant J, Zaret KS: Liver organogenesis promoted by endothelial cells prior to vascular function, Science 294:559–563, 2001. Matsuoka TA, Artner I, Henderson E, et al: The MafA transcription factor appears to be responsible for tissue-specific expression of insulin, Proc Natl Acad Sci U S A 101:2930–2933, 2004. Matsuoka TA, Zhao L, Artner I, et al: Members of the large Maf transcription family regulate insulin gene transcription in islet beta cells, Mol Cell Biol 23:6049–6062, 2003. McEvoy RC, Hegre OD: Morphometric quantitation of the pancreatic insulin-, glucagon-, and somatostatin-positive cell populations in normal and alloxan-diabetic rats, Diabetes 26:1140–1146, 1977. Means AL, Meszoely IM, Suzuki K, et al: Pancreatic epithelial plasticity mediated by acinar cell transdifferentiation and generation of nestin-positive intermediates, Development 132:3767–3776, 2005. Mettus RV, Rane SG: Characterization of the abnormal pancreatic development, reduced growth and infertility in Cdk4 mutant mice, Oncogene 22:8413–8421, 2003. Miyatsuka T, Kaneto H, Kajimoto Y, et al: Ectopically expressed PDX-1 in liver initiates endocrine and exocrine pancreas differentiation but causes dysmorphogenesis, Biochem Biophys Res Commun 310:1017–1025, 2003. Miyazaki S, Yamato E, Miyazaki J: Regulated expression of pdx-1 promotes in vitro differentiation of insulin-producing cells from embryonic stem cells, Diabetes 53:1030–1037, 2004. Murtaugh LC, Stanger BZ, Kwan KM, Melton DA: Notch signaling controls multiple steps of pancreatic differentiation, Proc Natl Acad Sci U S A 100:14920–14925, 2003. Nakajima-Nagata N, Sakurai T, Mitaka T, et al: In vitro induction of adult hepatic progenitor cells into insulin-producing cells, Biochem Biophys Res Commun 318:625–630, 2004. Naya FJ, Huang HP, Qiu Y, et al: Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in BETA2/neuroD-deficient mice, Genes Dev 11: 2323–2334, 1997. Nishimura W, Kondo T, Salameh T, et al: A switch from MafB to MafA expression accompanies differentiation to pancreatic beta-cells, Dev Biol 293:526–539, 2006. Obata J, Yano M, Mimura H, et al: p48 subunit of mouse PTF1 binds to RBP-Jkappa/CBF-1, the intracellular mediator of Notch signalling, and is expressed in the neural tube of early stage embryos, Genes Cells 6:345–360, 2001. Ober EA, Field HA, Stainier DY: From endoderm formation to liver and pancreas development in zebrafish, Mech Dev 120:5–18, 2003.

978

PANCREAS DEVELOPMENT AND STEM CELLS

Odorico JS, Kaufman DS, Thomson JA: Multilineage differentiation from human embryonic stem cell lines, Stem Cells 19:193–204, 2001. Offield MF, Jetton TL, Labosky PA, et al: PDX-1 is required for pancreatic outgrowth and differentiation of the rostral duodenum, Development 122:983–995, 1996. Ogata T, Park KY, Seno M, Kojima I: Reversal of streptozotocin-induced hyperglycemia by transplantation of pseudoislets consisting of beta cells derived from ductal cells, Endocr J 51: 381–386, 2004. Ogihara T, Watada H, Kanno R, et al: p38 MAPK is involved in activin A- and hepatocyte growth factor-mediated expression of pro-endocrine gene neurogenin 3 in AR42J-B13 cells, J Biol Chem 278:21693–21700, 2003. Olbrot M, Rud J, Moss LG, Sharma A: Identification of beta-cell-specific insulin gene transcription factor RIPE3b0001 as mammalian MafA, Proc Natl Acad Sci U S A 99: 6737–6742, 2002. Otonkoski T, Beattie GM, Rubin JS, et al: Hepatocyte growth factor/scatter factor has insulinotropic activity in human fetal pancreatic cells, Diabetes 43:947–953, 1994. Pang K, Mukonoweshuro C, Wong GG: Beta cells arise from glucose transporter type 2 (Glut2)expressing epithelial cells of the developing rat pancreas, Proc Natl Acad Sci U S A 91:9559–9563, 1994. Path G, Opel A, Knoll A, Seufert J: Nuclear protein p8 is associated with glucose-induced pancreatic beta-cell growth, Diabetes 53(Suppl 1):S82–S85, 2004. Percival AC, Slack JM: Analysis of pancreatic development using a cell lineage label, Exp Cell Res 247:123–132, 1999. Pictet RL, Clark WR, Williams RH, Rutter WJ: An ultrastructural analysis of the developing embryonic pancreas, Dev Biol 29:436–467, 1972. Pin CL, Rukstalis JM, Johnson C, Konieczny SF: The bHLH transcription factor Mist1 is required to maintain exocrine pancreas cell organization and acinar cell identity, J Cell Biol 155:519–530, 2001. Prado CL, Pugh-Bernard AE, Elghazi L, et al: Ghrelin cells replace insulin-producing beta cells in two mouse models of pancreas development, Proc Natl Acad Sci U S A 101:2924–2929, 2004. Prasadan K, Daume E, Preuett B, et al: Glucagon is required for early insulin-positive differentiation in the developing mouse pancreas, Diabetes 51:3229–3236, 2002. Rajagopal J, Anderson WJ, Kume S, et al: Insulin staining of ES cell progeny from insulin uptake, Science 299:363, 2003. Ramiya VK, Maraist M, Arfors KE, et al: Reversal of insulin-dependent diabetes using islets generated in vitro from pancreatic stem cells, Nat Med 6:278–282, 2000. Rane SG, Dubus P, Mettus RV, et al: Loss of Cdk4 expression causes insulin-deficient diabetes and Cdk4 activation results in beta-islet cell hyperplasia, Nat Genet 22:44–52, 1999. Rausa F, Samadani U, Ye H, et al: The cut-homeodomain transcriptional activator HNF-6 is coexpressed with its target gene HNF-3 beta in the developing murine liver and pancreas, Dev Biol 192:228–246, 1997. Risbud MV, Bhonde RR: Models of pancreatic regeneration in diabetes, Diabetes Res Clin Pract 58:155–165, 2002. Rooman I, Heremans Y, Heimberg H, Bouwens, L: Modulation of rat pancreatic acinoductal transdifferentiation and expression of PDX-1 in vitro, Diabetologia 43:907–914, 2000. Rooman I, Schuit F, Bouwens L: Effect of vascular endothelial growth factor on growth and differentiation of pancreatic ductal epithelium, Lab Invest 76:225–232, 1997. Rose SD, Swift GH, Peyton MJ, et al: The role of PTF1-P48 in pancreatic acinar gene expression, J Biol Chem 276:44018–44026, 2001. Rukstalis JM, Kowalik A, Zhu L, et al: Exocrine specific expression of Connexin32 is dependent on the basic helix-loop-helix transcription factor Mist1, J Cell Sci 116(Pt 16):3315–3325, 2003. Ryan EA, Lakey JR, Rajotte RV, et al: Clinical outcomes and insulin secretion after islet transplantation with the Edmonton protocol, Diabetes 50:710–719, 2001. Sander M, German MS: The beta cell transcription factors and development of the pancreas, J Mol Med 75:327–340, 1997. Sander M, Sussel L, Conners J, et al: Homeobox gene Nkx6.1 lies downstream of Nkx2.2 in the major pathway of beta-cell formation in the pancreas, Development 127:5533–5540, 2000. Scaglia L, Cahill CJ, Finegood DT, Bonner-Weir S: Apoptosis participates in the remodeling of the endocrine pancreas in the neonatal rat, Endocrinology 138:1736–1741, 1997.

REFERENCES

979 Schwitzgebel VM, Scheel DW, Conners JR, et al: Expression of neurogenin3 reveals an islet cell precursor population in the pancreas, Development 127:3533–3542, 2000. Seaberg RM, Smukler SR, Kieffer TJ, et al: Clonal identification of multipotent precursors from adult mouse pancreas that generate neural and pancreatic lineages, Nat Biotechnol 22:1115–1124, 2004. Selander L, Edlund H: Nestin is expressed in mesenchymal and not epithelial cells of the developing mouse pancreas, Mech Dev 113:189–192, 2002. Sellick GS, Barker KT, Stolte-Dijkstra I, et al: Mutations in PTF1A cause pancreatic and cerebellar agenesis, Nat Genet 36:1301–1305, 2004. Shapiro AM, Lakey JR, Ryan EA, et al: Islet transplantation in seven patients with type 1 diabetes mellitus using a glucocorticoid-free immunosuppressive regimen, N Engl J Med 343:230–238, 2000. Sharma A, Zangen DH, Reitz P, et al: The homeodomain protein IDX-1 increases after an early burst of proliferation during pancreatic regeneration, Diabetes 48:507–513, 1999. Shiozaki S, Tajima T, Zhang YQ, et al: Impaired differentiation of endocrine and exocrine cells of the pancreas in transgenic mouse expressing the truncated type II activin receptor, Biochim Biophys Acta 1450:1–11, 1999. Sipione S, Eshpeter A, Lyon JG, et al: Insulin expressing cells from differentiated embryonic stem cells are not beta cells, Diabetologia 47:499–508, 2004. Slack JM: Developmental biology of the pancreas, Development 121:1569–1580, 1995. Song SY, Gannon M, Washington MK, et al: Expansion of Pdx1-expressing pancreatic epithelium and islet neogenesis in transgenic mice overexpressing transforming growth factor alpha, Gastroenterology 117:1416–1426, 1999. Sonnenberg E, Meyer D, Weidner KM, Birchmeier C: Scatter factor/hepatocyte growth factor and its receptor, the c-met tyrosine kinase, can mediate a signal exchange between mesenchyme and epithelia during mouse development, J Cell Biol 123:223–235, 1993. Sosa-Pineda B, Chowdhury K, Torres M, et al: The Pax4 gene is essential for differentiation of insulin-producing beta cells in the mammalian pancreas, Nature 386:399–402, 1997. Sosa-Pineda B, Wigle JT, Oliver G: Hepatocyte migration during liver development requires Prox1, Nat Genet 25:254–255, 2000. St-Onge L, Sosa-Pineda B, Chowdhury K, et al: Pax6 is required for differentiation of glucagonproducing alpha-cells in mouse pancreas, Nature 387:406–409, 1997. Stoffers DA, Ferrer J, Clarke WL, Habener JF: Early-onset type-II diabetes mellitus (MODY4) linked to IPF1, Nat Genet 17:138–139, 1997. Stoffers DA, Stanojevic V, Habener JF: Insulin promoter factor-1 gene mutation linked to earlyonset type 2 diabetes mellitus directs expression of a dominant negative isoprotein, J Clin Invest 102:232–241, 1998. Sussel L, Kalamaras J, Hartigan-O-Connor DJ, et al: Mice lacking the homeodomain transcription factor Nkx2.2 have diabetes due to arrested differentiation of pancreatic beta cells, Development 125:2213–2221, 1998. Swenne I, Andersson A: Effect of genetic background on the capacity for islet cell replication in mice, Diabetologia 27:464–467, 1984. Swift GH, Liu Y, Rose SD, et al: An endocrine-exocrine switch in the activity of the pancreatic homeodomain protein PDX1 through formation of a trimeric complex with PBX1b and MRG1 (MEIS2), Mol Cell Biol 18:5109–5120, 1998. Taguchi M, Yamaguchi T, Otsuki M: Induction of PDX-1-positive cells in the main duct during regeneration after acute necrotizing pancreatitis in rats, J Pathol 197:638–646, 2002. Teratani T, Yamamoto H, Aoyagi K, et al: Direct hepatic fate specification from mouse embryonic stem cells, Hepatology 41:836–846, 2005. Treutelaar MK, Skidmore JM, Dias-Leme CL, et al: Nestin-lineage cells contribute to the microvasculature but not endocrine cells of the islet, Diabetes 52:2503–2512, 2003. Trivedi N, Hollister-Lock J, Lopez-Avalos MD, et al: Increase in beta-cell mass in transplanted porcine neonatal pancreatic cell clusters is due to proliferation of beta-cells and differentiation of duct cells, Endocrinology 142:2115–2122, 2001. Tsiotos GG, Barry MK, Johnson CD, Sarr MG: Pancreas regeneration after resection: does it occur in humans? Pancreas 19:310–313, 1999. Tsuchiya M, Taniguchi S, Yasuda K, et al: Potential roles of large mafs in cell lineages and developing pancreas, Pancreas 32:408–416, 2006. Tweedie E, Artner I, Crawtord L, et al: Maintenance of hepatic nuclear factor 6 in postnatal islets impairs terminal differentiation and function of beta-cells,, Diabetes 55:3264–3270, 2006.

980

PANCREAS DEVELOPMENT AND STEM CELLS

Wang J, Kilic G, Aydin M, et al: Prox1 activity controls pancreas morphogenesis and participates in the production of “secondary transition” pancreatic endocrine cells, Dev Biol 286:182–194, 2005. Wang RN, Kloppel G, Bouwens L: Duct- to islet-cell differentiation and islet growth in the pancreas of duct-ligated adult rats, Diabetologia 38:1405–1411, 1995. Wang X, Krupczak-Hollis K, Tan Y, et al: Increased hepatic Forkhead Box M1B (FoxM1B) levels in old-aged mice stimulated liver regeneration through diminished p27Kip1 protein levels and increased Cdc25B expression, J Biol Chem 277:44310–44316, 2002. Watada H, Kajimoto Y, Miyagawa J, et al: PDX-1 induces insulin and glucokinase gene expressions in alphaTC1 clone 6 cells in the presence of betacellulin, Diabetes 45:1826–1831, 1996. Watanabe H, Sumi S, Kitamura Y, et al: Immunohistochemical analysis of vascular endothelial growth factor and hepatocyte growth factor, and their receptors, in transplanted islets in rats, Surg Today 33:854–860, 2003. Wells JM: Genes expressed in the developing endocrine pancreas and their importance for stem cell and diabetes research, Diabetes Metab Res Rev 19:191–201, 2003. Wessels NK, Cohen JH: Early pancreas organogenesis: morphogenesis, tissue interactions, and mass effects, Dev Biol 15:237–270, 1967. Wilding L, Gannon M: The role of pdx1 and HNF6 in proliferation and differentiation of endocrine precursors, Diabetes Metab Res Rev 20:114–123, 2004. Wilson ME, Kalamaras JA, German MS: Expression pattern of IAPP and prohormone convertase 1/3 reveals a distinctive set of endocrine cells in the embryonic pancreas, Mech Dev 115:171–176, 2002. Wilson ME, Scheel D, German MS: Gene expression cascades in pancreatic development, Mech Dev 120:65–80, 2003. Wu KL, Gannon M, Peshavaria M, et al: Hepatocyte nuclear factor 3beta is involved in pancreatic beta-cell- specific transcription of the pdx-1 gene, Mol Cell Biol 17:6002–6013, 1997. Yoshitomi H, Zaret KS: Endothelial cell interactions initiate dorsal pancreas development by selectively inducing the transcription factor Ptf1a, Development 131:807–817, 2004. Zhang C, Moriguchi T, Kajihara M, et al: MafA is a key regulator of glucose-stimulated insulin secretion, Mol Cell Biol 25:4969–4976, 2005. Zhang H, Ackermann AM, Gusarova GA, et al: The Foxm1 transcription factor is required to maintain pancreatic beta-cell mass, Mol Endocrinol 20:1853–1866, 2006. Zhao L, Guo M, Matsuoka TA, et al: The islet beta cell-enriched MafA activator is a key regulator of insulin gene transcription, J Biol Chem 280:11887–11894, 2005.

FURTHER READING Jacquemin P, Pierreux CE, Fierens S, et al: Cloning and embryonic expression pattern of the mouse Onecut transcription factor OC-2, Gene Expr Patterns 3:639–644, 2003. Jensen J, Pedersen EE, Galante P, et al: Control of endodermal endocrine development by Hes-1, Nat Genet 24:36–44, 2000. Kim SK, Hebrok M, Melton DA: Pancreas development in the chick embryo, Cold Spring Harb Symp Quant Biol 6:377–383, 1997. Sander M, Neubuser A, Kalamaras J, et al: Genetic analysis reveals that PAX6 is required for normal transcription of pancreatic hormone genes and islet development, Genes Dev 11:1662–1673, 1997. Stoffers DA, Zinkin NT, Stanojevic V, et al: Pancreatic agenesis attributable to a single nucleotide deletion in the human IPF1 gene coding sequence, Nat Genet 15:106–110, 1997.

RECOMMENDED RESOURCES Bonner-Weir S: Life and death of the pancreatic beta cells, Trends Endocrinol Metab 11:375–378, 2000. Bonner-Weir S, Sharma A: Pancreatic stem cells, J Pathol 197:519–526, 2002. Edlund H: Developmental biology of the pancreas, Diabetes 50(Suppl 1):S5–S9, 2001. Edlund H: Pancreatic organogenesis—developmental mechanisms and implications for therapy, Nat Rev Genet 3:524–532, 2002.

RECOMMENDED RESOURCES

981

Gale EA: The discovery of type 1 diabetes, Diabetes 50:217–226, 2001. Jensen J: Gene regulatory factors in pancreatic development, Dev Dyn 229:176–200, 2004. Kahn BB: Type 2 diabetes: when insulin secretion fails to compensate for insulin resistance, Cell 92:593–596, 1998. Kania G, Blyszczuk P, Wobus AM: The generation of insulin-producing cells from embryonic stem cells—a discussion of controversial findings, Int J Dev Biol 48:1061–1064, 2004. Kim SK, MacDonald RJ: Signaling and transcriptional control of pancreatic organogenesis, Curr Opin Genet Dev 12:540–547, 2002. Risbud MV, Bhonde RR: Models of pancreatic regeneration in diabetes, Diabetes Res Clin Pract 58:155–165, 2002.

43

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS JAY D. KORMISH and KENNETH S. ZARET Cell and Developmental Biology Program, Fox Chase Cancer Center, Philadelphia, PA

INTRODUCTION The liver has many functions in health and disease, and thus there is much interest in the mechanisms of liver development and regeneration. The large size of the liver and its relative simplicity in terms of resident cell types make it an attractive experimental model for revealing general principles of development and progenitor cell biology. There are only two primary cell-type decisions that are made in the generation of the primary functional cell types of the liver. The first occurs when the embryonic endoderm generates the nascent liver cell, a hepatoblast, as opposed to other tissue progenitors, and the second occurs when the hepatoblast generates hepatocytes and cholangiocytes (bile duct cells). During development, the liver is a site of hematopoiesis, which is an essential function for the embryo. Thus, early liver growth is rapid, and vascular development is tightly integrated into hepatic morphogenesis. The rapid growth of the embryonic liver and its accessibility by relatively simple dissection techniques has enabled tissue explant and biochemical studies that have been unfeasible for other developing organ systems. Consequently, our understanding of liver organogenesis includes the genes and signals that induce cell-type decisions and mechanistic insights regarding how chromatin is regulated by transcription factors to execute the hepatic program. This chapter will focus on the experimental approaches and findings that have arisen from studies of early liver development in embryos and regenerating livers in adults and how they have provided insights that can be applied to stem cell differentiation. The adult liver secretes many serum proteins; they help control serum osmotic pressure and are carriers of lipids and other molecules. The adult liver also controls metabolite levels in the bloodstream and detoxifies ingested compounds that are taken up by the intestine and transferred to the liver via the hepatic portal vein. Acute toxicant ingestion can lead to the destruction of hepatocytes, but the cells have remarkable regenerative powers. Chronic liver

Principles of Developmental Genetics © 2007, Elsevier Inc. All rights reserved.

982

TWO ENDODERMAL ORIGINS OF EMBRYONIC LIVER CELLS

983

damage, whether elicited by toxicants, genetic disease, or viral infection, can lead to states in which mature hepatocyte proliferation is impaired and resident progenitor cells are activated. Because the liver is essential for viability and livers suitable for transplantation are in short supply, there is intense interest in generating hepatocytes from progenitor cells in the liver as well as from exogenous sources (e.g., transdifferentiation from other tissues or de novo differentiation from embryonic stem cells [ESCs]). In addition, being able to generate hepatocytes at will in vitro would greatly facilitate drug development, e.g., by allowing prospective compounds to be tested for their inhibitory activities, or their ability to elicit toxic molecular profiles. Furthermore, if hepatocytes could be readily generated from different human genetic backgrounds, it would allow drug leads to be prescreened for patient-specific sensitivities. An understanding of the factors required for liver development and progenitor cell biology will enhance the ability to generate hepatocytes for these purposes.

I. TWO ENDODERMAL ORIGINS OF EMBRYONIC LIVER CELLS The embryonic liver morphologically emerges from a single budding region of the developing ventral foregut. By contrast, the pancreas emerges from separate dorsal and ventral regions of the gut, and, later, the dorsal and ventral buds fuse to create the gland (Slack, 1995; see Chapter 42). However, when early time points of mouse foregut development were recently studied, the hepatoblasts that compose the liver bud were discovered to emanate from two spatially distinct populations of endoderm cells that are brought together, during gut closure, to create a single liver bud (Tremblay and Zaret, 2005; Figure 43.1). This finding arose from a fate-mapping study of

Distinct endoderm domains contribute to the liver bud. Upper panels, Frontal views of the anterior portion of mouse embryos, looking into the foregut. Lower panels, Parasagittal views. The figure depicts medial and lateral endoderm domains (pink) at different stages that contribute cells to the liver bud as determined by fate-mapping studies described in the text. The 1–3 and 4–6 somite pair stages precede liver specification. By the 7–8 somite pair stage, the prospective hepatic endoderm has converged at the ventral midline, and liver-specific gene expression commences. The green arrows depict the direction of the movement of cells. In the upper panels, the cells move ventrally and toward the midline. In the lower panels, the movement of cells enlarges the foregut. (See color insert.)

FIGURE 43.1

984

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

the foregut endoderm in which a lipophilic fluorescent dye was injected into patches of endoderm cells in different parts of the mouse embryonic foregut, before foregut tissue specification. The time of dye injection, which took place on embryonic day 8.0 (2–6 somite pairs), is the time at which the endoderm resides on the exterior, bottom surface of the embryo. This is during the beginning of foregut invagination, before gut tube closure. Embryos were cultured whole, in their yolk sac, until the organogenic phase of foregut gut development (embryonic day 9.0–9.5) and the tissue bud(s) in which the dye-injected cells resided were documented. The fate map results revealed paired lateral domains of endoderm cells on the left and right sides of the embryos, which, when labeled, primarily gave rise to liver bud cells (see Figure 43.1). The fate mapping also revealed a small domain of cells at the ventral midline that, when labeled, left a trail of descendants in the midline of various ventral foregut tissues and terminated within the rostral portion of the liver bud. This apparently multipotent medial progenitor population in the mouse appears to correspond with a similarly migrating population of cells seen in chick embryos (Kirby et al., 2003). In summary, liver bud cells arise from two functionally and spatially distinct progenitor domains (lateral and medial) that are morphogenetically brought together during foregut closure and that give rise to the liver bud. It is presently unknown whether common or distinct signals specify the liver in the medial and lateral hepatic progenitor domains. In this context, it is interesting to note that dorsal and ventral pancreas bud specification involves differences in signaling mechanisms (Deutsch et al., 2001; Kumar and Melton, 2003) and regulatory transcription factors (Ahlgren et al., 1997; Li et al., 1999; Bort et al., 2004). If there are differences in specifying the medial and lateral liver progenitors, it may indicate that there will be different ways to generate hepatocytes from nonliver progenitors and stem cells. Furthermore, genetic lineage studies will be necessary to determine whether adult descendants of the medial and lateral liver progenitors exhibit differences in hepatocyte function, including regenerative capacity. Considering that the discovery of different natural progenitors of the liver bud was made only recently, the rest of this chapter will consider the endoderm as a single embryonic source of liver progenitor cells.

II. DEVELOPMENTAL COMPETENCE OF THE VENTRAL FOREGUT ENDODERM, THE SPECIFICATION OF HEPATOBLASTS, AND THE EMERGENCE OF THE LIVER BUD A. Competence of the Endoderm to Initiate Hepatogenesis: Role of FOXA and GATA Transcription Factors The specification of the liver from the foregut endoderm has been defined as the time in development when the transcription of liver-specific genes initiates in endoderm cells and when it can be maintained by the cells outside of the normal embryonic context. During the foregut development of the mouse, the albumin gene is among the earliest to be expressed in the liver (Gualdi et al., 1996). Albumin expression is limited to the endoderm of the foregut region at the 7- to 8-somite stage (approximately embryonic day 8.5), and it remains active in foregut endoderm tissue explants (Cascio and Zaret, 1991; Gualdi et al., 1996). Although the midgut endoderm normally does not

DEVELOPMENTAL COMPETENCE OF THE VENTRAL FOREGUT ENDODERM

985

express albumin, isolating it from the mesoderm and ectoderm and culturing it in vitro causes the ectopic expression of albumin (Bossard and Zaret, 2000). Conversely, the in vitro association of midgut mesoderm inhibits albumin induction in foregut endoderm explants (Gualdi et al., 1996). However, by embryonic day 13.5, the midgut-associated endoderm has lost the capacity to activate albumin when removed from its associated mesoderm. These observations show that both the foregut and the midgut endoderm are competent to express the liver-specific gene before embryonic day E13.5, but they can be inhibited from doing so by mesodermal interactions. Recent studies in frogs and mice suggest that inhibition of Wnt signaling in the foregut but not the midgut is responsible for the observed difference in competence between the foregut and midgut (A. Zorn, personal communication). The expression of Wnt signaling inhibitors in the foregut but not the midgut during early endoderm development strengthens this model (Finley et al., 2003). Studies of factors controlling the expression of the albumin gene have revealed an explanation for developmental competence. The albumin gene contains an upstream enhancer element that is sufficient for liver-specific expression in transgenic mice (Pinkert et al., 1987). A variety of liver-enriched transcription factors have been found to bind and regulate the albumin enhancer in adult liver cells (DiPersio et al., 1991; Liu et al., 1991; Jackson et al., 1993; Bossard and Zaret, 1998; 2000). Interestingly, in vivo footprinting studies on E11.5 midgut, which does not express albumin but is competent to do so, showed that solely the FOXA and GATA binding sites of the enhancer are occupied in endoderm before albumin induction (Bossard and Zaret, 2000). Foxa1/Hnf3a, Foxa2/Hnf3b, Foxa3/Hnf3g, Gata4, and Gata6 transcription factor genes are expressed during the formation of the definitive endoderm, and they continue to be expressed in endodermal organs, including the liver (Ang et al., 1993; Monaghan et al., 1993; Sasaki and Hogan, 1993; Morrisey et al., 1998). Strikingly, the FOXA and GATA sites of the albumin enhancer become unoccupied on embryonic day 13.5, which is when the midgut endoderm loses its competence to express albumin (Bossard and Zaret, 2000). These observations suggest that FOXA and GATA factors play a key role in the endoderm’s competence to respond to developmental cues on embryonic day 13.5. Biochemical studies on in vitro reconstituted chromatin templates have shown that FOXA1 and, to a lesser extent, GATA4 can bind to target DNA sites in compacted chromatin and expose an underlying nucleosome (Cirillo et al., 2002). This has led to a model that stating that, in the endoderm, FOXA and GATA proteins act as “pioneer factors” by being among the first to bind to regulatory elements of genes during development and by helping to make the chromatin competent to be expressed in an endodermal organ (Zaret, 2002). Revealing the role of Foxa and Gata genes during liver development has been complicated by their functional redundancy and by their roles in gastrulation and the development of extraembryonic endoderm, before liver formation. The postgastrulation lethality associated with the Foxa2 homozygous mutant (Ang and Rossant, 1994; Weinstein et al., 1994) was initially overcome with the use of chimeric embryo technology (Dufort et al., 1998). In these experiments, Foxa2 homozygous null ESCs that were genetically marked with lacZ were injected into wild-type blastocysts, which were then implanted into a foster mother. The failure of Foxa2-/- cells to

986

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

survive in the foregut and midgut endoderm of the chimeric embryos demonstrated that Foxa2 is intrinsically necessary in those tissues (Dufort et al., 1998). Tissue-specific gene inactivation using the cre/lox system has also allowed for a bypass of early embryonic lethality. Here, the endogenous gene of interest is flanked by loxP sequences. In the presence of cre recombinase, which is driven by a tissue-specific promoter, the lox sites recombine and result in the excision of the intervening DNA. This technology has been used to inactivate Foxa2 function in the context of a Foxa1 deletion, the latter of which, by itself, is not lethal to embryos (Kaestner et al., 1999). Inactivation of both Foxa1 and Foxa2 in foregut endoderm demonstrated that the factors are redundantly required there for liver specification (Lee et al., 2005), thus supporting the chromatin competence model. GATA transcription factors have been found to play a similar critical role during early endoderm and liver development in the frog, fish, chicken, and mouse models of development (Stainier, 2002). In zebrafish, the morpholino knockdown of Gata4 has shown that it is necessary for liver specification (Holtzinger and Evans, 2005). Similar experiments with Gata6 have shown that it is not required for specification, but that it is required for the subsequent expansion of the fish liver bud (Holtzinger and Evans, 2005). Gata4 and Gata6 have been detected in the developing foregut and liver bud in the mouse (Bossard and Zaret, 1998; Morrisey et al., 1998; Jacobsen et al., 2002). Like those of the forkhead factors, the studies of Gata4 and Gata6 genes in mammalian development have been complicated by the early embryonic lethality associated with the mutants (Kuo et al., 1997; Morrisey et al., 1998). Chimeric embryo experiments showed that Gata4 is required in endoderm for foregut development, but a later role in liver development has not been investigated (Narita et al., 1997). The aggregation of Gata6 mutant ESCs and tetraploid wild-type ESCs allows a bypass of the yolk sac function of Gata factors, and this has demonstrated that Gata6 is required for liver bud expansion (Zhao et al., 2005). Gata4 may compensate for the early loss of Gata6 expression and allow for the liver specification that is observed in the Gata6 mutant. Such a prediction could be tested with the double inactivation of Gata4 and Gata6 during liver bud specification. In zebrafish and chicken, Gata5 is expressed in the developing foregut endoderm and liver (Laverriere et al., 1994; Reiter et al., 2001). Through the use of the morpholino-based knockdown of Gata5 in zebrafish, GATA5 has been found to be necessary for endoderm and liver development (Reiter et al., 2001). A similar function of Gata5 in mammalian foregut and liver development is unlikely, because Gata5 appears to not be expressed in the developing gut (Morrisey et al., 1998) B. Signaling Control of Liver Progenitor Cell Fate Embryo tissue transplant and explant systems in the chick and mouse have been fundamental for revealing signaling events that lead to liver specification. From such studies in the chick, it was initially determined that interactions with the cardiac mesoderm are necessary for the specification of the liver (Le Douarin, 1975; Houssaint, 1980; Fukuda-Taira, 1981; Gualdi et al., 1996). Further studies with the mouse determined that fibroblast growth factors (FGFs) produced by the cardiac mesoderm were essential for this induction (Gualdi et al., 1996; Jung et al., 1999, Calmont et al., 2006).

DEVELOPMENTAL COMPETENCE OF THE VENTRAL FOREGUT ENDODERM

987

In later experiments, it became apparent that tightly associated septum transversum mesenchyme cells were present in ventral foregut explants and that they were a source of bone morphogenetic protein (BMP) ligands (Rossi et al., 2001). BMPs alone could not induce liver gene expression, but they were found to be necessary in conjunction with FGFs for the endoderm to initiate hepatogenesis. Similar conclusions have been obtained with studies in the chick (Zhang et al., 2002; 2004). Interestingly, knowledge of the liver-inducing properties of BMP and FGF in embryos has been successfully applied to in vitro protocols for differentiating embryonic stem (ES) cell to liver-like cells (see Section V below). In subsequent studies of mouse FGF signaling during liver specification, it was determined that, in the absence of FGFs, the ventral foregut endoderm initiates the expression of genes for the ventral pancreas (Deutsch et al., 2001). The ventral pancreas is located immediately caudal to the liver domain, and it arises from the portion of the ventral foregut that normally has little or no contact with the cardiac endoderm. This has led to a model which states that the ventral foregut is bipotential for liver or pancreas fate, depending on contact with the cardiac mesoderm and the extent of exposure to FGF (Deutsch et al., 2001; Bort et al., 2004). This model has been extended by the observation that higher levels of FGF signaling are required for the specification of the lung just rostral to the liver region (Serls et al., 2005). The specification of the lung, liver, and ventral pancreas thus appears to be determined, at least in part, by the endoderm’s duration of association with the cardiac mesoderm and with different consequent levels of FGF signaling. In zebrafish, Wnt signaling has been found to be necessary for liver specification. An elegant forward genetic screen was performed to identify defects in endoderm organ development during zebrafish development (Ober et al., 2006). The mutant prometheus was identified and found to encode the Wnt2bb ligand. By various means, the authors showed that prometheus and b-catenin signaling is required for the early development of the zebrafish liver. Interestingly, even though the entire liver bud is missing cells during early development, a liver can still form a bit later than normal. It is believed that cells from the adjacent endoderm, possibly the gall bladder, can be recruited to initiate the formation of a liver bud in prometheus mutants, and once recruited possess the remarkable regenerative capacity to restore the developing liver. By this model, the Wnt2bb ligand would not be necessary for the specification of the later-appearing liver cells. Other signaling pathways have been studied with regard to their roles in the specification of foregut organs. In the zebrafish, retinoic acid is required for liver and pancreas specification (Stafford and Prince, 2002). In this system, it appears that retinoic acid acts as a morphogen to determine the caudal– rostral placement of the liver and pancreas near the end of gastrulation. When retinoic acid signaling is inhibited, the liver and pancreas are not specified. When retinoic acid signaling is enhanced by the addition of ectopic retinoic acid, the pancreas is specified at a more anterior position. Interestingly, intrinsic Hox gene expression does not seem to control foregut endoderm tissue specification; rather, as described previously, signaling from adjacent mesoderm has a primary role in cell-type specification. Furthermore, there is little evidence that a Hox code controls genes in the mesoderm that, in turn, specifically govern foregut patterning.

988

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

C. Concomitant Morphologic Transitions to the Liver Bud: Hex, Endothelial Cell Signaling, and Growth Zones After specification, there appear to be three stages of liver bud morphogenesis during which the nascent hepatoblasts transform from a simple epithelial layer into a mass of emergent liver bud cells (Bort et al., 2006). During the first phase, the nascent hepatoblast cells, which are contiguous with the endodermal epithelium, elongate away from the apical/luminal surface. This causes the epithelium to thicken and the cells to take on a columnar appearance. During the second phase, the nuclei of the hepatoblasts appear stratified as the result of the basal-to-apical migration of the nuclei progressing through cell division. During the third phase, the basal lamina degrades, and the hepatoblasts delaminate from the epithelium and migrate into the surrounding mesenchyme, creating a liver bud. Studies of the homeodomain gene Hex have shown that this transcription factor gene is critical for the second stage of hepatic bud development (Bort et al., 2006). In Hex null homozygotes, hepatoblast nuclei fail to undergo migration, and the cells consequently fail to delaminate. Later, Hex mutant hepatoblasts fail to maintain liver gene expression and eventually appear to take on a duodenal cell fate. The knockdown of hHex in zebrafish, using a morpholino approach, has also shown that the gene is required for liver development after hepatic specification (Wallace et al., 2001). The liver bud expression pattern of hHex in developing chick and frog embryos suggests that the gene plays similar roles in these organisms (Newman et al., 1997; Yatskievych et al., 1999). A second transcription factor gene, Prox1, appears to be required during the third phase of liver development, described above (Sosa-Pineda et al., 2000). In the Prox1 mutants, hepatic cells fail to delaminate and form a proper liver bud. This defect appears to result from defects in the normal sequence of E-cadherin downregulation and basal lamina breakdown. After the initial phases of liver bud outgrowth, the liver bud undergoes the extensive proliferation and migration of hepatic cells into the septum transversum mesenchyme (STM). The transcription factors H2.0-like homeobox gene (Hlx) and LIM-homeobox gene-2 (Lhx2) and the signaling molecules BMP4, hepatocyte growth factor (HGF), and transforming growth factor b are expressed by the STM, and they are implicated in hepatoblast proliferation (Schmidt et al., 1995; Hentsch et al., 1996; Amicone et al., 1997; Porter et al., 1997; Rossi et al., 2001; Weinstein et al., 2001; Kolterud et al., 2004; Wandzioch et al., 2004). Hlx and Lhx2 presumably promote the expression of the signaling factors expressed by the STM. In addition to the STM, endothelial cell precursors are recruited to or induced within the liver bud area, and these cells also promote liver bud migration and proliferation. A mouse mutant that is defective in endothelial cell differentiation, flk1, exhibits a failure in liver bud outgrowth (Matsumoto et al., 2001). The stimulatory effect of endothelial cells in wild-type liver bud explants, apart from the circulatory system, showed that the relevant factor(s) are produced locally by the endothelium. In adult livers, the release of HGF seems to be a paracrine factor released by endothelial cells to promote hepatocyte proliferation (LeCouter et al., 2003). The identification of additional signaling molecules involved in this endothelial–hepatocyte interaction may reveal factors that are crucial for liver regeneration.

HEPATOBLAST DIFFERENTIATION INTO HEPATOCYTES

989

Studies in chicken liver development have shed light on how the liver organ expands and undergoes further differentiation (Suksaweang et al., 2004). It appears that, in each liver lobe, differentiation from a hepatoblast to a functioning hepatocyte starts at the center of the lobe and expands to the periphery of the organs. The periphery of each lobe is associated with highly proliferative growth zones. The expanding growth zones appear to contain cells which are associated with mesenchymal cells. The alteration of the expression of b-catenin showed that the inhibition of Wnt signaling is required for the maintenance of the hepatic growth zones and further differentiation (Suksaweang et al., 2004).

III. HEPATOBLAST DIFFERENTIATION INTO HEPATOCYTES AND CHOLANGIOCYTES (BILIARY CELLS) A. Roles of HNF6, HNF-1b, and HNF4 Transcription Factors In the mouse, HNF-1b is expressed in the tubular structures of the liver, and it has been found to play an intrinsic role in biliary development (Coffinier et al., 2002). A cre recombinase driven by an alb/afp promoter and a floxed allele of endogenous HNF-1b gene has been used to specifically reduce the expression HNF-1b in the developing liver. Livers that are deficient for HNF-1b have large, malformed bile ducts and a failure of smaller bile ducts to differentiate. In these livers, there is also a concomitant loss of interlobal arteries. Because HNF-1b is not normally expressed in the interlobal arteries, it is thought that their defect is caused by the bile duct defect. This phenotype emphasizes the interplay between bile duct cells and endothelial cells during liver development. In addition to obvious defects in bile duct morphology, several genes associated with bile duct function (e.g., bile acid synthesis enzymes, fatty acid oxidation enzymes) are reduced in expression in the HNF-1b mutant. The role of HNF-1b in biliary development may be evolutionarily conserved in vertebrates, because the knockdown of HNF-1b in zebrafish shows a similar defect in biliary tree elaboration and bile secretion defects (Matthews et al., 2004). Further studies in zebrafish suggest that HNF-1b could play an even earlier role in liver development during foregut patterning. HNF-1b is expressed in the developing zebrafish foregut and later in the liver, pancreas, and hindgut (Sun and Hopkins, 2001; Gong et al., 2004). The morpholino knockdown of HNF-1b function during foregut development has shown that HNF-1b is absolutely necessary for pancreas development and that it influences liver specification (Sun and Hopkins, 2001). HNF6/OC1, which is a member of the onecut family, has been shown to play an essential role in liver development. Studies in the zebrafish have determined that HNF6 is necessary for proper biliary tract development (Matthews et al., 2004). In zebrafish, HNF6 is expressed in the developing liver and pancreas and later only in the liver, gall bladder, and proximal intestine. In the morpholino knockdown of HNF6, there is a defect in biliary tree development and bile secretion. Both mammalian and zebrafish studies have determined that HNF6 controls the expression of many other important transcription factors. HNF-1b appears to be a prime transcriptional target for HNF6, because in the HNF6 knockdown, HNF-1b transcription is reduced (Matthews et al., 2004). Indeed, HNF6 and HNF-1b morpholino knockdowns result in identical

990

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

phenotypes. This has led to the model that suggests that the phenotype observed in the HNF6 mutant is primarily the result of a defect in HNF-1b expression. B. Signaling Control: Role of Notch Signaling and Alagille Syndrome In both the zebrafish and the mouse, there is growing evidence that the Notch signaling pathway plays a significant role in the decision between hepatocyte and biliary cell fates. Hints for this role came from the study of Alagille syndrome in humans, which is characterized by defects in tubular structures throughout the body, including in the bile duct system in the liver. Human patients with Alagille syndrome frequently have mutations in the Notch receptor jagged1 (Li et al., 1997; Oda et al., 1997); this led to more in-depth studies of the Notch signaling pathway in mice. Mice that are double haploinsufficient for the jagged1 ligand and the Notch2 receptor or that are homozygous for a hypomorphic Notch2 allele display defects in bile duct differentiation (McCright et al., 2002). In these mutants, fewer cells form a bile duct precursor, and those that do form precursors never form a mature duct. A morpholino approach has been used to knock down the function of several members of the Notch pathway during zebrafish embryo development (Lorent et al., 2004). The Notch2/5 double knockdown results in a complete loss of the bile duct lineage and an increase in the hepatocyte lineage. The jagged 2/3 double knockdown results in a similar (although less penetrant) phenotype. In the reciprocal experiments, ectopic Notch expression results in ectopic biliary ducts. Taken together, the mouse and zebrafish studies show that the Notch pathway is required for the proper formation and differentiation of the bile duct cell lineage from a hepatoblast. In the mouse (McCright et al., 2002), jagged1 is expressed in the hepatic portal vein and arteries, and Notch2 is expressed in the surrounding hepatocytes. Thus, the notch pathway is likely activated in hepatoblasts that are in close proximity of the liver endothelial cells that express several jagged-type Notch receptors. C. Maturation of the Hepatocyte: Evidence for Complex Genetic Networks Initial evidence for liver transcription factor networks came from genetic studies. Using mouse genetics, HNF6 was found to be required for the stable expression of Foxa2 and HNF4 (Landry et al., 1997). In zebrafish morpholino knockdown studies, HNF6 is required for HNF-1b expression and for wildtype levels of HNF4 and Foxa2 expression (Matthews et al., 2004). In the mouse, HNF4, which is a nuclear receptor–type transcription factor, has been found to be required for liver differentiation after bud formation (Li et al., 2000; Parviz et al., 2003). A proximal-promoter region sufficient for high levels of liver expression of HNF4 has been identified. A combination of footprinting, DNA mobility shifting, and chromatin immunoprecipitation techniques have revealed binding sites for HNF-1a, HNF-1b, Sp-1, GATA6, HNF6, and COUP-TFII in the HNF4 promoter (Hatzis and Talianidis, 2001). In liverderived cell lines, high levels of HNF4 expression require synergism between HNF-1a and HNF6 or HNF1b and GATA6. These experiments and others involving chromatin immunoprecipitation (Kyrmizi et al., 2006) have led to a model depicting the complex interregulatory relationships among the transcription factors, i.e., a network that drives hepatic differentiation.

ADULT LIVER STEM CELLS

991

The above genetic and promoter studies have been vastly expanded by the recent advent of ChIP-on-chip studies. In this technique, DNA fragments that are isolated via chromatin immunoprecipitation are amplified and hybridized to microarrays displaying a tiled representation of the genome. The key advantage of this technique is that many transcription factor binding targets can be identified in one experiment. This has allowed for the expansion of known factor binding sites and the construction of regulatory networks of transcription factors required for liver development (Odom et al., 2004). In experiments performed by Odom et al. (2004), chromatin was isolated from human hepatocytes and pancreatic islets and immunoprecipitated with antibodies against HNF-1a, HNF4a, and HNF6. When liver and pancreatic targets were compared, HNF-1a, HNF6, and HNF4 were found to be bound to distinct sets of genes and a significant subset of common genes, apparently reflecting the close developmental relationship of the two tissues. HNF-1a and HNF4 consistently occupied the same promoters, and many genes were co-occupied by all three transcription factors. HNF6 was suggested to support a feed-forward loop in which HNF6 binds the HNF4 promoter and HNF4, in turn, binds many genes that are bound by both factors. The cross-regulatory relationship of the key transcription factors required for liver development has been expanded by additional ChIP-on-chip studies of HNF-1a, HNF4, FOXA2, HNF6, CREB1, and USF1 in liver chromatin (Odom et al., 2006). In all cases except USF1, the factors were found to bind their own promoters. In many cases, cross regulation was evident. For example, HNF4 and HNF-1a are bound to each other’s promoters, Foxa2 is bound by HNF6, and HNF6 and HNF-1a are bound to the HNF4 promoter. Furthermore, HNF-1a, HNF4, FOXA2, and HNF6 occupy the promoter of the transthyretin gene, which is a commonly studied liver gene. These results have further validated the combinatorial model for tissue-specific gene expression, and they indicate that the transcription factors that control multiple genes during development are likely to display autoregulation. This autoregulatory feedback is hypothesized to impart stability to the expression of key transcription factors driving organ differentiation. Interestingly, the auto- and cross-regulatory relationships of endodermal transcription factors, including FOXA1 and FOXA2, appear to have been fixed at the onset of metazoan evolution (Davidson and Erwin, 2006). From a comparison of the transcription factor networks that govern endoderm development, it was hypothesized that, within all metazoans, there exists an essentially immutable “kernel” network for endoderm development. This fundamental kernel provides the basis for “add-on” regulatory programs that differ among different metazoans. It remains to be determined to what extent the regulatory relationships of liver transcription factors unveiled by Odom et al. (2006) represent a fixed network for liver among other metazoans.

IV. ADULT LIVER STEM CELLS The mature liver is composed of functional units called lobules (Figure 43.2; Saxena et al., 1999; Fausto and Campbell, 2003). Lobules are approximately hexagonal-shaped structures that are composed of epithelial sheets or “plates” of hepatocytes. The hepatocyte plates are 1 to 2 cells thick, and they radiate from the central vein located in the center of the lobule to portal tracts at

992

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

FIGURE 43.2 Cross section of the liver lobule. Hepatocytes are organized into sheets of cells that radiate from the central vein to the portal tract. The portal tract (triad) is composed of the bile duct, the portal vein, and the portal artery. Blood flows from the portal tract through the sinusoidal spaces and to the central vein. Bile is secreted from the apical surface of the hepatocytes into the bile canaliculi. The canals of Hering are composed of cells that connect the bile ducts with the bile canaliculi. It is the cells of the canals of Hering that proliferate to generate oval cells during chronic liver damage. (Figure adapted from Bloom and Fawcett, 1994.)

the periphery. The portal tracts contain branches of the portal vein that carry blood from the intestine and the hepatic artery. Blood flows in an outward-in direction, from portal to central, through the sinusoidal spaces between the hepatic plates. Opposite from the sinusoidal/basal surface of the hepatocyte epithelium, bile is secreted into apical spaces called bile canaliculi. The bile canaliculi connect to intrahepatic bile ducts, which are composed of cholangiocytes in the portal tracts and which further connect to drain bile into the gall bladder. After acute liver damage (e.g., partial hepatectomy), hepatocytes rapidly replicate to reconstitute the liver mass (Rhim et al., 1994; Overturf et al., 1997; Laconi et al., 1998; Fausto and Campbell, 2003). On the basis of the high regenerative capacity of the liver and plate-like formation of hepatocytes within the lobule, it was originally thought that hepatocytes were generated from a stem cell niche located close the central vein, with descendants that stream distally to the portal tracts as they mature. However, this has proven to not be the case (Ng and Iannaccone, 1992; Bralet et al., 1994). Instead, it appears that virtually all hepatocytes have the capacity to replicate during normal growth and during acute liver damage (Rhim et al., 1994; Overturf et al., 1997). Bile duct cell replication appears to occur in response to an increase in biliary pressure, such as what occurs during biliary obstruction or damage (Slott et al., 1990). Biliary duct cell replication is evident in all parts of the tubular duct system, and the collective effect is to elongate the system. Like hepatocyte regeneration in response to acute damage, there has been no evidence for bile duct cell expansion via a stem cell niche (Slott et al., 1990).

ADULT LIVER STEM CELLS

993

In contrast with the situation with acute liver damage, recent studies have identified stem cell–like compartments in the adult liver that are activated in response to chronic liver damage (Thorgeirsson, 1996; Sell, 2001; Fausto and Campbell, 2003; Knight et al., 2005). Various chemical genetic models and human disease models have been used to mimic chronic liver damage, in which hepatocyte replication is compromised and a progenitor or stem cell population is activated (Shinozuka et al., 1978; Sell and Salman, 1984; Evarts et al., 1987; Germain et al., 1988; Dabeva and Shafritz, 1993; Factor et al., 1994; Sigal et al., 1995; Overturf et al., 1996). Careful work has shown the hepatic precursor cells/oval cells as originating from the canals of Hering, which is the ductular region that connects the hepatic canalicular system to the biliary tree (Paku et al., 2001). As the oval cells proliferate, they form tubular structures that maintain contact with the terminal bile ductual and distal hepatocytes. This structure is thought to maintain the flow of blood and bile while regeneration occurs. Furthermore, oval cell proliferation seems to be distinct from cholangiocyte proliferation, which occurs after bile duct obstruction (Slott et al., 1990; Paku et al., 2001). Oval cells are bipotential and have the capacity to form both hepatocytes and cholangiocytes (Pack et al., 1993; Nagy et al., 1994; Yin et al., 2002; Qin et al., 2004). However, proliferating biliary cells can only form other biliary cells (Slott et al., 1990). At this point, it is important to note that oval cells are likely not to be actual liver stem cells but are the fast replicating daughters of such a hypothesized cell type. The position and character of the true liver stem cell is still to be determined. Several laboratories have observed the stem cell–like properties of oval cells when they are transplanted into host livers (Yasui et al., 1997; Wang et al., 2003a). An example of the in vivo stem cell property of the oval cell came from experiments done by Wang et al. (2003a) with a mouse model. In these experiments, oval cell proliferation was induced with the carcinogen 3,5-diethoxycarbonyl-1,4-dihydrocollidine, and oval cells were isolated by size via fractionation and cell sorting on the basis of cell surface markers. The isolated cells were transplanted into mice that were genetically deficient for fumarylacetoacetate hydrolase (Fah; Overturf et al., 1996). Fah-deficient mice develop chronic liver damage, which in turn induces oval cell proliferation. Adding different levels of 2-(2-nitro-4-trifluoro-methyl-benzyolyl)-1,3cyclohexanedione (NTBC) reduces liver toxicity and prolongs the time over which liver regeneration must occur. When Fah-positive donor cells are introduced into a Fah-negative host, NTBC withdrawal imposes a selection for the proliferation of Fah-positive donor cells. The Fah mutant and wild-type alleles serve as genetic markers to distinguish between the host and donor cells, along with mismatched sex donors and hosts. When Fah-positive oval cells were introduced into a Fah-negative host, the donated cells were found to repopulate the liver. When differentially marked mature hepatocytes were coinjected with isolated oval cells, both oval cells and hepatocytes were able to equally contribute to liver regeneration. Overall, these results confirm the repopulating capacity of the oval cell in the compromised liver. Cell culture experiments demonstrated that oval cells can differentiate into both the hepatocyte and the bile duct cell lineages (Pack et al., 1993; Nagy et al., 1994; Yin et al., 2002; Qin et al., 2004). Future studies are needed to rigorously determine the capacity of oval cells to contribute to both hepatocytes and bile duct cells in the in vivo model.

994

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

The origin of oval cells has also been a subject of recent debate. Are oval cells a remnant of early embryonic hepatoblasts, are they generated later in development, or do they originate from non-liver sources? The morphologic changes of the hepatoblasts during development and the repopulating capacity of fetal hepatic cells have shed light on the potential developmental origins of oval cells. a-Fetoprotein (AFP) is expressed at high levels only during fetal development of the liver. The re-expression of AFP and other fetal markers during oval cell proliferation suggests that these cells may have a less mature hepatic character (Sell et al., 1974; Hayner et al., 1984; Sell and Salman, 1984; Lemire and Fausto, 1991; Alpini et al., 1992). Furthermore, during liver development, hepatoblasts near endothelial structures express markers that are indicative of both hepatocytes and bile duct cells. A similar expression of two differentiation markers is also seen in oval cells during their expansion after chronic liver injury (Yaswen et al., 1984; Germain et al., 1985). Like oval cells, hepatoblasts can be cultured and induced to form cells of both hepatocyte and cholangiocyte character (Rogler, 1997; Kamiya et al., 1999; Strick-Marchand and Weiss, 2002). Also, freshly isolated hepatoblasts and hepatoblasts that have been cultured can be reintroduced into the liver and repopulate both lineages of liver cells, under normal and chronic injury conditions (Dabeva et al., 2000; Sandhu et al., 2001; Malhi et al., 2002; Suzuki et al., 2002; Minguet et al., 2003; Strick-Marchand et al., 2004). Careful genetic lineage tracing during development will be needed to further refine the hepatoblast–oval cell lineage model. Oval cell/hepatic precursor cells (Omori et al., 1997; Petersen et al., 1998; Theise et al., 1999; Crosby et al., 2001; Petersen et al., 2003) and fetal hepatoblasts (Suzuki et al., 2002; Lazaro et al., 2003; Nava et al., 2005) express certain markers in common with hematopoietic stem cells (HSCs). In vitro experiments have demonstrated that a subset of bone marrow stem cells express AFP and c-kit (an HGF receptor) and that these stem cells can be induced to express hepatocyte and bile duct markers when they are exposed to appropriate stimuli (Oh et al., 2000; Miyazaki et al., 2002; Schwartz et al., 2002; Lee et al., 2004). However, in vivo experiments are not able to effectively recapitulate this in vitro transition. Well-controlled transplant experiments demonstrate that HSC conversion into hepatocytes is rare to negligible (Petersen et al., 1999; Alison et al., 2000; Fogt et al., 2002; Wagers et al., 2002; Ng et al., 2003; Menthena et al., 2004). Recently, the rare HSC transdifferentiation events observed in chronic injury models have been attributed to the fusion of HSCs to hepatocytes (Vassilopoulos et al., 2003; Wang et al., 2003b). The variable ploidy of hepatocytes and the ability of macrophages to fuse with other cells make this a plausible model (Vassilopoulos et al., 2003). Attempts to refute the fusion model have been made, but more sensitive fusion detection systems need to be employed to make this argument more convincing (Harris et al., 2004). Therefore, it appears that the majority of oval cells arise from the liver and that they express many markers in common with other stem cell tissues (Omori et al., 1997; Petersen et al., 1998; Theise et al., 1999; Crosby et al., 2001; Petersen et al., 2003). A poorly defined subset of cells begins to proliferate in the periductual structures of the liver lobule under conditions that stimulate oval cell production (Sell, 2001). These cells do not appear to form functional bile duct cells, and they have been speculated to be mesenchymal in origin, perhaps stellate cells (mobile mesenchymal cells of the mature liver) or infiltrating hematopoietic stem cells

DIFFERENTIATION OF LIVER-LIKE CELLS FROM EMBRYONIC STEM CELLS

995

(Paku et al., 2001). Although the origin and purpose of these cells remain unclear, the vicinity to proliferating oval cells suggests that they may be part of a niche that either stabilizes oval cells or that activates them during chronic damage. Given the important role of the STM and endothelial cells during embryonic hepatoblast proliferation and migration (Schmidt et al., 1995; Amicone et al., 1997; Matsumoto et al., 2001; Rossi et al., 2001; LeCouter et al., 2003; Wandzioch et al., 2004), the mesenchymal origin of the periductal cells near oval cells may not be surprising, and it suggests that signaling events resembling those in development could be involved.

V. DIFFERENTIATION OF LIVER-LIKE CELLS FROM EMBRYONIC STEM CELLS The difficultly in obtaining and isolating human oval cells and fetal hepatoblasts has led to the search for other sources of cells that could be used for liver regeneration via stem cell transplants. Mouse ESCs derived from the cells of the early blastocyst embryo can be cultured, manipulated genetically, and reintroduced into the developing blastocyst, where they can reconstitute all tissues of the developing embryo (Evans and Kaufman, 1981; Martin, 1981; Thomson et al., 1998; Reubinoff et al., 2000). Several groups have been able to promote the differentiation of ESCs into hepatocyte-like cells both in vitro and in vivo, albeit not very efficiently (Chinzei et al., 2002; Yamamoto et al., 2003; Teratani et al., 2005; Yamamoto et al., 2005). The differentiation of ESCs toward an hepatoblast phenotype reduces their frequency of teratoma formation (Chinzei et al., 2002). The partial differentiation of ESCs into hepatocyte precursors may allow for an expansion and enrichment of cells that will effectively contribute to liver tissue. A major hurdle in defining the final differentiation of hepatocytes has been the ability to distinguish hepatoblasts from visceral endoderm. Visceral endoderm is formed early during embryo development, and it later forms the extraembryonic tissue of the yolk sac. The visceral endoderm and yolk sac function together in a role that is similar to that of the liver before liver development (Meehan et al., 1984; Sellem et al., 1984; Duncan et al., 1997), and thus the yolk sac expresses many of the same genes that are diagnostic of liver differentiation (Meehan et al., 1984; Abe et al., 1996). More careful experiments have found genes that are differentially expressed in the liver and the yolk sac, and these markers have been used to confirm the formation of hepatocyte-like cells from ESC lines (Jones et al., 2002; Asahina et al., 2004; Kubo et al., 2004; Stamp et al., 2005; Tada et al., 2005). Signaling molecules that are known to influence endodermal cells during embryo development, such as activin, are being used to promote ESCs to differentiate along a mesendodermal fate (Schuldiner et al., 2000; Kubo et al., 2004; D’Amour et al., 2005; Tada et al., 2005). Brachyury, Foxa2, and goosecoid have been used as markers for this transition (Blum et al., 1992; Showell et al., 2004), along with cell sorting to enrich for definitive endoderm cells. Cells expressing markers of liver, lung, and intestine differentiated from such cells, although again at low efficiency (Kubo et al., 2004; Tada et al., 2005). Mimicking what has been found in embryology, many studies have found that FGFs promote the differentiation of ESC to liver-like cells

996

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

in vitro (Hamazaki et al., 2001; Kuai et al., 2003; Yamamoto et al., 2003, 2005; Teratani et al., 2005). Notably, Gouon-Evans et al. (2006) recently added BMPs (Rossi et al., 2001) to the differentiation protocol and more consistently and robustly generated hepatocyte-like cells from mouse ESC. Interestingly, they also saw co-differentiation of endothelial cells at the periphery of their hepatocyte-like cultures; as in development (Matsumoto et al., 2001), conditions that enhanced endothelial yield, enhanced hepatic cell yield. It thus seems likely that the further enhancement of hepatocyte differentiation from ESCs will be facilitated by the additional application of principles that apply to normal liver development.

SUMMARY

 The liver is specified from two functionally and spatially distinct domains 



 

of ventral foregut endoderm. FGF, BMP, and Wnt signaling appear to be required for liver specification. Hepatoblasts proliferate and migrate away from the ventral foregut. Intrinsic factors as well as signals from the septum transversum mesenchyme and endothelial-derived mesenchyme are required for liver bud development. After acute damage, hepatocytes undergo massive replication to reconstitute the damaged liver. Only during chronic liver damage, when hepatocyte replication is compromised, does a bipotential hepatic precursor cell/oval cell population begin to proliferate to repair the liver. Oval cells originate near the canals of Hering. Nonliver cell types, such as HSCs, do not appear to significantly contribute to liver cell repopulation. Presently, the most successful approaches to generating hepatic precursor cells from ESCs employ the signaling events and intrinsic transcription factors that are required for normal liver development.

GLOSSARY Canal of Hering A small population of cells that connect the bile ducts to hepatocytes. These cells allow for the flow of bile from the bile canaculi to the bile ducts. Cholangiocyte The cell type that comprises the bile duct system. Commitment The point in development when a specified tissue can no longer respond to development signals and becomes set in its gene expression profile. Competence The point in development at which a field of cells becomes competent to respond to signaling cues for cell-type specification. Hepatoblast The bipotential embryonic liver precursor cell for the adult hepatocyte or cholangiocyte.

997

REFERENCES

Hepatocytes The functional metabolic cells of the mature liver. These cells form polarized sheets or “plates” in the liver. The basal surface is in contact with endothelial cells, which allows for the transfer of metabolites to and from the blood stream, whereas the apical surface secretes bile into canaliculi, which form between adjacent hepatocytes. Oval cells A subset of cells that originate from the canals of Hering and that proliferate to reconstitute both hepatocytes and cholangiocytes during chronic liver damage. Oval cells are hypothesized to either constitute or to be related to the bipotential hepatic precursor cell population (hepatoblasts). Specification The point in development at which a field of cells, such as the one in the ventral foregut, starts to express markers that are specific to an organ and at which the expression of such markers is stable when the endoderm is cultivated outside of the embryo.

REFERENCES Abe K, Niwa H, Iwase K, et al: Endoderm-specific gene expression in embryonic stem cells differentiated to embryoid bodies, Exp Cell Res 229:27–34, 1996. Ahlgren U, Pfaff SL, Jessell TM, et al: Independent requirement for ISL1 in formation of pancreatic mesenchyme and islet cells, Nature 385:257–260, 1997. Alison MR, Poulsom R, Jeffery R, et al: Hepatocytes from non-hepatic adult stem cells, Nature 406:257, 2000. Alpini G, Aragona E, Dabeva M, et al: Distribution of albumin and alpha-fetoprotein mRNAs in normal, hyperplastic, and preneoplastic rat liver, Am J Pathol 141:623–632, 1992. Amicone L, Spagnoli FM, Spath G, et al: Transgenic expression in the liver of truncated Met blocks apoptosis and permits immortalization of hepatocytes, EMBO J 16:495–503, 1997. Ang SL, Rossant J: HNF-3b is essential for node and notochord formation in mouse development, Cell 78:561–574, 1994. Ang SL, Wierda A, Wong D, et al: The formation and maintenance of the definitive endoderm lineage in the mouse: involvement of HNF3/forkhead proteins, Development 119: 1301–1315, 1993. Asahina K, Fujimori H, Shimizu-Saito K, et al: Expression of the liver-specific gene Cyp7a1 reveals hepatic differentiation in embryoid bodies derived from mouse embryonic stem cells, Genes Cells 9:1297–1308, 2004. Bloom W, Fawcett DW: A textbook of histology, Chapman and Hall, New York, 1994. Blum M, Gaunt SJ, Cho KWY, et al: Gastrulation in the mouse: the role of the homeobox gene goosecoid, Cell 69:1097–1106, 1992. Bort R, Martinez-Barbera JP, Beddington RS, Zaret KS: Hex homeobox gene-dependent tissue positioning is required for organogenesis of the ventral pancreas, Development 131:797–806, 2004. Bort R, Signore M, Tremblay K, et al: Hex homeobox gene controls the transition of the endoderm to a pseudostratified, cell emergent epithelium for liver bud development, Dev Biol 290:44–56, 2006. Bossard P, Zaret KS: GATA transcription factors as potentiators of gut endoderm differentiation, Development 125:4909–4917, 1998. Bossard P, Zaret KS: Repressive and restrictive mesodermal interactions with gut endoderm: possible relation to Meckel’s diverticulum, Development 127:4915–4923, 2000. Bralet MP, Branchereau S, Brechot C, Ferry N: Cell lineage study in the liver using retroviral mediated gene transfer. Evidence against the streaming of hepatocytes in normal liver, Am J Pathol 144:896–905, 1994.

998

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

Calmont A, Wandzioch E, Tremblay KD, et al: An FGF response pathway that mediates hepatic gene induction in embryonic endoderm cells, Developmental Cell 11:339–348, 2006. Cascio S, Zaret KS: Hepatocyte differentiation initiates during endodermal-mesenchymal interactions prior to liver formation, Development 113:217–225, 1991. Chinzei R, Tanaka Y, Shimizu-Saito K, et al: Embryoid-body cells derived from a mouse embryonic stem cell line show differentiation into functional hepatocytes, Hepatology 36:22–29, 2002. Cirillo L, Lin FR, Cuesta I, et al: Opening of compacted chromatin by early developmental transcription factors HNF3 (FOXA) and GATA-4, Mol Cell 9:279–289, 2002. Coffinier C, Gresh L, Fiette L, et al: Bile system morphogenesis defects and liver dysfunction upon targeted deletion of HNF1beta, Development 129:1829–1838, 2002. Crosby HA, Kelly DA, Strain AJ: Human hepatic stem-like cells isolated using c-kit or CD34 can differentiate into biliary epithelium, Gastroenterology 120:534–544, 2001. Dabeva MD, Petkov PM, Sandhu J, et al: Proliferation and differentiation of fetal liver epithelial progenitor cells after transplantation into adult rat liver, Am J Pathol 156:2017–2031, 2000. Dabeva MD, Shafritz DA: Activation, proliferation, and differentiation of progenitor cells into hepatocytes in the D-galactosamine model of liver regeneration, Am J Pathol 143: 1606–1620, 1993. D’Amour KA, Agulnick AD, Eliazer S, et al: Efficient differentiation of human embryonic stem cells to definitive endoderm, Nat Biotech 23:1534–1541, 2005. Davidson EH, Erwin DH: Gene regulatory networks and the evolution of animal body plans, Science 311:796–800, 2006. Deutsch G, Jung J, Zheng M, et al: A bipotential precursor population for pancreas and liver within the embryonic endoderm, Development 128:871–881, 2001. DiPersio CM, Jackson DA, Zaret KS: The extracellular matrix coordinately modulates liver transcription factors and hepatocyte morphology, Mol Cell Biol 11:4405–4414, 1991. Dufort D, Schwartz L, Harpal K, Rossant J: The transcription factor HNF3b is required in visceral endoderm for normal primitive streak morphogenesis, Development 125:3015–3025, 1998. Duncan SA, Nagy A, Chan W: Murine gastrulation requires HNF-4 regulated gene expression in the visceral endoderm: tetraploid rescue of Hnf-4(-/-) embryos, Development 124:279–287, 1997. Evans MJ, Kaufman MH: Establishment in culture of pluripotential cells from mouse embryos, Nature 292:154–156, 1981. Evarts RP, Nagy P, Marsden E, Thorgeirsson SS: A precursor-product relationship exists between oval cells and hepatocytes in rat liver, Carcinogenesis 8:1737–1740, 1987. Factor VM, Radaeva SA, Thorgeirsson SS: Origin and fate of oval cells in dipin-induced hepatocarcinogenesis in the mouse, Am J Pathol 145:409–422, 1994. Fausto N, Campbell JS: The role of hepatocytes and oval cells in liver regeneration and repopulation, Mech Dev 120:117–130, 2003. Finley KR, Tennessen J, Shawlot W: The mouse secreted frizzled-related protein 5 gene is expressed in the anterior visceral endoderm and foregut endoderm during early postimplantation development, Gene Exp., Patt. 3:681–684, 2003. Fogt F, Beyser KH, Poremba C, et al: Recipient-derived hepatocytes in liver transplants: a rare event in sex-mismatched transplants, Hepatology 36:173–176, 2002. Fukuda-Taira S: Hepatic induction in the avian embryo: specificity of reactive endoderm and inductive mesoderm, J Embryol Exp Morph 63:111–125, 1981. Germain L, Blouin MJ, Marceau N: Biliary epithelial and hepatocytic cell lineage relationships in embryonic rat liver as determined by the differential expression of cytokeratins, a-fetoprotein, albumin, and cell surface-exposed components, Cancer Res 48:4909–4918, 1988. Germain L, Goyette R, Marceau N: Differential cytokeratin and alpha-fetoprotein expression in morphologically distinct epithelial cells emerging at the early stage of rat hepatocarcinogenesis, Cancer Res 45:673–681, 1985. Gong HY, Lin CJ, Chen MH, et al: Two distinct teleost hepatocyte nuclear factor 1 genes, hnf1alpha/tcf1 and hnf1beta/tcf2, abundantly expressed in liver, pancreas, gut and kidney of zebrafish, Gene 338:35–46, 2004. Gouon-Evans V, Boussemart L, Gadue P, et al: BMP-4 is required for hepatic specification of mouse embryonic stem cell-derived definitive endoderm, Nature Biotech 24:1402–1411, 2006. Gualdi R, Bossard P, Zheng M, et al: Hepatic specification of the gut endoderm in vitro: cell signaling and transcriptional control, Genes Dev 10:1670–1682, 1996. Hamazaki T, Iiboshi Y, Oka M, et al: Hepatic maturation in differentiating embryonic stem cells in vitro, FEBS Lett 497:15–19, 2001.

REFERENCES

999 Harris RG, Herzog EL, Bruscia EM, et al: Lack of a fusion requirement for development of bone marrow-derived epithelia, Science 305:90–93, 2004. Hatzis P, Talianidis I: Regulatory mechanisms controlling human hepatocyte nuclear factor 4alpha gene expression, Mol Cell Biol 21:7320–7330, 2001. Hayner NT, Braun L, Yaswen P, et al: Isozyme profiles of oval cells, parenchymal cells, and biliary cells isolated by centrifugal elutriation from normal and preneoplastic livers, Cancer Res 44:332–338, 1984. Hentsch B, Lyons I, Ruili L, et al: Hlx homeo box gene is essential for an inductive tissue interaction that drives expansion of embryonic liver and gut, Genes Dev 10:70–79, 1996. Holtzinger A, Evans T: Gata4 regulates the formation of multiple organs, Development 132:4005–4014, 2005. Houssaint E: Differentiation of the mouse hepatic primordium. I. An analysis of tissue interactions in hepatocyte differentiation, Cell Differ 9:269–279, 1980. Jackson DA, Rowader KE, Stevens K, et al: Modulation of liver-specific transcription by interactions between hepatocyte nuclear factor 3 and nuclear factor 1 binding DNA in close apposition, Mol Cell Biol 13:2401–2410, 1993. Jacobsen CM, Narita N, Bielinska M, et al: Genetic mosaic analysis reveals that GATA-4 is required for proper differentiation of mouse gastric epithelium, Dev Biol 241:34–46, 2002. Jones EA, Tosh D, Wilson DI, et al: Hepatic differentiation of murine embryonic stem cells, Exp Cell Res 272:15–22, 2002. Jung J, Zheng M, Goldfarb M, Zaret KS: Initiation of mammalian liver development from endoderm by fibroblasts growth factors, Science 284:1998–2003, 1999. Kaestner KH, Katz J, Liu Y, et al: Inactivation of the winged helix transcription factor HNF3alpha affects glucose homeostasis and islet glucagon gene expression in vivo, Genes Dev 13:495–504, 1999. Kamiya A, Kinoshita T, Ito Y, et al: Fetal liver development requires a paracrine action of oncostatin M through the gp130 signal transducer, EMBO J 18:2127–2136, 1999. Kirby ML, Lawson A, Stadt HA, et al: Hensen’s node gives rise to the ventral midline of the foregut: implications for organizing head and heart development, Dev Biol 253:175–188, 2003. Knight B, Matthews VB, Olynyk JK, Yeoh GC: Jekyll and Hyde: evolving perspectives on the function and potential of the adult liver progenitor (oval) cell, Bioessays 27:1192–1202, 2005. Kolterud A, Wandzioch E, Carlsson L: Lhx2 is expressed in the septum transversum mesenchyme that becomes an integral part of the liver and the formation of these cells is independent of functional Lhx2, Gene Expr Patterns 4:521–528, 2004. Kuai XL, Cong XQ, Li XL, Xiao SD: Generation of hepatocytes from cultured mouse embryonic stem cells, Liver Transpl 9:1094–1099, 2003. Kubo A, Shinozaki K, Shannon JM, et al: Development of definitive endoderm from embryonic stem cells in culture, Development 131:1651–1662, 2004. Kumar M, Melton D: Pancreas specification: a budding question, Curr Opin Genet Dev 13:401–407, 2003. Kuo CT, Morrisey EE, Anandappa R, et al: GATA4 transcription factor is required for ventral morphogenesis and heart tube formation, Genes Dev 11:1048–1060, 1997. Kyrmizi I, Hatzis P, Katrakili N, et al: Plasticity and expanding complexity of the hepatic transcription factor network during liver development, Genes Dev 20:2293–2305, 2006. Laconi E, Oren R, Mukhopadhyay DK, et al: Long-term, near-total liver replacement by transplantation of isolated hepatocytes in rats treated with retrorsine, Am J Pathol 153: 319–329, 1998. Landry C, Clotman F, Hioki T, et al: HNF-6 is expressed in endoderm derivatives and nervous system of the mouse embryo and participates to the cross-regulatory network of liver- enriched transcription factors, Dev Biol 192:247–257, 1997. Laverriere AC, MacNeill C, Mueller C, et al: GATA-4/5/6, a subfamily of three transcription factors transcribed in developing heart and gut, J Biol Chem 269:23177–23184, 1994. Lazaro CA, Croager EJ, Mitchell C, et al: Establishment, characterization, and long-term maintenance of cultures of human fetal hepatocytes, Hepatology 38:1095–1106, 2003. Le Douarin NM: An experimental analysis of liver development, Med Biol 53:427–455, 1975. LeCouter J, Moritz DR, Li B, et al: Angiogenesis-independent endothelial protection of liver: role of VEGFR-1, Science 299:890–893, 2003. Lee CS, Friedman JR, Fulmer JT, Kaestner KH: The initiation of liver development is dependent on Foxa transcription factors, Nature 435:944–947, 2005.

1000

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

Lee KD, Kuo TK, Whang-Peng J, et al: In vitro hepatic differentiation of human mesenchymal stem cells, Hepatology 40:1275–1284, 2004. Lemire JM, Fausto N: Multiple alpha-fetoprotein RNAs in adult rat liver: cell type-specific expression and differential regulation, Cancer Res 51:4656–4664, 1991. Li H, Arber S, Jessell TM, Edlund H: Selective agenesis of the dorsal pancreas in mice lacking homeobox gene Hlxb9, Nat Genet 23:67–70, 1999. Li J, Ning G, Duncan SA: Mammalian hepatocyte differentiation requires the transcription factor HNF-4alpha, Genes Dev 14:464–474, 2000. Li L, Krantz ID, Deng Y, et al: Alagille syndrome is caused by mutations in human Jagged1, which encodes a ligand for Notch1, Nat Genet 16:243–251, 1997. Liu JK, DiPersio CM, Zaret KS: Extracellular signals that regulate liver transcription factors during hepatic differentiation in vitro, Mol Cell Biol 11:773–784, 1991. Lorent K, Yeo SY, Oda T, et al: Inhibition of Jagged-mediated Notch signaling disrupts zebrafish biliary development and generates multi-organ defects compatible with an Alagille syndrome phenocopy, Development 131:5753–5766, 2004. Malhi H, Irani AN, Gagandeep S, Gupta S: Isolation of human progenitor liver epithelial cells with extensive replication capacity and differentiation into mature hepatocytes, J Cell Sci 115:2679–2688, 2002. Martin GR: Isolation of a pluripotent cell line from early mouse embryos cultured in medium conditioned by teratocarcinoma stem cells, Proc Natl Acad Sci U S A 78:7634–7638, 1981. Matsumoto K, Yoshitomi H, Rossant J, Zaret KS: Liver organogenesis promoted by endothelial cells prior to vascular function, Science 294:559–563, 2001. Matthews RP, Lorent K, Russo P, Pack M: The zebrafish onecut gene hnf-6 functions in an evolutionarily conserved genetic pathway that regulates vertebrate biliary development, Dev Biol 274:245–259, 2004. McCright B, Lozier J, Gridley T: A mouse model of Alagille syndrome: Notch2 as a genetic modifier of Jag1 haploinsufficiency, Development 129:1075–1082, 2002. Meehan RR, Barlow DP, Hill RE, et al: Pattern of serum protein gene expression in mouse visceral yolk sac and foetal liver, EMBO J 3:1881–1885, 1984. Menthena A, Deb N, Oertel M, et al: Bone marrow progenitors are not the source of expanding oval cells in injured liver, Stem Cells 22:1049–1061, 2004. Minguet S, Cortegano I, Gonzalo P, et al: A population of c-Kit(low)(CD45/TER119)- hepatic cell progenitors of 11-day postcoitus mouse embryo liver reconstitutes cell-depleted liver organoids, J Clin Invest 112:1152–1163, 2003. Miyazaki M, Akiyama I, Sakaguchi M, et al: Improved conditions to induce hepatocytes from rat bone marrow cells in culture, Biochem Biophys Res Commun 298:24–30, 2002. Monaghan AP, Kaestner KH, Grau E, Schu¨tz G: Postimplantation expression patterns indicate a role for the mouse forkhead/HNF-3a, b, and g genes in determination of the definitive endoderm, chordamesoderm and neuroectoderm, Development 119:567–578, 1993. Morrisey EE, Tang Z, Sigrist K, et al: GATA6 regulates HNF4 and is required for differentiation of visceral endoderm in the mouse embryo, Genes Dev 12:3579–3590, 1998. Nagy P, Bisgaard HC, Thorgeirsson SS: Expression of hepatic transcription factors during liver development and oval cell differentiation, J Cell Biol 126:223–233, 1994. Narita N, Bielinska M, Wilson DB: Wild-type visceral endoderm abrogates the ventral developmental defects associated with GATA-4 deficiency in the mouse, Dev Biol 189:270–274, 1997. Nava S, Westgren M, Jaksch M, et al: Characterization of cells in the developing human liver, Differentiation 73:249–260, 2005. Newman CS, Chia F, Krieg PA: The XHex homeobox gene is expressed during development of the vascular endothelium: overexpression leads to an increase in vascular endothelial cell number, Mech Dev 66:83–93, 1997. Ng IO, Chan KL, Shek WH, et al: High frequency of chimerism in transplanted livers, Hepatology 38:989–998, 2003. Ng YK, Iannaccone PM: Fractal geometry of mosaic pattern demonstrates liver regeneration is a self-similar process, Dev Biol 151:419–430, 1992. Ober EA, Verkade H, Field HA, Stainier DY: Mesodermal Wnt2b signalling positively regulates liver specification, Nature 442:688–691, 2006. Oda T, Elkahloun AG, Pike BL, et al: Mutations in the human Jagged1 gene are responsible for Alagille syndrome, Nat Genet 16:235–242, 1997. Odom DT, Dowell RD, Jacobsen ES, et al: Core transcriptional regulatory circuitry in human hepatocytes, Mol Syst Biol 2:1–5, 2006.

REFERENCES

1001 Odom DT, Zizlsperger N, Gordon DB, et al: Control of pancreas and liver gene expression by HNF transcription factors, Science 303:1378–1381, 2004. Oh SH, Miyazaki M, Kouchi H, et al: Hepatocyte growth factor induces differentiation of adult rat bone marrow cells into a hepatocyte lineage in vitro, Biochem Biophys Res Commun 279:500–504, 2000. Omori N, Omori M, Evarts RP, et al: Partial cloning of rat CD34 cDNA and expression during stem cell-dependent liver regeneration in the adult rat, Hepatology 26:720–727, 1997. Overturf K, al-Dhalimy M, Ou CN, et al: Serial transplantation reveals the stem-cell-like regenerative potential of adult mouse hepatocytes, Am J Pathol 151:1273–1280, 1997. Overturf K, Al-Dhalimy M, Tanguay R, et al: Hepatocytes corrected by gene therapy are selected in vivo in a murine model of hereditary tyrosinaemia type I, Nat Genet 12:266–273, 1996. Pack R, Heck R, Dienes HP, et al: Isolation, biochemical characterization, long-term culture, and phenotype modulation of oval cells from carcinogen-fed rats, Exp Cell Res 204:198–209, 1993. Paku S, Schnur J, Nagy P, Thorgeirsson SS: Origin and structural evolution of the early proliferating oval cells in rat liver, Am J Pathol 158:1313–1323, 2001. Parviz F, Matullo C, Garrison WD, et al: Hepatocyte nuclear factor 4alpha controls the development of a hepatic epithelium and liver morphogenesis, Nat Genet 34:292–296, 2003. Petersen BE, Bowen WC, Patrene KD, et al: Bone marrow as a potential source of hepatic oval cells, Science 284:1168–1170, 1999. Petersen BE, Goff JP, Greenberger JS, Michalopoulos GK: Hepatic oval cells express the hematopoietic stem cell marker Thy-1 in the rat, Hepatology 27:433–445, 1998. Petersen BE, Grossbard B, Hatch H, et al: Mouse A6-positive hepatic oval cells also express several hematopoietic stem cell markers, Hepatology 37:632–640, 2003. Pinkert CA, Ornitz DM, Brinster RL, Palmiter RD: An albumin enhancer located 10 kb upstream functions along with its promoter to direct efficient, liver-specific expression in transgenic mice, Genes Dev 1:268–276, 1987. Porter FD, Drago J, Xu Y, et al: Lhx2, a LIM homeobox gene, is required for eye, forebrain, and definitive erythrocyte development, Development 124:2935–2944, 1997. Qin AL, Zhou XQ, Zhang W, et al: Characterization and enrichment of hepatic progenitor cells in adult rat liver, World J Gastroenterol 10:1480–1486, 2004. Reiter JF, Kikuchi Y, Stainier DY: Multiple roles for Gata5 in zebrafish endoderm formation, Development 128:125–135, 2001. Reubinoff BE, Pera MF, Fong CY, et al: Embryonic stem cell lines from human blastocysts: somatic differentiation in vitro, Nat Biotechnol 18:399–404, 2000. Rhim JA, Sandgren EP, Degen JL, et al: Replacement of diseased mouse liver by hepatic cell transplantation, Science 263:1149–1152, 1994. Rogler LE: Selective bipotential differentiation of mouse embryonic hepatoblasts in vitro, Am J Pathol 150:591–602, 1997. Rossi JM, Dunn NR, Hogan BLM, Zaret KS: Distinct mesodermal signals, including BMP’s from the septum transversum mesenchyme, are required in combination for hepatogenesis from the endoderm, Genes Dev 15:1998–2009, 2001. Sandhu JS, Petkov PM, Dabeva MD, Shafritz DA: Stem cell properties and repopulation of the rat liver by fetal liver epithelial progenitor cells, Am J Pathol 159:1323–1334, 2001. Sasaki H, Hogan BLM: Differential expression of multiple fork head related genes during gastrulation and pattern formation in the mouse embryo, Development 118:47–59, 1993. Saxena R, Theise ND, Crawford JM: Microanatomy of the human liver-exploring the hidden interfaces, Hepatology 30:1339–1346, 1999. Schmidt C, Bladt F, Goedecke S, et al: Scatter factor/hepatocyte growth factor is essential for liver development, Nature 373:699–702, 1995. Schuldiner M, Yanuka O, Itskovitz-Eldor J, et al: Effects of eight growth factors on the differentiation of cells derived from human embryonic stem cells, Proc Natl Acad Sci U S A 97:11307–11312, 2000. Schwartz RE, Reyes M, Koodie L, et al: Multipotent adult progenitor cells from bone marrow differentiate into functional hepatocyte-like cells, J Clin Invest 109:1291–1302, 2002. Sell S: Heterogeneity and plasticity of hepatocyte lineage cells, Hepatology 33:738–750, 2001. Sell S, Reynolds RD, Reutter W: Rat alpha 1-fetoprotein: appearance after galactosamine-induced liver injury, J Natl Cancer Inst 53:289–291, 1974. Sell S, Salman J: Light- and electron-microscopic autoradiographic analysis of proliferating cells during the early stages of chemical hepatocarcinogenesis in the rat induced by feeding N-2-fluorenylacetamide in a choline-deficient diet, Am J Pathol 114:287–300, 1984.

1002

EARLY LIVER DEVELOPMENT AND HEPATIC PROGENITOR CELLS

Sellem CH, Frain M, Erdos T, Sala-Trepat JM: Differential expression of albumin and alphafetoprotein genes in fetal tissues of mouse and rat, Dev Biol 102:51–60, 1984. Serls AE, Doherty S, Parvatiyar P, et al: Different thresholds of fibroblast growth factors pattern the ventral foregut into liver and lung, Development 132:35–47, 2005. Shinozuka H, Lombardi B, Sell S, Iammarino RM: Early histological and functional alterations of ethionine liver carcinogenesis in rats fed a choline-deficient diet, Cancer Res 38:1092–1098, 1978. Showell C, Binder O, Conlon FL: T-box genes in early embryogenesis, Dev Dyn 229:201–218, 2004. Sigal SH, Rajvanshi P, Reid LM, Gupta S: Demonstration of differentiation in hepatocyte progenitor cells using dipeptidyl peptidase IV deficient mutant rats, Cell Mol Biol Res 41:39–47, 1995. Slack JMW: Developmental biology of the pancreas, Development 121:1569–1580, 1995. Slott PA, Liu MH, Tavoloni N: Origin, pattern, and mechanism of bile duct proliferation following biliary obstruction in the rat, Gastroenterology 99:466–477, 1990. Sosa-Pineda B, Wigle JT, Oliver G: Hepatocyte migration during liver development requires Prox1, Nat Genet 25:254–255, 2000. Stafford D, Prince V: Retinoic acid signaling is required for a critical early step in zebrafish pancreatic development, Curr Biol 12:1215–1220, 2002. Stainier DY: A glimpse into the molecular entrails of endoderm formation, Genes Dev 16:893–907, 2002. Stamp L, Crosby HA, Hawes SM, et al: A novel cell-surface marker found on human embryonic hepatoblasts and a subpopulation of hepatic biliary epithelial cells, Stem Cells 23:103–112, 2005. Strick-Marchand H, Morosan S, Charneau P, et al: Bipotential mouse embryonic liver stem cell lines contribute to liver regeneration and differentiate as bile ducts and hepatocytes, Proc Natl Acad Sci U S A 101:8360–8365, 2004. Strick-Marchand H, Weiss MC: Inducible differentiation and morphogenesis of bipotential liver cell lines from wild-type mouse embryos, Hepatology 36:794–804, 2002. Suksaweang S, Lin CM, Jiang TX, et al: Morphogenesis of chicken liver: identification of localized growth zones and the role of beta-catenin/Wnt in size regulation, Dev Biol 266:109–122, 2004. Sun Z, Hopkins N: vhnf1, the MODY5 and familial GCKD-associated gene, regulates regional specification of the zebrafish gut, pronephros, and hindbrain, Genes Dev 15:3217–3229, 2001. Suzuki A, Zheng YW, Kaneko S, et al: Clonal identification and characterization of self-renewing pluripotent stem cells in the developing liver, J Cell Biol 156:173–184, 2002. Tada S, Era T, Furusawa C, et al: Characterization of mesendoderm: a diverging point of the definitive endoderm and mesoderm in embryonic stem cell differentiation culture, Development 132:4363–4374, 2005. Teratani T, Yamamoto H, Aoyagi K, et al: Direct hepatic fate specification from mouse embryonic stem cells, Hepatology 41:836–846, 2005. Theise ND, Saxena R, Portmann BC, et al: The canals of Hering and hepatic stem cells in humans, Hepatology 30:1425–1433, 1999. Thomson JA, Itskovitz-Eldor J, Shapiro SS, et al: Embryonic stem cell lines derived from human blastocysts, Science 282:1145–1147, 1998. Thorgeirsson SS: Hepatic stem cells in liver regeneration, FASEB J 10:1249–1256, 1996. Tremblay KD, Zaret KS: Distinct populations of endoderm cells converge to generate the embryonic liver bud and ventral foregut tissues, Dev Biol 280:87–99, 2005. Vassilopoulos G, Wang PR, Russell DW: Transplanted bone marrow regenerates liver by cell fusion, Nature 422:901–904, 2003. Wagers AJ, Sherwood RI, Christensen JL, Weissman IL: Little evidence for developmental plasticity of adult hematopoietic stem cells, Science 297:2256–2259, 2002. Wallace KN, Yusuff S, Sonntag JM, et al: Zebrafish hhex regulates liver development and digestive organ chirality, Genesis 30:141–143, 2001. Wandzioch E, Kolterud A, Jacobsson M, et al: Lhx2-/- mice develop liver fibrosis, Proc Natl Acad Sci U S A 101:16549–16554, 2004. Wang X, Foster M, Al-Dhalimy M, et al: The origin and liver repopulating capacity of murine oval cells, Proc Natl Acad Sci U S A 100(Suppl 1):11881–11888, 2003a. Wang X, Willenbring H, Akkari Y, et al: Cell fusion is the principal source of bone-marrowderived hepatocytes, Nature 422:897–901, 2003b.

ADULT LIVER PROGENITOR CELLS

1003

Weinstein DC, Ruiz i Altaba A, Chen A, et al: The winged-helix transcription factor HNF-3b is required for notochord development in the mouse embryo, Cell 78:575–588, 1994. Weinstein M, Monga SP, Liu Y, et al: Smad proteins and hepatocyte growth factor control parallel regulatory pathways that converge on beta1-integrin to promote normal liver development, Mol Cell Biol 21:5122–5131, 2001. Yamamoto H, Quinn G, Asari A, et al: Differentiation of embryonic stem cells into hepatocytes: biological functions and therapeutic application, Hepatology 37:983–993, 2003. Yamamoto Y, Teratani T, Yamamoto H, et al: Recapitulation of in vivo gene expression during hepatic differentiation from murine embryonic stem cells, Hepatology 42:558–567, 2005. Yasui O, Miura N, Terada K, et al: Isolation of oval cells from Long-Evans Cinnamon rats and their transformation into hepatocytes in vivo in the rat liver, Hepatology 25:329–334, 1997. Yaswen P, Hayner NT, Fausto N: Isolation of oval cells by centrifugal elutriation and comparison with other cell types purified from normal and preneoplastic livers, Cancer Res 44:324–331, 1984. Yatskievych TA, Pascoe S, Antin PB: Expression of the homeobox gene Hex during early stages of chick embryo development, Mech Dev 80:107–109, 1999. Yin L, Sun M, Ilic Z, et al: Derivation, characterization, and phenotypic variation of hepatic progenitor cell lines isolated from adult rats, Hepatology 35:315–324, 2002. Zaret KS: Regulatory phases of early liver development: paradigms of organogenesis, Nat Rev Genet 3:499–512, 2002. Zhang W, Yatskievych TA, Baker RK, Antin PB: Regulation of Hex gene expression and initial stages of avian hepatogenesis by Bmp and Fgf signaling, Dev Biol 268:312–326, 2004. Zhang W, Yatskievych TA, Cao X, Antin PB: Regulation of Hex gene expression by a Smadsdependent signaling pathway, J Biol Chem 277:45435–45441, 2002. Zhao R, Watt AJ, Li J, et al: GATA6 is essential for embryonic development of the liver but dispensable for early heart formation, Mol Cell Biol 25:2622–2631, 2005.

RECOMMENDED READING LIVER DEVELOPMENT Zaret KS: Regulatory phases of early liver development: paradigms of organogenesis, Nat Rev Genet 3:499–512, 2002. Zhao R, Duncan SA: Embryonic development of the liver, Hepatology 41:956–967, 2005.

ADULT LIVER PROGENITOR CELLS Fausto N: Liver regeneration and repair: hepatocytes, progenitor cells, and stem cells, Hepatology 39:1477–1487, 2004. Fausto N, Campbell JS: The role of hepatocytes and oval cells in liver repopulation and regeneration, Mech Dev 120:117–130, 2003. Sell S: Heterogeneity and plasticity of hepatocyte lineage cells, Hepatology 33:738–750, 2001.

44

INTESTINAL STEM CELLS IN PHYSIOLOGIC REGENERATION AND DISEASE DAVID H. SCOVILLE,1,3 XI C. HE,1 GOO LEE,2 TOSHIRO SATO,1 TERRENCE A. BARRETT,2 and LINHENG LI1,3 1

Stowers Institute for Medical Research, Kansas City, MO

2

Northwestern University Feinberg School of Medicine, Department of Medicine and Microbiology/Immunology, Division of Gastroenterology, Evanston, IL 3 University of Kansas School of Medicine, Department of Pathology and Laboratory Medicine, Lawrence, KS

INTRODUCTION After normal development, many tissues maintain the ability to regenerate for repairing damaged tissues and lost cells by emulating earlier developmental pathways. The intestine represents an ideal system for the study of adult tissue stem cells because of its well-organized structure and its need for constant cellular regeneration. Intestinal stem cells, which are maintained in a lessdifferentiated state, are required for both the initial development and maintenance of the crypt–villus structure. In this chapter, we discuss the signaling pathways responsible for controlling stem cell behavior and epithelial cell differentiation. Finally, we discuss how alterations in these pathways contribute to cancer development in the intestine.

I. OVERVIEW OF INTESTINAL ANATOMY AND FUNCTIONAL HISTOLOGY The intestinal tract is composed of an endodermally derived epithelium and a mesodermally derived stroma. It differentiates from the primitive gut tube into several different regions with varying functions (see Chapter 40). beginning rostrally, the gastrointestinal tract is composed of the nasopharynx, the esophagus, and the stomach, which perform bulk transport and mechanical and enzymatic digestion. next, the small intestine is divided along its length into the duodenum, the jejunum, and the ileum, which aid with both digestion

1004

Principles of Developmental Genetics

© 2007, Elsevier Inc. All rights reserved.

OVERVIEW OF INTESTINAL ANATOMY AND FUNCTIONAL HISTOLOGY

1005

and absorption. the duodenum absorbs electrolytes, monosaccharides, and water- and fat-soluble vitamins; the jejunum absorbs protein and fat; the ileum is responsible for vitamin b12 and bile salt uptake. finally, the colon concentrates the remaining fecal contents. for the purposes of this chapter, our discussion will be limited to the small intestine and the colon. The small intestinal architecture is composed of contiguous villi and crypts. The villi, which are lined by a single layer of columnar epithelium, extend into the intestinal lumen and contain terminally differentiated cells; alternatively, the crypt, which harbors the proliferative potential of this tissue, is the result of epithelial invaginations into the gut mucosa. Intestinal stem cells (ISCs), which possess the capacity for self-renewal and the generation of multipotent transient amplifying progenitors, reside near the crypt base (Booth and Potten, 2000). Progressing upward from the intestinal stem cell position, the transient amplifying progenitors occupy the rest of the crypt structure, whereas functionally mature cells are located along the villus (Hermiston and Gordon, 1995). Thus, the crypt–villus architecture can be divided into proliferation and differentiation compartments (Figure 44.1). Mature intestinal epithelial cells are derived from transient amplifying progenitors that specify absorptive and secretory lineages. The absorptive lineage contains only one mature cell type, enterocytes, whereas the secretory lineage includes three differentiated cell types: goblet cells, enteroendocrine cells, and Paneth cells. Enterocytes, goblet cells, and enteroendocrine cells migrate upward from the crypt to the villus, eventually undergoing apoptosis at the villus tip and shedding into the lumen, a process that takes about 3 to 5 days in mice. Interestingly, Paneth cells, which differentiate from secretory progenitor cells within the crypt, do not follow the migrational pattern of the other cell types. Instead, they migrate in the opposite direction toward the crypt base. In addition, Paneth cells also have a longer life span. Enterocytes, which are located along the villus, are the most abundant cell type in the small intestine, and their primary function is to absorb nutrients from the intestinal lumen. To aid in this process, these cells secrete hydrolases

FIGURE 44.1 Intestinal structure and compartmentalization. The small intestine is composed of crypts and villi which correspond to compartments of proliferation and differentiation, respectively. Crypts contain ISCs and progenitors, as well as differentiated Paneth cells at the base. The other three mature cell types are located along the villus.

1006

INTESTINAL STEM CELLS IN PHYSIOLOGIC REGENERATION AND DISEASE

and contain an enzyme-rich, brush-like border along the apical microvilli that serves to dramatically increase the surface area available for digestion and absorption. These enzymes break down small peptides and carbohydrates into components that can be easily absorbed. Goblet cells function to secrete mucins that protect the epithelium from the harsh digestive environment of the intestinal lumen. These cells—and, correspondingly, the mucus content in the gut—increase in number from the duodenum to the colon. Enteroendocrine cells comprise less than 1% of the total population of intestinal epithelial cells, making them the rarest differentiated cell type. Different types of enteroendocrine cells exist, and they are classified on the basis of the molecules that they secrete, such as serotonin, cholecystokinin, substance P, and secretin (Evans and Potten, 1988; Hocker and Wiedenmann, 1998). These hormones and neuropeptides are secreted from the basolateral membrane, and they act in an endocrine fashion, in part to regulate gastrointestinal motility. Paneth cells function to control microbial growth within the gastrointestinal tract by secreting several antimicrobial agents, including cryptdin, defensins, and lysozyme (Ayabe et al., 2000; Porter et al., 2002). These secretions can also be used in immunohistochemical staining to distinguish this cell type. The structure of the colon is slightly different in that it lacks villi as well as Paneth cells. The stem cells are proposed to reside at the bottom of the crypts; progenitor cells occupy the bottom two thirds of the crypt up from the stem cell region, whereas that one third of the crypt closest to the lumen is occupied by terminally differentiated cells (Booth and Potten, 2000). In other words, the structure and division of cellular compartments in the colon are the same as that seen in the small intestine, with the exception that there are neither Paneth cells nor villi present.

II. INTESTINAL STEM CELLS Although direct functional proof of ISCs has yet to be demonstrated, significant evidence supports their existence. Historically, the presence and location of the putative ISCs have been recognized using clonal analysis techniques, DNA-labeling studies, and immunohistochemistry. Initially, the monoclonal nature of intestinal crypts was established using various chimeric or heterozygous mutant mice strains. It was observed that individual crypts were completely composed of cells either from only one of the mouse strains, such as in the case of chimeras (Hermiston et al., 1993; Roth et al., 1991; Schmidt et al., 1988), or they exhibited the same mutant gene expression profile, such as in the case of heterozygous mice (Bjerknes and Cheng, 1999; Roth et al., 1991). Furthermore, experiments in which the intestinal epithelium was subjected to a cytotoxic agent showed that a single surviving cell could regenerate the entire crypt (Potten, 1995). These results can be explained by a common stem cell yielding a monoclonal crypt cell population of multilineage progenitors, whereas villi, which are maintained by several crypts, are as a result polyclonal (Booth and Potten, 2000). Furthermore, DNA-label–retaining experiments using bromodeoxyuridine (BrdU) or 3H-thymidine both supported the existence of ISCs and localized them to a position near the crypt base (Potten et al., 2002). Experimentally, these labels incorporate into DNA and diminish with each cell division; however, cells that retain the label over long periods of time are thought to be predominantly slow cycling or in the quiescent state. Although

STEM CELL/NICHE INTERACTIONS

1007

progenitors are rapidly proliferating, adult tissue stem cells are thought to be slow cycling, and thus they can be distinguished using this method. Alternatively, long-term label retention can be explained by DNA segregation in asymmetric stem cell division. Recent evidence indicates that stem cell chromatids are preferentially segregated such that one chromatid is always retained in the parent stem cell, thereby yielding an “immortal chromatid” (Cairns, 1975). This could imply that label retention by these specific crypt cells identifies a characteristic stem cell function. Although mechanistic details remain unknown, this unique segregation would ensure that ISC DNA is preferentially protected from mutagenesis, thus decreasing the risk of carcinogenesis in this highly proliferative epithelium (Potten et al., 2002; Shinin et al., 2006). Immunohistochemical evidence has shown that long-term BrdU-retaining cells do not co-stain with lysozyme (a specific Paneth cell marker) or Ki67 (a cell proliferation marker), thus distinguishing these cells and designating them as putative slow-cycling ISCs (He et al., 2004). So far, much research has focused on finding novel and specific ISC markers in an attempt to further pinpoint their location. Molecules such as Musashi-1, phosphorylated phosphatase homologue of tensin (P-PTEN), phosphorylated Akt (P-Akt), and 14–3-3z have been found to be expressed in long-term BrdU retaining cells in the proposed ISC position as well as in some progenitor cells (Batlle et al., 2002; Booth and Potten, 2000; He et al., 2004; Korinek et al., 1998; Potten et al., 2003). These studies and others have been useful for localizing the ISC to a position about four to five cells above the crypt base (Potten, 1995). Alternatively, other experiments identified candidate stem cells intermingled with Paneth cells. These experiments showed that the chemical treatment of mice could mutate the lectin-expression profile of a single morphologically columnar cell. These cells, which are located at the crypt base, were found to give rise to mutant clones containing multiple cell types. In addition to their mixed cell population, the generation time for these clones indicated a stem cell or long-term progenitor origin (Bjerknes and Cheng, 1999). Despite these correlative experiments, demonstrative proof of putative stem cells using an appropriate in vivo method for functional characterization remains elusive. This is in part the result of technical challenges in isolating pure ISCs. However, recent reports have indicated the possibility of isolating fractions significantly enriched for ISCs using flow cytometry (Dekaney et al., 2005). Putative ISC markers have also provided insights into stem cell regulatory pathways. Potential interactions between the epithelium and adjacent stromal tissue have been an active area of research in an attempt to define the intestinal stem cell niche. Studies regarding signals emanating from the niche have and will provide insight into our understanding of stem cell self-renewal and proliferation as well as cell fate determination.

III. STEM CELL/NICHE INTERACTIONS The idea of a stem cell niche, which was conceived with regard for hematopoietic stem cells, proposed that a specific microenvironment exists for these unique cells (Schofield, 1978). This concept of a stem cell niche has carried over into the study of other adult stem cell systems. From pioneering experiments in Drosophila, Caenorhabditis elegans, and other mammalian systems, some basic and overarching concepts regarding stem cell niche function have been revealed (Doetsch, 2003; Fuchs et al., 2004; Li and Xie, 2005; Lin, 2002;

1008

INTESTINAL STEM CELLS IN PHYSIOLOGIC REGENERATION AND DISEASE

Spradling et al., 2001). The stem cell niche is composed of a special group of cells that provide a microenvironment to which stem cells directly attach. Although other mesenchymal microenvironments may exist along the length of the crypt–villus axis, niche interactions serve to specifically inhibit the differentiation and control the self-renewal and proliferation of stem cells. Because cellular regeneration is initiated at the stem cell level, regulating the properties of these cells is critical. Myoepithelial fibroblasts in the crypt epithelium are proposed to be a niche cell candidate in part as a result of both their proximity to and expression of prospective stem cell regulatory molecules (He et al., 2004; Mills and Gordon, 2001). Although incomplete, several stem cell–niche interactions have been well defined. In intestinal development, Hedgehog signaling from epithelial cells to mesenchymal cells serves to inhibit ectopic crypt formation within the villi and areas adjacent to already forming crypts. Thus, during development, Hedgehog signals ensure appropriate crypt spacing and numbers (Madison et al., 2005). Recently, Hedgehog signaling has also been shown to play a role in controlling proliferative signaling in adult intestinal tissue by restricting Wnt target gene expression in the colon (van den Brink et al., 2004). Although Wnt signaling is well known to promote small intestinal epithelial propagation, recent studies demonstrated that signaling through bone morphogenic proteins (BMPs) represents a mechanism for suppressing ISC proliferation (Haramis et al., 2004; He et al., 2004). In addition, various signals, including Wnt and Notch, are involved in cell fate determination. Unraveling these complex signaling mechanisms has led to a better understanding of the molecular and physiologic interactions that control stem cell fate as well as potential causes of tumorigenesis within the intestine (Radtke and Clevers, 2005).

IV. SIGNALING MECHANISMS CONTROLLING INTESTINAL STEM CELL SELF-RENEWAL, PROLIFERATION, AND DIFFERENTIATION A. Wnt Signaling Wnt signaling is known to play several roles within the intestine. First, it maintains stem cells via cell cycle control (Sancho et al., 2004). Second, Wnt signals mediate cell fate determination. Third, Wnt signaling controls the localization of cells along the crypt–villus axis. Although a variety of different types of Wnt signaling can occur, canonical Wnt pathway activation controls b-catenin localization. When the Wnt signal is not present, the complex of adenomatous polyposis coli/glycogen synthase kinase 3 beta (APC/GSK3b) can bind b-catenin (Munemitsu et al., 1995; Rubinfeld et al., 1996; Rubinfeld et al., 1993; Su et al., 1993). GSK3b, a protein kinase, can then phosphorylate b-catenin, thereby targeting this molecule for proteosomal destruction (Giles et al., 2003). Without the Wnt signal, T-cell factors (TCFs) act as transcriptional repressors. Conversely, Wnt receptor activation results in the nuclear localization of b-catenin, where it binds members of the TCF transcription factor family, activating target gene expression (Figure 44.2; Giles et al., 2003; Nusse et al., 1997; Peifer and Polakis, 2000; Sahl and Clever, 1994a, 1994b). Thus, upon ligand binding, Wnt/b-catenin activation in the ISC converts TCF into a transcriptional activator that initiates stem cell activation via the upregulation of genes such as cyclinD and c-myc, which are well known for promoting cell cycle progression.

SIGNALING MECHANISMS CONTROLLING INTESTINAL STEM

ISC signaling pathways. Wnt, Notch, BMP, and PTEN and PI3K pathways are involved in maintaining ISCs and differentiation programming within the intestine. Notch signaling results in transcriptional activation which drives differentiation of absorptive progenitors and is also a proliferative signal within the crypt. Wnt signaling results in nuclear localization of b-catenin which when bound to TCF factors converts them from repressors to transcriptional activators upregulating genes involved in activating the ISC cell cycle. PI3K signaling activates Akt which can assist Wnt in regulating b-catenin in certain types of cells including ISCs. PTEN controls Akt activation by inhibiting PI3K. Although the exact mechanism remains unknown, BMP signaling can inhibit proliferation in ISCs perhaps in part through the Smad transcription factors and in part via interaction with PTEN.

FIGURE 44.2

1009

1010

INTESTINAL STEM CELLS IN PHYSIOLOGIC REGENERATION AND DISEASE

Normally, stem cells undergo niche-dependent asymmetric division. This division results in a daughter cell, which migrates outside the niche, and a stem cell, which remains attached to the niche. Upon the completion of cell division, niche interaction ensures the return of TCF repression in the stem cell, whereas the more differentiated daughter cell progresses through the proliferation and differentiation program as a result of continued nuclear b-catenin/TCF-driven gene expression (Pinto et al., 2003; van de Wetering et al., 2002). Experimental evidence supports this model, because the deletion of TCF4 results in an absence of proliferating cells and a loss of crypt formation (Korinek et al., 1998). Additionally, APC has been shown to control crypt epithelial proliferation. In humans and mice, hereditary mutation of one of the APC-encoding alleles creates a predisposition for numerous intestinal adenomas and subsequent carcinoma development in familial adenomatous polyposis (FAP). This APC mutation allows for the constitutive nuclear localization of b-catenin, which results in overactive target gene expression and an enlargement in crypt size as a result of an increase in proliferating cells (Groden et al., 1991; Joslyn et al., 1991; Korinek et al., 1997; Miyoshi et al., 1992; Nakamura et al., 1991). In fact, mutations in the Wnt signaling cascade, including APC, may be the initiating event in most human colorectal cancers (Bienz and Clevers, 2000). Wnt signaling also affects cell fate determination by promoting general secretory lineage commitment as well as by driving Paneth cell gene expression (Figure 44.3). Evidence for this is based on several mouse intestinal

Intestinal cell fate determination. SP, Secretory progenitor; AP, absorptive progenitor. Notch and Wnt signaling are known to effect cell fate decision. Wnt activation induces early secretory programming. Although not linked to Wnt signaling, Math-1, a transcription factor, drives cells toward an initial secretory fate. Further expression of neurogenin-3 (ngn-3), Klf-4, and b-catenin induce terminal differentiation into enteroendocrine cell, goblet cells, and Paneth cells, respectively. Notch activated progenitor cells progress toward an absorptive fate through Hes-1 target gene expression. In fact, Hes-1 inhibits Math-1 expression, further strengthening this initial absorptive decision. (See color insert.)

FIGURE 44.3

SIGNALING MECHANISMS CONTROLLING INTESTINAL STEM

1011

phenotypes. The overexpression of Dkk1, a Wnt inhibitor, causes a loss of all secretory cell types, although absorptive enterocytes differentiate normally (Kuhnert et al., 2004; Pinto et al., 2003). In addition, nuclear b-catenin is required for the expression of functional Paneth cell molecules, including EphB3 and cryptdin/defensin (Andreu et al., 2005; Batlle et al., 2002). Finally, Wnt signaling has been shown to be important in cellular migration and localization along the crypt–villus epithelium. Eph receptor molecules are targets of b-catenin–mediated gene expression, and Eph/Ephrin signals aid in compartmentalizing proliferative cells in the crypt from the nonproliferative, more differentiated cells along the villus (Batlle et al., 2002; Shih et al., 2001; van de Wetering et al., 2002). Cells expressing transmembrane ephrin ligands and adjacent cells expressing Eph receptors exhibit repulsive properties, and they have been shown to be important for maintaining cellular boundaries and establishing migratory paths (Xu et al., 1999). In fact, EphB2 and EphB3 have been shown to be expressed on progenitor and Paneth cells, respectively, whereas ephrin B1 (one of the Eph ligands) is expressed at the crypt–villus junction. The deletion of EphB2 and EphB3 results in the dispersion of Paneth cells, which are normally restricted to the bottom of the crypt, along the crypt–villus axis. Additionally, cells expressing ephrinB1, which are normally restricted to the crypt–villus junction, can be found within the crypt (Batlle et al., 2002). Thus, Wnt signaling through nuclear b-catenin drives the expression of EphB2 in crypt cells as well as of EphB3 in Paneth cells, thereby ensuring the proper segregation of the proliferative cells and the Paneth cells of the crypt from the differentiated cells of the villus (Batlle et al., 2002; van de Wetering et al., 2002). B. Bone Morphogenic Protein Signaling Although Wnt signaling stimulates stem cell and progenitor proliferation in the crypt, BMPs inhibit cellular division in the ISCs and the villus epithelium. BMP signaling occurs via the dimerization of type I and II BMP receptor proteins, which have intrinsic Ser/Thr kinase activity. After dimerization, type II phosphorylates type I, which results in the activation of receptor-regulated Smad (R-Smad) via phosphorylation. R-Smad then forms a heterodimer with Smad4, which is a common Smad. This R-Smad/Smad4 complex is translocated to the nucleus, and it can act as a transcriptional regulator, eliciting both the activation and repression of target gene expression through coactivator or corepressor binding (see Figure 44.2; Shi and Massague, 2003; see Chapter 1). The expression of BMP2 and BMP4 is limited to the mesenchymal tissue surrounding both the villus epithelium and the stem cells within the crypt (Hardwick et al., 2004). Additionally, BMPR1A has also been identified in ISCs as well as villus epithelial cells, but not in transient amplifying progenitors (He et al., 2004). Active BMP signaling has been confirmed in these BMPR1Aþ cells by the identification of phosphorylated Smad1, Smad5, and Smad8 (these are R-Smads; He et al., 2004). Human genetic studies have reported BMPR1A and Smad4 mutations in juvenile polyposis, which is a syndrome that involves numerous hamartomatous polyps in the small intestine (Howe et al., 1998a; 1998b; 2001). Additionally, BMPR1A knockout mice and transgenic mice overexpressing the BMP antagonist Noggin both demonstrate an overabundance of crypt structures and enhanced Wnt/b-catenin activity (Haramis et al., 2004; He et al., 2004). As indicated previously, two active BMP signaling regions have been identified: one in the villus epithelium and the other within the putative intestinal

1012

INTESTINAL STEM CELLS IN PHYSIOLOGIC REGENERATION AND DISEASE

stem cells in the crypt (He et al., 2004). BMP signaling within the villus region must remain within these terminally differentiated cells to inhibit ectopic crypt formation (Haramis et al., 2004). Although the exact mechanism remains unknown, BMP signaling can inhibit proliferation in ISCs perhaps in part through the Smad transcription factors and in part via interaction with PTEN. However, the inhibition of stem cell self-renewal via BMP signaling must be balanced by other permissive signals and coordinately regulated, thus implying a more intricate system. C. PTEN/PI3K/Akt Signaling Wnt signaling, which is measured by phosphorylated low density lipoproteinreceptor-related protein 6 (LRP6), is found in all proliferating crypt cells, including stem cells, whereas only a small number of these cells actually show b-catenin–driven transcriptional activity using a Top-Gal transgenic mouse model in which a LacZ reporter gene is driven by a TCF optimal promoter (He et al., 2004). Therefore, other signals may assist Wnt in the regulation of b-catenin, particularly in stem cells. Indeed, PTEN, which is already reported to control stem cell maintenance in the hematopoietic stem cell, may play a similar role in the intestine (Zhang et al., 2006). PTEN is a tumor suppressor that displays tyrosine phosphatase and lipid phosphatase activity. Functionally, PTEN antagonizes PI3 kinase (PI3K) activity, thereby inhibiting Akt, a Ser/Thr kinase that is the main downstream mediator of the PI3K pathway. PTEN cannot be recruited to the membrane to suppress PI3K after its phosphorylation (see Figure 44.2; Vazquez et al., 2001). Interestingly, mutations in PTEN have been identified in patients with Cowden disease and Bannayan–Zonana syndrome, both of which share intestinal polyp development (Liaw et al., 1997; Marsh et al., 1997). Because defects in BMP, PTEN, and Wnt signaling pathways can all result in similar intestinal pathology, it is possible that these signaling pathways comprise a coordinated regulatory mechanism for intestinal stem/progenitor cell regulation. Indeed, Akt has been shown to effect b-catenin transcriptional activity. Evidence indicates that Akt can phosphorylate b-catenin, directly promoting 14–3-3z binding and facilitating b-catenin stabilization (Tian et al., 2004). Indeed, phosphorylated PTEN has been identified along with activated Akt and 14–3-3z in ISCs (He et al., 2004). Additionally, active PTEN has been shown to decrease nuclear levels of b-catenin (Persad et al., 2001), whereas recent work has shown that loss of PTEN in the crypt epithelium and surrounding stroma coordinately results in an increased number of ISCs and progenitors displaying active Akt signaling, nuclear b-catenin, and cyclinD1 and correspondingly increased cell cycle activity. This evidence establishes the role of PTEN as a negative regulator of PI3K/Akt signaling in ISCs and progenitors. Because BMP signaling has been shown to positively regulate PTEN (Waite and Eng, 2003), whereas Noggin antagonizes BMP signaling, these two signals can be used to regulate PTEN control on Akt activity which assists Wnt signaling in controlling b-catenin activity within the ISC (He et al., 2004; He et al., 2007). D. Notch Signaling Signaling mechanisms regulating cell fate determination within the intestine are just as important as those controlling proliferation. They ensure an appropriate mixture of fully differentiated cell types to facilitate proper intestinal

SIGNALING MECHANISMS CONTROLLING INTESTINAL STEM

1013

epithelial function. Notch signaling contributes to the decision-making apparatus developed to enable appropriate cellular specialization. Notch signaling occurs through the interaction of transmembrane molecules on adjacent cells. Delta and Jagged family transmembrane ligands on one cell can activate Notch receptors on an adjacent cell (Artavanis-Tsakonas et al., 1999). This activation results in a series of unique events involving the transmembrane and intracellular domains of the single-pass transmembrane Notch receptor. Upon ligand binding, g-secretase activity within the membrane cleaves the intracellular domain, thus freeing it from the cytoplasmic surface. The intracellular domain of Notch (NICD) translocates to the nucleus, where it binds CSL/RBPJk and activates target gene expression, including the expression of other transcriptional regulators such as Hes family proteins (see Figure 44.2; Artavanis-Tsakonas et al., 1999). Virtually all Notch expression in the intestine is limited to the crypt epithelium, which suggests that Notch signaling occurs among crypt cells (Schroder and Gossler, 2002). Logically, cell fate decisions would need to be made within the crypt progenitor population so that after cells migrated to the villus, they would already be programmed with a specific cell fate. Because two lineages exist within the intestinal epithelium (one secretory and the other absorptive), it stands to reason that the first fate decision is between these two pathways (see Figure 44.3). Evidence for Notch signaling in this initial determination has come from studies in which Notch pathway intermediates have been mutated or deleted. Notch inhibition via the deletion of CSL/RBPJk or treatment with g-secretase inhibitor results in an intestinal epithelium composed almost entirely of goblet cells (Milano et al., 2004; van Es et al., 2005). The inhibition of Notch signaling by deleting Hes also causes an overabundance of goblet and enteroendocrine cells (Jensen et al., 2000). Conversely, the constitutive expression of NICD results in an epithelium that is devoid of all secretory cells, including Paneth cells (Fre et al., 2005). Thus, Notchactivated cells are inhibited from progressing down the secretory lineage, and they are destined to become enterocytes. Furthermore, secretory cells avoid Notch signaling by expressing Notch ligands and inducing Notch activation in adjacent cells, thereby inhibiting neighboring cells from differentiating down a similar pathway and ensuring an appropriate ratio of secretory and absorptive cells (Crosnier et al., 2005). Mouse atonal homologue 1 (Math1) is a transcription factor that is expressed in secretory progenitors. Math1 mutant mice display a phenotype that is similar to that of transgenic NICD mice in that they have a complete lack of secretory cell types, although absorptive enterocytes develop normally (Yang et al., 2001). Hes1 in intestinal progenitor cells inhibits Math1 expression by inducing an absorptive fate determination (Jensen et al., 2000). Thus, Notchþ/Math1– cells become absorptive enterocytes, whereas Notch–/Math1þ cells progress toward a secretory fate. Furthermore, other factors are involved in later secretory fate decisions. Neurogenin-3 mutant mice lack enteroendocrine cells, but they contain all other cell types, which indicates that this transcription factor pushes development toward enteroendocrine cells that can specialize even further on the basis of the molecules that they are induced to produce and secrete (Jenny et al., 2002). Goblet cells express Kruppel-like factor 4, which, on the basis of Kruppel-like factor 4 mutant experiments, is important in the differentiation of this secretory cell type (Katz et al., 2002). Thus, although a complete understanding of the signals providing cues for

1014

INTESTINAL STEM CELLS IN PHYSIOLOGIC REGENERATION AND DISEASE

cellular fate decisions is lacking, our knowledge about these decision-making processes has grown tremendously during recent years (see Figure 44.3). Finally, Wnt and Notch pathways may also be linked, and this provides more evidence for multiple regulatory signaling mechanisms controlling stem cell maintenance as well as intestinal differentiation. Interestingly, when APCmin mice bearing a germline mutation in APC are treated with a g-secretase inhibitor to block Notch signaling, not only do all the intestinal epithelial cells become secretory in nature, but proliferation within the crypt ceases (van Es et al., 2005). In addition, if Notch signaling is overactive in an area of inactive Wnt signaling, proliferation is absent (Zecchini et al., 2005). Finally, increases in proliferative cell number occur when Notch is constitutively activated in areas of canonical Wnt activity within the crypt (Fre et al., 2005). Thus, it would seem that both Notch and Wnt signaling together are required for driving the proliferation of ISC/progenitor cells and for the control of appropriate cell fate determination.

V. INTESTINAL STEM CELLS IN HUMAN DISEASE A. Stem Cells in Colorectal Cancer Genetic analyses of sporadic and colitis-induced colorectal cancer (CRC) tissues indicate that genes regulating essential functions of intestinal stem cells are frequently mutated early during neoplastic transformation (Radtke and Clevers, 2005). The model of colonic neoplastic transformation proposed by Fearon and Vogelstein (1990; Figure 44.4) suggests that CRC results from gene mutations and chromosomal instability that occur in a preferred sequence (Radtke and Clevers, 2005). Mutations occur in genes that regulate the self-renewal, proliferation, and cell fate decisions of ISCs. Early mutations

FIGURE 44.4 Intestinal carcinoma development. This figure outlines the proposed model for tumorigenesis within the intestine, along with corresponding human colonoscopic images. Progression proceeds as various genetic mutations accumulate in a preferred sequence. Initially, mutations in Wnt pathway components such as APC, b-catenin, and Cox-2 lead to aberrant crypt foci. These ectopic crypts, outlined with an indigo carmine dye, can be detected via chromoendoscopy. APC mutations can also result in genomic instability, leading to development of an early adenoma. Further mutations in MAP kinase signaling proteins like K-Ras and then TGFb pathway components, such as DCC (deleted in colorectal carcinoma) and Smad4 genes result in progression toward an intermediate and late adenoma, respectively. Finally, mutations in PI3K subunits (i.e., PIK3CA) and p53 lead to cancer, shown here as a locally invasive exophytic carcinoma. (See color insert.)

INTESTINAL STEM CELLS IN HUMAN DISEASE

1015

of APC enhance the risk for other mutations by increasing genomic instability. As lesions advance, their severity correlates with additive mutations in K-ras, Smad4, and p53. The combination of these effects endows tumor cells with enhanced proliferation, resistance to apoptosis, and the capacity for local invasion and metastatic spread. At this time, it is unclear whether these neoplastic alterations (e.g., constitutive Wnt/b-catenin signaling) act through rare stem cells or promote the dedifferentiation of progenitor cells. In addition, clarification is needed regarding the behavior of activated intestinal stem cells and cancer stem cells that emerge in CRC. The earliest mutation associated with CRC development involves the APC gene. APC is mutated in familial adenomatous polyposis and in 85% of sporadic CRC (Grady and Markowitz, 2002; Miyaki et al., 1994). The mutational inactivation of APC causes epithelial hyperproliferation and precancerous aberrant crypt foci (Jen et al., 1994), which is the earliest lesion in colonic neoplasia. Thus, the subsequent induction of b-catenin target genes such as goblet-cell–associated intestinal trefoil factor and mucin 2 or Paneth cell matrilysin (matrix metalloproteinase 7) and cryptdin-1 may explain why some CRCs express “tumor markers” associated with secretory lineage cells (Blank et al., 1994). Alterations of Wnt/b-catenin signaling by mutated proteins other than APC provide clues to the early pathogenesis of CRC. For example, mutations in b-catenin that prevent degradation by GSK3b are less likely to progress to large adenomas and CRC than lesions with APC mutations (Samowitz et al., 1999). One explanation for the difference in outcome may relate to data that suggest that loss of APC increases genomic instability. Thus, lesions harboring APC mutations also carry the risk for polyploidization and chromosomal instability (Fodde et al., 2001; Kaplan et al., 2001). The impact of this effect is compounded by the localization of APC mutations and subsequent chromosomal alterations in long-lived cancer stem cells. Similar additive effects of mutations in K-ras and p53 have been described. Both genes are mutated in 50% of human CRCs. Recent results from murine models indicate that the transgenic expression of an oncogenic K-ras mutation (K-rasV12G) causes aberrant crypt foci with subsequent invasive adenocarcinoma (Janssen et al., 2002). Interestingly, more than 40% of K-rasV12G tumors harbored inactivating mutations in p53 or demonstrated a loss of heterozygosity (LOH; i.e., a loss of the wildtype allele). These data and others (Lin and Lowe, 2001) indicate that crosstalk between oncogenic Ras and p53 pathways contributes to CRC progression. However, lesions that harbor APC mutations before accumulating Ras and p53 mutations appear to carry a worse prognosis, and this is possibly related to the chromosomal instability associated with APC mutations. The existence of cancer stem cells in CRC has recently gained support. However, the origin of mutated stem cells in intestinal neoplasia continues to spark considerable debate among researchers. The “top-down” model formulated by Fearon and Vogelstein proposes that mutant cells originate in the upper crypt and surface epithelia of adenomas (Fearon and Volgelstein, 1990). In these studies, researchers reviewed tissue morphology, Ki-67 staining (a proliferative marker), nuclear b-catenin staining, and single nucleotide polymorphism analyses of DNA from microdissected cells to detect APC mutations and LOH in small (