Pharmacogenomics in Drug Discovery and Development (Methods in Molecular Biology, 2547) 1071625721, 9781071625729

Citation preview

Methods in Molecular Biology 2547

Qing Yan Editor

Pharmacogenomics in Drug Discovery and Development Third Edition

METHODS

IN

MOLECULAR BIOLOGY

Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, UK

For further volumes: http://www.springer.com/series/7651

For over 35 years, biological scientists have come to rely on the research protocols and methodologies in the critically acclaimed Methods in Molecular Biology series. The series was the first to introduce the step-by-step protocols approach that has become the standard in all biomedical protocol publishing. Each protocol is provided in readily-reproducible step-bystep fashion, opening with an introductory overview, a list of the materials and reagents needed to complete the experiment, and followed by a detailed procedure that is supported with a helpful notes section offering tips and tricks of the trade as well as troubleshooting advice. These hallmark features were introduced by series editor Dr. John Walker and constitute the key ingredient in each and every volume of the Methods in Molecular Biology series. Tested and trusted, comprehensive and reliable, all protocols from the series are indexed in PubMed.

Pharmacogenomics in Drug Discovery and Development Third Edition

Edited by

Qing Yan PharmTao, Santa Clara, CA, USA

Editor Qing Yan PharmTao Santa Clara, CA, USA

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-2572-9 ISBN 978-1-0716-2573-6 (eBook) https://doi.org/10.1007/978-1-0716-2573-6 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022, Corrected Publication 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover Illustration Caption: See Chapter 7 for more details. This Humana imprint is published by the registered company Springer Science+Business Media, LLC part of Springer Nature. The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.

Preface Pharmacogenomics is a rapidly developing area that may contribute to personalized medicine. Studies in pharmacogenomics and systems biology may help understand the complexity of the human-drug-environment interactions at various spatiotemporal levels. By changing gear from disease-centered to human-centric medicine, pharmacogenomics brings hope for achieving the optimal therapeutic effects. In this third edition, we approach the challenges in pharmacogenomics and personalized medicine by introducing novel ideas and cutting-edge methodologies. For example, Mendelian randomization (MR) approaches may help identify promising drug targets to overcome the barriers associated with translation (see Chapter 1). MR can be applied to predict disease outcomes and adverse drug events (ADEs). In another example, human leukocyte antigen (HLA) alleles have been related to higher risks of adverse drug reactions (ADRs) in some medications. Many new testing techniques in pharmacogenomics are now available for specific HLA alleles linked to ADRs (see Chapter 2). In cancers, supportive care therapies have contributed to higher tolerability for patients (see Chapter 3). Pharmacogenomics studies may differentiate patient groups with higher risks of toxicities from those who benefit most from the treatments. The most recent development in pharmacogenomics knowledge and clinical practices in supportive therapies is introduced in this volume (see Chapter 3). Furthermore, pharmacogenetic testing among cancer patients may provide the possibility of predicting, preventing, and alleviating chemotherapy-related toxicities (see Chapter 4). It is essential to study the drug-gene pairs in such testing efforts. As a powerful method to predict individual responses, the pharmacogenomics-based examination of chemotherapyinduced peripheral neurotoxicity (CIPN) has been developed to make genetic profiling more applicable (see Chapter 5). The enormous heterogeneity of cancer systems has made it very challenging to overcome drug resistance and ADRs. Recent developments, including the perception of cancer as the complex adaptive system (CAS), may help meet the challenges. The conceptual framework of “Yin-Yang dynamics” may help elucidate the CAS features for identifying effective biomarkers and drug targets for personalized therapies (see Chapter 6). Therapeutic cancer vaccines may be designed to incorporate individual tumor neoantigens (TNAs) for personalized treatment. The steps for the TNAs-carrying RNA vaccines, including individual tumor next generation sequencing (NGS) and drug delivery systems with relevant variations, are described in this edition (see Chapter 7). A pandemic like the one caused by SARS-CoV-2 cannot rely on discovering new drugs that may not provide immediate protection. Repurposing drugs may be a feasible approach for the treatment of COVID-19 infections. Pharmacogenomic analysis may help identify repurposed drug candidates for such instant care (see Chapter 8). The application of pharmacogenomics demands the exploration of the effects of genetic variations on pharmacokinetics and pharmacodynamics. Such understanding may be beneficial for supporting clinical decisions in cardiovascular pharmacotherapy about treatment options and risks of side effects (see Chapter 9).

v

vi

Preface

Moreover, calcium imaging has helped detect the mutations in genes associated with arrhythmia. A high-throughput methodology of calcium imaging for examining individual calcium transients is described in this book (see Chapter 10). Studies of genetic variants and systems biology have indicated that the Yin-Yang dynamics are especially meaningful for cardiovascular pharmacogenomics. The Yin-Yang imbalances in the complex adaptive systems (CAS) are essential for cardiovascular diseases (CVDs), including atherosclerosis, hypertension, and heart failure. The reestablishment of the Yin-Yang dynamical balances would empower the development of personalized medicine for CVDs (see Chapter 11). In addition, the guanosine-50 -O-3-thiotriphosphate (GTPγS) assay is reliable for detecting heterotrimeric G protein activation levels by G-protein coupled receptors (GPCRs) (see Chapter 12). Such measurements are helpful for the studies of heart functions and chronic heart failure. For Alzheimer’s disease (AD), pharmacogenetic studies have revealed that the therapeutic responses are genotype-specific related to the gene clusters under the regulation of epigenetic mechanisms. This book provides a thorough discussion of the current studies of AD pharmacogenomics and the personalized applications of anti-dementia drugs (see Chapter 13). For schizophrenia, the variabilities among individual responses to antipsychotics call for the personalized medicine approach to improve efficacy and avoid adverse effects. The promising pharmacogenomic discoveries are described here (see Chapter 14). For attention-deficit hyperactivity disorder (ADHD), each medication has different pharmacokinetic profiles. The pharmacogenomic information and the pharmacokinetic parameters discussed in the book may help clinicians and patients select the proper medications and doses (see Chapter 15). For the treatment of addictions, responses may rely on genetic, biosocial, and environmental factors. The illumination of the role of pharmacogenomics would support the discovery of new drugs and improve the efficacy of existing therapies, including those for cocaine, alcohol, and opioid dependences (see Chapter 16). Furthermore, studies of opioid medications via pharmacogenomics would benefit pain management in clinical practice to avoid addictions and improve efficacy. By understanding individual variations in pain susceptibility associated with genetic diversity, personalized therapeutic regimens may assist with the reduction of opioid dependency for better regulation of postoperative pain (see Chapters 17 and 18). Disease-modifying anti-rheumatic drugs (DMARDs) for rheumatoid arthritis (RA) have been found with significant variability in both efficacy and toxicity. A thorough discussion of the current status of the pharmacogenomics of both traditional and the newer biologic DMARDS is provided in this volume (see Chapter 19). Recent developments in therapy and the ethical construct of pediatric research have enabled pharmacogenomic evaluations for children. Such advances in pharmacogenomics would support better clinical decision-making for children, especially in childhood cancer treatment (see Chapter 20). Genetic ancestry inference has been applied to stratify patient cohorts to investigate pharmacogenomic variations within and between populations. A detailed guide using genome-wide genetic variant data sets is provided in this book, with a focus on the popular techniques, including principal components analysis (PCA) (see Chapter 21). This edition offers a state-of-the-art and integrative vision of pharmacogenomics by covering new concepts and practical methodologies focusing on disease treatments. The

Preface

vii

broad collection of theoretical and experimental approaches may enable problem-solving by tackling the complexity of personalized drug discovery and development. Written by leading experts in their fields, this edition aims at providing across-the-board resources to support the translation of pharmacogenomics into better individualized health care. I want to thank all authors for their valuable contributions to this exciting area. I also thank the series editor, Dr. John Walker, for his help with the editing. Santa Clara, CA, USA

Qing Yan

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v xi

1 Target Discovery for Drug Development Using Mendelian Randomization . . . . Daniel S. Evans 2 Human Leukocyte Antigen (HLA) Testing in Pharmacogenomics . . . . . . . . . . . . Ann M. Moyer and Manish J. Gandhi 3 Pharmacogenomics in Targeted Therapy and Supportive Care Therapies for Cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zahra Talebi, Alex Sparreboom, and Susan I. Colace 4 Pharmacogenomics in Cytotoxic Chemotherapy of Cancer. . . . . . . . . . . . . . . . . . . Zahra Talebi, Alex Sparreboom, and Susan I. Colace 5 Management of Side Effects in the Personalized Medicine Era: Chemotherapy-Induced Peripheral Neuropathy . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eleonora Pozzi and Paola Alberti 6 The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Qing Yan 7 Design of Personalized Neoantigen RNA Vaccines Against Cancer Based on Next-Generation Sequencing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ˜ a Alburquerque-Gonza´lez, Marı´a Dolores Lo pez-Abella´n, Begon Gine´s Luengo-Gil, Silvia Montoro-Garcı´a, and Pablo Conesa-Zamora 8 COVID-19 Pharmacotherapy: Drug Development, Repurposing of Drugs, and the Role of Pharmacogenomics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Rebecca Bock, Mariana Babayeva, and Zvi G. Loewy 9 Pharmacogenomics Informs Cardiovascular Pharmacotherapy . . . . . . . . . . . . . . . . Mariana Babayeva, Brigitte Azzi, and Zvi G. Loewy 10 Pharmacogenomic Screening of Drug Candidates Using Patient-Specific hiPSC-Derived Cardiomyocyte High-Throughput Calcium Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Malorie Blancard, K. Ashley Fetterman, and Paul W. Burridge 11 The Yin-Yang Dynamics in Cardiovascular Pharmacogenomics and Personalized Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Qing Yan 12 GTPγS Assay for Measuring Agonist-Induced Desensitization of Two Human Polymorphic Alpha2B-Adrenoceptor Variants . . . . . . . . . . . . . . . . . . . . . . . Jordana I. Borges, Alexandra M. Carbone, Natalie Cora, Anastasiya Sizova, and Anastasios Lymperopoulos

1

ix

21

47 63

95

141

165

187 201

241

255

267

x

Contents

13

Pharmacogenomics of Alzheimer’s Disease: Novel Strategies for Drug Utilization and Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ramon Cacabelos, Vinogran Naidoo, Olaia Martı´nez-Iglesias, Lola Corzo, Natalia Cacabelos, Rocı´o Pego, and Juan C. Carril 14 Pharmacogenetics of Antipsychotic Treatment in Schizophrenia . . . . . . . . . . . . . . ¨ ller, and Jennie G. Pouget Samar S. M. Elsheikh, Daniel J. Mu 15 The Pharmacogenetic Impact on the Pharmacokinetics of ADHD Medications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jacob T. Brown 16 Pharmacogenetics of Addiction Therapy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David P. Graham, Mark J. Harding, and David A. Nielsen 17 Pharmacogenomics of Opioid Treatment for Pain Management . . . . . . . . . . . . . . Sarahbeth Howes, Alexandra R. Cloutet, Jaeyeon Kweon, Taylor L. Powell, Daniel Raza, Elyse M. Cornett, and Alan D. Kaye 18 The Role of Pharmacogenomics in Postoperative Pain Management . . . . . . . . . . E. Paylor Sachtleben, Kelsey Rooney, Hannah Haddad, Victoria L. Lassiegne, Megan Boudreaux, Elyse M. Cornett, and Alan D. Kaye 19 Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis. . . . . . . . . . . . . . . . Atinuke Aluko and Prabha Ranganathan 20 Pharmacogenomics in Children. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael J. Rieder and Abdelbaset A. Elzagallaai 21 Genetic Ancestry Inference for Pharmacogenomics . . . . . . . . . . . . . . . . . . . . . . . . . I. King Jordan, Shivam Sharma, Shashwat Deepali Nagar, ˜ o-Ramı´rez Augusto Valderrama-Aguirre, and Leonardo Marin Correction to: Genetic Ancestry Inference for Pharmacogenomics . . . . . . . . . . . . . . . . I. King Jordan, Shivam Sharma, Shashwat Deepali Nagar, ˜ o-Ramı´rez Augusto Valderrama-Aguirre, and Leonardo Marin Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

275

389

427 437 491

505

527 569 595

C1

611

Contributors PAOLA ALBERTI • School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy; NeuroMI (Milan Center for Neuroscience), Milan, Italy BEGON˜A ALBURQUERQUE-GONZA´LEZ • Pathology and Histology Department Facultad de Ciencias de la Salud, UCAM Universidad Catolica San Antonio de Murcia, Murcia, Spain ATINUKE ALUKO • Division of Rheumatology, Department of Medicine, Washington University School of Medicine, St. Louis, MO, USA BRIGITTE AZZI • Touro College of Pharmacy, New York, NY, USA MARIANA BABAYEVA • Touro College of Pharmacy, New York, NY, USA MALORIE BLANCARD • Department of Pharmacology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA; Center for Pharmacogenomics, Northwestern University Feinberg School of Medicine, Chicago, IL, USA REBECCA BOCK • Stern College for Women, Yeshiva University, New York, NY, USA JORDANA I. BORGES • Laboratory for the Study of Neurohormonal Control of the Circulation, Department of Pharmaceutical Sciences, Nova Southeastern University, Fort Lauderdale, FL, USA MEGAN BOUDREAUX • School of Medicine, Louisiana State University Shreveport, Shreveport, LA, USA JACOB T. BROWN • Pharmacy Practice and Pharmaceutical Sciences, College of Pharmacy, University of Minnesota, Duluth, MN, USA PAUL W. BURRIDGE • Department of Pharmacology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA; Center for Pharmacogenomics, Northwestern University Feinberg School of Medicine, Chicago, IL, USA NATALIA CACABELOS • Department of Medical Documentation, International Center of Neuroscience and Genomic Medicine, EuroEspes Biomedical Research Center, Spain RAMO´N CACABELOS • Department of Genomic Medicine, International Center of Neuroscience and Genomic Medicine, EuroEspes Biomedical Research Center, Spain ALEXANDRA M. CARBONE • Laboratory for the Study of Neurohormonal Control of the Circulation, Department of Pharmaceutical Sciences, Nova Southeastern University, Fort Lauderdale, FL, USA JUAN C. CARRIL • Department of Genomics and Pharmacogenomics, International Center of Neuroscience and Genomic Medicine, EuroEspes Biomedical Research Center, Corunna, Spain ALEXANDRA R. CLOUTET • LSU Health Shreveport, Shreveport, LA, USA SUSAN I. COLACE • Division of Hematology, Oncology, and Blood & Marrow Transplant, Nationwide Children’s Hospital, Columbus, OH, USA; The Ohio State University, Columbus, OH, USA PABLO CONESA-ZAMORA • Pathology and Histology Department Facultad de Ciencias de la Salud, UCAM Universidad Catolica San Antonio de Murcia, Murcia, Spain; Laboratory Medicine Department, Group of Molecular Pathology and Pharmacogenetics, Biomedical Research Institute from Murcia (IMIB), Hospital Universitario Santa Lucı´a, Cartagena, Spain

xi

xii

Contributors

NATALIE CORA • Laboratory for the Study of Neurohormonal Control of the Circulation, Department of Pharmaceutical Sciences, Nova Southeastern University, Fort Lauderdale, FL, USA ELYSE M. CORNETT • Department of Anesthesiology, LSU Health Shreveport, Shreveport, LA, USA LOLA CORZO • Department of Medical Biochemistry, International Center of Neuroscience and Genomic Medicine, EuroEspes Biomedical Research Center, Spain SAMAR S. M. ELSHEIKH • The Centre for Addiction and Mental Health (CAMH), Toronto, ON, Canada ABDELBASET A. ELZAGALLAAI • Department of Physiology and Pharmacology, Schulich School of Medicine & Dentistry, Western University, London, ON, Canada DANIEL S. EVANS • California Pacific Medical Center Research Institute, San Francisco, CA, USA; Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, CA, USA K. ASHLEY FETTERMAN • Department of Pharmacology, Northwestern University Feinberg School of Medicine, Chicago, IL, USA; Center for Pharmacogenomics, Northwestern University Feinberg School of Medicine, Chicago, IL, USA MANISH J. GANDHI • Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, USA DAVID P. GRAHAM • Michael E. DeBakey Veterans Affairs Medical Center, and the Menninger Department of Psychiatry and Behavioral Sciences, Baylor College of Medicine, Houston, TX, USA HANNAH HADDAD • Kansas City University of Medicine and Biosciences, Kansas City, MO, USA MARK J. HARDING • Michael E. DeBakey Veterans Affairs Medical Center, and the Menninger Department of Psychiatry and Behavioral Sciences, Baylor College of Medicine, Houston, TX, USA SARAHBETH HOWES • LSU Health Shreveport, Shreveport, LA, USA I. KING JORDAN • School of Biological Sciences, Georgia Institute of Technology, Atlanta, GA, USA ALAN D. KAYE • Departments of Anesthesiology and Pharmacology, Toxicology and Neurosciences, LSU Health Shreveport, Shreveport, LA, USA JAEYEON KWEON • LSUHSC New Orleans, New Orleans, LA, USA VICTORIA L. LASSIEGNE • LSUHSC School of Medicine New Orleans, New Orleans, LA, USA ZVI G. LOEWY • Touro College of Pharmacy, New York, NY, USA; School of Medicine, New York Medical College, Valhalla, NY, USA MARI´A DOLORES LO´PEZ-ABELLA´N • Laboratory Medicine Department, Group of Molecular Pathology and Pharmacogenetics, Biomedical Research Institute from Murcia (IMIB), Hospital Universitario Santa Lucı´a, Cartagena, Spain GINE´S LUENGO-GIL • Laboratory Medicine Department, Group of Molecular Pathology and Pharmacogenetics, Biomedical Research Institute from Murcia (IMIB), Hospital Universitario Santa Lucı´a, Cartagena, Spain ANASTASIOS LYMPEROPOULOS • Laboratory for the Study of Neurohormonal Control of the Circulation, Department of Pharmaceutical Sciences, Nova Southeastern University, Fort Lauderdale, FL, USA LEONARDO MARIN˜O-RAMI´REZ • National Institute on Minority Health and Health Disparities, National Institutes of Health, Bethesda, MD, USA

Contributors

xiii

OLAIA MARTI´NEZ-IGLESIAS • Department of Medical Epigenetics, International Center of Neuroscience and Genomic Medicine, EuroEspes Biomedical Research Center, Spain SILVIA MONTORO-GARCI´A • Cell Culture Lab, Facultad de Ciencias de la Salud, UCAM Universidad Catolica San Antonio de Murcia, Murcia, Spain ANN M. MOYER • Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, USA DANIEL J. MU¨LLER • The Centre for Addiction and Mental Health (CAMH), Toronto, ON, Canada; Department of Psychiatry, University of Toronto, Toronto, ON, Canada SHASHWAT DEEPALI NAGAR • School of Biological Sciences, Georgia Institute of Technology, Atlanta, GA, USA VINOGRAN NAIDOO • Department of Neuroscience, International Center of Neuroscience and Genomic Medicine, EuroEspes Biomedical Research Center, Corunna, Spain DAVID A. NIELSEN • Michael E. DeBakey Veterans Affairs Medical Center, and the Menninger Department of Psychiatry and Behavioral Sciences, Baylor College of Medicine, Houston, TX, USA ROCI´O PEGO • Department of Neuropsychology, International Center of Neuroscience and Genomic Medicine, EuroEspes Biomedical Research Center, Corunna, Spain JENNIE G. POUGET • The Centre for Addiction and Mental Health (CAMH), Toronto, ON, Canada; Department of Psychiatry, University of Toronto, Toronto, ON, Canada TAYLOR L. POWELL • LSU Health Shreveport, Shreveport, LA, USA ELEONORA POZZI • School of Medicine and Surgery, University of Milano-Bicocca, Monza, Italy; NeuroMI (Milan Center for Neuroscience), Milan, Italy PRABHA RANGANATHAN • Division of Rheumatology, Department of Medicine, Washington University School of Medicine, St. Louis, MO, USA DANIEL RAZA • Tulane School of Medicine, New Orleans, LA, USA MICHAEL J. RIEDER • Division of Paediatric Clinical Pharmacology, Department of Paediatrics, Schulich School of Medicine & Dentistry, Western University, London, ON, Canada KELSEY ROONEY • LSU Health Shreveport, Shreveport, LA, USA E. PAYLOR SACHTLEBEN • LSU Health Shreveport, Shreveport, LA, USA SHIVAM SHARMA • National Institute on Minority Health and Health Disparities, National Institutes of Health, Bethesda, MD, USA ANASTASIYA SIZOVA • Laboratory for the Study of Neurohormonal Control of the Circulation, Department of Pharmaceutical Sciences, Nova Southeastern University, Fort Lauderdale, FL, USA ALEX SPARREBOOM • Division of Pharmaceutics and Pharmacology, College of Pharmacy & Comprehensive Cancer Center, The Ohio State University, Columbus, OH, USA ZAHRA TALEBI • Division of Pharmaceutics and Pharmacology, College of Pharmacy, The Ohio State University, Columbus, OH, USA AUGUSTO VALDERRAMA-AGUIRRE • Department of Biological Sciences, Faculty of Sciences, Universidad de Los Andes, Bogota´, DC, Colombia QING YAN • PharmTao, Santa Clara, CA, USA

Chapter 1 Target Discovery for Drug Development Using Mendelian Randomization Daniel S. Evans Abstract Making drug development more efficient by identifying promising drug targets can contribute to resource savings. Identifying promising drug targets using human genetic approaches can remove barriers related to translation. In addition, genetic information can be used to identify potentially causal relationships between a drug target and disease. Mendelian randomization (MR) is a class of approaches used to identify causal associations between pairs of genetically predicted traits using data from human genetic studies. MR can be used to prioritize candidate drug targets by predicting disease outcomes and adverse events that could result from the manipulation of a drug target. The theory behind MR is reviewed, including a discussion of MR assumptions, different MR analytical methods, tests for violations of assumptions, and MR methods that can be robust to some violations of MR assumptions. A protocol to perform two-sample MR (2SMR) with summary genome-wide association study (GWAS) results is described. An example of 2SMR examining the causal relationship between low-density lipoprotein (LDL) and coronary artery disease (CAD) is provided as an illustration of the protocol. Key words Mendelian randomization, Target discovery, Genetics, GWAS, Instrumental variables

1

Introduction The process of drug development is lengthy, costly, and subject to failure. It has been estimated that the clinical approval success rate is between 13% and 32%, depending on the therapeutic class and drug molecular type [1]. A subsequent study based on data from 835 drug developers between the years 2003 and 2011 found that the probability of FDA approval of the lead indication for drugs in phase one development was 15.3% [2]. The resources spent on drugs that fail to gain approval represent enormous losses that contribute to high costs and missed opportunities in drug development. Many factors contribute to the low success rate in drug development, but one key factor is the difficulty in identifying promising drug targets. A key feature of an ideal drug target is a

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_1, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

1

2

Daniel S. Evans

dose-dependent change in disease risk when the target’s activity is manipulated through a particular range, the so-called therapeutic window [3]. Dose-dependent relationships typically require a causal association between the drug target and the disease. In other words, a good drug target should be a component of a process that is required for disease development, not simply an effect of the disease itself. In addition, side effects and adverse events must also be taken into account. Within the therapeutic window of target activity, there must also not be adverse events or undesirable side effects that could prevent a drug from reaching FDA approval [3]. Unfortunately, many commonly used approaches to identify drug targets are subject to a number of limitations that result in less than satisfactory drug targets. Take, for example, preclinical research using non-human laboratory models. There are many challenges facing translation of non-human findings to humans [4]. Intervention treatment effects can often be discordant between humans and non-human model organisms. In a comparison of six interventions with well-characterized treatment effects in humans, only three interventions had similar outcomes in non-human model systems [5]. There could be many potential factors that contribute to the failure of non-human models to accurately reflect treatment effects in humans. One broad factor could be that non-human models of disease do not always capture human physiology and pathology relevant to the disease process [4]. A model system for a disease that is missing a physiological system known to be important in disease etiology would not be useful in the identification of targets participating in that aspect of disease development. A second broad factor could be biased treatment effect estimates from small non-human studies [4]. Studies in humans can provide useful evidence on the suitability of a target for drug development without the need for translation from non-human experimental laboratory models. One common target identification approach in human clinical research is to investigate whether biomarkers are associated with disease. This approach involves the measurement of biomarkers, such as proteins or lipids, in blood stored from participants in epidemiological studies of various designs, e.g., case-control or longitudinal cohort, followed by testing for associations with disease. Biomarkers identified in this manner can lead to candidate drug targets; however, determining which biomarkers are causally associated with disease can be challenging. Biomarker studies can be subject to multiple biases that impact effect estimates and call into question the causality of identified associations [6]. Confounding bias can occur when there exists a common cause between an exposure and an outcome of interest [7, 8]. Selection bias can occur when study participants in a study sample differ from the target population of interest. If characteristics that influence a biomarker’s association with an

Target Discovery Using Mendelian Randomization

3

outcome differ between the study sample and the target population, selection bias can result in a biased effect estimate [7, 9]. Reverse causality can induce a biomarker association with disease if the disease itself leads to an alteration in the biomarker [7, 10, 11]. Clearly, biomarkers identified in associations influenced by confounding factors or reverse causality are unlikely to be causally related to a disease outcome under study and thus are unlikely to affect disease risk when manipulated. These challenges all stem from the fact that biomarkers measured in research participants living in the community can change in response to various effects within a person. The use of genetic predictors of biomarkers can avoid some of the confounding influences of traditional biomarker studies by virtue of the fact that germline genetic variation is established at conception, with the notable exception of somatic mutations, and thus provides a causal anchor to associations [12]. Human genetic studies have been increasingly used to provide causal evidence supporting a drug target’s relationship with disease [13–17]. Not only can bias largely be avoided, but human genetics can also provide information useful to drug development and target identification. It has been shown that drugs with genetic support have approximately a two times higher chance of being approved [18]. An analytical framework that leverages human genetics is likely to provide substantial benefits to drug development.

2

Mendelian Randomization An analytical framework that takes advantage of genetics to identify causal associations falls under the term Mendelian randomization (MR) [13, 19]. MR is an application of instrumental variable (IV) analysis in which genetic variants are used as IVs [13]. IVs are factors that can operate like a randomization process to assign individuals into different groups of exposure levels that do not differ with respect to confounding factors [20]. An IV is valid so long as the IV is correlated with the exposure of interest, is not associated with the outcome of interest conditional on the exposure, and is not associated with confounders of the exposureoutcome relationship [20]. IVs do not necessarily need to be genetic variants. In fact, as originally conceived in economics research, IVs were not constructed using genetic variants. One application of IV analysis without genetics was in the investigation of whether starting school earlier in life affects educational attainment [21]. Family background was known to have a confounding influence on the relationship between age at school entry and educational attainment. Unfortunately, standard research tools fail to fully capture the influence of family background, so unmeasured residual

4

Daniel S. Evans

confounding would likely remain in ordinary least squares regression models. For this challenging research topic, IV analysis can be used to obtain unbiased estimates in the presence of strong confounding influences that are difficult to fully adjust for. Date of birth was used as an IV for age at school entry. Because the school start date is fixed and laws mandate children start school by a certain age, birth date randomly assigned children within a grade level into groups that started schooling earlier or later in life. Confounding influences related to family background were considered to be uncorrelated with birth date, so these influences were effectively randomized by the IV constructed with birth date. Using this IV, it was found that school starting age had a very modest effect on educational attainment beyond the effects of compulsory schooling up to a certain age [21]. MR is a specific application of IV analysis in which genetic variants are used as IVs. An individual’s genetic variants are determined at conception and, as such, are largely free from the effects of confounders. Very few potential confounders can alter a person’s DNA in all of the tissues relevant to the disease under study. Mendelian randomization refers to the fact that the inheritance of a trait-associated genetic allele at conception is analogous to the random assignment of treatment in a randomized controlled trial. Genetic predictors of traits can be used as instrumental variables to approximate a lifetime exposure to a risk factor of interest, e.g., increased LDL levels or increased expression of a particular gene, and these genetic instrumental variables (gIVs) can then be used to estimate the unconfounded association between the risk factor and the outcome. The comparison of individuals carrying gIVs for high trait levels and those with gIVs for low trait levels can be considered a natural experiment of a lifetime exposure to different trait levels [13]. Mendelian randomization can also be used to establish a doseresponse relationship between a risk factor and an outcome. Take, for example, a schematic diagram describing the relationship between LDL cholesterol levels and risk of heart disease (Fig. 1). Rare gain-of-function PCSK9 mutations were associated with high LDL levels and an increased risk of coronary heart disease [22]. Loss-of-function PCSK9 mutations were associated with low LDL levels and a lower risk of coronary heart disease [23– 25]. With only a single gene, a collection of variants that capture low and high activity levels of a gene can be used to establish a relationship between a gene product (PCSK9), an intermediate trait (LDL), and a clinical outcome (cardiovascular disease). By identifying causal associations between gene product activity, an intermediate trait, and a clinical outcome, MR can greatly aid in drug target identification.

High

5

PCSK9 gain -of-function alleles

Low

C a r d io v a s c u la r d is e a s e r i s k

Target Discovery Using Mendelian Randomization

PCSK9 loss -of-function alleles Low

High

LDL levels

Fig. 1 Schematic diagram of MR approach relating genetically predicted LDL and cardiovascular disease risk 2.1 Mendelian Randomization Terms

To define the terms used in this protocol, an example linear model will be referenced (Eq. 1). The dependent variable is y, the independent variable is x, the association effect parameter is β, and the residual error is ε. The dependent variable can sometimes also be referred to as the outcome. The independent variable can sometimes also be referred to as the predictor, explanatory variable, or exposure. y ¼βxþε

ð1Þ

In this protocol, the dependent variable will be referred to as the outcome, and the independent variable will be referred to as the exposure. Even though MR exposures might not typically be considered exposures in the sense of originating outside of the human body, this protocol will consistently use the term exposure for the independent variable. 2.2 Mendelian Randomization Assumptions and Limitations

The assumptions underlying IV analysis, as described above, also apply to MR. The IV assumptions have been tailored to MR, which makes use of genetic instrumental variables (gIVs). MR assumptions are grouped into three categories, some of which directly relate to this method’s limitations [26, 27]. The first MR assumption is that the genetic instrumental variable (gIV) is associated with the exposure. If this assumption is not satisfied, the gIV would not be a good proxy for the exposure. Using gIVs associated with the exposure at the genome-wide significance threshold (P  5  108) reduces the chance that gIV-exposure associations represent false positives. It should be

6

Daniel S. Evans

noted, however, that the magnitude of the gIV-exposure association effect estimate from GWAS is typically small, meaning that gIVs typically exert only a small influence on the exposure. Violations of this assumption can result in so-called weak instrument bias. When the gIV-exposure and gIV-outcome associations are estimated in the same study population (one-sample MR), weak instruments can bias MR estimates toward the confounded observational exposure-outcome association, which could lead one to incorrectly identify non-causal associations [28, 29]. However, when gIV associations with the exposure and the outcome are identified in different study populations (two-sample MR), weak instruments can bias the MR estimate toward the null, which leads to conservative conclusions rather than incorrectly identified non-causal associations [29, 30]. The weaker the gIV association with exposure is, the lower the statistical power of the MR test. Thus, non-significant MR results should be interpreted with caution and accompanied by a power calculation that takes into account the strength of the gIVs [31]. The second MR assumption is that the gIV is independent of the outcome, conditional on the exposure, and confounders of the exposure-outcome association. This assumption is also referred to as “the exclusion restriction.” Violations of the second assumption limit the ability to interpret MR results as providing a causal association between the exposure and outcome. If a direct path exists between a gIV and the outcome, one could not conclude that the gIV association with the outcome was mediated by the exposure. Violations of the second MR assumption can be induced by horizontal pleiotropy, linkage disequilibrium (LD) between gIVs, or population stratification [27]. The third MR assumption is that the gIV is not associated with confounders of the exposure and outcome association. In addition to violations of the second assumption, horizontal pleiotropy, LD, and population stratification can also induce violations of the third assumption [27]. Canalization is a limitation of MR that is not directly related to one of the three assumptions. Canalization refers to compensatory processes triggered by gIVs, perhaps during development, to counter the effect of gIVs. It would be expected that countering a gIV’s effects would weaken the gIV and thus bias the MR estimate toward the null, similar to weak instruments bias [26–28]. However, if a gIV-induced developmental compensatory process induces its own exposure separately from the exposure thought to be instrumented, the canalization-induced exposure could potentially have an effect on the outcome independently of the instrumented exposure, an example of pleiotropy and a violation of the second MR assumption.

Target Discovery Using Mendelian Randomization

7

2.3 Mendelian Randomization Exposures

One feature of MR that makes this method especially powerful for drug target identification is that any exposure can be used, as long as the exposure can be predicted by gIVs that fulfill all three MR assumptions. Exposures can include some of the thousands of traits that have been subjected to GWAS, e.g., LDL, weight, or glucose level [32]. Using results from the GTEx project, which identified genetic variants associated with gene expression, gIVs can be used to predict tissue-specific gene expression [33]. Gene expression is particularly useful in an MR context for drug discovery, as gene products can be candidate drug targets. Table 1 lists multiple repositories of GWAS results that can be mined for gIVs. To make this process more convenient, MR-Base has curated gIVs for using GWAS results for traits, gene expression, and also circulating metabolites, circulating proteins, and DNA methylation probes [34]. With large collections of gIVs for traits or gene expression, MR studies employing a one outcome to many exposures design can be performed to screen for potential drug targets for a single disease. Since gIVs have already been identified and cataloged in publicly available repositories, an MR screen can be performed with either publicly available GWAS results or genome-wide single nucleotide polymorphism (SNP) genotypes from a human study. With a collection of thousands of exposures (traits, gene expression, protein levels) that can be genetically predicted, an MR screen can be a costeffective way to leverage genome-wide genetic data to gain insight into causal associations with a disease or condition without the need to directly measure thousands of variables and biomarkers in clinical studies. MR using gIVs for gene expression has been used to identify gene associations with diseases and conditions, including cardiovascular disease, neurological diseases, psychiatric diseases, and human longevity [35–37].

2.4 Mendelian Randomization Outcomes

Much in the same way that catalogs of genetically predicted exposures have been assembled, many of these exposures can also be considered outcomes, depending on the research question. MR can be used to examine the potential causal association between one exposure and one outcome. However, it is also possible to examine many genetically predicted outcomes one at a time in an MR-based phenome-wide association study, or MR-PheWAS [38]. MR-PheWAS was used to identify potentially causal associations between body mass index (BMI) and many outcomes [38]. In addition, an MR-PheWAS was used to screen for traits associated with genetically predicted levels of BMI, C-reactive protein (CRP), and low-density lipoprotein (LDL) [39]. The concept of testing multiple outcomes in an MR framework has recently been expanded to also include multiple exposures with multiple outcomes. MR was used to test the association between the levels of 1002 genetically predicted proteins and 225 genetically predicted outcomes [40]. Similarly, associations between

8

Daniel S. Evans

Table 1 Sources of genetic associations with traits Consortium

Full consortium name

Web link

ALSKP

ALS Knowledge portal

http://alskp.org/informational/ data

http://www.cardiogramplusc4d. CARDIoGRAMplusC4D Coronary ARtery DIsease Genome org/data-downloads/ wide Replication and Meta-analysis (CARDIoGRAM) plus The Coronary Artery Disease (C4D) Genetics CDKP/ISGC

Cerebrovascular Disease Knowledge portal/International Stroke Genetics Consortium

https://cd.hugeamp.org/ downloads.html

CHARGE

Cohorts for Heart and Aging Research in Genetic Epidemiology

http://www.chargeconsortium. com/main/results

CKDGen

Chronic Kidney Disease Genetics Consortium

http://ckdgen.imbi.uni-freiburg.de

CMDKP

Common Metabolic Diseases Knowledge portal

https://hugeamp.org/downloads. html

CVDKP

Cardiovascular Disease Knowledge portal

https://cvd.hugeamp.org/ downloads.html

deCODE

deCODE genetics

https://www.decode.com/ summarydata/

Diagram

DIAbetes Genetics Replication And Meta-analysis

http://diagram-consortium.org/ downloads.html

EGG

Early Growth Genetics Consortium

http://egg-consortium.org/

eQTLGen

eQTL consortium

https://www.eqtlgen.org/

GEFOS

GEnetic Factors for OSteoporosis Consortium

http://www.gefos.org

GTEx portal

The Genotype-Tissue Expression project portal

https://gtexportal.org/home/

GIANT

Genetic Investigation of ANthropometric Traits

http://portals.broadinstitute.org/ collaboration/giant/index.php/ GIANT_consortium_data_files

GLGC

Global Lipids Genetics Consortium

http://csg.sph.umich.edu// abecasis/public/lipids2013/

GRASP

Genome-Wide Repository of Associations Between SNPs and Phenotypes

https://grasp.nhlbi.nih.gov/ FullResults.aspx

GWAS catalog

The NHGRI-EBI Catalog of human genome-wide association studies

https://www.ebi.ac.uk/gwas/ (continued)

Target Discovery Using Mendelian Randomization

9

Table 1 (continued) Consortium

Full consortium name

Web link

IBDGenetics

International Inflammatory Bowel Disease Genetics Consortium

https://www.ibdgenetics.org/ downloads.html

JENGER

Japanese ENcyclopedia of GEnetic associations by Riken

http://jenger.riken.jp/en/

MAGIC

Meta-Analyses of Glucose and Insulin-related traits Consortium

https://www.magicinvestigators. org/downloads/

MSKKP

Musculoskeletal Knowledge portal

https://msk.hugeamp.org/ downloads.html

NIAGADS

National Institute on Aging Genetics https://www.niagads.org/ genomics/showXmlDataContent. of Alzheimer’s Disease do?name¼XmlQuestions. Documentation#about

PGC

Psychiatric Genomic Consortium

https://www.med.unc.edu/pgc/ results-and-downloads

RGC

Reproductive Genetics Consortium

http://www.reprogen.org/data_ download.html

Sleep Disorder KP

Sleep Disorder Knowledge portal

https://sleep.hugeamp.org/ downloads.html

T2DKP

Type II Diabetes Knowledge portal

https://t2d.hugeamp.org/ downloads.html

UKBB

UK BioBank

http://geneatlas.roslin.ed.ac.uk

UKBB

UK BioBank

http://www.nealelab.is/uk-biobank

WTCC

Wellcome Trust Case Control Consortium (access by request)

https://www.wtccc.org.uk/ccc1/ summary_stats.html

AncestryDNA via EGA

AncestryDNA COVID-19 GWAS with Eight Phenotypes

https://ega-archive.org/studies/ EGAS00001005099

genetically predicted gene expression across 48 tissue types and 395 outcomes were estimated using MR [41]. These rich resources provide users with the opportunity to look up an exposure or outcome of interest from the catalog of precomputed results. 2.5 Mendelian Randomization for Drug Repurposing and Adverse Event Screening

An MR-PheWAS using one exposure that is a known target of an approved drug can identify causal associations with outcomes other than the lead indication. These associations could represent drug repurposing opportunities or potential adverse events [42, 43]. Reviewed below are two drug repurposing opportunities and one adverse event discovery using MR. In a drug repurposing strategy to identify additional indications for an approved drug,

10

Daniel S. Evans

MR was performed using a gIV for the target of tocilizumab, a monoclonal antibody approved for treatment of rheumatoid arthritis. Tocilizumab inhibits interleukin-6 receptor (IL6R), and the MR study used a non-synonymous variant (rs2228145, Asp358Ala) in IL6R as a gIV to represent lower IL6R activity. Results from their study suggested that inhibition of IL6R could result in lower risk of coronary heart disease, providing the rationale for clinical testing of tocilizumab for coronary heart disease indications [44]. The same gIV for IL6R was found to be associated with type-1 diabetes, and molecular mechanisms of this IL6R non-synonymous variant were identified to provide additional clues for engineering therapeutics to target IL6R activity [45]. With a similar strategy, an MR study was performed using gIVs for interleukin-18 (IL18). The results indicated that lower genetically predicted levels of IL18 were associated with lower risk of inflammatory bowel disease (IBD) [46]. The authors noted that other groups reported that a monoclonal antibody targeting IL18 that decreased free IL18 levels was found to be safe, yet ineffective for type-2 diabetes in a phase II trial. Thus, antibody therapies targeting IL18 that were deemed failures in drug development could potentially be repurposed with a lead indication for IBD if subsequent clinical trials support the findings. In an effort to screen for potential adverse events that could result from inhibitors of PCSK9, a PheWAS was performed using a genetic score for PCSK9. The PheWAS recapitulated the therapeutic effects of PCSK9 inhibitors but also identified an increased risk of type2 diabetes [47]. 2.6 Mendelian Randomization Analysis Methods

Many MR methods have been developed to address a variety of research questions. MR study designs can include estimation of gene-exposure and gene-outcome associations in the same clinical study (one-sample MR) or in separate studies (two-sample MR). One-sample MR requires individual-level genotype data, but two-sample MR can make use of individual-level genotype data or summary-level GWAS results [26–28]. This protocol will focus on two-sample MR (2SMR) using summary-level GWAS results because data and software for this method are publicly available, and the data relies on GWAS meta-analysis results, which are the largest and most robust set of genetic associations with exposures and outcomes (see Note 1). The reader is encouraged to read any of the review articles cited above for more details on other MR methods. Two-sample MR with summary GWAS results can be most easily understood with an example using a single gIV. With a single gIV, the gene-outcome association is scaled to reflect a genetically predicted one-unit change in the exposure. This scaled estimate is the Wald ratio. As an example for a single SNP gIV, if each gIV allele is associated with a 0.5 SD change in the exposure, and the same

Target Discovery Using Mendelian Randomization

11

gIV allele is associated with a 0.1 SD change in the outcome, then the Wald ratio estimate is 0.2 change in the outcome per one SD change in the exposure (0.1/0.5 ¼ 0.2) [27, 34]. The standard error (SE) of the Wald ratio estimate can be estimated in two ways: either the Wald ratio SE can be estimated as the SE of the gIV-outcome association divided by the gIV-exposure effect estimate or it can be estimated using the delta method to account for uncertainty in the gIV-outcome and gIV-exposure associations [27, 34, 48]. Modern GWAS meta-analysis typically identifies more than one genome-wide significant SNP association with a trait that could be considered an exposure in MR [32]. Two-sample MR can accommodate this additional information by combining Wald ratios from single SNP gIVs into a combined estimate [49]. A simple way to combine Wald ratios is to perform inverse variance weighted (IVW) fixed-effects meta-analysis [34]. IVW and other MR methods designed for use in 2SMR with summary GWAS results are described in the methods section of the protocol.

3

Methods This protocol will demonstrate the two-sample MR (2SMR) method using GWAS summary statistics, with an example based on LDL as an exposure [50] and coronary heart disease as an outcome [51] (see Note 2). Individual steps will be based on the TwoSampleMR R package and the extensively curated set of gIVs from MR-Base [34]. A web interface for this package exists, but the R package makes it possible to program the analysis to enable more customization and reproducible analysis. Furthermore, using the R package enables users to make use of outcomes and exposures that might not be available on the website. All graphs shown in the methods section were created using R v.4.0.5 on the x86_64-appledarwin17.0 64-bit platform and the TwoSampleMR package v.0.5.6 loaded.

3.1 Obtain Genetic Instruments

Genetic IVs can be obtained from the GWAS catalog or sources listed in Table 1. However, the most convenient approach is to obtain gIVs from MR-Base [34]. At a minimum, gIVs must include the SNP identifier (rs number), effect size, SE, and effect allele, and MR-Base has curated this information. To satisfy the first MR assumption, only genetic variants associated with the exposure at the genome-wide significance threshold (P-value 5  108) should be chosen as gIVs. For this first step, genome-wide significant SNP associations for the exposure are needed, but later steps will match up these SNPs in the outcome GWAS results. Linkage disequilibrium (LD) is correlation between SNPs, and if LD is not taken into account, gIVs from genomic regions with high LD will be more heavily weighted in the MR causal estimate.

12

Daniel S. Evans

Furthermore, dependence among Wald ratio estimates violates MR assumptions and assumptions of fixed effect meta-analysis. To ensure each gIV for the exposure is independent, LD clumping is performed to remove SNPs in LD with the most significant index SNP in an association region [52]. All gIVs curated in MR-Base have been LD clumped. The curated gIVs from MR-Base were previously stored in the R package MRInstruments. Now, all pre-LD-clumped gIVs are stored in a database, and the TwoSampleMR R package retrieves data from a database on a server. 3.2 Extract Exposure gIVs from Outcome GWAS and Harmonize

SNPs selected as gIVs for the exposure now must be identified in the outcome GWAS, regardless of whether the SNPs were genomewide significant in the outcome GWAS. This is why the full genome-wide GWAS results are needed, not just the SNPs that were genome-wide significant. SNP matching is first performed using rs numbers, which are accession numbers for genetic variants stored in the NCBI database dbSNP. Refinement and improvements to genome mapping result in new builds of the dbSNP database. Remapping between dbSNP builds can result in some rs numbers merging or being withdrawn [53]. When matching rs numbers from different GWAS results, it is critical that they come from the same dbSNP build. MR-Base has harmonized the GWAS results in their collection so that all rs numbers are mapped to dbSNP build 144. If a user is interested in harmonizing rs numbers outside of MR-Base, the Ensembl Variant Effect Predictor can lift rs numbers and identify merged and withdrawn rs numbers [54]. If a gIV for the exposure is not directly matched in the outcome GWAS, MR-Base will seek to identify SNPs that are LD proxies, i.e., highly correlated with the target SNP, using LD information from 1000 genomes genotypes. Once gIVs between exposure and outcome GWAS are matched, MR-Base attempts to harmonize the SNPs. Harmonization involves making sure that for each SNP, the effect on the exposure and outcome correspond to the same effect allele. This involves checking that the same effect alleles are listed for a SNP, that the alleles are not reported on opposite DNA strands, and that the frequency of palindromic SNP alleles, e.g., A/T and G/C, match [34].

3.3 Perform MR Analysis

Once gIVs for the exposure and outcome are matched and harmonized, MR analysis can be performed. The traditional MR analysis is an IVW meta-analysis of Wald ratios of SNP outcome/SNP exposure [34]. The IVW estimate is equivalent to a weighted linear regression of SNP-outcome associations on SNP-exposure associations with the intercept constrained to zero [28]. To visualize the IVW estimate, the weighted linear regression is plotted with the

Target Discovery Using Mendelian Randomization

13

Fig. 2 MR IVW test for LDL cholesterol and coronary heart disease (CHD) risk. Yaxis shows the SNP-CHD log OR, and the x-axis shows the SNP-LDL effect size in SD units. Points are gIV Wald ratio estimates with SE bars

individual Wald ratios, as demonstrated for coronary heart disease (CHD) and LDL cholesterol (Fig. 2). The IVW estimate shows a direct relationship between LDL and CHD (Fig. 2). MR-Base provides alternative MR methods that can be more resistant to violations of MR assumptions. In addition to IVW, the following methods can also be applied: MR Egger, simple mode, weighted mode, and weighted median. The TwoSampleMR R package also provides functions to perform additional MR methods. MR analysis of LDL and CHD using all five methods is shown in Fig. 3. Unlike IVW, MR-Egger does not constrain the regression Y-intercept to zero, which takes pleiotropy and violations of the second MR assumption into account [55]. The Y-intercept represents the SNP-outcome effect when the SNP-exposure effect is zero. If there are no direct effects due to pleiotropy, it is expected that the SNP-outcome effect would be zero when SNP exposure is zero. From visual inspection of Fig. 3, there does not appear to be evidence for pleiotropy. The simple (unweighted) and weighted mode methods provide a consistent estimate of the causal effect if the most common pleiotropy value across gIVs is zero, which is termed the zero modal pleiotropy assumption (ZEMPA) [56]. The weighted median method is defined as the median of a weighted

14

Daniel S. Evans

Fig. 3 MR tests for LDL cholesterol and coronary heart disease (CHD) risk. Y-axis shows the SNP-CHD log OR, and the x-axis shows the SNP-LDL effect size in SD units. Points are gIV Wald ratio estimates with SE bars. Different MR tests indicated with different colored lines

empirical density function of the Wald ratio estimates. This method can provide a consistent causal estimate if at least 50% of the gIVs are valid and do not violate MR assumptions [57]. With multiple MR methods, the user should evaluate whether the different methods provide consistent results, which would indicate that the MR estimate is valid under varying degrees of violations of MR assumptions. For the relationship between LDL and CHD, all methods consistently show a significant positive association between these two factors, albeit with varying magnitudes (Fig. 3). 3.4 Diagnostics and Sensitivity Analysis

Heterogeneity among gIVs can indicate violations of MR assumptions [58]. Heterogeneity can be visually assessed using a forest plot of the Wald ratio estimate for each gIV (Fig. 4). Statistical tests for heterogeneity (Cochrane Q) can also supplement visual inspection of plots. In addition to MR-Egger plots, funnel plots can be used to investigate the presence of horizontal pleiotropy. Asymmetry in the funnel plot could indicate a violation of the second MR assumption through pleiotropy [34]. Consistent with the MR-Egger intercept, the funnel plot of LDL-CHD Wald ratios does not provide evidence of asymmetry (Fig. 5).

Target Discovery Using Mendelian Randomization

Fig. 4 Forest plot of gIVs for LDL and CHD risk. X-axis shows each gIV’s Wald ratio estimates

15

16

Daniel S. Evans

Fig. 5 Funnel plot of gIVs for LDL and CHD risk. X-axis shows the Wald ratio estimate for each gIV, and the y-axis is the inverse of the SE of each Wald ratio estimate

To investigate whether the MR estimate is greatly influenced by a single gIV, the MR estimate can be re-estimated after sequentially leaving one gIV out. SNPs, or gIVs, that greatly change the MR estimate when dropped can provide information about the influence of outlier gIVs. A single gIV could potentially have a large horizontal pleiotropic effect [34]. For the LDL-CHD relationship, the MR estimate does not change outside of the confidence intervals of the overall MR estimate when any single gIV is dropped (Fig. 6).

4

Notes 1. Two-sample MR (2SMR) using GWAS results is a powerful approach to identify targets that could be promising starting points for drug development. In addition, 2SMR applied to known drug targets could be a useful approach for drug repurposing and prediction of adverse events.

Target Discovery Using Mendelian Randomization

17

Fig. 6 Leave-one-out (LOO) plot for LDL and CHD MR estimates. X-axis shows the IVW MR estimate for each LOO estimate and the overall MR estimate

18

Daniel S. Evans

2. In addition to the MR methods discussed in this protocol, new 2SMR methods are continuously being developed that can provide information to inform different research questions and that are robust to varying degrees of violations of MR assumptions. This protocol focused on 2SMR because it can utilize publicly available GWAS results that are based on very large sample sizes. Because of this, applying 2SMR does not require a great deal of resources, but can provide a tremendous amount of insight. As additional studies continue to relate genetic variants with molecular traits, the power of MR will continue to grow.

Acknowledgments This work was supported by NIH U24AG051129. References 1. DiMasi JA, Feldman L, Seckler A, Wilson A (2010) Trends in risks associated with new drug development: success rates for investigational drugs. Clin Pharmacol Ther 87:272– 277 2. Hay M, Thomas DW, Craighead JL et al (2014) Clinical development success rates for investigational drugs. Nat Biotechnol 32:40– 51 3. Plenge RM, Scolnick EM, Altshuler D (2013) Validating therapeutic targets through human genetics. Nat Rev Drug Discov 12:581–594 4. Ioannidis JPA (2012) Extrapolating from animals to humans. Sci Transl Med 4:151ps15 5. Perel P, Roberts I, Sena E et al (2007) Comparison of treatment effects between animal experiments and clinical trials: systematic review. BMJ 334:197 6. Smith GD, Ebrahim S (2002) Data dredging, bias, or confounding. BMJ 325:1437–1438 7. Lash TL, VanderWeele TJ, Haneause S, Rothman K (2020) Modern epidemiology, 4th edn. Wolters Kluwer Health 8. Forshed J (2017) Experimental design in clinical ‘omics biomarker discovery. J Proteome Res 16:3954–3960 9. Epidemiology for the uninitiated. https:// w w w. b m j . c o m / a b o u t - b m j / r e s o u r c e s readers/publications/epidemiology-uniniti ated. Accessed 10 Oct 2021 10. Bennett DA, Holmes MV (2017) Mendelian randomisation in cardiovascular research: an introduction for clinicians. Heart 103:1400– 1407

11. Savitz DA (2014) Invited commentary: interpreting associations between exposure biomarkers and pregnancy outcome. Am J Epidemiol 179:545–547 12. Schadt EE, Lamb J, Yang X et al (2005) An integrative genomics approach to infer causal associations between gene expression and disease. Nat Genet 37:710–717 13. Smith GD, Ebrahim S (2003) “Mendelian randomization”: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol 32: 1–22 14. Fang H, ULTRA-DD Consortium, De Wolf H, et al (2019) A genetics-led approach defines the drug target landscape of 30 immune-related traits. Nat Genet 51: 1082–1091 15. Plenge RM (2019) Priority index for human genetics and drug discovery. Nat Genet 51: 1073–1075 16. Estrada K, Froelich S, Wuster A et al (2021) Identifying therapeutic drug targets using bidirectional effect genes. Nat Commun 12:2224 17. Sanseau P, Agarwal P, Barnes MR et al (2012) Use of genome-wide association studies for drug repositioning. Nat Biotechnol 30:317– 320 18. Nelson MR, Tipney H, Painter JL et al (2015) The support of human genetic evidence for approved drug indications. Nat Genet 47: 856–860

Target Discovery Using Mendelian Randomization 19. Smith GD, Ebrahim S (2004) Mendelian randomization: prospects, potentials, and limitations. Int J Epidemiol 33:30–42 20. Maciejewski ML, Brookhart MA (2019) Using instrumental variables to address bias from unobserved confounders. JAMA 321:2124– 2125 21. Angrist JD, Krueger AB (1992) The effect of age at school entry on educational attainment: an application of instrumental variables with moments from two samples. J Am Stat Assoc 87:328–336 22. Abifadel M, Varret M, Rabe`s J-P et al (2003) Mutations in PCSK9 cause autosomal dominant hypercholesterolemia. Nat Genet 34: 154–156 23. Cohen J, Pertsemlidis A, Kotowski IK et al (2005) Low LDL cholesterol in individuals of African descent resulting from frequent nonsense mutations in PCSK9. Nat Genet 37:161– 165 24. Kotowski IK, Pertsemlidis A, Luke A et al (2006) A spectrum of PCSK9 alleles contributes to plasma levels of low-density lipoprotein cholesterol. Am J Hum Genet 78:410– 422 25. Cohen JC, Boerwinkle E, Mosley TH Jr, Hobbs HH (2006) Sequence variations in PCSK9, low LDL, and protection against coronary heart disease. N Engl J Med 354:1264– 1272 26. Davey Smith G, Hemani G (2014) Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet 23:R89–R98 27. Haycock PC, Burgess S, Wade KH et al (2016) Best (but oft-forgotten) practices: the design, analysis, and interpretation of Mendelian randomization studies. Am J Clin Nutr 103:965– 978 28. Zheng J, Baird D, Borges M-C et al (2017) Recent developments in mendelian randomization studies. Curr Epidemiol Rep 4:330–345 29. Pierce BL, Burgess S (2013) Efficient design for Mendelian randomization studies: subsample and 2-sample instrumental variable estimators. Am J Epidemiol 178:1177–1184 30. Tchetgen Tchetgen EJ, Walter S, Glymour MM (2013) Commentary: building an evidence base for Mendelian randomization studies: assessing the validity and strength of proposed genetic instrumental variables. Int J Epidemiol 42:328–331 31. Brion MJ, Shakhbazov K, Visscher PM (2013) Calculating statistical power in Mendelian randomization studies. Int J Epidemiol 42:1497– 1501

19

32. Buniello A, MacArthur JAL, Cerezo M et al (2019) The NHGRI-EBI GWAS Catalog of published genome-wide association studies, targeted arrays and summary statistics 2019. Nucleic Acids Res 47:D1005–D1012 33. GTEx Consortium (2020) The GTEx Consortium atlas of genetic regulatory effects across human tissues. Science 369:1318–1330 34. Hemani G, Zheng J, Elsworth B et al (2018) The MR-Base platform supports systematic causal inference across the human phenome. Elife 7. https://doi.org/10.7554/eLife. 34408 35. Taylor K, Davey Smith G, Relton CL et al (2019) Prioritizing putative influential genes in cardiovascular disease susceptibility by applying tissue-specific Mendelian randomization. Genome Med 11:6 36. Deelen J, Evans DS, Arking DE et al (2019) A meta-analysis of genome-wide association studies identifies multiple longevity genes. Nat Commun 10:3669 37. Baird DA, Liu JZ, Zheng J et al (2021) Identifying drug targets for neurological and psychiatric disease via genetics and the brain transcriptome. PLoS Genet 17:e1009224 38. Millard LAC, Davies NM, Timpson NJ et al (2015) MR-PheWAS: hypothesis prioritization among potential causal effects of body mass index on many outcomes, using Mendelian randomization. Sci Rep 5:16645 39. Evans DM, Brion MJ, Paternoster L et al (2013) Mining the human phenome using allelic scores that index biological intermediates. PLoS Genet 9:e1003919 40. Zheng J, Haberland V, Baird D et al (2020) Phenome-wide Mendelian randomization mapping the influence of the plasma proteome on complex diseases. Nat Genet 52:1122– 1131 41. Richardson TG, Hemani G, Gaunt TR et al (2020) A transcriptome-wide Mendelian randomization study to uncover tissue-dependent regulatory mechanisms across the human phenome. Nat Commun 11:185 42. Walker VM, Davey Smith G, Davies NM, Martin RM (2017) Mendelian randomization: a novel approach for the prediction of adverse drug events and drug repurposing opportunities. Int J Epidemiol 46:2078–2089 43. Gill D, Georgakis MK, Walker VM et al (2021) Mendelian randomization for studying the effects of perturbing drug targets. Wellcome Open Res 6:16 44. Interleukin-6 Receptor Mendelian Randomisation Analysis (IL6R MR) Consortium (2012) The interleukin-6 receptor as a target for

20

Daniel S. Evans

prevention of coronary heart disease: a Mendelian randomisation analysis. Lancet 379:1214– 1224 45. Ferreira RC, Freitag DF, Cutler AJ et al (2013) Functional IL6R 358Ala allele impairs classical IL-6 receptor signaling and influences risk of diverse inflammatory diseases. PLoS Genet 9: e1003444 46. Mokry LE, Zhou S, Guo C et al (2019) Interleukin-18 as a drug repositioning opportunity for inflammatory bowel disease: a Mendelian randomization study. Sci Rep 9:9386 47. Schmidt AF, Holmes MV, Preiss D et al (2019) Phenome-wide association analysis of LDL-cholesterol lowering genetic variants in PCSK9. BMC Cardiovasc Disord 19:240 48. Thomas DC, Lawlor DA, Thompson JR (2007) Re: estimation of bias in nongenetic observational studies using “Mendelian triangulation” by Bautista et al. Ann Epidemiol 17: 511–513 49. Burgess S, Butterworth A, Thompson SG (2013) Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol 37:658–665 50. Global Lipids Genetics Consortium, Willer CJ, Schmidt EM et al (2013) Discovery and refinement of loci associated with lipid levels. Nat Genet 45:1274–1283 51. Nikpay M, Goel A, Won H-H et al (2015) A comprehensive 1,000 genomes-based genome-

wide association meta-analysis of coronary artery disease. Nat Genet 47:1121–1130 52. Chang CC, Chow CC, Tellier LC et al (2015) Second-generation PLINK: rising to the challenge of larger and richer datasets. Gigascience 4:7 53. (2005) SNP FAQ archive. The dbSNP mapping process. In: National Center for Biotechnology Information (US), 2005–. https:// www.ncbi.nlm.nih.gov/books/NBK573560/ 54. McLaren W, Gil L, Hunt SE et al (2016) The ensembl variant effect predictor. Genome Biol 17:122 55. Bowden J, Davey Smith G, Burgess S (2015) Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol 44:512–525 56. Hartwig FP, Davey Smith G, Bowden J (2017) Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int J Epidemiol 46:1985–1998 57. Bowden J, Davey Smith G, Haycock PC, Burgess S (2016) Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator. Genet Epidemiol 40:304–314 58. Bowden J, Del Greco MF, Minelli C et al (2017) A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat Med 36:1783– 1802

Chapter 2 Human Leukocyte Antigen (HLA) Testing in Pharmacogenomics Ann M. Moyer and Manish J. Gandhi Abstract The genetic region on the short arm of chromosome 6 where the human leukocyte antigen (HLA) genes are located is the major histocompatibility complex. The genes in this region are highly polymorphic, and some loci have a high degree of homology with other genes and pseudogenes. Histocompatibility testing has traditionally been performed in the setting of transplantation and involves determining which specific alleles are present. Several HLA alleles have been associated with disease risk or increased risk of adverse drug reaction (ADR) when treated with certain medications. Testing for these applications differs from traditional histocompatibility in that the desired result is simply presence or absence of the allele of interest, rather than determining which allele is present. At present, the majority of HLA typing is done by molecular methods using commercially available kits. A subset of pharmacogenomics laboratories has developed their own methods, and in some cases, query single nucleotide variants associated with certain HLA alleles rather than directly testing for the allele. In this chapter, a brief introduction to the HLA system is provided, followed by an overview of a variety of testing technologies including those specifically used in pharmacogenomics, and the chapter concludes with details regarding specific HLA alleles associated with ADR. Key words HLA antigens, MHC, Pharmacogenomics, Pharmacogenetics, Hypersensitivity reactions, Immune-mediated adverse drug reaction, HLA typing

1

Introduction to the Human Leukocyte Antigen (HLA) System The major histocompatibility complex (MHC) was initially described in the 1930s by Peter Gorer and George Snell as an important determinant in tissue compatibility among inbred mice [1, 2]. Later, in the 1950s, Jean Dausett, Jon van Rood, and Rose Payne described the presence of alloantigens on leukocytes in multiparous women and patients who had received multiple transfusions [3–5]. Currently, HLA typing can be used for multiple applications, including solid organ and hematopoietic stem cell transplantation, disease predisposition, and risk of adverse drug reactions (ADR).

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_2, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

21

22

Ann M. Moyer and Manish J. Gandhi

The MHC is a group of polymorphic genes that encode cell surface proteins that are important for function of the adaptive immune system. MHC molecules bind antigens either from degraded “self” proteins or from “non-self” pathogens, which are then presented on the cell surface for recognition by T-cells. When a “self” antigen is presented, the immune system will prevent the cell from being targeted. In contrast, when a “non-self” antigen is displayed, the cell will be targeted for destruction by the immune system. The portion of the antigen or peptide that is presented is referred to as an epitope. Although the MHC exists and has similar structure in all jawed vertebrates, the number of genes varies. In humans, the MHC is known as human leukocyte antigen (HLA) system. The 224 genes encoding the classical human MHC are located on chromosome 6 and span 3.6 megabases [6]. However, the extended MHC was described in 2003 and expanded to include 421 coding genes across 7.6 megabases [7, 8]. The MHC and corresponding HLA system can be subdivided into three groups or classes, based on functional and structural differences. Class I and II encode the HLA molecules, while class III genes encode other immune proteins that are not directly involved in antigen processing and presentation. Examples of proteins encoded by class III genes include complement components, cytokines necessary for immune signaling, and heat shock proteins. Classical HLA molecules include an extracellular, membrane spanning, and intracellular domain. The extracellular domain is necessary for binding diverse antigenic peptides and interacting with variable T-cell receptors. 1.1

Class I

The class I proteins are expressed on most nucleated cells and platelets as a single α chain that non-covalently forms a heterodimer with the non-polymorphic β2-microglobulin, which is encoded by a gene located outside of the MHC on chromosome 15 and provides structural support. Class I molecules present antigens to CD8+ cytotoxic T-cells to induce cellular death when “non-self” is detected or to inhibitory receptors on NK cells to prevent killing when “self” is detected. Class I genes encode peptide-binding proteins that present peptides that have undergone intracellular proteasomal degradation, as well as the proteins involved in the intracellular degradation and processing. Class I proteins include classical MHC molecules HLA-A, HLA-B, and HLA-C that are encoded by highly polymorphic genes with the same respective names, as well as the non-classical HLA-E, HLA-F, and HLA-G that are much less polymorphic. The extracellular region of class I proteins includes an α1 and α2 domain, which correspond to the peptide binding domain. The high degree of polymorphism allows for binding diverse peptides and interaction with a variety of T-cell receptors [9].

HLA Testing in Pharmacogenomics

23

1.2

Class II

The class II proteins are typically expressed on antigen-presenting cells, including macrophages, dendritic cells, B-cells, and thymic epithelial cells. Class II proteins are comprised of an α chain and a β chain. Both chains have an extracellular hydrophilic domain with two regions (α1 and α2; β1 and β2), a transmembrane domain, and intracellular domain. The α1 and β1 regions form an antigenbinding domain, while the remainder of the extracellular portion forms an immunoglobulin-like domain. In contrast to class I proteins that present intracellular antigens, class II proteins present extracellular antigens to CD4-positive helper T-cells that stimulate the B-cells to produce antibodies. Class II proteins include the classical HLA-DP, HLA-DQ, and HLA-DR, as well as the non-classical class II proteins HLA-DM and HLA-DO. Because class II proteins have both an α chain and a β chain, at least two genes are necessary to encode each molecule. The gene names include an A or a B to indicate which type of chain is encoded. For example, HLA-DP is encoded by HLA-DPA1 and HLADPB1.

1.3

HLA Genetics

The MHC is region is highly polymorphic and is among the most complex of the human genome [6]. As of the writing of this chapter (August 2021), a total of 30,862 HLA alleles have been described. Of those 22,436 were class I alleles, while 8462 were class II alleles. Within class I, the HLA-B gene is somewhat more polymorphic with 8181 alleles, as compared to HLA-A and HLA-C with 6921 and 6779 alleles, respectively. However, given the location of the genes in a tight cluster on chromosome 6, recombination events between alleles are somewhat rare. Therefore, patterns of specific alleles are often inherited together as extended haplotypes that span genes [10]. The HLA genes function in a codominant fashion meaning that the allele from each parent will be expressed. As such, heterozygotes express a greater diversity of molecules than homozygotes. Within the HLA genes, certain regions are more polymorphic and have been the focus of testing. Exons 2 and 3 of the genes encoding class I proteins correspond to the α1 and α2 domains. Therefore, these two exons have been historically the most important for clinical testing as they include the majority of the variability and are responsible for antigen binding [9]. For class II proteins, the beta chain is significantly more polymorphic and has historically been the focus for typing – specifically, the HLA-DRB1, HLADRB3, HLA-DRB4, HLA-DRB5, HLA-DQB1, and HLA-DPB1 genes [11]. The greatest degree of variability in class II corresponds to the α1 and β1 domains, which are encoded by the second exons of the corresponding genes [12]. Therefore, exon 2 of the β chain has historically been primarily sequenced for class II. The DRB locus is of particular interest as there are multiple active genes and pseudogenes, different combinations of which can be inherited as

24

Ann M. Moyer and Manish J. Gandhi

extended haplotypes [13]. Each haplotype includes the active DRB1 gene and a DRB9 pseudogene. In addition, haplotypes may or may not contain one additional active gene – DRB3, DRB4, or DRB5 – as well as additional inactive pseudogenes. HLA typing can be challenging, in part due to the high degree of variability as demonstrated by the large number of alleles present at each locus, but also due to the presence of homologous genes, including pseudogenes [7]. In addition to individual differences in the encoded amino acid sequence for the antigen binding domains, genetic variation in other areas of the gene may also be significant. For example, genetic variants, such as those in the promoter of the gene, may also lead to differences in antigen expression. Non-genetic variables, such as epigenetic modifications that are not currently tested clinically, may also contribute to variable expression [14]. Due to the large number of individual genes and alleles that comprise the HLA system, a database was needed to catalog and name each allele to ensure consistent nomenclature across both research and clinical laboratories. Therefore, in 1998, the ImMunoGeneTics (IMGT)/HLA database was established [15]. It was later incorporated as a component of the Immuno Polymorphism Database (IPD). As the cost of sequencing has decreased, individuals have increasingly been sequenced beyond exons 2 and 3 of class I genes and beyond exon 2 of the class II β chain genes. Additionally, an increasing number of individuals have been sequenced, leading to a significant expansion in the number of known HLA alleles in recent years (Fig. 1). In many cases, newly identified alleles are rare and may only be detected once or within a family. The distribution of specific alleles and haplotypes varies among populations throughout the world. The Allele Frequency Net Database (AFND) curates the peer-reviewed literature, submissions from populations analyzed at International HLA and Immunogenetics Workshops, and submissions from individual laboratories to catalog and display the allele frequencies of HLA and other immune genes among worldwide populations [16]. As of August 2021, AFND included information on HLA allele frequency from 1289 population studies.

2

HLA Nomenclature Similar to other pharmacogenes, the HLA genes are described using specialized nomenclature. The HLA alleles are also described by a star-allele system but differ from other pharmacogenes in that individual alleles can be described to different levels, with colons separating the digits or levels (Fig. 2) [17]. The gene is described with the prefix “HLA” followed by a hyphen and then a letter

HLA Testing in Pharmacogenomics

25

Fig. 1 Number of HLA class I and class II alleles described over time in the ImMunoGeneTics (IMGT)/HLA database. (This figure has been reproduced with kind permission of Prof Steven GE Marsh, Anthony Nolan Research Institute, London (hla.alleles.org) [15, 55, 56])

corresponding to the specific gene. The first digits immediately after the star (asterisk) and before the first colon describe the allele group to a level that corresponds to the serological antigen. This is called the first field or allele group. The next digits after the first colon are considered the second field and describe the specific HLA protein. The subtypes, which differ from each other in amino acid sequence, were assigned in the order in which they were described. The second field is then divided from the next set of digits by another colon. The third set of digits, or third field, differentiates alleles with the same amino acid sequence (described at the second field) that differ only by coding nucleotide sequence. The variation at the third field is in essence describing synonymous or “silent” variants. Finally, the fourth field, present after the third colon, differentiates alleles that differ only by non-coding nucleotide changes in the introns or the 50 or 30 untranslated regions. In addition to the numbers described in the first through fourth fields, the allele name may also include an optional suffix to describe the

26

Ann M. Moyer and Manish J. Gandhi

Fig. 2 HLA nomenclature. The figure highlights each component of the HLA allele name. The gene is indicated after the HLA prefix and hyphen. Then an asterisk is used as a separator before the first field, which is represented by two digits. A colon is used to separate each of the fields, with up to four fields listed. The second field indicates the specific HLA protein, while the third field is used to describe a nucleotide change within the coding region that does not change the encoded amino acid. The fourth field describes changes in non-coding regions. Finally, a suffix may be utilized to indicate changes in expression, and the potential suffixes are listed in the table within the diagram. (This figure has been reproduced with kind permission of Prof Steven GE Marsh, Anthony Nolan Research Institute, London (hla.alleles.org) [15, 55, 56])

expression of the allele. For example, if an allele is denoted with an “N” (e.g., HLA-A*24:09N), the allele is a “null” allele that is not expressed. Of note, prior to the nomenclature revision in 2010, an asterisk (*) was not used. Since 2010, the asterisk indicates typing was performed by a molecular method, and the first field will have a minimum of two digits (e.g., HLA-A*02). When typing was performed by a serological method, no asterisk is included, and there may be only one digit (e.g., A2). HLA typing can be performed to different levels of resolution. Low-resolution typing refers to typing at the level of the first field or serologic equivalent. High resolution refers to the ability to differentiate between alleles with the different amino acid sequences for the antigen binding site, which corresponds to the second field. Finally, allelic resolution refers to the differentiation of unique nucleotide sequences, which corresponds to all of the digits currently used for an allele name. In some cases, this may be to the third or fourth field, such as when synonymous and/or non-coding changes have been identified, or to only the second field if no variants beyond amino acid changes have been defined. For most pharmacogenomic applications, typing to the second field is required.

HLA Testing in Pharmacogenomics

3

27

Examples and Mechanisms of HLA-Associated Adverse Drug Reactions

3.1 HLA and Adverse Drug Reactions (ADR)

Taken together, ADR are a significant problem in health care that account for approximately 6.5% of hospitalizations in the United States, Canada, and United Kingdom and can also result in a mortality rate of 0.13% [18–20]. Many examples in pharmacogenomics involve dose-related adverse drug reactions (ADR) due to genetic variants in drug metabolizing enzymes that result in a patient metabolizing a medication too quickly or too slowly. A subset of ADR are dose-independent idiosyncratic reactions that are immune-related and are more commonly encountered among patients with specific HLA types. These reactions tend to be rare, which makes them difficult to study and difficult to detect even during large clinical trials, and the mechanisms are not always well understood. However, they can be severe and, in some cases, fatal. Therefore, HLA typing prior to prescribing these medications and selecting an alternate medication for patients positive for the associated allele could result in avoidance of these reactions. HLA-related ADR are typically severe reactions resulting in skin reactions, hepatic failure, and/or significant leukopenia/agranulocytosis. Skin reactions can range from a somewhat mild rash or maculopapular exanthema (MPE) to more severe fixed drug eruption (FDE) or acute generalized exanthematous pustulosis (AGEP) to life-threatening reactions including drug reactions with eosinophilia and systemic symptoms (DRESS) and severe cutaneous drug reactions (SCARs), such as Stevens-Johnson syndrome (SJS), and toxic epidermal necrolysis (TEN). When the ADR impacts the liver, it is often referred to as drug-induced liver injury (DILI), which differs from an overdose-related hepatic injury in that it is not dose related [21]. DILI accounts for approximately 10–15% of the total cases of acute liver failure in the United States [22]. Fortunately, HLA-related ADR are generally rare reactions. However, this makes them more difficult to study and to identify the underlying genetic association. Most HLA-induced ADR have been identified through genome-wide association studies (GWAS) or case-control candidate gene studies. These studies typically require a large number of affected and unaffected individuals. Further complicating these studies is that the frequency of the HLA alleles differs among populations, so the HLA allele associated with a particular reaction may be influenced by the frequency of the allele in the population studied. To facilitate studies of HLA and ADR, the AFND catalogs not only provide the worldwide allele frequencies but also curate the peer-reviewed publications and allow users to search by drug and/or allele [23]. The drug of interest, proportion of cases and controls carrying the allele, and ethnicity of the population are summarized in the database entries.

28

Ann M. Moyer and Manish J. Gandhi

Table 1 HLA alleles reported to be associated with adverse drug reactions Drug

HLA-allele

References

Abacavir

B*57:01

[57]

Allopurinol

B*58:01, A*33:03, C*03:02

[58–61]

Ampicillin

A2, DRw52

[62]

Amoxicillinclavulanate

A*02:01, A*30:02, B*18:01, DRB1*15:01-DRB5*01:01DQB1*06:02, DQA1*01:02

[63–67]

Antithyroid drugs

B*27:05, B*38:02, DRB1*08:03

[68]

Aspirin

DPB*03:01, DRB1*13:02, DQB1*06:09

[69]

Carbamazepine

A*02:01, A*31:01, A*33:03, B*15:02, B*15:11, B*39:02, B*51: [26, 70–75] 01, B*57:01, B*58:01, C*03:02, C*08:01, DQB1*03:03, DRB1*07:01, DRB1*12:02

Clozapine

DQB1*05:02, B38, B39, B67, B*59:01

[76]

Dapsone

B*13:01

[77]

Feprazone

B22

[78]

Fenofibrate

A*33:01

[79]

Flucloxacillin

B*57:01, B*57:03, DRB1*07:01

[80, 81]

Flupirtine

DRB1*16:01-DQB1*05:02

[82]

Hydralazine

DR4

[83]

Infliximab

B*39:01

[84]

Lamotrigine

A*31:01, A*68:01, B*13:01, B*15:02, B*15:19, B*44:08, B*55: [85–87] 01, B*58:01, B*81:01, C*08:01, DQB1*06:09, DRB1*16:02

Lapatinib

DQA1*02:01, DRB1*07:01, DQB1*02:02

[88–90]

Levamisole

B27

[91]

Lumiracoxib

DRB1*15:01, DRB5*01:01, DQA1*01:02, DQB1*06:02

[92]

Methazolamide

B*59:01, C*03:02

[93, 94]

Minocycline

B*35:02

[95]

Nevirapine

B*14:02, B*35:05, B*58:01, C*04:01, C*08, DRB1*01:01, DRB1*01:02

[96–99]

Oxcarbazepine

B*13:02, B*15:02, B*15:19, B*15:27, B*15:58, B*27:04, B*27: [25, 86, 100, 09, B*38:02, B*48:04, B*56:01 101]

Oxicam

A2, B12

[102]

Pazopanib

B*57:01

[103]

Phenytoin

B*13:01, B*15:02, C*08:01, DRB1*16:02

[27, 86]

B*13:01, B*15:05, B*39:01

[104] (continued)

HLA Testing in Pharmacogenomics

29

Table 1 (continued) Drug

HLA-allele

References

Sulfamethoxazole

A29, B12, DR7, B*15:02, B*38, C*06:02, C*08:01

[105–107]

Terbinafine

A*33:01

[79]

Ticlopidine

A*33:01, A*33:03, B*44:03, C*14:03, DRB1*13:02, DQB1*06:04

[79, 108]

Ximelagatran

DRB1*07:01

[109]

Salazosulfapyridine (sulfasalazine)

At present, the most commonly tested alleles with the strongest evidence include HLA-B*57:01 for abacavir, HLA-B*58:01 for allopurinol, HLA-B*15:02 for phenytoin and carbamazepine, and HLA-A*31:01 for carbamazepine. Recommendations to consider patient genotype prior to prescription of these medications have been incorporated into the US Food and Drug medication labels and are the subject of guidelines published by the Clinical Pharmacogenetics Implementation Consortium (CPIC) [24–28]. Additional examples of HLA alleles associated with adverse drug reactions are provided in Table 1. 3.2

Mechanisms

Currently, there are four main mechanisms that have been proposed to explain how the drug and HLA molecule interact resulting in toxicity. First, in the “hapten/prohapten theory,” the drug or a metabolite binds covalently to an endogenous self-peptide and forms a complex with new epitopes. The drug is thought to be too small to elicit an immune response on its own but is larger when bound to a peptide. This results in the endogenous self-peptide being recognized as foreign, such that when the T-cell receptor (TCR) encounters the HLA molecule presenting drug-peptide complex, an immune response is mounted. Penicillin allergy represents an example of this model [29]. While there are some HLA associations with penicillin allergy, it is not yet well established, and clinically laboratories have generally not adopted HLA testing for this indication. The second mechanism is called “pharmacological interaction with immune receptors (p-i) concept.” In this mechanism, the drug or metabolite non-covalently binds the HLA molecule or the TCR. Unlike the hapten/prohapten theory, this interaction between the drug/HLA/TCR does not depend on the classic intracellular antigen processing pathway, but an endogenous peptide may be required to be presented on the HLA molecule for stability of the drug/HLA/TCR complex. The interaction of carbamazepine with HLA-B*15:02 leading to SJS/TEN is an

30

Ann M. Moyer and Manish J. Gandhi

example of this mechanism [30]. In the “altered peptide repertoire” theory, the drug is thought to bind in the peptide-binding groove of the HLA molecule and alter the shape of the binding cleft, which then would change the peptide specificity. This results in self-peptides appearing foreign and eliciting a T-cell response. The interaction of abacavir with HLA-B*57:01 follows this model and is one of the examples that is best understood [31–33]. Finally, the “altered TCR repertoire” model suggests that a small portion or metabolite of the drug could alter the shape of the TCR such that it may bind differently to HLA/peptide complexes, particularly those presenting a self-peptide that should not elicit a reaction, resulting in an immune response. Sulfamethoxazole may induce hypersensitivity reactions through this mechanism [34].

4

Overview of HLA Typing Methods Historically, serologic techniques were the gold standard for HLA typing based on the principle of complement-dependent cytotoxicity (CDC). Sera was collected from donors with known HLA antibodies. Cells were introduced into wells containing different known donor-derived antibodies, along with complement and a dye. When patient cells had the antigen corresponding to the donor antibody present, complement was activated, resulting in cell death and take-up of the dye. The CDC method was a challenge for Class II typing, and this resulted in moving toward molecular techniques and a better understanding of the HLA diversity. The CDC typing method or serologic equivalent method corresponds to the first field typing and is now considered low-resolution typing. Serologic techniques have been replaced by molecular or DNA-based methods, which have several advantages. First, these DNA-based methods are more specific because the reagents are based on the specific nucleotide sequence. As new alleles are discovered, new reagents can be developed to detect these sequences and incorporated into tests, or the new alleles may be amplified with existing reagents and identified when the sequence generated is compared to reference databases. DNA-based methods allow for typing at different levels of resolution depending on the clinical indication. Finally, DNA-based methods do not require a specific cell type and are generally not impacted by the health of the patient or viability of cells to be tested. HLA typing may differ depending on the indication. For example, in the setting of transplantation, the HLA type of each locus of interest is determined (e.g., at the HLA-A locus, the testing will determine which alleles are present). In contrast, when typing for specific alleles associated with disease (e.g., DQ2 and DQ8 for celiac disease) or with an increased risk for an adverse drug reaction (e.g., HLA-B*57:01 and abacavir hypersensitivity), a specific test

HLA Testing in Pharmacogenomics

31

for the allele(s) of interest may be performed and resulted as “positive” or “negative.” Additionally, the resolution may vary. For solid organ transplantation, low-resolution typing may be acceptable, whereas high resolution is required for stem cell transplantation. For pharmacogenomic purposes, typically typing to the second field is required for positive identification of the allele. This section will provide an overview of molecular typing methods that are currently used, including several that are specifically used in pharmacogenomics. 4.1 Typing by Sequence-Specific Primers (PCR-SSP)

Testing using the sequence-specific primers (SSP) approach allows for determination of presence or absence of the targeted genetic variant directly as part of the PCR process (Fig. 3a). When performing a PCR reaction, a primer that completely matches the target sequence will amplify DNA with greater efficiency than a primer that has one or more mismatches. If the mismatch is at the 30 end of the primer, it will not amplify DNA under standard conditions because Taq polymerase lacks 30 to 50 exonuclease proofreading ability [35]. The PCR-SSP technique, which was originally termed amplification refractory mutation system (ARMS), takes advantage of this principle by utilizing PCR primers that are specifically designed to detect a specific genetic variant [36, 37]. When the allele is present, amplification will occur, and the PCR product can be detected using agarose gel electrophoresis. When the allele is absent, no PCR product will be produced for that reaction. This technique was later applied and adapted for histocompatibility testing and ultimately re-named PCR-SSP [38]. PCR-SSP was the first method developed for histocompatibility testing because it interrogates small regions of DNA. Early Taq polymerases were only capable of producing small amplicons, which are sufficient for this technique. PCR-SSP can be used either as a high-resolution method or as a low-resolution method, depending on the genetic variants targeted. Due to the highly polymorphic nature of the HLA locus, multiple individual reactions are required. While it is possible for a laboratory to develop the necessary primer sets and optimize the reactions, commercial SSP trays are now readily available and commonly used. Commercial trays are typically available separately for class I or class II. Laboratories only querying a specific allele or several alleles, as may be the case for PGx applications, may choose to design their own reagents. The individual PCR reactions are first set up and performed. Each reaction includes both an allele-specific reaction and a control reaction that amplifies a housekeeping gene. The primers for the housekeeping gene are usually at a lower concentration to ensure that its amplification does not outcompete the allele-specific reaction. Once the PCR is completed, the products are run on an agarose gel, along with a sizing ladder. Each reaction is evaluated as positive or negative, based on whether a

32

Ann M. Moyer and Manish J. Gandhi

A.

B.

SSP

Amplificaon

SSO

C.

Real-Time PCR

F.

Targeted Typing by Linkage Disequilibrium

No Amplificaon

Specific Band

Control Band

D.

Sanger Sequencing Exon2

Exon3

E.

Next-Generaon Sequencing

Exon4

HLA-B*15:02 Non-B*15:02

Query database for match and allele call

Query database for match and allele call

Fig. 3 Techniques used for HLA typing. (a) The sequence-specific primer (SSP) technique uses a primer set that is specific to the allele of interest. When the allele is present (blue lines with orange mark), amplification occurs, and the PCR product can be detected by presence of a band on agarose gel electrophoresis. When the allele is absent (blue lines with green mark), the primers (purple arrows) do not produce a product, and no band is visualized. (b) In the sequence-specific oligo (SSO) technique, primers amplify the region of interest. A hybridization step may use a bead-based approach (red spheres representing beads with black lines representing probes). When the complementary amplicon and probe are present, they hybridize. The beads are passed through an instrument with two lasers. The red laser excites the bead, allowing for identification of the specific bead and attached probe, and the green laser excites the hybridized probe/amplicon pair, which generates a signal (bright green symbol). (c) Real-time PCR amplifies the region or allele of interest. A TaqMan hydrolysis probe (left) with a signal (green) and quencher (red) can specifically bind to the amplicon. During the PCR reaction, the bound probe will be degraded, separating the signal from the quencher, which results in a signal. Alternatively, a specific amplicon may be detected using SYBR green, which intercalates between the strands of double-stranded DNA (right). In both cases, increasing signal is detected as the reaction progresses. (d) In Sanger sequencing, an amplicon containing the region(s) of interest is generated using PCR primers (purple arrows). Then a second reaction is performed using sequencing primers (blue arrows) and with incorporation of fluorescently labeled dideoxynucleotides that terminate the reaction. The products are separated by capillary electrophoresis and detected. (e) In next-generation sequencing, the target region may be enriched by a capture-based method (green bars) or PCR reactions (purple arrows). The region of interest is interrogated many times generating “reads” (blue bars, also depicted by height of grey boxes under each nucleotide) and the nucleotide sequence. The data is compared to databases to identify the allele (s) present. (f) Some laboratories select a variant (purple symbol) that is present when the allele of interest is present and absent (green symbol) when the HLA allele of interest is absent and then test for the presence of the variant instead of directly testing the allele. The variant may be an intergenic variant distant from the HLA locus. This approach is not recommended because the linkage patterns may not be consistent across all populations

HLA Testing in Pharmacogenomics

33

band of the expected size for the allele-specific reaction is detected. The control band may be absent if outcompeted by the allelespecific reaction; however, if both the control band and the allelespecific band are absent, the reaction has failed and must be repeated. The pattern of positive reactions is then compared against a chart containing the expected pattern for each allele to determine which alleles are present in the sample tested. PCR-SSP is an excellent technique to type a small number of samples in a short amount of time. In addition, both the forward and the reverse primer can be used to detect the presence or absence of variants, allowing for some phasing (determining whether variants are on the same or opposite allele) to resolve ambiguities. This technique is not well suited for automation or high-throughput settings due to the use of gel electrophoresis. Other detection methods, such as fluorescent probes, may be used for detection rather than a gel. SSP is also not ideal for high throughput due to the requirement for many individual reactions to achieve high resolution. In addition, new alleles are being identified regularly due to sequencing-based techniques (described later in this chapter), which would require incorporation of additional reactions to avoid ambiguous typing. Despite these limitations, this remains a suitable option for PGx testing, given that positive or negative results for only one or a few alleles are necessary. 4.2 Typing by Sequence-Specific Oligonucleotide Probes (PCR-SSO)

Another early method applied to histocompatibility testing was the sequence-specific oligonucleotide probes approach (Fig. 3b). This technique utilizes a polymerase chain reaction (PCR) to first generate an amplicon for the region of interest and then combines it with a hybridization step where probes with specific known sequences are used to determine which allele is present. This technique was first applied to class II typing due to the limitations of serology at the DR locus [39, 40]. PCR-SSO requires a longer PCR amplicon than PCR-SSP and was not feasible until more robust Taq polymerase enzymes were available. For PCR-SSO-based typing, primers are designed in a conserved region of an exon for each locus to be typed. The primers are used to amplify that locus, and the amplicons generated include any variants located between the primers. For typing class I loci, typically exons 2, 3, and 4 are amplified, while for class II, only exons 2 and 3 are included. Next, the negatively charged DNA amplicons generated are blotted onto a membrane, such as a positively charged nylon membrane that allows for strong binding. Then the membrane with bound DNA amplicon is exposed to a hybridization buffer and oligonucleotide probe with known specific sequence corresponding to an allele or group of alleles. The oligonucleotide probe, which is usually approximately 20 base pairs long, only hybridizes to the amplicon when the sequence matches. A different probe is required for each variant or allele that is being

34

Ann M. Moyer and Manish J. Gandhi

tested. Bound probe is detected, while unbound probe is washed away. Enzyme-linked and chemiluminescent systems are used for detection. Finally, the HLA genotype can be determined based on the pattern of positive reactions. PCR-SSO typing is useful for histocompatibility typing of many samples simultaneously but is time-consuming and thus not well suited to meet quick turnaround times that may be required in some clinical scenarios. The SSO method requires a large number of probes to obtain unambiguous results for certain combinations of heterozygous alleles. Alternatively, ambiguous typings could be resolved by use of a two-step process where the probes used in the second step depend on the findings from the first step. An additional advantage of SSO is that if a previously undescribed allele is encountered, it will likely be amplified and result in a pattern that does not match the expected pattern for known alleles. However, if the novel allele has variants outside of the regions included in this assay (exons 2, 3, and 4 for class I; exons 2 and 3 for class II), it may not be recognized as a novel allele and may be mis-classified as a previously described allele. To allow for typing a low number of samples, a modification of SSO, referred to as reverse-SSO (rSSO), was developed [41, 42]. In contrast to SSO where the DNA amplicon is bound to the membrane, in rSSO, the oligonucleotide probes are bound to membranes or to beads. The membranes are pre-made and contain all of the probes that are needed for a locus. When PCR amplification is completed on the sample, a label is incorporated, and then the amplicons are hybridized to the membrane-bound probes and detected. Alternative approaches that do not require a membrane have also been developed. For example, a 96-well plate may be used where the PCR product is adhered to the walls of the well and the binding of the oligonucleotide probe triggers detection through a chemiluminescent or enzyme-linked color-based system. Another variation utilizes the Luminex xMAP® technology (Fig. 3b). In this version, the oligonucleotide probes are covalently linked to polystyrene microbeads. The microbeads have two fluorophores in differing proportions, resulting in approximately 100 distinct beads. A different oligonucleotide probe is attached to each bead type. After the PCR amplification step, the amplicons are introduced to the set of beads, and each amplicon hybridizes with the matching probe. Biotin is incorporated into the amplicons, which has high affinity for streptavidin. The Luminex system uses a streptavidin phycoerythrin (SA-PE) conjugate to then fluorescently label the amplicons. Then, the sample is run on the Luminex instrument, which is similar to a flow cytometer. The instrument uses a red laser to excite the dyes that are in the microbeads and a green laser to excite the SA-PE on the amplicons. As each bead passes through the detection step, the instrument recognizes the pattern of the internal dye to determine which bead it is (and thus which

HLA Testing in Pharmacogenomics

35

corresponding probe is attached) and the signal from the SA-PE to determine if the amplicon is present (positive) or absent (negative). Similar to traditional PCR-SSO, the pattern is used to determine the HLA genotype of the sample. This can all occur within one well of a 96-well PCR plate, allowing multiple samples to be tested simultaneously, and can be automated. Commercial kits are available, making this technique relatively simple for laboratories to adopt. 4.3 SequencingBased Typing (SBT)

The SSO and SSP techniques may be used in combination with sequencing-based typing (SBT). In sequencing-based typing, the region of interest is first amplified and then is sequenced to determine each nucleotide present between the primers using either Sanger sequencing or massively parallel sequencing (also known as next-generation sequencing, NGS). The sequence generated is compared to the IMGT/HLA database to identify the allele present. An SSO or SSP approach can provide a non-specific result that determines which primers should be used for the SBT approach for definitive typing. SBT can also be used independently from SSO and SSP, though in some cases where the phase of variants present is ambiguous, then SSO or SSP can be used to complement SBT. Sequencing has historically been performed primarily by Sanger sequencing using a dye terminator chemistry (Fig. 3d). In Sanger sequencing, a PCR reaction is first performed to generate a template. Sanger sequencing has focused primarily on exons 2, 3, and 4 for class I loci and exons 2 and 3 for class II because the bulk of the clinically significant variation is located in these regions, and Sanger sequencing is a relatively expensive technique. Once the PCR reaction has generated the template for the region(s) of interest, then another PCR reaction is performed, but in this case, the reaction contains fluorescently labeled dideoxynucleotides (ddNTPs) in addition to the deoxynucleotides that are present in a typical PCR. The ddNTPs terminate the reaction when they are incorporated, so the amplicons are of varying lengths that each end with a fluorescently labeled A, C, G, or T – each nucleotide has a different label so they can be differentiated from one another. Then these products are separated using capillary electrophoresis. The amplicons of variable length are injected into the capillary and then migrate through with smaller fragments moving more quickly and larger fragments more slowly. At the end, each fragment passes through a laser that results in fluorescence of the dye, which is detected. The sequence can then be assembled based on the order of the fluorescent dyes detected, which corresponds to the nucleotide sequence. Once the nucleotide sequence has been determined, the alleles present are identified by comparing the sequence against a database of all known alleles. Commercial kits have been developed and, in most cases, amplify a larger region (more exons) to allow for better resolution. An advantage of Sanger sequencing is

36

Ann M. Moyer and Manish J. Gandhi

that it is highly accurate. A disadvantage is that when multiple heterozygous variants are present, Sanger sequencing does not allow for phasing of the variants, which may lead to some ambiguity. In recent years, sequencing has begun to transition from Sanger to massively parallel sequencing, which is also known as nextgeneration sequencing (NGS), due to its accuracy and efficiency [43] (Fig. 3e). Although the most widely used NGS platforms at present rely on short reads of 100–200 base pairs, a major advantage over Sanger sequencing is that the phase can be determined for variants detected within the same read. Phasing becomes an even greater advantage with long-read NGS platforms; however, in some cases, phasing remains a challenge leading to ambiguous results. In addition, rather than focusing on exons 2, 3, and 4 for class I or exons 2 and 3 for class II genes, NGS allows for greater multiplexing such that entire genes may be sequenced. Therefore, using NGS-based approaches and sequencing entire genes (class I and DRB1 or a major part of the genes DQB1 and DPB1) has led to identification of many new alleles and may allow for better understanding of the impact of non-coding variants on expression and regulation of HLA genes. When NGS is used for histocompatibility testing, the target region is first enriched by either capture-based method, long-range PCR, or creating short PCR amplicons across the region of interest. When a PCR amplicon-based method is used, there is risk for allele dropout, which could occur for a variety of technical reasons, including a variant under a primer [44]. Commercial kits are available with carefully designed primer sets to minimize dropout. Allele dropout is generally less problematic for capture-based methods because the baits can tolerate a few mismatches while still allowing for hybridization. The disadvantage of capture-based methods is that they may allow for off-target capture, decreasing specificity. After target enrichment, a library preparation is performed to ligate the DNA fragments to be sequenced to indexed adaptors. After sequencing, the data must be analyzed using specialized software to align the reads to the HLA region and call the alleles present. In some cases, genetic sequencing may have originally been performed for non-HLA purposes, but there is reason to evaluate the HLA region. For example, exome sequencing is performed on a research basis in a large cohort. Later, investigators realize there is a signal in the HLA region associated with a disease, or perhaps the cohort is used to study a disease that may be in part immunemediated. Therefore, the investigators would like to study whether any specific HLA alleles are associated with risk of developing the disease phenotype. Alternatively, a rare adverse drug reaction may be noted in the cohort, and the investigators wish to determine whether an HLA allele is associated. Having the ability to perform HLA typing on an existing exome or genome sequencing dataset

HLA Testing in Pharmacogenomics

37

may allow for additional research beyond the original intent of the dataset without requiring significant additional funding. At present, there are a number of software packages that allow for this capability with a high degree of accuracy. There are differences among the programs in accuracy among the loci as well as which specific loci are included [45]. 4.4 Testing for Presence or Absence of Specific Alleles

While the same techniques used for traditional HLA typing could also be used for pharmacogenomic purposes, when performing testing to predict toxicity of a particular drug, often, it is sufficient to determine whether the allele(s) associated with toxicity is present or absent. A positive or negative result is sufficient for these purposes without needing to determine whether the allele is heterozygous or homozygous when present. Multiple techniques have been used for pharmacogenomic applications, with differing degrees of specificity. Patch testing has the lowest specificity at 60–70%, followed by flow cytometry for monoclonal antibodies at about 80%. Molecular techniques, which are most commonly used in current practice, have higher specificities of >95% for SSO, >97% for SSP, and >99% for real-time PCR [46].

4.4.1 Monoclonal Antibody Technique

Historically, a monoclonal antibody, mAb 3E12, was used for typing that could detect HLA-B17 (HLA-B*57 and HLA-B*58 are both members of the B17 group) [47, 48]. The antibody was labeled with phycoerythrin and incubated with blood from the patient. Then, the sample was analyzed using flow cytometry to determine if the antibody bound to patient cells, indicating the presence of the B17 antigen. While this was an inexpensive and quick technique to perform, it was less specific and sensitive than current PCR-based strategies, and this technique could not identify which allele (B*57 or B*58) was present. Therefore, patients who were positive required further testing to specifically determine whether the positive signal corresponded to a B*57:01, B*58:01, or another allele. Use of the monoclonal antibody approach has now been replaced by other techniques, such as real-time PCR.

4.4.2

Several approaches to real-time PCR have been used. Several commercial methods are available for HLA typing that provide accurate first field typing. Similar to the monoclonal antibody technique, these methods can be used to screen out the negative cases and proceed to high-resolution typing on only the subset of cases that are potentially positive for a pharmacogenomically relevant allele. Alternatively, real-time PCR primers may be designed to specifically amplify the allele of interest, similar to the SSP technique (Fig. 3c). The amplicon can then be detected either using SYBR green, which is a fluorescent dye that intercalates between the two strands of double-stranded DNA, or using TaqMan hydrolysis probes. When SYBR green is used, the fluorescent signal increases with each

Real-Time PCR

38

Ann M. Moyer and Manish J. Gandhi

subsequent PCR cycle until the reaction plateaus. When the two strands dissociate in subsequent cycles, the SYBR green signal is lost, resulting in a curve that can be visualized. The shape of the melt curve, or dissociation plot, is specific to the expected amplicon and can be used as additional information to support the presence or absence of the specific allele. When TaqMan is used, a probe is designed that is specific to the expected amplicon. The probe has a fluorophore and a quencher. During the PCR cycle, when the target is present and the probe can bind, it will be cleaved, separating the quencher from the fluorophore and resulting in a signal that increases with each subsequent cycle. The dissociation of the two DNA strands cannot be detected, however, and a melt curve is not produced. With both the SYBR green and the TaqMan detection methods, the crossing threshold – the point when the fluorescent signal is greater than background indicating the presence of the amplicon – for the target amplicon is compared against an internal control, which is typically a housekeeping gene with little individual genetic variation. The purpose of the internal control is to ensure that there are no PCR inhibitors present. Therefore, when the internal control is positive and the HLA target is negative or has a high crossing point, the sample is considered negative. If both are positive, the patient sample is positive for the HLA target. If both are negative, the reaction has failed and is repeated. The Ct (cycle threshold or crossing threshold) for the HLA reaction must be within a specified number of cycles of the internal control, defined during assay development, because a variant under the primer for the HLA target reaction could result in amplification, but it would typically be less efficient and have a higher Ct. Given the high degree of polymorphism, it is possible that the primers may amplify more than one allele; however, if a variant is present between the primers (within the amplicon), the melt curve will likely have a different shape in the case of SYBR green, or the probe will not bind in the case of TaqMan chemistry. While these assays are typically highly sensitive and specific, when results are unclear, it is best to follow up with a traditional typing technique to specifically identify the alleles present. 4.4.3 Targeted Genotyping Using Linkage Disequilibrium

Many laboratories performing pharmacogenomic testing are using multiplexed methods to detect single nucleotide variants (SNV) or small insertions/deletions across many genes simultaneously to offer multi-gene panels. Therefore, having a technique to test for the relevant HLA alleles on the same platform is attractive due to its convenience. Some laboratories have begun to use a tagging-SNV approach where they identify an SNV in linkage disequilibrium (LD) with the HLA allele of interest [49, 50] (Fig. 3f). This means that the variant nucleotide is thought to be present when the allele of interest is present and absent when the HLA allele of

HLA Testing in Pharmacogenomics

39

interest is absent. In some cases, the SNV may be within or near the HLA gene, while in others, it may be an intergenic variant distant from the HLA locus. While at first glance this may seem to be an excellent approach, the HLA region is highly polymorphic, and while the SNV may be in LD in some populations, it may not be associated with the HLA allele of interest in all populations or may not have been studied across populations, and thus the degree of association across populations may be unknown [51]. Therefore, utilizing a direct HLA sequencing or other approach outlined above may be preferable to avoid false-negative and/or falsepositive results [52]. 4.5 Imputing HLA Types Using SNP Data from Genome-Wide Association Studies (GWAS)

5

The HLA type of an individual can be predicted based on SNVs, such as those generated by a chip used for a genome-wide association study (GWAS) that are located near and within the HLA region by taking advantage of the extended haplotype structure. Similar to the discussion of exome sequencing above, there are many GWAS datasets containing large cohorts of patients that were not originally performed with HLA applications in mind, but could potentially be leveraged for additional research without needing to incur costs from repeating the genotyping if HLA data could be imputed [53]. To impute HLA types from GWAS data, a training set is required that provides information regarding how the SNVs are associated with HLA alleles in the population. Then the HLA types of new samples from the same population or cohort may be determined based on the SNVs present. In general, this is a much less accurate method than direct sequencing but can be up to 92–99% accurate to the second field for cohorts of individuals of European descent. This technique tends to perform less well though for populations of other ancestry unless a training set is used that corresponds to the population being studied [53, 54]. A major limitation of this technique is that rare alleles are difficult or impossible to predict.

Conclusions Multiple HLA alleles have been associated with increased risk for adverse drug reactions. These reactions may be severe or fatal, suggesting that HLA testing is an important tool to predict and avoid these reactions. While pharmacogenomic laboratories are familiar with highly polymorphic genes and homology, the HLA region presents an even greater challenge when implementing clinical testing. A variety of techniques are available for histocompatibility testing, including many commercially available kits, and could also be applied to pharmacogenomics. Although use of a targeted genotyping strategy that relies on linkage disequilibrium is attractive to pharmacogenomic laboratories, this method should be used

40

Ann M. Moyer and Manish J. Gandhi

with great caution due to differences in haplotype structures among populations. Due to the complexity of HLA testing, pharmacogenomic laboratories may benefit from partnering with a traditional histocompatibility testing laboratory for assistance with test design as well as confirmatory testing to clarify ambiguous results. References 1. Klein J (1986) Seeds of time: fifty years ago Peter A. Gorer discovered the H-2 complex. Immunogenetics 24(6):331–338. https:// doi.org/10.1007/BF00377947 2. Snell GD, Higgins GF (1951) Alleles at the histocompatibility-2 locus in the mouse as determined by tumor transplantation. Genetics 36(3):306–310 3. Dausset J (1958) [iso-leuko-antibodies]. Acta Haematol 20(1–4):156–166. https://doi. org/10.1159/000205478 4. Payne R, Rolfs MR (1958) Fetomaternal leukocyte incompatibility. J Clin Invest 37(12): 1756–1763. https://doi.org/10.1172/ JCI103768 5. Van Rood JJ, Eernisse JG, Van Leeuwen A (1958) Leucocyte antibodies in sera from pregnant women. Nature 181(4625): 1735–1736. https://doi.org/10.1038/ 1811735a0 6. The MHC Sequencing Consortium (1999) Complete sequence and gene map of a human major histocompatibility complex. Nature 401(6756):921–923. https://doi. org/10.1038/44853 7. Horton R, Wilming L, Rand V et al (2004) Gene map of the extended human MHC. Nat Rev Genet 5(12):889–899. https://doi.org/ 10.1038/nrg1489 8. Mungall AJ, Palmer SA, Sims SK et al (2003) The DNA sequence and analysis of human chromosome 6. Nature 425(6960): 805–811. https://doi.org/10.1038/ nature02055 9. Geraghty DE, Koller BH, Hansen JA et al (1992) The HLA class I gene family includes at least six genes and twelve pseudogenes and gene fragments. J Immunol 149(6): 1934–1946 10. Yunis EJ, Larsen CE, Fernandez-Vina M et al (2003) Inheritable variable sizes of DNA stretches in the human MHC: conserved extended haplotypes and their fragments or blocks. Tissue Antigens 62(1):1–20. https:// doi.org/10.1034/j.1399-0039.2003. 00098.x

11. Kaufman JF, Auffray C, Korman AJ et al (1984) The class II molecules of the human and murine major histocompatibility complex. Cell 36(1):1–13. https://doi.org/10. 1016/0092-8674(84)90068-0 12. Giles RC, Capra JD (1985) Biochemistry of MHC class II molecules. Tissue Antigens 25(2):57–68. https://doi.org/10.1111/j. 1399-0039.1985.tb00415.x 13. Andersson G, Andersson L, Larhammar D et al (1994) Simplifying genetic locus assignment of HLA-DRB genes. Immunol Today 15(2):58–62. https://doi.org/10.1016/ 0167-5699(94)90134-1 14. Carey BS, Poulton KV, Poles A (2019) Factors affecting HLA expression: a review. Int J Immunogenet 46(5):307–320. https://doi. org/10.1111/iji.12443 15. Robinson J, Barker DJ, Georgiou X et al (2020) IPD-IMGT/HLA database. Nucleic Acids Res 48(D1):D948–D955. https://doi. org/10.1093/nar/gkz950 16. Gonzalez-Galarza FF, McCabe A, Santos E et al (2020) Allele frequency net database (AFND) 2020 update: gold-standard data classification, open access genotype data and new query tools. Nucleic Acids Res 48(D1): D783–D788. https://doi.org/10.1093/ nar/gkz1029 17. Nunes E, Heslop H, Fernandez-Vina M et al (2011) Definitions of histocompatibility typing terms. Blood 118(23):e180–e183. https://doi.org/10.1182/blood-201105-353490 18. Lazarou J, Pomeranz BH, Corey PN (1998) Incidence of adverse drug reactions in hospitalized patients: a meta-analysis of prospective studies. JAMA 279(15):1200–1205. https:// doi.org/10.1001/jama.279.15.1200 19. Kvasz M, Allen IE, Gordon MJ et al (2000) Adverse drug reactions in hospitalized patients: a critique of a meta-analysis. MedGenMed 2(2):E3 20. Miguel A, Azevedo LF, Araujo M et al (2012) Frequency of adverse drug reactions in hospitalized patients: a systematic review and metaanalysis. Pharmacoepidemiol Drug Saf

HLA Testing in Pharmacogenomics 21(11):1139–1154. https://doi.org/10. 1002/pds.3309 21. Daly AK (2020) Pharmacogenomics of druginduced liver injury. Adv Mol Pathol 3: 107–115. https://doi.org/10.1016/j.yamp. 2020.07.010 22. Ostapowicz G, Fontana RJ, Schiodt FV et al (2002) Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med 137(12): 947–954. https://doi.org/10.7326/00034819-137-12-200212170-00007 23. Gonzalez-Galarza FF, McCabe A, Melo dos Santos EJ et al (2018) Allele frequency net database. In: Boegel S (ed) HLA typing: methods and protocols, vol 1802. Springer Science, part of Springer Nature, pp 49–62. https://doi.org/10.1007/978-1-49398546-3_4 24. U.S. Food and Drug Administration (2021) Table of pharmacogenomic biomarkers in drug labeling. https://www.fda.gov/drugs/ science-and-research-drugs/tablepharmacogenomic-biomarkers-drug-label ing. Accessed 14 Aug 2021 25. Martin MA, Hoffman JM, Freimuth RR et al (2014) Clinical pharmacogenetics implementation consortium guidelines for HLA-B genotype and abacavir dosing: 2014 update. Clin Pharmacol Ther 95(5):499–500. https://doi.org/10.1038/clpt.2014.38 26. Phillips EJ, Sukasem C, Whirl-Carrillo M et al (2018) Clinical pharmacogenetics implementation consortium guideline for HLA genotype and use of carbamazepine and oxcarbazepine: 2017 update. Clin Pharmacol Ther 103(4):574–581. https://doi.org/10. 1002/cpt.1004 27. Karnes JH, Rettie AE, Somogyi AA et al (2021) Clinical pharmacogenetics implementation consortium (CPIC) guideline for CYP2C9 and HLA-B genotypes and phenytoin dosing: 2020 update. Clin Pharmacol Ther 109(2):302–309. https://doi.org/10. 1002/cpt.2008 28. Saito Y, Stamp LK, Caudle KE et al (2016) Clinical pharmacogenetics implementation consortium (CPIC) guidelines for human leukocyte antigen B (HLA-B) genotype and allopurinol dosing: 2015 update. Clin Pharmacol Ther 99(1):36–37. https://doi.org/10. 1002/cpt.161 29. Padovan E, Bauer T, Tongio MM et al (1997) Penicilloyl peptides are recognized as T cell antigenic determinants in penicillin allergy. Eur J Immunol 27(6):1303–1307. https:// doi.org/10.1002/eji.1830270602

41

30. Wei CY, Chung WH, Huang HW et al (2012) Direct interaction between HLA-B and carbamazepine activates T cells in patients with Stevens-Johnson Syndrome. J Allergy Clin Immunol 129(6):1562–1569 e1565. https://doi.org/10.1016/j.jaci.2011. 12.990 31. Illing PT, Vivian JP, Purcell AW et al (2013) Human leukocyte antigen-associated drug hypersensitivity. Curr Opin Immunol 25(1): 81–89. https://doi.org/10.1016/j.coi. 2012.10.002 32. Norcross MA, Luo S, Lu L et al (2012) Abacavir induces loading of novel self-peptides into HLA-B*57: 01: an autoimmune model for HLA-associated drug hypersensitivity. AIDS 26(11):F21–F29. https://doi.org/10. 1097/QAD.0b013e328355fe8f 33. Ostrov DA, Grant BJ, Pompeu YA et al (2012) Drug hypersensitivity caused by alteration of the MHC-presented self-peptide repertoire. Proc Natl Acad Sci U S A 109(25): 9959–9964. https://doi.org/10.1073/pnas. 1207934109 34. Watkins S, Pichler WJ (2013) Sulfamethoxazole induces a switch mechanism in T cell receptors containing TCRVbeta20-1, altering pHLA recognition. PLoS One 8(10):e76211. https://doi.org/10.1371/journal.pone. 0076211 35. Russo MW, Galanko JA, Shrestha R et al (2004) Liver transplantation for acute liver failure from drug induced liver injury in the United States. Liver Transpl 10(8): 1018–1023. https://doi.org/10.1002/lt. 20204 36. Wu DY, Ugozzoli L, Pal BK et al (1989) Allele-specific enzymatic amplification of beta-globin genomic DNA for diagnosis of sickle cell anemia. Proc Natl Acad Sci U S A 86(8):2757–2760. https://doi.org/10. 1073/pnas.86.8.2757 37. Newton CR, Graham A, Heptinstall LE et al (1989) Analysis of any point mutation in DNA. The amplification refractory mutation system (ARMS). Nucleic Acids Res 17(7): 2503–2516. https://doi.org/10.1093/nar/ 17.7.2503 38. Olerup O, Zetterquist H (1992) HLA-DR typing by pcr amplification with sequencespecific primers (PCR-SSP) in 2 hours: an alternative to serological DR typing in clinical practice including donor-recipient matching in cadaveric transplantation. Tissue Antigens 39(5):225–235. https://doi.org/10.1111/j. 1399-0039.1992.tb01940.x 39. Mickelson E, Smith A, McKinney S et al (1993) A comparative study of HLA-DRB1

42

Ann M. Moyer and Manish J. Gandhi

typing by standard serology and hybridization of non-radioactive sequence-specific oligonucleotide probes to pcr-amplified DNA. Tissue Antigens 41(2):86–93. https://doi.org/10. 1111/j.1399-0039.1993.tb01984.x 40. Wordsworth BP, Allsopp CE, Young RP et al (1990) HLA-DR typing using DNA amplification by the polymerase chain reaction and sequential hybridization to sequence-specific oligonucleotide probes. Immunogenetics 32(6):413–418. https://doi.org/10.1007/ BF00241635 41. Erlich H, Bugawan T, Begovich AB et al (1991) HLA-DR, DQ and DP typing using PCR amplification and immobilized probes. Eur J Immunogenet 18(1–2):33–55. https://doi.org/10.1111/j.1744-313x. 1991.tb00005.x 42. Bugawan TL, Apple R, Erlich HA (1994) A method for typing polymorphism at the HLA-A locus using PCR amplification and immobilized oligonucleotide probes. Tissue Antigens 44(3):137–147. https://doi.org/ 10.1111/j.1399-0039.1994.tb02371.x 43. Gandhi MJ, Ferriola D, Huang Y et al (2017) Targeted next-generation sequencing for human leukocyte antigen typing in a clinical laboratory: metrics of relevance and considerations for its successful implementation. Arch Pathol Lab Med 141(6):806–812. https://doi.org/10.5858/arpa.20160537-RA 44. Walsh PS, Erlich HA, Higuchi R (1992) Preferential PCR amplification of alleles: mechanisms and solutions. PCR Methods Appl 1(4): 241–250. https://doi.org/10.1101/gr.1. 4.241 45. Moyer AM, Dukek B, Duellman P et al (2020) Concordance between predicted HLA type using next generation sequencing data generated for non-HLA purposes and clinical HLA type. Hum Immunol 81(8): 423–429. https://doi.org/10.1016/j. humimm.2020.06.002 46. Fan WL, Shiao MS, Hui RC et al (2017) HLA association with drug-induced adverse reactions. J Immunol Res 2017:3186328. https://doi.org/10.1155/2017/3186328 47. Kostenko L, Kjer-Nielsen L, Nicholson I et al (2011) Rapid screening for the detection of HLA-B57 and HLA-B58 in prevention of drug hypersensitivity. Tissue Antigens 78(1): 1 1 – 2 0 . h t t p s : // d o i . o r g / 1 0 . 1 1 1 1 / j . 1399-0039.2011.01649.x 48. Martin AM, Krueger R, Almeida CA et al (2006) A sensitive and rapid alternative to HLA typing as a genetic screening test for abacavir hypersensitivity syndrome.

Pharmacogenet Genomics 16(5):353–357. h t t p s : // d o i . o r g / 1 0 . 1 0 9 7 / 0 1 . f p c . 0000197468.16126.cd 49. Erlichster M, Goudey B, Skafidas E et al (2019) Cross-ethnicity tagging SNPs for HLA alleles associated with adverse drug reaction. Pharmacogenomics J 19(3):230–239. https://doi.org/10.1038/s41397-0180039-z 50. He Y, Hoskins JM, Clark S et al (2015) Accuracy of SNPs to predict risk of HLA alleles associated with drug-induced hypersensitivity events across racial groups. Pharmacogenomics 16(8):817–824. https://doi.org/10. 2217/pgs.15.41 51. Zhu GD, Brenton AA, Malhotra A et al (2015) Genotypes at rs2844682 and rs3909184 have no clinical value in identifying HLA-B*15:02 carriers. Eur J Clin Pharmacol 71(8):1021–1023. https://doi.org/ 10.1007/s00228-015-1879-y 52. Saksit N, Nakkam N, Konyoung P et al (2017) Comparison between the HLA-B(*) 58 : 01 allele and single-nucleotide polymorphisms in chromosome 6 for prediction of allopurinol-induced severe cutaneous adverse reactions. J Immunol Res 2017: 2738784. https://doi.org/10.1155/2017/ 2738784 53. Meyer D, Nunes K (2017) HLA imputation, what is it good for? Hum Immunol 78(3): 239–241. https://doi.org/10.1016/j. humimm.2017.02.007 54. Zheng X, Shen J, Cox C et al (2014) HIBAG--HLA genotype imputation with attribute bagging. Pharmacogenomics J 14(2):192–200. https://doi.org/10.1038/ tpj.2013.18 55. Marsh SG, Albert ED, Bodmer WF et al (2010) Nomenclature for factors of the HLA system, 2010. Tissue Antigens 75(4): 291–455. https://doi.org/10.1111/j. 1399-0039.2010.01466.x 56. Robinson J, Malik A, Parham P et al (2000) IMGT/HLA database--a sequence database for the human major histocompatibility complex. Tissue Antigens 55(3):280–287. https://doi.org/10.1034/j.1399-0039. 2000.550314.x 57. Mallal S, Phillips E, Carosi G et al (2008) HLA-B*5701 screening for hypersensitivity to abacavir. N Engl J Med 358(6):568–579. https://doi.org/10.1056/NEJMoa0706135 58. Hung SI, Chung WH, Liou LB et al (2005) HLA-B*5801 allele as a genetic marker for severe cutaneous adverse reactions caused by allopurinol. Proc Natl Acad Sci U S A

HLA Testing in Pharmacogenomics 102(11):4134–4139. https://doi.org/10. 1073/pnas.0409500102 59. Kaniwa N, Saito Y, Aihara M et al (2008) HLA-B locus in Japanese patients with antiepileptics and allopurinol-related StevensJohnson Syndrome and toxic epidermal necrolysis. Pharmacogenomics 9(11): 1617–1622. https://doi.org/10.2217/ 14622416.9.11.1617 60. Cristallo AF, Schroeder J, Citterio A et al (2011) A study of HLA class I and class II 4-digit allele level in Stevens-Johnson syndrome and toxic epidermal necrolysis. Int J Immunogenet 38(4):303–309. https://doi. org/10.1111/j.1744-313X.2011.01011.x 61. Kang HR, Jee YK, Kim YS et al (2011) Positive and negative associations of HLA class I alleles with allopurinol-induced SCARs in Koreans. Pharmacogenet Genomics 21(5): 303–307. https://doi.org/10.1097/FPC. 0b013e32834282b8 62. Romano A, De Santis A, Romito A et al (1998) Delayed hypersensitivity to aminopenicillins is related to major histocompatibility complex genes. Ann Allergy Asthma Immunol 80(5):433–437. https://doi.org/10. 1016/s1081-1206(10)62997-3 63. O’Donohue J, Oien KA, Donaldson P et al (2000) Co-amoxiclav jaundice: clinical and histological features and HLA class II association. Gut 47(5):717–720. https://doi.org/ 10.1136/gut.47.5.717 64. Donaldson PT, Daly AK, Henderson J et al (2010) Human leucocyte antigen class II genotype in susceptibility and resistance to co-amoxiclav-induced liver injury. J Hepatol 53(6):1049–1053. https://doi.org/10. 1016/j.jhep.2010.05.033 65. Lucena MI, Molokhia M, Shen Y et al (2011) Susceptibility to amoxicillin-clavulanateinduced liver injury is influenced by multiple HLA class I and II alleles. Gastroenterology 141(1):338–347. https://doi.org/10.1053/ j.gastro.2011.04.001 66. Stephens C, Lopez-Nevot MA, Ruiz-Cabello F et al (2013) HLA alleles influence the clinical signature of amoxicillin-clavulanate hepatotoxicity. PLoS One 8(7):e68111. https:// doi.org/10.1371/journal.pone.0068111 67. Hautekeete ML, Horsmans Y, Van Waeyenberge C et al (1999) HLA association of amoxicillin-clavulanate--induced hepatitis. Gastroenterology 117(5):1181–1186. https://doi.org/10.1016/s0016-5085(99) 70404-x 68. Chen WT, Chi CC (2019) Associations of HLA genotypes with antithyroid drug-

43

induced agranulocytosis: a systematic review and meta-analysis of pharmacogenomics studies. Br J Clin Pharmacol 85(9):1878–1887. https://doi.org/10.1111/bcp.13989 69. Kim SH, Choi JH, Lee KW et al (2005) The human leucocyte antigen-DRB1*1302DQB1*0609-DPB1*0201 haplotype may be a strong genetic marker for aspirin-induced urticaria. Clin Exp Allergy 35(3):339–344. https://doi.org/10.1111/j.1365-2222. 2004.02197.x 70. Man CB, Kwan P, Baum L et al (2007) Association between HLA-B*1502 allele and antiepileptic drug-induced cutaneous reactions in Han Chinese. Epilepsia 48(5):1015–1018. https://doi.org/10.1111/j.1528-1167. 2007.01022.x 71. Mockenhaupt M, Wang CW, Hung SI et al (2019) HLA-B*57:01 confers genetic susceptibility to carbamazepine-induced SJS/TEN in Europeans. Allergy 74(11):2227–2230. https://doi.org/10.1111/all.13821 72. Li LJ, Hu FY, Wu XT et al (2013) Predictive markers for carbamazepine and lamotrigineinduced maculopapular exanthema in Han Chinese. Epilepsy Res 106(1–2):296–300. https://doi.org/10.1016/j.eplepsyres.2013. 05.004 73. Hung SI, Chung WH, Jee SH et al (2006) Genetic susceptibility to carbamazepineinduced cutaneous adverse drug reactions. Pharmacogenet Genomics 16(4):297–306. h t t p s : // d o i . o r g / 1 0 . 1 0 9 7 / 0 1 . f p c . 0000199500.46842.4a 74. Kashiwagi M, Aihara M, Takahashi Y et al (2008) Human leukocyte antigen genotypes in carbamazepine-induced severe cutaneous adverse drug response in Japanese patients. J Dermatol 35(10):683–685. https://doi.org/ 10.1111/j.1346-8138.2008.00548.x 75. Hsiao YH, Hui RC, Wu T et al (2014) Genotype-phenotype association between HLA and carbamazepine-induced hypersensitivity reactions: strength and clinical correlations. J Dermatol Sci 73(2):101–109. https://doi.org/10.1016/j.jdermsci.2013. 10.003 76. Legge SE, Walters JT (2019) Genetics of clozapine-associated neutropenia: recent advances, challenges and future perspective. Pharmacogenomics 20(4):279–290. https:// doi.org/10.2217/pgs-2018-0188 77. Zhang FR, Liu H, Irwanto A et al (2013) HLA-B*13:01 and the dapsone hypersensitivity syndrome. N Engl J Med 369(17): 1620–1628. https://doi.org/10.1056/ NEJMoa1213096

44

Ann M. Moyer and Manish J. Gandhi

78. Pellicano R, Lomuto M, Ciavarella G et al (1997) Fixed drug eruptions with feprazone are linked to HLA-B22. J Am Acad Dermatol 36(5 Pt 1):782–784. https://doi.org/10. 1016/s0190-9622(97)80347-7 79. Nicoletti P, Aithal GP, Bjornsson ES et al (2017) Association of liver injury from specific drugs, or groups of drugs, with polymorphisms in HLA and other genes in a genome-wide association study. Gastroenterology 152(5):1078–1089. https://doi.org/ 10.1053/j.gastro.2016.12.016 80. Daly AK, Donaldson PT, Bhatnagar P et al (2009) HLA-B*5701 genotype is a major determinant of drug-induced liver injury due to flucloxacillin. Nat Genet 41(7):816–819. https://doi.org/10.1038/ng.379 81. Nicoletti P, Aithal GP, Chamberlain TC et al (2019) Drug-induced liver injury due to flucloxacillin: relevance of multiple human leukocyte antigen alleles. Clin Pharmacol Ther 106(1):245–253. https://doi.org/10.1002/ cpt.1375 82. Nicoletti P, Werk AN, Sawle A et al (2016) HLA-DRB1*16: 01-DQB1*05: 02 is a novel genetic risk factor for flupirtine-induced liver injury. Pharmacogenet Genomics 26(5): 218–224. https://doi.org/10.1097/FPC. 0000000000000209 83. Batchelor JR, Welsh KI, Tinoco RM et al (1980) Hydralazine-induced systemic lupus erythematosus: influence of HLA-DR and sex on susceptibility. Lancet 1(8178): 1107–1109. https://doi.org/10.1016/ s0140-6736(80)91554-8 84. Bruno CD, Fremd B, Church RJ et al (2020) HLA associations with infliximab-induced liver injury. Pharmacogenomics J 20(5): 681–686. https://doi.org/10.1038/ s41397-020-0159-0 85. Kazeem GR, Cox C, Aponte J et al (2009) High-resolution HLA genotyping and severe cutaneous adverse reactions in lamotriginetreated patients. Pharmacogenet Genomics 19(9):661–665. https://doi.org/10.1097/ FPC.0b013e32832c347d 86. Hung SI, Chung WH, Liu ZS et al (2010) Common risk allele in aromatic antiepilepticdrug induced Stevens-Johnson syndrome and toxic epidermal necrolysis in Han Chinese. Pharmacogenomics 11(3):349–356. https:// doi.org/10.2217/pgs.09.162 87. Shi YW, Min FL, Liu XR et al (2011) HLA-B alleles and lamotrigine-induced cutaneous adverse drug reactions in the Han Chinese population. Basic Clin Pharmacol Toxicol 109(1):42–46. https://doi.org/10.1111/j. 1742-7843.2011.00681.x

88. Spraggs CF, Parham LR, Hunt CM et al (2012) Lapatinib-induced liver injury characterized by class II HLA and Gilbert’s syndrome genotypes. Clin Pharmacol Ther 91(4):647–652. https://doi.org/10.1038/ clpt.2011.277 89. Spraggs CF, Budde LR, Briley LP et al (2011) HLA-DQA1*02:01 is a major risk factor for lapatinib-induced hepatotoxicity in women with advanced breast cancer. J Clin Oncol 29(6):667–673. https://doi.org/10.1200/ JCO.2010.31.3197 90. Schaid DJ, Spraggs CF, McDonnell SK et al (2014) Prospective validation of HLA-DRB1*07:01 allele carriage as a predictive risk factor for lapatinib-induced liver injury. J Clin Oncol 32(22):2296–2303. https://doi.org/10.1200/JCO.2013.52. 9867 91. Schmidt KL, Mueller-Eckhardt C (1977) Agranulocytosis, levamisole, and HLA-B27. Lancet 2(8028):85. https://doi.org/10. 1016/s0140-6736(77)90082-4 92. Singer JB, Lewitzky S, Leroy E et al (2010) A genome-wide study identifies HLA alleles associated with lumiracoxib-related liver injury. Nat Genet 42(8):711–714. https:// doi.org/10.1038/ng.632 93. Li X, Jin S, Fan Y et al (2019) Association of HLA-C*03:02 with methimazole-induced liver injury in Graves’ disease patients. Biomed Pharmacother 117:109095. https:// doi.org/10.1016/j.biopha.2019.109095 94. Kim SH, Kim M, Lee KW et al (2010) HLA-B*5901 is strongly associated with methazolamide-induced Stevens-Johnson syndrome/toxic epidermal necrolysis. Pharmacogenomics 11(6):879–884. https://doi. org/10.2217/pgs.10.54 95. Urban TJ, Nicoletti P, Chalasani N et al (2017) Minocycline hepatotoxicity: clinical characterization and identification of HLA-B*35:02 as a risk factor. J Hepatol 67(1):137–144. https://doi.org/10.1016/j. jhep.2017.03.010 96. Littera R, Carcassi C, Masala A et al (2006) HLA-dependent hypersensitivity to nevirapine in Sardinian HIV patients. AIDS 20(12):1621–1626. https://doi.org/10. 1097/01.aids.0000238408.82947.09 97. Martin AM, Nolan D, James I et al (2005) Predisposition to nevirapine hypersensitivity associated with HLA-DRB1*0101 and abrogated by low CD4 T-cell counts. AIDS 19(1): 9 7 – 9 9 . h t t p s : // d o i . o r g / 1 0 . 1 0 9 7 / 00002030-200501030-00014

HLA Testing in Pharmacogenomics 98. Yuan J, Guo S, Hall D et al (2011) Toxicogenomics of nevirapine-associated cutaneous and hepatic adverse events among populations of African, Asian, and European descent. AIDS 25(10):1271–1280. https://doi.org/ 10.1097/QAD.0b013e32834779df 99. Phillips E, Bartlett JA, Sanne I et al (2013) Associations between HLA-DRB1*0102, HLA-B*5801, and hepatotoxicity during initiation of nevirapine-containing regimens in South Africa. J Acquir Immune Defic Syndr 62(2):e55–e57. https://doi.org/10.1097/ QAI.0b013e31827ca50f 100. He N, Min FL, Shi YW et al (2012) Cutaneous reactions induced by oxcarbazepine in Southern Han Chinese: incidence, features, risk factors and relation to HLA-B alleles. Seizure 21(8):614–618. https://doi.org/10. 1016/j.seizure.2012.06.014 101. Lv YD, Min FL, Liao WP et al (2013) The association between oxcarbazepine-induced maculopapular eruption and HLA-B alleles in a Northern Han Chinese population. BMC Neurol 13:75. https://doi.org/10. 1186/1471-2377-13-75 102. Roujeau JC, Huynh TN, Bracq C et al (1987) Genetic susceptibility to toxic epidermal necrolysis. Arch Dermatol 123(9): 1171–1173 103. Xu CF, Johnson T, Wang X et al (2016) HLA-B*57:01 confers susceptibility to pazopanib-associated liver injury in patients with cancer. Clin Cancer Res 22(6): 1371–1377. https://doi.org/10.1158/ 1078-0432.CCR-15-2044 104. Yang F, Gu B, Zhang L et al (2014) HLA-B*13:01 is associated with

45

salazosulfapyridine-induced drug rash with eosinophilia and systemic symptoms in Chinese Han population. Pharmacogenomics 15(11):1461–1469. https://doi.org/10. 2217/pgs.14.69 105. Roujeau JC, Bracq C, Huyn NT et al (1986) HLA phenotypes and bullous cutaneous reactions to drugs. Tissue Antigens 28(4): 251–254. https://doi.org/10.1111/j. 1399-0039.1986.tb00491.x 106. Kongpan T, Mahasirimongkol S, Konyoung P et al (2015) Candidate HLA genes for prediction of co-trimoxazole-induced severe cutaneous reactions. Pharmacogenet Genomics 25(8):402–411. https://doi.org/10.1097/ FPC.0000000000000153 107. Lonjou C, Borot N, Sekula P et al (2008) A European study of HLA-B in StevensJohnson syndrome and toxic epidermal necrolysis related to five high-risk drugs. Pharmacogenet Genomics 18(2):99–107. h t t p s : // d o i . o r g / 1 0 . 1 0 9 7 / F P C . 0b013e3282f3ef9c 108. Hirata K, Takagi H, Yamamoto M et al (2008) Ticlopidine-induced hepatotoxicity is associated with specific human leukocyte antigen genomic subtypes in Japanese patients: a preliminary case-control study. Pharmacogenomics J 8(1):29–33. https://doi.org/10. 1038/sj.tpj.6500442 109. Kindmark A, Jawaid A, Harbron CG et al (2008) Genome-wide pharmacogenetic investigation of a hepatic adverse event without clinical signs of immunopathology suggests an underlying immune pathogenesis. Pharmacogenomics J 8(3):186–195. https://doi.org/10.1038/sj.tpj.6500458

Chapter 3 Pharmacogenomics in Targeted Therapy and Supportive Care Therapies for Cancer Zahra Talebi, Alex Sparreboom, and Susan I. Colace Summary Targeted therapies have significantly altered the landscape of available cancer therapies across all diagnoses and patient populations, and supportive care therapies have steadily improved throughout the years to make therapy more tolerable for patients. Even so, these therapies have varied efficacy and toxicity among patients with cancer, and pharmacogenomics presents an opportunity to identify which patients are most at risk of toxicities and most likely to benefit from them. While the field of pharmacogenomics in targeted cancer therapy is still growing, we review current knowledge, hypotheses, and clinical practices in this chapter, along with a brief review of pharmacogenomics in supportive therapies in cancer treatment. Key words Cancer, Pharmacogenomics, Targeted therapies, Chemotherapy, Toxicity, Efficacy

1

Introduction The discipline of pharmacogenomics describes differences in the pharmacokinetics and pharmacodynamics of drugs as a result of inherited variation in drug metabolizing enzymes, drug transporters, and drug targets between patients. These inherited differences are occasionally responsible for extensive interpatient variability in drug disposition (systemic exposure) or effects (normal tissue and tumor exposure). Severe toxicity might occur in the absence of typical metabolism of active compounds, while the therapeutic effect of a drug could be diminished in the case of absence of activation of a prodrug. The importance and detectability of polymorphisms for a given enzyme depend on the contribution of the variant gene product to pharmacological response, the availability of alternative pathways of metabolism, and the frequency of occurrence of the least common variant allele. Although many substrates have been identified for the known polymorphic drug metabolizing enzymes and transporters, the contribution of a genetically determined source of inter-individual pharmacokinetic variability has

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_3, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

47

48

Zahra Talebi et al.

been established for only very few cancer chemotherapeutic agents. Most of these cases involve agents for which elimination is critically dependent on a rate-limiting breakdown by a polymorphic enzyme (e.g., 6-mercaptopurine by thiopurine-S-methyltransferase; 5-flurouracil by dihydropyrimidine dehydrogenase; SN-38 by UDP-glucuronosyltransferase 1A1) or when a polymorphic enzyme is involved in the formation of a toxic metabolite (e.g., tamoxifen by CYP2D6) [1]. Genetically determined variation in ADME and therapeutic target genes is increasingly recognized to have a significant role as a determinant of inter-subject variability in response to common targeted therapeutics (Table 1). This chapter explores what is currently known, theorized, and unknown regarding the utility of pharmacogenetics in the rational use of targeted agents and adjunct therapies in clinical oncology practice.

2 2.1

Pharmacogenomics in Targeted Therapies Bevacizumab

Bevacizumab is an antiangiogenesis agent which is a monoclonal antibody against vascular endothelial growth factor (VEGF). Interindividual variability exists in response to and toxicity from bevacizumab. [2]. Common drug toxicities associated with bevacizumab are hypertension, hemorrhage, and proteinuria, with hypertension being the most common toxicity [3]. VEVGFA rs2010963 has been associated in some studies with increased risk of thrombo-hemorrhagic events and toxicity in general (risk allele C) as well as increased risk of grade >3 hypertension (risk allele G). However, data from these association studies are inconsistent, and underlying mechanisms for these increased risks are unclear [3]. One retrospective study examined breast cancer tumor VEGF genotypes, which had previously been shown to reflect the germline VEGF genotype, and found an increased risk of hypertension associated with bevacizumab and VEGF-634 CC (rs2010963) and VEGF-1498 TT, while superior overall survival was associated with genotypes VEGF-2578 AA (rs699947) and the VEGF-1154 A allele, with an additive effect [2]. Polymorphisms rs1799983 and rs2070744 in the nitric oxide synthase 3 (NOS3) gene have been associated with risk of grade >3 hypertension and proteinuria in patients receiving bevacizumab. These SNPs are related to nitric oxide production, which helps regulate vascular tone, and so has biologic plausibility in regard to mechanism of effect on the inter-individual variation in these toxicities [3]. A polymorphism in a gene related to autophagy, RB1-inducible coiled-coil 1 (RB1CC1) rs1129660, and germline polymorphisms (rs9381299 and rs834576) upstream of the NO signaling-related gene heat shock protein 90 alpha family class B member

Drug class

Xanthine oxidase inhibitor

HDAC inhibitor

Epidermal growth gactor receptor inhibitor; monoclonal antibody

BRAF inhibitor

Fibroblast growth factor receptor inhibitor

Antiandrogen

Drug

Allopurinol

Belinostat

Cetuximab

Dabrafenib

Erdafitinib

Flutamide

Y

Y Y Y

Y

Y

Y

N

HLA-B HLA-A HLA-C UGT1A1

EGF

G6PD

CYP2C9

G6PD

Gene

N

N

N

N

N N

Y

N/A N

B/C

B/C

D

B

B/C C

A

Y

Y

Y

N

Y

N N

Y

Recommendation/ guideline for patients with variants

Monitor closely for toxicity

Monitor closely or avoid drug

None

(continued)

Consider Increased risk of monitoring methemoglobinemia in methemoglobin patients with G6PD deficiency levels

Poor metabolizers may have increased systemic exposure and increased risk of toxicity

Increased risk of hemolytic anemia in patients with G6PD deficiency

Some findings of improved survival with certain variants; data mixed

Increased risk of adverse reaction Reduce starting dose in patients with homozygous *28

Increased risk of severe cutaneous Monitor closely or avoid drug adverse reaction; avoid use in variant carriers N/A None N/A None

CPICIn Changes to risks or drug identified CPIC CPIC FDA efficacy in patients with DGI level guideline label variants

Table 1 Drug-gene interactions (DGIs) with associated CPIC guidelines and FDA label recommendations for targeted cancer therapies and supportive care therapies

Pharmacogenomics in Targeted Cancer Therapy 49

BCR-ABL tyrosine kinase inhibitor

Nilotinib

Ondansetron Selective 5-HT3 receptor antagonist

Anti-HER2 tyrosine kinase inhibitor

Lapatinib

Y

Y

Y

Y Y

HLA-DRB1

UGT1A1

CYP2D6 ABCB1

N

HLA-DQA1

Immune checkpoint HLA-A inhibitor; antiCTLA4 monoclonal antibody

Ipilimumab

N

Y

N

N

N

C/D N

A

B/C

C

B/C

N/A N

B/C

N

N

Y

Y

Y

Y

Y

None

Monitor closely for toxicity

Recommendation/ guideline for patients with variants

Rapid metabolizers may have decreased response to drug Cer tain genotypes may af fect response to drug

Poor metabolizers may be at increased risk for adverse reaction such as hyperbilirubinemia

Select alternate drug None

Monitor closely for toxicity

Monitor liver Cer tain genotypes may be at function for all increased risk of hepatotoxicity genotypes Cer tain genotypes may be at Monitor liver increased risk of hepatotoxicity function for all genotypes

Specific genotype facilitates the immune presentation of the investigational peptide vaccine – one study enrolled only patients of this genotype

Poor metabolizers may have increased systemic exposure and increased risk of toxicity

Y

CYP2D6

Epidermal growth factor receptor inhibitor

Gefitinib

Gene

Drug class

CPICIn Changes to risks or drug identified CPIC CPIC FDA efficacy in patients with level guideline label variants DGI

Drug

Table 1 (continued)

50 Zahra Talebi et al.

Sacituzumab Anti-Trop2 monoclonal govitecan antibody; topoisomerase I inhibitor

PARP inhibitor

Rucaparib

Enzyme, urate-oxidase CYB5R (Recombinant)

Rasburicase

Anti-CD20 monoclonal antibody

Y

UGT1A1

Rituximab

Y

HLA-B

Tyrosine kinase inhibitor; vascular endothelial growth factor inhibitor

Pazopanib

Y

Y

N N N

G6PD

FCGR3A

CYP1A2 CYP2D6 UGT1A1

N

Y

CYP2D6

Palonesetron Selective 5-HT3 receptor antagonist

N

N

N

N

Y

N/A N

N/A N

N/A N

D

A

N/A N

B/C

B/C

C

Y

Y

Y

N

Y

Y

Y

Y

N

None

Poor metabolizers may be at increased risk for higher systemic concentrations and adverse reaction such as neutropenia

No clinically significant findings in variant carriers No clinically significant findings in variant carriers

Genotype may af fect response to drug

(continued)

Monitor closely

None

None

None

Patients with CYB5R deficiency Select alternate may be at increased risk of drug hemolysis and methemoglobinemia Patients with G6PD deficiency are Select alternate at increased risk of hemolysis drug and methemoglobinemia

Specific genotype may result in Monitor liver increased risk of hepatotoxicity function for all genotypes Poor metabolizers may be at Monitor closely increased risk for adverse reaction such as hyperbilirubinemia

Rapid metabolizers may have decreased response to drug

Pharmacogenomics in Targeted Cancer Therapy 51

Estrogen receptor antagonist

MEK inhibitor

Tamoxifen

Trametinib

N/A N A

F5 (Factor V Leiden) N Y

N

CYP2D6

G6PD N/A N

Y

N/A N

N

F2 (Prothrombin)

Gene

Y

Y

Y

Y

Patients with G6PD deficiency excluded from clinical trial, ef fect unknown

No clinical benefit to screening for mutations No clinical benefit to screening for mutations Intermediate and poor metabolizers may have lower active metabolite concentrations, but impact on efficacy not well established

CPICIn Changes to risks or drug identified CPIC CPIC FDA efficacy in patients with level guideline label variants DGI

CPIC level is based on level of evidence, with Level A carr ying the highest preponderance of evidence and Level D carr ying the least

Drug class

Drug

Table 1 (continued)

None

None

None

None

Recommendation/ guideline for patients with variants

52 Zahra Talebi et al.

Pharmacogenomics in Targeted Cancer Therapy

53

1 (HSP90AB1) have also been reported as associated with risk of bevacizumab-related hypertension. A polymorphism in synaptic vesicle glycoprotein 2C (SV2C), rs6453204, has also been associated with grade >3 hypertension related to bevacizumab [3]. An intronic SNP in KCNAB1 (rs6770663) has been identified as a potential pharmacogenetic marker for increased risk of severe hypertension with bevacizumab [4]. All of these SNPs require further study and better understanding of their potential mechanisms before they can be considered to be clinically useful. 2.2 EGFR-Targeted Tyrosine Kinase Inhibitors

EGFR tyrosine kinase inhibitors like gefitinib, erlotinib, afatinib, and osimertinib selectively inhibit EGFR tyrosine kinase and are used for cancer treatment in patients whose tumors overexpress EGFR or the EGFR receptor. EGFR is also expressed in non-malignant cells, and so EGFR TKIs can cause toxicities related to their effect on normal cells. Common toxicities include skin and gastrointestinal toxicities, while drug-induced interstitial lung disease is a severe but rare side effect in patients treated with EGFR TKIs [3]. While differences in the incidence of EGFR TKI-related drug reactions have been noted between ethnic groups, suggesting that genetic differences may play a role in this inter-individual variability, there have not been any conclusive findings that can be put to clinical use [3]. Patients with shorter CA repeats in intron-1 of EGFR, which is associated with mRNA expression and protein levels, have been shown to have improved progression-free survival, but this polymorphism has not been associated with toxicities. An Italian cohort study identified three different germline EGFR polymorphisms (
216 G>T,
191C>A, and R497K) associated with grade >2 diarrhea, but not with skin rash. Variants in genes commonly associated with inter-individual variability in drug response, such as the ABC transporter family and cytochrome P450 enzymes, have also been studied, but results remain preliminary and have not been replicated [3]. In a study of Japanese patients, ABCG2 rs2231137 was associated with skin rash, while ABCG2 rs2231142 and ABCB1 rs1045642 were not. In a Chinese population, CYP4F11 rs1064796 and UGT3A1 rs10045685 were associated with erlotinib-related skin and gastrointestinal toxicities [3]. Pharmacogenetic studies of EGFR TKI-related interstitial lung disease have been limited, though one small study in 13 Japanese patients demonstrated some potential associations with SNVs rs75399069, rs417168, rs442281, rs17690253, rs184448987, rs10165147, and rs1348851. Due to the size of the study, none of the associations were statistically significant [3].

54

Zahra Talebi et al.

More studies are needed before these associations can be considered useful in clinical practice. 2.3 BCR-ABL Targeted Tyrosine Kinase Inhibitors (TKIs)

BCR-ABL TKIs like imatinib, nilotinib, dasatinib, bosutinib, and ponatinib are highly efficacious, but inter-individual variability exists in response, with some patients developing resistance after initial response [5]. While point mutations in the target BCR-ABL gene are the main source of resistance to TKIs, pharmacokinetic changes based on the bioavailability of TKIs may also contribute [6]. Imatinib is a substrate of P-glycoprotein-mediated efflux, and P-glycoprotein is encoded by ABCB1, a multidrug resistance gene. SNPs which alter function of P-glycoprotein may influence the efficiency of absorption or elimination of imatinib [5]. ATP-binding cassette (ABC) transporters are membrane glycoproteins which play a key role in the efflux of TKIs; thus, changes in the expression of these transporters can affect the safety and efficacy of TKIs [6]. ABCB1 and ABCG2 play a prominent role in multidrug resistance, and excretion of a number of endogenous and exogenous compounds, including chemotherapeutics, from hematopoietic progenitor cells, as well as cancer stem cells. ABCB1 encodes the P-glycoprotein MDR1, and ABCG2 encodes the breast cancer resistance protein (BCRP). Overexpression of multidrug resistance proteins (MPR1 and 2), which are encoded by ABCC1 and ABCC2, has also been reported to play a role in the varied bioavailability of TKIs between patients [6]. ABCB1 is highly polymorphic, with three extensively studied SNPs: 1236C>T (rs1128503) and 3435C>T (rs104564), which are synonymous, and 22677G>T/A (rs2032582), a missense mutation. These three SNPs define a haplotype which has been associated with increased expression of MDR1. Variants in ABCG2 have been shown to reduce protein expression through an effect of the drug that interacts with the BCRP transporter. SNPs identified in ABCC2 (
24C>T, 1249G>A, 3972C>T) have been found to affect expression level and transport activity of MPR2, leading to difference in pharmacokinetics of various drug [6]. In a study of 90 CML patients treated with imatinib, three ABCB1 polymorphisms (1236C>T or rs1128503, 2677G>T/A or rs2032582, and 3435C>T or rs1045642) were examined for relationship to response to therapy, defined as achievement of major molecular response (MMR), as well as relationship to imatinib concentration. Patients with genotype 1236 TT were found to have higher imatinib concentrations, while patients with at least one G allele in 25677G>T/A had a worse response to therapy. Haplotypes were also examined for the same endpoints, and haplotype 1236C-2677G-3435C was associated with less frequent MMR [5].

Pharmacogenomics in Targeted Cancer Therapy

55

One study of 90 Caucasian adult CML patients treated with nilotinib examined the relationship between the achievement and/or loss of molecular response and seven SNPs in four ABC transporter genes: ABCC1 rs212090 (5463T>A); ABCC2 rs3740066 (3972C>T), rs4148386 (G>A), and rs1885301 (1549G>A); ABCG2 rs2231137 (34G>A) and rs2231142 (G>C); and ABCB1 rs1045642 (343C>T). Three genotypes were associated with higher probability of achieving MR3 response in a shorter timeframe (ABCC2 rs3740066 CC and CT, ABCB1 rs1045642 TT), and three genotypes were associated with a higher probability of achieving MR4 (ABCC2 rs3740066 CC, ABCB1 rs1045642 CC and TT). ABCG2 rs2231142 GG genotype was associated with a decreased risk of MR3 loss. In contrast, ABCG2 rs2231137 was associated with a lower probability of achieving MR3. The study performed a haplotype analysis to develop two models predictive of response using a combination of the genotypes studied. This analysis identified four genotypes predictive of probability of MR3 achievement (ABCC2 rs374006 CT and CC, ABCG2 rs2231137 GG, ABCB1 rs1045642 TT) and three genotype predictive of MR4 achievement (ABCC2 rs374006 CC, ABCB1 rs1045642 CC and TT) [6]. Hyperbilirubinemia has also been reported with UGT1A1 mutations in patients taking nilotinib [7]. HLA-B*57:01 has been implicated as a risk factor for hepatotoxicty related to other drugs, such as abacavir. Computational modeling of both drugs suggests that they may bind and interact with HLA-B*57:01 in a way that alters antigen recognition and triggers immune self-reactivity, resulting in damage to hepatocytes [8]. In a study with 517 and 79 patients treated with first-line imatinib and nilotinib, respectively, ASXL1 rs4911231 and BIM rs686952 variants were independent predictors of early molecular response (MR), major MR, deep MRs (MR4 and MR4.5), and failure-free survival (FFS) with imatinib treatment, but these variants did not consistently predict MR or FFS with nilotinib treatment. A model developed using the Sokal risk score, a predictive risk score using clinical features, combined with the ASXL1 and BIM variants identified an ultrahigh-risk group, representing 10% of patients, which predicted inferior overall survival, progression to acute phase/blast crisis, FFS, and MRs [9]. Further studies are required to confirm these observations and develop models with clinical utility [6]. 2.4 Immune Checkpoint Inhibitors

Immune checkpoint inhibitors have not been studied at great length for pharmacogenomic predictors of toxicity or efficacy. While one cohort identified the polymorphism rs2227981 in the gene that encodes PD-1, PDCD1, as a potential predictor of toxicity, this finding was not validated in other studies [3]. In a small

56

Zahra Talebi et al.

case-control study of 11 patients receiving checkpoint inhibitors (nivolumab, pembrolizumab, ipilumumab), patients with HLADR15, B52, and Cw12 were more likely to have pituitary-related toxicities, and in a different study, patients with HLA types that have been associated with development of type 1 diabetes mellitus in the general population (HLA DRB01*03 or *04, DR3-DQ2, DR4-DQ8) were more likely to develop immune checkpointinduced type 1 diabetes mellitus [3]. Trastuzumab

Trastuzumab is a humanized monoclonal antibody which binds the extracellular domain of human epidermal growth factor type 2 (HER2), preventing the activation of HER2 signaling and inducing antibody-dependent cellular cytotoxicity. It is used to treat HER2-positive cancers. Approximately 5% of patients treated with trastuzumab experience cardiotoxicity with a decline in left ventricular function [3]. Among European populations, germline mutations in ERBB2 (Ile655Val and Pro1170Ala) have been shown in some studies to predict cardiotoxicity related to trastuzumab. However, these associations have not been replicated in other populations, and data remains inconsistent even in European populations [3]. Other GWAS studies in European populations have identified a potential genetic risk factors for cardiotoxicity, including LDB2 rs55756123, BRINP1 rs10117876, RAB22A rs707557, TRPC6 rs77679196, LINC01060 rs7698718, and intergenic region 6p22.3 rs4305714. A rare variant associated with cardiotoxicity secondary to trastuzumab in the Japanese population is EYS rs139944387 [3]. In a Japanese cohort, a GWAS study identified five germline loci which can be used in a predictive scoring system for trastuzumab-related cardiotoxicity: chr13q14.3 rs9316695, chr15q26.3 rs28415722, chr17q25.3 rs7406710, chr4q25 rs11932853, and chr15q26.3 rs8032978 ( p ¼ 7.82  10
15) [10]. Despite these findings, further studies are required to develop models with clear clinical utility across diverse populations.

2.6 Vascular Endothelial Growth Factor (VEGF) Inhibitors

VEGF inhibitors such as pazopanib and sunitinib have significantly improved the treatment of diseases such as advanced renal cell carcinoma, but are not without significant side effects which can limit their use [11]. Pazopanib inhibits UGT1A1, which metabolizes bilirubin for elimination, and the UGT1A1 genetic variant TA7 (*28) causes reduced expression of UGT1A1, predisposing individuals to Gilbert’s syndrome, a benign syndrome of episodic hyperbilirubinemia. 28 polymorphisms in 11 genes involved in the pharmacokinetics and pharmacodynamics of pazopanib were studied in relationship to pazopanib-associated hyperbilirubinemia. UGT1A1 TA-repeat polymorphism was strongly associated with hyperbilirubinemia in patients taking pazopanib [12]. In another

2.5

Pharmacogenomics in Targeted Cancer Therapy

57

study, clinical and genetic data from 2190 patients treated with pazopanib was used to identify a 1.5–2 times increased risk of elevated ALT in patients with HLA-B*57:01 [8]. Sunitinib is a small molecule multi-kinase inhibitor which targets a range of tyrosine kinase receptors, including VEGFR1, VEGFR2, VEGFR3, as well as platelet-derived growth factors, Kit receptor, FLT3, and the receptor encoded by the ret protooncogene, RET. The most common toxicities related to sunitinib are liver injury, hypertension, diarrhea, mucositis, myelotoxicity, and hand-foot syndrome, all of which can result in treatment delays, dose changes, and discontinuation of therapy. Patients of Asian descent have been noted to have more toxicities related to sunitinib than European patients [3]. The germline polymorphism ABCG2 rs2231142 has been significantly associated with grade >3 thrombocytopenia in Japanese and Korean patients and grade >3 neutropenia and hand-foot syndrome in Korean patients [3]. A meta-analysis of eight studies examining associations between patients receiving sunitinib for renal cell carcinoma and sunitinib-related outcomes and toxicities concluded that the ABCG2 rs2231142 polymorphism may be useful as a predictor of thrombocytopenia related to sunitinib and hand-foot syndrome in Asians. ABCB1 rs1128503 was identified as a potential predictor of hypertension related to sunitinib and was also associated with PFS in renal cell carcinoma. ABCB1 rs2032585 was associated with PFS in the same population [13]. In a study of 333 patients with metastatic renal cell carcinoma from the USA, Spain, and the Netherlands, 22 SNPs and six haplotypes in ten candidate genes related to PK/PD of sunitinib were tested for association with sunitinib-related toxicity and efficacy. CYP3A5*1 was associated with dose reductions in this patient population, while the ABCB1 haplotypes containing CGT were associated with an increased PFS [14]. In European patients, two studies found associations between CYP3A4 rs4646437 and grade >3 hypertension as well any toxicity grade >3 [3]. Prospective validation studies are required to confirm clinical utility of these findings [14]. 2.7

Tamoxifen

Tamoxifen is a selective estrogen receptor modulator which has been used to treat women with estrogen receptor (ER)-positive breast cancer for over 40 years and has resulted in significant improvements in relapse rate and mortality for both pre- and postmenopausal breast cancer patients. CYP2D6 is the enzyme responsible for transforming tamoxifen to 4-hydroxytamoxifen and 4-hydroxy N-desmethyl tamoxifen (endoxifen), which have greater anti-estrogenic effect than tamoxifen. As a result, patients with CYP2D6 genetic polymorphisms and patients who take strong CYP2D6 inhibitors during therapy with tamoxifen demonstrate lower endoxifen concentrations and higher risk of disease recurrence [15].

58

Zahra Talebi et al.

In 908 pre- and postmenopausal ER-positive breast cancer patients treated with tamoxifen, 68–82% inter-patient variability of the metabolic ratio of endoxifen/desmethyl-TAM (E/DMT) was explained by CYP2D6 diplotype, while plasma endoxifen was predictable by 39–58% across all ethnicities [16]. CYP2D6 is highly polymorphic, and over 100 alleles have been reported across a geographically, racially, and ethnically diverse spectrum. The most common alleles based on function are no function alleles (*3, *4, *5, *6), decreased function alleles (*9, *10, *17, *41), and normal function alleles (*1, *2). Some clinical laboratories also report copy number variations, which are expressed as “xN,” where N represents the number of CYP2D6 gene copies. Alleles with two or more normal function gene copies are considered increased function alleles. Due to the potential for more than one gene copy on each allele, the CPIC guidelines are based on an activity score, in which a score of 0 represents a poor metabolizer, 0.5 represents intermediate metabolizers, 1.5 or 2.0 represents normal metabolizers, and >2.0 represents ultrarapid metabolizers. A score of 1.0 has less activity than 1.5 or 2.0 and may be classified as an intermediate metabolizer than some laboratories and others as a normal metabolizer [15]. Based on these scores, CPIC recommends starting with the standard dose for normal and ultrarapid metabolizers to and alternative hormonal therapy for intermediate and poor metabolizers [15].

3 Pharmacogenomics in Drugs Commonly Used to Manage Chemotherapy Side Effects 3.1

Allopurinol

Allopurinol inhibits the conversion of hypoxanthine and xanthine to uric acid by xanthine oxidase and is commonly used to prevent and treat the hyperuricemia component of tumor lysis syndrome in adults and children with leukemia, lymphoma, and solid tumors. Severe cutaneous adverse reactions (SCAR), such as drug hypersensitivity syndrome, Stevens-Johnson syndrome, and toxic epidermal necrolysis, are a rare but serious toxicity in patients treated with allopurinol. HLA genes, and particularly HLA-B, are highly polymorphic, with over 5000 known HLA-I and HLA-II alleles and over 1500 HLA-B alleles. The HLA-B58*01 allele has been associated with increased risk of SCAR in patients taking allopurinol, but the HLAB gene does not affect the pharmacokinetics or pharmacodynamics of allopurinol. The mechanism of this reaction is unclear, but the overarching hypothesis is that the interaction between the drug and the HLA-B58*01 allele results in an alteration in the HLA allele which causes T-cell activation. CPIC guidelines recommend that allopurinol be avoided in any patient who carries at least one HLAB58*01 allele.

Pharmacogenomics in Targeted Cancer Therapy

59

The HLA-B58*01 allele is more common in Asian vs European populations. However, the positive predictive value for HLAB58*01 as related to development of SCAR is ~1.5%, while the negative predictive value in Asian populations is 100%. Therefore, more studies regarding risk factors for development of SCAR related to allopurinol are warranted [17, 18]. 3.2

Rasburicase

3.3 5-HT3 Antagonists

4

Rasburicase is a recombinant urate oxidase enzyme used to break down uric acid for prophylaxis and treatment of hyperuricemia related to chemotherapy in adults and children with leukemia, lymphoma, and solid tumors. Patients with G6PD deficiency are at risk for development of acute hemolytic anemia when exposed to a number of drugs, including rasburicase. Screening for G6PD deficiency prior to rasburicase use is typically done through a quantitative enzyme assay, and CPIC guidelines offer a flow chart for interpretation of G6PD genotype and recommendations for use of rasburicase. In general, if a patient is determined to be G6PD deficient, use of rasburicase is contraindicated [19]. Ondansetron and tropisetron are 5-hydroxytryptamine type 3 (5-HT3) antagonists commonly used in the prevention and treatment of chemotherapy-induced nausea and vomiting, and variability in their efficacy has been linked to CYP2D6 genotype. CYP2D6, along with other CYP enzymes, metabolizes ondansetron and tropisetron to their inactive metabolites. Dolasetron, palonosetron, and ramosetron are also metabolized in part by CYP2D6, though there is limited evidence to guide use of these drugs based on CYP2D6 genotype. Granisetron is primarily metabolized by CYP3A4 and CYP1A1, and studies have not demonstrated a significant role for CYP2D6 in its metabolism. Patients who have a CYP2D6 ultra-metabolizer phenotype may experience decreased efficacy of ondansetron and granisetron, and so use of an alternate antiemetic agent not predominantly metabolized by CYP2D6 is recommended in the CPIC guidelines [20].

Conclusion The last two decades have provided evidence that targeted therapies can be curative in subsets of patients with advanced malignant diseases such as chronic myelocytic leukemia, and the expectation is that the list of cancers effectively treated and possibly even cured using modalities that include these agents will continue to expand in the near future. While targeted therapies have revolutionized the field of oncology, the contribution of inter-individual pharmacokinetic variability known to exist with many of these agents currently

60

Zahra Talebi et al.

in the clinic or in development to treatment outcome and the degree to which tumors may be intrinsically sensitive or resistant remains relatively poorly understood. It is anticipated that incorporation of pharmacogenomic concepts in the development and use of routine patient care will ultimately result in the discovery of improved treatment modalities for multiple malignant diseases. References 1. Talebi Z, Baker SD, Sparreboom A (2023) Pharmacology of small-molecule anticancer agents. In: Bast RC Jr, Byrd J, Croce CM (eds) Holland-Frei cancer medicine, 10th edn. Wiley Blackwell, USA 2. Schneider BP, Wang M, Radovich M et al (2008) Association of vascular endothelial growth factor and vascular endothelial growth factor receptor-2 genetic polymorphisms with outcome in a trial of paclitaxel compared with paclitaxel plus bevacizumab in advanced breast cancer: ECOG 2100. J Clin Oncol 26(28): 4672–4678 3. Udagawa C, Zembutsu H (2020) Pharmacogenetics for severe adverse drug reactions induced by molecular-targeted therapy. Cancer Sci 111(10):3445–3457 4. Diasio RB, Innocenti F, Offer SM (2021) Pharmacogenomic-guided therapy in colorectal cancer. Clin Pharmacol Ther 110(3): 616–625 5. Dulucq S, Bouchet S, Turcq B et al (2008) Multidrug resistance gene (MDR1) polymorphisms are associated with major molecular responses to standard-dose imatinib in chronic myeloid leukemia. Blood 112(5):2024–2027 6. Loscocco F, Visani G, Ruzzo A et al (2021) Clinical relevance of ABCB1, ABCG2, and ABCC2 gene polymorphisms in chronic myeloid leukemia patients treated with nilotinib. Front Oncol 11:672287 7. Tan Y, Ye Y, Zhou X (2020) Nilotinib-induced liver injury: a case report. Medicine (Baltimore) 99(36):e22061 8. Xu CF, Johnson T, Wang X et al (2016) HLA-B*57:01 confers susceptibility to pazopanib-associated liver injury in patients with cancer. Clin Cancer Res 22(6): 1371–1377 9. Marum JE, Yeung DT, Purins L (2017) ASXL1 and BIM germ line variants predict response and identify CML patients with the greatest risk of imatinib failure. Blood Adv 1(18): 1369–1381

10. Nakano MH, Udagawa C, Shimo A et al (2019) A genome-wide association study identifies five novel genetic markers for trastuzumab-induced cardiotoxicity in japanese population. Biol Pharm Bull 42(12): 2045–2053 11. Appleby L, Morrissey S, Bellmunt J et al (2011) Management of treatment-related toxicity with targeted therapies for renal cell carcinoma: evidence-based practice and best practices. Hematol Oncol Clin North Am 25(4):893–915 12. Xu CF, Reck BH, Xue Z et al (2010) Pazopanib-induced hyperbilirubinemia is associated with Gilbert’s syndrome UGT1A1 polymorphism. Br J Cancer 102(9):1371–1377 13. Sun F, Chen Z, Yao P et al (2021) Metaanalysis of ABCG2 and ABCB1 polymorphisms with sunitinib-induced toxicity and efficacy in renal cell carcinoma. Front Pharmacol 12: 641075 14. Diekstra MH, Swen JJ, Boven E et al (2015) CYP3A5 and ABCB1 polymorphisms as predictors for sunitinib outcome in metastatic renal cell carcinoma. Eur Urol 68(4):621–629 15. Goetz MP, Sangkuhl K, Guchelaar HJ et al (2018) Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for CYP2D6 and tamoxifen therapy. Clin Pharmacol Ther 103(5):770–777 16. Schroth W, Winter S, Murdter T et al (2017) Improved prediction of endoxifen metabolism by CYP2D6 genotype in breast cancer patients treated with tamoxifen. Front Pharmacol 8: 582 17. Hershfield MS, Callaghan JT, Tassaneeyakul W et al (2013) Clinical pharmacogenetics implementation consortium guidelines for human leukocyte antigen-B genotype and allopurinol dosing. Clin Pharmacol Ther 93(2):153–158 18. Saito Y, Stamp LK, Caudle KE et al (2016) Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines for human leukocyte antigen B (HLA-B) genotype and allo-

Pharmacogenomics in Targeted Cancer Therapy purinol dosing: 2015 update. Clin Pharmacol Ther 99(1):36–37 19. Relling MV, McDonagh EM, Chang T et al (2014) Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines for rasburicase therapy in the context of G6PD deficiency genotype. Clin Pharmacol Ther 96(2):169–174

61

20. Bell GC, Caudle KE, Whirl-Carrillo M et al (2017) Clinical Pharmacogenetics Implementation Consortium (CPIC) guideline for CYP2D6 genotype and use of ondansetron and tropisetron. Clin Pharmacol Ther 102(2): 213–218

Chapter 4 Pharmacogenomics in Cytotoxic Chemotherapy of Cancer Zahra Talebi, Alex Sparreboom, and Susan I. Colace Abstract Pharmacogenetic testing in patients with cancer requiring cytotoxic chemotherapy offers the potential to predict, prevent, and mitigate chemotherapy-related toxicities. While multiple drug-gene pairs have been identified and studied, few drug-gene pairs are currently used routinely in the clinical status. Here we review what is known, theorized, and unknown regarding the use of pharmacogenetic testing in cancer. Key words Cancer, Pharmacogenomics, Pharmacogenetics, Chemotherapy, Toxicity, Efficacy

1

Introduction Balancing chemotherapeutic efficacy with toxicity is an ongoing challenge facing oncologists and cancer patients, and the field of pharmacogenomics offers the hope of tools to better achieve that balance. A literature search in PubMed for “cancer pharmacogenomics” starting in 1990 yields 2290 results, with over 100 papers consistently published per year starting in 2012 and 189 published in 2020 alone. An increased ability to predict, prevent, and mitigate side effects through pharmacogenomics can improve adherence to and completion of life-saving chemotherapeutic regimens [1]. While many studies investigate somatic biomarkers, frequently used to identify targets for or to optimize therapy, less attention has been paid to germline biomarkers for efficacy and toxicity [1, 2]. And while many genetic mutations have been incorporated into FDA-approved drug labels, most of these are somatic tumor mutations, as opposed to germline mutations [2]. There are currently 112 drugs with “pharmacogenomic biomarkers” identified in FDA labels, comprising 207 drug-gene pairs. However, only 30 drug-gene pairs in 23 drugs refer to germline pharmacogenes which could potentially affect clinical efficacy, toxicity, drug choice, or drug dose. Of these 30 drug-gene pairs, the

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_4, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

63

64

Zahra Talebi et al.

following categories of recommendations are made in the FDA label based on genotype: 12 identify a need to watch more closely for toxicity, 6 recommend a dose adjustment to prevent toxicity, 3 recommend an alternate drug, 2 are simply mentions of clinical trial details related to pharmacogenomics, and 7 discuss hypothetical interactions based on PK/PD pathways, but ultimately are without clinical significance. Only 9 of the 30 drug-gene pairs have CPIC guidelines [3] (Table 1). Pharmacogenetic studies in cancer therapy are limited because such studies can be challenging to perform for anticancer agents; as it is unethical to give anticancer medications to those without cancer, so much of the research has been limited to cell or animal models derived from human tumor samples. In some studies, drugs in cell or animal models demonstrate less than or equal heritability of response as in control water treatment, suggesting that cytotoxicity is more likely to be due to environmental rather than genetic variability [2]. Even when studies can be completed in the context of a clinical trial with human subjects, conclusive evidence for drug choice and dose adjustment based on pharmacogenetics can be difficult to obtain, as studies are often heterogeneous in terms of genes, endpoints, and drugs. There are no significant studies in pharmacogenomics and long-term side effects [1]. Despite these challenges, the field of pharmacogenetics has made significant contributions to the care of patients with cancer and continues to work toward identification of drug-gene interactions which can be replicated and used in clinical practice. This chapter explores what is currently known, theorized, and unknown regarding the utility of pharmacogenetics in clinical oncology practice.

2 2.1

Pharmacogenomics in Cytotoxic Chemotherapy Azathioprines

The thiopurine drugs azathioprine, mercaptopurine, and thioguanine are key components of therapy for pediatric patients with acute lymphoblastic leukemia, inflammatory bowel disease, and autoimmune disorders. The principal cytotoxic effect of thiopurine drugs is the result of the production of 6-thioguanine nucleotides, thioguanine mono- and di-phosphates, which are converted to thioguanine triphosphates. Thioguanine triphosphates are incorporated into RNA, while thio-deoxyguanosine triphosphates are incorporated into DNA, resulting in cytotoxicity [4]. The metabolism of thiopurine drugs is highly complex, involving multiple competing enzymatic steps of the salvage purine pathway [4]. Azathioprine, mercaptopurine, and thioguanine are all inactive prodrugs that require intracellular activation by multiple enzymes. The first extracellular step of azathioprine activation involves conversion to mercaptopurine via metabolism by GSTM1, GSTA1, and GSTA2 [5].

Alkylating agent – Platinum analog

Alkylating agent – Platinum analog

Carboplatin

Cisplatin

N

N

Y

Y

Y

Y

TPMT COMT ERCC1 GSTM1 NQO1 XPC

N

N

N

Y

D

D

D

D

N

N

N

N

N/A N

D

C

D

Y Y

A

Y

MTHFR

Antimetabolite – DPYD Pyrimidine analog TYMS

Capecitabine

Gene

Drug class

Drug

N

N

N

N

N

Y

N

N

Y

None

Avoid drug

Patients with certain genotype may have increased risk of ototoxicity Children with certain genotype may have increased risk of hearing loss Increased risk of adverse reaction in patients with certain genotype Patients with certain genotype may have increased risk of ototoxicity Some findings of improved survival with certain variants; data mixed

None

None

None

None

None

None

(continued)

Recommendation/guideline for patients with variants

Some findings of improved None survival with certain variants; data mixed

Increased risk of adverse reaction in intermediate or poor metabolizers Potentially increased risk of adverse reaction

CPICIn Changes to risks or drug identified CPIC CPIC FDA efficacy in patients with Level guideline Label variants DGI

Table 1 Drug-gene interactions (DGIs) with associated CPIC guidelines and FDA label recommendations for cytotoxic cancer chemotherapies

Pharmacogenomics in Chemotherapy 65

Drug class

Doxorubicin

Daunorubicin

HAS3

SLC28A3 Anthracycline, topoisomerase II inhibitor CBR3

HAS3

D

D

Y

Y

B/C

D

Y

Y

D

D

Y

Y

SOD2

D

B/C

Y

NQO1

D

Y

Y

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

N

Recommendation/guideline for patients with variants

Patients with certain genotypes None may have increased risk of cardiotoxicity Certain alleles associated with None increased risk of cardiotoxicity over others Patients with certain genotypes None may have increased risk of cardiotoxicity

Patients with certain genotypes None may have increased risk of cardiotoxicity Certain alleles associated with None increased risk of cardiotoxicity over others Patients with certain genotypes None may have increased risk of cardiotoxicity

Patients with certain genotype None may have increased risk of nausea Some findings of improved None survival with certain variants; data mixed Some findings of improved None survival with certain variants; data mixed

Patients with certain genotype may have increased risk of ototoxicity

CPICIn Changes to risks or drug identified CPIC CPIC FDA efficacy in patients with DGI Level guideline Label variants

GSTP1

Gene

Anthracycline, SLC28A3 topoisomerase II inhibitor CBR3

Cyclophosphamide Alkylating agent

Drug

Table 1 (continued)

66 Zahra Talebi et al.

Fluorouracil

Etoposide

Epirubicin

Y

Y

Y Y

GSTP1 NQO1 TYMS UMPS

Topoisomerase II DYNC2H1 Y inhibitor Y

D

Y

NQO1

DPYD

D

Y

HAS3

Antimetabolite pyrimidine analog

D

Y

D

D

D

D

A

D

D

Y

CBR3 Anthracycline, topoisomerase II inhibitor GSTP1

D

Y

NQO1

N

N

N

N

Y

N

N

N

N

N

N

N

N

N

N

Y

N

N

N

N

N

N

None

None

None

None

(continued)

Increased risk of severe adverse Reduce dose by 50% in reaction in intermediate and intermediate metabolizers, poor metabolizers select alternate drug in poor metabolizers Some genotypes associated with None increased risk of toxicity and outcomes, data is mixed Some genotypes associated with None increased risk of toxicity and outcomes, data is mixed Some genotypes associated with None increased risk of toxicity Some genotypes with improved None response, data mixed

Patients with certain genotypes None may have increased risk of death. Data limited

Certain alleles associated with increased risk of cardiotoxicity over others Certain genotypes may be at increased risk of neutropenia, data limited Patients with certain genotypes may have increased risk of cardiotoxicity Some findings of worse outcome with certain variants

None Some findings of worse outcome with cer tain variants

Pharmacogenomics in Chemotherapy 67

Y

Y

Antimetabolite – NUDT15 Purine analog

TPMT

Y

SEMA3C

Mercaptopurine

Y

C8orf34

Topoisomerase I inhibitor

Irinotecan

A

A

D

D

A

D

Y

Y

D

Y

Y

Y

N

N

N

N

N

Y

Y

N

N

Y

N

N

Recommendation/guideline for patients with variants

Intermediate and poor metabolizers may have increased systemic exposure and increased risk for myelosuppression Intermediate and poor metabolizers may have increased systemic exposure and increased risk for myelosuppression

Reduce dose for intermediate and poor metabolizers

Reduce dose for intermediate and poor metabolizers

Reduce starting dose Poor metabolizers may have increased systemic exposure and increased risk of toxicity None Certain genotypes may be at increased risk of diarrhea, data limited None Certain genotypes may be at increased risk of neutropenia, data limited

Certain alleles associated with None increased risk of cardiotoxicity over others Patients with certain genotypes None may have increased risk of cardiotoxicity

CPICIn Changes to risks or drug identified CPIC CPIC FDA efficacy in patients with DGI Level guideline Label variants

UGT1A1

Anthracycline, CBR3 topoisomerase II inhibitor HAS3

Idarubicin

Gene

Drug class

Drug

Table 1 (continued)

68 Zahra Talebi et al.

Y

Y Y

MTRR

GSTM1 GSTP1

Y

A

A

D

D

D

D

C

C

C

Y

Y

N

N

N

N

N

N

N

Y

Y

N

N

N

N

N

N

N

Intermediate and poor metabolizers may have increased systemic exposure and increased risk for myelosuppression Intermediate and poor metabolizers may have increased systemic exposure and increased risk for myelosuppression

Reduce dose for intermediate and poor metabolizers

Reduce dose for intermediate and poor metabolizers

Certain genotypes may affect None response to drug Certain genotypes may increase None risk of toxicity

Certain genotypes have been None associated with increased risk of toxicity. Data is limited Certain genotypes have been None associated with response and toxicity. Data is limited

Certain genotypes associated None with increased risk of toxicity and efficacy, data mixed Some genotypes are associated None with increased exposure and risk of toxicity; data is mixed Some genotypes have been None associated with decreased drug clearance and increased risk of toxicity. Data is mixed

CPIC level is based on level of evidence, with Level A carrying the highest preponderance of evidence and Level D carrying the least

TPMT

Y

Y

ATIC

Antimetabolite – NUDT15 Purine analog

Y

SLCO1B1

Thioguanine

Y

MTHFR

Alkylating agent – Platinum analog

Y

Antimetabolite – ABCB1 Anti-folate

Oxaliplatin

Methotrexate

Pharmacogenomics in Chemotherapy 69

70

Zahra Talebi et al.

Once mercaptopurine or thioguanine is transported into the hepatocyte, there are multiple metabolic pathways. One important pathway involves thiopurine methyltransferase (TPMT), which methylates mercaptopurine and thioguanine compounds [6]. Another important component of thiopurine metabolism involves the NUDT15 enzyme, which converts the active 6-thioguanine nucleotide into inactive metabolites [7]. The complex and competing nature of the thiopurine metabolic pathway can make achieving the clinical balance between efficacy and toxicity challenging. To date, preemptive testing of two genes, TPMT and NUDT15, has been clinically implemented to achieve that balance. There is an inverse relationship between TPMT activity, a heritable trait, and the level of the cytotoxic 6-thioguanine nucleotide metabolites [8]. There are three wellcharacterized variants in TPMT that result in unstable TPMT protein and enhanced protein degradation. These three variants account for 90% of TPMT low-activity phenotypes and are present in 5% of white populations, 3% of Asian populations, and 6% of African American populations [6, 9, 10]. Patients who are TPMT poor metabolizers have a significantly increased risk of hematopoietic toxicity and cytopenia when compared to patients with normal TPMT activity, with dose reductions due to mercaptopurine toxicity required in 100% of poor metabolizers versus 35% of intermediate metabolizers and 7% of normal metabolizers [8]. Thus, the recommendation for TPMT poor metabolizers requiring this drug for treatment of malignancy is to drastically reduce the thiopurine dose (e.g., a tenfold dose reduction) and to consider alternative therapy for nonmalignant conditions [11]. TPMT intermediate metabolizers also have a significantly increased risk of hematopoietic toxicity and cytopenia when compared to patients with normal TPMT activity [8]. However, 40–70% of TPMT intermediate metabolizers tolerate full doses of thiopurine drugs, demonstrating the complex nature of thiopurine metabolism [11]. Therefore, the dosing recommendation for TPMT intermediate metabolizers is a more moderate dose reduction (e.g., 50–80% of the typical dose, depending on dose and indication) [11]. All active Children’s Oncology Group protocols for acute lymphoblastic leukemia currently recommend testing for TPMT variants at diagnosis and adjusting initial doses of thiopurine drugs accordingly. Clinical decision support for TPMT-based thiopurine dosing is available via the PharmGKB website (https://www.pharmgkb.org/ ) for 34 different TPMT star alleles and their combinations, based on CPIC guidelines. Some institutions have also incorporated electronic health record-based decision support into their practice, which automatically advises providers ordering thiopurine drugs for patients with TPMT variants that their patient may require a dose adjustment based on their genotype or phenotype and provides links to current guidelines, such as CPIC [11].

Pharmacogenomics in Chemotherapy

71

NUDT15, which inactivates 6-thioguanine nucleotides, was initially recognized as a clinically important enzyme in the thiopurine pathway via genome-wide association study of patients treated with thiopurine drugs for acute lymphocytic leukemia and inflammatory bowel disease [7, 12]. The NUDT15 variant rs116855232 was associated with thiopurine-related hematopoietic toxicity in several studies of mercaptopurine, and further studies identified a similar toxicity profile for azathioprine and thioguanine. Additional variants have been identified, with varying degrees of effect on NUDT15 activity [11]. The frequency of poor metabolizer phenotypes ranges from 1% to 10%, with higher frequencies reported among Asian and Hispanic populations [13, 14]. NUDT15 poor metabolizers who require thiopurines for treatment of malignancy are recommended to start at a significantly decreased dose, and those being treated for nonmalignant conditions are recommended to use an alternate therapy. For NUDT15 intermediate metabolizers and those with variants with uncertain functional activity, reduced dosing is recommended, with careful dose adjustment after 2–4 weeks based on toxicity and response [11]. CPIC guidelines also include guidance for patients for whom both TPMT and NUDT15 genotypes are known. All recommended dose reductions are based on a standard mercaptopurine starting dose of 75 mg/m2/day; lower starting doses may not require a dose reduction, particularly in intermediate metabolizers [11]. While preemptive testing of both TPMT and NUDT15 prior to thiopurine drug exposure can mitigate some of the toxicity associated with these drugs, thiopurine metabolism is complex and involves multiple additional enzymes of potential significance (e.g., XO, ITPA, MTHFR, IMPDH1, and IMPDH2) [15– 17]. Therefore, it is crucial that providers initiating thiopurine therapy continue to monitor patients for toxicity rather than interpreting normal metabolizer status for TPMT or NUDT15 as a guarantee against encountering significant toxicity. 2.2

Anthracyclines

Anthracyclines disrupt DNA and RNA synthesis by intercalating in DNA, inhibiting the topoisomerase II enzyme, which is highly expressed in rapidly proliferating cells, thereby preventing DNA replication and inducing apoptosis preventing the replication of rapidly dividing cells [18, 19]. Anthracyclines form the backbone of therapy for many solid tumors and leukemia in both pediatric and adult diseases and can be used in combination with many other chemotherapeutic agents [1]. Anthracyclines are highly effective, but limited in use by their toxicities, including gastrointestinal, myelosuppressive, and cardiotoxicity [1]. Anthracycline-induced cardiotoxicity (AIC) can manifest as asymptomatic cardiac dysfunction in as many as 57% of patients and restrictive or dilated cardiomyopathy with congestive heart failure (CHF) in up to 16–26% of pediatric and adult patients

72

Zahra Talebi et al.

[18, 19]. CHF in this patient population can have a mortality rate of up to 79%. Most patients develop cardiotoxicity more than a year after therapy, though a small percentage of patients experience cardiotoxicity within the first year [18]. Various studies report a cumulative incidence of cardiac disease (heart failure, valvular disease, coronary artery disease, arrhythmia) in anywhere from 5 to 10% of childhood cancer survivors by age 45 [20]. There is high inter-individual variability in AIC despite efforts to mitigate it including continuous infusion, liposomal formulations, and protective medications such as dexrazoxane, which may contribute to secondary malignancies [19, 20]. Non-pharmacogenomic risk factors for AIC in both adult and pediatric cancer survivors include age, hypertension, radiation exposure near the heart, cumulative anthracycline dose, and combination therapy with other cardiotoxic agents [19]. Compared to adult cancer survivors, there are also more traditional risk factors for cardiac disease in childhood cancer survivors – therefore, early detection and screening are important for mitigating risk [20]. There are multiple theories regarding the pathophysiology of AIC, including reactive oxygen species formation during treatment leading to lipid peroxidation and DNA damage of cardiomyocytes, accumulation of cardiotoxic anthracycline metabolites in the heart, disruption of calcium homeostasis, mitochondrial damage, and induction of apoptosis [18, 19]. Mitoxantrone has traditionally been classified as an anthracycline, but the mechanism by which it causes cardiotoxicity is different. One rat study demonstrated that mitoxantrone-induced cardiotoxicity may be a result of interference with cardiac energetic metabolism instead of oxidative stress. Mitoxantrone disrupts cardiac mitochondrial function which depletes ATP, resulting in a dysfunctional heart and impaired cardiac function [21]. It is 10 times more cardiotoxic than other anthracyclines with a nonlinear dose-response relationship with heart failure risk [20]. Many genes have been implicated and studied as potential pharmacogenetic risk factors for AIC, but no single gene has been identified [20]. Currently, over 20 genes have been proposed to be associated with AIC, but many studies require review and replication [19]. The pharmacogenomic markers with the strongest evidence for association with AIC to date are RARG rs2229774 and UGT1A6*4 rs17863783, which confer risk, and SLC28A3 rs7853758 and rs885004, which are associated with a protective effect [18, 19]. RARG normally represses expression of TOP2B, which codes for topoisomerase II-beta (TOP2B). Increased expression of TOP2B has been associated with risk of AIC, thus lending biologic plausibility to the association between decreased or loss of function variants in RARG with increased risk of AIC, which has also been

Pharmacogenomics in Chemotherapy

73

supported by studies in mouse and human cell lines [19]. RARG rs2229774 is a non-synonymous coding variant which alters RARG function leading to decreased repression of TOP2B. It has been associated with AIC in children and replicated in European, African, East Asian, Hispanic, and Aboriginal Canadian patient populations [18]. UGT1A6 is a key player in the glucuronidation pathway, and variants resulting in decreased enzyme activity may be linked to reduced elimination and increased accumulation of anthracycline or toxic metabolites [18, 19]. UGT1A6*4 (rs17863783) is a variant that results in 30–100% decreased enzyme activity [18]. The SLC super family is involved in anthracycline transport, lending biological plausibility to these genes as pharmacogenomic markers for AIC. SLC28A3 has been identified as a drug transporter in cancer cell line studies, and variants in SLC28A3 may reduce transport of anthracyclines into cardiomyocytes [18, 19]. SLC28A3 rs7853758 is associated with altered SLC28A3 mRNA levels, which may suggest a functional effect leading to decreased anthracycline transport into cardiomyocytes, and subsequently a reduced risk of AIC [18]. Beyond these three genes with the strongest evidence, a multitude of other genes have been identified as potentially associated with risk of AIC, with a range of strength in supporting evidence. Five candidate gene studies and two GWAS studies in women with breast cancer looked for genes associated with cardiotoxicity. While all studies used different definitions of cardiotoxicity and different endpoints, the candidate gene studies identified five SNPs with statistically significant associations – rs246221 in ABCC1, rs1045642 in ABCB1, rs1056892 in CBR3, rs10838611 in ATG13, and UGT2B7–161. Two studies looked at a different SNP in CBR3 and found conflicting data regarding association. In the GWAS studies, intergenic rs28714259 and the electron transfer flavoprotein beta gene (EFTB) were found to be significantly associated with AIC, but these studies have not been replicated at this time [1]. A systematic review and meta-analysis through 2016 examined 28 studies which looked at 84 different genes and 147 SNPs; 87 SNPs were identified as significantly associated with AIC by at least one study. Most SNPs were in transporter genes in the ABC or SLC families. The most studied genes involving metabolism were from the AKR (aldo/keto reductase) and CBR (carbonyl reductase) families [22]. Three risk variants significantly increased the risk for AIC: ABCC2 rs8187710 (pooled odds ratio 2.20; 95% CI 1.36–3.54), CYBA rs4673 (1.55; 1.05–2.30), and RAC2 rs13058338 (1.79; 1.27–2.52). However, this meta-analysis only involved a handful of studies with inconsistencies in methodology, populations, and reported results [22]. The ABC family of genes, which includes ABCC1, ABCC2, ABCC5, and ABCB4, has been associated with AIC in multiple

74

Zahra Talebi et al.

studies, but findings have been inconsistent, and the quality and quantity of evidence is limited. The association of this gene family with AIC is biologically plausible given the role these transporters play in transporting many drugs, including anthracyclines, but more studies are warranted before considering this gene useful for clinical practice [18]. Functional studies of CBR3 suggest that variants may affect metabolism of doxorubicin into the cardiotoxic metabolite doxorubicinol, but clinical association studies have had inconsistent results [18]. RAC2, NCF4, CYPBA are part of the nicotinamide adenine dinucleotide phosphate (NADPH) multi-enzyme complex. RAC2 is a mitochondrial protein required in the electron transfer reaction of NADPH oxidase during the formation of reactive oxygen species (ROS). Alteration in RAC2 leads to mitochondrial dysfunction resulting in increased ROS production. RAC2 rs13058338 is an intronic variant which has been associated with increased risk of cardiotoxicity in four studies [22]. Polymorphisms in other genes in this complex may also result in altered NADPH oxidase activity and production of ROS, but overall, evidence for associations with these polymorphisms and AIC remains limited and conflicting [18]. Many other variants have been studied with some identified associations, but without replication or validation required to justify clinical use. GSTP1 variants have been associated with AIC in two small studies, but these have not been replicated [18]. A small pediatric study identified an intronic variant (rs10836235) in CAT as marginally associated with AIC, but a larger GWAS did not identify any associations [18]. SULT2B1 is involved in the sulfate conjugation of anthracyclines, and rs10426377 was identified as marginally significant in association with AIC, but more evidence required [18]. CYPOR/POR are involved in the cytochrome P450 system and the biotransformation of a variety of drugs; intronic variants have been reported in association with AIC in one small pediatric study, but these findings have not replicated [18]. HAS3 is involved in cardiac remodeling; the coding variant rs2232228 was reported to be associated with cardiomyopathy, particularly in patients with higher cumulative anthracycline dose, in a case-only cohort, but was not replicated in a GWAS [18]. HFE is the gene associated with hemochromatosis. Variants rs1799945 and rs1800562 have been associated with dose-dependent myocardial iron overload and AIC in a few studies, but no other studies have observed this association [18]. HNMT rs17583889 is associated with AIC in childhood cancer survivor cohorts and in case studies, but more evidence is needed [18]. NOS3 has an essential role in cardiac function, and variant rs1799983 was associated with doxorubicin toxicity in childhood ALL survivors, but this was not replicated in a GWAS which included other tumor types [18].

Pharmacogenomics in Chemotherapy

75

Genetic risk factors may help identify children and adults at highest risk of cardiotoxicity secondary to anthracyclines prior to treatment. Modifications to treatment for patients at most risk may include liposomal anthracyclines, with addition of the cardioprotective agent dexrazoxane, and longer infusion time for anthracyclines, all of which have demonstrated some clinical benefit in mitigating risk of cardiotoxicity. Patients with higher risk may also undergo more frequent or sensitive screening than is typically recommended for their constellation of risk factors (dose, exposure to chest radiation therapy, etc.). In contrast, for patients identified to be at low risk of cardiotoxicity, the burden of surveillance may be decreased [20]. Despite this potential for clinical utility, neither an AHA statement from 2018 nor an article summarizing American and European guidelines regarding cardio-oncology management of cardiotoxicity related to cancer therapy make any reference to pharmacogenomic risks related to AIC [19, 23]. While multiple risk prediction models have been developed which include demographic, clinical, and pharmacogenetic factors, none have been validated for clinical use at this time [20]. A recent case series demonstrated how pharmacogenomic testing using RARG, UGT1A6, and SLC28A3 variants can be used a priori to adjust treatment in patients with pediatric cancers in an effort to balance risk of cardiotoxicity with desired therapeutic outcome [24]. Further studies such as these may provide sufficient evidence for full clinical implementation of pharmacogenetic testing to predict AIC. 2.3

Irinotecan

Irinotecan is a pro-drug which is converted to its active metabolite, SN-38, via hydrolysis. SN-38 is then inactivated via glucuronidation by UDP-glucuronosyltransferase (UGT1A1) [25]. Pharmacokinetic studies of irinotecan demonstrate large inter-individual variability in PK parameters of approximately 30% [26]. UGT1A1 has over 135 reported genetic variants. Many variant alleles result in decreased enzyme production, the most common of which are *28 (rs8175347) in Caucasian and African American populations, with a minor allele frequency (MAF) of 26–39% and 30–56%, respectively, and *6 (rs4148323) in Asian populations, with MAF of up to 47%. Decreased UGT1A1 activity can lead to increased serum concentration of SN-38 and thus increased risk of toxicity, such as neutropenia, hepatotoxicity, and diarrhea [25, 26]. Some studies have identified an increased dose-independent risk of neutropenia for patients both carrying and homozygous for *28, predominantly in Caucasian populations; the risk for severe diarrhea is dose dependent. Asians with at least one *6 allele also demonstrate an increased dose-independent risk of neutropenia and also an increased risk of severe diarrhea, though no analysis to determine dose dependence has been performed. There has been

76

Zahra Talebi et al.

no identified association with response based on genotype in patients with *28 or *6 alleles, but in a meta-analysis including eight studies, primarily phase 1 prospective dose finding studies, there were some lower case response rates in patients with irinotecan doses adjusted for UGT1A1 genotype [25, 26]. Other common UGT1A1 variant alleles could theoretically result in increased risk of toxicity. *60 (rs4124874;3279 T > G) is in linkage with *28, and the haplotype consisting of *28, *60, and *93 has been significantly associated with grade 4 neutropenia. *93 (rs10929302, 3156G > A) is in linkage disequilibrium with *28, results in reduced UGT1A1 expression, and has been clinically associated with increased bilirubin levels, increased SN-38 AUC, increased hematologic toxicities, diarrhea, and vomiting. The haplotype of *28, *60, *90, and *3 has been associated with increased response rate [26]. Polymorphisms in other uridine diphosphate glucuronosyltransferase (UGT) isoforms have also been associated with risk of toxicity with irinotecan. UGT1A9*22 demonstrates higher enzyme expression leading to higher SN-38 glucuronidation and increased risk for diarrhea. UGT1A7*3 and *4 demonstrate lower enzyme expression, and patients with genotype UGT1A7 *3/*3 have been shown to have an increased risk of toxicities with irinotecan. The haplotype of UGT1A1*28, *60, *93, UGT1A7*3, and UGT1A9*1 has been associated with severe neutropenia [26]. Irinotecan and SN-38 are substrates of ABC transporters, lending biologic plausibility to studies which have found associations between polymorphisms in the ABC family and irinotecan-related toxicities. A multivariate analysis including ABCC1 SNPs rs6498588 and rs17501331 found these SNPs to be associated with increased SN-38 plasma concentrations and decreased absolute neutrophil count (ANC), while ABCB1 rs12720066 has been associated with decreased SN-38 exposure and increased ANC. ABCB1 rs1045642 has been associated with increased risk for early toxicity and lower treatment response [26]. Some studies have associated ABCC2 polymorphisms with a protective effect on diarrhea, while others have associated them with toxicity in patients without UGT1A1*28, suggesting that this could function as a secondary screen for toxicity in patients with WT UGT1A1 [26, 27]. ABCG2 has not been associated with irinotecan exposure [26]. Polymorphisms in other genes involved in the metabolic pathway of irinotecan have been studied to determine whether they might also be useful in identification of patients at increased risk for toxicity secondary to irinotecan. SLCO1B1 encodes transporter OATP1B1, which is involved in the hepatic uptake of SN-38. SLCO1B1*1b is associated with increased ANC, likely due to increased hepatic uptake of SN-38 and reduced SN-38 concentration; this allele has also been associated with improved progression-

Pharmacogenomics in Chemotherapy

77

free survival. SCLO1B1*5 (rs4149056) results in reduced transporter activity and has been associated with increased SN-38 concentration and increased risk of neutropenia in conjunction with UGT1A1*28. Drug transporter polymorphisms may be useful predictors for toxicity, but no guidelines or definitive evidence for clinical utility currently exist [26]. CYP3A4 inactivates irinotecan through conversion to metabolite APC [27]. CYP3A4 polymorphisms have not been studied despite known drug-drug interactions involving CYP3A4 substrates [26]. While genotype-based dosing guidelines exist, and many national organizations recommend dose reduction for patients with homozygous variant risk alleles, upfront UGT1A1 genotyping is not routinely done in patients receiving irinotecan [26]. This is primarily due to the heterogeneity and conflicting evidence of currently available studies regarding risk of irinotecan toxicity based on UGT1A1 genotype, which do not provide clear guidance as to how to safely reduce the dose of irinotecan to mitigate toxicity without sacrificing drug efficacy and tumor response [25]. For example, when irinotecan is given at lower doses and greater frequency, there is no association between genotype and toxicity based on data in reviews and meta-analyses. The FDA added a recommendation of reduced initial dosage of irinotecan in patients homozygous for the UGT1A1*28 allele, but this recommendation does not provide guidance in how much to reduce the dose. The Dutch Pharmacogenomics Working Group (DPWG) recommends to reduce the starting dose of irinotecan by 30% of standard dose for patients with homozygous variant genotype, but some of the studies reviewed in one meta-analysis indicate that doses may need to be reduced further, as the difference in MTD for wild type vs homozygous variant in some studies was >30% [25]. A trial in the Netherlands performed prospective genotypeguided irinotecan dosing based on UGT1A1*28 and *93 genotype status. Recruitment for this trial was completed in February 2021, but as of September 2021, no data has been published [21]. More data from trials such as these are necessary before the true clinical utility of genotype-guided dosing is known for irinotecan. 2.4

Taxanes

The dose-limiting toxicity of many tubulin poisons, including paclitaxel, is a chronic, dose-dependent sensory peripheral neuropathy that is characterized by tingling, numbness, increased sensitivity to cold and touch, and burning pain of the distal extremities. The incidence of this side effect is particularly high in the case of paclitaxel, as it occurs in up to 70–80% in patients with breast cancer [28]. With continued dosing, the painful symptoms increase in severity and can persist for years, or even cause a lifelong functional

78

Zahra Talebi et al.

impairment that impacts quality of life [29, 30]. The mechanistic basis of this side effect remains uncertain, although prior studies have demonstrated that pharmacokinetic properties of paclitaxel may be statistically significantly correlated with markers of paclitaxel-induced peripheral neurotoxicity [31, 32]. These prior findings have triggered multiple investigations to evaluate the association of variants in ADME pathway genes as predictors of paclitaxel pharmacokinetics. The elimination of paclitaxel is predominantly mediated by hepatic metabolism, which is dependent on hepatic uptake by organic anion transporting polypeptides, followed by CYP2C8 and CYP3A-mediated biotransformation to inactive metabolites. The key transporters involved in the hepatocellular uptake of paclitaxel and docetaxel are OATP1B1 and OATP1B3 in humans and the single functional homologue OATP1B2 in mice and rats [33– 38]. These findings have been independently verified and are consistent with in vitro studies that have identified paclitaxel as a potent inhibitor of OATP1B1- and OATP1B3-mediated transport [39– 47]. Based on in vitro uptake studies, multiple functionally different haplotypes, including OATP1B1*5 and OATP1B1*15, were found to have a detrimental impact on the transport of taxanes in vitro [36]. This finding is consistent with previous studies showing substantially diminished transport activity of several OATP1B1 substrates by these particular variants when transfected into mammalian cells [48]. In vivo, these variants have been associated with altered systemic exposure and toxicity in response to multiple substrate drugs [49]. Interestingly, the relevance of these genetic variants in OATP1B1 was confirmed in a previous clinical study, but not in a prospectively conducted pharmacogeneticassociation study done in a group of predominantly white cancer patients [36, 50]. It is possible that additional rare genetic variants or haplotypes in OATP1B1 of importance to the transport of taxanes in the study population are yet to be discovered and that much larger numbers of patients are then needed to more precisely quantify genotype-phenotype associations. Similarly, several genetic variants in OATP1B3 were not significantly associated with the pharmacokinetics of taxanes in predominantly white patients [35, 36]. It should be pointed out that these findings are at odds with several other investigations performed in patients of Asian descent. For example, homozygosity (GG) for rs11045585 was associated with reduced clearance of docetaxel, compared with patients carrying the AA or AG genotypes [51]. In another study, a particular OATP1B3 genotype combination comprising the reference allele at IVS4 + 76G > A (rs4149118) and variant alleles at 699G > A (rs7311358), IVS12 + 5676A > G (rs11045585), and *347_*348insA (rs3834935) indel was also linked with reduced clearance of docetaxel [52]. It is possible that differences in outcome are associated with the fact that such

Pharmacogenomics in Chemotherapy

79

variants may occur at different frequencies between Asians and Caucasians, and/or on different, ethnicity-dependent haplotype structures. Regardless of any potential ethnic considerations, the existence of at least two potentially redundant uptake transporters in the human liver with similar affinity for paclitaxel and docetaxel supports the possibility that functional defects in both of these proteins may be required to confer substantially altered disposition phenotypes similar to those observed in OATP1B2-deficient rodents. While complete functional deficiency of either OATP1B1 or OATP1B3 has been recorded to occur, deficiency of both transporters is very rare, with an estimated frequency in the human population of about 1 in a million [53, 54]. It can thus be postulated that intrinsic physiologic and environmental variables influencing OATP1B1- or OATP1B3-mediated uptake of taxanes into hepatocytes may have a more profound influence on clearance in the general population than do defective genetic variants. This recognition is particularly relevant in the context of the recent guidelines offered by The International Transporter Consortium regarding preclinical criteria needed to trigger the conduct of clinical studies to evaluate drug-transporter interactions [55]. Indeed, it is conceivable that instances of idiosyncratic hypersensitivity to taxanes are the result of currently unrecognized drug-drug interactions at the level of hepatocellular uptake mechanisms [56]. Following transporter-mediated uptake into hepatocytes, CYP2C8 contributes the most significantly to metabolic inactivation of paclitaxel but not docetaxel [57], and polymorphisms in the CYP2C8 gene have been shown to associate with toxicity of certain anticancer drugs in affected individuals. Preliminary investigations suggest that the CYP2C8*3 variant is associated with reduced plasma concentrations of certain CYP2C8 substrate drugs like repaglinide, but this has not been consistently observed for paclitaxel in patients with cancer [50, 58, 59]. A CYP2C8*4 (792C > G; Ile264Met) allele with decreased activity has been reported, and a potentially inactive CYP2C8*5 allele (475delA; frameshift) was found in one Japanese individual. The clinical consequences of these variant alleles in relation to the clinical pharmacology of paclitaxel are yet to be determined in a prospective, adequately powered study. In addition to CYP2C8, paclitaxel can undergo hepatic metabolism by members of the CYP3A subfamily, of which CYP3A4 and CYP3A5 are the most important members. CYP3A4 activity in vivo shows extensive inter-individual variation, up to 14-fold, which may be caused by health status and environmental, hormonal, or genetic factors. It is thought that genetic differences may explain up to 90% of the observed variation in drug metabolizing capacity of patients, although “null” alleles are very rare. Most of the reported SNPs for CYP3A4 occur with allele frequencies less than 5%. Decreased CYP3A4 activity has been demonstrated for various

80

Zahra Talebi et al.

alleles, including CYP3A4*17 allele (566 T > C; Phe189Ser) in vitro, and there are indications that the CYP3A4*4 (352A > G; Ile118Val), *5 (653C > G; Pro218Arg), and *6 (831insA; frameshift) alleles encode proteins with decreased catalytic activity. The impact of these alleles on in vivo CYP3A4 activity and clearance of paclitaxel remain unclear. The next important member of the CYP3A subfamily is CYP3A5, which is expressed in only 10 to 40% of Caucasians. Due to the fact that there is a large overlap in substrate specificity between CYP3A4 and CYP3A5, the contribution of each isozyme to total CYP3A activity will depend on both the drug under investigation and a patient’s genotype. In addition, CYP3A5 is expressed extra-hepatically in the prostate, kidney, adrenal, and pituitary glands, while CYP3A4 activity is more restricted to liver and intestine. Consequently, depending on the site of action of a drug, the role of CYP3A5 may be larger than anticipated. The discovery of an allele (CYP3A5*3) that encodes the absence of a protein and occurs with a high frequency in an enzyme family known to be involved in the metabolism of many drugs supports the notion that identification of anticancer drugs that may benefit from CYP3A5 genetic screening is indicated. One of the first examples pointing in this direction is the metabolism of agents that have higher intrinsic clearance for CYP3A5 compared with CYP3A4, such as the immunosuppressive drug tacrolimus and the vinca alkaloid vincristine. For most drugs, however, including paclitaxel, the common polymorphisms in CYP3A4 and CYP3A5 do not appear to have important functional significance (62). Formerly known as MDR1 or PGY1, ABCB1 was the first human ABC transporter gene cloned and characterized through its ability to confer a multidrug resistance phenotype to cancer cells that had developed resistance to certain chemotherapy drugs, including paclitaxel. An analysis of the potential functional consequences of different most common synonymous and non-synonymous coding ABCB1 variants (i.e., Asn21Asp, Phe103Leu, Ser400Asn, Ala893Ser/Thr, and Ala998Thr) has been performed and indicated that the substrate specificity of the variant proteins was not substantially affected. This suggests that these SNPs result in mutant proteins with a distribution and function similar to the wild-type protein. In line with this observation, studies investigating clinical consequences of the ABCB1 polymorphisms in terms of their ability to modulate the pharmacokinetic profile of paclitaxel and docetaxel have been essentially negative [35, 60–62]. In spite of the limited data currently available, the collective findings suggest that ABCB1 may play a role in the pharmacokinetic profile of anticancer drug substrates that are given orally and/or undergo substantial renal excretion [63]. Studies employing more agnostic approaches such as the Illumina DMET chip, which includes several thousand variants in ADME genes of relevance to drug disposition, have identified

Pharmacogenomics in Chemotherapy

81

multiple-SNP signatures that can predict impaired clearance of paclitaxel [64], but not docetaxel [65]. Interestingly, SNPs included in the paclitaxel signature were within genes with no known relevance to paclitaxel pharmacokinetics. Similarly, a recent genome-wide association study (GWAS) reported that carriers of an intergenic SNP (rs17130142) were significantly associated with the pharmacokinetics of paclitaxel [66], but this SNP has not been connected with a functional effect on any ADME-relevant gene. Another GWAS was unable to identify any variants associated with paclitaxel pharmacokinetics [67]. Overall, inconsistent findings across these pharmacogenetic-pharmacokinetic studies indicate that no single SNP or SNP signatures have been identified with a sufficiently strong association with the pharmacokinetics of paclitaxel or docetaxel to warrant their use in guiding personalized dosing [31]. Beyond pharmacogenetic studies aimed at evaluating variation in ADME genes as predictors of taxane pharmacokinetics, various alternative approaches have been proposed to predict, prevent, and/or treat paclitaxel-induced peripheral neuropathy [68]. The predictive strategies have predominantly focused on the search for hereditary biomarkers that could identify patients at increased risk of toxicity through candidate gene [58, 61, 69–75] or genomewide association studies [76–79]. These studies have been recently reviewed [80] and have fairly consistently reported associations for variants in genes related to hereditary neuropathy conditions. However, the collective findings from these studies done to date have identified non-overlapping single or pathway biomarker associations that preclude immediate clinical implementation in order to improve the safety of taxane-based chemotherapy regimens [81– 85]. It should be pointed out that the decision to act on a neurotoxicity biomarker is further hampered in many diseases by the lack of available alternative treatments to replace paclitaxel or docetaxel and/or by the need for a patient-tailored reduction in the chemotherapy dose to prevent toxicity, which will have negative effects on disease management. Given the current state of the field of taxane pharmacogenomics, the development of neuroprotective strategies that could effectively afford tissue protection through therapeutic intervention [86], or the implementation of therapeutic drug monitoring [31], is urgently needed. 2.5

Methotrexate

Methotrexate (MTX) competitively inhibits dihydrofolate reductase, a key enzyme in the reduction of dihydrofolate to tetrahydrofolate. This disrupts the de novo biosynthesis of purines and pyrimidines, thus preventing the formation of one-carbon donors required for DNA methylation [87]. Toxicities of methotrexate include bone marrow suppression, gastrointestinal mucositis, liver toxicity, neurotoxicity, and nephrotoxicity [88].

82

Zahra Talebi et al.

There is a high variability of MTX pharmacokinetics among patients, and a large number of pharmacogenetic studies have been published in an attempt to explain and potentially predict toxicities related to methotrexate. This is particularly true of patients who receive high-dose methotrexate, which has been especially important to improve outcomes in pediatric acute lymphoblastic leukemia, the most common type of childhood cancer [87]. High-dose methotrexate is also commonly used in the treatment of sarcoma. Genes evaluated in clinical studies typically fall under the category of folate pathway genes (MTHFR, MTR, MTRR, and DHFR) or PK/transporter genes (SLCO1B1, SLC19A1, OAT1, and OAT3); however, evidence for the clinical use of most of these variants is varied, and at times, the clinical utility is unclear [87]. In a systematic review examining 58 articles and 24 different genes associated with methotrexate transporter pharmacology or the folate transport pathway, SLCO1B1 was found to be the only gene that reliably demonstrates an effect on methotrexate pharmacokinetics [87]. SLCO1B1 (solute carrier organic anion transporter family member 1B1) is expressed almost exclusively in hepatocytes, and substrates include bilirubin, estrogens, and drugs such as statins and methotrexate. SLCO1B1 rs4149056 is the SNP most commonly associated with variability in MTX clearance. This SNP reduces the localization of the transporter to the surface of the cells, resulting in a reduction in transport of MTX and decreased clearance of MTX. SLCO1B1 rs4149056 is included in two * alleles, *5, a loss of function allele, and *15, a reduced function allele. Other reduced function alleles include *23 and *31. The SLCO1B1 variant rs2306283 is included in multiple * alleles, such as *14 and *35, and results in increased transporter expression. While most studies classify this variant as one that increases MTX clearance, one study associated it with slower clearance [87]. Variants in ABC transporter family (ABCB1, ABCG2, ABCC1, ABCC2, ABCC3, ABCC4, ABCC5, ABCC10) have been demonstrated to affect methotrexate exposure in some studies. However, no variant in any ABC transporter has enough consistent, quality evidence to suggest that it could be used clinically to predict risk of toxicity. More studies are needed before these variants would offer any kind of clinical utility [87]. Polymorphisms in SLC19A1, SLC22A6, SLCO22A8, SLCO1A2, and SLCO1B3 have been evaluated in a large number of studies given their roles in folate metabolism and homeostasis. None of these polymorphisms have been significantly associated with methotrexate PK, despite their biologic plausibility [87]. Multiple genes in the folate pathway have been studied, and while some have been associated with increased or decreased clearance of methotrexate, none have demonstrated a consistent

Pharmacogenomics in Chemotherapy

83

clinically significant association with methotrexate pharmacokinetics and toxicity (ATIC, DHFR, MTHFD1, MTR, MTRR, MTHFR, TYMS) [87]. Even in studies where some significant effects of genotype on MTX PK/PD were noted, the effect size was small, and PK measures may ultimately be more effective in predicting toxicity than pharmacogenetic variants. More studies are needed to study rare variants and attempt to replicate studies where some effect on MTX clearance was noted [87]. 2.6

Platinums

The antitumor effect of platinum compounds is due to the formation of intra-and inter-strand DNA adducts [89]. Significant toxicities of platinum drugs include nephrotoxicity, ototoxicity, and myelosuppression. Platinum drugs are used to treat a wide range of malignancies in both adult and pediatric cancer patients, including lung cancer, sarcomas, brain tumors, and breast cancer. Studies have primarily examined DNA repair pathway genes (ERCC1, ERCC2 (XPD), ERCC5, XRCC1, XRCC3, and XPC) and genes involved in cisplatin metabolism (GSTP1, GSTM3, GSTM1, and GSTT1) [90]. However, multiple GWAS studies have not revealed any germline polymorphisms associated with platinum sensitivity [89]. Variants in ERRC1, a key player in the DNA cross-linking repair process, have been proposed as predictors of nephrotoxicity with platinum compounds, but evidence to date is conflicting [1]. Some evidence from somatic studies of tumor cells identifies two SNPS (rs11615 and rs3212986) in ERCC1 as associated with response to platinum-based therapy, but there is insufficient evidence to support the clinical use of these SNPs to predict response [89]. ERCC2 rs13181 has been associated with overall increased toxicity related to platinum [91]. One potential mechanism of platinum resistance is the inactivation of intracellular platinum by glutathione; therefore, polymorphisms in GSTP1, which increase intracellular glutathione, were proposed to be a mechanism of platinum resistance. However, a study of 108 patients did not reveal any association between GSTP1 haplotype and platinum resistance. The study did identify that patients with GSTP1 rs1695 (Ile105Val) had an increased risk of hematologic toxicity, a finding replicated in a case-control study when GSTP1 rs1695 was found in combination with ERCC1 (rs11615) and DPYD (rs1801265) in patients with gastric cancer [91, 92]. There is currently no evidence that GSTP1 rs1695 is associated with oxaliplatin-related neurotoxicity, though a protective role of the G allele in GSTP1 rs1695 in terms of overall toxicity is noted [91, 93]. In one meta-analysis, GSTP1 rs1695 was associated with treatment efficacy in patients with gastrointestinal cancer, but there was a significant amount of heterogeneity in the studies examined,

84

Zahra Talebi et al.

making it difficult to reliably carry this data forward into clinical practice. The same meta-analysis identified an association between the GSTT1 null variant and treatment, which was also seen in a study identifying GSTP1 polymorphic alleles and GSTM1 and GSTT1 null alleles as associated with improved survival, with less glutathione S-transferase activity associated with better treatment response. Overall, this data is most consistent in gastric cancer, and more studies with less heterogeneity are required for confirmation of these findings [90, 94]. Cisplatin exposure has been shown to result in outer hair cell death, which is then unable to regenerate in mammals, leading to hearing loss [95]. In a study of 206 adult head and neck cancer patients, >60% experience hearing loss and reported rates of cisplatin-related hearing loss in children ranging from 42 to 88%, with a great deal of inter-individual variability [95, 96]. Initially, variants in both TPMT and COMT were associated with cisplatinrelated hearing loss in children, but attempts to replicate these findings have had mixed results, and it is unclear what role these genes play in predicting this toxicity [95, 96]. A more recent GWAS study identified ACYP2 rs1872328 as a risk allele, which was then replicated within the original cohort and with an outside cohort. The same investigators then performed meta-analysis along with their own data, which showed a significant association with this ACYP2 SNP, suggesting a true association. In this study, the COMT SNP was not significant in their study cohort alone, but a pooled odds ratio was statistically significant for the associations with the COMT SNP rs464316 [96]. MATE1 (multidrug and toxin extrusion protein 1) is a bidirectional antiporter, expressed on the apical membrane of renal tubular epithelial cells, suggesting it may play an important role in modulating clearance and systemic exposure of cisplatin. In one study, patients with genotype MATE1 A/A at rs2289669 were protected, while COMT T carriers were at higher risk of hearing loss [95]. Patients with this MATE1 genotype may have enhanced cisplatin clearance, though mechanism of MATE1-related platinum toxicities is not well understood [95]. Overall, the role of pharmacogenomics in the prediction of platinum drug response and toxicity is incompletely understood, and more studies are required to confirm these associations before they can be used in clinical practice [96]. 2.7 Fluoropyrimidines

5-fluorouracil (5-FU) is an antimetabolite which mimics uracil; the anabolite 5-flurodeoxyuridine monophosphate is a better substrate for thymidylate synthase than uracil, and so this blocks the formation of thymidine monophosphate (dTMP). As this is the sole pathway for de novo dTMP formation, this results in depletion of dTMP in rapidly dividing cells and thus cell death. 5-FU metabolites are also incorporated into transcribed RNA and replicating

Pharmacogenomics in Chemotherapy

85

DNA, causing cell death. 5-FU is catabolized by dihydropyrimidine dehydrogenase (DPYD), forming inactive metabolites. There are two oral prodrugs available in the USA, capecitabine and trifluridine and tipiracil hydrochloride, which depend on the same anabolic and catabolic steps as 5-FU [97]. 10–40% of patients receiving chemotherapy with fluoropyrimidines develop severe toxicities, including myelosuppression, cardiotoxicity, hand-foot syndrome, nausea, and diarrhea, and fluoropyrimidines have a treatment-related mortality of 0.5–1% [93]. Cardiotoxicity, both primary and secondary to severe coronary vasospasm, is an infrequent but well-reported adverse effect of fluoropyrimidines [98]. Life-threatening cardiotoxicity secondary to FU is reported at a frequency of less than 1% across all age groups [99]. Continuous infusion and concomitant use of cisplatin increase risk cardiotoxicity related to fluoropyrimidines [98]. While current pharmacogenetic testing for toxicities related to 5-FU has good positive predictive value, the presence of toxicities in patients without these pharmacogenetic markers indicates that other factors contribute [97]. An umbrella review examined four systematic reviews of genes associated with fluoropyrimidine-related toxicities: two examined DPYD variants, three examined TYMS variants, and three examined MTHFR variants. MTHFR variants were not found to be of clinical significance in the prediction of fluoropyrimidine-associated toxicities in any study [93]. Dihydropyrimidine dehydrogenase (DPYD) is the initial and rate-limiting enzyme in the catabolism of fluoropyrimidines [100]. Pharmacogenomic studies have associated DPYD deficiency with an increased risk of FU cardiotoxicity [101]. In the umbrella review, the two studies which examined DPYD variants concluded that the four commonly tested variants (rs3918290 or *2A, rs55886062 or *13, rs67376798, and rs75017182, an intronic variant which promotes alternative splicing and has been shown to the causal variant in the DPYD haplotype HapB3, which is associated with toxicity) are predictive of severe fluoropyrimidineassociated toxicity [93, 97]. Evidence for a biomarker predictive of fluoropyrimidine-related toxicity was strongest for DPYD*2A (IVS14:1G > A), which predicted global and specific toxicities for capecitabine and infused fluoropyrimidines, but DPYD rs67376798 (2846A > T) also increased the likelihood of toxicity in patients on fluoropyrimidine therapy. These two variants are loss of function mutations, with DPYD*2A being the most common [93]. DPYD *2A and *13 have the most deleterious effect on DPYD activity, while c.2846A > T (rs67376798) and HapB3 (rs75017182) result in moderately reduced activity [102]. The DPYD haplotype HapB3 spans intron 5 to exon 11 and is likely a result of the causal variant *2A, which is located at the intron boundary of exon 14, and results in the entire exon being skipped and thus a non-functional protein [102].

86

Zahra Talebi et al.

These four DPYD variants are currently the only ones tested and considered validated risk variants for toxicity and are predominantly found in people of European descent, though other variants have recently been identified which have been shown to result in DPYD dysfunction, such as rs115232898, which was found in 3–5% of African Americans studied. HapB3 is the most common decreased function variant in Europeans, and among all four common variants combined, there is a MAF of ~7% among Europeans [102]. Other variants found in other non-European populations highlight the need for geographic and racial diversification of pharmacogenetic studies. Preemptive testing for these DPYD variants has not yet been approved by the FDA but is currently done in several European countries, and CPIC guidelines have been published for this DGI [97]. In the CPIC guidelines, a normal metabolizer has no DPYD alleles with decreased or loss of function, an intermediate metabolizer has one loss of function allele or two decreased function alleles, and a poor metabolizer carries at least one loss of function allele with the other allele being no function or decreased function. CPIC guidelines are based on the DPYD activity score, which ranges from 0 (two no function alleles) to 2 (no variants), and dose reduction or avoidance of 5-FU or F-FU prodrugs is recommended depending on the score. The guidelines note that ~50% of decreased function DPYD variant allele carriers develop toxicities when treated with standard doses of 5-FU or 5-FU prodrugs; however, patients without any DPYD variant may still develop toxicities due to other factors [102]. Regarding the TYMS variants discussed in the umbrella review (TYMS 5’UTR 2R/3R and 3’ UTR bp ins/del, both of which result in decreased expression), while there was a statistically significant association between the polymorphisms and toxicity, the effect size was small and of limited clinical utility [93]. These common variants in TYMS have a variable number of tandem repeats in the TYMS promoter/enhanced region, which in conjunction with single nucleotide variants in the repeat motif determines the number of binding sites for the transcription factor USF-1. The number of USF-1 binding sites available has been associated with 5-FU-related gastrointestinal toxicity, but more studies are required before this can be considered a reliable biomarker for clinical practice [97]. Other studies have also shown heterogeneity in the association, with variants in TYMS being associated with up to a 2.5-fold risk of toxicities in some studies and no association in others [100]. There is a case report of a pediatric patient heterozygous for two TYMS variants (deletion in the 30 region, rs16430; G > C substitution in the promoter enhancer region of the allele with two tandem repeats, rs183205964) who developed cardiotoxicity with 5FU therapy [100]. Nevertheless, the clinical utility of TYMS

Pharmacogenomics in Chemotherapy

87

genotyping to predict risk of fluoropyrimidine-related toxities is unclear, and dosing strategies and guidelines are not yet in place based on them [102]. 2.8

Asparaginase

Asparaginase is a key component to treatment of leukemia and lymphoma, and immune reactions to asparaginase in its various formulations can result in treatment delays and decreased exposure to the drug, resulting in increased risk of relapse. Asparaginase is a non-human enzyme which depletes asparagine in the serum [103]. Leukemic cells lose the ability to produce asparagine, and so the lack of exogenous asparagine results in cell death [104]. Patients receiving asparaginase products can have hypersensitivity reactions, the most frequent toxicity related to asparaginase, or can develop neutralizing antibodies which decrease the exposure to the drug [104]. Asparaginase hypersensitivity reactions typically develop with the second dose of the drug and beyond, suggesting that patients are sensitized to the drug during initial doses, with asparaginase processing by antigen presenting cells, epitope presentation on major histocompatibility complex class 2 molecules, CD41 T-cell activation/differentiation, and T-cell–dependent B-cell differentiation to plasma and memory B cells, resulting in a hypersensitivity reaction with subsequent doses [105]. Alterations in HLA-DRB1*07:01 have been associated with asparaginase hypersensitivity in patients of diverse backgrounds [106]. High-risk amino acids located in the binding pocket of the HLA protein of patients with HLA-DRB1*07:01 may alter the interaction between the drug and the HLA-DRB1 protein, increasing affinity and resulting in increased immune response [103]. In a study of 359 pediatric Hungarian ALL patients, patients with HLA-DRB1*07:01 and HLA-DQB1*02:02 alleles had significantly higher risk of developing asparaginase hypersensitivity compared to non-carriers. The haplotype HLADRB1*07:01-HLADQB1*02:02 was also associated with increased risk, and the haplotype HLA-DRB1*07:01–HLA-DQA1*02:01–HLA-DQB1*02: 02 was associated with the highest risk. The HLA-DQB1*02:02 allele was carried by significantly fewer T-cell ALL patients than B-cell [104]. A GWAS study of 4259 pediatric patients of diverse ancestry who received PEG asparaginase as part of leukemia therapy found a similar association between asparaginase hypersensitivity and the HLA haplotype with DRB1*07:01, DQA1*02:01, and DQB1*02: 02; this association was strongest for patients of European ancestry, and partially replicated in patients with non-European ancestry. An association between ARHGAP28 rs9958628, which has been linked to immune response in children, and asparaginase hypersensitivity was also identified in patients of non-European ancestry [107].

88

Zahra Talebi et al.

In a study of 284 Caucasian children who received asparaginase for leukemia, HLA-DRB1*07:01 and DQA1*02:01, which are in linkage disequilibrium, were significantly associated with asparaginase hypersensitivity. Haplotype analysis demonstrated that the association was dependent on the presence of the DQB1*02:02 in the HLA-DRB1*07:01-DQA1*02:01-DQB1*02:02 haplotype [106]. A study of 541 patients of European ancestry who were receiving native E coli asparaginase (Elspar) or PEGylated E coli asparaginase (Oncaspar) as part of leukemia therapy found a significant association between HLA-DRB1*07:01 alleles and risk of hypersensitivity reaction [103]. And in a study of 59 cases of asparaginase hypersensitivity with lack of detectable enzyme activity and 772 controls from the Nordic Society for Pediatric Hematology and Oncology (NOPHO), HLA-DQA1 rs9272131 was associated with PEG-asparaginase allergy, along with TAP2 rs115360810 and CNOT3 rs73062673 on 19q13.42 [108]. The TAP2 variant is in close proximity to the HLA-DQ1 variant. The TAP2 gene is involved in the transport of antigens into the cytoplasmic reticulum and has been shown to contribute to asthma and allergy. CNOT3 is part of the CCR4-not complex, which regulates gene expression and cellular signals. CNOT3 knockdown increases the transcription of major histocompatibility class (MHC) II molecules and thus has a regulatory effect on HLA genes [108]. The nuclear factor of activated T cells (NFAT) family of transcription factors has an important role in immunity and can affect antibody response to protein-based therapeutics like asparaginase. In a mouse study using NFATC2-deficient mice, mice with NFATC2-deficiency had a decreased risk of asparaginase hypersensitivity, with impaired T helper response. An intronic variant of NFATC2 (rs6021191) increases NFATC2 mRNA and has been associated with an increased risk of asparaginase hypersensitivity. In contrast, patients with Trisomy 21 overexpress the NFAT inhibitors DYRK1A and DSCR19 and have a lower risk of developing asparaginase hypersensitivity [105]. While further studies are required before these pharmacogenetic biomarkers can be considered reliable for use, the ability to predict patients at higher risk for asparaginase hyper-reactivity may allow clinicians to better prevent these reactions, thus maintaining asparaginase as a therapeutic option for this patient population and improving risk of relapse.

Pharmacogenomics in Chemotherapy

3

89

Conclusion Substantial progress has been made in recent years toward optimization of cytotoxic chemotherapeutics with the use of pharmacogenetics, although various aspects of anticancer drug pharmacology deserve more work before they become more useful clinically. Indeed, incorporation of pharmacologic principles in drug development, clinical trials, and routine care are essential to maximize the clinical potential of anticancer agents, as improvements in outcome have been observed using these principles to individualize anticancer drug administration for some agents. In addition to the application of genetic screening as a prognostic test, pharmacogenomic studies are now rapidly elucidating the inherited nature of differences in drug disposition and pharmacodynamic effects and will provide a stronger scientific basis for optimizing drug therapy on the basis of each patient’s or tumor’s genetic constitution.

References 1. Al-Mahayri ZN, Patrinos GP, Ali BR (2020) Toxicity and pharmacogenomic biomarkers in breast cancer chemotherapy. Front Pharmacol 11:445 2. Weng L, Zhang L, Peng Y et al (2013) Pharmacogenetics and pharmacogenomics: a bridge to individualized cancer therapy. Pharmacogenomics 14(3):315–324 3. FDA (2017) Table of pharmacogenomic biomarkers in drug labeling. https://www.fda. gov/Drugs/ScienceResearch/ucm572698. htm. Accessed 2 Feb 2018 4. de Beaumais TA, Jacqz-Aigrain E (2012) Intracellular disposition of methotrexate in acute lymphoblastic leukemia in children. Curr Drug Metab 13(6):822–834 5. Zaza G, Cheok M, Krynetskaia N et al (2010) Thiopurine pathway. Pharmacogenet Genomics 20(9):573–574 6. Weinshilboum R (2001) Thiopurine pharmacogenetics: clinical and molecular studies of thiopurine methyltransferase. Drug Metab Dispos 29(4 Pt 2):601–605 7. Yang JJ, Landier W, Yang W et al (2015) Inherited NUDT15 variant is a genetic determinant of mercaptopurine intolerance in children with acute lymphoblastic leukemia. J Clin Oncol 33(11):1235–1242 8. Relling MV, Hancock ML, Rivera GK et al (1999) Mercaptopurine therapy intolerance and heterozygosity at the thiopurine S-methyltransferase gene locus. J Natl Cancer Inst 91(23):2001–2008

9. Relling MV, Klein TE (2011) CPIC: clinical pharmacogenetics implementation consortium of the pharmacogenomics research network. Clin Pharmacol Ther 89(3):464–467 10. Relling MV, Gardner EE, Sandborn WJ et al (2013) Clinical pharmacogenetics implementation consortium guidelines for thiopurine methyltransferase genotype and thiopurine dosing: 2013 update. Clin Pharmacol Ther 93(4):324–325 11. Relling MV, Schwab M, Whirl-Carrillo M, Suarez-Kurtz G, Pui CH, Stein CM, Moyer AM, Evans WE, Klein TE, Antillon-Klussmann FG, Caudle KE, Kato M, Yeoh AEJ, Schmiegelow K, Yang JJ (2019) Clinical Pharmacogenetics Implementation Consortium Guideline for Thiopurine Dosing Based on TPMT and NUDT15 Genotypes: 2018 Update. Clin Pharmacol Ther 105 (5):1095–1105. https://doi.org/10.1002/ cpt.1304. Epub 2019 Jan 20. PMID: 30447069; PMCID: PMC6576267 12. Walker GJ, Harrison JW, Heap GA et al (2019) Association of Genetic Variants in NUDT15 with Thiopurine-induced myelosuppression in patients with inflammatory bowel disease. JAMA 321(8):773–785 13. Tanaka Y, Kato M, Hasegawa D et al (2015) Susceptibility to 6-MP toxicity conferred by a NUDT15 variant in Japanese children with acute lymphoblastic leukaemia. Br J Haematol 171(1):109–115 14. Moriyama T, Nishii R, Perez-Andreu V et al (2016) NUDT15 polymorphisms alter

90

Zahra Talebi et al.

thiopurine metabolism and hematopoietic toxicity. Nat Genet 48(4):367–373 15. Matimba A, Li F, Livshits A et al (2014) Thiopurine pharmacogenomics: association of SNPs with clinical response and functional validation of candidate genes. Pharmacogenomics 15(4):433–447 16. Stocco G, Franca R, Verzegnassi F et al (2012) Multilocus genotypes of relevance for drug metabolizing enzymes and therapy with thiopurines in patients with acute lymphoblastic leukemia. Front Genet 3:309 17. Hareedy MS, El Desoky ES, Woillard JB et al (2015) Genetic variants in 6-mercaptopurine pathway as potential factors of hematological toxicity in acute lymphoblastic leukemia patients. Pharmacogenomics 16(10): 1119–1134 18. Aminkeng F, Ross CJ, Rassekh SR et al (2016) Recommendations for genetic testing to reduce the incidence of anthracyclineinduced cardiotoxicity. Br J Clin Pharmacol 82(3):683–695 19. Scott E, Hasbullah JS, Ross CJ et al (2018) Reducing anthracycline-induced cardiotoxicity through pharmacogenetics. Pharmacogenomics 19(15):1147–1150 20. Leerink JM, de Baat EC, Feijen EAM et al (2020) Cardiac disease in childhood cancer survivors: risk prediction, prevention, and surveillance: JACC CardioOncology state-ofthe-art review. JACC Cardio Oncol 2(3): 363–378 21. Rossato LG, Costa VM, Dallegrave E et al (2014) Mitochondrial cumulative damage induced by mitoxantrone: late onset cardiac energetic impairment. Cardiovasc Toxicol 14(1):30–40 22. Leong SL, Chaiyakunapruk N, Lee SW (2017) Candidate Gene Association studies of anthracycline-induced cardiotoxicity: a systematic review and meta-analysis. Sci Rep 7(1):39 23. Alexandre J, Cautela J, Ederhy S et al (2020) Cardiovascular toxicity related to cancer treatment: a pragmatic approach to the American and European cardio-oncology guidelines. J Am Heart Assoc 9(18):e018403 24. Loucks CM, Yan K, Tanoshima R, Ross CJD, Rassekh SR, Carleton BC (2022) Pharmacogenetic testing to guide therapeutic decisionmaking and improve outcomes for children undergoing anthracycline-based chemotherapy. Basic Clin Pharmacol Toxicol 30(Suppl 1):95–99. https://doi.org/10.1111/bcpt. 13593. Epub 2021 May 11. PMID: 33900042

25. Argevani L, Hughes C, Schuh MJ (2020) Dosage adjustment of irinotecan in patients with UGT1A1 polymorphisms: a review of current literature. Innov Pharm 11 (3). https://doi.org/10.24926/iip.v11i3. 3203. PMID: 34007623; PMCID: PMC8075136 26. de Man FM, Goey AKL, van Schaik RHN et al (2018) Individualization of irinotecan treatment: a review of pharmacokinetics, pharmacodynamics, and pharmacogenetics. Clin Pharmacokinet 57(10):1229–1254 27. Marsh S, Hoskins JM (2010) Irinotecan pharmacogenomics. Pharmacogenomics 11(7): 1003–1010 28. De Iuliis F, Taglieri L, Salerno G et al (2015) Taxane induced neuropathy in patients affected by breast cancer: literature review. Crit Rev Oncol Hematol 96(1):34–45 29. Mielke S, Sparreboom A, Mross K (2006) Peripheral neuropathy: a persisting challenge in paclitaxel-based regimes. Eur J Cancer 42(1):24–30 30. Peters CM, Jimenez-Andrade JM, Jonas BM et al (2007) Intravenous paclitaxel administration in the rat induces a peripheral sensory neuropathy characterized by macrophage infiltration and injury to sensory neurons and their supporting cells. Exp Neurol 203(1):42–54 31. Hertz DL (2021) Exploring pharmacogenetics of paclitaxel- and docetaxel-induced peripheral neuropathy by evaluating the direct pharmacogenetic-pharmacokinetic and pharmacokinetic-neuropathy relationships. Expert Opin Drug Metab Toxicol 17(2): 227–239 32. Carozzi VA, Canta A, Chiorazzi A (2015) Chemotherapy-induced peripheral neuropathy: what do we know about mechanisms? Neurosci Lett 596:90–107 33. Franke RM, Carducci MA, Rudek MA et al (2010) Castration-dependent pharmacokinetics of docetaxel in patients with prostate cancer. J Clin Oncol 28(30):4562–4567 34. Nieuweboer AJ, Hu S, Gui C et al (2014) Influence of drug formulation on OATP1Bmediated transport of paclitaxel. Cancer Res 74(11):3137–3145 35. Baker SD, Verweij J, Cusatis GA et al (2009) Pharmacogenetic pathway analysis of docetaxel elimination. Clin Pharmacol Ther 85(2):155–163 36. de Graan AJ, Lancaster CS, Obaidat A et al (2012) Influence of polymorphic OATP1Btype carriers on the disposition of docetaxel. Clin Cancer Res Off J Am Associat Cancer Res 18(16):4433–4440

Pharmacogenomics in Chemotherapy 37. Smith NF, Marsh S, Scott-Horton TJ et al (2007) Variants in the SLCO1B3 gene: interethnic distribution and association with paclitaxel pharmacokinetics. Clin Pharmacol Ther 81(1):76–82 38. Smith NF, Acharya MR, Desai N et al (2005) Identification of OATP1B3 as a high-affinity hepatocellular transporter of paclitaxel. Cancer Biol Ther 4(8):815–818 39. Gui C, Miao Y, Thompson L et al (2008) Effect of pregnane X receptor ligands on transport mediated by human OATP1B1 and OATP1B3. Eur J Pharmacol 584(1): 57–65 40. Letschert K, Faulstich H, Keller D et al (2006) Molecular characterization and inhibition of amanitin uptake into human hepatocytes. Toxicol Sci Off J Soc Toxicol 91(1): 140–149 41. Yamaguchi H, Kobayashi M, Okada M et al (2008) Rapid screening of antineoplastic candidates for the human organic anion transporter OATP1B3 substrates using fluorescent probes. Cancer Lett 260(1–2): 163–169 42. Iusuf D, Hendrikx JJ, van Esch A et al (2015) Human OATP1B1, OATP1B3 and OATP1A2 can mediate the in vivo uptake and clearance of docetaxel. Int J Cancer 136(1):225–233 43. Marada VV, Florl S, Kuhne A et al (2015) Interaction of human organic anion transporter polypeptides 1B1 and 1B3 with antineoplastic compounds. Eur J Med Chem 92: 723–731 44. Sun X, Li J, Guo C et al (2016) Pharmacokinetic effects of curcumin on docetaxel mediated by OATP1B1, OATP1B3 and CYP450s. Drug Metab Pharmacokinet 31(4):269–275 45. van de Steeg E, van Esch A, Wagenaar EE et al (2013) Influence of human OATP1B1, OATP1B3, and OATP1A2 on the pharmacokinetics of methotrexate and paclitaxel in humanized transgenic mice. Clin Cancer Res 19(4):821–832 46. van de Steeg E, van Esch A, Wagenaar E et al (2011) High impact of Oatp1a/1b transporters on in vivo disposition of the hydrophobic anticancer drug paclitaxel. Clin Cancer Res 17(2):294–301 47. Gui C, Wahlgren B, Lushington GH et al (2009) Identification, Ki determination and CoMFA analysis of nuclear receptor ligands as competitive inhibitors of OATP1B1mediated estradiol-17beta-glucuronide transport. Pharmacol Res 60(1):50–56

91

48. Kameyama Y, Yamashita K, Kobayashi K et al (2005) Functional characterization of SLCO1B1 (OATP-C) variants, SLCO1B1*5, SLCO1B1*15 and SLCO1B1*15+C1007G, by using transient expression systems of HeLa and HEK293 cells. Pharmacogenet Genomics 15(7): 513–522 49. Niemi M, Pasanen MK, Neuvonen PJ (2011) Organic anion transporting polypeptide 1B1: a genetically polymorphic transporter of major importance for hepatic drug uptake. Pharmacol Rev 63(1):157–181 50. Marcath LA, Kidwell KM, Robinson AC et al (2019) Patients carrying CYP2C8*3 have shorter systemic paclitaxel exposure. Pharmacogenomics 20(2):95–104 51. Chew SC, Singh O, Chen X et al (2011) The effects of CYP3A4, CYP3A5, ABCB1, ABCC2, ABCG2 and SLCO1B3 single nucleotide polymorphisms on the pharmacokinetics and pharmacodynamics of docetaxel in nasopharyngeal carcinoma patients. Cancer Chemother Pharmacol 67(6):1471–1478 52. Chew SC, Sandanaraj E, Singh O et al (2011) Influence of SLCO1B3 haplotype-tag SNPs on docetaxel disposition in Chinese nasopharyngeal cancer patients. Br J Clin Pharmacol. https://doi.org/10.1111/j.1365-2125. 2011.04123.x 53. van de Steeg E, Stranecky V, Hartmannova H et al (2012) Complete OATP1B1 and OATP1B3 deficiency causes human rotor syndrome by interrupting conjugated bilirubin reuptake into the liver. J Clin Invest 122(2): 519–528 54. Kim SR, Saito Y, Sai K et al (2007) Genetic variations and frequencies of major haplotypes in SLCO1B1 encoding the transporter OATP1B1 in Japanese subjects: SLCO1B1*17 is more prevalent than *15. Drug Metab Pharmacokinet 22(6):456–461 55. Giacomini KM, Huang SM, Tweedie DJ et al (2010) Membrane transporters in drug development. Nat Rev Drug Discov 9(3):215–236 56. Hu S, Mathijssen RH, de Bruijn P et al (2014) Inhibition of OATP1B1 by tyrosine kinase inhibitors: in vitro-in vivo correlations. Br J Cancer 110(4):894–898 57. Leblanc AF, Sprowl JA, Alberti P et al (2018) OATP1B2 deficiency protects against paclitaxel-induced neurotoxicity. J Clin Invest 128(2):816–825 58. Green H, Soderkvist P, Rosenberg P et al (2009) Pharmacogenetic studies of paclitaxel in the treatment of ovarian cancer. Basic Clin Pharmacol Toxicol 104(2):130–137

92

Zahra Talebi et al.

59. Bergmann TK, Brasch-Andersen C, Gre´en H et al (2011) Impact of CYP2C8*3 on paclitaxel clearance: a population pharmacokinetic and pharmacogenomic study in 93 patients with ovarian cancer. Pharmacogenomics J 11(2):113–120 60. Henningsson A, Marsh S, Loos WJ et al (2005) Association of CYP2C8, CYP3A4, CYP3A5, and ABCB1 polymorphisms with the pharmacokinetics of paclitaxel. Clin Cancer Res Off J Am Associat Cancer Res 11(22): 8097–8104 61. de Graan AJ, Elens L, Sprowl JA et al (2013) CYP3A4*22 genotype and systemic exposure affect paclitaxel-induced neurotoxicity. Clin Cancer Res Off J Am Associat Cancer Res 19(12):3316–3324 62. Stage TB, Hu S, Sparreboom A et al (2021) Role for drug transporters in chemotherapyinduced peripheral neuropathy. Clin Transl Sci 14(2):460–467 63. Franke RM, Gardner ER, Sparreboom A (2010) Pharmacogenetics of drug transporters. Curr Pharm Des 16(2):220–230 64. de Graan AJ, Elens L, Smid M et al (2013) A pharmacogenetic predictive model for paclitaxel clearance based on the DMET platform. Clin Cancer Res Off J Am Associat Cancer Res 19(18):5210–5217 65. Nieuweboer AJ, Smid M, de Graan AM et al (2016) Role of genetic variation in docetaxelinduced neutropenia and pharmacokinetics. Pharmacogenomics J 16(6):519–524 66. Gao B, Lu Y, Nieuweboer AJM et al (2018) Genome-wide association study of paclitaxel and carboplatin disposition in women with epithelial ovarian cancer. Sci Rep 8(1):1508 67. Sato Y, Yamamoto N, Kunitoh H et al (2011) Genome-wide association study on overall survival of advanced non-small cell lung cancer patients treated with carboplatin and paclitaxel. J Thorac Oncol 6(1):132–138 68. Scripture CD, Figg WD, Sparreboom A (2006) Peripheral neuropathy induced by paclitaxel: recent insights and future perspectives. Curr Neuropharmacol 4(2):165–172 69. Boora GK, Kanwar R, Kulkarni AA et al (2016) Testing of candidate single nucleotide variants associated with paclitaxel neuropathy in the trial NCCTG N08C1 (Alliance). Cancer Med 5(4):631–639 70. Hertz DL, Motsinger-Reif AA, Drobish A et al (2012) CYP2C8*3 predicts benefit/risk profile in breast cancer patients receiving neoadjuvant paclitaxel. Breast Cancer Res Treat 134(1):401–410 71. Hertz DL, Roy S, Motsinger-Reif AA et al (2013) CYP2C8*3 increases risk of

neuropathy in breast cancer patients treated with paclitaxel. Annals Oncol Off J Eur Soc Med Oncol ESMO 24(6):1472–1478 72. Leskela S, Jara C, Leandro-Garcia L et al (2011) Polymorphisms in cytochromes P450 2C8 and 3A5 are associated with paclitaxel neurotoxicity. Pharmacogenomics Journal 11(2):121–129 73. Sissung TM, Mross K, Steinberg SM et al (2006) Association of ABCB1 genotypes with paclitaxel-mediated peripheral neuropathy and neutropenia. Eur J Cancer 42(17): 2893–2896 74. Tanabe Y, Shimizu C, Hamada A et al (2017) Paclitaxel-induced sensory peripheral neuropathy is associated with an ABCB1 single nucleotide polymorphism and older age in Japanese. Cancer Chemother Pharmacol 79(6):1179–1186 75. Abraham JE, Guo Q, Dorling L, Tyrer J, Ingle S, Hardy R, Vallier AL, Hiller L, Burns R, Jones L, Bowden SJ, Dunn JA, Poole CJ, Caldas C, Pharoah PP, Earl HM (2014) Replication of genetic polymorphisms reported to be associated with taxane-related sensory neuropathy in patients with early breast cancer treated with Paclitaxel. Clin Cancer Res 20(9):2466–2475. https://doi. org/10.1158/1078-0432.CCR-13-3232. Epub 2014 Mar 5. PMID: 24599932 76. Baldwin RM, Owzar K, Zembutsu H et al (2012) A genome-wide association study identifies novel loci for paclitaxel-induced sensory peripheral neuropathy in CALGB 40101. Clin Cancer Res Off J Am Associat Cancer Res 18(18):5099–5109 77. Bergmann TK, Vach W, Feddersen S, Eckhoff L, Gre´en H, Herrstedt J, Brosen K (2013) GWAS-based association between RWDD3 and TECTA variants and paclitaxel induced neuropathy could not be confirmed in Scandinavian ovarian cancer patients. Acta Oncol 52(4):871–874. https://doi.org/10. 3109/0284186X.2012.707787. Epub 2012 Aug 9. PMID: 22877241 78. Lam SW, Frederiks CN, van der Straaten T et al (2016) Genotypes of CYP2C8 and FGD4 and their association with peripheral neuropathy or early dose reduction in paclitaxel-treated breast cancer patients. Br J Cancer 115(11):1335–1342 79. Sucheston L, Zhao H, Yao S et al (2011) Genetic predictors of taxane-induced neurotoxicity in a SWOG phase III intergroup adjuvant breast cancer treatment trial (S0221), vol 130. Springer, Netherlands 80. Al-Mahayri ZN, AlAhmad MM, Ali BR (2021) Current opinion on the

Pharmacogenomics in Chemotherapy pharmacogenomics of paclitaxel-induced toxicity. Expert Opin Drug Metab Toxicol 17(7): 785–801 81. Brewer JR, Morrison G, Dolan ME et al (2016) Chemotherapy-induced peripheral neuropathy: current status and progress. Gynecol Oncol 140(1):176–183 82. Schneider BP, Hershman DL, Loprinzi C (2015) Symptoms: chemotherapy-induced peripheral neuropathy. Adv Exp Med Biol 862:77–87 83. Schneider BP, Li L, Radovich M et al (2015) Genome-wide association studies for Taxaneinduced peripheral neuropathy in ECOG5103 and ECOG-1199. Clin Cancer Res Off J Am Associat Cancer Res 21(22):5082–5091 84. Frederiks CN, Lam SW, Guchelaar HJ et al (2015) Genetic polymorphisms and paclitaxel- or docetaxel-induced toxicities: a systematic review. Cancer Treat Rev 41(10): 935–950 85. Hertz DL (2013) Germline pharmacogenetics of paclitaxel for cancer treatment. Pharmacogenomics 14(9):1065–1084 86. Hu S, Huang KM, Adams EJ et al (2019) Recent developments of novel pharmacologic therapeutics for prevention of chemotherapyinduced peripheral neuropathy. Clin Cancer Res Off J Am Associat Cancer Res 25(21): 6295–6301 87. Taylor ZL, Vang J, Lopez-Lopez E, Oosterom N, Mikkelsen T, Ramsey LB (2021) Systematic review of pharmacogenetic factors that influence high-dose methotrexate pharmacokinetics in pediatric malignancies. Cancers (Basel) 13(11):2837. https://doi.org/ 10.3390/cancers13112837. PMID: 34200242; PMCID: PMC8201112 88. Schmiegelow K (2009) Advances in individual prediction of methotrexate toxicity: a review. Br J Haematol 146(5):489–503 89. Mao CX, Li M, Zhang W et al (2020) Pharmacogenomics for the efficacy of platinumbased chemotherapy: old drugs, new integrated perspective. Biomed Pharmacother 126:110057 90. Roco A, Cayun J, Contreras S et al (2014) Can pharmacogenetics explain efficacy and safety of cisplatin pharmacotherapy? Front Genet 5:391 91. Cordova-Delgado M, Bravo ML, Cumsille E et al (2021) A case-control study of a combination of single nucleotide polymorphisms and clinical parameters to predict clinically relevant toxicity associated with fluoropyrimidine and platinum-based chemotherapy in gastric cancer. BMC Cancer 21(1):1030

93

92. Booton R, Ward T, Heighway J et al (2006) Glutathione-S-transferase P1 isoenzyme polymorphisms, platinum-based chemotherapy, and non-small cell lung cancer. J Thorac Oncol 1(7):679–683 93. Campbell JM, Bateman E, Peters M et al (2016) Fluoropyrimidine and platinum toxicity pharmacogenetics: an umbrella review of systematic reviews and meta-analyses. Pharmacogenomics 17(4):435–451 94. Sun Y, Pan J, Tong X et al (2019) Glutathione S-transferases genes variants and chemotherapy efficacy in gastrointestinal cancer patients: a meta-analysis based on 50 pharmacogenetic studies. J Cancer 10(13):2915–2926 95. Teft WA, Winquist E, Nichols AC et al (2019) Predictors of cisplatin-induced ototoxicity and survival in chemoradiation treated head and neck cancer patients. Oral Oncol 89: 72–78 96. Thiesen S, Yin P, Jorgensen AL et al (2017) TPMT, COMT and ACYP2 genetic variants in paediatric cancer patients with cisplatininduced ototoxicity. Pharmacogenet Genomics 27(6):213–222 97. Diasio RB, Innocenti F, Offer SM (2021) Pharmacogenomic-guided therapy in colorectal cancer. Clin Pharmacol Ther 110(3): 616–625 98. Saif MW (2013) Dihydropyrimidine dehydrogenase gene (DPYD) polymorphism among Caucasian and non-Caucasian patients with 5-FU- and capecitabine-related toxicity using full sequencing of DPYD. Cancer Genomics Proteomics 10(2):89–92 99. Keefe DL, Roistacher N, Pierri MK (1993) Clinical cardiotoxicity of 5-fluorouracil. J Clin Pharmacol 33(11):1060–1070 100. Belsky JA, Yeager ND, Fitch J et al (2019) Case of severe cardiotoxicity in a pediatric patient after fluorouracil administration. JCO Precis Oncol 3:1–4 101. Van Kuilenburg AB, Meinsma R, Zoetekouw L et al (2002) High prevalence of the IVS14 + 1G>a mutation in the dihydropyrimidine dehydrogenase gene of patients with severe 5-fluorouracil-associated toxicity. Pharmacogenetics 12(7):555–558 102. Amstutz U, Henricks LM, Offer SM et al (2018) Clinical pharmacogenetics implementation consortium (CPIC) guideline for Dihydropyrimidine dehydrogenase genotype and Fluoropyrimidine dosing: 2017 update. Clin Pharmacol Ther 103(2):210–216 103. Fernandez CA, Smith C, Yang W et al (2014) HLA-DRB1*07:01 is associated with a

94

Zahra Talebi et al.

higher risk of asparaginase allergies. Blood 124(8):1266–1276 104. Kutszegi N, Yang X, Gezsi A et al (2017) HLA-DRB1*07:01-HLA-DQA1*02:01HLA-DQB1*02:02 haplotype is associated with a high risk of asparaginase hypersensitivity in acute lymphoblastic leukemia. Haematologica 102(9):1578–1586 105. Rathod S, Ramsey M, Finkelman FD et al (2020) Genetic inhibition of NFATC2 attenuates asparaginase hypersensitivity in mice. Blood Adv 4(18):4406–4416 106. Gagne V, St-Onge P, Beaulieu P et al (2020) HLA alleles associated with asparaginase hypersensitivity in childhood ALL: a report from the DFCI consortium. Pharmacogenomics 21

107. Liu Y, Yang W, Smith CA, Cheng C, Karol SE, Larsen EC, Winick N, Carroll WL, Loh ML, Raetz EA, Hunger SP, Winter SS, Dunsmore KP, Devidas M, Yang JJ, Evans WE, Jeha S, Pui CH, Inaba H, Relling MV (2021) Class II human leukocyte antigen variants associate with risk of pegaspargase hypersensitivity. Clin Pharmacol Ther 110 (3):794–802. https://doi.org/10.1002/cpt. 2241. Epub 2021 Apr 21. PMID: 33768542; PMCID: PMC8790808 108. Hojfeldt SG, Wolthers BO, Tulstrup M et al (2019) Genetic predisposition to PEG-asparaginase hypersensitivity in children treated according to NOPHO ALL2008. Br J Haematol 184(3):405–417

Chapter 5 Management of Side Effects in the Personalized Medicine Era: Chemotherapy-Induced Peripheral Neurotoxicity Eleonora Pozzi and Paola Alberti Abstract Pharmacogenomics is a powerful tool to predict individual response to treatment, in order to personalize therapy, and it has been explored extensively in oncology practice. Not only efficacy on the malignant disease has been investigated but also the possibility to predict adverse effects due to drug administration. Chemotherapy-induced peripheral neurotoxicity (CIPN) is one of those. This potentially severe and longlasting/permanent side effect of commonly administered anticancer drugs can severely impair quality of life (QoL) in a large cohort of long survival patients. So far, a pharmacogenomics-based approach in CIPN regard has been quite delusive, making a methodological improvement warranted in this field of interest: even the most refined genetic analysis cannot be effective if not applied correctly. Here we try to devise why it is so, suggesting how THE “bench-side” (pharmacogenomics) might benefit from and should cooperate with THE “bed-side” (clinimetrics), in order to make genetic profiling effective if applied to CIPN. Key words Pharmacogenomics, Chemotherapy-induced peripheral neurotoxicity, Personalized medicine

1 1.1

Introduction: Pharmacogenomics and CIPN: What and Why? The Identikit

Pharmacogenomics could be a powerful tool to predict individual response to therapy on the basis of inter-individual genetic differences. It could be employed to define a genetic signature that can predict either efficacy or adverse effect of a given treatment. In respect to genetic variability, single nucleotide polymorphisms (SNPs) account for more than 90% of genetic variations in the human genome; remaining alterations are due to insertions and deletions, tandem repeats, and microsatellites [1]. A genetic profile on the basis of SNPs would be more than useful for chemotherapy-induced peripheral neurotoxicity (CIPN) risk stratification, prior to start treatment. CIPN is a potentially severe and long-lasting side effect of commonly employed

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_5, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

95

96

Eleonora Pozzi and Paola Alberti

anticancer drugs: platinum compounds, taxanes, proteasome inhibitors, vinca alkaloids, and epothilones; they are used every day to treat the big killers: breast, colorectal, lung cancer and multiple myeloma [2–5]. CIPN influences chemotherapy administration: it can lead to dose reduction or even discontinuation in case of more severe neurological toxicity. Quality of life (QoL) can be even impaired in a large population of cancer patients due to CIPN symptoms/signs. Affected subjects experience mainly sensory alterations at limb extremities (hypoesthesia, paresthesia, neuropathic pain) and, less frequently, mild distal limb weakness [6]. 1.2

2

Issues in CIPN

One of the main still unmet clinical and scientific needs in this field is the lack of a gold standard for CIPN diagnosis and grading [6– 8]. As a consequence, in clinical trials designed so far, there was a difficulty in appropriate endpoint(s) selection; moreover, the absence of precise data on incidence and prevalence of CIPN, due to this lack, made study design less solid than would be required. That could partially explain why no preventive or curative strategy for CIPN has been found as efficacious. Thus, stratifying patients in high/low risk to develop neuropathy, thanks to pharmacogenomics, is an even more urgent matter. But the aforementioned methodological pitfalls also reflect themselves on pharmacogenomics, when applied to CIPN: no clear gold standard to define CIPN presence and severity means a potentially poor population stratification before proceeding to pharmacogenomics analysis. Even more importantly, identification of target gene(s) is a crucial issue in any toxicogenomic analysis. For what regard CIPN, most studies selected gene targets on the basis of mechanistic hypotheses mainly relevant to cancer cells, instead of genes involved in neurons and glial cells.

CIPN and Pharmacogenomics So Far Gene target selection in the vast majority of studies reported in literature was performed mainly looking for pathways related to cancer cells. Therefore, genes that have a role in drug disposition, metabolism, and detoxification, DNA repair, and cancer-cell resistance have been studied much more extensively than the ones directly related to the peripheral nervous system; to give a general idea, principal genes analyzed were (see Table 1 for extensive description): • GSTP1 gene: it is part of the glutathione S-transferases, a family of enzymes that catalyze conjugation of many hydrophobic and electrophilic compounds with reduced glutathione, covering an important role in detoxification [9–11].

CHT

Colorectal, rectosigmoid, stomach

Breast (872 in tAnGo trial, 431 in Neo-tAnGo trial)

Breast, ovarian, lung, colorectal (692 in N08Cx trial,

Abraham (2014) [66]

Adjei (2021) [67]

Not stated

Not stated

Not stated

Previous neurotoxic CHT

N08C1 trial: 2.5% of N08Cx paclitaxel and patients carboplatin; reported a N08CA trial: Grade paclitaxel; N08CB 1 baseline trial: oxaliplatin neuropathy

Paclitaxel

FOLFOX-4

Childhood acute Vincristine regimens lymphoblastic leukemia

Cancer (n of patients)

Abbasian (2020) [65]

Abaji (2018) [64]

Reference

Table 1 Summary of pharmacogenomic findings in literature

GWAS

TYMS-rs16430

WES, EWAS

Genetic study

Most patients GWAS were White, others were Black or African

Caucasian

Not stated

Caucasian

Ethnic origin

QLQ-CIPN20

NCI-CTCAE v2

CTCAE v4

CTCAE v3

CIPN assessment

CIPN patients number

(continued)

PDE6C Not stated rs56360211 had a very strong association with CIPN in N08Cx but not in the

ABCB1 rs3213619 Grades 2–4 ¼ was associated 360, grades 0–1 with a decreased ¼ 943 risk of CIPN. TUBB2A rs9501929 was associated with an increased risk of CIPN

TYMS rs16430 was Any grade ¼ 57 significantly associated with increased grade III neurotoxicity

SYNE2 rs2781377 Grades 3–4 ¼ 35 and MRPL4 -rs10513762 had an increased risk of VIPN BAHD1 rs3803357 had a protective effect against highgrade CIPN

Results

Pharmacogenomics and CIPN 97

Colorectal (86)

Colorectal (55)

Breast (8 and 228 from independent cohorts)

Antonacopoulou (2010) [69]

Apellániz-Ruiz (2015) [70]

381 in MCBDR trial)

Cancer (n of patients)

Anon (2018) [68]

Reference

Table 1 (continued)

Paclitaxel

FOLFOX-4

FOLFOX-4 or XELOX for at least 6 months

CHT

Not stated

CHT naive

Patients with pre-existing neuropathy or those who received previous chemotherapy were excluded

(NCI CTCAE v 4)

Previous neurotoxic CHT

Caucasian

Not stated

Not stated

American, Asian

Ethnic origin

Oxaliplatin-specific scale

CIPN assessment

WES

NCI-CTC v4

ITGB3 Leu59Pro TNSc

KCNN3

Genetic study

C/C ¼ 3; T/T ¼ 18; C/T ¼ 13 (any grade)

Grade 1 ¼ 33, grade 2 ¼ 33, grade 3 ¼ 20

CIPN patients number

CYP3A4 defective Grade 3 ¼ 8 variants were related to more severe CIPN (loss-of-function CYP3A4*20 and missense variants CYP3A4*25 were related to extreme CIPN)

Association with CIPN severity for T/T genotype (C/T and C/C)

None

MCBDR, while TMEM150C rs113807868 was significantly associated with CIPN in the MCBDR but not in N08Cx

Results

98 Eleonora Pozzi and Paola Alberti

Colorectal (200) FOLFOX-4, FOLFOX-6, XELOX

Colorectal (151) FOLFOX-4, XELOX

Argyriou (2013) [60]

Argyriou (2019) [73]

Not stated

CHT naive

CHT naive

Docetaxel, paclitaxel 23 previous taxane therapy (docetaxel and paclitaxel)

Breast (133)

Arbitrio (2019) [72]

Paclitaxel

Breast, ovarian (228 and 202 from independent cohorts)

Apellániz-Ruiz (2017) [71]

Not stated

Caucasian

Caucasian

Caucasian

NCI-CTC v4

CAG repeat region on the KCNN3 gene

SCN10A rs12632942 SCN10A rs6800541, SCN9A rs6746030, SCN4A rs2302237

(continued)

Acute CIPN grade 1 ¼ 43, grade 2 ¼ 34, grade 3 ¼ 53, no CIPN ¼ 21

SCN4A rs2302237 169 (acute CIPN), predictive of 145 (chronic acute and CIPN) chronic CIPN

NR1I3 Grades 2–3 ¼ 44 rs11584174 and UGT2B7 rs7438284 , rs7662029, rs7439366, rs7668258 were protective against grades 2–3 CIPN

CYP2C8 Grades 0–1 ¼ rs1058930 and 97, grades 2–3 PRX rs268674 ¼ 131 were associated with CIPN. Significantly higher CIPN risk for EPHA5/ 6/ 8 low-frequency variant carriers

OXA-neuropathy Relationship questionnaire for between the acute CIPN KCNN3 CAG TNSc for chronic repeat CIPN polymorphism and acute CIPN

TNSc, NCI-CTC v 3.0, 11-item specific oxaliplatin acute toxicity questionnaire

Identification of NCI-CTC v2 SNPs in ADME genes significantly related to the onset of grades  23 CIPN

Exon deepsequencing, GWAS

Pharmacogenomics and CIPN 99

Multiple Myeloma (33), non-Hodgkin lymphoma (12)

Breast (855)

Colorectal (40)

Ovarian (119)

Baldwin (2012) [50]

Basso (2010) [75]

Bergmann (2011) [76]

Cancer (n of patients)

Azoulay (2019) [74]

Reference

Table 1 (continued)

Paclitaxel, carboplatin

FOLFOX-4, FOLFOX-6

Paclitaxel

Vincristine, bortezomib

CHT

Caucasian, African American, Asian

Not stated

Ethnic origin

Not stated

Not stated

CHT naive and Caucasian not CHT naı¨ve (but neurotoxicity naı¨ve)

CHT naive

Not stated

Previous neurotoxic CHT

ABCB1 Ser893Ala/ Thr, CYP2C8*3 (exon 3) CYP3A5*3

NCI-CTC (version not specified)

Neurophysiological examination, acute toxicity symptoms scale

NCI-CTC v2

GWAS

CAG motif of SK3

TNSc, FACTGOG-NTx

CIPN assessment

BDNF Val66MetSNP

Genetic study

CIPN patients number

None

Sensor y neuropathy: grade 1 ¼ 58, grade 2 ¼ 20, grade 3 ¼ 3 Ataxia due to

A short CAG 28 (11, severe) repeats allele (13–15 repeats) associated with higher incidence of severe acute OXA-induced neuro-toxicity

EPHA5 rs7349683 See reference for C>T and FGD4 details rs10771973 G>A are associated with an increased probability of developing grade 2 or greater sensory peripheral neuropathy

ACT-GOG-NTx Not stated score was higher in Val-BDNF than Met-BDNF patients

Results

100 Eleonora Pozzi and Paola Alberti

NSCLC (118)

No selection for Paclitaxel cancer type, but selected among cancer patients to be treated with paclitaxel (269)

Pre-B acute lymphoblastic leukemia (1)

Booton (2006) [44]

Boora (2016) [77]

Bosilkovska (2016) [78]

Vincristine

Docetaxel +/cisplatin or carboplatin

Colorectal (346) FOLFOX-6

Boige (2010) [39]

Not stated

Not stated

CHT naive

CHT-naive and previous CHT patients (no previous neurotoxic CHT)

African

Not stated

Not stated

Not stated NCI-CTC v2

CYP3A5,ABCB1, Not stated CEP72

See reference for CIPN20 complete list of 22 SNPs tested

GSTP1 Ile105Val NCI-CTC V2

GSTP1 Ile105Val, GSTM1 deletion, ERCC1 Asn118Asn

Not stated

Grades 2/3/4 ¼ 111 (first-line treatment); grades 2/3/4 ¼ 112 (second line treatment); grades 2/3/4 ¼ 31 (third-line treatment)

CYP3A5*3/*3 (CYP3A5 non expressor)

(continued)

Total ¼ 1

EPHA5 rs7349683 Not stated and ABCB1 rs3213619 were found to be significantly associated with CIPN. CYP2C8*3 rs10509681 and rs11572080 showed a CIPN risk phenotype, while CYP1B1 rs1056836 showed a protective effect

None

None

neuropathy: grade 1 ¼ 25, grade 2 ¼ 9, grade 3 ¼ 1

Pharmacogenomics and CIPN 101

CHT naive

Not stated

Bortezomib, vincristine

Bortezomib

Multiple myeloma (329 for gene expression, 369 for SNP analysis)

Multiple myeloma (646)

Campo (2017) [81]

Previous neurotoxic CHT

Broyl (2010) [80]

CHT

Docetaxel (n ¼ 70), Not stated paclitaxel (n ¼ 43)

Cancer (n of patients)

Boso´ (2014) [79] Breast (113)

Reference

Table 1 (continued)

Caucasian

Not stated

Caucasian

Ethnic origin

CIPN assessment

GWAS

Not stated

See reference for NCI-CTC v3 complete list of genes and SNPs tested

See reference for CTCAE v4 complete list of 47 SNPs tested

Genetic study

Docetaxel: grade 0 ¼ 47 Grade 1 ¼ 20 Grade 2 ¼ 3 Paclitaxel: Grade 0 ¼ 13 Grade 1 ¼ 17 Grade 2 ¼ 12 Grade 3 ¼ 1

CIPN patients number

PSMB4 rs7172, IL17RD rs154598, EDN1 rs5370, MED20 rs2274578, BTRC rs4151060, F2 rs3136516,

Grades 0–1 ¼ 544, grades 2–4 ¼ 102

40 genes Bortezomib: differentially CIPN after one expressed in cycle grade 2–4 early- vs late¼ 20, CIPN onset CIPN in after 2–3 cycles bortezomib¼ 63; treated or vincristine: vincristineCIPN after one treated patients; cycle grade 2–4 { 59 SNPs ¼ 11, CIPN associated with after 2–3 cycles bortezomib or ¼ 17 vincristine CIPN {

Paclitaxel: ERCC1 Gln504Lys rs rs3212986 was associated with CIPN grade 2

Results

102 Eleonora Pozzi and Paola Alberti

Childhood acute Vincristine lymphoblastic leukemia (320)

Ceppi (2014) [84]

FOLFOX-4, bortezomib

Colorectal (12)

Caponigro (2009) [83]

Bortezomib

Multiple myeloma (646)

Campo (2018) [82]

Not stated

CHT-naive and previous CHT patients (no oxaliplatin based CHT)

Not stated

Caucasian

Not stated

Caucasian

TUBB1, MAP4, ACTG1, CAPG, ABCB1 and CYP3A5

Not stated

GWAS

CTCAE v3

NCI-CTC v3

Not stated

12

Grades 0–1 ¼ 547, grades 2–3 ¼ 102

(continued)

ACTG1 rs1135989 Not stated and CAPG rs2229668 were associated with higher risk of CIPN grade 3–4. CAPG

None

Chr4 rs6552496, Chr5 rs12521798, Chr16 rs8060632, Chr18 rs17748074 were associated with CIPN

IL10RA rs2229113, NFATC4 rs2228233, ABCC6 rs8058696, ABCC1 rs2384937/ rs35604, and NFATC1 rs9954562 were significantly associated with CIPN

Pharmacogenomics and CIPN 103

Colorectal (166) FOLFOX-4

Chen (2010) [86]

Paclitaxel

CHT

Breast (121)

Cancer (n of patients)

Chang (2009) [85]

Reference

Table 1 (continued)

CHT naive

CHT-naive and previous CHT patients (no previous neurotoxic CHT)

Previous neurotoxic CHT

Asian

Not stated

Ethnic origin

GSTP1 Ile105Val, ERCC1 Asn118Asn

ABCB1 Ser893Ala/ Thr

Genetic study

NCI-CTC (version not specified)

NCI-CTC v3

CIPN assessment

Grade 3 ¼ 12

CIPN patients number

Significant GSTP1 Ile105Val: association with A/A ¼ higher incidence 116, A/G and of grade 3–4 G/G ¼ CIPN for 33 (grade 0–2 GSTP1 after eight Ile105Val G/G cycles); A/A ¼ +A/G vs A/A 9, A/G and G/G ¼ 8 (grade 3–4 after eight cycles); A/A ¼ 107, A/G and G/G ¼ 26 (grade 0–2 after 12 cycles);

None

rs377010 had a protective effect against highCIPN grades, whereas ABCB1 rs4728709 against low-CIPN grades

Results

104 Eleonora Pozzi and Paola Alberti

CHT naive

Breast, ovarian, Paclitaxel, docetaxel genitourinary, gastric (35)

Childhood acute Vincristine lymphoblastic leukemia (321)

Breast (150)

Di Francia (2017) [88]

Diouf (2015) [89]

Eckhoff (2015) [48]

Docetaxel

Not stated

Custodio (2014) Colorectal (206) FOLFOX, CAPOX [87]

Not stated

Not stated

CHT naive

Cho (2010) [47] B-cell lymphoma R-CHOP (94)

Not stated

Not stated

Not stated

Not stated

Asian

GSTP1, ABCB1

GWAS

ABCB1, CYP2C83, CYP3A41B, XRCC3

ee reference selected

NCI-CTCAE v2

NCI-CTCAE v1

NCI-CTCAE v4

NCI-CTCAE v2

GSTP1 Ile105Val NCI-CTC v3 GSTM1 deletion

Grades 3–4 ¼ 2

Grades 2–4 ¼ 86

(continued)

The occurrence of Grades 0–1 ¼ grade 75, grade 2 ¼ 2 neurotoxicity 46, grade 3 ¼ was significantly 19, grade 4 ¼ more frequent in 10 patients carrying GSTP1 rs1138272 C/T or T/T

The severity of CIPN was greater for the CEP72 rs924607 risk allele (TT genotype)

XRCC3 rs1799794 Grades 0 ¼ AG+GG allele 29, grades 1–4 showed a weak ¼6 significant trend of risk of CIPN

CCNH rs2230641 Grade 1 ¼ and ABCG2 81, grades 2–3 rs3114018 were ¼ 70 associated with a higher risk of severe CIPN

None

A/A ¼ 18, A/G and G/G ¼ 15 (grade 3–4 after 12 cycles)

Pharmacogenomics and CIPN 105

Carboplatin, paclitaxel

Ovarian (112)

Colorectal (135) FOLFOX-4

Ferracini (2020) [91]

Gamelin (2007) [35]

CHT

Bortezomib

Cancer (n of patients)

Favis (2011) [90] Multiple myeloma (351)

Reference

Table 1 (continued)

CIPN assessment

AGXT Pro11Leu, NCI-CTC v1 and AGXT oxaliplatin Ile340Met, specific scale GSTP1 Ile105Val

CTCAE v5

See reference for NCI-CTC (version complete list of not specified) genes and SNPs tested

Genetic study

Non-white ¼ 8 GST, ABCB1 White ¼ 104

Not stated

Ethnic origin

CHT-naive or Caucasian previous CHT patients (no oxaliplatinbased CHT)

Not stated

CHT naive

Previous neurotoxic CHT

Grades 2–3 ¼ 35

227

CIPN patients number

Significant AGXT Pro11Leu: association with C/C ¼ CIPN severity 42, C/T and for AGXT T/T ¼ Pro11Leu C/T 12 (grade 1); and T/T vs C/C ¼ 4, C/T C/C, and for and T/T ¼ AGXT 14 (grade 2); Ile340Met A/G C/C ¼ 0, C/T and G/G vs A/A and T/T ¼ 6 (grade 3) AGXT Ile340Met: A/A ¼ 41, A/G and G/G ¼ 12 (grade 1); A/A ¼ 4, A/G and G/G ¼ 13 (grade 2); A/A ¼ 0, A/G and G/G ¼ 6 (grade 3)

ABCB1 c.3435C>T genotype had increased risk of grade 2 and 3 CIPN

Five genes differentially expressed in patients with earlier onset of CIPN{

Results

106 Eleonora Pozzi and Paola Alberti

Multiple myeloma (172)

Breast (109)

Hertz (2012) [94]

Paclitaxel

Ovarian (30), Paclitaxel, peritoneal (5), carboplatin uterus (1), cervix (1), uncertain (1)

FLO or FLP

Bortezomib or bortezomib and thalidomide

Green (2009) [93]

Goekkurt (2009) Gastric (134) [30]

Garcı´a-Sanz (2017) [92]

CHT naive

Not stated

Neurotoxic CHT was an exclusion criterion

Not stated

Caucasian AfricanAmerican, other

Caucasian

Not stated

Not stated

CYP1B1*3 (rs1056836, 4326C[G), CYP2C8*3 (rs11572080, 416G[A and rs10509681, 1196A[G), CYP3A4*1B (rs2740574, -392A[G), CYP3A5*3C (rs776746, 6986A[G), and ABCB1*2 (rs1045642,

ABCB1 Ser893Ala/Thr CYP2C8*3 (exon 3)

ERCC1 Asn118Asn, GSTM1 deletion, GSTP1 Ile105Val

GWAS, Axiom Exome Genotyping array

NCI-CTC v4

CI-CTC selfcreated questionnaire

Oxaliplatin specific scale

CTCAE v3

Grade 2 ¼ 33, grade 3 ¼ 7

CYP2C8*3 ariant revealed a trend toward increased neuropathy; no associations for other markers

12

(continued)

Patient Not stated heterozygous for CYP2C8*3 had higher risk of motor neuropathy

Significant GSTP1 Ile105Val: association with A/A ¼ 54, A/G higher incidence ¼ 46, G/G ¼ of grade 3–4 20 (grade 0–2); CIPN for A/A ¼ 10, A/G GSTP1 ¼ 0, G/G ¼ Ile105Val A/A 2 (grade 3–4) vs A/G or G/G

No SNPs significantly associated with CIPN were found

Pharmacogenomics and CIPN 107

Prostate (623)

Breast (412)

Hertz (2014) [96]

Cancer (n of patients)

Hertz (2016) [95]

Reference

Table 1 (continued)

Paclitaxel

Docetaxel, prednisone

CHT

Not stated

Not stated

Previous neurotoxic CHT

Caucasian (288), AfricanAmerican (124)

Caucasian

Ethnic origin

NCI-CTCAE v3

CIPN assessment

See reference for NCI-CTCAE v4 complete list of genes and SNPs

GWAS

3435C[T, rs2032582, 2677G[T/A, and rs1128503, 1236C[T)

Genetic study

The risk of CIPN was significantly greater in the CYP2C8 low-metabolizer patients. ABCG1 rs492338 surpassed the exploratory significance threshold for an association with CIPN in the Caucasian cohort but not in the non-Caucasian group

VAC14 rs875858 significantly increased CIPN risk

Results

Grade 2 ¼ 71

Grade 3 ¼ 50

CIPN patients number

108 Eleonora Pozzi and Paola Alberti

Colorectal (52)

Ovarian (75)

Colorectal (51)

NSCLC (62)

Multiple Myeloma (1495)

Hong (2011) [31]

Hu (2016) [97]

Inada (2010) [98]

Isla (2004) [99]

Johnson (2011) [100]

CHT naive

Not stated

CHT-naive and previous CHT patients

Vincristine, thalidomide

Not stated

Docetaxel + cisplatin Not stated

FOLFOX-6

Paclitaxel/ carboplatin

Oxaliplatin

Not stated

Not stated

Asian

Not stated

Asian

WHO criteria

NCI-CTC v3

NCI-CTC v2

NCI-CTC v3

See reference for NCI-CTC v2 complete list of (assessed only in induction phase)

ERCC1 Asn118Asn

ERCC1 Asn118Asn, GSTP1 Ile105Val

CYP3A5

ERCC1 Asn118Asn, GSTP1 Ile105Val, AGXT Ile340Met

Not stated

Five SNPs were cross-validated in two different

None

(continued)

Significant association with grade 2 CIPN

36 patients developed grade 2–4 neurological toxic effects

Grade 1 CIPN ERCC1 developed earlier Asn118Asn: in patients with C/C ¼ ERCC1 20, C/T and Asn118Asn C/T T/T ¼ and T/T vs C/C 16 (grade 1); and in those with C/C ¼ 7, C/T GSTP1 and T/T ¼ Ile105Val A/A 8 (grade 2–3) than in those GSTP1 with A/G and Ile105Val: A/A G/G, but no ¼ 27, A/G and increased risk of G/G ¼ 9 (grade grade 2 or 1); A/A ¼ higher CIPN 11, A/G and was reported G/G ¼ 4 (grade 2–3)

None

Significant STP1 Ile105Val: association with A/G and G/G higher incidence ¼ 5, A/A ¼ of grades 2–3 3 (grades 2–4) CIPN for GSTP1 Ile105Val A/G or G/G vs A/A

Pharmacogenomics and CIPN 109

Previous neurotoxic CHT

Colorectal (381) Modified FOLFOX6 Not stated

Caucasian

Ethnic origin

Asian

CHT naive and Asian previous CHT patients (no oxaliplatinbased CHT)

Kanai (2016) [102]

Modified FOLFOX6

Colorectal (82)

Kanai (2010) [40]

Not stated Platinum or platinumcombination drug chemotherapy

CHT

Lung (400)

Cancer (n of patients)

Johnson (2015) [101]

Reference

Table 1 (continued)

CIPN assessment

Oxaliplatin specific scale

See reference for NCI-CTC v1 and complete list of 2 and CTCAE genes and version 3 SNPs tested

GSTP1 Ile105Val, AGXT Pro11Leu, AGXT Ile340Met

See reference for CTCAE complete list of genes and SNPs tested

genes and SNPs tested

Genetic study

(number of patients not reported but overall patients with CIPN ¼ 446)

CIPN patients number

None

None

Grades 0–1 ¼ 204, grade 2 ¼ 147, grade 3 ¼ 30

Grade 1 ¼ 38, grade 2 ¼ 43, grade 3 ¼ 1

GPX7 rs3753753, CIPN patients ¼ ABCC4 141, non-CIPN rs1729786, patients ¼ 259 GSTA4, MGMT, XPC, RAD51, and RRMI showed significant associations with the development of CIPN

series of thalidomidetreated patients, and nine SNPs were crossvalidated in two different series vincristinetreated patients{

Results

110 Eleonora Pozzi and Paola Alberti

Gastric (73)

Keam (2008) [36]

Paclitaxel

Solid tumors (183)

Breast (119)

Komatsu (2015) [104]

Kulkarni (2015) [105]

Paclitaxel

Paclitaxel, carboplatin, docetaxel, cisplatin

Kim (2009) [45] Ovarian (118)

Modified FOLFOX6

Acute Vincristine lymphoblastic leukemia (72), solid tumors (43)

Kayiliog˘lu (2017) [103]

Modified FOLFOX

Colorectal (105), gastric (16), pancreatic (1)

Katayanagi (2019) [49]

Not stated

Not stated

CHT naive

CHT-naive and previous CHT patients

Not stated

Not stated

Most patients were Caucasian

Asian

Asian

Not stated

Not stated

Asian

RWDD3 rs2296308, TECTA rs1829

GWAS**

GSTP1 Ile105Val, ERCC1 8092C!A, GSTM1 deletion

GSTP1 Ile105Val, ERCC1 Asn118Asn

CYP3A5*1,*3, *6, *7

GSTP1

CIPN20

NCI-CTC v2

NCI-CTC v2

NCI-CTC v3

CTCAE v3

NCI-CTC v4

Grade 1/2 ¼ 12, grade 3/4 ¼ 1

Grade 3 ¼ 25

None

AIPL1 rs7214517 associated with peripheral neuropath in Asian patients

(continued)

Not stated

No CIPN:121, grade 1 ¼ 38, grade 2 ¼ 24

Higher rate of Grade 3/4 ¼ 18 grade 3/4 sensory or motor CIPN for ERCC1 8092C!A C/C genotype vs C/A or A/A

None

None

Significantly Grade 3 ¼ 10 increased onset of grade3–4 CIPN in AG type patients, compared with AA-type patients

Pharmacogenomics and CIPN 111

NSCLC (32)

Breast (219)

Breast (188)

Kus (2016) [107]

Lam (2016) [108]

Cancer (n of patients)

Kumpiro (2016) [106]

Reference

Table 1 (continued)

Not stated

Paclitaxel and bevacizumab

Not stated

Not stated

Ethnic origin

Not stated

CHT naive

Previous neurotoxic CHT

Docetaxel, paclitaxel CHT naive

Carboplatin

CHT

CYP2C8*3, CYP3A4*22, TUBB2A, FGD4, EPHA5

ABCB1 rs1045642, CYP3A4 rs2740574, ERCC1 rs3212935, ERCC2 rs13181, FGFR4 rs351855, TP53 rs1042522, ERBB2 rs1136201, CYP2C8 rs1934951

CTR1 rs12686377

Genetic study

NCI-CTCAE v3

CTCAE v4.03

CTCAE v4.03

CIPN assessment

CIPN patients number

CYP2C8*3 carriers Grade 1 ¼ 68, grade 2 ¼ had an increased risk of CIPN 41, grade 3 ¼ 17

ABCB1 TT No CIPN ¼ genotype 180, CIPN compared to TC grade 2 ¼ 39 and CC genotype, and CYP3A4 AA and AG genotype compared to GG genotype had significantly higher risk for grade 2 CIPN. For FDGFR gene with AG and GG genotype compared to AA genotype, there was significant association with regard to any grade of CIPN risk

CTR1 rs12686377 Not clearly stated was associated with CIPN

Results

112 Eleonora Pozzi and Paola Alberti

Ovarian (321)

Ovary, fallopian tube, peritoneum, lung, uterus, breast (144)

Colorectal (59), pancreas (4), stomach (2)

Lambrechts (2015) [110]

Leandro-Garcia (2013) [111]

Lecomte (2006) [27]

Lee (2014) [112] Breast (85)

NSCLC (86)

Lamba (2014) [109]

Paclitaxel

FOLFOX-4, FOLFOX-6, FOLFOX-7, GEMOX, TOMOX

Paclitaxel + carboplatin

Paclitaxel, carboplatin

Carboplatin, cisplatin

CT naive

CHT-naive and pretreated patients (no previous neurotoxic CHT)

See reference for NCI-CTC v4 complete list of 63 SNPs tested

Asian

Caucasian (59), African (4), Asian (1)

Oxaliplatin specific scale

NCI-CTC v2

See reference for Not stated complete list of 38 SNPs tested

GSTP1 Ile105Val, GSTM1 deletion

GWAS

99% Caucasian See reference for CTCAE v4 complete list of 26 SNPs selected

Caucasian

CHT naı¨ve and Caucasian not CHT naı¨ve (6)

Not stated

Not stated

See reference for details

Grade 1 ¼ 114, grade 2 ¼ 40, grade 3 ¼ 9

(continued)

For RRM1 rs9937, Not stated the AA genotype had a significantly higher incidence of CIPN compared to AG and GG genotypes

Significant GSTP1 Ile105Val: association with A/A ¼ CIPN grade 13 (grade 3), 3 severity for A/G and G/G GSTP1 ¼ 2 (grade 3) Ile105Val A/A vs A/G and G/G (p ¼ 0.02)

EPHA5rs7349683 and XKR4rs4737264 detect risk of neuropathy

None

Significant Not stated association of ABCG2 rs13120400 and XPC rs2228001 with CIPN

Pharmacogenomics and CIPN 113

Lung (39), Paclitaxel breast (38), ovarian (24), uterus (6), head and neck (4), others (7)

Gastric (85)

Acute Vincristine lymphoblastic leukemia (1103 in POG trial and 70 in

Leskela (2011) [115]

Li (2010) [32]

Li (2019) [116]

FOLFOX-4

Paclitaxel

Breast (35)

Leibovici (2018) [114]

CHT

Paclitaxel

Cancer (n of patients)

Lee (2015) [113] Breast (304), ovarian (39)

Reference

Table 1 (continued)

Not stated

CHT naive

CHT naive and previous CHT patients

Not stated

Not stated

Previous neurotoxic CHT

Caucasian

Not stated

Caucasian

Caucasian

Caucasian

Ethnic origin NCI-CTC v4

CIPN assessment

NCI-CTC v2

GWAS

Not stated

Grade 1 ¼ 38, grade 2 ¼ 86, grade 3 ¼ 166

CIPN patients number

Not stated

More severe CIPN Grade 1 ¼ severity for 29, grade 2 ¼ GSTP1 14, grade 3 ¼ Ile105Val A/A 12 vs A/G and G/G

Significant Sensory association for neuropathy: increased risk of grade 1 ¼ CIPN with 17, grade 2 ¼ polymorphisms 44, grade 3 ¼ CYP2C8*3, and 14 reduced risk for Motor CYP2C8 neuropathy: haplotype C and grade 1 ¼ CYP3A5*3 7, grade 2 ¼ 5, grade 3 ¼ 2

Met-BDNF SNP was associated only with pre-existing CIPN

None

Results

NCI-CTCAE v2 in COCH s1045466 the POG trial and and s7963521 Total were Neuropathy significantly Score-pediatric associated with Vincristine CIPN

GSTP1 Ile105Val NCI-CTC v2

CYP2C8*3, CYP2C8 haplotype C, CYP3A5*3, ABCB1 Ser893Ala/ Thr

BDNF Val66Met TNSc

CYP2C8*3

Genetic study

114 Eleonora Pozzi and Paola Alberti

Acute lymphoblastic leukemia (150)

Colorectal (2183)

Multiple myeloma (469)

Multiple myeloma (983)

Breast (58)

Ovarian (914)

Lopez-Lopez (2016) [117]

Madi (2018) [118]

Magrangeas (2016) [119]

Mahmoudpour (2018) [120]

Marcath (2020) [121]

Marsh (2007) [46]

ADVANCE trial)

Carboplatin, paclitaxel, docetaxel

Paclitaxel

Bortezomib thalidomide, vincristine

Bortezomib

Oxaliplatin

Vincristine

Not stated

Mainly Caucasian

Not stated

CHT-naive and previous CHT patients

European

European, African, Asian

Not stated

Mainly Caucasian

Not stated

Not stated

CT naive

Not stated

CYP2C8*3, CYP3A5*5, ABCB1 Ser893Ala/

PKNOX1 rs2839629 was significantly associated with CIPN

NCI-CTC v2

Grades 2–4 ¼ 148

Grade 2 ¼ 155

None

(continued)

Grade 0–1 ¼ 710, grade 2–4 ¼ 204

EPHA5 rs7349683 Not stated was associated with increased CIPN

Not stated NCI-CTC, different versions

NCI-CTC v3

EPHA4, EPHA5, CIPN8, CIPN20 EPHA6, EPHA8

GWAS

GWAS

DCLRE1A Asp317His was significantly associated with CIPN but not after Bonferroni correction

See reference for Not stated complete list of SNPs tested

Not stated

ABCC2 rs3740066 In induction GG and rs12826 phase: GG genotypes low-grade ¼ 29; were associated high-grade ¼ with increased 18 CIPN (grade In later phase: 1–4) low-grade ¼ 5; high-grade ¼ 4

See reference for WHO criteria complete list of 150 SNPs tested

(TNS-PV) in ADVANCE trial

Pharmacogenomics and CIPN 115

Cancer (n of patients)

McLeod (2010) [33]

Vincristine

CHT

Colorectal (520) FOLFOX-4, IROX

Martin-Guerrero Acute (2019) [122] lymphoblastic leukemia (133)

Reference

Table 1 (continued)

CHT-naive and previous CHT patients

Not stated

Previous neurotoxic CHT

TUBB1,TUBB3, TUBB2A, TUBB2, TUBB4, MAPT

Thr, GSTP1 Ile105Val, ERCC1 Asn118Asn

Genetic study

Caucasian GSTP1 (450), Black Ile105Val, (36), Asian ABCB1 (9), Ser893Ala/ Hispanic Thr, ERCC1 (16), Asn118Asn, Others (9) CYP3A5*3

Caucasian

Ethnic origin

NCI-CTC v2

WHO criteria

CIPN assessment

CIPN patients number

T/T genotype was more likely to discontinue FOLFOX treatment; IROX (but not FOLFOX) patients with T/T genotype had more grade 3–4 CIPN

FOLFOX discontinuation rate 24% vs 10%; 8/43 vs 0/54 IROX patients (grade 3–4)

TUBB2B In induction rs12355840 was phase: low significantly grade ¼ 28; associated with high grade ¼ 18 grades 1–2 In later phase: low CIPN and grade ¼ 5; high MAPT grade ¼ 4 s11867549 with grades 3–4 CIPN in induction but not after FDR correction

Results

116 Eleonora Pozzi and Paola Alberti

Testicular (238)

Oldenburg (2007) [29]

Cisplatin +/vinblastine not stated

Not stated

Not stated

Not stated

Colorectal (120) FOLFOX, CAPOX, COI, FOLFOXIRI

Nichetti (2019) [125]

Not stated

Not stated

Not stated

Colorectal (88)

Negarandeh (2020) [124]

Not stated

Not stated

FOLFOX, FOLFIRI Not stated

Multiple Thalidomide myeloma (82)

Mlak (2019) [123]

Docetaxel

Breast (16), non-small cell lung cancer (14), prostate (16), other (12)

Mir (2009) [43] Association with GSTP1 Ile105Val: increased A/A ¼ 17, A/G incidence of and G/G ¼ grade 2 or 29 (grade 0–1), higher CIPN for A/A ¼ 8, A/G GSTP1 and G/G ¼ Ile105Val A/A 2 (grade 2 or or A/G vs G/G higher)

NCI-CTCAE v4

NCI-CTCAE v5

GSTM1 deletion, SCIN GSTP1 Ile105Val

BCC2, ABCG2 ABCB, SLC31A1

DPYD

GSTP1 Ile105Val: A/A ¼ 23 moderate and 16 severe CIPN; A/G 29 moderate and 20 severe CIPN; A/A 10 moderate without severe CIPN

More severe longterm CIPN for GSTP1 Ile105Val A/A or A/G vs G/G

(continued)

Grades 2–3 ¼ 38

Not stated Trend toward increased grade 2–3 CIPN in patients with single SNPs but not after univariable analysis

None

CRBN NCI-CTCAE v4.03 CRBN CC FOLFOX: grade rs6768972 and genotype 1 ¼ 89; grade rs167275 (rs1672753) had 2 ¼ 69; grade more than a 3¼1 14-fold higher FOLFIRI: grade risk of CIPN 1 ¼ 16; grade 2 ¼ 20; grade 3¼0

GSTM1 deletion, NCI-CTC v2 GSTP1 Ile105Val

Pharmacogenomics and CIPN 117

Colorectal (70)

Digestive tract (228)

Palugulla (2017) [63]

Cancer (n of patients)

Oguri (2013) [41]

Reference

Table 1 (continued)

CAPOX, EOX, FOLFOX, GEMOX

FOLFOX-6

CHT

Ethnic origin

CHT naive

Asian

CHT naı¨ve and Asian not CHT naive (not stated)

Previous neurotoxic CHT CIPN assessment

SCNA

NCI-CTC v4.03

ERCC1 C118T NCI-CTC v3 rs11615, GSTP1 Ile105Val rs1695, TAC1, FOXC1, ITGA1, ACYP2, DLEU7, BTG4, CAMK2N1, FARS2

Genetic study

CIPN patients number

SCN9A rs6746030 Grade 1 ¼ was significantly 73, grade 2 ¼ associated with a 55, grade 3 ¼ 2 higher incidence of chronic CIPN, while the rs6754031 variant was linked with a lower incidence. The SCN10A polymorphic variant was

Association for 48 severity for A/G and G/G genotypes of rs17140129 in FARS2, ERCC1 C118T rs11615 related to time of onset, rs10486003 C/C genotype related to shorter onset, negative association for GSTP1 Ile105Val rs1695

Results

118 Eleonora Pozzi and Paola Alberti

Ovarian (454)

Breast (95)

Park (2017) [128]

Rizzo (2010) [129]

Paclitaxel, docetaxel

Paclitaxel/ carboplatin

CHT-naive and previous CHT patients

Not stated

Not stated

Paclitaxel

Breast (21)

CHT naive

Park (2014) [127]

GEMOX, Modified FOLFOX-4, CAPOX, EOX

CHT naive

Digestive tract (228)

Pare` (2008) [37] Colorectal (126) FOLFOX-4

Palugulla (2018) [126]

Caucasian

Caucasian

Not stated

Not stated

Asian Questionnaire with a yes/no response format

NCI-CTCAE, TNSc

ABCB1 Ser893Ala/ Thr, CYP2C8*3

NCI-CTC v3

MAPT rs242557 NCI-CTCAE v3 and rs1052553, GSK3β rs6438552 and rs3755557, CEP72 rs924607, TUBB2 rs909961

MAPT haplotype 1 and rs242557, GSK3β rs6438552

GSTP1 Ile105Val Oxaliplatin-specific scale

GSTP1 rs1965, ABCG2 rs3114018, CCNH rs2230641, rs309381, AGXT rs4426527

None

MAPT additive polymorphisms were associated with patientreported CIPN, and GSK3β additive polymorphisms were associated with clinician reported CIPN

Significant relationship between the GSK-3β rs6438552 and CIPN

None

(continued)

Grade 1 ¼ 1, grade 2¼6

Grades 1–2 ¼ 96, grade 3 ¼ 352

Grade 1 ¼ 56%, grade 2 ¼ 3%1, grade 3 ¼ 13%

Grade 1 ¼ 42, grade 2 ¼ 66, grade 3 ¼ 5

CCNH rs2230641 Grade 1 ¼ and rs3093816 71 grade 2 ¼ were 35, grade 3 ¼ 5 significantly associated with the incidence and severity of acute CIPN

associated with severity of CIPN

Pharmacogenomics and CIPN 119

Colorectal (166) FOLFOX-4

Colorectal (517) FOLFOX-4, XELOX

B-cell lymphoma R-CHOP (56)

Ruzzo (2007) [28]

Ruzzo (2014) [131]

Sawaki (2020) [132]

Previous neurotoxic CHT

Not stated

Not stated

CHT-naive and previous CHT patients

Docetaxel, paclitaxel CHT naive

CHT

Breast (176)

Cancer (n of patients)

Rua (2018) [130]

Reference

Table 1 (continued)

Asian

Caucasian

Not stated

Caucasian

Ethnic origin

CIPN assessment

Oxaliplatin specific scale

CEP72 rs924607, ETAA1 rs17032980, MTNR1B rs12786200, CYP3A5 rs776746, -rs7963521

NCI-CTCAE v3 and v4

TS, MTHFR, NCI-CTCAE v2 ERCC1, XRCC1, XRCC3, XPD, GSTT1, GSTP1, GSTM1, ABCC1, ABCC2

ERCC1 Asn118Asn, GSTP1, Ile105Val, GSTM1 deletion

CAG motif repeat NCI-CTCAE of KCNN3

Genetic study

CIPN patients number

None

None

Significant association with CIPN severity GSTP1 Ile105Val G/G>A/G>A/ A (pT SNP

NCI-CTCAE v4

CIPN assessment

AUC area under the curve, CAPOX capecitabine, oxaliplatin, COI capecitabine, oxaliplatin, CIPN chemotherapy-induced peripheral neurotoxicity, CHT chemotherapy, CTCAE Common Terminology Criteria for Adverse Events, EOX Epirubicin, capecitabine, oxaliplatin, FLO fluorouracil, leucovorin, oxaliplatin. FLP fluorouracil, leucovorin, cisplatin, FOLFOX leucovorin, fluorouracil, oxaliplatin, FOLFOXIRI/FOLFIRINOX 5-fluorouracil, leucovorin, oxaliplatin, irinotecan, GEMOX gemcitabine, oxaliplatin, GWAS** Genome-wide association study, IROX irinotecan, oxaliplatin, NCI-CTC National Cancer Institute common toxicity criteria, NSCLC non-small-cell lung cancer, PAD bortezomib, epirubicin, dexamethasone, PCD bortezomib, cyclophosphamide, dexamethasone, PTD bortezomib, thalidomide, dexamethasone, R-CHOP rituximab, cyclophosphamide, doxorubicin, vincristine, prednisone, SCIN scale for chemotherapy-induced neurotoxicity, SNPs single nucleotide polymorphisms, TNSc Total Neuropathy Score, clinical version, XELOX capecitabine, oxaliplatin, TOMOX raltitrexed, oxaliplatin, WES whole exome sequencing, {Retrospective study; {None of the reported associations refer to other targets described in this table. § Leu11 allele was not found in this population of Japanese patients

Vincristine

CHT

Acute lymphoblastic leukemia (133)

Cancer (n of patients)

Zgheib (2018) [147]

Reference

Table 1 (continued)

126 Eleonora Pozzi and Paola Alberti

Pharmacogenomics and CIPN

127

• GSTM1 and GSTM3 genes: they are part of the μ class of glutathione S-transferases crucial for detoxification through glutathione conjugation of electrophilic compounds, including carcinogens, therapeutic drugs, environmental toxins, and products of oxidative stress. Genetic variations can change an individual’s susceptibility to carcinogens and toxins and affect toxic effects and effectiveness of specific drugs [9, 10, 12, 13]. • ERCC1 gene: Excision repair cross-complementing group 1 (ERCC1) is part of the nucleotide excision-repair pathway and is required for repair of DNA lesions, such as those induced by ultraviolet light or formed by electrophilic compounds including cisplatin. Polymorphisms that alter expression of ERCC1 might have a role in carcinogenesis, and this gene has been investigated extensively for its role in cancer cell resistance to platinum drugs [14]. • AGXT gene: Alanine-glyoxylate aminotransferase (AGXT) prevents accumulation of glyoxylate in the cytosol by converting it into glycolate, which is subsequently metabolized into oxalate by lactate dehydrogenase [15]. Because of this role in oxalate metabolism, the AGTX Ile340Met polymorphism (rs4426527; NP_000021.1) has been investigated in two studies of patients with colorectal cancer treated with oxaliplatin. • ABCB1 gene: they are part of the ATP-binding cassette proteins; they transport various molecules across extracellular and intracellular membranes. The membrane-associated protein ABCB1 (also known as P-gp or MDR1) is part of the MDR/TAP subfamily that reduces drug accumulation in multidrug-resistant cells [16–18]. Paclitaxel and docetaxel are known substrates of ABCB1-mediated efflux from cancer cells, and the ABCB1 Ser893Ala and Ser893Thr SNPs (rs2032582; NP_000918.2) have been investigated in patients treated with taxanes. • CYP2C8 and CYP3A5 genes: The cytochrome P450 superfamily is a large group of enzymes that are involved in organic substances and xenobiotic oxidation. They are the major enzymes of drug metabolism and bioactivation, accounting for about 75% of total metabolic reactions [19, 20]. Although no clear data are available for their distribution and activity in the peripheral nervous system, their possible contribution to development of CIPN due to altered pharmacokinetics of taxanes has been hypothesized; in particular the enzymes CYP2C8 and CYP3A5 help eliminate paclitaxel through successive hydroxylation reactions [21]. • ITGB3 gene: Integrin B3 (ITGB3) belongs to the large family of integrins, which are integral cell-surface proteins composed of an α and β chain and known to participate in cell adhesion and

128

Eleonora Pozzi and Paola Alberti

cell-surface-mediated signaling. The ITGB3 Leu59Pro polymorphism (rs5918; NP_000203.2) has been associated with different activation of the MAPK3 and MAPK1 subgroup of mitogen-activated protein kinases, and reduced activation of MAPK3 and MAPK1 has been seen in in vitro models of neurotoxic effects of platinum drugs [22]. Studies reported in literature regarding these most represented targets are quite conflicting as extensively reported in Table 1. We will focus our attention on two different targets to highlight possible pitfalls and strength in addressing CIPN pharmacogenomics, GSTP1 which is an example of the old-fashioned approach, and in the next section we will present a different point of view which can solve the matter, focused on sodium-voltage operated channel genes. 2.1 The GSTP1 Gene Saga: Starting Point Away from Peripheral Nervous System

Glutathione S-transferases are a family of enzymes that catalyze conjugation of many hydrophobic and electrophilic compounds with reduced glutathione: they are of primary importance in detoxification. GSTP1 (glutathione S-transferase P1) belongs to the π class and plays a part in detoxification of platinum drugs [9–11]. A SNP in GSTP1 (562A!G; rs1695; NM_000852.3) causing substitution of isoleucine for valine diminishes the enzyme’s activity, whereas homozygous deletion of the entire gene abolishes its action; this might be relevant to CIPN due to the role of oxidative stress in this disorder’s onset and course [23–26]. The GSTP1 Ile105Val SNP (rs1695; NP_000843.1) were investigated in relation to peripheral neurotoxicity of platinum drugs in several studies. In some of these, a positive association with CIPN was reported; instead, in others, no correlative evidence was recorded. In 2006, Lecomte and colleagues [27] described a cohort of white European, African, and Asian patients with colorectal, pancreatic, or gastric cancer, who were receiving treatment with oxaliplatinbased chemotherapy. In these individuals, a significant association was noted between occurrence of the A/A genotype and grade 3 CIPN, assessed with an oxaliplatin-specific scale. The number of patients available for comparison was very low, and the result was soon challenged by findings of a study undertaken in a larger cohort of patients affected by colorectal cancer and treated with oxaliplatin. The result of Lecomte’s work was made more unclear when, in another study [28], a correlation with CIPN severity was established, but with the G rather than the A allele. Subsequently, Oldenburg and co-workers [29] used a symptom questionnaire to identify an association between more severe long-term CIPN and the A/A or A/G versus G/G genotype. In a study done in 2009 in 134 patients with gastric cancer, the A/A genotype was associated significantly with grade 3 CIPN, although only 12 patients were eventually available for comparison [30]. In 2010, a significant

Pharmacogenomics and CIPN

129

association was described in three independent studies between the A/A genotype and either more severe CIPN [31, 32] or earlier onset of grade 1 CIPN scored with the National Cancer Institute’s common toxicity criteria (NCI-CTC) without any effect on development of more severe CIPN grades; however, as in previous work, only a few patients were available for comparison. In the largest study reported so far in individuals with colorectal cancer treated with oxaliplatin, McLeod and colleagues [33] studied two GSTP1 SNPs—Ile105Val and Ala114Val (rs1138272; NP_000843.1 [590C!T, NM_000852.3]). They stated that the T/T genotype was a predictor for more frequent discontinuation of FOLFOX (leucovorin, fluorouracil, oxaliplatin) and for more severe CIPN after treatment with irinotecan and oxaliplatin (but, rather surprisingly, not FOLFOX). Instead, Hong and collaborators [31] observed a more pronounced risk for grade 2 oxaliplatin-related neurotoxicity for A/A genotype (NCI-CTC was applied for grading). The Ala114Val SNP was not associated with CIPN in a study by Khrunin and co-workers [34]. In disagreement with these positive results, negative findings of an investigation of the GSTP1 Ile105Val SNP in patients treated with a platinum drug [35] were replicated in several subsequent studies [36–38]. In 2010, four independent clinical trials undertaken in individuals affected either by colorectal cancer and treated with oxaliplatin-based chemotherapy [13, 39, 40] or by ovarian cancer and treated with cisplatin [34] also had negative results. These prospective data accorded with those described in a previous retrospective study done in a mixed population of white European, Hispanic, Asian, and African individuals with colorectal cancer treated with oxaliplatin [12]. Also, Oguri and colleagues [41] in 2013 reported a negative association for GSTP1 Ile105Val rs1695 polymorphism with oxaliplatinrelated CIPN in a cohort of Asian patients. A genome-wide association study (GWAS) performed by Won and colleagues [42] in a population of 247 patients treated with oxaliplatin showed that rs1695 in GSTP1 had no association with oxaliplatin-related CIPN. Rates of the GSTP1 Ile105Val polymorphism were also investigated in patients treated with docetaxel [43], taxanes and platinum regimens [44–46], and vincristine [47]. In these studies, the A/A genotype was associated positively with a higher incidence of NCI-CTC grade 2 CIPN only in docetaxel-treated patients, whereas no association was reported in those treated with taxanes and carboplatin regimens or with vincristine. Eckhoff and collaborators [48] explored the role of GSTP1 in 150 breast cancer patients treated with docetaxel and found a significant association with grade 2 neurotoxicity with rs1138272 C/T or T/T. Katayanagi and colleagues [49] analyzed data from nearly 150 gastrointestinal neoplasm cancer patients and found an increased risk of higher CIPN grade in patients carrier of the variant AG in GSTP1.

130

Eleonora Pozzi and Paola Alberti

In summary, the GSTP1 example embodies why definitive inferences cannot be inferred from data reported in literature up to now. This quite confusing picture, so well depicted for GSTP1, can be extended to all candidate genes enlisted in Table 1. Several reasons might explain these findings. The first one is related to CIPN assessment and population selection/data collection. As it was yet said, there is still not a consensus on the best method(s) to measure neurological side effects due to chemotherapy. Different selected neurological endpoints were applied among studies reported here: comparison among different works is quite difficult as a consequence. In the major part, NCI-CTC was the outcome measure: it is well-known that it is a poor instrument to graduate CIPN correctly, not containing detailed neurological examinations [8]. Another important issue is that some studies applied a selfcreated scale (without a previous proper validation). Also, incidence, type, and severity of other risk factors for peripheral nerve damage were rarely assessed (e.g., diabetes, alcohol intake, pre-existent neuropathy due to other causes than chemotherapy [CHT]); moreover, several studies included a mixed population at baseline (CHT naı¨ve and not CHT naı¨ve), without a clear basal neurological evaluation before starting CHT, even if the previous CT regimen was a neurotoxic one. Rarely precise data on the actual cumulative dose administered were recorded, making it difficult to ascertain if cohorts in different studies were exposed to comparable amounts of the neurotoxic drug. Another aspect to be pointed out is that even if sample size is quite considerable at baseline, in many studies, the subgroup of patients with CIPN was remarkably sparse, making possible that statistics were underpowered. Second, this review makes clear that a shift in the gene selection strategy is crucial to obtain valuable results from pharmacogenomics. Genes related to neurons would be the right choice to be employed in regard to specific side effects as CIPN is. Apart from delusive data about genes such as GSTP1, first inferences from GWAS support this point of view; it is becoming clear that factors contributing to the function and repair of peripheral nerves are more important than alterations in pharmacokinetics for determining genetic susceptibility to this toxicity. An extensive example of this was given by Baldwin and colleagues in 2012 [50] in a GWAS aimed to identify new loci for paclitaxel-related CIPN in a cohort of breast cancer patients. In their work, it is interesting to note that previously tested target genes were not associated with neuropathy development; differently, it was observed a positive association with a marker related to neurons. It is the case of FGD4: it encodes for the protein FGD1-related F-actin-binding protein (Frabin), and previous studies have shown specific point mutations in FGD4 can cause the congenital peripheral neuropathy CMT (CMT4H) [51–54]. Frabin is a guanine nucleotide exchange factor for cdc42, a Rho-GTPase that regulates cellular morphogenesis, including

Pharmacogenomics and CIPN

131

myelination [51]. The observed association between the FGD4 SNP rs10771973 and paclitaxel-induced sensory peripheral neuropathy is consistent with the hypothesis that common FGD4 polymorphisms subtly affect the development and/or maintenance of Schwann cell function. In this case, carriers of common FGD4 polymorphisms would have pre-existing subclinical abnormalities and a predisposition for toxicity [55]. Alternatively, authors suggested that FGD4 polymorphisms could lead to impaired repair processes such as Schwann cell re-myelination and/or axonal regeneration after paclitaxel exposure. Obviously, more extensive studies are needed, but this observation confirms that maybe we should move toward such a target as this. 2.2 An Opposite Approach: Starting from Peripheral Nervous System

More informative studies can be performed when the starting point is the peripheral nervous system itself, by selecting grading tools more apt to actually measure neuropathy and selecting target genes related to the pathophysiology of peripheral nerves. For what regard CIPN assessment, even if its measurement is still a matter of debate, the CI-PeriNomS [56] study gave precious information. First validity and reliability findings were obtained for a set of outcome measures in CIPN detection and grading, among which the most notable is the Total Neuropathy Score scale, clinical version (TNSc®) [57]. Apart from being valid and reliable, it is a scale that either comprehends a neurological examination, in contrast with NCI-CTC [58], and is simple and rapid enough to be applied in an everyday setting, in contrast with other neurological scales tested as mIss [modified Incat Sensory Score® (55)]; therefore, TNSc® seems an appropriate instrument applicable even in large cohorts of patients to be correctly described for a further pharmacogenomics analysis. In CI-PeriNomS also patient-reported outcome measures (PRO) and pain assessment were suggested for a full CIPN evaluation; these tools enable to encompass the wholeness of this phenomenon: since it consists mainly of sensory alterations, subjective point of view and pain evaluation cannot be underestimated. They should also be integrated to precisely describe the study population. Very recently, the CI-PeriNomS group published a novel paper calculating for the first time the minimal clinical important difference of TNSc® and its shorter “nurse” version, making even more robust the suggestion to use this tool in CIPN trials [59]. • The application of these tools in CIPN grading is embodied in a work published by Argyriou and colleagues that not only relied on a solid CIPN assessment but also decided to explore a target gene family directly related to nerve functioning: SCNA genes. The voltage-gated sodium channels (SCNA) are key in the initiation and propagation of action potentials in electrically excitable cells. They are membrane proteins and are encoded

132

Eleonora Pozzi and Paola Alberti

by different genes. Mutations in SCNAs are known to cause diseases in the central and peripheral nervous system (PNS). Focusing on mutations related to peripheral nerves and muscle dysfunction, we can enlist SCN4A gene, expressed in the skeletal muscle, whose mutations have been linked to several myotonias and periodic paralysis disorders. The integral membrane protein encoding the SCN5A gene is found primarily in the cardiac muscle, and defects in this gene cause autosomaldominant cardiac disease; SCN9A gene, which plays a significant role in nociception signaling, has been related to insensitivity to pain and paroxysmal extreme pain disorder; SCN10A is hypothesized to be involved in painful neuropathies [60]. Argyriou et al. [60] aimed at identifying SNPs of SCNAs genes. There are, in fact, robust preclinical studies suggesting that a sodium channelopathy might lead to oxaliplatin peripheral neurotoxicity [61, 62]. Argyriou and colleagues used the TNSc® as the pivotal element to assess CIPN. A total of 200 patients with CRC were genotyped. SCN4A-rs2302237 emerged as being predictive of the clinical severity of neuropathy. The results of the study need to be further confirmed in larger cohorts, as authors state, however, it is a convincing example of the methodology that should be applied: a thoughtful selection of genes of interest and a careful, yet simple, valid, and reproducible, neurological assessment of subjects. In line with these observations, the role of SCNA genes was more recently confirmed by Palugulla and collaborators [63] SCN9A who observed a significant association of SCN9A rs6746030 with a higher incidence of chronic CIPN.

3

Future Perspectives: Crosstalk Between Bench and Bedside From the overview here presented, it can be concluded that for future studies, a golden rule could be proposed: the bench side (pharmacogenomics) should be employed rigorously as rigorously the bedside (clinimetrics) should be managed (see Fig. 1): the two sides of the same coin should cooperate. This methodological strategy can thus be proposed: • Patients should be evaluated in a refined way. Tools like TNSc® are to be applied. Population should be clearly stratified at baseline for pre-existing and/or co-existent risk factors for neuropathy development. Precise actual cumulative dose data should be analyzed. Sample size should consider the real amounts of patients that have developed neurological signs/ symptoms. Moreover, patient-reported outcome measures (PRO) and pain assessment should be considered.

Pharmacogenomics and CIPN

133

Fig. 1 Virtuous crosstalk between bench and bedside in CIPN research

• Genes related to the peripheral nervous system and CIPN pathogenesis are to be investigated. Mechanisms of peripheral nerve damage in CIPN are not yet completely understood; however, hints from preclinical animal models could be considered as a guidance. GWAS might also be helpful to identify new potential predictors of neuropathy development. In conclusion, a very powerful instrument as pharmacogenomics, being it delusive so far, is the confirmation that an appropriate clinimetric approach to CIPN is warranted: i.e., appropriate, valid, reliable, and feasible clinical instrument(s) for its measure are needed. That would probably be the key to building up a strategy to contrast an adverse effect that negatively impacts QoL in a large population of long survival cancer patients. References 1. Syva¨nen AC (2005) Toward genome-wide SNP genotyping. Nat Genet 37(Suppl):S5– S10 2. Velasco R, Alberti P, Bruna J et al (2019) Bortezomib and other proteosome inhibitors-induced peripheral neurotoxicity: From pathogenesis to treatment. J Peripher Nerv Syst 24(Suppl 2):S52–S62

3. Islam B, Lustberg M, Staff NP et al (2019) Vinca alkaloids, thalidomide and eribulininduced peripheral neurotoxicity: From pathogenesis to treatment. J Peripher Nerv Syst 24 (Suppl 2):S63–S73 4. Tamburin S, Park SB, Alberti P et al (2019) Taxane and epothilone-induced peripheral neurotoxicity: From pathogenesis to

134

Eleonora Pozzi and Paola Alberti

treatment. J Peripher Nerv Syst 24(Suppl 2): S40–S51 5. Alberti P (2019) Platinum-drugs induced peripheral neurotoxicity: clinical course and preclinical evidence. Expert Opin Drug Metab Toxicol 15(6):487–497 6. Park SB, Alberti P, Kolb NA et al (2019) Overview and critical revision of clinical assessment tools in chemotherapy-induced peripheral neurotoxicity. J Peripher Nerv Syst 24(Suppl 2):S13–S25 7. Argyriou AA, Park SB, Islam B et al (2019) Neurophysiological, nerve imaging and other techniques to assess chemotherapy-induced peripheral neurotoxicity in the clinical and research settings. J Neurol Neurosurg Psychiatry 8. Cavaletti G, Frigeni B, Lanzani F et al (2010) Chemotherapy-Induced Peripheral Neurotoxicity assessment: a critical revision of the currently available tools. Eur J Cancer 46(3): 479–494 9. Ekhart C, Rodenhuis S, Smits PH et al (2009) An overview of the relations between polymorphisms in drug metabolising enzymes and drug transporters and survival after cancer drug treatment. Cancer Treat Rev 35(1): 18–31 10. Strange RC, Spiteri MA, Ramachandran S et al (2001) Glutathione-S-transferase family of enzymes. Mutat Res 482(1-2):21–26 11. Funke S, Timofeeva M, Risch A et al (2010) Genetic polymorphisms in GST genes and survival of colorectal cancer patients treated with chemotherapy. Pharmacogenomics 11(1):33–41 12. Stoehlmacher J, Park DJ, Zhang W et al (2002) Association between glutathione S-transferase P1, T1, and M1 genetic polymorphism and survival of patients with metastatic colorectal cancer. J Natl Cancer Inst 94(12):936–942 13. Zarate R, Rodrı´guez J, Bandres E et al (2010) Oxaliplatin, irinotecan and capecitabine as first-line therapy in metastatic colorectal cancer (mCRC): a dose-finding study and pharmacogenomic analysis. Br J Cancer 102(6): 987–994 14. Martin LP, Hamilton TC, Schilder RJ (2008) Platinum resistance: the role of DNA repair pathways. Clin Cancer Res 14(5):1291–1295 15. Se´gurel L, Lafosse S, Heyer E et al (2010) Frequency of the AGT Pro11Leu polymorphism in humans: Does diet matter? Ann Hum Genet 74(1):57–64 16. Levi F, Perpoint B, Garufi C et al (1993) Oxaliplatin activity against metastatic

colorectal cancer. A phase II study of 5-day continuous venous infusion at circadian rhythm modulated rate. Eur J Cancer 29A (9):1280–1284 17. Crowley ECallaghan R (2010) Multidrug efflux pumps: drug binding--gates or cavity? FEBS J 277(3):530–539 18. Sharom FJ (2008) ABC multidrug transporters: structure, function and role in chemoresistance. Pharmacogenomics 9(1):105–127 19. Henningsson A, Marsh S, Loos WJ et al (2005) Association of CYP2C8, CYP3A4, CYP3A5, and ABCB1 polymorphisms with the pharmacokinetics of paclitaxel. Clin Cancer Res 11(22):8097–8104 20. Zhou SF, Liu JP, Chowbay B (2009) Polymorphism of human cytochrome P450 enzymes and its clinical impact. Drug Metab Rev 41(2):89–295 21. Mielke S, Sparreboom A, Steinberg SM et al (2005) Association of Paclitaxel pharmacokinetics with the development of peripheral neuropathy in patients with advanced cancer. Clin Cancer Res 11(13):4843–4850 22. Scuteri A, Galimberti A, Maggioni D et al (2009) Role of MAPKs in platinum-induced neuronal apoptosis. Neurotoxicology 30(2): 312–319 23. Durand JP, Deplanque G, Montheil V et al (2012) Efficacy of venlafaxine for the prevention and relief of oxaliplatin-induced acute neurotoxicity: results of EFFOX, a randomized, double-blind, placebo-controlled phase III trial. Ann Oncol 23(1):200–205 24. Argyriou AA, Chroni E, Koutras A et al (2006) Preventing paclitaxel-induced peripheral neuropathy: a phase II trial of vitamin E supplementation. J Pain Symptom Manage 32(3):237–244 25. Argyriou AA, Chroni E, Koutras A et al (2006) A randomized controlled trial evaluating the efficacy and safety of vitamin E supplementation for protection against cisplatininduced peripheral neuropathy: final results. Support Care Cancer 14(11):1134–1140 26. Pace A, Giannarelli D, Galie` E et al (2010) Vitamin E neuroprotection for cisplatin neuropathy: a randomized, placebo-controlled trial. Neurology 74(9):762–766 27. Lecomte T, Landi B, Beaune P et al (2006) Glutathione S-transferase P1 polymorphism (Ile105Val) predicts cumulative neuropathy in patients receiving oxaliplatin-based chemotherapy. Clin Cancer Res 12(10):3050–3056 28. Ruzzo A, Graziano F, Loupakis F et al (2008) Pharmacogenetic profiling in patients with advanced colorectal cancer treated with first-

Pharmacogenomics and CIPN line FOLFIRI chemotherapy. Pharmacogenomics J 8(4):278–288 29. Oldenburg J, Kraggerud SM, Brydøy M et al (2007) Association between long-term neuro-toxicities in testicular cancer survivors and polymorphisms in glutathione-s-transferase-P1 and -M1, a retrospective cross sectional study. J Transl Med 5:70 30. Goekkurt E, Al-Batran SE, Hartmann JT et al (2009) Pharmacogenetic analyses of a phase III trial in metastatic gastroesophageal adenocarcinoma with fluorouracil and leucovorin plus either oxaliplatin or cisplatin: a study of the arbeitsgemeinschaft internistische onkologie. J Clin Oncol 27(17):2863–2873 31. Hong J, Han SW, Ham HS et al (2011) Phase II study of biweekly S-1 and oxaliplatin combination chemotherapy in metastatic colorectal cancer and pharmacogenetic analysis. Cancer Chemother Pharmacol 67(6): 1323–1331 32. Li QF, Yao RY, Liu KW et al (2010) Genetic polymorphism of GSTP1: prediction of clinical outcome to oxaliplatin/5-FU-based chemotherapy in advanced gastric cancer. J Korean Med Sci 25(6):846–852 33. McLeod HL, Sargent DJ, Marsh S et al (2010) Pharmacogenetic predictors of adverse events and response to chemotherapy in metastatic colorectal cancer: results from North American Gastrointestinal Intergroup Trial N9741. J Clin Oncol 28(20): 3227–3233 34. Khrunin AV, Moisseev A, Gorbunova V et al (2010) Genetic polymorphisms and the efficacy and toxicity of cisplatin-based chemotherapy in ovarian cancer patients. Pharmacogenomics J 10(1):54–61 35. Gamelin L, Capitain O, Morel A et al (2007) Predictive factors of oxaliplatin neurotoxicity: the involvement of the oxalate outcome pathway. Clin Cancer Res 13(21):6359–6368 36. Keam B, Im SA, Han SW et al (2008) Modified FOLFOX-6 chemotherapy in advanced gastric cancer: Results of phase II study and comprehensive analysis of polymorphisms as a predictive and prognostic marker. BMC Cancer 8:148 37. Pare´ L, Marcuello E, Alte´s A et al (2008) Pharmacogenetic prediction of clinical outcome in advanced colorectal cancer patients receiving oxaliplatin/5-fluorouracil as firstline chemotherapy. Br J Cancer 99(7): 1050–1055 38. Seo BG, Kwon HC, Oh SY et al (2009) Comprehensive analysis of excision repair complementation group 1, glutathione S-transferase,

135

thymidylate synthase and uridine diphosphate glucuronosyl transferase 1A1 polymorphisms predictive for treatment outcome in patients with advanced gastric cancer treated with FOLFOX or FOLFIRI. Oncol Rep 22(1): 127–136 39. Boige V, Mendiboure J, Pignon JP et al (2010) Pharmacogenetic assessment of toxicity and outcome in patients with metastatic colorectal cancer treated with LV5FU2, FOLFOX, and FOLFIRI: FFCD 2000-05. J Clin Oncol 28(15):2556–2564 40. Kanai M, Yoshioka A, Tanaka S et al (2010) Associations between glutathione S-transferase pi Ile105Val and glyoxylate aminotransferase Pro11Leu and Ile340Met polymorphisms and early-onset oxaliplatininduced neuropathy. Cancer Epidemiol 34(2):189–193 41. Oguri T, Mitsuma A, Inada-Inoue M et al (2013) Genetic polymorphisms associated with oxaliplatin-induced peripheral neurotoxicity in Japanese patients with colorectal cancer. Int J Clin Pharmacol Ther 51(6): 475–481 42. Won HH, Lee J, Park JO et al (2012) Polymorphic markers associated with severe oxaliplatin-induced, chronic peripheral neuropathy in colon cancer patients. Cancer 118(11):2828–2836 43. Mir O, Alexandre J, Tran A et al (2009) Relationship between GSTP1 Ile(105)Val polymorphism and docetaxel-induced peripheral neuropathy: clinical evidence of a role of oxidative stress in taxane toxicity. Ann Oncol 20(4):736–740 44. Booton R, Booten R, Ward T et al (2006) Glutathione-S-transferase P1 isoenzyme polymorphisms, platinum-based chemotherapy, and non-small cell lung cancer. J Thorac Oncol 1(7):679–683 45. Kim HS, Kim MK, Chung HH et al (2009) Genetic polymorphisms affecting clinical outcomes in epithelial ovarian cancer patients treated with taxanes and platinum compounds: a Korean population-based study. Gynecol Oncol 113(2):264–269 46. Marsh S, Paul J, King CR et al (2007) Pharmacogenetic assessment of toxicity and outcome after platinum plus taxane chemotherapy in ovarian cancer: the Scottish Randomised Trial in Ovarian Cancer. J Clin Oncol 25(29):4528–4535 47. Cho HJ, Eom HS, Kim HJ et al (2010) Glutathione-S-transferase genotypes influence the risk of chemotherapy-related toxicities and prognosis in Korean patients with

136

Eleonora Pozzi and Paola Alberti

diffuse large B-cell lymphoma. Cancer Genet Cytogenet 198(1):40–46 48. Eckhoff L, Feddersen S, Knoop AS et al (2015) Docetaxel-induced neuropathy: a pharmacogenetic case-control study of 150 women with early-stage breast cancer. Acta Oncol 54(4):530–537 49. Katayanagi S, Katsumata K, Mori Y et al (2019) as a potential predictive factor for adverse events associated with platinumbased antitumor agent-induced peripheral neuropathy. Oncol Lett 17(3):2897–2904 50. Baldwin RM, Owzar K, Zembutsu H et al (2012) A genome-wide association study identifies novel loci for paclitaxel-induced sensory peripheral neuropathy in CALGB 40101. Clin Cancer Res 18(18):5099–5109 51. Delague V, Jacquier A, Hamadouche T et al (2007) Mutations in FGD4 encoding the Rho GDP/GTP exchange factor FRABIN cause autosomal recessive Charcot-Marie-Tooth type 4H. Am J Hum Genet 81(1):1–16 52. Stendel C, Roos A, Deconinck T et al (2007) Peripheral nerve demyelination caused by a mutant Rho GTPase guanine nucleotide exchange factor, frabin/FGD4. Am J Hum Genet 81(1):158–164 53. Fabrizi GM, Taioli F, Cavallaro T et al (2009) Further evidence that mutations in FGD4/ frabin cause Charcot-Marie-Tooth disease type 4H. Neurology 72(13):1160–1164 54. Houlden H, Hammans S, Katifi H et al (2009) A novel Frabin (FGD4) nonsense mutation p.R275X associated with phenotypic variability in CMT4H. Neurology 72(7):617–620 55. Lee JJ, Swain SM (2006) Peripheral neuropathy induced by microtubule-stabilizing agents. J Clin Oncol 24(10):1633–1642 56. Cavaletti G, Cornblath DR, Merkies IS et al (2013) The chemotherapy-induced peripheral neuropathy outcome measures standardization study: from consensus to the first validity and reliability findings. Ann Oncol 24(2):454–462 57. Cavaletti G, Frigeni B, Lanzani F et al (2007) The Total Neuropathy Score as an assessment tool for grading the course of chemotherapyinduced peripheral neurotoxicity: comparison with the National Cancer Institute-Common Toxicity Scale. J Peripher Nerv Syst 12(3): 210–215 58. Trotti A, Colevas AD, Setser A et al (2003) CTCAE v3.0: development of a comprehensive grading system for the adverse effects of cancer treatment. Semin Radiat Oncol 13(3): 176–181

59. Alberti P, Bernasconi DP, Cornblath DR et al (2021) Prospective Evaluation of Health Care Provider and Patient Assessments in Chemotherapy-Induced Peripheral Neurotoxicity. Neurology 97(7):e660–e672 60. Argyriou AA, Cavaletti G, Antonacopoulou A et al (2013) Voltage-gated sodium channel polymorphisms play a pivotal role in the development of oxaliplatin-induced peripheral neurotoxicity: results from a prospective multicenter study. Cancer 119(19): 3570–3577 61. Grolleau F, Gamelin L, Boisdron-Celle M et al (2001) A possible explanation for a neurotoxic effect of the anticancer agent oxaliplatin on neuronal voltage-gated sodium channels. J Neurophysiol 85(5):2293–2297 62. Alberti P, Canta A, Chiorazzi A et al (2020) Topiramate prevents oxaliplatin-related axonal hyperexcitability and oxaliplatin induced peripheral neurotoxicity. Neuropharmacology 164:107905 63. Palugulla S, Thakkar DN, Kayal S et al (2017) Association of Voltage-Gated Sodium Channel Genetic Polymorphisms with OxaliplatinInduced Chronic Peripheral Neuropathy in South Indian Cancer Patients. Asian Pac J Cancer Prev 18(11):3157–3165 64. Abaji R, Ceppi F, Patel S et al (2018) Genetic risk factors for VIPN in childhood acute lymphoblastic leukemia patients identified using whole-exome sequencing. Pharmacogenomics 19(15):1181–1193 65. Abbasian MH, Ansarinejad N, Abbasi B et al (2020) The Role of. Avicenna J Med Biotechnol 12(3):157–164 66. Abraham JE, Guo Q, Dorling L et al (2014) Replication of genetic polymorphisms reported to be associated with taxane-related sensory neuropathy in patients with early breast cancer treated with Paclitaxel. Clin Cancer Res 20(9):2466–2475 67. Adjei AA, Lopez CL, Schaid DJ et al (2021) Genetic Predictors of ChemotherapyInduced Peripheral Neuropathy from Paclitaxel, Carboplatin and Oxaliplatin: NCCTG/Alliance N08C1, N08CA and N08CB Study. Cancers (Basel) 13(5) 68. Anon B, Largeau B, Girault A et al (2018) Possible association of CAG repeat polymorphism in KCNN3 encoding the potassium channel SK3 with oxaliplatin-induced neurotoxicity. Cancer Chemother Pharmacol 82(1): 149–157 69. Antonacopoulou AG, Argyriou AA, Scopa CD et al (2010) Integrin beta-3 L33P: a new insight into the pathogenesis of chronic

Pharmacogenomics and CIPN oxaliplatin-induced peripheral neuropathy? Eur J Neurol 17(7):963–968 70. Apellániz-Ruiz M, Lee MY, Sánchez-Barroso L et al (2015) Whole-exome sequencing reveals defective CYP3A4 variants predictive of paclitaxel dose-limiting neuropathy. Clin Cancer Res 21(2):322–328 71. Apellániz-Ruiz M, Tejero H, Inglada-Pe´rez L et al (2017) Targeted Sequencing Reveals Low-Frequency Variants in. Clin Cancer Res 23(5):1227–1235 72. Arbitrio M, Scionti F, Altomare E et al (2019) Polymorphic Variants in NR1I3 and UGT2B7 Predict Taxane Neurotoxicity and Have Prognostic Relevance in Patients With Breast Cancer: A Case-Control Study. Clin Pharmacol Ther 106(2):422–431 73. Argyriou AA, Antonacopoulou AG, Alberti P et al (2019) Liability of the voltage-gated potassium channel KCNN3 repeat polymorphism to acute oxaliplatin-induced peripheral neurotoxicity. J Peripher Nerv Syst 24(4): 298–303 74. Azoulay D, Giryes S, Nasser R et al (2019) Prediction of Chemotherapy-Induced Peripheral Neuropathy in Patients with Lymphoma and Myeloma: the Roles of Brain-Derived Neurotropic Factor Protein Levels and A Gene Polymorphism. J Clin Neurol 15(4): 511–516 75. Basso M, Modoni A, Spada D et al (2011) Polymorphism of CAG motif of SK3 gene is associated with acute oxaliplatin neurotoxicity. Cancer Chemother Pharmacol 67(5): 1179–1187 76. Bergmann TK, Gre´en H, Brasch-Andersen C et al (2011) Retrospective study of the impact of pharmacogenetic variants on paclitaxel toxicity and survival in patients with ovarian cancer. Eur J Clin Pharmacol 67(7):693–700 77. Boora GK, Kanwar R, Kulkarni AA et al (2016) Testing of candidate single nucleotide variants associated with paclitaxel neuropathy in the trial NCCTG N08C1 (Alliance). Cancer Med 5(4):631–639 78. Bosilkovska M, Ing Lorenzini K, Uppugunduri CR et al (2016) Severe Vincristineinduced Neuropathic Pain in a CYP3A5 Nonexpressor With Reduced CYP3A4/5 Activity: Case Study. Clin Ther 38(1):216–220 79. Boso´ V, Herrero MJ, Santaballa A et al (2014) SNPs and taxane toxicity in breast cancer patients. Pharmacogenomics 15(15): 1845–1858 80. Broyl A, Corthals SL, Jongen JL et al (2010) Mechanisms of peripheral neuropathy associated with bortezomib and vincristine in

137

patients with newly diagnosed multiple myeloma: a prospective analysis of data from the HOVON-65/GMMG-HD4 trial. Lancet Oncol 11(11):1057–1065 81. Campo C, Da Silva Filho MI, Weinhold N et al (2017) Genetic Susceptibility to Bortezomib-Induced Peripheral Neuroropathy: Replication of the Reported Candidate Susceptibility Loci. Neurochem Res 42(3): 925–931 82. Campo C, da Silva Filho MI, Weinhold N et al (2018) Bortezomib-induced peripheral neuropathy: A genome-wide association study on multiple myeloma patients. Hematol Oncol 36(1):232–237 83. Caponigro F, Lacombe D, Twelves C et al (2009) An EORTC phase I study of Bortezomib in combination with oxaliplatin, leucovorin and 5-fluorouracil in patients with advanced colorectal cancer. Eur J Cancer 45(1):48–55 84. Ceppi F, Langlois-Pelletier C, Gagne´ V et al (2014) Polymorphisms of the vincristine pathway and response to treatment in children with childhood acute lymphoblastic leukemia. Pharmacogenomics 15(8):1105–1116 85. Chang H, Rha SY, Jeung HC et al (2009) Association of the ABCB1 gene polymorphisms 2677G>T/A and 3435C>T with clinical outcomes of paclitaxel monotherapy in metastatic breast cancer patients. Ann Oncol 20(2):272–277 86. Chen YC, Tzeng CH, Chen PM et al (2010) Influence of GSTP1 I105V polymorphism on cumulative neuropathy and outcome of FOLFOX-4 treatment in Asian patients with colorectal carcinoma. Cancer Sci 101(2): 530–535 87. Custodio A, Moreno-Rubio J, Aparicio J et al (2014) Pharmacogenetic predictors of severe peripheral neuropathy in colon cancer patients treated with oxaliplatin-based adjuvant chemotherapy: a GEMCAD group study. Ann Oncol 25(2):398–403 88. Di Francia R, Atripaldi L, Di Martino S et al (2017) Assessment of Pharmacogenomic Panel Assay for Prediction of Taxane Toxicities: Preliminary Results. Front Pharmacol 8: 797 89. Diouf B, Crews KR, Lew G et al (2015) Association of an inherited genetic variant with vincristine-related peripheral neuropathy in children with acute lymphoblastic leukemia. JAMA 313(8):815–823 90. Favis R, Sun Y, van de Velde H et al (2011) Genetic variation associated with bortezomib-

138

Eleonora Pozzi and Paola Alberti

induced peripheral neuropathy. Pharmacogenet Genomics 21(3):121–129 91. Ferracini AC, Lopes-Aguiar L, Lourenc¸o GJ et al (2021) GSTP1 and ABCB1 Polymorphisms Predicting Toxicities and Clinical Management on Carboplatin and Paclitaxel-Based Chemotherapy in Ovarian Cancer. Clin Transl Sci 14(2):720–728 92. Garcı´a-Sanz R, Corchete LA, Alcoceba M et al (2017) Prediction of peripheral neuropathy in multiple myeloma patients receiving bortezomib and thalidomide: a genetic study based on a single nucleotide polymorphism array. Hematol Oncol 35(4):746–751 93. Gre´en H, So¨derkvist P, Rosenberg P et al (2009) Pharmacogenetic studies of Paclitaxel in the treatment of ovarian cancer. Basic Clin Pharmacol Toxicol 104(2):130–137 94. Hertz DL, Motsinger-Reif AA, Drobish A et al (2012) CYP2C8*3 predicts benefit/risk profile in breast cancer patients receiving neoadjuvant paclitaxel. Breast Cancer Res Treat 134(1):401–410 95. Hertz DL, Owzar K, Lessans S et al (2016) Pharmacogenetic Discovery in CALGB (Alliance) 90401 and Mechanistic Validation of a VAC14 Polymorphism that Increases Risk of Docetaxel-Induced Neuropathy. Clin Cancer Res 22(19):4890–4900 96. Hertz DL, Roy S, Jack J et al (2014) Genetic heterogeneity beyond CYP2C8*3 does not explain differential sensitivity to paclitaxelinduced neuropathy. Breast Cancer Res Treat 145(1):245–254 97. Hu L, Lv QL, Guo Y et al (2016) Genetic variation of CYP3A5 influences paclitaxel/ carboplatin-induced toxicity in Chinese epithelial ovarian cancer patients. J Clin Pharmacol 56(3):349–354 98. Inada M, Sato M, Morita S et al (2010) Associations between oxaliplatin-induced peripheral neuropathy and polymorphisms of the ERCC1 and GSTP1 genes. Int J Clin Pharmacol Ther 48(11):729–734 99. Isla D, Sarries C, Rosell R et al (2004) Single nucleotide polymorphisms and outcome in docetaxel-cisplatin-treated advanced nonsmall-cell lung cancer. Ann Oncol 15(8): 1194–1203 100. Johnson DC, Corthals SL, Walker BA et al (2011) Genetic factors underlying the risk of thalidomide-related neuropathy in patients with multiple myeloma. J Clin Oncol 29(7): 797–804 101. Johnson C, Pankratz VS, Velazquez AI et al (2015) Candidate pathway-based genetic association study of platinum and platinum-

taxane related toxicity in a cohort of primary lung cancer patients. J Neurol Sci 349(1-2): 124–128 102. Kanai M, Kawaguchi T, Kotaka M et al (2016) Large-scale prospective pharmacogenomics study of oxaliplatin-induced neuropathy in colon cancer patients enrolled in the JFMC41-1001-C2 (JOIN Trial). Ann Oncol 27(6):1143–1148 103. Kayiliog˘lu H, Kocak U, Kan Karaer D et al (2017) Association of CYP3A5 Expression and Vincristine Neurotoxicity in Pediatric Malignancies in Turkish Population. J Pediatr Hematol Oncol 39(6):458–462 104. Komatsu M, Wheeler HE, Chung S et al (2015) Pharmacoethnicity in PaclitaxelInduced Sensory Peripheral Neuropathy. Clin Cancer Res 21(19):4337–4346 105. Kulkarni AA, Boora G, Kanwar R et al (2015) RWDD3 and TECTA variants not linked to paclitaxel induced peripheral neuropathy in North American trial Alliance N08C1. Acta Oncol 54(8):1227–1229 106. Kumpiro S, Sriuranpong VAreepium N (2016) Impact of the Copper Transporter Protein 1 (CTR1) Polymorphism on Adverse Events among Advanced NonSmall Cell Lung Cancer Patients Treated with a Carboplatin/ Gemcitabine Regimen. Asian Pac J Cancer Prev 17(9):4391–4394 107. Kus T, Aktas G, Kalender ME et al (2016) Polymorphism of CYP3A4 and ABCB1 genes increase the risk of neuropathy in breast cancer patients treated with paclitaxel and docetaxel. Onco Targets Ther 9:5073–5080 108. Lam SW, Frederiks CN, van der Straaten T et al (2016) Genotypes of CYP2C8 and FGD4 and their association with peripheral neuropathy or early dose reduction in paclitaxel-treated breast cancer patients. Br J Cancer 115(11):1335–1342 109. Lamba JK, Fridley BL, Ghosh TM et al (2014) Genetic variation in platinating agent and taxane pathway genes as predictors of outcome and toxicity in advanced non-smallcell lung cancer. Pharmacogenomics 15(12): 1565–1574 110. Lambrechts S, Lambrechts D, Despierre E et al (2015) Genetic variability in drug transport, metabolism or DNA repair affecting toxicity of chemotherapy in ovarian cancer. BMC Pharmacol Toxicol 16:2 111. Leandro-Garcı´a LJ, Inglada-Pe´rez L, Pita G et al (2013) Genome-wide association study identifies ephrin type A receptors implicated in paclitaxel induced peripheral sensory neuropathy. J Med Genet 50(9):599–605

Pharmacogenomics and CIPN 112. Lee SY, Im SA, Park YH et al (2014) Genetic polymorphisms of SLC28A3, SLC29A1 and RRM1 predict clinical outcome in patients with metastatic breast cancer receiving gemcitabine plus paclitaxel chemotherapy. Eur J Cancer 50(4):698–705 113. Lee MY, Apellániz-Ruiz M, Johansson I et al (2015) Role of cytochrome P450 2C8*3 (CYP2C8*3) in paclitaxel metabolism and paclitaxel-induced neurotoxicity. Pharmacogenomics 16(9):929–937 114. Leibovici A, Sharon RAzoulay D (2018) BDNF Val66Met is Associated with Pre-existing but not with Paclitaxel-induced Peripheral Neuropathy in an Israeli Cohort of Breast Cancer Patients. Isr Med Assoc J 20(12):746–748 115. Leskela¨ S, Jara C, Leandro-Garcı´a LJ et al (2011) Polymorphisms in cytochromes P450 2C8 and 3A5 are associated with paclitaxel neurotoxicity. Pharmacogenomics J 11(2): 121–129 116. Li L, Sajdyk T, Smith EML et al (2019) Genetic Variants Associated With Vincristine-Induced Peripheral Neuropathy in Two Populations of Children With Acute Lymphoblastic Leukemia. Clin Pharmacol Ther 105(6):1421–1428 117. Lopez-Lopez E, Gutierrez-Camino A, Astigarraga I et al (2016) Vincristine pharmacokinetics pathway and neurotoxicity during early phases of treatment in pediatric acute lymphoblastic leukemia. Pharmacogenomics 17(7):731–741 118. Madi A, Fisher D, Maughan TS et al (2018) Pharmacogenetic analyses of 2183 patients with advanced colorectal cancer; potential role for common dihydropyrimidine dehydrogenase variants in toxicity to chemotherapy. Eur J Cancer 102:31–39 119. Magrangeas F, Kuiper R, Avet-Loiseau H et al (2016) A Genome-Wide Association Study Identifies a Novel Locus for BortezomibInduced Peripheral Neuropathy in European Patients with Multiple Myeloma. Clin Cancer Res 22(17):4350–4355 120. Mahmoudpour SH, Bandapalli OR, da Silva Filho MI et al (2018) Chemotherapy-induced peripheral neuropathy: evidence from genome-wide association studies and replication within multiple myeloma patients. BMC Cancer 18(1):820 121. Marcath LA, Kidwell KM, Vangipuram K et al (2020) Genetic variation in EPHA contributes to sensitivity to paclitaxel-induced peripheral neuropathy. Br J Clin Pharmacol 86(5):880–890

139

122. Martin-Guerrero I, Gutierrez-Camino A, Echebarria-Barona A et al (2019) Variants in vincristine pharmacodynamic genes involved in neurotoxicity at induction phase in the therapy of pediatric acute lymphoblastic leukemia. Pharmacogenomics J 19(6):564–569 123. Mlak R, Szudy-Szczyrek A, Mazurek M et al (2019) Polymorphisms in the promotor region of the CRBN gene as a predictive factor for peripheral neuropathy in the course of thalidomide-based chemotherapy in multiple myeloma patients. Br J Haematol 186(5): 695–705 124. Negarandeh R, Salehifar E, Saghafi F et al (2020) Evaluation of adverse effects of chemotherapy regimens of 5-fluoropyrimidines derivatives and their association with DPYD polymorphisms in colorectal cancer patients. BMC Cancer 20(1):560 125. Nichetti F, Falvella FS, Miceli R et al (2019) Is a pharmacogenomic panel useful to estimate the risk of oxaliplatin-related neurotoxicity in colorectal cancer patients? Pharmacogenomics J 19(5):465–472 126. Palugulla S, Devaraju P, Kayal S et al (2018) Genetic polymorphisms in cyclin H gene are associated with oxaliplatin-induced acute peripheral neuropathy in South Indian digestive tract cancer patients. Cancer Chemother Pharmacol 82(3):421–428 127. Park SB, Kwok JB, Loy CT et al (2014) Paclitaxel-induced neuropathy: potential association of MAPT and GSK3B genotypes. BMC Cancer 14:993 128. Park SB, Kwok JB, Asher R et al (2017) Clinical and genetic predictors of paclitaxel neurotoxicity based on patient- versus clinicianreported incidence and severity of neurotoxicity in the ICON7 trial. Ann Oncol 28(11): 2733–2740 129. Rizzo R, Spaggiari F, Indelli M et al (2010) Association of CYP1B1 with hypersensitivity induced by taxane therapy in breast cancer patients. Breast Cancer Res Treat 124(2): 593–598 130. Rua C, Gue´guinou M, Soubai I et al (2018) SK3 Gene Polymorphism Is Associated with Taxane Neurotoxicity and Cell Calcium Homeostasis. Clin Cancer Res 24(21): 5313–5320 131. Ruzzo A, Graziano F, Galli F et al (2014) Genetic markers for toxicity of adjuvant oxaliplatin and fluoropyrimidines in the phase III TOSCA trial in high-risk colon cancer patients. Sci Rep 4:6828 132. Sawaki A, Miyazaki K, Yamaguchi M et al (2020) Genetic polymorphisms and

140

Eleonora Pozzi and Paola Alberti

vincristine-induced peripheral neuropathy in patients treated with rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone therapy. Int J Hematol 111(5): 686–691 133. Schneider BP, Li L, Radovich M et al (2015) Genome-Wide Association Studies for Taxane-Induced Peripheral Neuropathy in ECOG-5103 and ECOG-1199. Clin Cancer Res 21(22):5082–5091 134. Schneider BP, Lai D, Shen F et al (2016) Charcot-Marie-Tooth gene, SBF2, associated with taxane-induced peripheral neuropathy in African Americans. Oncotarget 7(50): 82244–82253 135. Sereno M, Gutie´rrez-Gutie´rrez G, Rubio JM et al (2017) Genetic polymorphisms of SCN9A are associated with oxaliplatininduced neuropathy. BMC Cancer 17(1):63 136. Sims RP (2016) The effect of race on the CYP3A-mediated metabolism of vincristine in pediatric patients with acute lymphoblastic leukemia. J Oncol Pharm Pract 22(1):76–81 137. Sissung TM, Mross K, Steinberg SM et al (2006) Association of ABCB1 genotypes with paclitaxel-mediated peripheral neuropathy and neutropenia. Eur J Cancer 42(17): 2893–2896 138. Sissung TM, Baum CE, Deeken J et al (2008) ABCB1 genetic variation influences the toxicity and clinical outcome of patients with androgen-independent prostate cancer treated with docetaxel. Clin Cancer Res 14(14): 4543–4549 139. Skiles JL, Chiang C, Li CH et al (2018) CYP3A5 genotype and its impact on vincristine pharmacokinetics and development of neuropathy in Kenyan children with cancer. Pediatr Blood Cancer 65(3) 140. Stock W, Diouf B, Crews KR et al (2017) An Inherited Genetic Variant in CEP72 Promoter Predisposes to Vincristine-Induced Peripheral Neuropathy in Adults With Acute Lymphoblastic Leukemia. Clin Pharmacol Ther 101(3):391–395 141. Sucheston-Campbell LE, Clay-Gilmour AI, Barlow WE et al (2018) Genome-wide metaanalyses identifies novel taxane-induced

peripheral neuropathy-associated loci. Pharmacogenet Genomics 28(2):49–55 142. Tanabe Y, Shimizu C, Hamada A et al (2017) Paclitaxel-induced sensory peripheral neuropathy is associated with an ABCB1 single nucleotide polymorphism and older age in Japanese. Cancer Chemother Pharmacol 79(6):1179–1186 143. Tanabe Y, Shiraishi S, Hashimoto K et al (2020) Taxane-induced sensory peripheral neuropathy is associated with an SCN9A single nucleotide polymorphism in Japanese patients. BMC Cancer 20(1):325 144. Terrazzino S, Argyriou AA, Cargnin S et al (2015) Genetic determinants of chronic oxaliplatin-induced peripheral neurotoxicity: a genome-wide study replication and metaanalysis. J Peripher Nerv Syst 20(1):15–23 145. van Rossum AGJ, Kok M, McCool D et al (2017) Independent replication of polymorphisms predicting toxicity in breast cancer patients randomized between dose-dense and docetaxel-containing adjuvant chemotherapy. Oncotarget 8(69):113531–113542 146. Wright GEB, Amstutz U, Dro¨gemo¨ller BI et al (2019) Pharmacogenomics of Vincristine-Induced Peripheral Neuropathy Implicates Pharmacokinetic and Inherited Neuropathy Genes. Clin Pharmacol Ther 105(2):402–410 147. Zgheib NK, Ghanem KM, Tamim H et al (2018) Genetic polymorphisms in candidate genes are not associated with increased vincristine-related peripheral neuropathy in Arab children treated for acute childhood leukemia: a single institution study. Pharmacogenet Genomics 28(8):189–195 148. Zhong J, Guo Z, Fan L et al (2019) ABCB1 polymorphism predicts the toxicity and clinical outcome of lung cancer patients with taxane-based chemotherapy. Thorac Cancer 10(11):2088–2095 149. Zhou W, An G, Jian Y et al (2015) Effect of CYP2C19 and CYP3A4 gene polymorphisms on the efficacy of bortezomib-based regimens in patients with multiple myeloma. Oncol Lett 10(2):1171–1175

Chapter 6 The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine Qing Yan Abstract The enormous heterogeneity of cancer systems has made it very challenging to overcome drug resistance and adverse reactions to achieve personalized therapies. Recent developments in systems biology, especially the perception of cancer as the complex adaptive system (CAS), may help meet the challenges by deciphering the interactions at various levels from the molecular, cellular, tissue-organ, to the whole organism. The ubiquitous Yin-Yang interactions among the coevolving components, including the genes and proteins, decide their spatiotemporal features at various stages from cancer initiation to metastasis. The Yin-Yang imbalances across different systems levels, from genetic mutations to tumor cells adaptation, have been related to the intra- and inter-tumoral heterogeneity in the micro- and macro-environments. At the molecular and cellular levels, dysfunctional Yin-Yang dynamics in the cytokine networks, mitochondrial activities, redox systems, apoptosis, and metabolism can contribute to tumor cell growth and escape of immune surveillance. Up to the organism and system levels, the Yin-Yang imbalances in the cancer microenvironments can lead to different phenotypes from breast cancer to leukemia. These factors may be considered the systems-based biomarkers and treatment targets. The features of adaptation and nonlinearity in Yin-Yang dynamical interactions should be addressed by individualized drug combinations, dosages, intensities, timing, and frequencies at different cancer stages. The comprehensive “Yin-Yang dynamics” framework would enable powerful approaches for personalized and systems medicine strategies. Key words Biomarkers, Cancer, Complex adaptive systems (CAS), Epigenetics, Immune, Pharmacogenomics, Personalized medicine, Systems biology, Tumor, Yin-Yang

1

Challenges in Cancer Pharmacogenomics and Personalized Medicine Although progress has been made in cancer therapies in recent decades, the treatment capacity has not been achieved for all cancer patients because of drug toxicity and disease resistance [1]. While pharmacogenomics focuses on individual differences in drug responses, it is crucial to understand the underlying mechanisms of drug resistance and various side effects. Diseases such as cancers are multicellular systems with tremendous complexity [2]. While certain cancers respond to the

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_6, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

141

142

Qing Yan

treatments, other types do not. For instance, the usages of immunotherapies for glioblastoma have reached the bottleneck [3]. While some patients respond to the therapies, others do not. The capabilities of glioblastoma tumors to nullify antitumor immunity have been suggested as multifaceted with complex adaptive features [3]. Furthermore, acquired drug resistance has been a critical hurdle blocking the success of targeted cancer treatments [4]. It has been recognized that targeting tumors with kinase inhibitors may lead to complex adaptive activities that promote inhibitor-resistant tumor clones. The enormous heterogeneity of cancer systems has made it very challenging to overcome adverse reactions to achieve personalized therapies [5]. These challenges have called for urgent amendments from the reductionism-based system to the systems-based, dynamical, and human-centric approaches [6]. The components interact via elaborate networks at each level in the hierarchical systems to maintain homeostasis. The interactions need to be elucidated at various levels from the molecular, cellular, tissue-organ, to the whole organism [6, 7]. Factors such as genetic mutations and stress may lead to homeostatic disturbances and impairments that overpower the standard adaptive safeguard. In these processes, systems do not react linearly to treatment augmentations but rather reflect the overall responses as disease symptoms or drug resistance [6, 7].

2

Complex Adaptive Systems (CAS) in Cancer Recent developments in systems biology, especially the perception of cancer as the complex adaptive system (CAS), may help meet the challenges by deciphering the interactions at various systems levels [6, 8]. The CAS concepts may help clarify the dynamic links between tumor growth and drug resistance [1]. Intracellular molecular networks, extracellular matrix, and environmental factors, including stress, may influence early and late cancer stages from initiation to metastasis [9]. Unlike the conventional static view, the features of CAS may better characterize the constantly changing and evolving pathophysiological systems. CAS are open systems with nonlinear interactions among the components [6, 8]. These components also communicate with and adapt to the micro- or macro-environment. The overall collective behaviors and functions result from the interconnections and co-evolution of the elements. In systems theories, such a feature is often summarized as “the whole is greater than the sum of its parts” [6]. For instance, head and neck squamous cell carcinoma (HNSCC) has been recognized as an evolving system. The path of its somatic evolution may contribute to the difficulties in

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

143

improving survival among patients [10]. The adaptation and robustness of the cancer cells in response to the genetic, epigenetic, and micro-environmental stresses have made it challenging to promote patients’ quality of life. Because cancers are complex adaptive ecosystems, the microenvironmental interactions among the components, including the genes and proteins, may contribute to their heterogeneous spatiotemporal features [11, 12]. As the members coevolve and adapt to environmental changes, new features may arise in the subsystems, such as drug resistance in the later stages of cancer [8]. Such an eco-evolutionary standpoint is necessary for understanding cancer pathology and personalized treatments. Dynamical activities in cancer development decide that different biomarkers and therapeutics are needed at various stages of the evolvement for the same patient and different patients with the same disease [6, 8]. Such perception is the basic principle in personalized medicine. In addition, the feedback loops among the communicating elements at different levels from genes, proteins, to the environment provide the connections between the structure-function and genotype-phenotype [6]. Specifically, at the molecular level, changes and different isoforms in the structures and functions may influence the downstream cellular networks and cancer phenotypes at the higher levels [8]. On the other hand, factors at higher levels may impact the lower levels. An example is that the environmental stress factor of excess sunlight may result in DNA damages at the molecular level and skin cancer at the organism level [6]. These cross-level feedback loops and spatiotemporal interconnections request the elucidation of the complex mechanisms within and between different levels. Biomarkers and drug targets should no longer just focus on polymorphisms or genes but need to include systemic feedback loops from genetic pathways to gene-environment interactions during different stages.

3

Yin-Yang Dynamics in CAS and Cancer Understanding these critical CAS properties would assist the design of more effective approaches in prevention and treatment in personalized medicine [6, 8]. For instance, continuous boosting of the dosages in cancer chemotherapy may have adverse results but not improvements [6]. Studies have shown that chemotherapy can help shrink the tumor sizes initially but may provoke secondary tumors if given with excess dosages [12]. The remarkable feature of this nonlinearity reveals the transformation of the opposite characteristics in the same course, pointing to Yin-Yang dynamics in the adaptation and evolutionary processes [8]. As discussed in the

144

Qing Yan

following sections, such Yin-Yang dynamics are critical in psychophysiological homeostasis and disease development and treatment. The complex adaptive features of the nonlinear dynamical balancing, counteracting, interdependent, and transforming components in CAS can best be integrated and represented by the concepts of “Yin” and “Yang” [8]. Many scientific publications have used the concepts of “Yin” and “Yang” to illustrate the opposite yet complementary and converting aspects of similar entities or processes. The “Yin-Yang dynamics” conceptual framework may help identify effective biomarkers and drug targets for personalized and systems medicine [8]. As discussed in the examples below, the imbalances in the immune mechanisms, redox conditions, metabolic processes, and mitochondrial functions have been linked to tumor growth, the escape of immune surveillance, and drug resistance. Systems-based personalized strategies should focus on these adaptive circumstances to reestablish Yin-Yang dynamical balances.

4

Yin-Yang Dynamics and miRNAs in Cancer Table 1 lists examples of the Yin-Yang roles of microRNAs (miRNAs) in cancers. The elucidation of the Yin and Yang of miRNAs has been suggested useful for therapeutic strategies for obesity, diabetes, and cancer [13]. The miRNAs have essential functions in the insulin signaling cascade as they may promote or inhibit the production of insulin or insulin-like growth factor 1 (IGF1) (see Table 1). They may influence the activities of the insulin/IGF-1 receptor and the associated signaling network. They may bind to target mRNAs to suppress mRNA translation. They can also enhance mRNA degradation [13]. Depending on the cellular environments, miR-26a can act as a tumor inhibitor or oncogene with the opposite Yin-Yang roles [14]. For example, miR-26a has lower levels in hepatocellular carcinoma (HCC), with reduced tumor inhibitory functions on cyclins D2 and E2 (see Table 1). However, higher levels of miR-26a expression have been observed in high-grade glioma (GBM) with inhibitory functions on the tumor suppressor PTEN, leading to tumor progression [14]. The Yin-Yang pattern has been found in miRNAs including miR-125b and miR-146a [15]. While they are essential in supporting normal physiological activities, they may also be associated with autoimmune diseases and cancer (see Table 1). Specifically, miR-125b may have immediate and direct effects. In contrast, miR-146a has slow effects, indicating chronic changes in cellular activities [15]. The interactions between reactive oxygen species (ROS) and miRNA in cancer also have the Yin-Yang modes [16]. ROS have

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

145

Table 1 Yin-Yang dynamical interactions in miRNAs and epigenetics Factors

Conditions

miRNAs

Obesity, type Promotion vs. inhibition of insulin 2 diabetes, cancer

[13]

miR-26a

HCC, glioma

A tumor inhibitor vs. oncogene

[14]

miRNAs, miR-125b, miR-146a

Autoimmune diseases, cancer

Normal vs. diseases; quick vs. slow effects [15]

ROS and miRNAs

Cancer

ROS regulates miRNAs, supports tumor [16] growth; miRNAs affect ROS

The lin28/let-7 axis

HCC and HCC metastasis

The imbalance associated with tumor development

[17]

Histone demethylases and Cancer methyltransferases

Regulation of oncogenes vs. tumor suppressors

[18]

PRDM isoforms

Oncogenes vs. tumor inhibitors

[19, 20]

Multiple tumor types

The RIZ family, RIZ1 and Cancers including prostate cancer RIZ2 PRMT5

Yin-Yang interactions

References

Tumor suppressor vs. proto-oncoprotein [21]

Acute myeloid Promotion vs. inhibition of AML leukemia (AML)

[22]

crucial roles in cell proliferation and apoptosis. They are also essential for immune actions. Higher levels of ROS may induce tumor development through DNA damage and oncogenic signaling networks [16]. ROS may also affect miRNA biogenesis and epigenetic regulations to support tumor growth (see Table 1). On the other hand, miRNAs may regulate cellular ROS homeostasis by controlling the networks associated with ROS generation and removal [16]. In addition, the lin28/let-7 signaling networks show Yin-Yang balancing features [17]. Although they both are involved in tumor development, lin28 and the let-7 miRNA have contrasting expression styles. The lin28/let-7 axis imbalance has been associated with oncogenesis (see Table 1). Lower levels of let-7 and elevated levels of lin28 have been observed in different tumors such as hepatocellular carcinoma (HCC) and HCC metastasis [17]. Such Yin-Yang imbalance may be crucial in tumor progression and abnormal transformation, as the opposite expression modes of lin28 and let-7 have only been observed in HCC during the later phases [17]. In HCC and ovarian cancer cells, the block of lin28 or re-stimulation of let-7 may suppress tumor cell proliferation. Studies have also related the tumorigenic effects of lin28 to Myc and NF-κB associated networks [17].

146

5

Qing Yan

The Yin and Yang in Epigenetics and Cancer The epigenetic regulators of oncogenes and tumor suppressors, especially histone demethylases and methyltransferases in cancer, may have the Yin-Yang pattern [18]. Abnormal histone methylation may affect oncogenic and tumor inhibition mechanisms (see Table 1). Specifically, lysine methyltransferases and lysine demethylases have opposite functions [18]. However, they may complement each other in a complex network between cancer metabolism and epigenetics. The upregulations of PR domain zinc finger proteins (PRDMs) have been found in multiple tumor types, making these proteins important anticancer targets. The PRDM proteins may have essential roles in mutations and epigenetic silencing in cancer cells [19]. Different PRDMs isoforms have shown Yin-Yang features [19, 20]. Specific PRDM isoforms may have the functions of oncogenes, while others may be tumor inhibitors (see Table 1). Such dualities have been associated with the complex regulations of alternative splicing or promoter usage. These mechanisms may generate full-length or PR-deficient isoforms with opposite effects [20]. Retinoblastoma protein-interacting zinc-finger gene 1(RIZ1) and RIZ2 proteins belong to the nuclear histone/protein methyltransferase superfamily. RIZ1 has the N-terminal PR domain, but RIZ2 does not, differentiating the two proteins. RIZ1 is a tumor suppressor, while RIZ2 is a proto-oncoprotein, showing the YinYang pattern [21]. During stage IV of eight cancers and stage III of prostate cancer, higher levels of RIZ1 may be detected (see Table 1). Although RIZ1 has been associated with tumor metastasis, its PR domain may have anticancer effects [21]. Moreover, protein arginine methyltransferase (PRMT) has essential roles in epigenetic regulation. Altered PRMT5 has been related to cancer, especially acute myeloid leukemia (AML) [22]. Elevated levels of PRMT5 may promote AML growth, and lower levels of PRMT5 may inhibit the development. However, recent findings have also shown that PRMT5 may be involved in blocking and activating genes in a Yin-Yang manner in signaling cascades associated with leukemia development [22] (see Table 1).

6

Yin-Yang Dynamics in Cytokines and Chemokines The Yin-Yang functions have been found for the cytokine interleukin-21 (IL-21) signaling network that influences the naive B cells, decided by the circumstances [23]. IL-21 has pro-apoptotic effects on B and NK cells. It may lead to apoptosis in naive B cells with pathogen-associated signals to toll-like receptors (TLR). On

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

147

the other hand, upon stimulated by the B cell antigen receptor and IL-4, IL-21 may enhance the generation of naive B cells. In addition, IL-21 is involved in immunoglobulin generation and Th17 differentiation [23]. While IL-21 has antitumor functions, it also has a role in developing autoimmune diseases (see Table 2). IL-27 has both antitumor and protumor potentials [24] (see Table 2). It can promote Th1/Tc1 reactions and antitumor T-cell functions, and suppress tumor growth and inflammation. IL-27 can promote T-cell survival in the tumor microenvironment. On the other hand, IL-27 can activate pathways including IL-10, PD-L1, CD39, and Th1-Treg and suppress tumor-specific T-cell reactions [24]. Its antitumor functions may be more potent than its protumor functions. Interleukin-33 (IL-33) belongs to the IL-1 family. IL-33 is involved in several cancers with both pro- and anti-tumorigenic effects [25]. As a prognostic biomarker, IL-33 may enhance the proliferation and metastasis of tumor cells. This cytokine may stimulate tumor growth by changing the tumor microenvironment (TME) and promoting angiogenesis [25]. It can stimulate M2 macrophage polarization and tumor infiltration and promote immunosuppressive cells. However, IL-33 also has antitumor effects associated with T helper 1 (TH1) reactions [25]. Such Yin and Yang functions of IL-33 request more comprehensive strategies when using this cytokine in cancer immunotherapies (see Table 2). As a member of the transforming growth factor β (TGFβ) superfamily, bone morphogenetic proteins (BMPs) are growth proteins that are critical for homeostasis [26]. The interactions between BMPs and their receptors have been associated with carcinogenesis and tumor development. However, BMPs may also have tumor-inhibitory effects in some conditions, showing the Yin-Yang mechanisms [26] (see Table 2). In breast cancer, C-X-C Motif Chemokine Receptor 4 (CXCR4) and CXCR7 are important in tumor progression and metastasis [27]. CXCR4 may promote the progression by the interaction with MMP12 upon the stimulation of CXCL12. CXCR7 may be involved in the tumor microenvironment. However, the co-expression of CXCR4 and CXCR7 may lead to the suppression of CXCL12-associated tumor progression and reduced lung metastases (see Table 2). The Yin-Yang modes of CXCR4 and CXCR7 may help block progression and metastasis in breast cancer [27].

7

The Yin and Yang of Immune Cells: Inflammation and Tumor Microenvironments The communications between tumor cells and innate immune cells gear tumor inflammation toward pro- or anti-tumorigenic

148

Qing Yan

Table 2 Yin-Yang dynamical interactions in the immune system Factors

Conditions

Yin-Yang interactions

References

IL-21

Tumor, autoimmune Antitumor functions vs. development of diseases autoimmune diseases

[23]

IL-27

Tumor

The antitumor vs. protumor effects

[24]

IL-33

Several cancers

Protumor vs. antitumor functions

[25]

BMPs

Tumor progression

Tumor development vs. tumor suppressor [26]

CXCR4, CXCR7, CXCL12

Breast cancer (BC)

Promotion vs. inhibition of tumor progression

[27]

Innate immunity

Inflammation

Antitumor vs. protumor reactions

[28]

NF-κB

Tumor growth, drug Immunosurveillance vs. tumorigenic resistance effects

[29]

NF-κB and GR

Breast cancer (BC)

Pro- vs. anti-inflammation; cancer cell survival vs. apoptosis

[30]

Monocytes

Cancers

Protumor vs. antitumor effects

[31]

TAMs

Cancers, chemo-and radiotherapy

Tumor-promoting vs. antineoplastic effects

[32]

M1 and M2macrophages

Tumor progression

Antitumor vs. cancer progression

[33]

Tumor-associated neutrophils

Lung cancer

Suppression vs. enhancement of cancer progression

[34]

ILC1, ILC2, ILC3

Tumor, inflammation Protumor vs. antitumor effects

NK cells and monocytes

Multiple myeloma

Monocytes promote NK cells; NK cells inhibit monocytes

[36]

iNKT

Ulcerative colitis

Intestinal inflammation, tumor formation vs. immunosurveillance

[37]

Macrophages, T/NK cell subsets

Follicular lymphoma

Antitumor vs. protumor and drug resistance effects

[38]

B cells, CXCR4/ CXCL12, Tregs

Breast cancer

Antitumor vs. tumor promotion effects

[39]

Tregs

Autoimmunity and cancer

Beneficial in autoimmunity vs. harmful in [40] cancers

Toll-like receptors (TLRs)

Tumor inflammation Antitumor vs. tumor-generating inflammation

BLT1

Cancers

Tumor inflammation vs. immune surveillance

[42]

MSCs

GVHD, AML

Beneficial in GVHD vs. harmful in AML

[43]

Angiogenesis and immune responses

Cancer

Antitumor effects vs. angiogenesis

[44]

[35]

[41]

(continued)

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

149

Table 2 (continued) Factors

Conditions

Yin-Yang interactions

References

Extracellular vesicles (EVs)

Tumor progression

Immune activation vs. immunosuppression

[45]

reactions. The pattern recognition-mediated sensing systems may lead to antitumor immunosurveillance or inflammation that promotes tumor development [28]. Tumor-related inflammation may have the Yin and Yang dual paths (see Table 2). Signaling networks may inhibit innate immunity, forming a negative feedback loop in the tumor microenvironments. However, the same network may also target nuclear factor-κB (NF-κB)-associated anti-apoptosis and inflammation signals, negatively controlling tumor growth [28]. Specifically, the innate immune systems have antitumor mechanisms by generating cytokines to promote innate and adaptive effector cells for antitumor functions. On the other hand, the innate immune systems may also have protumor means by producing cytokines such as IL-6, TNF-α, and IL-1b TLR to support pro-tumorigenic inflammation [28]. Nuclear factor-κB (NF-κB) is critical in immune regulation with immunosurveillance effects [29]. However, it has also been associated with tumor growth, metastasis, and drug resistance (see Table 2). In certain conditions, NF-κB may be transformed to be tumorigenic, opposite to its original functions with the Yin-Yang fashion [29]. The involvement of NF-κB and glucocorticoid receptor (GR) in inflammation has the Yin-Yang pattern [30]. While NF-κB has pro-inflammatory functions, GR has anti-inflammatory effects (see Table 2). The interactions can decide the survival or apoptosis of breast cancer (BC) cells between these effects [30]. Because NF-κB has been associated with the start and development of BC, the suppressors of NF-κB have been applied in the essential or adjuvant therapeutics. Glucocorticoids (GCs) have been used in adjuvant treatments to stimulate apoptosis for different cancers, especially for leukemia and lymphoma. However, GCs may enhance cancer cell survival and drug resistance [30]. Therefore, complex interactions should be considered for more precise and personalized therapies. Monocytes are essential in innate and adaptive immunity [31]. They can influence the tumor microenvironment by stimulating immune tolerance, angiogenesis, and higher proliferation of tumor cells. However, monocytes also have antitumor functions by promoting antigen-presenting cells (APCs) [31]. Such Yin-Yang dual functions are the reactions of monocytes upon environmental stress (see Table 2).

150

Qing Yan

Macrophages have crucial roles in tumor-generating inflammation. Tumor-associated macrophages (TAMs) are involved in tumor development at various levels. TAMs can support cancer stem cells and metastasis and inhibit adaptive protective immunity [32]. In addition, TAMs have Yin-Yang effects on chemotherapies and radiotherapies (see Table 2). While having suppressing functions on the antitumor reactions, they may also promote antineoplastic mechanisms [32]. Thus, TAMs are critical for achieving personalized medicine in immunotherapy and chemotherapy. Many studies have confirmed the Yin-Yang roles of macrophages during tumor development [33]. M1 macrophages have the controlling part of antitumor effects in the initial stages of cancer. They can provoke certain adaptive antitumor immune functions, leading to the M1 polarization [33]. For example, mononuclear phagocytes may have antitumor functions via killing cancer cells and triggering tissue disruptive responses (M1). During the late stages of cancers, M2-polarized tumor-related macrophages take the controlling part to inhibit adaptive immune functions (see Table 2). Such dynamical “macrophage balance” is critical in cancer progression, associated with innate immunity-related inflammation and adaptive immunity-related cancer surveillance [33]. Cancer-associated inflammation has important influences on the tumor microenvironment. In lung cancer, tumor-related neutrophils can polarize to “N1” or “N2” phenotypes that suppress or enhance cancer progression, the Yin-Yang-like effects [34] (see Table 2). As the innate counterparts of T cells, innate lymphoid cells (ILCs) have significant roles in the processes of tissue homeostasis and inflammation. ILCs contain cytokine-generating cells, including ILC1, ILC2, and ILC3. Decided by the tumor stage and the microenvironment, the ILCs can have protumor and antitumor functions [35] (see Table 2). ILC1, ILC2, and ILC3 can support tumor growth by generating tumor stimulation cytokines, inhibiting immune responses, and promoting tumor immune escape. However, they can have the opposite antitumor functions by activating and gathering leukocytes to the tumor site [35]. They may also promote the apoptosis of cancer cells, support the intra-tumoral tertiary lymphoid structure (TLS), and enhance the tumor-antigen presentation. Natural killer (NK) cells are cytotoxic with anticancer functions. Complicated interactions are critical between monocytes/ macrophages and NK cells. The growth of the NK cells needs the activation of monocytes [36]. However, NK cells may inhibit the dendrimer-activated monocytes in the anti-inflammatory pathway [36]. The Yin-Yang patterns in the interactions between dendrimer-activated monocytes and NK cells need to be considered

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

151

in NK-cell-based anticancer therapies such as those for multiple myeloma (see Table 2). Invariant natural killer T (iNKT) cells have dual roles in the cancer microenvironment [37]. In human ulcerative colitis, the iNKT cells can inhibit T helper 1 (TH1)’s antitumor immunity and enhance intestinal inflammation. However, they may also promote TH1 functions and the immunosurveillance mechanism to inhibit metastasis (see Table 2). In follicular lymphoma (FL), T/NK cells and macrophages may have double functions in the dynamic interactions with B cells in affected lymph nodes and bone marrow [38]. Such bidirectional communications may result in antitumor immune reactions. On the other hand, the microenvironment heterogeneity, including the Treg/Th17 polarization imbalance, may establish a tumorpromoting network for growth and drug resistance [38] (see Table 2). In the heterogeneous breast tumor microenvironment, B cells have the Yin-Yang functions [39]. They can generate immunoglobulin and kill tumor cells through CXCR4/CXCL12 and perforin pathways. On the other hand, they can also promote Tregs or myeloid-derived suppressor cells (MDSCs) to suppress antitumor immune reactions [39]. The complexity of the actions of B cells makes it necessary to understand the systemic interactions for the application of B cell-based immunotherapy (see Table 2). The CD4(+) regulatory T cells (Tregs) may have dual functions in autoimmunity and cancer [40]. Tregs participate in the processes of immune tolerance and inhibiting pro-inflammatory reactions. Such roles may be beneficial for autoimmune diseases but damaging in cancers (see Table 2). The Yin-Yang patterns are prominent in many immune networks. For example, toll-like receptors (TLRs) are involved in both protumor and antitumor pathways [41]. Specifically, TLR3, TLR5, TLR7/TLR8, and TLR9 have shown antitumor functions by transforming immune tolerance into antitumor immune responses. High expression of TLRs may stimulate dendritic cells, natural killer (NK) cells, and cytotoxic T lymphocytes (CTL) associated with type I interferon (IFN) [41]. However, chronic low levels of TLRs including TLR2 and TLR4 may result in tumor-generating inflammation and the inhibition of tumor apoptosis. Tumor-associated macrophages (TAMs) and fibroblasts may be involved in such processes [41]. Depending on the tumor microenvironment, TLRs may have both Yin and Yang roles in protumor or antitumor processes (see Table 2). In another example, leukotriene B4 (LTB4) interacts with G-protein coupled receptors (GPCR) BLT1 and BLT2 [42]. The LTB4-BLT1 axis has a Yin-Yang pattern in cancer regulation. While BLT1 is involved in tumor-enhancing inflammation, it also communicates with cytotoxic T lymphocytes (CTLs) to activate

152

Qing Yan

immune surveillance antitumor mechanisms [42]. On the other hand, BLT2 is associated with pro-tumorigenic networks to support drug resistance, angiogenesis, and metastasis (see Table 2). With the Yin and Yang functions in the bone marrow microenvironment, mesenchymal stromal cells (MSCs) have immunoregulatory functions [43]. Such processes have been applied in graft versus host disease (GVHD). They may be beneficial by suppressing innate and adaptive immune mechanisms (see Table 2). However, MSCs may be harmful by supporting immune tolerance for developing acute myeloid leukemia (AML). They are also involved in drug resistance in AML [43]. Furthermore, angiogenesis and immune responses are not independent mechanisms but have the Yin-Yang mode interactions [44]. The suppression of angiogenesis and the promotion of immune responses may lead to the inhibition of tumor growth. With the host cells playing critical roles, the activities of tumor angiogenesis may suppress immune responses. The blocking of angiogenesis may improve immune reactions [44]. Such interactions may help therapeutic design strategies (see Table 2). Extracellular vesicles (EVs) significantly influence the tumor microenvironment [45]. EVs are involved in immune-tumor interactions and immunosuppression, contributing to tumor development. However, tumor-originated EVs may also stimulate host immunity and initiate therapeutic responses [45] (see Table 2).

8

Yin-Yang Dynamical Balances in the Redox Systems Redox homeostasis has a crucial role in cellular metabolism and significantly influences cancer development. Reactive oxygen and nitrogen species (ROS/RNS) have the Yin-Yang functions on cancer, impacting cellular growth and apoptosis [46]. During the early and later stages of cancers, ROS/RNS have protumor or antitumor parts, respectively (see Table 3). They are involved in the interactions among tumor components and affect the transformation between anti- to pro-tumorigenic processes. During various phases of tumorigenesis, they affect the oxidation mechanisms and tumor microenvironment, resulting in different outcomes [46]. In cancers, the complex interactions between mitochondrial dynamics and reactive oxygen species (ROS) have been recognized as the Yin-Yang type [47]. Higher mitochondria-caused ROS and redox signaling may lead to cell death (see Table 3). Decided by the cellular environment, mitochondrial ultrastructure has been associated with ROS production [47]. Abnormal mitochondrial morphology can promote ROS formation. However, such changes can damage mitochondrial functions and worsen oxidative stress conditions, forming a negative loop [47].

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

153

Table 3 Yin-Yang dynamical interactions in the redox systems Factors

Conditions

Yin-Yang interactions

References

ROS/RNS

Solid tumors

Protumor vs. antitumor

[46]

Mitochondria, ROS

Cancer progression

ROS production, cell death

[47]

Mitochondrial metabolism

Tumor progression

Promotion vs. inhibition of cancer metastasis

[48]

Nitric oxide (NO)

Prostate cancer, tumor progression

Cancer promotion vs. anticancer effects

[49, 50]

Nitric oxide (NO)

Head and neck cancers

Beneficial vs. harmful effects

[51]

Mitochondria are critical in the metabolic mechanisms in tumor development. The Yin-Yang patterns of mitochondrial activities have been identified in metastasis and drug resistance [48]. Depending on the microenvironment, mitochondrial functions may be beneficial or damaging in these activities (see Table 3). As a signaling molecule, nitric oxide (NO) may have different functions with complex protumor or antitumor roles, decided by its concentrations and the tumor microenvironment (TME) [49, 50]. Low concentrations of NO may result in castrationresistant prostate cells, angiogenesis, and higher levels of cancer growth [49, 50]. At medium concentrations, NO may enhance metastasis and inhibit apoptosis. However, high NO concentrations have been shown to stimulate cytotoxicity and apoptosis with anticancer effects [49, 50]. NO may suppress androgen receptors and block prostate cancer development [49]. Such different mechanisms may be decided by the TME and inflammatory conditions (see Table 3). This is also true in head and neck cancers, as many factors decide if nitric oxide (NO) is beneficial or harmful in the YinYang manner [51]. These factors include the tissues, the concentrations, and the redox conditions (see Table 3). NO may change enzymatic mechanisms by interacting with covalent, redox factors, and metallic functional sites [51].

9

The Yin-Yang Interactions in Autophagy and Apoptosis The common understanding is that stimulating apoptotic networks or the lack of apoptotic suppressors would improve cancer treatment [52]. However, the suppression of apoptosis may reduce ischemia-caused impairments. Some studies about apoptosisregulating drugs have shown unpredicted effects in various clinical situations [52]. The Yin and Yang roles of autophagy and the

154

Qing Yan

interactions among apoptotic and autophagic networks may contribute to such results (see Table 4). Such mechanisms need to be noticed in the development of cancer treatments. Autophagy and cell growth can suppress each other in the YinYang manner [53]. In the environment of cell stresses, including nutrient deficiency, autophagy may be triggered by decreased or halted cell growth (see Table 4). Ribosomes on the endoplasmic reticulum surface may inhibit autophagy. Autophagy may inhibit cell growth. Increased levels of autophagy elements, including ULK1 and WIPI1, may reduce the mammalian target of rapamycin (mTOR) [53]. Signaling networks may have different influences on autophagy and cell growth, e.g., the mTOR pathway has been related to both processes. The endoplasmic reticulum stress (ERS) responses and autophagy in cancers, including glioblastoma, also have the Yin-Yang patterns [54, 55]. When the ERS responses and autophagy are at low to moderate levels, they may have tumor protective and chemoresistance effects. However, more severe situations may trigger apoptosis (see Table 4). Higher levels of ER stress may result in tumor cell death [54, 55]. Such mechanisms suggest that combination approaches may be needed to develop treatments.

10

Yin-Yang Dynamical Balances in the p53 and c-Myc Pathways The Yin-Yang positions of the tumor suppressor p53 pathway for cell protection or cell death rely on the intensity of the stress [56]. High stress levels have been found to trigger p53-related genes that are pro-apoptotic and pro-oxidant in the direction of cell death or senescence (see Table 4). Conversely, low or mild stress strength can trigger the p53-related pro-survival cascades to protect cells from impairments [56]. The Yin-Yang dynamic balance of the p53 can be critical in cancer prevention [57]. Upregulated p53 may have anticancer effects to reduce the risks of cancer development. However, such protection may result in the early occurrence of specific aging phenotypes (see Table 4). On the other hand, reduced p53 expressions have been associated with higher cancer risks and the initial development of tumors [57]. c-Myc (proto-oncoprotein) and p53 (the tumor suppressor) have been suggested as a Yin-Yang pair of proteins in various cellular functions, including cell proliferation and tissue homeostasis (see Table 4). The protein c-Myc may activate the ARF tumor suppressor and p53 [58]. On the other hand, p53 may inhibit c-Myc via various processes, including microRNA-associated suppression. In addition, c-Myc may inhibit p53 via the c-Myc-Inducible Long noncoding RNA Inactivating P53 (MILIP) [58].

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

155

Table 4 Yin-Yang interactions in p53, c-Myc, apoptosis, and autophagy Factors

Conditions

Yin-Yang interactions

References

Apoptosis, autophagy

Cancers and cancer treatments

Autophagy in cell survival vs. cell death

[52]

Autophagy, cell growth, mTOR

Cell growth

Autophagy vs. cell growth

[53]

ERS and autophagy

Glioblastoma, chemoresistance

Tumor protective and chemoresistance vs. cell death

[54, 55]

p53, stress

Cancers

Cell protection vs. cell death

[56]

p53

Tumor, premature aging

Anticancer vs. early aging; cancer development

[57]

p53 and c-Myc

Cancers

Proto-oncoprotein vs. tumor inhibitor

[58, 59]

p53 and MDM2

Cell growth and proliferation

MDM2 suppresses p53; p53 disturbs MDM2 production

[60]

E2F-1

Apoptosis

Apoptosis promotion vs. inhibition

[61]

Myc

Cancer development

Tumor development vs. tumor suppression

[62]

SerRS and c-Myc

Vascular development Inhibition vs. activation in the vasculature [63] development

An inhibitory axis targeting p53 via c-Myc may be critical in cancer development [58]. Moreover, both p53 and c-Myc are controlled by some E3 ubiquitin ligases and their regulator ARF [59]. These proteins may also have opposite effects in the regulation of cell proliferation. The complex ubiquitylation is critical in cancer [59]. Furthermore, proteins p53 and mouse double minute 2 homolog (MDM2) interact in the Yin-Yang negative feedback loop [60]. The MDM2 may suppress p53 activities, while the transcriptional factor p53 may disturb the MDM2 production (see Table 4). Such interactions compose a negative feedback loop. MDM2 may overcome the p53-associated inhibition of cell proliferation in certain conditions and support cell survival [60]. The Yin and Yang features have been identified in the E2 factor (E2F) that is involved in cell cycles [61] (see Table 4). Upon the stimulations, including DNA damage, apoptosis may be triggered, and E2F subunits such as E2F-1 may be critical in such processes. On the other hand, in an environment lacking p53, E2F-1 may inhibit apoptosis [61]. When the p53 responses are activated by the DNA damage networks, the apoptosis inhibition by E2F-1 may be overcome. While both E2F-1 and p53 may interact with apoptosis

156

Qing Yan

target genes, the p53 responses may dominate the E2F-1-caused inhibition [61]. Although Myc is a proto-oncoprotein, its Yin-Yang dual functions have been recognized [62]. The deregulation of Myc may transform the cell into a hyperproliferative condition for tumor development. On the other hand, Myc activities may trigger tumor inhibition processes such as apoptosis to curb tumor progression [62]. Such understanding would be helpful in the design of individualized therapeutic targets (see Table 4). Moreover, the Yin-Yang model has been identified in the interactions between seryl-tRNA synthetase (SerRS) and c-Myc during vascular development [63]. SerRS is a crucial element in vascular development. SerRS may interact with the Vascular Endothelial Growth Factor A (VEGFA) promoter, causing the SIRT2 histone deacetylase to inhibit VEGFA activities [63]. SerRS may block c-Myc’s functions of supporting VEGFA, keeping the dynamical balance during vasculature development (see Table 4). Therefore, SerRS may regulate VEGFA and block vascular overgrowth by inhibiting c-Myc, as c-Myc is critical for activating VEGFA [63].

11

Yin-Yang Interactions in Various Networks Acyl-CoA Synthetase 3 (ACSL3) and ACSL4 are essential in fatty acid metabolism, and both belong to the ACSLs family. These two proteins are controlled by the androgen-AR network with opposite effects [64]. The inhibition of the androgen-AR network may lead to lower levels of ACSL3 but higher levels of ACSL4. ACSL3 is plentiful in prostate cancer cells with reduced efficiency in catalyzing the generation of fatty acyl-CoA. In comparison, lower levels of ACSL4 have been found with greater catalytic efficiency (see Table 5). Such Yin-Yang functions may be related to cancer cells survival [64]. As a regulator of the defense systems, NF-E2-related factor2 (Nrf2) has been associated with preventing many diseases, including cancer [65]. Nrf2 may target the selenoproteins, including thioredoxin reductase-1 (TrxR1) and glutathione peroxidase2 (GPx2). Increased levels of TrxR1 in tumor cells are harmful as they may enhance cellular proliferation (see Table 5). Although GPx2 may have the anti-inflammatory effects to suppress tumor growth, it may also promote cellular growth and tumor development, showing the Yin-Yang effects [65]. The phosphatase and tensin homolog (PTEN) is a lipid phosphatase with tumor suppressor functions that offset the activities of the oncogenic phosphoinositide 3-kinase (PI3K). PTEN is often mutated in tumors [66]. However, the suppression of PTEN has been found beneficial in some other human diseases, including type

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

157

Table 5 Yin-Yang dynamical interactions in various pathways Factors

Conditions

Yin-Yang interactions

ACSL3 and ACSL4

Prostate cancer

Oppositely controlled by the androgen-AR [64] network

Nrf2, TrxR1, GPx2

Cancer cells

Anti-inflammation, tumor inhibition vs. tumor promotion

[65]

PTEN

Tumor, type 2 diabetes, neurodegeneration diseases

Beneficial for tumor suppression; harmful in type 2 diabetes, neurodegeneration diseases

[66]

The SMC5/6 complex

Cancer cells replication

Genome stability and anticancer vs. cancer [67] cells replication

STAT1, STAT3

Colorectal cancer

Tumor inhibitor vs. tumor enhancer

[68]

Sulfs

Tumor

Protumor vs. antitumor functions

[69]

The VEGF and Tumor angiogenesis Notch networks

VEGF promotes DLL4/Notch; DLL4/ Notch controls VEGF

[70]

Wnt signaling

Tumor development vs. suppression

[71]

Melanoma

References

Pancreatic cancer Sympathetic and parasympathetic nerves

Tumor angiogenesis vs. tumor suppression [72]

The Ras/MAPK pathway, drugs

Thyroid carcinoma

Tumor inhibition vs. tumor promotion

[75]

Valproic acid (VPA)

Glioma

Tumor cell survival vs. pro-apoptotic reactions

[76]

2 diabetes and neurodegeneration diseases, indicating its dual roles in different microenvironments [66] (see Table 5). Yin-Yang patterns have also been identified in the ATPase SMC5/6 complex, a member of the structural maintenance of chromosomes (SMC) complexes [67]. The SMC5/6 complex may support genetic stability via homologous recombination (HR) with anticancer functions (see Table 5). However, the SMC5/6 complex may also enhance telomere elongation via HR to support alternative lengthening of telomeres (ALT) and cancer cells replication [67]. In colorectal cancer (CRC), Yin-Yang pattern interactions between the signal transducer and activator of transcription 1 (STAT1) and STAT3 need to be considered, especially for cancer prognosis [68]. STAT1 has been deemed an independent tumor inhibitor and a positive prognostic indicator. STAT3 may contribute to cancer progression with negative clinical results (see Table 5). STAT1 may inhibit oncogenic STAT3 functions [68]. Such

158

Qing Yan

Yin-Yang relationships suggest that these two should be considered together rather than separately for cancer prognosis. The Sulfs are extracellular sulfatases with protumor and antitumor functions [69]. They may interact with the polysaccharide heparan sulfate (HS) (see Table 5). The vascular endothelial growth factor (VEGF) and Notch signaling networks have the Yin-Yang mode interactions, especially in the processes of angiogenesis [70]. The inhibition of the VEGF network may suppress cancer angiogenesis. The Delta-like ligand 4 (DLL4)/Notch network may reduce angiogenesis by inhibiting endothelial tip cells [70]. The inhibition of the DLL4/Notch network may enhance non-productive angiogenesis but suppress tumor growth (see Table 5). The VEGF signaling interacts with the DLL4/Notch network in vasculature. VEGF may promote the DLL4/Notch network, while the DLL4/Notch may control the VEGF network [70]. Combination treatment approaches of silencing the DLL4/Notch and VEGF networks may suppress tumor growth. The opposite Yin-Yang patterns have been identified in the Wnt signaling networks in melanoma [71]. Canonical Wnt networks, including Wnt1 and Wnt3A, have essential roles in the early phases of tumor development with the transformation of melanocytes to radial growth phase melanoma (RGP). However, the molecules may suppress metastasis during later phases, including vertical growth phase melanoma (VGP) (see Table 5). The Wnt5A signaling may not be detectable during the early phases of the tumor. In the later phases with the transformation of melanocytes, Wnt5A may decrease b-catenin and enhance metastasis [71]. The Yin and Yang features have been found in neural processes in pancreatic cancer. Sensory and sympathetic nerves may affect the tumor microenvironment via neurotrophic factors to trigger the development of pancreatic cancer cells [72]. They may stimulate the production of neurotransmitters, including substance P (SP) and noradrenaline (NA). These molecules then stimulate the neurokinin-1 receptor (NK1R) and the adrenergic receptor beta 2 (ADRβ2), respectively, to promote tumor angiogenesis [72]. On the other hand, parasympathetic nerves may eventually suppress pancreatic cancer development by producing acetylcholine (ACh) and cholinergic muscarinic receptor 1 (CHRM1). These molecules may suppress PI3K/AKT and EGFR/ERK [72]. The cholinergic cascade may help inhibit the cancer stem cells, TNFα levels, and metastasis. The opposite functions of sensory and sympathetic versus parasympathetic nerves indicate that the dynamical balance of the neural activities may influence pancreatic cancer (see Table 5).

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

12

159

Conclusion: Yin-Yang Dynamics and Personalized Cancer Therapy With the understanding that cancers are not just diseases but complex adaptive systems (CAS) with nonlinear connections among different levels from genotypic variants to phenotypes, systemic therapeutic approaches have been demanded [73, 74]. As summarized in Fig. 1, such methodologies should embrace the evolutionary-based mechanism and address the features of adaptation, nonlinearity, and feedback loops on various spatiotemporal scales. Across-level components from genetic mutations to tumor cell adaptation contribute to the intra- and inter-tumoral heterogeneity, adverse reactions, and drug resistance in tumor microenvironments [74]. Complex adaptive therapeutic strategies (CATS) should focus on targeting the disturbance of the components and connections for the reconstruction of the balanced dynamics [8, 74] (see Fig. 1). The above evidences have demonstrated that the counteracting and interconnecting Yin-Yang pattern is ubiquitous that may not just represent and interpret the features of CAS but also enable a more dynamical and integrative view (see Fig. 1). Such understanding may provide novel insights into developing personalized and systems-based diagnosis, prevention, and therapeutic strategies. At the molecular and cellular levels, the Yin-Yang imbalances in miRNAs, epigenetics, mitochondrial activities, apoptosis, and redox systems can lead to systemic malfunctions in cancer. At the organism and system levels, the disturbance of the Yin-Yang dynamics in the immune functions and various networks can contribute to multiple phenotypes of cancer (see Fig. 1). Such YinYang imbalances may be considered systems-based biomarkers and treatment targets for restoring the normal spatiotemporal dynamics [6, 8]. The features of adaptation and nonlinearity in the ubiquitous Yin-Yang interactions should be addressed by individualized dosages, intensities, timing, and frequencies at different cancer stages (see Fig. 1). The feedback loops at various levels require the treatments considering both internal and external stressors and environmental factors. For instance, in thyroid carcinoma, antitumor drugs targeting the canonical signaling pathways may result in Yin-Yang consequences [75]. While the drugs may inhibit tumor growth in many cases, they may also promote tumor cell migration. Specifically, drugs targeting the PI3/Akt and MEK/ERK1/2 pathways may reduce the proliferation of tumor cells. In some other cases, drugs suppressing the MEK/ERK1/2 pathways may significantly enhance tumor cell migration [75]. Such opposite functions of the antitumor drugs need to be carefully considered (see Table 5).

160

Qing Yan

Fig. 1 Yin-Yang dynamics framework for complex adaptive therapeutic strategies and personalized medicine

In another example, valproic acid (VPA) is a histone deacetylase inhibitor considered a promising glioma drug. VPA may promote the unfolded protein response (UPR). However, VPA has shown a Yin-Yang pattern during different stages [76]. In the early stages, it may enhance the survival of cancer cells. Longer-lasting stimulation may lead to pro-apoptotic reactions (see Table 5). Thus, VPA may influence the balance between tumor adaptation and apoptosis related to protein homeostasis and proteotoxicity [76]. Furthermore, combination treatments based on patient stratification have been suggested to tackle the complex adaptive mechanisms such as those in melanoma therapy [77]. Combination therapies may target multiple signaling and metabolic networks and embrace epigenetic processes and immune functions. Such novel treatments need to restore the dynamical Yin-Yang balances in the immunosurveillance. Not just immunosuppression should be targeted; the mechanisms of inflammation, the tumor microenvironment, and the disease/treatment stages should also be thoroughly studied [1, 8]. Overall, the comprehensive “Yin-Yang dynamics” framework would enable powerful approaches for personalized and systems medicine strategies (see Fig. 1).

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine

161

References 1. Derbal Y (2018) The adaptive complexity of cancer. Biomed Res Int 2018:5837235 2. Fong EJ, Strelez C, Mumenthaler SM (2020) A perspective on expanding our understanding of cancer treatments by integrating approaches from the biological and physical sciences. SLAS Discov 25:672–683 3. Reardon DA, Wucherpfennig K, Chiocca EA (2017) Immunotherapy for glioblastoma: on the sidelines or in the game? Discov Med 24: 201–208 4. Rusan M, Li K, Li Y, Christensen CL et al (2018) Suppression of adaptive responses to targeted cancer therapy by transcriptional repression. Cancer Discov 8:59–73 5. Meyer AS, Heiser LM (2019) Systems biology approaches to measure and model phenotypic heterogeneity in cancer. Curr Opin Syst Biol 17:35–40 6. Yan Q (2014) From pharmacogenomics and systems biology to personalized care: a framework of systems and dynamical medicine. Methods Mol Biol 1175:3–17 7. Feinendegen LE, Neumann RD (2006) The issue of risk in complex adaptive systems: the case of low-dose radiation induced cancer. Hum Exp Toxicol 25:11–17 8. Yan Q (2018) Stress and systemic inflammation: Yin-Yang dynamics in health and diseases. Methods Mol Biol 1781:3–20 9. Feller L, Khammissa RAG, Lemmer J (2017) Biomechanical cell regulatory networks as complex adaptive systems in relation to cancer. Cancer Cell Int 17:16 10. Masuda M, Toh S, Wakasaki T et al (2013) Somatic evolution of head and neck cancer biological robustness and latent vulnerability. Mol Oncol 7:14–28 11. Uthamacumaran A (2020) Cancer: A turbulence problem. Neoplasia 22:759–769 12. Reynolds BA, Oli MW, Oli MK (2020) Eco-oncology: applying ecological principles to understand and manage cancer. Ecol Evol 10:8538–8553 13. Liu J, Liu F (2021) The Yin and Yang function of microRNAs in insulin signaling and cancer. RNA Biol 18:24–32 14. Huse JT, Holland EC (2009) Yin and Yang: cancer-implicated miRNAs that have it both ways. Cell Cycle 8:3611–3612 15. So AY-L, Zhao JL, Baltimore D (2013) The Yin and Yang of microRNAs: leukemia and immunity. Immunol Rev 253:129–145

16. Babu KR, Tay Y (2019) The Yin-Yang regulation of reactive oxygen species and microRNAs in cancer. Int J Mol Sci 20:E5335 17. Ji J, Wang XW (2010) A Yin-Yang balancing act of the lin28/let-7 link in tumorigenesis. J Hepatol 53:974–975 18. Filipp FV (2017) Crosstalk between epigenetics and metabolism-Yin and Yang of histone demethylases and methyltransferases in cancer. Brief Funct Genomics 16:320–325 19. Mzoughi S, Tan YX, Low D, Guccione E (2016) The role of PRDMs in cancer: one family, two sides. Curr Opin Genet Dev 36: 83–91 20. Di Tullio F, Schwarz M, Zorgati H et al (2021) The duality of PRDM proteins: epigenetic and structural perspectives. FEBS J. https://doi. org/10.1111/febs.15844 21. Sun W, Qiao L, Liu Q, Chen L et al (2011) Anticancer activity of the PR domain of tumor suppressor RIZ1. Int J Med Sci 8:161–167 22. Tarighat SS, Santhanam R, Frankhouser D et al (2016) The dual epigenetic role of PRMT5 in acute myeloid leukemia: gene activation and repression via histone arginine methylation. Leukemia 30:789–799 23. Spolski R, Leonard WJ (2008) The Yin and Yang of interleukin-21 in allergy, autoimmunity and cancer. Curr Opin Immunol 20:295– 301 24. Li M-S, Liu Z, Liu J-Q, Zhu X et al (2015) The Yin and Yang aspects of IL-27 in induction of cancer-specific T-cell responses and immunotherapy. Immunotherapy 7:191–200 25. Fournie´ J-J, Poupot M (2018) The pro-tumorigenic IL-33 involved in antitumor immunity: a Yin and Yang cytokine. Front Immunol 9:2506 26. Singh A, Morris RJ (2010) The Yin and Yang of bone morphogenetic proteins in cancer. Cytokine Growth Factor Rev 21:299–313 27. Hawkins OE, Richmond A (2012) The dynamic Yin-Yang interaction of CXCR4 and CXCR7 in breast cancer metastasis. Breast Cancer Res 14:103 28. Jinushi M (2014) Yin and Yang of tumor inflammation: how innate immune suppressors shape the tumor microenvironments. Int J Cancer 135:1277–1285 29. Xiao G, Fu J (2011) NF-κB and cancer: a paradigm of Yin-Yang. Am J Cancer Res 1:192–221 30. Ling J, Kumar R (2012) Crosstalk between NFκB and glucocorticoid signaling: a potential

162

Qing Yan

target of breast cancer therapy. Cancer Lett 322:119–126 31. Ugel S, Cane` S, De Sanctis F, Bronte V (2021) Monocytes in the tumor microenvironment. Annu Rev Pathol 16:93–122 32. Mantovani A, Marchesi F, Malesci A et al (2017) Tumor-associated macrophages as treatment targets in oncology. Nat Rev Clin Oncol 14:399–416 33. Allavena P, Sica A, Garlanda C, Mantovani A (2008) The Yin-Yang of tumor-associated macrophages in neoplastic progression and immune surveillance. Immunol Rev 222:155– 161 34. Mantovani A (2009) The Yin-Yang of tumorassociated neutrophils. Cancer Cell 16:173– 174 35. Carrega P, Campana S, Bonaccorsi I, Ferlazzo G (2016) The Yin and Yang of innate lymphoid cells in cancer. Immunol Lett 179:29–35 36. Poupot M, Turrin CO, Caminade AM et al (2016) Poly(phosphorhydrazone) dendrimers: Yin and Yang of monocyte activation for human NK cell amplification applied to immunotherapy against multiple myeloma. Nanomedicine 12(8):2321–2330 37. Wang Y, Cardell SL (2017) The Yin and Yang of invariant natural killer T cells in tumor immunity-suppression of tumor immunity in the intestine. Front Immunol 8:1945 38. Ame´-Thomas P, Tarte K (2014) The Yin and the Yang of follicular lymphoma cell niches: role of microenvironment heterogeneity and plasticity. Semin Cancer Biol 24:23–32 39. Zhang Z, Zhu Y, Wang Z et al (2017) Yin-Yang effect of tumor infiltrating B cells in breast cancer: from mechanism to immunotherapy. Cancer Lett 393:1–7 40. Galgani M, Di Giacomo A, Matarese G, La Cava A (2009) The Yin and Yang of CD4(+) regulatory T cells in autoimmunity and cancer. Curr Med Chem 16:4626–4631 41. Pradere J-P, Dapito DH, Schwabe RF (2014) The Yin and Yang of toll-like receptors in cancer. Oncogene 33:3485–3495 42. Jala VR, Bodduluri SR, Satpathy SR et al (2017) The Yin and Yang of leukotriene B4 mediated inflammation in cancer. Semin Immunol 33:58–64 43. Ciciarello M, Corradi G, Loscocco F et al (2019) The Yin and Yang of the bone marrow microenvironment: pros and cons of mesenchymal stromal cells in acute myeloid leukemia. Front Oncol 9:1135 44. Szala S, Mitrus I, Sochanik A (2010) Can inhibition of angiogenesis and stimulation of immune response be combined into a more

effective antitumor therapy? Cancer Immunol Immunother 59:1449–1455 45. Yamauchi T, Moroishi T (2021) The Yin and Yang of tumor-derived extracellular vesicles in tumor immunity. J Biochem 169:155–161 46. Mijatovic´ S, Savic´-Radojevic´ A, Pljesˇa-Ercegovac M et al (2020) The double-faced role of nitric oxide and reactive oxygen species in solid tumors. Antioxidants (Basel) 9:E374 47. Jezˇek J, Cooper KF, Strich R (2018) Reactive oxygen species and mitochondrial dynamics: the Yin and Yang of mitochondrial dysfunction and cancer progression. Antioxidants (Basel) 7: E13 48. Valcarcel-Jimenez L, Gaude E, Torrano V et al (2017) Mitochondrial metabolism: Yin and Yang for tumor progression. Trends Endocrinol Metab 28:748–757 49. Soni Y, Softness K, Arora H, Ramasamy R (2020) The Yin Yang role of nitric oxide in prostate cancer. Am J Mens Health 14: 1557988320903191 50. Burke AJ, Sullivan FJ, Giles FJ, Glynn SA (2013) The Yin and Yang of nitric oxide in cancer progression. Carcinogenesis 34:503– 512 51. Bentz BG, Simmons RL, Haines GK, Radosevich JA (2000) The Yin and Yang of nitric oxide: reflections on the physiology and pathophysiology of NO. Head Neck 22:71–83 52. Jain MV, Paczulla AM, Klonisch T et al (2013) Interconnections between apoptotic, autophagic and necrotic pathways: implications for cancer therapy development. J Cell Mol Med 17: 12–29 53. Neufeld TP (2012) Autophagy and cell growth--the Yin and Yang of nutrient responses. J Cell Sci 125:2359–2368 54. Scho¨nthal AH, Chen TC, Hofman FM et al (2011) Preclinical development of novel antiglioma drugs targeting the endoplasmic reticulum stress response. Curr Pharm Des 17:2428– 2438 55. Scho¨nthal AH (2009) Endoplasmic reticulum stress and autophagy as targets for cancer therapy. Cancer Lett 275:163–169 56. Budanov AV (2014) The role of tumor suppressor p53 in the antioxidant defense and metabolism. Subcell Biochem 85:337–358 57. Campisi J (2002) Cancer and aging: Yin, Yang, and p53. Sci Aging Knowledge Environ 2002: 1pe–1 58. Feng YC, Liu XY, Teng L et al (2020) c-Myc inactivation of p53 through the pan-cancer lncRNA MILIP drives cancer pathogenesis. Nat Commun 11:4980

The Yin-Yang Dynamics in Cancer Pharmacogenomics and Personalized Medicine 59. Dai M-S, Jin Y, Gallegos JR, Lu H (2006) Balance of Yin and Yang: ubiquitylationmediated regulation of p53 and c-Myc. Neoplasia 8:630–644 60. Lu H (2017) p53 and MDM2: their Yin-Yang intimacy. J Mol Cell Biol 9:1–2 61. La Thangue NB (2003) The Yin and Yang of E2F-1: balancing life and death. Nat Cell Biol 5:587–589 62. Larsson L-G, Henriksson MA (2010) The Yin and Yang functions of the Myc oncoprotein in cancer development and as targets for therapy. Exp Cell Res 316:1429–1437 63. Shi Y, Xu X, Zhang Q et al (2014) tRNA synthetase counteracts c-Myc to develop functional vasculature. eLife 3:e02349 64. Ma Y, Zhang X, Alsaidan OA et al (2021) Long-chain acyl-CoA synthetase 4–mediated fatty acid metabolism sustains androgen receptor pathway–independent prostate cancer. Mol Cancer Res 19:124–135 65. Brigelius-Flohe´ R, Mu¨ller M, Lippmann D, Kipp AP (2012) The Yin and Yang of nrf2regulated selenoproteins in carcinogenesis. Int J Cell Biol 2012:486147 66. Pulido R (2015) PTEN: a Yin-Yang master regulator protein in health and disease. Methods 77–78:3–10 67. Potts PR (2009) The Yin and Yang of the MMS21–SMC5/6 SUMO ligase complex in homologous recombination. DNA Repair 8: 499–506 68. Friedrich K, Dolznig H, Han X, Moriggl R (2017) Steering of carcinoma progression by the YIN/YANG interaction of STAT1/ STAT3. Biosci Trends 11:1–8 69. Vive`s RR, Seffouh A, Lortat-Jacob H (2014) Post-synthetic regulation of HS structure: the

163

Yin and Yang of the Sulfs in cancer. Front Oncol 3:331 70. Li J-L, Harris AL (2009) Crosstalk of VEGF and Notch pathways in tumor angiogenesis: therapeutic implications. Front Biosci (Landmark Ed) 14:3094–3110 71. O’Connell MP, Weeraratna AT (2009) Hear the Wnt Ror: how melanoma cells adjust to changes in Wnt. Pigment Cell Melanoma Res 22:724–739 72. Faulkner S, Jobling P, March B et al (2019) Tumor neurobiology and the war of nerves in cancer. Cancer Discov 9:702–710 73. Ye CJ, Regan S, Liu G et al (2018) Understanding aneuploidy in cancer through the lens of system inheritance, fuzzy inheritance and emergence of new genome systems. Mol Cytogenet 11:31 74. Cho YW, Kim SY, Kwon IC, Kim I-S (2014) Complex adaptive therapeutic strategy (CATS) for cancer. J Control Release 175:43–47 75. Glassmann A, Winter J, Kraus D et al (2014) Pharmacological suppression of the Ras/MAPK pathway in thyroid carcinoma cells can provoke opposite effects on cell migration and proliferation: the appearance of Yin-Yang effects and the need of combinatorial treatments. Int J Oncol 45:2587–2595 76. Cattaneo M, Baronchelli S, Schiffer D et al (2014) Down-modulation of SEL1L, an unfolded protein response and endoplasmic reticulum-associated degradation protein, sensitizes glioma stem cells to the cytotoxic effect of valproic acid. J Biol Chem 289:2826–2838 77. Cheng P, Levesque MP, Dummer R, Mangana J (2020) Targeting complex, adaptive responses in melanoma therapy. Cancer Treat Rev 86:101997

Chapter 7 Design of Personalized Neoantigen RNA Vaccines Against Cancer Based on Next-Generation Sequencing Data Begon˜a Alburquerque-Gonza´lez, Marı´a Dolores Lo´pez-Abella´n, Gine´s Luengo-Gil, Silvia Montoro-Garcı´a, and Pablo Conesa-Zamora Abstract The good clinical results of immune checkpoint inhibitors (ICIs) in recent cancer therapy and the success of RNA vaccines against SARS-nCoV2 have provided important lessons to the scientific community. On the one hand, the efficacy of ICI depends on the number and immunogenicity of tumor neoantigens (TNAs) which unfortunately are not abundantly expressed in many cancer subtypes. On the other hand, novel RNA vaccines have significantly improved both the stability and immunogenicity of mRNA and its efficient delivery, this way overcoming past technique limitations and also allowing a quick vaccine development at the same time. These two facts together have triggered a resurgence of therapeutic cancer vaccines which can be designed to include individual TNAs and be synthesized in a timeframe short enough to be suitable for the tailored treatment of a given cancer patient. In this chapter, we explain the pipeline for the synthesis of TNA-carrying RNA vaccines which encompasses several steps such as individual tumor next-generation sequencing (NGS), selection of immunogenic TNAs, nucleic acid synthesis, drug delivery systems, and immunogenicity assessment, all of each step comprising different alternatives and variations which will be discussed. Key words Vaccines, Neoantigen, mRNA, Cancer, High-throughput nucleotide sequencing, Immune response, Immunotherapy, Algorithm, Drug delivery systems

1

Introduction Over the past decades, the immune evasion exerted by tumor cells seems to be of paramount importance for cancer growth and promotion. Conversely, the use of immune checkpoint inhibitors (ICIs) that increase the level of recognition of tumor neoantigens (TNA) is considered as a major breakthrough in recent cancer therapy. However, the efficacy of ICI is highly dependent on the tumor mutation burden (TMB) and/or the abundance of sufficient immunogenic TNA. Therefore, the number of patients who could benefit from ICI is limited by the relatively low TMB exhibited by

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_7, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

165

166

Begon˜a Alburquerque-Gonza´lez et al.

most cancers. In this scenario, the appearance of RNA vaccines against SARS-nCoV-2 has made researchers and oncologists consider, more than ever, the prospect of efficient development of personalized TNA-carrying therapeutic vaccines against cancer. Moreover, personalized antitumoral RNA vaccines seem feasible as anti-COVID vaccines were developed and manufactured in less than 1 year, which is a reasonable time for the tailored treatment of a given cancer patient. Previous failure history of cancer vaccine considered the use of tumor-associated antigens (TAA), which are expressed in a healthy tissue but overexpressed in tumor or cancer germline antigens, which are expressed only in germline cells and tumor cells. Nevertheless, improved DNA and RNA sequencing techniques, such as next-generation sequencing (NGS), have allowed affordable and quick high-throughput scrutiny of TNA in a given tumor tissue. In parallel, ICI and chimerical antigen receptor (CAR) T-cell successful experiences have taught us that TNA-carrying vaccines may be more effective than those containing TAA or germline antigens. In addition, TNAs offer the possibility of priming the immune system with tens of antigens at the same time which is an important advantage to tackle tumor heterogeneity [1]. However, tumors with high TMB, such as lung cancer and melanoma, can harbor hundreds of TNA, thus making necessary the selection of a few of them to synthesize the RNA vaccine. There are several bioinformatic tools aiming to predict which TNA will be more immunogenic, mainly based on its potential binding to major histocompatibility complex (MHC), and, therefore, good candidate to be included in the vaccine. In any case, the concomitant use of ICIs and TNA-carrying vaccines seems promising as the former unblock the immune system, whereas the latter prime T-cells against the tumor [1, 2]. Once the TNA have been selected in silico, the RNA vaccine developing process commences with the construction of the plasmid DNA (pDNA) which contains the TNA-coding DNA sequence. The problems previously encountered with mRNA vaccines have been solved to improve translational efficiency, mRNA stability, and protein production by taking into consideration the critical quality attributes (CQA) which include 7-methylguanosine (7mG) 50 capping; UTR structure, length, and regulatory elements; codon optimization and modifications (e.g., pseudouridine or N-1methylpseudouridine); poly-A tail properties; and purification of the mRNA from contaminants such as double-stranded RNA. Additionally, the delivery of mRNA is crucial for vaccine efficiency as it must protect mRNA from nuclease degradation, lead the vaccine to the target organ, facilitate cellular uptake, and prevent from endosomal degradation. In this regard, lipid nanoparticles formulations seem to sufficiently achieve all of these requirements

Anti-Tumoral Neoantigen RNA Vaccines

167

[3]. Of note, an important verification before stepping forward into in vivo studies is the performance of assays aiming to evaluate the immunogenicity of synthesized vaccines. Given the novelty of this approach, this chapter, rather than having a fixed protocol format, will deal with the different procedures available in each step of the briefly commented pipeline. Finally, a summary of preclinical and clinical trials using TNA-carrying vaccines will be commented. The production and delivery of neoantigen-based RNA vaccines are still far from a straightforward pipeline, and different alternatives in each of its steps demonstrate that there is still much work to do on optimization and standardization. Nevertheless, a growing number of companies, including Thermofisher, Creative Biolabs, and Sartorius (see Note 1), are providing reagents and kits for each step of the workflow following GMP criteria. Although important limitations in the RNA stability and delivery have been overcome, the selection of which mutations could code for the best neoantigens is probably the most uncertain factor in neoantigen-based vaccine development because the understanding of how immunogenic a somatic mutation can be is as complex as the immune system works. In this scenario, the use of artificial intelligence for a competent neoantigen prediction based on NGS data will certainly be an inflection point in the development process of efficient RNA vaccines.

2

Methods A general representation of the workflow for producing RNA vaccines based on tumor neoantigens is illustrated in Fig. 1. Likewise, a summary of the different alternatives associated with each step of this process is included in Table 1.

2.1 NGS Technologies for Cancer Vaccine Development 2.1.1 Sequencing Technology for Personalized Medicine

The fact that NGS has become faster and affordable has boosted the development of personalized medicine and, in particular, neoantigen prediction and cancer vaccine development. For this purpose, NGS provides data from the entire genome (whole genome sequencing, (WGS)), protein-coding genes (whole exome sequencing (WES)), RNA sequencing (transcriptome sequencing or RNAseq), and exome sequencing of certain genes (targeted exome sequencing (TES) or specific hot spots (panel sequencing, (PS)). There are two main platforms to perform NGS in clinical settings: Illumina and Ion Torrent. These platforms present a similar workflow (library preparation, sequencing, and data analysis) and differ in their core chemistry method of assessing base sequencing. Illumina offers several systems: HiSeq2000, MiSeq, NextSeq 550, and NovaSeq 6000, being the last two specifically for neoantigen detection [4]. This technology is based upon bridge

168

Begon˜a Alburquerque-Gonza´lez et al.

Fig. 1 General pipeline for the synthesis of RNA vaccines based on tumor neoantigens. WES whole genome sequencing, MHC major histocompatibility complex. (Created with BioRender.com)

amplification and immobilization on a flow cell, and their sequencing methodology consists in cyclic reversible dye chain termination. Ion Torrent PGM or Ion Torrent S5XL includes a DNA template from an arranged library into a droplet, i.e., the bead emulsion, and the sequencing method depends on pH measurement [5]. Illumina stands as a well-known option for sequencing short sequences up to 300 base pair length, and it takes 4–144 h to perform the complete task, whereas Ion Torrent can read up to 600 base pairs, and its run time is much shorter, from 2 to 10 h approximately [6]. Lately, new generations of sequencing methods are arising. Third-generation sequencing, including the MinION from Oxford Nanopore Technologies (ONT) and the single-molecule real-time (SMRT) technology from Pacific Bioscience (PacBio), aims to read long sequences up to 30,000 bases (30 kb) in length in real time [7, 8]. There is also a fourth-generation sequencing system, which permits in situ sequencing in fixed tissue and cells, though it is necessary to improve the efficiency of this technique in terms of standardization, cost-effectiveness, practicality, and full integration into the current sequencing system [9]. Nevertheless, NGS is still considered the “gold-standard” preferred high-throughput option for clinical applications.

Anti-Tumoral Neoantigen RNA Vaccines

169

Table 1 Descriptions of steps and alternatives in the synthesis of neoantigen-based RNA vaccines Next-generation sequencing Sequencing technology for personalized NGS (gold standard) medicine Illumina Ion Torrent Third-generation sequencing MinION from ONT SMRT technology from PacBio Fourth-generation sequencing system: in situ sequencing in fixed tissue and cells Variant detection or somatic mutation Bioinformatic analysis: calling WGS, WES, and RNAseq data as input HLA typing

Obtain patient HLA haplotype Compare RNAseq or WES obtained by NGS Predict neoantigen-MHC binding affinity

Integrated pipelines

pVAC-Se CloudNe TIminer INTEGRATE-Neo

Neoantigen selection Predicting peptide processing and MHC ligandome prediction

Form a stable peptide-MHC complex Select proteolyis predictors NetChop20S ProteaSMM PepCleaveCD4 High-throughput approaches: calculating the affinity of wildtype and mutant peptides to the patient’s MHC Class I and II

mRNA vaccine synthesis mRNA construct types

Non-replicating mRNA (NRM) Encodes the coding sequence (CDS) Flanked by 50 and 30 untranslated regions (UTRs), a 50 -cap, and a 30 -poly-(A) tail Formulated in lipid nanoparticles (LNPs) Translated immediately by ribosomes to produce the protein of interest Self-amplifying mRNA (SAM) Encodes additionally a replicase able to direct intracellular mRNA amplification Formulated in lipid nanoparticles (LNPs) Translated immediately by ribosomes to produce the replicase needed for self-amplification of the mRNA (continued)

170

Begon˜a Alburquerque-Gonza´lez et al.

Table 1 (continued) Production of mRNA cancer vaccines Design of the antigen in silico Design and synthesis of linear DNA Manufacturing process: Generation of a plasmid DNA (pDNA) Linearized pDNA serves as a template for the DNA-dependent RNA polymerase to transcribe the mRNA In vitro transcription of mRNA: Addition of the 50 cap and the 30 poly (A) tail Removal of DNA template by DNase Purification of mRNA: The in vitro transcribed mRNA is purified using: LiCl/NaCl-EtOH precipitation dT microbeads Administration and delivery to antigen-presenting cells (APCs)/ host cell Formulation and delivery strategies Formulation

Simple formulation Naked plasmid DNA (ex. ZyCoV-D) Complex formulation Lipid or proteolipid nanoparticles Calcium phosphate nanoparticles Polymer-based nanostructured biomaterials known as polymeric nanoparticles (NPs) PLGA PLA Advanced formulations Lipid-polymer hybrid nanoparticles Encapsulated in gold nanoparticles Magnetic vectors (SPIONs) Polyethylenimine (PEI) Hyaluronic acid (HA) Encapsulated in virus-like particles (VLPs) Protein-DNA complexed nanoparticle Using living organisms. poxviruses and adenoviruses

Cell-targeted delivery (mainly antigen-presenting cell)

Antibody-based Ligand-based

Administration routes

Classical intramuscular injection. In vivo electroporation (Cliniporator) Needle-free injection systems (NFIS) Intravenous or intralesional injection Orally Intratumoral injection (continued)

Anti-Tumoral Neoantigen RNA Vaccines

171

Table 1 (continued) Immunogenicity assessment In vivo measures of antigen-specific immunity

Delayed-type hypersensitivity testing (DTH test)

In vitro measures of antigen-specific immunity

Phenotypic measures of antigen-specific cellular immune responses Flow cytometry using peptide major histocompatibility complex tetramers In Vitro functional measures of antigen-specific immune responses Lymphoproliferation assay ELISA and ELISpot Cytokine flow cytometry Quantitative reverse transcriptase polymerase chain reaction Direct cytotoxicity assay Limiting dilution analysis

2.1.2 NGS for Neoantigen Vaccine Development

The first step in developing a neoantigen vaccine is to identify and confirm patient-specific immunogenic non-synonymous somatic mutations expressed in the tumor, such as single nucleotide variations (SNV), deletion-insertions, or intron retentions, comparing the sequences of the tumor and matched healthy tissues or blood samples. Tumor biopsy and healthy tissue from peripheral blood mononuclear cells or the surrounding area of the lesion are the starting material for DNA and RNA extraction [10]. After sample collection, sequencing by NGS is performed on both tumor and healthy cells from the same patient. WES data from the tumor sample and non-transformed cells is to detect non-synonymous somatic mutations in a process called variant calling, whereas RNAseq analysis constricts the analysis to mutations of expressed genes [11]. Likewise, NGS technology can be used to assess changes in biomarkers of ICIs efficiency, such as TMB, microsatellite instability, and PD-L1 amplification and other therapeutic effects, such as drug resistance and progression-related genetic mutations [12]. An example of this technology would be the Oncomine Tumor Mutations Load Assay, a NGS-based assay for the assessment of tumor mutational burden [13].

2.1.3 NGS for Variant Detection or Somatic Mutation Calling

During the bioinformatic analysis of the raw data obtained by NGS, there are two main processes to perform: sequence alignment and variant calling. Both of them use WGS, WES, or RNAseq data as input. First, the alignment of tumor/normal sequencing data to the human reference genome permits the identification of variants or differences from the non-cancerous normal reference genome. Some available alignment tools are Bowtie2, Novoalign, or SOAPv2 [14]. Following the sequence alignment, genetic

172

Begon˜a Alburquerque-Gonza´lez et al.

alterations in the tumor can be detected and annotated using variant-calling algorithms, such as GATK, VarScan2, FACTERA, ANNOVAR MUTECT, or EBCall [15]. 2.1.4

NGS for HLA Typing

Before manufacturing the neoantigen vaccine with the variants obtained, it is vital to select the neoepitopes encoded by somatic mutations in the tumor with the highest likelihood of being presented by specific MHC molecules of a given patient [16]. To this effect, the HLA haplotype from the patient must be obtained, either by clinical assays or with HLA typing tools. HLA typing algorithms compare RNAseq or WES obtained by NGS from the patient peripheral blood nuclear cells with published reference sequences, selecting for alignments with the least number of mismatches to determine HLA type [15]. HLAminer and Seq2HLA are examples of HLA typing tools that can be used, though they are prone to sequencing errors and low accuracy [17]. Other options are ATHLATES and SOAP-HLA [18, 19], whereas more recent programs are Poysolver or Optitype, with improved accuracy and sensitivity [20]. Afterward, it will be necessary to predict neoantigen-MHC binding affinity, using sequence-based methods that focus on the primary protein sequences, such as NetMHC and NetMHCpan (also TMHMM) [21].

2.1.5 Integrated Pipelines

Many of the previously described steps have been incorporated as neoantigen prediction pipelines. In general, these tools require an annotated list of non-synonymous somatic mutations obtained from NGS data (WGS, WES, or RNAseq) and the patient’s HLA haplotype. Examples are pVAC-Seq [22], CloudNeo [23], TIminer [24], or INTEGRATE-Neo [25].

2.1.6 NGS Challenges for Neoantigen Vaccine Development

Despite NGS greatly reducing the cost and time required for whole-exome and RNA sequencing, the sequencing performance is still highly dependent on the quality of the biopsy and the heterogeneity of the tumor, thus representing a challenge to identify neoantigens relevant to the entire tumor [26]. Besides, a large patient sample is needed, which is not always available. It is also worth mentioning that clonal hematopoiesis somatic mutation can lead to likely misattribution of tumor mutation results on unpaired NGS experiments [27] and even using also blood paired sequencing along with that from the solid tumor; an inconvenient of variant calling tools would be the generation of false-positive mutations due to sequencing errors and misalignment, which can be reduced by increasing the number of algorithms used to determine potential variants [16]. Finally, another source of error is that current algorithms are heavily biased toward detection of SNVs over other mutations and the low expression of HLA class II among tumors that makes it challenging to identify HLA class II epitopes [28].

Anti-Tumoral Neoantigen RNA Vaccines

173

2.2 Neoantigen Selection

Numerous preclinical studies in mice with immunotherapy have given a beacon of hope for cancer treatment. However, the clinical application of this approach, as well as the identification of neoantigens that cause therapeutic immunity, has proven to be difficult, hampering the rapid development of neoantigen-based personal vaccines. Neoantigens are created by cancer-specific DNA somatic mutations, resulting in specific peptide sequences. Indeed, neoantigens are ideal targets since they are not expressed in normal tissues, they have not been previously recognized by the body’s immune system, and they are then more easily recognized by the T-cells. However, certain criteria must be met for mutations to create neoantigens: (1) they must be present within peptides that are processed by the antigen presentation machinery; (2) the mutated peptide must bind to MHC with enough affinity to present itself to the T-cells; and (3) the mutated peptide bound to MHC must be recognized by the patient’s T-cells. In the following section, we will describe several successful bioinformatic tools that have appeared in recent years for neoepitope prediction and neoantigen selection for personalized immunotherapy. Besides, we will also mention the current challenges that handicap the predictive performance of these approaches and open the possibilities to improve the efficacy of such personalized treatments. The prediction of neoantigens is a multi-step screening process which includes somatic mutation identification, HLA haplotyping, peptide processing, and peptide-MHC binding prediction.

2.2.1 Somatic Single Nucleotide Variants

Tools available for tumor-related SNVs have been implemented over the last years. Accordingly, different variant callers such as MuTect2 or VarScan2 have been extensively used [29]. Interesting comparative studies have shown that both softwares have remarkable sensitivity to detect somatic mutations for variants with high allele fractions [30–32]. Other popular somatic SNV callers are SomaticSniper and Strelka [33, 34]. It has been suggested the utility of using the consensus yields from multiple algorithms to obtain a more trustworthy prediction for a missense variant [31], because when results from different variant callers agree, they lean toward agreeing for true positive variants, but when they disagree, these variants are more probably false positives due to PCR artefacts, tissue impurity, somatic allele frequencies, and genetic heterogeneity [35]. Of note, no consensus has been reached about how to process the DNA for the mutanome definition.

2.2.2 Human Leukocyte Antigen (HLA) Haplotyping

Human MHC (major histocompatibility complex), also known as HLA, is the entire set of genes that code for highly polymorphic proteins involved in antigenic peptide presentation to the epitopespecific T-cells. T lymphocyte recognition of neoantigens depends on a well-fitting interaction of HLA molecule and the epitope. This

174

Begon˜a Alburquerque-Gonza´lez et al.

represents a challenge to the prediction of MHC-neoantigens since accurate HLA haplotyping is a key factor for binding preferences and should not be relegated to conventional clinical typing [36]. This step was incorporated in the trials during the last decade, since before, HLA genotype determination was only used in six out of 45 trials (13%) [37]. HLA typing is currently being analyzed based on NGS technology due to the high variability of the HLA loci. Initially, Warren et al. proposed an algorithm (HLAminer) based on allele-specific scoring for WESm [38]. Then, Seq2HLA algorithm used standard RNA-Seq reads [38]. Back then, several modifications have been included to the following algorithms using RNA sequencing data such as Optitype [20] and Polysover [17]. The accurate somatic HLA mutations detection and the copy number variations in the HLA locus [39] based on NGS are still accompanied by labor-intensive preparations and remain timeconsuming. Kiyotani et al. nicely compared several bioinformatic algorithms and reported Optitype as the highest performing approach with 97.2% accuracy [40]. As happened with SNVs detection, combination of HLA-typing approaches may be used for optimizing different aspects of HLA mutation detection performance in order to increase the accuracy at the cost of sensitivity. 2.2.3 Predicting Peptide Processing

For a target peptide to serve as a natural T-cell neoantigen, the parent protein must first be processed so that its peptides can be presented in a molecule of MHC and, thus, accessible for T-cell recognition. Biological processing and presentation of an antigen are complex and multifactorial processes. This is one of the main causes why most peptides that are predicted to display high-affinity MHC binding affinity are not able to cause T-cell responses. For a peptide to form a stable peptide-MHC complex and be clinically relevant, it should strive to overcome the immune proteasome processing machinery. Indeed, even if a peptide has a strong MHC binding affinity prediction, the upstream immune processing is crucial for the transport of the peptides to the appropriate cellular compartments and for the interaction with MHC molecules [41]. Several elements are at the core of the proteasome such as the transporter associated with antigen processing (TAP), ER aminopeptidase (ERAPs), and tapasin, among others. To study proteolysis and to aid MHC-I ligand predictions, different proteolyis predictors were developed long ago such as NetChop20S [42] and ProteaSMM [43]. PepCleaveCD4 has been more recently used for predicting naturally processed MHC II ligands [44]. Other prediction methods for proteasomal cleavages also include TAP transport/MHC I which outperform earlier methods based on in vitro cleavage data [42]. Computational tools for MHC II-peptide processing prediction are encouraged due to their implication in cancer therapeutic research.

Anti-Tumoral Neoantigen RNA Vaccines

175

2.2.4 MHC Ligandome Prediction

As it has been already mentioned, the neoepitope candidates still need to be selected using other combination of factors such as the neoepitope abundance and the MHC ligandome prediction (after HLA typing). High-throughput approaches have showed a great interest to rank the candidates by calculating the affinity of wild-type and mutant peptides to the patient’s MHC Class I and II (NetMHCpan 4.0 and NetMHCIIpan 3.2) [45]. However, it is important to mention that MHC II displays open binding pockets which lead to a more challenging affinity prediction. In fact, several reports have selected hundreds of tumor-specific neoantigens, but after in vitro experiments, only few of them were able to stimulate T-cells. This fact, together with the huge amounts of mutated peptides recognized as exogenous peptides (CD4+ T-cell recognition), encourages the rapid development of bioinformatic approaches that implement in silico data for MHC affinity binding prediction. Indeed, traditionally NetMHC was based on the sequence of the MHC gene for the prediction, while now, the last versions are much more accurate because they include pan-allele specific machine learning methods including 3D protein such as MHCflurry and LC-MS/MS datasets [46, 47]. These tools currently obtain MHC-binding affinity of untrained alleles, on the basis of neural network-based machine learning NNAlign algorithms [21]. As occurs with the above factors, the comparison between results on MHC-binding and MHC-elution epitopes seems to present the best appropriate option for neoepitope prediction [48].

2.3 mRNA Cancer Vaccines

The manufacture of mRNA vaccines is cell free and simpler than that of DNA vaccines, because it generally uses linear DNA molecules or libraries of cDNA for in vitro transcription, followed by purification of the chemically synthesized mRNAs [49]. mRNA vaccines can deliver a high number of antigens and co-stimulatory signals, with no risk of infection or insertional mutagenesis, and manufacturing is rapid and inexpensive [50]. As the majority of neoantigens are unique to each patient’s cancer, a vaccine technology that is flexible and potent is required to develop personalized neoantigen vaccines [51]. Two types of mRNA constructs are being currently evaluated for the design of mRNA cancer vaccines: • Non-replicating mRNA (NRM) whose constructs encode the coding sequence (CDS) and are flanked by a 50 -cap structure, a 30 -poly-(A) tail, and 50 and 30 untranslated regions (UTRs). • Self-amplifying mRNA (SAM) whose constructs additionally encode a replicase able to trigger intracellular mRNA amplification.

176

Begon˜a Alburquerque-Gonza´lez et al.

Both construct types contain a cap structure, 50 and 30 untranslated regions (UTRs), open-reading frame (ORF), and a 30 poly (A) tail [3]. 2.3.1 Production of mRNA Cancer Vaccines

mRNA cancer vaccine synthesis commences with the in silico design of the antigen (selection of amino acids sequences of neoantigens) which has the advantage of rapid production and preclinical evaluation. Despite sequence differences for each neoantigen, the production process is standard, reducing the timing and cost. The first step is the generation of a pDNA which contains a DNA-dependent RNA polymerase promoter, such as T7, followed by the appropriate sequence for the mRNA construct. Next step is the linearization of the DNA template to produce multiple copies of the coded mRNA using RNA polymerases (T3, T7, SP6) to transcribe the mRNA [51]. After DNA linearization, the in vitro transcription of mRNA is produced. In this step, the incorporation of the 50 cap and the 30 poly (A) tail can be carried out. The enzymatic addition of the cap can be achieved by using guanylyl transferase and 20 -O-methyltransferase to produce a Cap 0 (N7MeGpppN) or Cap 1 (N7MeGpppN20 -OMe) structure, respectively. In turn, the addition of poly-A tail can be accomplished via enzymatic addition via poly-A polymerase [3]. Then the DNase is used to remove the DNA template, and the in vitro transcribed mRNA is purified by means of dT microbeads or precipitation with LiCl/NaCl-EtOH. Contaminants such as doublestranded RNA (dsRNA) that could result from RNA polymerase over-activity can be eliminated with different treatments such as cellulose purification, HPLC purification, and RNase III enzymatic treatment [51]. The drug substance obtained is then formulated under sterility, identity, purity, and potency testing conditions which allow good manufacturing practice (GMP) facilities to switch to a new vaccine type within a very short period of time, given that the reagents and materials are the same. Finally, the administration and delivery to antigen-presenting cells (APCs)/host cell was taken [3].

2.4 Formulation and Delivery Strategies

Modern tailored vaccines in cancer treatment can be DNA or RNA. RNA vaccines can be formulated naked, complexed, or inside of some types of molecular transporters [52]. RNA vaccines began development in the 1990s. Due to the labile nature of RNA, this kind of vaccines are, in general terms, more unstable than those of DNA, and, therefore, more complex formulations, straightforward manufacturing, and expensive conditions for storage and transport are needed. There are two main approaches of RNA vaccines: non-replicating mRNAs and virally derived, self-amplifying RNAs. Although both encode the selected antigen/s, the self-amplifying RNAs encode, in addition, the viral

Anti-Tumoral Neoantigen RNA Vaccines

177

Fig. 2 Description of different formulation and delivery strategies for RNA vaccines. (Created with BioRender. com)

replication machinery that enables RNA amplification and protein expression [53]. Some of the possible formulations and delivery strategies or RNA vaccines are described in Fig. 2. 2.4.1 RNA Vaccine Formulation

1. Naked and Unmodified RNA Naked RNA is the simplest formulation but in this case is exposed to extracellular RNases, and their entries into cells are quite limited. This methodology has demonstrated effectivity for immunization in vivo, especially when is administered by intradermal or intranodal injection and without any modification of the desired RNA [54–56].

178

Begon˜a Alburquerque-Gonza´lez et al.

2. Associated with Protamine Although RNA can be administered naked, formulation with cationic peptides like protamine has demonstrated improvements in RNA protection against RNAses [57], and, therefore, this strategy may increase the half-life of the RNA into the body, although it has demonstrated limited protein expression and efficacy as anticancer vaccine. This setback is possibly due to a strong association between protamine and mRNA that limits their binding capacity to ribosomes [58]. In order to overcome this problem some improvements in this formulation have been carried out [59]. 3. Associated with Cationic Particles Most complex formulations include lipid or proteolipid cationic nanoparticles which improve delivery of RNA to host cells by promotion of nanoparticle-driven vesicle fusion and internalization of the nucleic acid [60]. Another option is the use of calcium phosphate nanoparticles, which are biodegradable vehicles with low toxicity and high loading capacity of RNA [61]. Recent biotechnological advances have enabled the use of polymer-based nanostructured biomaterials known as polymeric nanoparticles (NPs). For example, the polyester-based nano-vehicles, including poly(lacticco-glycolic acid) (PLGA) and polylactic acid (PLA) NPs, showed effectivity for delivering nucleic acids, including RNA [62]. More advanced formulations under development are lipid-polymer hybrid NPs, which offer the ability to combine the advantages of both polymer and lipid materials in a single design for efficient transport and delivery of nucleic acids [63]. RNA can also be encapsulated in gold nanoparticles, such as partially PEGylated dendrimer-entrapped gold nanoparticles [64–66] and even in magnetic vectors such as super-paramagnetic iron oxide NPs (SPIONs), polyethylenimine (PEI), and hyaluronic acid (HA) [67, 68]. As one of the disadvantages of RNA is their short half-life, injectable hydrogels of graphene oxide and PEI can release mRNA and adjuvants for at least 30 days after subcutaneous injection, extending mRNA half-life, increasing the number of antigen-specific CD8+ T-cells, and inhibiting the tumor growth with only one shot. Other suitable vehicles for nucleic acid formulation are encapsulated in virus-like particles (VLPs), and protein-RNA complexed nanoparticles [69–71], which are both under development [72]. 2.4.2 Cell-Targeted Delivery

Therapeutic approaches for achieving strong immune responses against solid tumors, especially metastatic, may require celltargeted rather than untargeted delivery. Biodistribution of nanoparticles carrying mRNAs can be modulated by further adapting and optimizing the surface of the nanoparticle. For example, nanoparticles can be covered with antibodies or ligands to deliver

Anti-Tumoral Neoantigen RNA Vaccines

179

mRNA molecules into desired cell targets like antigen-presenting cells (APC, i.e., dendritic cells, Langerhans cells, macrophages, or B cells) [73, 74]. Nanoparticles of oligopeptide end-modified poly (β-amino esters) (OM-PBAEs) can be complexed with mRNA and form discrete nanoparticles. Using an appropriate end-oligopeptide modification, this system enables the specific targeting and major transfection of APC in vivo, after intravenous administration [75, 76]. Organ targeted therapy can also be achieved by modifying the ratios of lipid components in the case of lipid nanoparticles [73] or even using organ-specific antibodies or ligands. 2.4.3 Administration Routes

Delivery of RNA vaccines can be performed by classical intramuscular injection in the form of naked RNA or, alternatively, with nucleic acids complexed in cationic nanoparticle vectors. This approach has been used by modern SARS-CoV-2 vaccines like Comirnaty (BioNTech/Pfizer) or mRNA-1273 (Moderna). One way of improving efficiency is using in vivo electroporation medical devices such as Cliniporator [77]. There are several clinical trials using electroporation for gene therapy and for vaccination, in particular against SARS-CoV-2 (NCT04447781, NCT04788459, NCT04627675). Microbiota should be taken into account since antitumor effectiveness of a neoantigen cancer vaccine delivered by electroporation is influenced by microbiota composition [78]. Recent advances in subcutaneous injections have provided novel methods for delivery of nucleic acids, such as needle-free injection systems (NFIS), which reduces pain and anxiety and eliminates safety risks associated with needlestick injuries in patients and in sanitary professionals [79]. These injector systems produce a stream of pressurized solution that infiltrates up to 2 mm in the skin at high speed, resulting in uniform dispersion and higher uptake of the molecules in the surface of the skin [80]. Intranodal as well as intradermal injection has been demonstrated as useful delivery routes for RNA vaccines [54–56], especially for naked RNA. Finally, ex vivo loading of primary dendritic cells transfected in vitro with mRNA is another interesting method to generate cellular immunity against cancer, and there are several research lines in development [53, 81], as well as some clinical trials (NCT04335890, NCT04105582).

2.5 Immunogenicity Assessment

Common assays for measuring the immune response of cancer vaccines need to be established so that these assays can 1 day serve as surrogate markers for clinical response. However, increases in the number of cytotoxic T-cells through immunization have not been correlated with clinical tumor regression. Therefore, a T-cell assay not only needs to be sensitive, specific, reliable, reproducible, simple, and quick to perform but must also demonstrate close correlation with clinical outcome. Assays currently used to measure T-cell response are discussed below.

180

Begon˜a Alburquerque-Gonza´lez et al.

2.5.1 Measures to Assess In Vivo AntigenSpecific Immunity

1. Delayed-type hypersensitivity testing (DTH test). This assay usually consists of the following steps: • Small quantities of antigen are injected intradermally. • The response was mediated by CD4+T-helper and CD8+Tcells. • The diameter of the erythema at the site of injection is measured after 48–72 h [82].

2.5.2 Measures to Assess In Vitro AntigenSpecific Immunity

1. Phenotypic measures of cellular immune responses against specific antigens. • Flow cytometry using peptide tetramers of the MHC. Fluorochrome-labeled streptavidin is added to purified, biotinylated MHC/peptide “monomers” to form soluble tetrameric complexes. Fluorescent MHC/peptide tetramers bind to T-cells bearing MHC/peptide-specific T-cell receptors (TCRs) and can be detected by flow cytometry [82]. Though this method detects and quantifies changes in peripheral T-cell proportions, its application in cancer vaccine trials has not yet been associated with clinical outcome [83]. 2. In vitro functional measures of antigen-specific immune responses • Lymphoproliferation assay Purified T-cells or peripheral blood mononuclear cells (PBMCs) are exposed to antigen with or without the presence of stimulating cells. Lastly, after 72–120 h [3H]-Thymidine is added to quantify DNA synthesis by the amount of incorporated radioisotope. As an advantage, the assay is directly performed on peripheral blood samples, thus reporting the in vivo T-cell activity. Still, it has not been proven to correlate with disease outcome, and its performance can vary among patients depending on the nonspecific immune function [82]. • Detection of secreted cytokines by enzyme-linked immunosorbant assay (ELISA) and enzyme-linked immunospot assay (ELISpot) – ELISA Cytokine secretion by T-cells in response to antigen may be detected by measuring either bulk cytokine production. The PBMC supernatant is incubated with the antigen to be subsequently harvested and placed to microtiter plates coated with the antibody against the cytokine of interest. Linked antibodies to a reporter molecule are added, and the plates are washed and read.

Anti-Tumoral Neoantigen RNA Vaccines

181

– ELISpot Antigen-induced cytokine secretion by T-cells may be detected by enumerating individual cytokineproducing T-cells. In a microtiter plate, an antibody against the cytokine of interest attached to a solid-phase membrane is used. Unseparated PBMCs or isolated CD8+or CD4+ T-cells are treated in the antibody-coated wells with an antigen. Once the antigen recognition occurs, T-cells release the cytokine of interest, which will be bound by the well-coated antibody. After washing, the release of the cytokine is determined by an enzyme-labeled detection antibody and the corresponding chromogenic substrate [82]. • Cytokine flow cytometry Flow cytometry is used for the direct detection of intracellular cytokine expression with antibodies against the fluorochrome-conjugated cytokine after periodical activation driven by various stimuli. The stimulation is carried out with mononuclear cells obtained from peripheral blood, lymph nodes, or biological fluids. After this stimulus, the cells are fixed, permeabilized, and stained with several antibodies [84]. • Quantitative reverse transcriptase polymerase chain reaction The peptide used in the vaccine or an irrelevant peptide as control is incubated with the cells of interest for 2 h. Total RNA isolation and subsequent quantitation by qRT-PCR of the cytokine mRNA expression in the samples are then carried out [82]. • Direct cytotoxicity assay Cytotoxicity assays are interesting because the capacity of CD8+ CTLs to lyse tumor is considered as a relevant surrogate of in vivo antitumor activity. This method includes: • Treating the T-cells or PBMCs specimen with antigenexpressing targets labelled with chromium-51 or europium. • Measuring the release of chromium or europium after cell lysis [84]. • After incubations, the lysis percentage of the targets is calculated by comparing it with the maximum possible lysis of the target. • Limiting dilution analysis T-cell number is correlated from a functional activity. The technique steps involve serial dilution of T-cells in a large number of wells, an in vitro stimulation a finally target lysis [82].

182

Begon˜a Alburquerque-Gonza´lez et al.

Several advanced monitoring assays are used nowadays to characterize phenotypical and functional features of antitumor T-cells in immuno-oncology trials. They include T-cell proliferation, cytokine profile, cytotoxic T lymphocytes (CTL) assays, CTL-associated molecules (CD107, perforin, granzyme B, and CD154), and MHC-multimer analysis. Still, as commented before, these assays do not efficiently correlate immune response and clinical outcome which make evident the methodological limits of immunologic assays or the post-vaccination absence of antitumor effects appropriately consistent to increase progression-free or overall survival [85]. Although the gold standard in vaccine efficacy evaluation is the clinical benefit, assessing the antitumoral effects of the RNA vaccine on animal models may provide a proper understanding of dose optimization, delivery, scheduling, and combination therapy before clinical trials [82].

3

Notes 1. Useful links describing available resources and services offered by biotechnological companies for RNA synthesis (accessed September 2021): Thermofisher:https://www.thermofisher.com/es/es/home/ biotech-lab-solutions/therapeutics-research-develop ment-solutions/vaccine-research-development-tools.html Creative Biolabs: https://www.creative-biolabs.com/vaccine/ custom-gmp-grade-mrna-synthesis-service.html Sartorius:https://www.sartorius.com/en/applications/bio pharmaceutical-manufacturing/vaccines/vaccine-develop ment/mrna-vaccines Trilink:https://www.trilinkbiotech.com/custom-mrnasynthesis

References 1. Mullard A (2016) The cancer vaccine resurgence. Nat Rev Drug Discov 15:663–665 2. Tahtinen S, Feola S, Capasso C et al (2020) Exploiting preexisting immunity to enhance oncolytic cancer immunotherapy. Cancer Res 80:2575–2585 3. Jackson NAC, Kester KE, Casimiro D et al (2020) The promise of mRNA vaccines: a biotech and industrial perspective. NPJ Vaccines 5:11 4. Garofalo A, Sholl L, Reardon B et al (2016) The impact of tumor profiling approaches and

genomic data strategies for cancer precision medicine. Genome Med 8:79 5. Mehrotra M, Duose DY, Singh RR et al (2017) Versatile ion S5XL sequencer for targeted next generation sequencing of solid tumors in a clinical laboratory. PLoS One 12:e0181968 6. Xuan J, Yu Y, Qing T et al (2013) Nextgeneration sequencing in the clinic: promises and challenges. Cancer Lett 340:284–295 7. Lu H, Giordano F, Ning Z (2016) Oxford Nanopore MinION sequencing and genome

Anti-Tumoral Neoantigen RNA Vaccines assembly. Genomics Proteomics Bioinformatics 14:265–279 8. Eid J, Fehr A, Gray J et al (2009) Real-time DNA sequencing from single polymerase molecules. Science 323:133–138 9. Yin Q, Tang J, Zhu X (2019) Next-generation sequencing technologies accelerate advances in T-cell therapy for cancer. Brief Funct Genomics 18:119–128 10. Tay BQ, Wright Q, Ladwa R et al (2021) Evolution of cancer vaccines-challenges, achievements, and future directions. Vaccines (Basel) 9:535 11. Kvistborg P, Clynes R, Song W et al (2016) Immune monitoring technology primer: whole exome sequencing for neoantigen discovery and precision oncology. J Immunother Cancer 4:22 12. Wu TM, Liu JB, Liu Y et al (2020) Power and promise of next-generation sequencing in liquid biopsies and cancer control. Cancer Control 27:1073274820934805 13. Mankor JM, Paats MS, Groenendijk FH et al (2020) Impact of panel design and cut-off on tumour mutational burden assessment in metastatic solid tumour samples. Br J Cancer 122: 953–956 14. Vestergaard LK, Oliveira DNP, Hogdall CK et al (2021) Next generation sequencing technology in the clinic and its challenges. Cancers (Basel) 13:1751 15. Lancaster EM, Jablons D, Kratz JR (2020) Applications of next-generation sequencing in neoantigen prediction and cancer vaccine development. Genet Test Mol Biomarkers 24: 59–66 16. Hu Z, Ott PA, Wu CJ (2018) Towards personalized, tumour-specific, therapeutic vaccines for cancer. Nat Rev Immunol 18:168–182 17. Shukla SA, Rooney MS, Rajasagi A et al (2015) Comprehensive analysis of cancer-associated somatic mutations in class I HLA genes. Nat Biotechnol 33:1152–1158 18. Liu C, Yang X, Duffy B et al (2013) ATHLATES: accurate typing of human leukocyte antigen through exome sequencing. Nucleic Acids Res 41:e142 19. Li R, Yu C, Li Y et al (2009) SOAP2: an improved ultrafast tool for short read alignment. Bioinformatics 25:1966–1967 20. Szolek A, Schubert B, Mohr C et al (2014) OptiType: precision HLA typing from nextgeneration sequencing data. Bioinformatics 30:3310–3316 21. Nielsen M, Andreatta M (2016) NetMHCpan-3.0; improved prediction of binding to MHC class I molecules integrating

183

information from multiple receptor and peptide length datasets. Genome Med 8:33 22. Hundal J, Carreno BM, Petti AA et al (2016) pVAC-Seq: a genome-guided in silico approach to identifying tumor neoantigens. Genome Med 8:11 23. Bais P, Namburi S, Gatti DM et al (2017) CloudNeo: a cloud pipeline for identifying patient-specific tumor neoantigens. Bioinformatics 33:3110–3112 24. Tappeiner E, Finotello F, Charoentong P et al (2017) TIminer: NGS data mining pipeline for cancer immunology and immunotherapy. Bioinformatics 33:3140–3141 25. Zhang J, Mardis ER, Maher CA (2017) INTEGRATE-neo: a pipeline for personalized gene fusion neoantigen discovery. Bioinformatics 33:555–557 26. Supabphol S, Li L, Goedegebuure SP et al (2021) Neoantigen vaccine platforms in clinical development: understanding the future of personalized immunotherapy. Expert Opin Investig Drugs 30:529–541 27. Coombs CC, Gillis NK, Tan X et al (2018) Identification of clonal hematopoiesis mutations in solid tumor patients undergoing unpaired next-generation sequencing assays. Clin Cancer Res 24:5918–5924 28. Chen I, Chen MY, Goedegebuure SP et al (2021) Challenges targeting cancer neoantigens in 2021: a systematic literature review. Expert Rev Vaccines 20:827–837 29. Cibulskis K, Lawrence MS, Carter SL et al (2013) Sensitive detection of somatic point mutations in impure and heterogeneous cancer samples. Nat Biotechnol 31:213–219 30. Van Horebeek L, Hilven K, Mallants K et al (2019) A robust pipeline with high replication rate for detection of somatic variants in the adaptive immune system as a source of common genetic variation in autoimmune disease. Hum Mol Genet 28:1369–1380 31. Cai L, Yuan W, Zhang Z et al (2016) In-depth comparison of somatic point mutation callers based on different tumor next-generation sequencing depth data. Sci Rep 6:36540 32. Chen Z, Yuan Y, Chen X et al (2020) Systematic comparison of somatic variant calling performance among different sequencing depth and mutation frequency. Sci Rep 10:3501 33. Larson DE, Harris CC, Chen K et al (2012) SomaticSniper: identification of somatic point mutations in whole genome sequencing data. Bioinformatics 28:311–317 34. Saunders CT, Wong WS, Swamy S et al (2012) Strelka: accurate somatic small-variant calling

184

Begon˜a Alburquerque-Gonza´lez et al.

from sequenced tumor-normal sample pairs. Bioinformatics 28:1811–1817 35. Alioto TS, Buchhalter I, Derdak S et al (2015) A comprehensive assessment of somatic mutation detection in cancer using whole-genome sequencing. Nat Commun 6:10001 36. Adams SD, Barracchini KC, Simonis TB et al (2001) High throughput HLA sequence-based typing (SBT) utilizing the ABI Prism 3700 DNA Analyzer. Tumori 87:S40–S43 37. Nagorsen D, Thiel E (2008) HLA typing demands for peptide-based anti-cancer vaccine. Cancer Immunol Immunother 57:1903–1910 38. Boegel S, Lower M, Schafer M et al (2012) HLA typing from RNA-Seq sequence reads. Genome Med 4:102 39. McGranahan N, Rosenthal R, Hiley CT et al (2017) Allele-specific HLA loss and immune escape in lung cancer evolution. Cell 171(1259–1271):e1211 40. Kiyotani K, Mai TH, Nakamura Y (2017) Comparison of exome-based HLA class I genotyping tools: identification of platformspecific genotyping errors. J Hum Genet 62: 397–405 41. Calis JJ, Reinink P, Keller C et al (2015) Role of peptide processing predictions in T cell epitope identification: contribution of different prediction programs. Immunogenetics 67:85–93 42. Nielsen M, Lundegaard C, Lund O et al (2005) The role of the proteasome in generating cytotoxic T-cell epitopes: insights obtained from improved predictions of proteasomal cleavage. Immunogenetics 57:33–41 43. Tenzer S, Peters B, Bulik S et al (2005) Modeling the MHC class I pathway by combining predictions of proteasomal cleavage, TAP transport and MHC class I binding. Cell Mol Life Sci 62:1025–1037 44. Paul S, Karosiene E, Dhanda SK et al (2018) Determination of a predictive cleavage motif for eluted major histocompatibility complex class II ligands. Front Immunol 9:1795 45. Jurtz V, Paul S, Andreatta M et al (2017) NetMHCpan-4.0: improved peptide-MHC class I interaction predictions integrating eluted ligand and peptide binding affinity data. J Immunol 19:3360–3368 46. O’Donnell TJ, Rubinsteyn A, Bonsack M et al (2018) MHCflurry: open-source class I MHC binding affinity prediction. Cell Syst 7(129–132):e124 47. Bulik-Sullivan B, Busby J, Palmer CD et al (2018) Deep learning using tumor HLA peptide mass spectrometry datasets improves neoantigen identification. Nat Biotechnol. https://doi.org/10.1038/nbt.4313

48. Zhao W, Sher X (2018) Systematically benchmarking peptide-MHC binding predictors: from synthetic to naturally processed epitopes. PLoS Comput Biol 14:e1006457 49. Jahanafrooz Z, Baradaran B, Mosafer J et al (2020) Comparison of DNA and mRNA vaccines against cancer. Drug Discov Today 25: 552–560 50. Jou J, Harrington KJ, Zocca MB et al (2021) The changing landscape of therapeutic cancer vaccines-novel platforms and neoantigen identification. Clin Cancer Res 27:689–703 51. Esprit A, de Mey W, Bahadur Shahi R et al (2020) Neo-Antigen mRNA vaccines. Vaccines (Basel) 8:776 52. Rahman MM, Zhou N, Huang J (2021) An overview on the development of mRNA-based vaccines and their formulation strategies for improved antigen expression in vivo. Vaccines (Basel) 9:244 53. Pardi N, Hogan MJ, Porter FW et al (2018) mRNA vaccines – a new era in vaccinology. Nat Rev Drug Discov 17:261–279 54. Selmi A, Vascotto F, Kautz-Neu K et al (2016) Uptake of synthetic naked RNA by skinresident dendritic cells via macropinocytosis allows antigen expression and induction of T-cell responses in mice. Cancer Immunol Immunother 65:1075–1083 55. Sahin U, Derhovanessian E, Miller M et al (2017) Personalized RNA mutanome vaccines mobilize poly-specific therapeutic immunity against cancer. Nature 547:222–226 56. Bialkowski L, van Weijnen A, Van der Jeught K et al (2016) Intralymphatic mRNA vaccine induces CD8 T-cell responses that inhibit the growth of mucosally located tumours. Sci Rep 6:22509 57. Hoerr I, Obst R, Rammensee HG et al (2000) In vivo application of RNA leads to induction of specific cytotoxic T lymphocytes and antibodies. Eur J Immunol 30:1–7 58. Schlake T, Thess A, Fotin-Mleczek M et al (2012) Developing mRNA-vaccine technologies. RNA Biol 9:1319–1330 59. Kallen KJ, Heidenreich R, Schnee M et al (2013) A novel, disruptive vaccination technology: self-adjuvanted RNActive((R)) vaccines. Hum Vaccin Immunother 9:2263–2276 60. Alfagih IM, Aldosari B, AlQuadeib B et al (2020) Nanoparticles as adjuvants and nanodelivery systems for mRNA-based vaccines. Pharmaceutics 13:45 61. Wang Y, Zhang L, Xu Z et al (2018) mRNA vaccine with antigen-specific checkpoint blockade induces an enhanced immune response

Anti-Tumoral Neoantigen RNA Vaccines against established melanoma. Mol Ther 26: 420–434 62. Piperno A, Sciortino MT, Giusto E et al (2021) Recent advances and challenges in gene delivery mediated by polyester-based nanoparticles. Int J Nanomedicine 16:5981–6002 63. Lim M, Badruddoza AZM, Firdous J et al (2020) Engineered nanodelivery systems to improve DNA vaccine technologies. Pharmaceutics 12:30 64. Hou W, Wei P, Kong L et al (2016) Partially PEGylated dendrimer-entrapped gold nanoparticles: a promising nanoplatform for highly efficient DNA and siRNA delivery. J Mater Chem B 4:2933–2943 65. Li B, Zhang X, Dong Y (2019) Nanoscale platforms for messenger RNA delivery. Wiley Interdiscip Rev Nanomed Nanobiotechnol 11:e1530 66. Graczyk A, Pawlowska R, Jedrzejczyk D et al (2020) Gold nanoparticles in conjunction with nucleic acids as a modern molecular system for cellular delivery. Molecules 25:204 67. Al-Deen FM, Xiang SD, Ma C et al (2017) Magnetic nanovectors for the development of DNA blood-stage malaria vaccines. Nanomaterials (Basel) 7:30 68. Luo X, Zhao W, Li B et al (2018) Co-delivery of mRNA and SPIONs through amino-ester nanomaterials. Nano Res 11:5596–5603 69. Shibata T, Fujita Y, Ohno H et al (2017) Protein-driven RNA nanostructured devices that function in vitro and control mammalian cell fate. Nat Commun 8:540 70. Gomes AC, Roesti ES, El-Turabi A et al (2019) Type of RNA packed in VLPs impacts IgG class switching-implications for an influenza vaccine design. Vaccines (Basel) 7:47 71. Jekhmane S, de Haas R, da Silva P, Filho O et al (2017) Virus-like particles of mRNA with artificial minimal coat proteins: particle formation, stability, and transfection efficiency. Nucleic Acid Ther 27:159–167 72. Le DT, Muller KM (2021) In vitro assembly of virus-like particles and their applications. Life (Basel) 11:334 73. Hou X, Zaks T, Langer R et al (2021) Lipid nanoparticles for mRNA delivery. Nat Rev Mater 6(12):1078–1094 74. Rosenblum D, Gutkin A, Kedmi R et al (2020) CRISPR-Cas9 genome editing using targeted

185

lipid nanoparticles for cancer therapy. Sci Adv 6:eabc9450 75. Fornaguera C, Guerra-Rebollo M, Angel Lazaro M et al (2018) mRNA Delivery System for Targeting Antigen-Presenting Cells In Vivo. Adv Healthc Mater 7:e1800335 76. Fornaguera C, Guerra-Rebollo M, Lazaro MA et al (2019) In vivo retargeting of poly(beta aminoester) (OM-PBAE) nanoparticles is influenced by protein corona. Adv Healthc Mater 8:e1900849 77. Marino M, Luján E, Mocskos E et al (2021) OpenEP: an open-source simulator for electroporation-based tumor treatments. Sci Rep 11:1423 78. Lione L, Salvatori E, Petrazzuolo A et al (2021) Antitumor efficacy of a neoantigen cancer vaccine delivered by electroporation is influenced by microbiota composition. Onco Targets Ther 10:1898832 79. Kelley EL, Smith RH, Corcoran G et al (2021) Advances in subcutaneous injections: PRECISE II: a study of safety and subject preference for an innovative needle-free injection system. Drug Deliv 28:1915–1922 80. Dey A, Chozhavel Rajanathan TM, Chandra H et al (2021) Immunogenic potential of DNA vaccine candidate, ZyCoV-D against SARSCoV-2 in animal models. Vaccine 39:4108– 4116 81. Joshi A, Tandel N, Tyagi P et al (2019) RNA-loaded dendritic cells: more than a tour de force in cancer therapeutics. Immunotherapy 11:1129–1147 82. Lyerly HK (2003) Quantitating cellular immune responses to cancer vaccines. Semin Oncol 30:9–16 83. Lee KH, Wang E, Nielsen MB et al (1999) Increased vaccine-specific T cell frequency after peptide-based vaccination correlates with increased susceptibility to in vitro stimulation but does not lead to tumor regression. J Immunol 163:6292–6300 84. Hobeika AC, Clay TM, Mosca PJ et al (2001) Quantitating therapeutically relevant T-cell responses to cancer vaccines. Crit Rev Immunol 21:287–297 85. Macchia I, Urbani F, Proietti E (2013) Immune monitoring in cancer vaccine clinical trials: critical issues of functional flow cytometry-based assays. Biomed Res Int 2013:726239

Chapter 8 COVID-19 Pharmacotherapy: Drug Development, Repurposing of Drugs, and the Role of Pharmacogenomics Rebecca Bock, Mariana Babayeva, and Zvi G. Loewy Abstract The SARS-CoV-2 virus has been the subject of intense pharmacological research. Various pharmacotherapeutic approaches including antiviral and immunotherapy are being explored. A pandemic, however, cannot depend on the development of new drugs; the time required for conventional drug discovery and development is far too lengthy. As such, repurposing drugs is being used as a viable approach for identifying pharmacological agents for COVID-19 infections. Evaluation of repurposed drug candidates with pharmacogenomic analysis is being used to identify near-term pharmacological remedies for COVID-19. Key words Antiviral therapy, Immunotherapy, Repurposing, Pharmacogenomics

1

Introduction The cause of the COVID-19 pandemic is the highly infectious severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) belonging to the genus of Betacoronavirus which consists of six additional species of the virus [1]. These coronaviruses are spherical in structure, contain single-stranded positive-sense RNA, and are surrounded by protein “spikes” which appear like a crown, giving the virus its name “Corona” [2, 3]. SARS-CoV-2 consists of 29,903 nucleotides with 79–89% sequence similarity to SARSCoV and 50–60% sequence similarity to MERS-CoV [1, 4], two additional forms of coronavirus which have caused previous epidemics [2]. The genome of the virus includes 12 functional open-reading frames (ORFs) which code for both structural and nonstructural proteins which are crucial for virion synthesis and assembly [2]. Some of the structural proteins include the spike (S) protein, the membrane (M) protein, the envelope (E) protein, and the

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_8, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

187

188

Rebecca Bock et al.

nucleocapsid (N) protein, each of which join to form the outer layers of the virus [2]. Additionally, nonstructural proteins, such as 3-chymotrypsin-like protease (3CLpro), papain-like protease (PLpro), helicase (Hel), RNA-dependent RNA polymerase (RdRp), endoribonuclease, and multiple accessory proteins, are encoded for use during viral replication [2, 4]. The S proteins are transmembrane glycoproteins which can be found as trimers on the surface of the virion [2]. They can be characterized as a class I fusion protein and are comprised of two subunits, one of which (S1 subunit) allows the virion to attach to receptors on the surface of the host cell and one of which (S2 subunit) causes the membranes of the virion and host cell to fuse, resulting in entry of the viral RNA into the host cell [2]. The S1 subunit includes one signal peptide, an N-terminal domain (NTD), and a receptor-binding domain (RBD), providing the ability to bind to the host cell receptors as needed. The S2 subunit includes the C-terminal, fusion peptide (FP), heptad repeat (HR) 1 and 2, transmembrane domain (TM), and cytoplasmic domain (CP), providing the ability to fuse with the host cell [2]. These S proteins show around a 76% amino acid sequence homology to the S proteins of SARS-CoV, revealing similar viral characteristics, and thus similar activity, between the two viruses [4]. To be successful in receptor binding, the SARS-CoV-2 S protein must have a clear binding path to the angiotensin-converting enzyme 2 (ACE2) receptor of the host cell. Because alveolar epithelial cells on the surface of the lungs and many cells on extrapulmonary tissues are cells with ACE2 receptors, SARS-CoV-2 can easily invade these areas and use them as the starting point for further replication, distribution, and destruction [2]. Due to a mutation in the S protein of SARS-CoV-2, the virus has a greater receptor binding affinity and is, therefore, more transmittable than other coronaviruses [4]. The world has been struck by an unprecedented crisis with the emergence of the severe acute respiratory syndrome coronavirus 2 (SARS-Cov-2) and its rapid spread and escalation to the COVID19 pandemic. While the need is clear for new drugs to combat the deadly virus, research and development of new medications can take many years. Looking to existing drugs that can be repurposed to treat patients with COVID-19 will be critical to mounting a timely and effective pandemic response and to saving lives. Pharmacogenomics, an emerging field at the intersection of genetics and pharmacotherapy, may be a key player in enhancing the success of repurposing existing drugs to alleviate the urgent global situation.

COVID-19 Pharmacotherapy

2

189

COVID-19 Therapeutic Methods To address the disease, many therapeutic methods have been suggested as a means of treating infected patients and ultimately contributing to bringing this worldwide pandemic to a near end. Some strategies involve targeting the virus by either inhibiting viral replication and translation of viral RNA or preventing the virus from binding to host cell receptors, while others involve a focus on the host by either an improvement of the host’s immune response or preventing certain host cell enzymes from carrying out their proper function [1]. In each case, many factors must be considered to determine the most effective strategy, including various characteristics of both the virus and the drugs or other therapies which will be used, such as the composition, functionality, and mechanism of action [2]. By categorizing the numerous types of therapies, the effectiveness of each can be more easily evaluated, helping to find the most efficient way to treat COVID-19 infections [5].

2.1 Antiviral Drug Therapy

Antiviral therapy aims to inhibit viral processes which allow the virus to infect the host. Certain drugs do so by targeting functional regions of the virus which are necessary for initial infection by the virus, such as receptor-binding sites. Others inhibit crucial steps in the replication and translation processes, preventing further or proper synthesis of the virions. By causing changes in the steps which typically provide the virus with its functionality, the virus is unable to infect and replicate [1]. The first method of inhibiting binding, and therefore entry, of the virus into the host cell aims to prevent initial infection of the host by the virus. To do so, TMPRSS2 (serine protease) activity must be stopped, as it usually cleaves or primes the S protein so that it can bind to the ACE2 receptor. By preventing TMPRSS2 from performing its job, proper binding cannot occur, thereby stopping the virus from infecting its host. One way to do so is with the use of a nonselective TMPRSS2 inhibitor, such as camostat mesilate, or with the use of a selective TMPRSS2 inhibitor, such as bromhexine [2]. An alternative approach to inhibit infection by the virus is by preventing the fusion of the virus with the host cells and the release of viral RNA into the host cells. While the virus requires a specific endosomal pH to fuse with the host cell and for its RNA to enter the cell, some antivirals, including chloroquine, can increase the endosomal pH, thereby stopping the viral RNA from entering the host cell and creating additional copies of the virus [2]. In a case where the viral RNA has already entered the host cell, RdRp inhibitors, such as remdesivir, can be used to interfere with the proper function of the RNA polymerase in the cell [2]. By doing so, the RNA of the virus that has entered the host cell cannot

190

Rebecca Bock et al.

be reverse transcribed properly, or at all, and the virus will not be replicated. Additionally, by inhibiting 3CLpro which is the main protease necessary for replication of the virus, important polyproteins (PP1A and PP1AB) cannot be cleaved to form two independent protein sections which play a role in replication. Therefore, the replication process of the virus would not occur by its proper mechanism which would help limit the number of overall virions formed and the infectivity of the virus. 2.2

Immunotherapy

Unlike antiviral therapy, immunotherapy attempts to strengthen the host’s humoral immunity to prevent infection more passively, rather than directly targeting the virus itself. However, the two approaches are similar in that, ultimately, the results of the immunotherapy provide antiviral effects. In the immunotherapy process, the antibodies which are part of the host’s humoral immune system bind to the antigens of the virus, thus preventing further infection by the virus [4].

2.2.1 Convalescent Plasma

In many cases of infectious diseases, including COVID-19, antibodies formed in individuals who were infected with a pathogen and have recovered are used to fight off infection in other patients [6]. The concept of convalescent plasma suggests providing suffering COVID-19 patients with the plasma, containing IgG antibodies, from previously infected COVID-19 patients. In doing so, the antibodies will recognize the virus in the new host and will help fight off the infection.

2.2.2 Monoclonal Antibodies

Monoclonal antibodies (mAbs) are the homogeneous products of a single B-cell clone which binds to a specific epitope, i.e., part of the viral antigen [4]. They are effective at stopping viral entry into the host cells by binding to the virus’s receptor-binding domain on the S protein, thereby preventing attachment of the virus to the host cell receptors. In addition, mAbs can suppress inflammation caused by increased levels of pro-inflammatory factors as a result of infection by the virus which helps limit the severity of patients’ symptoms [7].

2.2.3

Current and potential vaccines for COVID-19 come in many forms, including live attenuated vaccines, inactivated virus vaccines, recombinant protein vaccines, adenovirus vector vaccines, and nucleic acid (mRNA and DNA) vaccines [6]. They each use different methods, such as encoding the viral S protein with viral mRNA or gene fragments, or distributing portions of the S protein, to produce neutralizing antibodies and trigger a strong immune response for fighting the viral infection [7]. Table 1 lists the various therapeutic agents that have either been approved or are in development for use by COVID-19 patients. These drugs consist of a variety of types, including antivirals, anti-

Vaccines

COVID-19 Pharmacotherapy

191

Table 1 Therapeutic agents for COVID-19 Stage of development Source reference

Drug

Type

Mechanism of action

Remdesivir

Antiviral

Inhibits RdRp and viral Phase replication 3 clinical trial

Clin Exp Med (2021) 21: 167–179

Lopinavir/ritonavir

Antiviral

Inhibits proteases 3CLpro or PLpro

Clin Exp Med (2021) 21: 167–179

Favipiravir

Antiviral

Inhibits RdRp and viral Phase 2/3 replication clinical trial

Ribavirin

Antiviral

Inhibits polymerases, interferes with RNA capping, causes random mutations, and enhances T cell response

Chemostat mesylate

Antiviral

TMPRSS2 inhibitor

Nafamostat mesylate

Antiviral

TMPRSS2 inhibitor – inhibits S-mediated membrane fusion

Ivermectin

Antiviral/ antiparasitic

Reduces viral RNA

Chloroquine/ Antiparasite hydroxychloroquine

Convalescent plasma

Phase 2 clinical trial

Drug repurposing – hypothesis, molecular aspects and therapeutic applications. IntechOpen (2020)

Phase 1/2 clinical trial

Int Emerg Med (2020)

FDA approved

Int J Mol Sci (2020) 21:5559

Inhibitis mitogenactivated protein kinases (MAPK); interferes with glycosylation of ACE2, viral assembly, and budding, proteolytic processing of the M protein, and changes pH; degrades S protein

Immunotherapy Neutralizes the virus by detecting epitopes of the virus (continued)

192

Rebecca Bock et al.

Table 1 (continued)

Drug

Type

Mechanism of action

Intravenous immunoglobin

Immunotherapy Interferes with B-cell antigen presentation, immunemodulation, and immune substitution

Corticosteroids

Corticosteroids

Stage of development Source reference

Activates ACE2 and suppresses cytokine storm, thereby playing a protective role in respiratory and digestive systems

Azithromycin

Induces expression of interferon and pro-inflammatory cytokines (IL-6 and IL-8)

Nitric oxide

Inhibits viral RNA and Phase 2 clinical protein synthesis, trial inhibits viral replication by inhibiting palmitoylation of S-protein which affects binding to ACE2 receptor

Brain Behav Immun (2020) 87:59–73

Baricitinib

Monoclonal antibodies

JAK antagonist

Sarilumab

Monoclonal antibodies

IL-6 receptor antagonist

Camrelizumab

Monoclonal antibodies

Blocks PD-1 signaling to rescue exhausted CD8+T cell; restores CD8+T cell activity during viral infections

Adalimumab

Monoclonal antibodies

Targets TNF-α

Tocilizumab

Monoclonal antibodies

FDA Expert Revi Clin IL-6 receptor approved Pharm (2020) antagonist – inhibits inducement of for 13(9):957–975; inflammatory storm emergency Fact sheet for use healthcare

Phase 2/3 clinical trial

Int J Mol Sci (2020) 21:5559

(continued)

COVID-19 Pharmacotherapy

193

Table 1 (continued)

Drug

Type

Mechanism of action

Stage of development Source reference providers: emergency use authorization for Actemra® (tocilizumab) (2021)

Sources: Ijms, Yang, Zhang, Springer, Dr. Loewy article

parasites, immunotherapies, corticosteroids, monoclonal antibodies, and more. Each drug has a different mechanism of action which allows it to prevent or fight viral infection by SARS-CoV-2. By targeting major enzymes or proteins in or on the host cells or the virions, these drugs are able to inhibit crucial processes which allow the virus to infect the host cell. While some of the drugs have been FDA approved and can be used on COVID-19 patients, others are still being tested in clinical trials to ensure the safe and effective use against SARS-CoV-2 infection.

3

Repurposing Drugs for COVID-19 With the escalation of COVID-19, the virus was rampant, spreading rapidly from region to region. The seriousness and severity of the disease did not go unnoticed, and it was soon categorized as a pandemic. Scientists knew that for there to be any hope for survival for many who would become infected, a quick and effective solution would be necessary. Therefore, they turned to the concept of repurposing already existent drugs, i.e., applying developed drugs for certain indications to treat other illnesses and diseases, thus providing a faster and less expensive route to treating the highly contagious and potentially life-threatening coronavirus [8]. With repurposed drugs, all the relevant information about the drugs themselves is known, including their mechanisms of action, their adverse effects, and what they have been used for previously. By understanding their level of effectiveness for other diseases that have similar characteristics and symptoms to COVID-19, scientists have been able to identify candidate repurposed drugs to treat the virus. Through this process, the scientific community can potentially help bring an end to the pandemic [8]. Generally, drug discovery is very time-consuming, costly, and a process which can last anywhere from 10 to 16 years [9]. However, repurposing drugs is a more efficient process which consists of

194

Rebecca Bock et al.

finding new uses for already existing drugs, whether those drugs are old, failed, being investigated, FDA approved, currently on the market, or being used for more recent diseases. Because the process involves the use of drugs which have already been studied, repurposing drugs can shorten drug discovery by between 5 and 7 years while also significantly decreasing the total cost of development and production. Additionally, the familiarity with the drugs being tested allows for a reduced failure risk, as any toxicity or general safety concerns have already been identified and addressed [9]. Originally, repurposed drugs were a serendipitous phenomenon which served great and unexpected benefits to the world of medicine [8]. With the increase in advanced technology and scientific development, repurposing drugs has become an involved process, allowing for accelerated discoveries [9]. Repurposing drugs consists of four stages, including compound identification, compound acquisition, development, and FDA post-market safety monitoring. Because information about the drugs being used for repurposing is already disclosed, including information on their previous clinical efficacy and level of safety, the process does not require initial years of drug development. This allows for direct entry of the drugs into preclinical testing. Therefore, repurposed drugs can be used to treat rapidly and re-emerging infectious diseases, such as COVID-19, and difficult to treat or neglected diseases. Repurposed drugs can work as either on-target or off-target drugs [8]. When on-target, the mechanism of action of the particular drug and, therefore, the biological target remain the same. However, the resulting therapeutic outcome can differ from a previous outcome, allowing the drug to be used for a different disease [9]. Contrastingly, off-target drugs do not have the same mechanism of action and, therefore, do not have the same biological target. Consequently, these drugs provide new therapeutic indications which were not previously discovered [9]. To develop these drugs and test for their efficacy, the drugs are first screened in silico and then are investigated in vitro, or in vivo, in desired biological targets. Once successful with regard to both stages, the drugs are further evaluated in clinical trials in the hopes of being a safe and effective drug for the disease that requires treating. Table 2 lists several drugs which have the potential to be repurposed for treatment of COVID-19. Each listed drug has been previously and/or is currently being used to treat other viruses and diseases, including influenza, malaria, rheumatoid arthritis, HIV, and more. If successful, these candidate drugs may have the ability to treat COVID-19 patients, ranging from those who have more mild symptoms to those who are critically ill and hospitalized. While all the listed drugs have the potential to be successful choices for repurposed drugs, some might be more

COVID-19 Pharmacotherapy

195

Table 2 Candidate repurposed drugs for COVID-19 Drug

Original use

Favipiravir

Influenza

Treatment/usage of drug

Empirical studies

Hydroxychloroquine Malaria, rheumatoid For treatment of patients in arthritis, systemic clinical trial and hospitalized lupus erythematosus patients when necessary

Increased efficacy when used with azithromycin

Chloroquine

Malaria, extraintestinal For treatment of patients in amebiasis clinical trial

Less effective than hydroxychloroquine

Ivermectin

HIV, dengue, influenza, RSV, rabies

Lopinavir/ritonavir

HIV/AIDS

Remdesivir

Influenza, ebola

Tocilizumab, IL-6 inhibitor

Rheumatoid arthritis

Colchicine

Gout, pericarditis, HIV, hepatitis, zika

Baricitinib and Ruxolitinib

Rheumatoid arthritis

Can be effective together with ribavirin For patients with severe disease

Sources: Dr. Loewy’s article, Springer

effective than others or be more effective when combined for use with another drug. By comparing the treatment status and effectiveness of each drug, scientists can determine which drugs are most advantageous for use against COVID-19.

4

Pharmacogenomics With the need for a fast and efficient system for repurposing drugs to help treat COVID-19, all possible limitations must be considered. One of these includes the relevance of the human genome in instructing the body’s interactions with the different therapeutic drugs that may be used to fight off the virus [10]. With knowledge regarding the way in which variations in expression of certain crucial human genes relate to the different outcomes resulting from use of various drugs, repurposing drugs can be used in combatting COVID-19 in a quicker, safer, and more successful approach. This concept, employing pharmacogenomics, can lead to a handful of emergency treatment methods, helping to suppress the virus before it has the chance to do further harm [10].

196

Rebecca Bock et al.

Pharmacogenomics is based on monogenic variants, i.e., genetic polymorphisms, and how these changes affect drug response and efficacy. Genetic variation may result from a single nucleotide polymorphism (SNP) which is the most common of changes. These changes in the genetic material can alter the way in which drugs are recognized, transported, and metabolized within the body which can potentially have harmful results for specific recipients based on the way in which their genetic material is expressed. Knowing how a patient’s body will react to the intake of specific drugs can help determine the favorable doses of such drugs for each person and can help prepare for any potential adverse effects [11]. As many existing drugs are being considered for potential treatment of COVID-19, it is important to determine whether or not, or to what extent, these drugs will be effective in fighting off the virus. Determining the genetic makeup of infected patients will help in designing the optimal personalized pharmacotherapy; each drug will have a different impact on different patients [12]. Some drugs will be more quickly recognized and metabolized by the host’s enzymes, and others might even go as far as to cause adverse effects [11]. Additionally, certain drugs might only need to be administered in small doses, while others may be more effective in the presence of another drug. In any case, the expression of the specific genome will help determine which drugs should be used for each patient in order to provide a quicker and safer treatment for the patient. Considering how genetic material relates to the interaction of a drug with the host’s body will help identify ways to eliminate any harmful and toxic effects and further suffering of the patient. While variations in the genome of the genes that encode enzymes which metabolize drugs play an extremely significant role in treating COVID-19, changes in the genome of enzymes which are directly related to infection by SARS-CoV-2 can determine how badly one will be infected with the virus [13]. Certain enzymes are more relevant than others with regard to infection by SARS-CoV-2, and, therefore, the genetic makeup of the genes that express those enzymes should be closely studied. For example, ACE2 which controls the magnitude of virus-host cell binding and, thus infection by the virus, is a very important enzyme in the process of viral entry into the host cell. A polymorphism in this enzyme can either help prevent infection or can result in the promotion of infection based on the change that was made in the amino acids which are expressed. The same is true of TMPRSS2 which primes the S protein on the surface of the virus, thereby permitting the virus to bind to the host cell’s ACE2 receptor and allow for entry of the viral RNA into the host cell. Variations in the genotype of the gene which expresses TMPRSS2 can limit the level of expression or effect of the enzyme’s ability to prime the S

COVID-19 Pharmacotherapy

197

protein, in effect limiting or preventing binding of the virus to the host cell and entry of viral RNA into the host cell [13]. Therefore, it is important to assess all factors that might impact the level of infection, including variations in enzymes that are important for drug uptake and also enzymes which directly impact viral infection. When determining the optimal drugs to use for COVID-19 patients, it is extremely important to consider the patients’ backgrounds. The world consists of many different populations, with people in each having varied genetic makeup from people of other populations. This is relevant because drastic differences in genetic makeup can result in varied responses to certain drugs [11]. Therefore, categorizing genetic material based on different populations can help scientists get a better sense of how and to what extent crucial enzymes will be expressed in specific populations so that certain drugs can be used specifically for patients of that population. By focusing on the patients’ backgrounds and determining their origins and respective population, specific drugs can be used to most effectively target infection by SARS-CoV-2 [10]. Because people of different origins would react differently to varying drugs due to variations in their genetic makeup, it is important to also recognize potential adverse effects which can help in choosing the most effective and safest drugs for each population [11, 12]. This can help limit the use by patients of ineffective drugs or drugs that have the potential to cause adverse effects for them specifically. Table 3 lists multiple drugs which are potential repurposed therapeutic agents for COVID-19, in addition to their pharmacogenomic indications. Knowing the genotypes of each drug allows for recognition of any polymorphisms within the genetic material which may result in different forms of expression of critical metabolic enzymes. Each type of alteration in the gene sequence of these important enzymes in the host cells can have a great impact on the effectiveness of the drug. While these polymorphisms can have a negative impact, including lowering enzyme activity or causing adverse effects upon metabolization of the drug, they can also result in greater drug concentrations. Associating specific genotypes with the outcome when using certain drugs allows for determining which drugs are most effective for each infected individual.

5

Design Considerations for a Pharmacogenomic Assay Pharmacogenomics, the intersection of pharmacology and genomics, entails analysis of a specific segment of genomic DNA that contains polymorphic region(s). Common polymorphisms or variants include single nucleotide polymorphisms (SNP), insertions or deletions, and copy number variants [14].

198

Rebecca Bock et al.

Table 3 Pharmacogenomic indications of COVID-19 candidate drugs Drug

Genotype

Phenotype

Hydroxychloroquine CYP2D6, CYP3A4, CYP3A5, CYP2C8, Polymorphism in CYP2C8*4, CYP1A1 substrate CYP2C8*2, CYP2C8*3 causes lower enzyme activity, poor or intermediate CYP2D6 metabolizers (CYP2D6*4, CYP2D6*10) result in high hydroxychloroquine concetrations, SNPs in G6PD cause lower enzyme activity result in greater risk of hemolysis Chloroquine

CYP2D6, CYP3A4, CYP2C8 CYP1A1 substrate

Polymorphism in CYP2C8*4, CYP2C8*2, CYP2C8*3 causes lower enzyme activity, SNPs in G6PD cause lower enzyme activity result in greater risk of hemolysis

Lopinavir/ritonaivr

CYP3A4 inhibitor and substrate, CYP2D6 substrate, CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19 inducer, Pgp substrate, UGT1A1 inducer

Polymorphisms in UGT1A1, UGT1A7, APOE, APOC3 causes gastrointestinal adverse effects, hepatotoxicity, pancreatitis, and cardiac conduction abnormalities

Ribavirin

CYP3A4 inhibitor and substrate, CYP2D6 substrate, CYP1A2, CYP2C9

Polymorphisms in ITPA, VDR, SLC28A2 result in toxicity and adverse effects, variants of VDR, SLC29A1, IFNL3, MICB-OASL result in increased response

Azithromycin

CYP2C9 inhibitor, biliar excretion

Two polymorphisms in ABCB1 result in lower peak drug concentrations

Remdesivir

CYP2D6, CYP3A4, CYP2C8, CYP27B Unknown inducer, OATP1B1

Favipiravir

CYP2C8 and aldehyde oxidase inhibitor Unknown

Dexamethasone

CYP3A4, CYP2C8 inducer

Variants of CYP3A7, CYP3A5, and CYP3A4 can affect corticosteroid response, variants of ATF5, MIR3683, CTNNB1, PNPLA3

Sources: Nature, Dovepress, Wiley

Numerous genotyping technologies are available to determine an individual’s genotype. A technology platform is typically chosen based on access to instrumentation, technical capabilities, throughput requirements, the nature of the polymorphism, the speed at which the genotyping must be performed, and the cost [15]. For clinical pharmacogenomic (PGx) applications, the ideal platform should reliably and accurately detect an individual molecular change (variants) occurring in each selected gene associated with drug response (Pharmacogene). In addition, the gene(s)-drug pair

COVID-19 Pharmacotherapy

199

needs to be curated at the allele level to properly confer proven clinical validity and utility. There are regulatory organizations like the Federal Drug and Food Administration (FDA) and professional organizations such as the Clinical Pharmacogenomics International Consortium (CPIC) that produce standards for the proper and safe implementation of pharmacogenomic testing in clinical practice. Any test that is used for patient pharmacotherapy decisions must be performed in a certified clinical laboratory. In the United States, the certification must be compliant with Clinical Laboratory Improvement Amendments (CLIA) to ensure that the clinical laboratory operates under standards for the overall quality management system to ensure best clinical laboratory practices. The most common methodological platforms used in the field of PGx are next-generation sequencing (NGS), polymerase chain reaction (PCR), and microarrays. Each approach offers a unique combination of scale, accuracy, throughput, and cost and has distinct technical and operational advantages and disadvantages [16]. References 1. Rohani N, Moughari FA, Eslahchi C et al (2021) Discovering potential candidates of RNAi-based therapy for COVID-19 using computational methods. PeerJ. https://doi. org/10.7717/peerj.10505 2. Depfenhart M, de Villiers D, Lemperle G et al (2020) Potential new treatment strategies for COVID-19: is there a role for bromhexine as add-on therapy? Intern Emerg Med 15: 801–812 3. Loewy ZG (2020) The journey to a COVID19 diagnostic test. J Precis Med 2020 4. Zhand S, Jazi MS, Mohammadi S et al (2020) COVID-19: the immune responses and Clinial therapy candidates. Int J Molec Sci. https:// doi.org/10.3390/ijms21155559 5. Gavriatopoulou M, Ntanasis-Stathopoulos I, Korompoki E et al (2021) Emerging treatment strategies for COVID-19 infection. Clin Exp Med 21:167–179 6. Yang X, Liu Y, Yang Q et al (2020) Medication therapy strategies for the coronavirus disease 2019 (COVID-19): recent progress and challenges. Expert Rev Clin Pharm 13(9):957–975 7. Zhang J, Xie B, Hashimoto K (2020) Current status of potential therapeutic candidates for the COVID-19 crisis. Brain Behav Imm 87: 59–73 8. Parvathaneni V, Gupta V (2020) Utilizing drug repurposing against COVID-19 – efficacy, limitations, and T challenges. Life Sci. https://doi.org/10.1016/j.lfs.2020.118275 9. Rudrapal M, Khairnar SJ, Jadhav AG (2020) Drug repurposing (DR): an emerging

approach in drug discovery. https://doi.org/ 10.5772/IntechOpen.93193 10. Badary OA (2021) Pharmacogenomics and COVID-19: clinical implications of human genome interactions with repurposed drugs. Pharm J 21:275–284 11. Cafiero C, Re A, Micera A et al (2020) Pharmacogenomics and pharmacogenetics: in Silico prediction of drug effects in treatments for novel coronavirus SARS-CoV2 disease. Pharm Pers Med 13:463–484 12. Babayeva M, Loewy Z (2020) Repurposing drugs for COVID-19: pharmacokinetics and pharmacogenomics of chloroquine and Hydroxychloroquine. Pharm Pers Med 13:531–542 13. Al-Eitan LN, Alahmad SZ (2021) Pharmacogenomics of genetic polymorphism within the genes responsible for SARS CoV-2 susceptibility and the drug-metabolising genes used in treatment. Rev Med Virol. https://doi.org/ 10.1002/rmv.2194 14. Liu J, Zhou Y, Liu S et al (2018) The co-existence of copy number variations (CNVs) and single nucleotide polymorphisms (SNPs) at a locus can result in distorted calculations of the significance in associating SNPs to disease. Hum Genet 137:553–567 15. Offit K (2011) Personalized medicine: new genomics, old lessons. Hum Genet 130:3–14 16. Zhang J, Chiodini R, Badr A et al (2011) The impact of next-generation sequencing on genomics. J Genet Genomics 38:95–109

Chapter 9 Pharmacogenomics Informs Cardiovascular Pharmacotherapy Mariana Babayeva, Brigitte Azzi, and Zvi G. Loewy Abstract Precision medicine exemplifies the emergence of personalized treatment options which may benefit specific patient populations based upon their genetic makeup. Application of pharmacogenomics requires an understanding of how genetic variations impact pharmacokinetic and pharmacodynamic properties. This particular approach in pharmacotherapy is helpful because it can assist in and improve clinical decisions. Application of pharmacogenomics to cardiovascular pharmacotherapy provides for the ability of the medical provider to gain critical knowledge on a patient’s response to various treatment options and risk of side effects. Key words Personalized medicine, Polymorphism, Genotype

1

Introduction Cardiovascular disease is an umbrella term for a host of various morbidities that affect the heart and blood vessels such as thromboembolic disorders, hyperlipidemia, arrhythmias, hypertension, and heart failure. Pharmacists, physicians, and the entire healthcare community work tirelessly to address and overcome challenges in the pharmacotherapy surrounding cardiovascular disease. These efforts include improving patient adherence to medication regimens, improving selection of more optimal initial treatments, and improving monitoring and avoidance of adverse events or drug reactions. Despite the undoubted progress made in these fields, there still exists barriers to achieving higher rates of success in treating those with cardiovascular disease that are central to the individual variability in response to drug therapy. The advent of personalized medicine has come with the potential for improving the efficacy and safety of existing medications used to treat cardiovascular disease. The etiology of cardiovascular

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_9, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

201

202

Mariana Babayeva et al.

disease can be complex, and variations of genes in any of the pathways involved in it can affect a patient’s prognosis and response to therapy. Knowledge of these personal genetic variations can help in precisely identifying the best treatment drug and dose for each individual, thereby reducing complications, hospitalizations, and cost. Knowledge of the pharmacogenomics of cardiovascular and stroke drugs is necessary for effective and safe dosing and to avoid treatment failure and severe complications. This review focuses on recent advances in personalized medicine for the treatment of cardiovascular disease and stroke. In this paper, mechanisms underlying the variability in therapeutic response to cardiovascular and stroke pharmaceuticals in terms of mediating beneficial and adverse effects are reviewed and analyzed. Knowledge of the pharmacogenomics of these treatments is necessary for effective and safe dosing and to avoid treatment failure and severe complications.

2

Hypertension Therapy Hypertension is the leading chronic condition for which medications are prescribed [1]. At the present time, 31% of adults worldwide and 46% of US adults are projected to have hypertension, 20% of which are asymptomatic. The disease is associated with greater than 8 million worldwide deaths annually [2]. Although hypertension is a strong modifiable risk factor for cardiovascular disease, many individuals do not follow a pharmacological treatment [3]. The high prevalence of asymptomatic and pharmacologically uncontrolled individuals, coupled with low adherence, is associated with a significant clinical challenge. Genetic studies have suggested that heritability accounts for 30–50% of interindividual hypertension variability. The low efficacy of some therapies may be related to the interindividual genetic variability [4]. The challenge of blood pressure control likely has many causes, including poor adherence to therapy, lack of success of clinicians in achieving blood pressure control in their patients, and poor response to hypertensive pharmacotherapy. Approaches for selecting optimal pharmacotherapy for individuals are needed. To that end, pharmacogenomics is being used to guide treatment options with other cardiovascular drugs and the promise holds for hypertensive drugs. Genetic association studies have demonstrated associations between blood pressure lowering and cardiovascular outcomes. NEDD4L has been studied with thiazide diuretics and ADRB1 has been investigated with β-blockers. These two genes provide guidance for the clinical efficacy of antihypertensive drugs in individuals.

Cardiovascular Pharmacogenomics

203

2.1 Thiazide Diuretics

NEDD4L NEDD4L encodes the NEDD4-2 protein which contributes to the control of the cell surface expression of the sodium channel (ENaC), thus influencing sodium transport [5]. The genetic polymorphism rs4149601G>A results in a splice site in NEDD4L. The G allele leads to greater sodium retention. The G allele is associated with hypertension [6]. Several studies have confirmed that patients with the rs4149601G>A polymorphism exhibit enhanced response to thiazides. The Nordic Diltiazem (NORDIL) study reported that blood pressure lowering by a thiazide/β-blocker combination was greater in carriers of the G allele [7]. The Pharmacogenomic Evaluation of Antihypertensive Responses (PEAR) study demonstrated improved blood pressure lowering in patients with the G allele in response to hydrochlorothiazide [8].

2.2

Beta-Blockers

ADRB1 ADRB1 encodes the β1-adrenergic receptor. ADRB1 mediates the physiological response to noradrenaline and adrenaline. ADRB1 has been associated with blood pressure and hypertension in several studies. In a study of approximately 90,000 participants, the allele Arg389 was associated with hypertension [9]. Genome-wide association studies with greater than 60,000 participants identified an association between ADRB1 and hypertension [10]. The Arg389 allele is believed to be associated with enhanced β-blocker efficacy [11]. It has been associated with improved clinical outcomes with β-blocker treatment in patients with heart failure, atrial fibrillation, and ventricular arrhythmias [12–14]. Recently, two large genome-wide meta-analyses for antihypertensive response to β-blocker treatment have been reported. Several randomized controlled trials for discovery and randomized controlled trials for replication were analyzed [15, 16]. Together, these studies suggest that variation in ADRB1 may influence blood pressure response and clinical outcomes in patients with hypertension and cardiovascular disease.

2.3 ReninAngiotensinAldosterone System (RAAS)

The renin-angiotensin-aldosterone system (RAAS) regulates blood pressure and fluid/electrolyte balance as well as systemic vascular resistance [17]. The blood pressure regulation begins with the release of the hepatic protein angiotensinogen. The angiotensinogen is cleaved to hormone angiotensin I by renin, which is secreted by the kidneys. Angiotensin I is then converted to angiotensin II by angiotensin-converting enzyme (ACE) [18]. Vasoconstrictor angiotensin II increases blood pressure via two mechanisms: constricting vascular smooth muscle by stimulating angiotensin II type 1 receptor (AT1R) and promoting the production of aldosterone [19]. Aldosterone raises blood pressure by stimulating renal reabsorption of sodium and water and, thus, increases blood volume.

204

Mariana Babayeva et al.

Another protein, ACE2 counteracts the effects of the RAAS system by catalyzing the breakdown of angiotensin II back into angiotensin I. Medications influencing the RAAS system are ACE inhibitors, ARBS, and renin inhibitors. 2.4 AngiotensinConverting Enzyme (ACE) Inhibitors

Angiotensin-converting enzyme (ACE) inhibitors are a family of drugs used to treat cardiovascular and renal diseases, including hypertension and heart failure. ACE inhibitors, encoded by ACE gene, prevent the activity of ACE enzyme, which converts angiotensin I to angiotensin II and increases the level of vasodilator bradykinin [18]. Commonly prescribed ACE inhibitors include benazepril, zofenopril, perindopril, trandolapril, captopril, enalapril, lisinopril, and ramipril. ACE There is a high interindividual variability in ACE levels. The most investigated ACE genetic variant is the insertion/deletion (I/D) in intron 16. This ACE polymorphism happens in 27% of the general population and explains half of the phenotypic variations of ACE expression [20, 21]. The II, ID, and DD genotype frequencies were estimated as 22.5%, 47%, and 30.5%, respectively [22]. While ethnicity is a major contributing factor to the polymorphism, the genetic diversity is particularly high in people of African descent [23–25]. The reported prevalence of the D allele is 39.1% in Asian, 56.2% in Caucasians, and 60.3% in African American populations [22]. However, associations between ACE variants and blood pressure have been inconsistent. Moreover, clinical studies have observed conflicting results regarding ACE polymorphism and reactions to the therapy. Some studies have demonstrated greater response to ACE inhibitors in carriers with the I allele, whereas others have shown better response with the D genotype [26–40]. Furthermore, several studies have demonstrated no association between the ACE I/D polymorphism and blood pressure response or vascular events with ACE inhibition therapy showing no effects of this polymorphism on blood pressure responses to ACE inhibitors or ARB [38, 41–45]. The associations of the ACE I/D polymorphism with hypertension and cardiovascular complications have been inconsistent as well [46–51]. However, the D allele of the ACE I/D polymorphism is correlated with increased risk of diabetic nephropathy [52–54]. Overall, there is currently no evidence to support a role of the ACE I/D polymorphism in predicting risk of cardiovascular events or blood pressure response to ACE inhibitors.

Cardiovascular Pharmacogenomics

205

AGT Angiotensinogen, encoded by AGT gene, is the precursor for the formation of angiotensin I, which is associated with the ACE inhibitor’s mechanism of action [55]. The most common polymorphism of AGT is Met235Thr. Studies have shown that individuals with 235Thr alleles have higher plasma levels of angiotensin compared to people with Met alleles [55]. However, the associations between Met235Thr polymorphism in AGT gene and the response to ACE inhibitors or cardiovascular events, such as stroke or myocardial infarction, remain unclear [39, 56–63]. NOS3 The positive effect of ACE inhibitors was also associated with the vasodilation produced by nitric oxide (NO) due to nitric oxide synthase (NOS3) activation [64, 65]. The NOS3 gene polymorphisms were linked to hypertension and other cardiovascular disorders [66–69]. The importance of NOS3 polymorphism for ACE inhibitors therapy has been validated in some trials. A study in hypertensive patients treated with enalapril demonstrated better response to the drug in carriers with the C allele of the NOS3 786T/C (rs2070744) [70]. Enalapril also produced greater response in the carriers with T allele of the NOS3, 665C/T (rs3918226), while the A allele of the NOS3 tagSNP rs3918188 and the CAG haplotype were associated with an inverse result [71]. Some other genes including bradykinin receptor B2 (BDKRB2), protein kinase C (PKC), and vascular endothelial growth factor (VEGF) have been shown to contribute to NOS3 activation by ACE inhibitors [65, 72, 73]. Therefore, the polymorphisms in these genes may affect the antihypertensive responses to ACE medications [64, 70, 74, 75]. However, this assumption requires further investigation. 2.5 Angiotensin Receptor Blockers (ARB)

The angiotensin receptor blockers (ARB), also called angiotensin (AT1) receptor antagonists or sartans (losartan, valsartan, candesartan, irbesartan, telmisartan, and olmesartan), are medications inhibiting the effects of angiotensin II and lowering blood pressure. In addition, ARBs increase salt and water excretion, thus reducing the plasma volume and subsequently blood pressure [76]. ARBs are widely used drugs with indications for hypertension, chronic heart failure, secondary stroke prevention, and diabetic nephropathy and [77]. The ARBs have fewer side effects than ACE inhibitors [78]. CYP2C9 Almost all ARBs are metabolized by CYP2C9 enzyme, encoded by CYP2C9 gene [79]. However, up to date only losartan and irbesartan pharmacokinetics were shown to be dependent on polymorphisms of CYP2C9 gene. The oxidation of losartan to a more potent antihypertensive metabolite, E3174, was significantly reduced in CYP2C9*2 or CYP2C9*3 variants, especially in individuals hetero- or homozygous for the CYP2C9*3 allele or homozygous for the CYP2C9*2 allele [80, 81]. The finding suggests that

206

Mariana Babayeva et al.

the CYP2C9 genotype can contribute to losartan activation and carriers with CYP2C9*2 or *3 allele may have more noticeable antihypertensive effect than wild-type carriers. The similar trend was seen for irbesartan. The blood pressure reduction in patients with genotype CYP2C9*1 was 7.5% and 14.4% in patients with CYP2C9*2 allele [82]. ACE The effect of the ACE I/D polymorphism on ARBs treatment was also studied. The Swedish Irbesartan Left Ventricular Hypertrophy Investigation versus Atenolol (SILVHIA) trial showed irbesartan produced a larger decrease in blood pressure in I/I genotype compared to D/D or I/D carriers [83]. However, no conclusive results were found in other studies [55, 58]. CYP11B2 and REN Studies have investigated the effect of the CYP11B2 C344T (rs179998) and REN C5312T polymorphisms on ARB treatment. The CYP11B2 gene encoding aldosterone synthase, which catalyzes the final step of aldosterone production, was shown to modulate aldosterone level, hypertension susceptibility, and blood pressure responses to ARB [84–87]. However, the findings for CYP11B2 C344T polymorphism are conflicting, since in some studies, the C allele was related to blood pressure response, while in others, the blood pressure lowering effect was attributed to the T allele. Same trend was seen with REN C5312T polymorphisms [87–91]. Overall, the results of the studies were controversial and do not have a predictive value. AT1R ARBs work by blocking the activation of angiotensin II receptors (AT1R) [20, 39]. Polymorphism in angiotensin II receptor type 1 (AGTR1) gene encoding AT1R was shown to correlate with the response to ARBs [20]. The most common polymorphism is the AGTR1 A1166C. A study has demonstrated that ARB candesartan produced a greater decrease in systolic and diastolic blood pressure in homozygotes of the AGTR1 A1166 allele compared to C1166 allele carriers [92]. However, other investigations presented inconsistent results and currently there are no solid conclusions regarding the effect of this polymorphism on therapeutic response to ARB therapy [55, 58, 93, 94]. NOS3 The hypertensive effect of ARBs was also associated with NOS3 activation [95]. A study showed that ARBs increase formation of NO. Among all ARBs olmesartan produces greatest increase (30%) in production of NO. Moreover, olmesartan significantly increased NO formation in homozygous C allele of NOS3 SNP
786T/C (rs2070744) endothelial cells, suggesting that this ARB medication has more significant effects in the carriers with C allele of the NOS3 786T/C compared to T allele [78, 96].

Cardiovascular Pharmacogenomics

207

Other genetic polymorphisms studied in relation to ARB treatment include the GG genotype for the SNP rs10752271 in the CAMK1D gene (encoding calcium/calmodulin-dependent protein kinase 1D, involved in aldosterone synthesis) and the SNP rs3814995 in the NPHS1 gene (encoding nephrin, an important contributor to blood pressure regulation) associated with improved blood pressure responses to losartan [97, 98]. Polymorphisms in other genes, SNPs rs11020821 in FUT4 gene (encoding fucosyltransferase 4), rs3758785 in GPR83 gene (encoding G proteincoupled receptor 83), and rs11649420 in SCNN1G gene (encoding sodium channel, non-voltage-gated 1, gamma subunit) were associated with response to candesartan [99]. SNPs in SCNN1G were found to be strongly related to blood pressure. The GG genotype for the rs11649420 polymorphism was associated with a threefold greater blood pressure decrease to candesartan compared to combined AA+AG group. Moreover, the rs3758785 GG carriers had 16-fold greater response to candesartan compared to AA carriers [78, 99, 100]. There have been many studies in relation to genetic polymorphisms and ARB therapeutic responses; however, the majority have conflicting results and/or several limitations, thus affecting the credibility of the results. 2.6 Direct Renin Inhibitors

Aliskiren, the direct renin inhibitor (DRI), belongs to the new class of antihypertensive drugs. Aliskiren binds to the active site of renin and blocks angiotensinogen cleavage, thus preventing the formation of angiotensin I. Clinical studies have demonstrated at least equivalent or superior blood pressure lowering efficacy compared with existing drugs with a favorable side effect profile [101– 103]. Data of pharmacogenomic effect on antihypertensive response to aliskiren are very limited. ABCB1 and CYP3A4 Aliskiren is slightly metabolized (20%) by CYP3A4 and a substrate for Pgp transporter (encoded by ABCB1 gene) [104, 105]. Based on available studies, polymorphism in the ABCB1 gene does not affect pharmacokinetics and therapeutic effect of aliskiren [108]. Most probably, CYP3A4 polymorphism would not affect pharmacokinetics and efficacy of this medication too, as only small portion undergo metabolism. REN A study revealed that a common genetic variant of renin REN 5312C/T contributes to blood pressure variations. Carriers of the
5312T allele had higher systolic and diastolic blood pressure. Aliskiren produced lesser blood pressure reductions in
5312T allele carriers than in CC homozygotes carriers [106].

208

Mariana Babayeva et al.

OATP2B1 Current data point to the chance that organic anion transporting polypeptide 2B1 (OATP2B1), encoded by SLCO2B1 gene, is one of the modulators of aliskiren pharmacokinetics [107, 109]. However, results of a prospective pharmacokinetic study showed that the SLCO2B1 c.935G>A SNP has no clinically significant effect on the pharmacokinetics of aliskiren [110]. The results of various studies revealed that genetic polymorphisms may affect the blood pressure reduction by RAAS antihypertensive medications. However, only a small fraction of the genes probably contributes to the antihypertensive responses. 2.7

Hydralazine

Hydralazine, Apresoline, is a vasodilator used in the treatment of resistant hypertension. This medication can be used alone or in combination with isosorbide dinitrate (BiDil). Hydralazine is metabolized by an acetylation reaction mediated by N-acetyltransferase 2 (NAT2), encoded by NAT2 gene. The activity of NAT2 enzyme depends on polymorphisms of NAT2 gene. NAT2 Polymorphisms in the NAT2 gene generate rapid, intermediate, and slow acetylator phenotypes and can result in altered efficacy and/or toxicity of NAT2 substrates [111–113]. Intermediate acetylators are often undifferentiated from fast metabolizers and can be considered functionally comparable [114]. NAT2*4 allele is associated with rapid acetylation, while the most common NAT2*5, *6, *7, and *14 alleles are linked to reduced N-acetylation activities with lower clearance values compared to NAT24 [115–120]. Carriers with two copies of the *6 or *7 allele have more reduced acetylation rate than those with the *5 allele and are considered as ultra-slow metabolizers [121–123]. The allele frequency distribution of NAT2 in the general population is approximately 1:1 in slow acetylators versus intermediate or fast acetylators [124]. Slow metabolizer genotypes (NAT2*5, *6, *7, and *14) are common in Caucasians and African Americans. In these populations, most people carry at least one copy of a slow acetylator allele, and less than 10% of the groups are homozygous for the wild type (fast acetylator) [125]. In the general population the frequencies for NAT2*5 and NAT2*14 are 0.43 and 0.004, respectively. Individuals of African ancestry have a higher frequency of the NAT2*5 and *14 genotypes, 0.67 and 0.08, respectively [126, 127]. The occurrences of ultra-slow acetylators are 29.05–54.27% in Europeans, Africans, and Southeast Asians, while Japanese and East Asian populations showed lower frequencies, 4.75% and 11.11%, respectively [117].

Cardiovascular Pharmacogenomics

209

Variability in the enzymatic activity of N-acetyltransferase 2 is an important contributor to interindividual differences in hydralazine response and toxicity. A substantial correlation exists between acetylator phenotype and hydralazine concentrations. Same dose of hydralazine produced significantly higher AUCs and plasma concentrations and longer tmax in slow acetylators compared to fast acetylators [128, 129]. Other studies demonstrated the NAT2 genotype-adjusted hydralazine doses produced almost the same pharmacokinetic parameters in slow and fast metabolizers [130, 131]. Doses of 83 mg and 182 mg produced AUC of 1410 and 1446 ng*h/mL in slow and fast metabolizers, respectively [132]. Similar result was demonstrated in oncology population. Despite the over twofold difference in dose, the plasma hydralazine levels were similar between fast and slow acetylators (239.1 ng/ mL vs. 259.2 ng/mL) [133]. Acetylation status remains relevant in pregnancy. In pregnant women, dose-normalized AUC and Cmax of hydralazine were higher in slow acetylators than in fast acetylators [134]. The hydralazine plasma levels correlate with a reduction in arterial pressure. A significant reduction of diastolic blood pressure was greater in slow acetylators than in fast acetylators [135]. The dose required to reduce supine systolic blood pressure in slow acetylators should be less than in fast acetylators [130, 131]. A study suggested 140–225% increase in dose to fast or intermediate acetylators compared to slow acetylators to produce the same hydralazine exposure and therapeutic effect [128]. Most studies suggest a 50–100% higher dose for fast acetylators [124]. The slow acetylation phenotype demonstrates not only increased antihypertensive efficacy but also a higher potential risk of adverse events attributable to hydralazine [128, 135]. One of the adverse effects is drug-induced systemic lupus erythematosus (SLE) [135–137]. Among patients with drug-induced SLE treated with hydralazine, 96% were phenotyped as slow acetylators [138]. Dose reduction should be considered to reduce/avoid the development of hydralazine adverse events in NAT2 slow metabolizer phenotypes. At the same time, fast acetylators are at risk for inefficacy at lower doses of hydralazine and are likely to benefit from a higher starting dose [124]. The US Food and Drug Administration (FDA) recommends the use of acetylation status to predict clinical outcome and avoid toxicity for Apresoline (hydralazine hydrochloride) monotherapy and the combination therapy BiDil (isosorbide dinitrate and hydralazine hydrochloride). In 2020, the FDA issued an updated information that NAT2 enzyme activity is a predictor of hydralazine systemic concentration [139]. With appropriate guidance clinicians can implement a personalized approach to hydralazine dosing, enabling more efficient and safe treatment of resistant hypertension.

210

3

Mariana Babayeva et al.

Cholesterol Medications One of the primary risk factors identified for coronary heart disease (CHD) is elevated levels of low-density lipoprotein (LDL) cholesterol. High levels of LDL are prevalent in more than a quarter of all adult Americans [140]. Statins, 3-hydroxy-3-methyl-glutarylcoenzyme A reductase inhibitors, have been observed to lower LDL levels. As a consequence, millions of Americans use statins [141]. However, two issues have affected the clinical utility of statins: (1) myopathy-muscle toxicity and (2) low levels of medication adherence [142]. SLCO1B1 SLCO1B1 is currently the only clinically relevant pharmacogenomic test regarding statin efficacy and toxicity. The anion transporting polypeptide 1B1 (OATP1B1) encoded by SLCO1B1 is a liver-specific transporter expressed on hepatocytes. The protein is involved in hepatic uptake of statins. In a genome-wide association study it was demonstrated that a variant of SLCO1B1, 521C, which associates with decreased OATP1B1 transporter activity and decreased hepatic uptake resulted in an odds ratio for myopathy of 4.5 compared to wild-type patients, all of which took simvastatin 80 mG daily [143]. In variant patients on simvastatin 40 mG, the relative risk was 2.6, suggesting that the strength of SLCO1B1 myotoxicity association increases with dose. The evidence for an association with other statins is less clear. Clinical guidelines for simvastatin-SLCO1B1 have been developed [144]. For patients that encode the variant, it is recommended to start with an alternate statin or to use the lower simvastatin dose of 20 mG/day. The benefit of integrating SLCO1B1 testing into routine clinical care has yet to be determined. In an approach to determine the benefit of testing, in a randomized clinical trial, preemptive SLCO1B1 testing with statinnaı¨ve patients was reported to be noninferior to no testing in reducing LDL-C. Moreover, no physician prescribed simvastatin to a patient known to have decreased or poor SLCO1B1 transporter function genotype [145]. NAT2 The UK Biobank project was a population cohort of greater than 500,000 participants that represented a general population of the United Kingdom with no enrichment for specific disorders [146]. A broad array of phenotypic information and diverse biological specimens were collected for each of the Biobank participants. Biobank participants provided blood, urine, and saliva samples. They also answered questions on lifestyle and healthrelated factors and completed physical measures. One Biobank study entailed analysis of approximately 300 medications;

Cardiovascular Pharmacogenomics

211

associations between pharmacogenes and the use of medications resulted in the identification of medication-pharmacogene relationships as prognostic indicators of therapeutic outcome [147]. Haplotype variation in NAT2 and the use of statins were found to produce clinically useful decreases in LDL-C concentrations. Research on genes and molecular pathways and their role in modulating the efficacy of lipid-lowering drugs may result in defining more effective treatments with statins and consequently lead to improved adherence. Evaluation of the effects of genetic variation by pharmacogenomic testing may pave the way for a personalized medicine approach for statin pharmacotherapy.

4

Acute Coronary Syndrome (ACS) Medications The common mechanism of acute coronary syndromes (ACS) is the rupture of an atherosclerotic plaque followed by partial or complete thrombotic occlusion of an artery. Mechanism of action of most ACS medications, by use of blood thinners is inhibition of formation of blood clots, inhibition of expansion of already formed clots, and reduction the risk of embolization of blood clots to other organs. Anticoagulant and antiplatelet drugs are commonly used to treat ACS. Anticoagulants are used to treat blood clots such as deep vein thrombosis and pulmonary embolism and to prevent stroke in people who have atrial fibrillation, valvular heart disease, or artificial heart valves. Anticoagulant drugs that are routinely used in the treatment of ACS include warfarin and novel oral anticoagulant drugs (NOAC): dabigatran, rivaroxaban, apixaban, edoxaban, and betrixaban. Antiplatelet drugs are often prescribed to individuals who have had a cardiac event or who are at an increased risk for thromboembolism, ischemia, infarction, recurrent cardiovascular events, and death. Pharmacologic antiplatelet agents that are usually used in the treatment of ACS include P2Y12 inhibitors (clopidogrel, prasugrel, ticagrelor), glycoprotein IIb/IIIa inhibitors (abciximab, eptifibatide, tirofiban), and NSAID (aspirin). Some of the ASC medications have interindividual variability in their therapeutic response. Genetic polymorphisms may play a role in the differences seen in anti-platelet drug efficacy and safety.

5 5.1

Anticoagulants Warfarin

Warfarin (Coumadin) is one of the most used oral anticoagulants. It decreases blood clotting by blocking vitamin K epoxide reductase that activates vitamin K1. Without sufficient active vitamin K1,

212

Mariana Babayeva et al.

clotting factors II, VII, IX, and X have decreased clotting ability. The effect and safety of warfarin can be monitored by international normalized ratio (INR) [148]. Warfarin’s narrow therapeutic index (NTI) requires careful dosing consideration to achieve a target INR and to avoid bleeding. Genetic polymorphisms in some genes involved in the pharmacokinetics and/or pharmacodynamics may impact its safety and therapeutic effect of warfarin therapy. Genetic testing can reveal the presence of such variants and help to increase the efficacy of therapy and decrease the risks of adverse events. Warfarin consists of a racemic mixture of two active enantiomers—R- and S-forms. S-warfarin is up to 5 times more potent than the R-isomer in producing an anticoagulant response [149]. Both warfarin enantiomers undergo CYP-mediated metabolism by many different enzymes. Major enzyme for S-warfarin is CYP2C9 and for R-warfarin is CYP3A4 [150]. Warfarin activity, at least partially, depends on genetic factors. Polymorphisms in two genes (VKORC1 and CYP2C9) play an important role in response to warfarin. CYP2C9 CYP2C9 polymorphisms can explain 10% of the dose variation between patients [151]. In comparison to the normally functioning CYP2C9*1 allele, common loss-of-function variants CYP2C9*2 (rs1799853) and CYP2C9*3 (rs1057910) are associated with approximately 30–40% and 80–90% less metabolizing power, respectively [152]. The increased concentrations of plasma warfarin can put such patients at an increased risk of bleeding and require a significant decrease in the warfarin dose [152, 153]. Patients with CYP2C9 *1/*2, *1/*3, *2/*2, *2/*3, or *3/*3 have lower warfarin dose needs by 19% for *2 allele and 33% for *3 allele compared with patients with CYP2C9 *1/*1 [154–156]. The loss-of-function CYP2C9*2 and *3 alleles have high frequencies in Caucasians (up to 18%) and low rates in African Americans and most Asians, suggesting that these variations may be of little or no relevance in the later populations [157, 158]. While the CYP2C9*2 and *3 polymorphisms are less common in African descent, other loss-of-function CYP2C9 variants, *5, *6, *8, and *11, have greater implications in this population [152]. Most of the currently available FDA-approved CYP2C9 genetic tests only include detection parameters for the *2 and *3 alleles. The Clinical Pharmacogenetics Implementation Consortium (CPIC) does not advise pharmacogenetic testing for individuals of self-identified African ancestry unless CYP2C9 *5, *6, *8, and *11 polymorphisms are also tested [159].

Cardiovascular Pharmacogenomics

213

CYP2C CYP2C rs12777823 is associated with a clinically relevant effect on warfarin therapy [160]. The SNP rs12777823 in the CYP2C gene has demonstrated a decreased metabolizing capacity of warfarin that is independent of variations in CYP2C9 (CYP2C9*2, *3) [159, 160]. While the rs12777823 SNP frequently appears in other ethnic populations, the effect of this polymorphism is greater in African Americans [160]. African American heterozygous and homozygous carriers of the mutant rs12777823 A allele required a respective decrease in weekly warfarin dose by approximately 7 mg and 9 mg [159]. The association was only detected for African Americans, suggesting that it is not the underlying cause but likely inherited with other variant(s) on a haplotype that influences warfarin dose in this population [159]. Most clinical laboratories do not yet include genotyping for the presence of the rs12777823 SNP. VKORC1 VKORC1 polymorphisms explain up to 34% of the dose variation between patients [151, 161]. Variations in VKORC1 can change sensitivity to warfarin and result in either decreased or increased dosing, depending on the variant [162]. Two key combinations of alleles (haplotypes) explain 25% of variations: low-dose haplotype group A and a high-dose haplotype group B [163]. The -1639G>A allele (rs9923231) of the VKORC1 was found to decrease the expression of the VKORC1 and increase circulating vitamin K [159]. There is a need for reduction by 30% per A allele in warfarin maintenance doses compared to wild-type -1639G/G homozygotes allele [155]. The -1173C>T allele (rs9934438) of the VKORC1 was also associated with a 1.7 mg/day dose reduction of warfarin, though there was conflicting evidence regarding VKORC1 1173C>T allele use in predicting warfarin dosing [152, 159, 161]. African Americans have higher proportion of group B haplotypes and require highdose warfarin, while Asian Americans commonly require low-dose warfarin due to higher proportion of group A haplotypes [163, 164]. Caucasian patients with A haplotype group (14%) also require low warfarin doses, while Caucasians with B haplotype group (25%) required high warfarin dose. Such patients with A/B haplotype group (61%) required intermediate warfarin dose [165]. The VKORC1 polymorphism can improve warfarin dose prediction for racially diverse populations [166]. CYP4F2 Up to 4% of the variation in warfarin response could be attributed to CYP4F2 polymorphism [167]. Hepatic metabolizing enzyme CYP4F2 catalyzes vitamin K1 into hydroxy-vitamin K1, thus reducing the accumulation of vitamin K [152]. Decreased function CYP4F2*3 (1297G>A, V433M, or rs2108622) polymorphism results in decreased oxidation of vitamin K and increased

214

Mariana Babayeva et al.

blood levels of vitamin K and requires higher warfarin doses for more effective anticoagulation [152, 155]. Carriers of the decreased function A allele (*3) have been shown to benefit from doses increased by 8–13% compared to carriers with the wild-type G allele (*1) [168]. Furthermore, CYP2C9*2 was more prevalent in subjects with the CYP4F2*3 variant [168]. Moreover, patients carrying A allele of CYP4F2 1347G>A polymorphism exhibited a fivefold increased risk of PVT [169]. In addition, carriers for T (CT/TT) variant allele have decreased the CYP4F2 enzyme activity and therefore require an increase in the warfarin dose compared to C (CC) variant [170, 172]. There was a 4–12% increase in the warfarin dose per T allele [173]. However, the differences were not found to be statistically significant [174]. Genotyping has shown the CYP4F2*3 variant was up to 38% in whites and Asians and only up to 8% among black ethnic group [168, 171, 175]. The relatively high percentage of patients of European and Asian descent may account for the utility of CYP4F2 in this population [167]. But the data do not support an impact of CYP4F2 polymorphism on warfarin dosing in African ancestry and so no recommendation is made for the use of CYP4F2 genotype data in blacks [159]. The Clinical Pharmacogenetics Implementation Consortium (CPIC) updated guideline for pharmacogenetics warfarin dosing. Based on the guidelines, in individuals with genotypes associated with CYP2C9 poor metabolism (e.g., CYP2C9 *2/*3, *3/*3) or both increased sensitivity (VKORC1-1639 A/A) and CYP2C9 poor metabolism, an alternative oral anticoagulant might be considered. If CYP2C9*5, *6, *8, or*11 variant alleles are detected, dose should be decreased by 15–30% per variant allele, or an alternative agent should be considered. Larger dose reductions might be needed in patients homozygous for variant alleles (20–40%, e.g., CYP2C9*2/*5/*6). If the CYP4F2*3 allele is also detected, the dose should be increased by 5–10% [159]. Despite the promise of pharmacogenomic testing in warfarin dosing, its use in clinical practice is controversial. The European Pharmacogenetics of Anticoagulant Therapy (EU-PACT) group found that patients treated by genotype-guided warfarin therapy resulted in higher rates of treatment time within the therapeutic range (67.4% vs. 60.3%), less time to reach target INR (21 days vs. 29 days), and significantly fewer cases of overanticoagulation (INR > 4) [176]. On the other hand, the Clarification of Optimal Anticoagulation through Genetics (COAG) trial conversely found that there was no significant difference between genotype-guided therapy and clinically guided therapy in terms of treatment time within the therapeutic range (45.2% vs. 45.4%, respectively) or outcomes such as over-anticoagulation and major bleeding or under-coagulation (INR < 2) and thromboembolism

Cardiovascular Pharmacogenomics

215

[177]. Moreover, an investigation of nine randomized clinical trials demonstrated that genotype-guided dosing strategy did not confer benefit in terms of time within therapeutic range, excessive anticoagulation (INR greater than 4), or a reduction in either major bleeding or thromboembolic events [178]. 5.2

Dabigatran

Dabigatran etexilate, Pradaxa, is a novel oral anticoagulant (NOAC), or direct oral anticoagulant (DOAC), that is commonly used in patients with nonvalvular atrial fibrillation. People with atrial fibrillation have an increased risk for forming blood clots and, therefore, risk for stroke and systemic embolism. The formation of blood clots begins when serine protease thrombin (factor IIa) catalyzes the conversion of the soluble protein, fibrinogen, to the insoluble protein, fibrin. Fibrin forms a fibrous mesh that complexes with platelets to form blood clots. Dabigatran prevents the formation of these blood clots by acting as a competitive, direct thrombin (factor IIa) inhibitor, thereby halting the progression of the coagulation cascade [179]. A major advantage of dabigatran over warfarin is that dabigatran can be administered as a fixed dose without the need for monitoring coagulation. However, dabigatran therapy still poses risk for adverse effects associated with over-anticoagulation such as bleeding. Plasma concentrations of dabigatran are highly variable between individuals with systemic variation as high as 30% [180]. Studies have been able to partly attribute these disparities to genetic polymorphisms. Pharmacogenetic variations in CES1 enzyme and P-glycoprotein may play an important role in the effectiveness and safety of dabigatran anticoagulant therapy [181]. CES1 Dabigatran etexilate is a prodrug that requires bioactivation into active metabolite, dabigatran. Two enzymes (CES1 and CES2) hydrolyze dabigatran etexilate into intermediate metabolites M1 and M2 and then to dabigatran. The rates of dabigatran formation were much more correlated with CES1 activity relative to CES2 activity, indicating that CES1 is the primary enzyme activating the prodrug with approximately 90% of the total activity [182, 183]. Polymorphisms of the CES1 gene have been associated with interindividual variability in dabigatran plasma concentrations, efficacy, and safety. An in vitro study conducted on human livers revealed that heterozygote carriers of the loss-of-function CES1 SNP G143E (rs71647871) affect the activity of CES1 and thus the activation of dabigatran [184]. The study found that the M2 activation rate was 41% less on average in G143E carriers. Additionally, gender played a role in M2 activation and female livers had a higher rate of the activation [184]. This finding confirmed previous data that demonstrated greater expression of CES1 in female livers and higher plasma concentrations of dabigatran in females [183, 185].

216

Mariana Babayeva et al.

Both CES1 single-nucleotide polymorphisms, rs2244613 and rs8192935, were associated with significantly lower trough concentrations of the active drug [181]. The SNP rs2244613 was also linked to a lower risk of bleeding [160, 186]. A study found that at least one minor C allele of the rs2244613 was present in 32.8% in the participants and corresponded to a decreased risk of bleeding during dabigatran therapy [9]. Each minor C allele of the rs2244613 was correlated with a 15% decrease in trough concentration of dabigatran compared to the wild-type AA genotype [181, 186]. However, the minor allele A on the CES1 SNP rs2244613 was associated with increased trough levels and higher risk for minor bleeding in patients treated with dabigatran [187]. Plasma trough concentrations of dabigatran have been shown to have a greater correlation to clinical outcomes including adverse events such as thrombosis and bleeding than peak concentrations [188]. Another SNP on CES1, rs8192935 have greater impact on both peak and trough concentrations of dabigatran [181, 186, 189]. Each minor T allele decreased peak concentrations by 12% compared to the wild-type CC genotype [181, 186]. Trough concentrations of dabigatran adjusted for variations in creatinine clearance and sex were significantly lower for the minor T allele compared to the wild-type C allele, with mean concentrations at 86.3 ng/dl for CC homozygotes, 62.1 ng/dl for heterozygotes, and 53.5 ng/dl for TT homozygotes [181]. However, this SNP was found to have no association with bleeding or ischemic events [186]. The genotype frequencies of CES1 rs2244613 and CES1 rs8192935 were significantly different between Asians and Caucasians. The genotypes of healthy Chinese subjects for CES1 rs2244613 allele C accounted for 61.1% and CES1 rs8192935 allele G accounted for 24%. While CES1 rs2244613 allele C accounted for 15.3–28.3%, CES1 rs8192935 allele G accounted for 67–68.5% among Caucasians [189]. The data demonstrate that polymorphisms in the CES1 gene have a significant impact on interindividual variability and can contribute to the safety of dabigatran in patients [190]. A personalized dabigatran treatment approach based on patient-specific CES1 genotypes and sex may have the potential to improve the efficacy and safety of dabigatran pharmacotherapy. ABCB1 Dabigatran transport is mediated by P-glycoprotein (P-gp) transporter encoded by polymorphic ABCB1 gene. The effect of ABCB1 gene on dabigatran therapy is controversial. Based on a study, carriers of the ABCB1 rs1045642 CC and rs2032582 GG genotypes had lower maximum plasma concentration and area under the curve (AUC) than TT and the A/T allele carriers, respectively [191]. The allele A of rs1045642 was

Cardiovascular Pharmacogenomics

217

significantly associated with the increased incidence of bleeding events. Conversely, other studies reported that no significant association was found between the ABCB1 polymorphisms rs4148738 and rs1045642 and dabigatran PK/PD [187, 190]. The genotype frequencies of ABCB1 rs1045642 and rs4148738 were similar in Asian and Caucasian populations [189]. ABCB1 rs1045642 allele A accounted for 43.8% and ABCB1 rs414738 allele C accounted for 43.8 in Asian population. Among Caucasians ABCB1 rs1045642 allele A accounted for 50.8% and ABCB1 rs414738 allele C accounted for 52.7–54.2% [189, 190]. 5.3

Rivaroxaban

Rivaroxaban is mainly metabolized by cytochrome P4503A4, 3A5, and 2J2 and is transported by P-glycoprotein, encoded by ABCB1 gene, and the breast cancer resistance protein (BCRP), encoded by the ABCG2 gene [192]. Both transporters are responsible for active renal secretion of rivaroxaban [193]. The peak and trough rivaroxaban concentrations depend on CYP3A4 activity [194]. CYP3A4 Some CYP3A4 polymorphisms (CYP3A4*22, rs35599367, and CYP3A4*17, rs4987161) were related to decreased metabolic activity and higher rivaroxaban plasma levels [195, 196]. However, a recent study showed no significant difference in rivaroxaban concentrations between the haplotypes CYP3A4-rs35599367/ABCB1-rs1045642 and CYP3A4rs35599367/ABCB1-rs4148738 compared to the wild haplotypes [197]. ABCB1 ABCB1 gene polymorphisms were investigated in only a few clinical studies. The studies demonstrated rivaroxaban-induced hemorrhage in patients with ABCB1 homozygous TT genotype for rs2032582 and rs1045642, due to higher peak concentrations and greater AUC produced by the variants [191, 193]. However, these two variants did not show a significant increase in rivaroxaban peak concentrations in healthy volunteers [198]. Pharmacogenomic screening of this haplotype mutation may be warranted particularly in patients with risk factors such as renal impairment [199].

5.4

Apixaban

Apixaban is metabolized mainly by CYP3A4 and CYP3A5, but also by CYP1A2, CYP2C8, CYP2C9, CYP2C19, and CYP2J2, and the sulfotransferases SULT1A1 and SULT1A2, mainly SULT1A1 [199]. Apixaban is transported by P-glycoprotein, encoded by ABCB1 gene and BCRP, encoded by the ABCG2 gene [192, 199].

218

Mariana Babayeva et al.

ABCB1 ABCB1 rs4148738 is significantly associated with the variability of peak levels of apixaban. Particularly, AA genotype of rs4148738 produced higher peak levels as compared to carriers with G allele [200]. Peak concentrations of apixaban in carriers of ABCB1 rs4148738 G variant were decreased by 26% in heterozygotes and 32% in homozygotes [200]. However, another study showed no significant association of ABCB1 SNPs (rs1045642, rs4148738, rs1128503, and rs2032582) with pharmacokinetics of apixaban in patients with atrial fibrillation and stroke [201, 202]. Variant ABCG2 421C>A with reduced transporter activity produced higher peak and trough blood levels of apixaban [203]. A significant increase in the concentration/dose ratio was reported for ABCG2 variant, 421A>A (rs2231142), as well [204]. CYP3A5 Among P450 enzymes CYP3A5*1/*3 or *3/*3 (rs776746) produced increased apixaban plasma levels [204]. However, CYP3A5*3 (rs776746) had no effect on the pharmacokinetics of the drug [202]. SULT1A1 One more enzyme involved in apixaban metabolism is the sulfotransferase SULT1A1. Loss-of-function SULT1A*3 (667A>G) allele can produce moderate changes in the metabolism and, therefore, in anticoagulant effect of apixaban, whereas SULT1A*2 (638G>A) does not have potential to alter the efficacy of apixaban [199, 205]. To date, no studies have investigated the impact of these variants on the efficacy or toxicity of apixaban. 5.5 Edoxaban and Betrixaban

Edoxaban is a substrate for P-glycoprotein encoded by ABCB1 gene and metabolic enzymes mainly CES1 encoded by CES1 gene. One of edoxaban active metabolites M4 is transported by organic-anion-transporting polypeptide 1B1 (OATP1B1) encoded by the LCO1B1 gene [192]. Variations in drug plasma concentrations during edoxaban therapy could be related to the CES1 and ABCB1 polymorphisms [206]. To date, only one study has investigated the rs1045642 (3435C>T) variants of ABCB1 and rs4149056 (521T>C) of SLCO1B1 [204]. These variants do not seem to impact the pharmacokinetics of edoxaban [207, 208]. To date, there is no data on the effect of genetic polymorphisms on betrixaban pharmacokinetics and pharmacodynamics. However, it would be expected that ABCB1 polymorphisms could impact plasma concentrations of betrixaban as betrixaban is a substrate for P-gp [208, 209]. The interindividual variability in NOACs is high and could result from genetic polymorphisms. However, NOACs are not subject to pharmacogenetic testing in clinical practice [210].

Cardiovascular Pharmacogenomics

6 6.1

219

Antiplatelets Clopidogrel

Clopidogrel, Plavix, is the medication that was most widely studied in antiplatelet pharmacogenomics. Clopidogrel is usually considered as the first-line antiplatelet agent because of its proven efficacy and cost-effectiveness [211, 212]. Clopidogrel belongs to the class of P2Y12 inhibitors and is used in the treatment of acute coronary syndrome, recent myocardial infarction or stroke, peripheral arterial disease, bypass surgery, and stent placement. Clopidogrel often is prescribed in combination with aspirin and remains the most widely prescribed anti-platelet drug [213]. Clopidogrel inhibits platelet aggregation through the binding of its active metabolite to the adenosine diphosphate (ADP) receptor encoded by P2RY12 gene [152, 214]. ABCB1 Clopidogrel transport is mediated by P-glycoprotein (P-gp) transporter encoded by polymorphic ABCB1 gene. Polymorphisms in the ABCB1 gene have been shown to affect clopidogrel plasma concentrations. Variants C1236T (rs1128503), G2677T (rs2032582), and C3435T (rs1045642) were detected in patients with the frequencies of 42.9, 44.8, and 52.8%, respectively [215]. These alleles were associated with increased transporter activity and decreased clopidogrel concentrations. Carriers of the T-T-T haplotype, a T allele at each of the SNPs, were seen to have significantly decreased clopidogrel levels and antiplatelet activity [216]. In addition, patients with two copies of the T-T-T haplotype had a greater risk of myocardial infarction, stroke, and death. Individuals with at least one C allele on one of the SNPs could achieve much higher clopidogrel plasma concentrations [155]. However, other studies have found no association between ABCB1 variations and clopidogrel response [217–220]. The clinical importance of these polymorphisms is still unclear and requires further investigation. Clopidogrel is a prodrug converting to its metabolites by CES1 and several P450 enzymes: CYP 2C19, CYP1A2, CYP3A4, CYP2B6, and CYP3A [214]. 85% of the absorbed dose is hydrolyzed to inactive metabolite by CES1. The clopidogrel active metabolite is also metabolized by CES1 to inactive hydrolytic metabolite [221]. The remaining 15% of the absorbed dose undergoes two-step oxidative process in the liver to form intermediate metabolite, 2-oxo-clopidogrel, and then active metabolite, clopidogrel thiol H4. CYP2C19 enzyme plays the main role in the production of active metabolite [221].

220

Mariana Babayeva et al.

CYP2C19 CYP2C19 genetic polymorphisms may contribute to variability in clopidogrel pharmacokinetics and result in differential effects on platelet aggregation. CYP2C19*1, CYP2C19*2, CYP2C19*3, CYP2C*4, CYP2C*5, CYP2C19*10, and CYP2C19*17 are different alleles of CYP2C19 gene. CYP2C19*1 is the normal function (wild-type) allele, and CYP2C19*2 (c.681G>A, rs4244285), CYP2C19*3 (c.636G>A, rs498693), CYP2C*4 (rs28399504), CYP2C*5 (rs72552267), and CYP2C19*10 (rs6413438) are loss-of-function alleles, while CYP2C19*17 (rs1248560) is a gain-of-function allele [222]. These variations produce diverse CYP2C19 phenotypes such as ultrarapid, extensive, intermediate, and poor metabolizers. A loss-of-function genetic polymorphism, particularly CYP2C19*2, is responsible for the reduced metabolism of clopidogrel and lower plasma concentrations of the active metabolite in healthy volunteers and in patients with ACS [223–225, 228]. Therefore, loss-of-function CYP2C19 polymorphism is associated with decreased antiplatelet activity and increased risk of adverse cardiovascular events. Poor metabolizers with two *2 alleles and intermediate metabolizers with only one *2 allele are at an increased risk of adverse cardiovascular outcomes (myocardial infarction, stroke, and death) [217–219, 224, 225, 229]. This risk of these adverse effects for poor metabolizers is increased by 1.8-fold, while the risk for intermediate metabolizers is increased by 1.5-fold [230]. The risk for stent thrombosis also increases for intermediate and poor metabolizers by 2.6- and 4-fold, respectively [230]. The loss-of-function alleles CYP2C19*3 have significantly reduced active clopidogrel metabolite levels as compared to wildtype subjects or the overall population [228]. CYP2C19*3 was also associated with increased risk of adverse cardiovascular outcomes [219, 224, 231]. Other loss-of-function allele variations of CYP2C19*4 and *5 seem to contribute to clopidogrel variations in safety and efficacy too, but their extent is not clearly defined [232]. The presence of CYP2C19*10 has significant clinical implications. The CYP2C19*10 variation affects bioactivation of clopidogrel and leads to decreased activity of the metabolizing enzyme by 75% compared to the wild-type CYP2C19*1 allele [233]. Moreover, the *10 allele has been found to interfere with certain CYP2C19 genotyping assays—particularly the CYP2C19*2 TaqMan assay—leading to misidentifying CYP2C19*10/*2 as CYP2C19*2/*2 [234]. This is especially important since the *10 retains some metabolizing capacity, while the *2 variant retains none.

Cardiovascular Pharmacogenomics

221

One gain-of-function allele, CYP2C19 *17 has been associated with increased antiplatelet activity and potential increased risk of bleeding [152, 234–236]. This variation was associated with increased production of the active metabolite and enhanced inhibition of platelet aggregation. However, the precise role of CYP2C19*17 on clopidogrel safety and efficacy is still unclear [152]. Some clinical studies have reported lower cardiovascular event rates and an increased risk of bleeding in patients with the CYP2C19*17 allele [152, 220, 236, 237], but other studies have not replicated the results [238, 239]. Some data suggest that the combination of CYP2C19 and ABCB1 polymorphisms may provide complementary information regarding clopidogrel clinical outcomes [226, 240]. Polymorphism of CYP2C19 depends on race. The CYP2C19*2 allele frequency is 36.8% in Indians, 28.4% in Asians, 16% in African Americans, and 13.3% in Caucasians [241]. The distribution of loss-of-function alleles (CYP2C19*2 and *3) shows greater inter- and intra-ethnic variations among Indian populations (37% and 1.9%) compared to other major ethnics such as Asians (28.4% and 10.1%) and Caucasians (13.3% and 0.2%) [241]. The *10 variant is less common with frequencies of 0.8%, 0.25%, and 0% in African Americans, Hispanics, and Caucasians, respectively [233]. The CYP2C19*17 allele is common in Caucasians (18%) and African Americans (18%), but not in Asians (4%) [242]. PON1 Another enzyme, esterase paraoxonase-1 (PON1) was identified as an important enzyme involved in the second step of the clopidogrel metabolism. The genetic variations in PON1 may affect the rate of clopidogrel metabolism [243]. Homozygous carriers of PON1-QQ192 with coronary artery disease and stent implantation displayed lower plasma concentrations of clopidogrel’s active metabolite, decreased antiplatelet activity, and increased risk for stent thrombosis compared to homozygous noncarriers (RR192) [243, 244]. However, later studies were unable to replicate these findings [60]. The Food and Drug Administration (FDA) included a black boxed warning for clopidogrel use for poor metabolizers [245]. The poor metabolizers, homozygotes for the *2 allele, have most clinically significant variations in clopidogrel safety and efficacy. However, genetic testing prior to clopidogrel use was not required. Various studies have released conflicting data on the significance of the CYP2C19 genotype for clopidogrel therapy. Available data indicate that CYP2C19 polymorphisms account for only a small portion of variability in clopidogrel response (12%); and the predictive value of CYP2C19 genotyping for adverse

222

Mariana Babayeva et al.

clinical events is low too [246]. This confusion has sparked debate on the clinical usefulness of pharmacogenetic testing in patients receiving clopidogrel treatment [228]. The ACC/AHA/SCAI PCI guidelines have recommended CYP2C19 genetic testing only for patients at high risk for poor clinical outcomes [247]. The Clinical Pharmacogenetics Implementation Consortium (CPIC) suggests that CYP2C19 pharmacogenomic testing should be “indication-specific” and is most clinically significant for poor metabolizers [222]. The CPIC rules recommend PGx testing for all patients undergoing percutaneous coronary intervention (PCI) or for patients at moderate or high risk for poor outcomes (patients undergoing high-risk PCI procedures or patients with stent thrombosis or diabetes). The guidelines recommended standard clopidogrel doses for ultrarapid or extensive metabolizers and increased clopidogrel doses or alternative drug therapies for intermediate or poor metabolizers with the CYP2C19*2 allele [221]. CES1 CES1 is another enzyme involved in metabolism not only clopidogrel but also its active metabolite. Data backs assumption that loss-of-function CES1 G143E, 428G>A, rs71647871 carriers have significantly higher plasma concentrations of both clopidogrel and its active metabolite. A 50% higher active metabolite concentration was observed in G143E carriers compared with noncarriers [185]. In addition, the inhibition of adenosine diphosphateinduced platelet aggregation was 24% higher in G143E carriers (reduced to 71% from baseline) relative to noncarriers (reduced to 57% from baseline) [248–250]. Another clinical studies reported that the AUC0–1 ratio of clopidogrel inactive metabolite to clopidogrel was 53% less in G143E carriers and the average inhibition of P2Y12-mediated platelet aggregation in the carriers was 19% higher than in noncarriers [221, 251]. As a result, patients with CES1 dysfunction would have a higher concentration of clopidogrel active metabolite compared with normal CES1 metabolizers when taking the same dose. Clopidogrel dose adjustment may be necessary to prevent potential toxicity (i.e., bleeding) in patients with CES1 dysfunction [185]. 6.2 Prasugrel, Ticagrelor

Other P2Y12 inhibitors, prasugrel or ticagrelor, can be used in the treatment of ASC as alternative medications. Common genetic variations do not affect markedly prasugrel or ticagrelor pharmacokinetics or clinical outcomes. CYP2C19 The CYP2C19 enzyme is involved in the metabolism of prasugrel, but CYP2C19 genetic polymorphisms (CYP2C19 *2 and *17) do not significantly impact the pharmacokinetics or the clinical outcomes of the drug. Studies on the effect of CYP2C19

Cardiovascular Pharmacogenomics

223

genetic polymorphisms on the prasugrel demonstrated that prasugrel provides great platelet inhibition, regardless of the genotypes [252]. Carriers of CYP2C19*2 alleles who received prasugrel had better efficacy compared with those who received clopidogrel [253]. Prasugrel caused a better decrease in the rate of the CV mortality, stroke, or MI in reduced metabolizers compared to clopidogrel [252, 254–256]. However, a study in ACS/PCI patients with CYP2C19*17 (rs12248560) allele revealed a higher risk of bleeding compared to noncarrier patients on prasugrel [257]. In addition, no associations were found between other prasugrel metabolizing enzymes (CYP2C9, CYP2B6, CYP3A5, and CYP1A2) or efflux transporter MDR1 (ABCB1) and safety and efficacy of prasugrel [226, 227]. Unlike prasugrel and clopidogrel, ticagrelor does not require metabolic activation. Ticagrelor is not a substrate for CYP2C19 enzyme or P-gp transport. Therefore, CYP or active transport (ABCB1) polymorphisms do not affect safety and efficacy of this drug [258–260]. Ticagrelor produces a more stable antiplatelet effect than other P2Y12 inhibitors and most of its variable platelet reactions are independent of genetic factors [261, 262]. While the simple solution may be to prescribe non-CYP2C19dependent medications such as prasugrel or ticagrelor to all ACS patients, these agents have restrictions such as specific contraindications, limited approved indications, and bleeding risks [263].

7

Aspirin Aspirin is acetylsalicylic acid (ASA) that belongs to a class of non-steroidal anti-inflammatory drug (NSAID). It is widely used either alone or in combination with other antiplatelet medications for the treatment and prevention of heart attack and stroke, particularly in individuals with atherosclerosis. The mechanism action of aspirin is irreversible inhibition of cyclooxygenase COX-1 and COX-2 enzymes. These enzymes metabolize arachidonic acid into prostaglandin and the potent platelet activator and vasoconstrictor, thromboxane A2 (TXA2) [264]. The inhibition of TXA2 production leads to a lower expression of glycoprotein GP IIb/IIIa, causing an inhibition of platelet activation. Aspirin is very effective in the prevention of platelet aggregation and reducing the risk of cardiovascular events. However, serious vascular events may occur in patients on aspirin. Some individuals have an aspirin resistance, where aspirin is unable to effectively inhibit platelet aggregation due to its failure to inhibit the COX-related platelet activation pathway [265].

224

Mariana Babayeva et al.

Aspirin resistance was found in 15.8% of patients and associates with 3.85 and 6 times greater risk of developing recurrent cardiovascular events and vascular-related death, respectively [266, 267]. Available data suggest that certain genetic variations may have a significant role in aspirin resistance [268]. Many candidate genes have been studied for their influence over aspirin efficacy. GPIa GPIa is a unit of a receptor for collagen on platelet GPIa/ IIa. Studies have reported controversial data on the participation of GPIa in clinical outcome of aspirin. Some data have shown no effect of the GPIa (C807T) polymorphism on the response of aspirin [269]. However, other studies supported involvement of the gene variation in aspirin effect. GPIa gene polymorphism at 807C>T (rsl126643) locus was associated with aspirin resistance in Chinese females [270]. Moreover, a considerable fraction of Jordanian population was resistant to the antiplatelet effect of aspirin due to GPIba C-5T polymorphism [266]. PGIIIa PGIIIa is a subunit of the PGIIb/IIIa fibrinogen receptor, encoded by ITGB3 gene. Some studies estimated that GPIIIa may be responsible for reduced aspirin response [264, 265, 271]. The P1A1/A2 polymorphism of the ITGB3 gene, T1565—>C (rs5918), was associated with reduced efficacy of aspirin in both healthy individuals and patients with cardiovascular disease [264, 269]. Another PlA1/A1 genotype was linked to a twofold increased risk of high platelet [272]. Furthermore, patients with the PlA2/A2 genetic constitution had even higher platelet aggregation than patients with the PlA1/A1 or PlA1/A2 genotypes and showed a significantly increased risk of the composite outcome of death or myocardial infarction and stent thrombosis [273, 274]. However, some studies demonstrated conflicting results and showed no significant association between the PlA1/A2 polymorphism and aspirin resistance [269, 273, 275]. COX-1 and COX2 The influence of several genetic variants of COX-1 and COX-2 enzymes, encoded by PTGS1and PTGS2 genes, on the response to aspirin has been studied. Majority of the studies reported a nonsignificant association between COX-1 and COX-2 polymorphism and aspirin resistance or adverse effect [266, 269, 270, 273, 275–278]. However, a study in elderly Chinese patients with cardiovascular disease demonstrated that the COX-1 G-allele (A-1676G, rs1330344) caused almost twofold increase of the risk of aspirin resistance [279]. In another study COX-1 haplotype GCGCC was associated with significant increased platelet aggregation [278]. Research demonstrated a higher frequency of the PTGS2 (COX-2) G-765C (rs20417) allele

Cardiovascular Pharmacogenomics

225

carriers in aspirin-resistant patients compared to aspirin-sensitive patients [278]. But the same polymorphism (PTGS2 rs20417) was correlated with a reduced risk of myocardial infarction [281]. In addition, individual patients with the combination of COX-1 (rs3842787) and COX-2 (rs20417) had a significantly higher risk of aspirin resistance and were associated with lower reduction of platelet aggregation activity [282]. PEAR1 Limited studies were conducted on the effect of platelet endothelial aggregation receptor 1 (PEAR1) on aspirin resistance. A large study found that carriers of the polymorphic allele have higher PEAR1 protein expression and higher platelet activity with aspirin compared to noncarriers [153, 283]. PEAR1 rs12041331 polymorphism was associated with an increased risk of myocardial infarction in Caucasians [284]. LPA Apolipoprotein (a), LPA, gene was strongly associated with an increased level of lipoprotein and an increased risk of coronary disease [285]. A placebo-controlled clinical trial revealed a substantial link between polymorphism in this gene and aspirin use [153, 285]. The LPA, rs3798220, was found to reduce the risk of the composite outcome of MI, ischemic stroke, or death by twofold compared to noncarriers [285]. Polymorphism of other genes CYP2C19*2 and *17, P2Y1, P2Y12, GPVI, and TBXA2R was investigated with respect to aspirin resistance [269, 286–288]. Up to date there was not found significant association with aspirin resistance.

8

Digoxin Digoxin is a frequently prescribed cardiac glycoside used to treat heart failure, atrial fibrillation, and certain cardiac arrhythmias. Digoxin is a medication with well-known cardiotoxic effect [289]. Digoxin is a drug with a narrow therapeutic index. Small changes in serum concentrations can result in significant alterations in therapeutic effect and toxicity. Digoxin therapy is associated with large variability in pharmacokinetics, which are highly correlated with genetic variance [290]. Metabolism does not play a major role in elimination of digoxin. Only about 13% of a digoxin dose is metabolized [291]. Digoxin is mainly eliminated by renal excretion, where P-glycoprotein (MDR1) transporter, encoded by ABCB1 gene, plays an important role [292]. The impact of pharmacogenomics on digoxin pharmacokinetics was demonstrated in mdr1a (P-gp) knockout mice (
/
). Digoxin achieved a 68-fold higher level in the brain of knockout mice than in normal mice [293].

226

Mariana Babayeva et al.

ABCB1 The expression and function of the drug transporter P-glycoprotein are highly variable. Many single-nucleotide polymorphisms have been identified in the MDR-1 gene. The most common SNPs of the ABCB1 gene are C1236T, G2677T/A, and C3435T [294]. Digoxin pharmacokinetics is significantly different between individuals with the TTT-TTT haplotype and those with TGC-CGC [295]. Effects of the SNPs of the ABCB1 gene, C1236T, G2677T/A, and C3435T on digoxin pharmacokinetics are inconclusive. Although some trials associate the TTT haplotype with increased digoxin serum levels, other studies do not support the finding [296–300]. The most studied polymorphism occurs at nucleotide 3435 in exon 26. The C/T 3435 polymorphism showed a statistically significant correlation with P-gp expression and activity [301]. A study reported that subjects homozygous for 3435T had more than twofold lower expression of P-gp in the small intestine and 38% higher digoxin plasma concentrations compared to 3435C subjects [302]. Accordingly, another study demonstrated that homozygous TT subjects had 20% higher digoxin concentrations than CT and CC subjects [303]. These results were supported by other trials [295, 296, 294, 302–304]. However, some data back the opinion that C3435T variant does not affect the plasma levels of digoxin or even 3435T polymorphism can decrease concentrations of digoxin [306, 307]. Reported AUC0–4 h values were 4.11, 3.20, and 3.27 ng h/ml for the subjects with C/C and C/T or T/T. respectively [308]. Limitations of the studies were that renal function, which can affect digoxin excretion and plasma concentrations, was not considered. Additionally, the association of the C3435T genotype with digoxin pharmacokinetics may depend on ethnicity [298]. Two studies in Japanese subjects found that carriers with the 3435C genotype had higher drug levels [308, 309]. The frequency of the 3435C>T polymorphism has been shown to vary significantly between ethnic populations. SNP 3435C>T occurs at high frequencies (60–72%) in Asians and less frequencies (34–42%) in Caucasians [294]. Africans have significantly higher C allele frequency (>74%) and at least 50% carry both C alleles [294]. The high frequency of the C allele in the African group suggests overexpression of Pgp transporter, which may affect pharmacokinetics, efficacy, and safety of digoxin in individuals of African origin [310]. Moreover, there was found a relationship between three ABCB1 single-nucleotide polymorphisms (SNPs) (3435C>T, 1236C>T, and 2677G>T), digoxin concentration, and gender. Females had higher serum concentrations of digoxin most probably due to higher frequency of T allele [311]. This correlation could explain higher digoxin concentrations in females compared to males as well as higher mortality in women treated with digoxin [298, 312].

Cardiovascular Pharmacogenomics

227

OAT transporter genes Digoxin has also been described as a substrate of various organic transporters: OATP2 (SLCO1A2 gene), OATP4C1 transporter (SLCO4C1 gene), OSTα, and OSTβ (SLC51A and SLC51B genes) [313–319]. Although in vivo studies have shown that these transporters contribute to the disposition of digoxin, the role of the proteins and corresponding genes in the pharmacokinetics of digoxin in humans has not been clearly defined [320].

References 1. http://www.who.int/chp/knowledge/ publicaions/adherence_report/en 2 . h t t p s : // p r o f e s s i o n a l h e a r t . o r g / p r o fessionalScienceNews/UCM_503383_ Heart-Disease-and-Stroke-Statistics-2019Update.jsp 3. Torrellas C, Carril JC, Cacabelos R (2014) Benefits of pharmacogenetics in the management of hypertension. J Pharmacogenomics Pharmacoproteomics 5:126 4. Munroe PB, Barnes MR, Caufield MJ (2013) Advances in blood pressure genomics. Circ Res 1112:1365–1379 5. Rizzo F, Staub O (2015) NEDD4-2 and saltsensitive hypertension. Curr Opin Nephrol Hypertens 24:111–116 6. Dahlberg J, Sjogren M, Hedblad B et al (2014) Genetic variation in NEDD4L, an epithelial sodium channel regulator, is associated with cardiovascular disease and cardiovascular death. J Hypertens 32:294–299 7. Svensson-Farbom P (2011) A functional variant of the NEDD4L gene is associated with beneficial treatment response with β-blockers and diuretics in hypertensive patients. J Hypertens 29:388–395 8. McDonough CW (2013) Association of variants of NEDD4L with blood response and adverse cardiovascular outcomes in hypertensive patients. J Hypertens 31:698–704 9. Johnson AD (2011) Association of hypertension drug target genes with blood pressure and hypertension in 86,588 individuals. Hypertension 57:903–910 10. Ganesh SK (2013) Loci influencing blood pressure identified using a cardiovascular gene-centric array. Hum Mol Genet 22: 1663–1678 11. Wu D (2015) Associations between ADRB1 and CYP2D6 gene polymorphisms and the response to β-blocker therapy in hypertension. J Int Med Res 43:424–434

12. Liggett SB (2006) A polymorphism with a conserved β1-adrengeric receptor motif alters cardiac function and β-blocker response in human heart failure. Proc Natl Acad Sci U S A 103:11288–11293 13. Aleong RG (2013) Prevention of atrial fibrillation by bucindolol is dependent on the β1389 Arg/Gly adrenergic receptor polymorphism. JACC Heart Fail 1:338–344 14. Aleong RG (2013) Adrenergic receptor polymorphisms and prevention of ventricular arrhythmias with bucindolol in patients with chronic heart failure. Circ Arrythm Electrophysiol 6:137–143 15. Singh S, Warren HR, Hiltunen TP et al (2019) Genome-wide meta-analysis of blood pressure response to β1-blockers: results from ICAPS (International Consortium of Antihypertensive Pharmacogenomics Studies). J Am Heart Assoc 8:e013115 16. McDonough CW, Warren HR, Jack JR et al (2021) Adverse cardiovascular outcomes and antihypertensive treatment: a genome-wide interaction met-analysis in the international consortium for antihypertensive pharmacogenomic studies. Clin Pharmacol Therap 110: 723–732 17. Fountain JH, Lappin SL (2021) Physiology, renin angiotensin system. [Updated 2020 Jul 27]. In: StatPearls [Internet]. StatPearls Publishing, Treasure Island. Available from: https://www.ncbi.nlm.nih.gov/books/ NBK470410/ 18. Byrd JB, Ram CVS, Lerma EV (2019) Pharmacologic treatment of hypertension. In: Nephrology secrets. Elsevier, Netherlands, pp 477–482 19. van Thiel BS, van der Pluijm I, te Riet L et al (2015) The renin-angiotensin system and its involvement in vascular disease. Eur J Pharmacol 763:3–14 20. Momary K (2007) Cardiovascular pharmacogenomics. J Pharm Pract 20:265–276

228

Mariana Babayeva et al.

21. Rigat B, Hubert C, Alhenc-Gelas F et al (1990) An insertion/deletion polymorphism in the angiotensin I-converting enzyme gene accounting for half the variance of serum enzyme levels. J Clin Invest 86:1343–1346 22. Staessen J, Ginocchio G, Wang JG et al (1997) Genetic variability in the reninangiotensin system: prevalence of alleles and genotypes. J Cardiovasc Risk 4:401–422 23. Sayed-Tabatabaei FA, Oostra BA, Isaacs A et al (2006) ACE polymorphisms. Circ Res 98:1123–1133 24. Kramers C, Danilov SM, Deinum J et al (2001) Point mutation in the stalk of angiotensin-converting enzyme causes a dramatic increase in serum angiotensinconverting enzyme but no cardiovascular disease. Circulation 104:1236–1240 25. Rieder MJ, Taylor SL, Clark AG et al (1999) Sequence variation in the human angiotensin converting enzyme. Nat Genet 22:59–62 26. Scharplatz M, Puhan MA, Steurer J et al (2005) Does the angiotensin-converting enzyme (ACE) gene insertion/deletion polymorphism modify the response to ACE inhibitor therapy? – a systematic review. Trials 6:16 27. Anwar MS, Iskandar MZ, Parry HM et al (2015) The future of pharmacogenetics in the treatment of heart failure. Pharmacogenomics 16:1817–1827 28. McDonough CW, Warren HR, Jack JR et al (2021) Adverse cardiovascular outcomes and antihypertensive treatment: a genome-wide interaction meta-analysis in the International Consortium for Antihypertensive Pharmacogenomics Studies. Clin Pharmacol Ther 110: 723–732 29. Pinheiro DS, Santos RS, Jardim PCBV et al (2019) The combination of ACE I/D and ACE2 G8790A polymorphisms revels susceptibility to hypertension: a genetic association study in Brazilian patients. PLoS One 14: e0221248 30. Fløe A, Hoffmann HJ, Nissen PH et al (2014) Genotyping increases the yield of angiotensinconverting enzyme in sarcoidosis--a systematic review. Dan Med J 61:A4815 31. He Q, Fan C, Yu M et al (2013) Associations of ACE gene insertion/deletion polymorphism, ACE activity, and ACE mRNA expression with hypertension in a Chinese population. PLoS One 8:e75870 32. Krishnan R, Sekar D, Karunanithy S et al (2016) Association of angiotensin converting enzyme gene insertion/deletion polymorphism with essential hypertension in south Indian population. Genes Dis 3:159–163

33. Shanmuganathan R, Kumaresan R, Giri P (2015) Prevalence of angiotensin converting enzyme (ACE) gene insertion/deletion polymorphism in south Indian population with hypertension and chronic kidney disease. J Postgrad Med 61:230–234 34. Ohmichi N, Iwai N, Uchida Y et al (1997) Relationship between the response to the angiotensin converting enzyme inhibitor imidapril and the angiotensin converting enzyme genotype. Am J Hypertens 10:951–955 35. Kohno M, Yokokawa K, Minami M et al (1999) Association between angiotensinconverting enzyme gene polymorphisms and regression of left ventricular hypertrophy in patients treated with angiotensin-converting enzyme inhibitors. Am J Med 106:544–549 36. Stavroulakis GA, Makris TK, Krespi PG et al (2000) Predicting response to chronic antihypertensive treatment with fosinopril: the role of angiotensin-converting enzyme gene polymorphism. Cardiovasc Drugs Ther 14: 427–432 37. Li X, Du Y, Du Y et al (2003) Correlation of angiotensin-converting enzyme gene polymorphism with effect of antihypertensive therapy by angiotensin-converting enzyme inhibitor. J Cardiovasc Pharmacol Ther 8: 25–30 38. Dudley C, Keavney B, Casadei B et al (1996) Prediction of patient responses to antihypertensive drugs using genetic polymorphisms: investigation of renin-angiotensin system genes. J Hypertens 14:259–262 39. Schelleman HK, Witteman OH, Breteler JC et al (2007) Angiotensinogen M235T polymorphism and the risk of myocardial infarction and stroke among hypertensive patients on ACE-inhibitors or beta-blockers. Eur J Hum Genet 15:478–484 40. Weinberger J (2003) Randomised trial of a perindopril-based blood-pressure-lowering regimen among 6105 individuals with previous stroke or transient ischaemic attack. Curr Cardiol Rep 5:140 41. Harrap SB, Tzourio C, Cambien F et al (2003) The ACE gene I/D polymorphism is not associated with the blood pressure and cardiovascular benefits of ACE inhibition. Hypertension 42:297–303 42. Tsikouris JP, Peeters MJ (2007) Pharmacogenomics of renin angiotensin system inhibitors in coronary artery disease. Cardiovasc Drugs Ther 21:121–132 43. Sasaki M, Oki T, Iuchi A et al (1996) Relationship between the angiotensin converting enzyme gene polymorphism and the effects of

Cardiovascular Pharmacogenomics enalapril on left ventricular hypertrophy and impaired diastolic filling in essential hypertension: M-mode and pulsed doppler echocardiographic studies. J Hypertens 14: 1403–1408 44. Hingorani AD, Jia H, Stevens PA et al (1995) Renin-angiotensin system gene polymorphisms influence blood pressure and the response to angiotensin converting enzyme inhibition. J Hypertens 13:1602–1609 45. Konoshita T (2011) Genomic Disease Outcome Consortium Study I. Do genetic variants of the renin-angiotensin system predict blood pressure response to renin-angiotensin system-blocking drugs? A systematic review of pharmacogenomics in the renin-angiotensin system. Curr Hypertens Rep 13:356–361 46. Wu CK, Luo JL, Tsai CT et al (2010) Demonstrating the pharmacogenetic effects of angiotensin-converting enzyme inhibitors on long-term prognosis of diastolic heart failure. Pharmacogenomics J 10:46–53 47. Arnett DK, Davis BR, Ford CE et al (2005) Pharmacogenetic association of the angiotensin-converting enzyme insertion/ deletion polymorphism on blood pressure and cardiovascular risk in relation to antihypertensive treatment: the genetics of hypertension-associated treatment (GenHAT) study. Circulation 111:3374–3383 48. Hollenberg NK (2003) The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). Major outcomes in high-risk hypertensive patients randomized to angiotensin-converting enzyme inhibitor or calcium channel blocker vs diuretic. Curr Hypertens Rep 5:183–185 49. Zintzaras E, Raman G, Kitsios G et al (2008) Angiotensin-converting enzyme insertion/ deletion gene polymorphic variant as a marker of coronary artery disease: a meta-analysis. Arch Int Med 168:1077–1089 50. Tao HM, Shao B, Chen GZ (2009) Metaanalysis of the ACE gene polymorphism in cerebral infarction. Can J Neurol Sci 36: 20–25 51. Medica I, Kastrin A, Peterlin B (2007) Genetic polymorphisms in vasoactive genes and preeclampsia: a meta-analysis. Eur J Obstet Gynecol Reprod Biol 131:115–126 52. Lin C, Yang HY, Wu CC et al (2014) Angiotensin-converting enzyme insertion/ deletion polymorphism contributes high risk for chronic kidney disease in Asian male with hypertension--a meta-regression analysis of 98 observational studies. PLoS One 9:e87604

229

53. Mooyaart AL, Valk EJ, van Es LA et al (2011) Genetic associations in diabetic nephropathy: a meta-analysis. Diabetologia 54:544–553 54. Ruggenenti P, Bettinaglio P, Pinares F et al (2008) Angiotensin converting enzyme insertion/deletion polymorphism and renoprotection in diabetic and nondiabetic nephropathies. Clin J Am Soc Nephrol 3: 1511–1525 55. Rosskopf D, Michel MC (2008) Pharmacogenomics of G protein-coupled receptor ligands in cardiovascular medicine. Pharmacol Rev 60:513–535 56. Schelleman H, Klungel OH, Witteman JC et al (2006) The influence of the alphaadducin G460W polymorphism and angiotensinogen M235T polymorphism on antihypertensive medication and blood pressure. Eur J Hum Genet 14:860–866 57. Iyngkaran P, Thomas MC, Johnson R et al (2016) Contextualizing genetics for regional heart failure care. Curr Cardiol Rev 12: 231–242 58. Barone C, Mousa SS, Mousa SA (2009) Pharmacogenomics in cardiovascular disorders: steps in approaching personalized medicine in cardiovascular medicine. Pharmgenomics Pers Med 2:59–67 59. Amici S, Paciaroni M, Agnelli G et al (2011) Gene-drug interaction in stroke. Stroke Res Treat 2011:212485 60. Dandona S (2014) Cardiovascular drugs and the genetic response. Methodist Debakey Cardiovasc J 10:13–17 61. Gatti RR, Santos PS, Sena AA et al (2013) The interaction of AGT and NOS3 gene polymorphisms with conventional risk factors increases predisposition to hypertension. J Renin-Angiotensin-Aldosterone Syst 14: 360–368 62. Konopka A, Szperl M, Piotrowski W et al (2012) Polymorphisms of angiotensin II type 1 receptor gene and those of angiotensinogen point at culprit artery in ST-segment elevation myocardial infarction. Gene 500: 148–150 63. Vivenza D, Feola M, Garrone O et al (2013) Role of the renin-angiotensin-aldosterone system and the glutathione S-transferase Mu, Pi and Theta gene polymorphisms in cardiotoxicity after anthracycline chemotherapy for breast carcinoma. Int J Biol Markers 28: e336–e347 64. Oliveira-Paula GH, Luizon MR, Lacchini R et al (2017) Gene-gene interactions among

230

Mariana Babayeva et al.

PRKCA, NOS3 and BDKRB2 polymorphisms affect the antihypertensive effects of enalapril. Basic Clin Pharmacol Toxicol 120: 284–291 65. Linz W, Wohlfart P, Scholkens BA et al (1999) Interactions among ACE, kinins and NO. Cardiovasc Res 43:549–561 66. Oliveira-Paula GH, Lacchini R, Tanus-Santos JE (2016) Endothelial nitric oxide synthase: from biochemistry and gene structure to clinical implications of NOS3 polymorphisms. Gene 575:584–599 67. Oliveira-Paula GH, Lacchini R, Tanus-Santos JE (2017) Clinical and pharmacogenetic impact of endothelial nitric oxide synthase polymorphisms on cardiovascular diseases. Nitric Oxide 63:39–51 68. Oliveira-Paula GH, Lacchini R, Pinheiro LC et al (2018) Endothelial nitric oxide synthase polymorphisms affect the changes in blood pressure and nitric oxide bioavailability induced by propofol. Nitric Oxide 75:77–84 69. Oliveira-Paula GH, Pinheiro LC, Ferreira GC et al (2018) Angiotensin converting enzyme inhibitors enhance the hypotensive effects of propofol by increasing nitric oxide production. Free Radic Biol Med 115:10–17 70. Silva PS, Fontana V, Luizon MR et al (2013) eNOS and BDKRB2 genotypes affect the antihypertensive responses to enalapril. Eur J Clin Pharmacol 69:167–177 71. Oliveira-Paula GH, Lacchini R, Luizon MR et al (2016) Endothelial nitric oxide synthase tagSNPs influence the effects of enalapril in essential hypertension. Nitric Oxide 55-56: 62–69 72. Li P, Kondo T, Numaguchi Y et al (2008) Role of bradykinin, nitric oxide, and angiotensin II type 2 receptor in imidapril-induced angiogenesis. Hypertension 51:252–258 73. Yang X, Zhu Q, Fong J et al (1996) Enalaprilat, an angiotensin-converting enzyme inhibitor, enhances functional preservation during long-term cardiac preservation. Possible involvement of bradykinin and PKC. J Mol Cell Cardiol 28:1445–1452 74. Oliveira-Paula GH, Lacchini R, Fontana V et al (2015) Polymorphisms in VEGFA gene affect the antihypertensive responses to enalapril. Eur J Clin Pharmacol 71:949–957 75. Oliveira-Paula GH, Pereira SC, Tanus-Santos JE et al (2019) Pharmacogenomics and hypertension: current insights. Pharmgenomics Pers Med 12:341–359 76. Israili ZH (2000) Clinical pharmacokinetics of angiotensin II (AT1) receptor blockers in hypertension. J Hum Hypertens 14:S73–S86

77. Aulakh GK, Sodhi RK, Singh M (2007) An update on non-peptide angiotensin receptor antagonists and related RAAS modulators. Life Sci 81:615–639 78. Flaten HK, Monte AA (2017) The Pharmacogenomic and Metabolomic predictors of ACE inhibitor and angiotensin II receptor blocker effectiveness and safety. Cardiovasc Drugs Ther 31:471–482 79. Kamiyama E, Yoshigae Y, Kasuya A et al (2007) Inhibitory effects of angiotensin receptor blockers on CYP2C9 activity in human liver microsomes. Drug Metab Pharmacokinet 22:267–275 80. Yasar U, Tybring G, Hidestrand M et al (2001) Role of CYP2C9 polymorphism in losartan oxidation. Drug Metab Dispos 29: 1051–1056 81. Cabaleiro T, Roma´n M, Ochoa D et al (2013) Evaluation of the relationship between sex, polymorphisms in CYP2C8 and CYP2C9, and pharmacokinetics of angiotensin receptor blockers. Drug Metab Dispos 41:224–229 82. Hallberg P, Karlsson J, Kurland L et al (2002) The CYP2C9 genotype predicts the blood pressure response to irbesartan: results from the Swedish Irbesartan left ventricular hypertrophy investigation vs atenolol (SILVHIA) trial. J Hypertens 20:2089–2093 83. Kurland L, Melhus H, Karlsson J et al (2001) Angiotensin converting enzyme gene polymorphism predicts blood pressure response to angiotensin II receptor type 1 antagonist treatment in hypertensive patients. J Hypertens 19:1783–1787 84. Freel EM, Ingram M, Friel EC et al (2007) Phenotypic consequences of variation across the aldosterone synthase and 11-beta hydroxylase locus in a hypertensive cohort: data from the MRC BRIGHT study. Clin Endocrinol 67:832–838 85. Davies E, Holloway CD, Ingram MC et al (1999) Aldosterone excretion rate and blood pressure in essential hypertension are related to polymorphic differences in the aldosterone synthase gene CYP11B2. Hypertension 33: 703–707 86. Lacchini R, Sabha M, Coeli FB et al (2009) T allele of
344 C/T polymorphism in aldosterone synthase gene is not associated with resistant hypertension. Hypertens Res 32: 159–162 87. Kurland L, Melhus H, Karlsson J et al (2002) Aldosterone synthase (CYP11B2)
344 C/T polymorphism is related to antihypertensive response: result from the Swedish Irbesartan left ventricular hypertrophy investigation

Cardiovascular Pharmacogenomics versus atenolol (SILVHIA) trial. Am J Hypertens 15:389–393 88. Kurland L, Liljedahl U, Karlsson J et al (2004) Angiotensinogen gene polymorphisms: relationship to blood pressure response to antihypertensive treatment. Results from the Swedish Irbesartan left ventricular hypertrophy investigation vs atenolol (SILVHIA) trial. Am J Hypertens 17:8–13 89. Ortlepp JR, Hanrath P, Mevissen V et al (2002) Variants of the CYP11B2 gene predict response to therapy with candesartan. Eur J Pharmacol 445:151–152 90. Konoshita T, Kato N, Fuchs S et al (2009) Genetic variant of the renin-angiotensin system and diabetes influences blood pressure response to angiotensin receptor blockers. Diabetes Care 32:1485–1490 91. Ji X, Qi H, Li DB et al (2015) Associations between human aldosterone synthase CYP11B2 (
344T/C) gene polymorphism and antihypertensive response to valsartan in Chinese patients with essential hypertension. Int J Clin Exp Med 8:1173–1177 92. de Denus S, Zakrzewski-Jakubiak M, Dube´ MP et al (2008) Effects of AGTR1 A1166C gene polymorphism in patients with heart failure treated with candesartan. Ann Pharmacother 42:925–932 93. de Denus S, Dube´ MP, Fouodjio R et al (2018) A prospective study of the impact of AGTR1 A1166C on the effects of candesartan in patients with heart failure. Pharmacogenomics 19:599–612 94. Sun Y, Liao Y, Yuan Y et al (2014) Influence of autoantibodies against AT1 receptor and AGTR1 polymorphisms on candesartanbased antihypertensive regimen: results from the study of optimal treatment in hypertensive patients with anti-AT1-receptor autoantibodies trial. J Am Soc Hypertens 8:21–27 95. Artham S, Fouda AY, El-Remessy AB et al (2015) Vascular protective effects of angiotensin receptor blockers: beyond blood pressure. Receptors Clin Investig 2:e774 96. Mason RP, Jacob RF, Kubant R et al (2012) Effects of angiotensin receptor blockers on endothelial nitric oxide release: the role of eNOS variants. Br J Clin Pharmacol 74: 141–146 97. Hiltunen TP, Donner KM, Sarin AP et al (2015) Pharmacogenomics of hypertension: a genome-wide, placebo-controlled crossover study, using four classes of antihypertensive drugs. J Am Heart Assoc 4:e001521 98. Frau F, Zaninello R, Salvi E et al (2014) Genome-wide association study identifies

231

CAMKID variants involved in blood pressure response to losartan: the SOPHIA study. Pharmacogenomics 15:1643–1652 99. Turner ST, Bailey KR, Schwartz GL et al (2012) Genomic association analysis identifies multiple loci influencing antihypertensive response to an angiotensin II receptor blocker. Hypertension 59:1204–1211 100. Rysz J, Franczyk B, Rysz-Go´rzyn´ska M et al (2020) Pharmacogenomics of hypertension treatment. Int J Mol Sci 21:4709 101. Oh BH (2007) Aliskiren, the first in a new class of direct renin inhibitors for hypertension: present and future perspectives. Expert Opin Pharmacother 8:2839–2849 102. Volpe M, Pontremoli R, Borghi C (2011) Direct renin inhibition: from pharmacological innovation to novel therapeutic opportunities. High Blood Press Cardiovasc Prev 18: 93–105 103. Musini VM, Lawrence KA, Fortin PM et al (2017) Blood pressure lowering efficacy of renin inhibitors for primary hypertension. Cochrane Database Syst Rev 4:CD007066 104. Buczko W, Hermanowicz JM (2008) Pharmacokinetics and pharmacodynamics of aliskiren, an oral direct renin inhibitor. Pharmacol Rep 60:623–631 105. Vaidyanathan S, Camenisch G, Schuetz H et al (2008) Pharmacokinetics of the oral direct renin inhibitor aliskiren in combination with digoxin, atorvastatin, and ketoconazole in healthy subjects: the role of P-glycoprotein in the disposition of aliskiren. J Clin Pharmacol 48:1323–1338 106. Moore N, Dicker P, O’Brien JK et al (2007) Renin gene polymorphisms and haplotypes, blood pressure, and responses to reninangiotensin system inhibition. Hypertension 50:340–347 107. Vaidyanathan S, Jarugula V, Dieterich HA et al (2008) Clinical pharmacokinetics and pharmacodynamics of Aliskiren. Clin Pharmacokinet 47:515–531 108. Tapaninen T, Neuvonen PJ, Niemi M (2010) Rifampicin reduces the plasma concentrations and the renin-inhibiting effect of aliskiren. Eur J Clin Pharmacol 66:497–502 109. Tapaninen T, Neuvonen PJ, Niemi M (2011) Orange and apple juice greatly reduce the plasma concentrations of the OATP2B1 substrate aliskiren. Br J Clin Pharmacol 71: 718–726 110. Tapaninen T, Karonen T, Backman JT et al (2013) SLCO2B1 c.935G>a single nucleotide polymorphism has no effect on the

232

Mariana Babayeva et al.

pharmacokinetics of montelukast and aliskiren. Pharmacogenet Genomics 23:19–24 111. Weber WW, Hein DW (1985) N-acetylation pharmacogenetics. Pharmacol Rev 37:25–79 112. McDonagh EM, Boukouvala S, Aklillu E et al (2014) PharmGKB summary: very important pharmacogene information for N-acetyltransferase 2. Pharmacogenet Genomics 24:409–425 113. Hein DW, Millner LM (2021) Arylamine N-acetyltransferase acetylation polymorphisms: paradigm for pharmacogenomic-guided therapy- a focused review. Expert Opin Drug Metab Toxicol 17:9–21 114. Weinshilboum R, Wang L (2004) Pharmacogenomics: bench to bedside. Nat Rev Drug Discov 3:739–748 115. Hein DW (2009) N-acetyltransferase SNPs: emerging concepts serve as a paradigm for understanding complexities of personalized medicine. Expert Opin Drug Metab Toxicol 5:353–366 116. Parkin DP, Vandenplas S, Botha FJ et al (1997) Trimodality of isoniazid elimination: phenotype and genotype in patients with tuberculosis. Am J Respir Crit Care Med 155:1717–1722 117. Fukunaga K, Kato K, Okusaka T et al (2021) Functional characterization of the effects of N-acetyltransferase 2 alleles on N-acetylation of eight drugs and worldwide distribution of substrate-specific diversity. Front Genet 12: 652704 118. Allen CE, Doll MA, Hein DW (2017) Nacetyltransferase 2 genotype-dependent Nacetylation of hydralazine in human hepatocytes. Drug Metab Dispos 45:1276–1281 119. Hein DW, Doll MA (2012) Accuracy of various human NAT2 SNP genotyping panels to infer rapid, intermediate and slow acetylator phenotypes. Pharmacogenomics 13:31–41 120. Hein DW, Zhang X, Doll MA (2018) Role of N-acetyltransferase 2 acetylation polymorbis phism in 4, 40 -methylene (2-chloroaniline) biotransformation. Toxicol Lett 283:100–105 121. Cascorbi I, Drakoulis N, Brockmo¨ller J et al (1995) Arylamine N-acetyltransferase (NAT2) mutations and their allelic linkage in unrelated Caucasian individuals: correlation with phenotypic activity. Am J Hum Genet 57:581–592 122. Ruiz JD, Martı´nez C, Anderson K et al (2012) The differential effect of NAT2 variant alleles permits refinement in phenotype inference and identifies a very slow acetylation genotype. PLoS One 7:e44629

123. Selinski S, Blaszkewicz M, Getzmann S et al (2015) N-acetyltransferase 2: ultra-slow acetylators enter the stage. Arch Toxicol 89: 2445–2447 124. Collins KS, Raviele ALJ, Elchynski AL et al (2020) Genotype-guided hydralazine therapy. Am J Nephrol 51:764–776 125. Walker K, Ginsberg G, Hattis D et al (2009) Genetic polymorphism in N-acetyltransferase (NAT): population distribution of NAT1 and NAT2 activity. J Toxicol Environ Health B Crit Rev 12:440–472 126. Sherry ST, Ward M, Sirotkin K (1999) dbSNP-database for single nucleotide polymorphisms and other classes of minor genetic variation. Genome Res 9:677–679 127. Bell DA, Taylor JA, Butler MA et al (1993) Genotype/phenotype discordance for human arylamine N-acetyltransferase (NAT2) reveals a new slow-acetylator allele common in African-Americans. Carcinogenesis 14: 1689–1692 128. Gonzalez-Fierro A, Vasquez-Bahena D, TajaChayeb L et al (2011) Pharmacokinetics of hydralazine, an antihypertensive and DNA-demethylating agent, using controlled-release formulations designed for use in dosing schedules based on the acetylator phenotype. Int J Clin Pharmacol Ther 49: 519–524 129. Zacest R, Koch-Weser J (1972) Relation of hydralazine plasma concentration to dosage and hypotensive action. Clin Pharmacol Ther 13:420–425 130. Jounela AJ, Pasanen M, Mattila MJ (1975) Acetylator phenotype and the antihypertensive response to hydralazine. Acta Med Scand 197:303–306 131. Kalowski S, Hua AS, Whitworth JA et al (1979) Hydrallazine with beta-blocker and diuretic in the treatment of hypertension. A double-blind crossover study. Med J Aust 2: 439–440 132. Garce´s-Eisele SJ, Cedillo-Carvallo B, Reyes˜ ez V et al (2014) Genetic selection of Nu´n volunteers and concomitant dose adjustment leads to comparable hydralazine/valproate exposure. J Clin Pharm Ther 39:368–375 133. Timbrell JA, Harland SJ, Facchini V (1981) Effect of dose on acetylator phenotype distribution of hydralazine. Clin Pharmacol Ther 29:337–343 134. Han LW, Ryu RJ, Cusumano M et al (2019) Effect of N-acetyltransferase 2 genotype on the pharmacokinetics of hydralazine during pregnancy. J Clin Pharmacol 59:1678–1689

Cardiovascular Pharmacogenomics 135. Spinasse LB, Santos AR, Suffys PN et al (2014) Different phenotypes of the NAT2 gene influences hydralazine antihypertensive response in patients with resistant hypertension. Pharmacogenomics 15:169–178 136. Cameron HA, Ramsay LE (1984) The lupus syndrome induced by hydralazine: a common complication with low dose treatment. Br Med J 289:410–412 137. Ihle BU, Whitworth JA, Dowling JP et al (1984) Hydralazine and lupus nephritis. Clin Nephrol 22:230–238 138. Batchelor JR, Welsh KI, Tinoco RM et al (1980) Hydralazine-induced systemic lupus erythematosus: influence of HLA-DR and sex on susceptibility. Lancet 1:1107–1109 139. FDA. https://www.fda.gov/medicaldevices/precision-medicine/tablepharmacogenetic-associations 140. Roger VL, Go AS, Lloyd-Jones DM, Benjamin EJ et al (2012) Heart disease and stroke statistics – 2012 update: a report from the American heart association. Circulation 125: e2–e220 141. http://www.imshealth.com/ims/Global/ Content/Insights/IMS%20Institute%20for %20Healthcare%20 Infor matics/IHIIMedicines_in_U.S._Report-2011.pdf 142. Andrade SE, Walker AM, Gottlieb LK et al (1995) Discontinuation of antihyperlipidemic drugs – do rates reported in clinical trials reflect rates in primary care settings? N Engl J Med 332:1125–1131 143. Wilhelmsen L, Link EM, Parish S et al (2008) SLCO1B1 variants and statin-induced myopathy – a genome-wide study. N Engl J Med 359:789–799 144. Ramsey LB, Johnson SG, Caudle KE et al (2014) The clinical pharmacogenetics implementation consortium guideline for SLCO1B1 and simvastatin-induced myopathy:2014 update. Clin Pharmaco Ther 96: 423–428 145. Vassy JL, Gaziano JM, Green RC et al (2020) Effect of pharmacogenetic testing for statin myopathy risk vs usual care on blood cholesterol: a randomized clinical trial. JAMA Netw Open 3:e2027092 146. Bycroft C, Freeman C, Petkova D et al (2018) The UK biobank resource with deep phenotyping and genomic data. Nature 562: 203–228 147. Wendt FR, Koller D, Pathak GA et al (2021) Biobank scale pharmacogenomics informs the genetic underpinnings of simvastatin use. Clin Pharmacol Ther 110:777–785

233

148. RxList. Coumadin (warfarin sodium): uses, dosage, side effects, interactions, warning (rxlist.com) 149. Hirsh J, Fuster V, Ansell J et al (2003) American Heart Association/American College of Cardiology Foundation guide to warfarin therapy. J Am Coll Cardiol 41:1633–1652 150. Shaik AN, Grater R, Lulla M et al (2016) Comparison of enzyme kinetics of warfarin analyzed by LC-MS/MS QTrap and differential mobility spectrometry. J Chromatogr B Analyt Technol Biomed Life Sci 1008: 164–173 151. Wadelius M, Chen LY, Downes K et al (2005) Common VKORC1 and GGCX polymorphisms associated with warfarin dose. Pharmacogenomics J 5:262–270 152. Weeke P, Roden DM (2014) Applied pharmacogenomics in cardiovascular medicine. Annu Rev Med 65:81–94 153. Liu J, Jiang HH, Wu DK et al (2017) Effect of gene polymorphims on the warfarin treatment at initial stage. Pharmacogenomics J 17:47–52 154. Takeuchi M, Kobayashi T, Biss T et al (2020) CYP2C9, VKORC1, and CYP4F2 polymorphisms and pediatric warfarin maintenance dose: a systematic review and meta-analysis. Pharmacogenomics J 20: 306–319 155. Voora D, Ginsburg GS (2012) Clinical application of cardiovascular pharmacogenetics. J Am Coll Cardiol 60:9–20 156. Yin T, Miyata T (2007) Warfarin dose and the pharmacogenomics of CYP2C9 and VKORC1 - rationale and perspectives. Thromb Res 120:1–10 157. Gage BF, Eby C, Johnson JA et al (2008) Use of pharmacogenetic and clinical factors to predict the therapeutic dose of warfarin. Clin Pharmacol Ther 84:326–331 158. Wadelius M, Chen LY, Lindh JD et al (2009) The largest prospective warfarin-treated cohort supports genetic forecasting. Blood 113:784–792 159. Johnson JA, Caudle KE, Gong L et al (2017) Clinical Pharmacogenetics implementation consortium (CPIC) guideline for Pharmacogenetics-guided warfarin dosing: 2017 update. Clin Pharmacol Ther 102: 397–404 160. Perera MA, Cavallari LH, Limdi NA et al (2013) Genetic variants associated with warfarin dose in African-American individuals: a genome-wide association study. Lancet 382: 790–796 161. Henderson LM, Robinson RF, Ray L et al (2019) VKORC1 and novel CYP2C9

234

Mariana Babayeva et al.

variation predict warfarin response in Alaska native and American Indian people. Clin Transl Sci 12:312–320 162. Rost S, Fregin A, Ivaskevicius V et al (2004) Mutations in VKORC1 cause warfarin resistance and multiple coagulation factor deficiency type 2. Nature 427:537–541 163. Rieder MJ, Reiner AP, Gage BF et al (2005) Effect of VKORC1 haplotypes on transcriptional regulation and warfarin dose. N Engl J Med 352:2285–2293 164. Schelleman H, Brensinger CM, Chen J et al (2010) New genetic variant that might improve warfarin dose prediction in African Americans. Br J Clin Pharmacol 70:393–399 165. Sipeky C, Melegh B (2008) Haplogroup analysis of vitamin-K epoxide reductase (VKORC1) gene: novel element in the optimization of anticoagulant therapy. Orv Hetil 149:1839–1844 166. Limdi NA, Beasley TM, Crowley MR et al (2008) VKORC1 polymorphisms, haplotypes and haplotype groups on warfarin dose among African-Americans and EuropeanAmericans. Pharmacogenomics 9:1445–1458 167. Sagreiya H, Berube C, Wen A et al (2010) Extending and evaluating a warfarin dosing algorithm that includes CYP4F2 and pooled rare variants of CYP2C9. Pharmacogenet Genomics 20:407–413 168. Shendre A, Brown TM, Liu N et al (2016) Race-specific influence of CYP4F2 on dose and risk of Hemorrhage among warfarin users. Pharmacotherapy 36:263–272 169. Kalpana SR, Bharath G, Simran Jain NM et al (2019) Prosthetic valve thrombosis – association of genetic polymorphisms of VKORC1, CYP2C9 and CYP4F2 genes. Medicine (Baltimore). 98:e14365 170. Sun X, Yu WY, Ma WL et al (2016) Impact of the CYP4F2 gene polymorphisms on the warfarin maintenance dose: a systematic review and meta-analysis. Biomed Rep 4:498–506 171. Caldwell MD, Awad T, Johnson JA et al (2008) CYP4F2 genetic variant alters required warfarin dose. Blood 111: 4106–4112 172. Nakamura K, Obayashi K, Araki T et al (2012) CYP4F2 gene polymorphism as a contributor to warfarin maintenance dose in Japanese subjects. J Clin Pharm Ther 37: 481–485 173. Liang R, Wang C, Zhao H et al (2012) Influence of CYP4F2 genotype on warfarin dose requirement-a systematic review and metaanalysis. Thromb Res 130:38–44 174. Al-Eitan LN, Almasri AY, Alnaamneh AH et al (2021) Influence of CYP4F2, ApoE,

and CYP2A6 gene polymorphisms on the variability of warfarin dosage requirements and susceptibility to cardiovascular disease in Jordan. Int J Med Sci 18:826–834 175. Singh O, Sandanaraj E, Subramanian K et al (2011) Influence of CYP4F2 rs2108622 (V433M) on warfarin dose requirement in Asian patients. Drug Metab Pharmacokinet 26:130–136 176. Pirmohamed M, Burnside G, Eriksson N et al (2013) A randomized trial of genotypeguided dosing of warfarin. N Engl J Med 369:2294–2303 177. Kimmel SE, French B, Kasner SE et al (2013) A pharmacogenetic versus a clinical algorithm for warfarin dosing. N Engl J Med 369: 2283–2293 178. Stergiopoulos K, Brown DL (2014) Genotype-guided vs clinical dosing of warfarin and its analogues: meta-analysis of randomized clinical trials. JAMA Intern Med 174: 1330–1338 179. RxList. Pradaxa. https://www.rxlist.com/ pradaxa-drug.htm#indications ˜ o E et al (2018) 180. Cullell N, Carrera C, Muin Pharmacogenetic studies with oral anticoagulants. Genome-wide association studies in vitamin K antagonist and direct oral anticoagulants. Oncotarget 9:29238–29258 181. Dimatteo C, D’Andrea G, Vecchione G et al (2016) Pharmacogenetics of dabigatran etexilate interindividual variability. Thromb Res 144:1–5 182. Merali Z, Ross S, Pare´ G (2014) The pharmacogenetics of carboxylesterases: CES1 and CES2 genetic variants and their clinical effect. Drug Metabol Drug Interact 29:143–151 183. Shi J, Wang X, Nguyen JH et al (2016) Dabigatran etexilate activation is affected by the CES1 genetic polymorphism G143E (rs71647871) and gender. Biochem Pharmacol 119:76–84 184. Shi J, Wang X, Eyler RF et al (2016) Association of Oseltamivir Activation with gender and carboxylesterase 1 genetic polymorphisms. Basic Clin Pharmacol Toxicol 119: 555–561 185. Her L, Zhu HJ (2020) Carboxylesterase 1 and precision pharmacotherapy: Pharmacogenetics and nongenetic regulators. Drug Metab Dispos 48:230–244 186. Pare´ G, Eriksson N, Lehr T et al (2013) Genetic determinants of dabigatran plasma levels and their relation to bleeding. Circulation 127:1404–1412 187. Ji Q, Zhang C, Xu Q et al (2021) The impact of ABCB1 and CES1 polymorphisms on dabigatran pharmacokinetics and

Cardiovascular Pharmacogenomics pharmacodynamics in patients with atrial fibrillation. Br J Clin Pharmacol 87: 2247–2255 188. Reilly PA, Lehr T, Haertter S et al (2014) The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients: the RE-LY trial (randomized evaluation of Long-term anticoagulation therapy). J Am Coll Cardiol 63:321–328 189. Liu Y, Yang C, Qi W et al (2021) The impact of ABCB1 and CES1 polymorphisms on dabigatran pharmacokinetics in healthy Chinese subjects. Pharmgenomics Pers Med 14: 477–485 190. Sychev D, Skripka A, Ryzhikova K et al (2020) Effect of CES1 and ABCB1 genotypes on the pharmacokinetics and clinical outcomes of dabigatran etexilate in patients with atrial fibrillation and chronic kidney disease. Drug Metab Pers Ther 35(1) 191. Xie Q, Xiang Q, Mu G et al (2018) Effect of ABCB1 genotypes on the pharmacokinetics and clinical outcomes of new Oral anticoagulants: a systematic review and meta-analysis. Curr Pharm Des 24:3558–3565 192. O’connor CT, Kiernan TJ, Yan BP (2017) The genetic basis of antiplatelet and anticoagulant therapy: a pharmacogenetic review of newer antiplatelets (clopidogrel, prasugrel and ticagrelor) and anticoagulants (dabigatran, rivaroxaban, apixaban and e´doxaban) expert Opin. Drug Metab. Toxicology 13: 725–739 193. Ing Lorenzini K, Daali Y, Fontana P et al (2016) Rivaroxaban-induced Hemorrhage associated with ABCB1 genetic defect. Front Pharmacol 7:494 194. Sychev DA, Vardanyan A, Rozhkov A et al (2018) CYP3A activity and rivaroxaban serum concentrations in Russian patients with deep vein thrombosis. Genet Test Mol Biomark 22:51–54 195. Wang D, Guo Y, Wrighton SA et al (2011) Intronic polymorphism in CYP3A4 affects hepatic expression and response to statin drugs. Pharm J 11:274–286 196. Dai D, Tang J, Rose R et al (2001) Identification of variants of CYP3A4 and characterization of their abilities to metabolize testosterone and chlorpyrifos. J Pharmacol Exp Ther 299:825–831 197. Sychev D, Minnigulov R, Bochkov P et al (2019) Effect of CYP3A4, CYP3A5, ABCB1 gene polymorphisms on rivaroxaban pharmacokinetics in patients undergoing total hip

235

and knee replacement surgery. High Blood Press Cardiovasc Prev 26:413–420 198. Gouin-Thibault I, Delavenne X, Blanchard A et al (2017) Interindividual variability in dabigatran and rivaroxaban exposure: contribution of ABCB1 genetic polymorphisms and interaction with clarithromycin. J Thromb Haemost 15:273–283 199. Kanuri SH, Kreutz RP (2019) Pharmacogenomics of novel direct oral anticoagulants: newly identified genes and genetic variants. J Pers Med 9:7 200. Dimatteo C, D’Andrea G, Vecchione G et al (2016) ABCB1 SNP rs4148738 modulation of apixaban interindividual variability. Thromb Res 145:24–26 201. Kryukov AV, Sychev DA, Andreev DA et al (2018) Influence of ABCB1 and CYP3A5 gene polymorphisms on pharmacokinetics of apixaban in patients with atrial fibrillation and acute stroke. Pharm Pers Med 11:43–49 202. Ueshima S, Hira D, Fujii R et al (2017) Impact of ABCB1, ABCG2, and CYP3A5 polymorphisms on plasma trough concentrations of apixaban in Japanese patients with atrial fibrillation. Pharm Genom 27:329–336 203. Gulilat M, Keller D, Linton B et al (2020) Drug interactions and pharmacogenetic factors contribute to variation in apixaban concentration in atrial fibrillation patients in routine care. J Thromb. Thrombolysis. 49: 294–303 204. Sherry ST, Ward MH, Kholodov M et al (2001) dbSNP: the NCBI database of genetic variation. Nucleic Acids Res 29:308–311 205. Nagar S, Walther S, Blanchard RL (2006) Sulfotransferase (SULT) 1A1 polymorphic variants *1, *2, and *3 are associated with altered enzymatic activity, cellular phenotype, and protein degradation. Mol Pharmacol 69: 2084–2092 206. Asˇic´ A, Marjanovic´ D, Mirat J et al (2018) Pharmacogenetics of novel oral anticoagulants: a review of identified gene variants & future perspectives. Per Med 15:209–221 207. Vandell AG, Lee J, Shi M et al (2018) An integrated pharmacokinetic/pharmacogenomic analysis of ABCB1 and SLCO1B1 polymorphisms on edoxaban exposure. Pharm J 18:153–159 208. Raymond J, Imbert L, Cousin T et al (2021) Pharmacogenetics of direct Oral anticoagulants: a systematic review. J Pers Med 11:37 209. Palladino M, Merli G, Thomson L (2013) Evaluation of the oral direct factor Xa inhibitor–Betrixaban. Expert Opin Investig Drugs 22:1465–1472

236

Mariana Babayeva et al.

210. Non-vitamin K antagonist Oral Anticoagulant (NOAC) guidelines. Updated 2017. Non-vitamin K antagonist Oral Anticoagulant (NOAC) Guidelines (nsw.gov.au) 211. Roe MT, Armstrong PW, Fox KA et al (2012) Prasugrel versus clopidogrel for acute coronary syndromes without revascularization. N Engl J Med 367:1297–1309 212. Wallentin L, Becker RC, Budaj A et al (2009) Ticagrelor versus clopidogrel in patients with acute coronary syndromes. N Engl J Med 361:1045–1057 213. Patti G, Micieli G, Cimminiello C et al (2020) The role of Clopidogrel in 2020: a reappraisal. Cardiovasc Ther 2020:8703627 214. Cattaneo M (2012) Response variability to clopidogrel: is tailored treatment, based on laboratory testing, the right solution? J Thromb Haemost 10:327–336 215. Mugosa S, Todorovic Z, Cukic J et al (2021) ABCB1 polymorphism in clopidogrel-treated Montenegrin patients. Open Life Sci 16: 142–149 216. Karaz´niewicz-Łada M, Danielak D, Rubis´ B et al (2015) Impact of common ABCB1 polymorphism on pharmacokinetics and pharmacodynamics of clopidogrel and its metabolites. J Clin Pharm Ther 40:226–231 217. Shuldiner AR, O’Connell JR, Bliden KP et al (2009) Association of cytochrome P450 2C19 genotype with the antiplatelet effect and clinical efficacy of clopidogrel therapy. JAMA 302:849–857 218. Peters BJ, Harmsze AM, ten Berg JM et al (2011) CYP2C19 and ABCB1 genes and individualized treatment with clopidogrel. Pharmacogenomics 12:141–144 219. Harmsze AM, van Werkum JW, Ten Berg JM et al (2010) CYP2C19*2 and CYP2C9*3 alleles are associated with stent thrombosis: a case-control study. Eur Heart J 31: 3046–3053 220. Tiroch KA, Sibbing D, Koch W et al (2010) Protective effect of the CYP2C19 *17 polymorphism with increased activation of clopidogrel on cardiovascular events. Am Heart J 160:506–512 221. Zhu HJ, Wang X, Gawronski BE et al (2013) Carboxylesterase 1 as a determinant of clopidogrel metabolism and activation. J Pharmacol Exp Ther 344:665–672 222. Scott SA, Sangkuhl K, Stein CM et al (2013) Clinical Pharmacogenetics implementation consortium guidelines for CYP2C19 genotype and clopidogrel therapy: 2013 update. Clin Pharmacol Ther 94:317–323

223. Umemura K, Furuta T, Kondo K (2008) The common gene variants of CYP2C19 affect pharmacokinetics and pharmacodynamics in an active metabolite of clopidogrel in healthy subjects. J Thromb Haemost 6:1439–1441 224. Mega JL, Close SL, Wiviott SD et al (2009) Cytochrome p-450 polymorphisms and response to clopidogrel. N Engl J Med 360: 354–362 225. Mega JL, Simon T, Collet JP et al (2010) Reduced-function CYP2C19 genotype and risk of adverse clinical outcomes among patients treated with clopidogrel predominantly for PCI: a meta-analysis. JAMA 304: 1821–1830 226. Mega JL, Close SL, Wiviott SD et al (2009) Cytochrome P450 genetic polymorphisms and the response to prasugrel: relationship to pharmacokinetic, pharmacodynamic, and clinical outcomes. Circulation 119: 2553–2560 227. Mega JL, Close SL, Wiviott SD et al (2010) Genetic variants in ABCB1 and CYP2C19 and cardiovascular outcomes after treatment with clopidogrel and prasugrel in the TRITON-TIMI 38 trial: a pharmacogenetic analysis. Lancet 376:1312–1319 228. Pereira NL, Rihal CS, So DYF et al (2019) Clopidogrel Pharmacogenetics. Circ Cardiovasc Interv 12:e007811 229. Lee CR, Sriramoju VB, Cervantes A et al (2018) Clinical outcomes and sustainability of using CYP2C19 genotype-guided antiplatelet therapy after percutaneous coronary intervention. Circ Genom Precis Med 11: e002069 230. Dahl M-L, Arzu Gunes A (2010) Implications of inter-individual differences in Clopidogrel metabolism, with focus on Pharmacogenetics. Pharmaceuticals 3: 782–794 231. Yamamoto K, Hokimoto S, Chitose T et al (2011) Impact of CYP2C19 polymorphism on residual platelet reactivity in patients with coronary heart disease during antiplatelet therapy. J Cardiol 57:194–201 232. Jeong YH, Tantry US, Kim IS et al (2011) Effect of CYP2C19*2 and *3 loss-of-function alleles on platelet reactivity and adverse clinical events in east Asian acute myocardial infarction survivors treated with clopidogrel and aspirin. Circ Cardiovasc Interv 4: 585–594 233. Langaee TY, Zhu HJ, Wang X et al (2014) The influence of the CYP2C19*10 allele on clopidogrel activation and CYP2C19*2

Cardiovascular Pharmacogenomics genotyping. Pharmacogenet Genomics 24: 381–386 234. Fre´re C, Cuisset T, Gaborit B et al (2009) The CYP2C19*17 allele is associated with better platelet response to clopidogrel in patients admitted for non-ST acute coronary syndrome. J Thromb Haemost 7:1409–1411 235. Sibbing D, Gross L (2018) CYP2C19 genotyping in percutaneous coronary intervention-treated patients: ready for prime time? JACC Cardiovasc Interv 11: 192–194 236. Sibbing D, Gebhard D, Koch W et al (2010) Isolated and interactive impact of common CYP2C19 genetic variants on the antiplatelet effect of chronic clopidogrel therapy. J Thromb Haemost 8:1685–1693 237. Mao L, Jian C, Changzhi L et al (2013) Cytochrome CYP2C19 polymorphism and risk of adverse clinical events in clopidogrel-treated patients: a meta-analysis based on 23,035 subjects. Arch Cardiovasc Dis 106:517–527 238. Simon T, Verstuyft C, Mary-Krause M et al (2009) Genetic determinants of response to clopidogrel and cardiovascular events. N Engl J Med 360:363–375 239. Geisler T, Bigalke B, Schwab M (2011) CYP2C19 genotype and outcomes of clopidogrel treatment. N Engl J Med 364:481 240. Zhuo ZL, Xian HP, Long Y et al (2018) Association between CYP2C19 and ABCB1 polymorphisms and clopidogrel resistance in clopidogrel-treated Chinese patients. Anatol J Cardiol 19:123–129 241. Gurusamy U, Shewade DG (2014) Chapter 46. Pharmacogenomics in India. In: Padmanabhan S (ed) Handbook of pharmacogenomics and stratified medicine. Academic Press, Amsterdam, pp 1037–1059 242. Jarrar M, Behl S, Manyam G et al (2016) Cytochrome allelic variants and clopidogrel metabolism in cardiovascular diseases therapy. Mol Biol Rep 43:473–484 243. Lei HP, Yu XY, Wu H et al (2018) Effects of PON1 gene promoter DNA methylation and genetic variations on the clinical outcomes of dual antiplatelet therapy for patients undergoing percutaneous coronary intervention. Clin Pharmacokinet 57:817–829 244. Bouman HJ, Scho¨mig E, van Werkum JW et al (2011) Paraoxonase-1 is a major determinant of clopidogrel efficacy. Nat Med 17: 110–116 245. Pereira N, Jeffrey B, Geske JB et al (2016) Pharmacogenetics of clopidogrel. An unresolved issue. Circ Cardiovasc Genet 9: 185–188

237

246. Frelinger AL 3rd, Bhatt DL, Lee RD et al (2013) Clopidogrel pharmacokinetics and pharmacodynamics vary widely despite exclusion or control of polymorphisms (CYP2C19, ABCB1, PON1), noncompliance, diet, smoking, co-medications (including proton pump inhibitors), and pre-existent variability in platelet function. J Am Coll Cardiol 61: 872–879 247. Levine GN, Bates ER, Blankenship JC et al (2011) 2011 ACCF/AHA/SCAI guideline for percutaneous coronary intervention. A report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines and the Society for Cardiovascular Angiography and Interventions. J Am Coll Cardiol 58: e44–e122 248. Lewis JP, Horenstein RB, Ryan K et al (2013) The functional G143E variant of carboxylesterase 1 is associated with increased clopidogrel active metabolite levels and greater clopidogrel response. Pharmacogenet Genomics 23:1–8 249. Bozzi LM, Mitchell BD, Lewis JP et al (2016) The pharmacogenomics of anti-platelet intervention (PAPI) study: variation in platelet response to Clopidogrel and aspirin. Curr Vasc Pharmacol 14:116–124 250. Jiang XL, Samant S, Lewis JP et al (2016) Development of a physiology-directed population pharmacokinetic and pharmacodynamic model for characterizing the impact of genetic and demographic factors on clopidogrel response in healthy adults. Eur J Pharm Sci 82:64–78 251. Tarkiainen EK, Holmberg MT, Tornio A et al (2015) Carboxylesterase 1 c.428G>a single nucleotide variation increases the antiplatelet effects of clopidogrel by reducing its hydrolysis in humans. Clin Pharmacol Ther 97: 650–658 252. Grosdidier C, Quilici J, Loosveld M et al (2013) Effect of CYP2C19*2 and *17 genetic variants on platelet response to clopidogrel and prasugrel maintenance dose and relation to bleeding complications. Am J Cardiol 111: 985–990 253. Ma´chal J, Hlinomaz O (2019) Efficacy of P2Y12 receptor blockers after myocardial infarction and genetic variability of their metabolic pathways. Curr Vasc Pharmacol 17: 35–40 254. Alexopoulos D, Dimitropoulos G, Davlouros P et al (2011) Prasugrel overcomes high on-clopidogrel platelet reactivity poststenting more effectively than high-dose (150 mg) clopidogrel: the importance of

238

Mariana Babayeva et al.

CYP2C19*2 genotyping. JACC Cardiovasc Interv 4:403–410 255. Sorich MJ, Vitry A, Ward MB et al (2010) Prasugrel vs. clopidogrel for cytochrome P450 2C19-genotyped subgroups: integration of the TRITON-TIMI 38 trial data. J Thromb Haemost 8:1678–1684 256. Ge Z, Baber U, Claessen BE et al (2019) Associations between use of prasugrel vs clopidogrel and outcomes by type of acute coronary syndrome: an analysis from the PROMETHEUS registry. J Thromb Thrombolysis 48:42–51 257. Cuisset T, Loosveld M, Morange PE et al (2012) CYP2C19*2 and *17 alleles have a significant impact on platelet response and bleeding risk in patients treated with prasugrel after acute coronary syndrome. JACC Cardiovasc Interv 5:1280–1287 258. NCBI. CYP2C19 cytochrome P450 family 2 subfamily C member 19 [ Homo sapiens (human)] CYP2C19 cytochrome P450 family 2 subfamily C member 19 [Homo sapiens (human)] – Gene – NCBI (nih.gov) 259. Tantry US, Bliden KP, Wei C et al (2010) First analysis of the relation between CYP2C19 genotype and pharmacodynamics in patients treated with ticagrelor versus clopidogrel: the ONSET/OFFSET and RESPOND genotype studies. Circ Cardiovasc Genet 3:556–566 260. Wallentin L, James S, Storey RF et al (2010) Effect of CYP2C19 and ABCB1 single nucleotide polymorphisms on outcomes of treatment with ticagrelor versus clopidogrel for acute coronary syndromes: a genetic substudy of the PLATO trial. Lancet 376:1320–1328 261. Mohammed Ahmed A, Abubakar S, Nur Aizati Athirah D et al (2020) The role of pharmacogenetics and pharmacometabonomics in the personalization of ticagrelor antiplatelet therapy. Figshare. Conference contribution. https://doi.org/10.6084/m9. figshare.13426031.v1 262. Zhu Q, Zhong W, Wang X et al (2019) Pharmacokinetic and Pharmacogenetic factors contributing to platelet function recovery after single dose of Ticagrelor in healthy subjects. Front Pharmacol 10:209 263. Schu¨pke S, Neumann F-J, Maurizio Menichelli M et al (2019) Ticagrelor or Prasugrel in patients with acute coronary syndromes. N Engl J Med 381:1524–1534 264. Wurtz M, Lordkipanidze M, Grove EL (2013) Pharmacogenomics in cardiovascular disease: focus on aspirin and ADP receptor

antagonists. J Thromb Haemost 11: 1627–1639 265. Ahmad T, Voora D, Becker RC (2011) The pharmacogenetics of antiplatelet agents: towards personalized therapy? Nat Rev Cardiol 8:560–571 266. Al-Azzam SI, Alzoubi KH, Khabour OF et al (2013) The contribution of platelet glycoproteins (GPIa C807T and GPIba C-5T) and cyclooxygenase 2 (COX-2G-765C) polymorphisms to platelet response in patients treated with aspirin. Gene 526:118–121 267. Krasopoulos G, Brister SJ, Beattie WS et al (2008) Aspirin “resistance” and risk of cardiovascular morbidity: systematic review and meta-analysis. BMJ 336:195–198 268. Guirgis M, Thompson P, Jansen S (2017) Review of aspirin and clopidogrel resistance in peripheral arterial disease. J Vasc Surg 66: 1576–1586 269. Goodman T, Ferro A, Sharma P (2008) Pharmacogenetics of aspirin resistance: a comprehensive systematic review. Br J Clin Pharmacol 66:222–232 270. Wang H, Sun X, Dong W et al (2018) Association of GPIa and COX-2 gene polymorphism with aspirin resistance. J Clin Lab Anal 32:e22331 271. Feher G, Feher A, Pusch G et al (2009) The genetics of antiplatelet drug resistance. Clin Genet 75:1–18 272. Abderrazek F, Chakroun T, Addad F et al (2010) The GPIIIa PlA polymorphism and the platelet hyperactivity in Tunisian patients with stable coronary artery disease treated with aspirin. Thromb Res 125:e265–e268 273. Lordkipanidze´ M, Diodati JG, Palisaitis DA et al (2011) Genetic determinants of response to aspirin: appraisal of 4 candidate genes. Thromb Res 128:47–53 274. Kastrati A, Koch W, Gawaz M et al (2000) PlA polymorphism of glycoprotein IIIa and risk of adverse events after coronary stent placement. J Am Coll Cardiol 36:84–89 275. Voora D, Horton J, Shah SH et al (2011) Polymorphisms associated with in vitro aspirin resistance are not associated with clinical outcomes in patients with coronary artery disease who report regular aspirin use. Am Heart J 162:166–172 276. Clappers N, van Oijen MG, Sundaresan S et al (2008) The C50T polymorphism of the cyclooxygenase-1 gene and the risk of thrombotic events during low-dose therapy with acetyl salicylic acid. Thromb Haemost 100: 70–75

Cardiovascular Pharmacogenomics 277. Jefferson BK, Foster JH, McCarthy JJ et al (2005) Aspirin resistance and a single gene. Am J Cardiol 95:805–808 278. Xu ZH, Jiao JR, Yang R et al (2012) Aspirin resistance: clinical significance and genetic polymorphism. J Int Med Res 40:282–292 279. Fan L, Cao J, Liu L et al (2013) Frequency, risk factors, prognosis, and genetic polymorphism of the cyclooxygenase-1 gene for aspirin resistance in elderly Chinese patients with cardiovascular disease. Gerontology 59: 122–131 280. Maree AO, Curtin RJ, Chubb A et al (2005) Cyclooxygenase-1 haplotype modulates platelet response to aspirin. J Thromb Haemost 3:2340–2345 281. Lemaitre RN, Rice K, Marciante K et al (2009) Variation in eicosanoid genes, non-fatal myocardial infarction and ischemic stroke. Atherosclerosis 204:e58–e63 282. Yi X, Cheng W, Lin J et al (2016) Interaction between COX-1 and COX-2 variants associated with aspirin resistance in Chinese stroke patients. J Stroke Cerebrovasc Dis 25: 2136–2144 283. Faraday N, Yanek LR, Yang XP et al (2011) Identification of a specific intronic PEAR1 gene variant associated with greater platelet aggregability and protein expression. Blood 118:3367–3375 284. Lewis JP, Ryan K, O’Connell JR et al (2013) Genetic variation in PEAR1 is associated with platelet aggregation and cardiovascular outcomes. Circ Cardiovasc Genet 6:184–192 285. Chasman DI, Shiffman D, Zee RY et al (2009) Polymorphism in the apolipoprotein (a) gene, plasma lipoprotein(a), cardiovascular disease, and low-dose aspirin therapy. Atherosclerosis 203:371–376 286. Gurbel PA, Shuldiner AR, Bliden KP et al (2011) The relation between CYP2C19 genotype and phenotype in stented patients on maintenance dual antiplatelet therapy. Am Heart J 161:598–604 287. Timur AA, Murugesan G, Zhang L et al (2012) P2RY1 and P2RY12 polymorphisms and on-aspirin platelet reactivity in patients with coronary artery disease. Int J Lab Hematol 34:473–483 288. Kunicki TJ, Williams SA, Nugent DJ et al (2009) Lack of association between aspirin responsiveness and seven candidate gene haplotypes in patients with symptomatic vascular disease. Thromb Haemost 101:123–133 289. Patocka J, Nepovimova E, Wu W et al (2020) Digoxin: pharmacology and toxicology-a

239

review. Environ Toxicol Pharmacol 79: 103400 290. Birkenfeld AL, Jordan J, Hofmann U et al (2009) Genetic influences on the pharmacokinetics of orally and intravenously administered digoxin as exhibited by monozygotic twins. Clin Pharmacol Ther 86:605–608 291. FDA Approved Drug Products: Lanoxin (digoxin) oral tablets. label (fda.gov) 292. Gozalpour E, Wittgen HG, van den Heuvel JJ et al (2013) Interaction of digitalis-like compounds with p-glycoprotein. Toxicol Sci 131: 502–511 293. Lee VHL, Sporty JL, Fandy TE (2001) Pharmacogenomics of drug transporters: the next drug delivery challenge. Adv Drug Deliv Rev 50:S33–S40 294. Fung KL, Gottesman MM (2009) A synonymous polymorphism in a common MDR1 (ABCB1) haplotype shapes protein function. Biochim Biophys Acta 1794:860–871 295. Xu P, Jiang ZP, Zhang BK et al (2008) Impact of MDR1 haplotypes derived from C1236T, G2677T/a and C3435T on the pharmacokinetics of single-dose oral digoxin in healthy Chinese volunteers. Pharmacology 82: 221–227 296. Aarnoudse AJ, Dieleman JP, Visser LE et al (2008) Common ATP-binding cassette B1 variants are associated with increased digoxin serum concentration. Pharmacogenet Genomics 18:299–305 297. Hsin CH, Stoffel MS, Gazzaz M et al (2020) Combinations of common SNPs of the transporter gene ABCB1 influence apparent bioavailability, but not renal elimination of oral digoxin. Sci Rep 10:12457 298. Neuvonen AM, Palo JU, Sajantila A (2011) Post-mortem ABCB1 genotyping reveals an elevated toxicity for female digoxin users. Int J Legal Med 125:265–269 299. Kurzawski M, Bartnicka L, Florczak M et al (2007) Impact of ABCB1 (MDR1) gene polymorphism and P-glycoprotein inhibitors on digoxin serum concentration in congestive heart failure patients. Pharmacol Rep 59: 107–111 300. Oni-Orisan A, Lanfear DE (2014) Pharmacogenomics in heart failure: where are we now and how can we reach clinical application? Cardiol Rev 22:193–198 301. Larsen UL, Hyldahl Olesen L, Guldborg Nyvold C et al (2007) Human intestinal P-glycoprotein activity estimated by the model substrate digoxin. Scand J Clin Lab Invest 67:123–134

240

Mariana Babayeva et al.

302. Hoffmeyer S, Burk O, von Richter O et al (2000) Functional polymorphisms of the human multidrug-resistance gene: multiple sequence variations and correlation of one allele with P-glycoprotein expression and activity in vivo. Proc Natl Acad Sci U S A 97:3473–3478 303. Verstuyft C, Schwab M, Schaeffeler E et al (2003) Digoxin pharmacokinetics and MDR1 genetic polymorphisms. Eur J Clin Pharmacol 58:809–812 304. Johne A, Kopke K, Gerloff T et al (2002) Modulation of steady-state kinetics of digoxin by haplotypes of the P-glycoprotein MDR1 gene. Clin Pharmacol Ther 72:584–594 305. Comets E, Verstuyft C, Lavielle M et al (2007) Modelling the influence of MDR1 polymorphism on digoxin pharmacokinetic parameters. Eur J Clin Pharmacol 63: 437–449 306. Gerloff T, Schaefer M, Johne A et al (2002) MDR1 genotypes do not influence the absorption of a single oral dose of 1 mg digoxin in healthy white males. Br J Clin Pharmacol 54:610–616 307. Chowbay B, Li H, David M et al (2005) Meta-analysis of the influence of MDR1 C3435T polymorphism on digoxin pharmacokinetics and MDR1 gene expression. Br J Clin Pharmacol 60:159–171 308. Sakaeda T, Nakamura T, Horinouchi M et al (2001) MDR1 genotype-related pharmacokinetics of digoxin after single oral administration in healthy Japanese subjects. Pharm Res 18:1400–1404 309. Horinouchi M, Sakaeda T, Nakamura T et al (2002) Significant genetic linkage of MDR1 polymorphisms at positions 3435 and 2677: functional relevance to pharmacokinetics of digoxin. Pharm Res 19:1581–1585 310. Ameyaw MM, Regateiro F, Li T et al (2001) MDR1 pharmacogenetics: frequency of the C3435T mutation in exon 26 is significantly influenced by ethnicity. Pharmacogenetics 11: 217–221 311. Adams KF Jr, Patterson JH, Gattis WA et al (2005) Relationship of serum digoxin concentration to mortality and morbidity in women in the digitalis investigation group trial: a retrospective analysis. J Am Coll Cardiol 46:497–504

312. Ahmed S, Zhou Z, Zhou J et al (2018) Pharmacogenomics of Drug metabolizing enzymes and transporters: relevance to precision medicine. Genomics Proteomics Bioinformatics 14:298–313 313. Yamaguchi H, Sugie M, Okada M et al (2010) Transport of estrone 3-sulfate mediated by organic anion transporter OATP4C1: estrone 3-sulfate binds to the different recognition site for digoxin in OATP4C1. Drug Metab Pharmacokinet 25:314–317 314. Scotcher D, Jones CR, Galetin A et al (2017) Delineating the role of various factors in renal disposition of digoxin through application of physiologically based kidney model to renal impairment populations. J Pharmacol Exp Ther 360:484–495 315. Hagenbuch B, Adler ID, Schmid TE (2000) Molecular cloning and functional characterization of the mouse organic-anion-transporting polypeptide 1 (Oatp1) and mapping of the gene to chromosome X. Biochem J 345: 115–120 316. Seward DJ, Koh AS, Boyer JL et al (2003) Functional complementation between a novel mammalian polygenic transport complex and an evolutionarily ancient organic solute transporter, OSTalpha-OSTbeta. J Biol Chem 278:27473–27482 317. Ballatori N, Li N, Fang F et al (2009) OST alpha-OST beta: a key membrane transporter of bile acids and conjugated steroids. Front Biosci (Landmark Ed) 14:2829–2844 318. Malinen MM, Ali I, Bezencon J et al (2018) Organic solute transporter OSTalpha/beta is overexpressed in nonalcoholic steatohepatitis and modulated by drugs associated with liver injury. Am J Physiol Gastrointest Liver Physiol 314:G597–G609 319. van Montfoort JE, Schmid TE, Adler ID et al (2002) Functional characterization of the mouse organic-anion-transporting polypeptide 2. Biochim Biophys Acta 1564:183–188 320. Taub ME, Mease K, Sane RS et al (2011) Digoxin is not a substrate for organic aniontransporting polypeptide transporters OATP1A2, OATP1B1, OATP1B3, and OATP2B1 but is a substrate for a sodiumdependent transporter expressed in HEK293 cells. Drug Metab Dispos 39 (11):2093–2102. https://doi.org/10.1124/ dmd.111.040816

Chapter 10 Pharmacogenomic Screening of Drug Candidates using Patient-Specific hiPSC-Derived Cardiomyocyte High-Throughput Calcium Imaging Malorie Blancard, K. Ashley Fetterman, and Paul W. Burridge Abstract Calcium imaging is an invaluable technique to detect and characterize calcium flux in cells. The use of calcium dye provides information on the concentration and spatial distribution of calcium. Calcium imaging is a well-established technique to assess the calcium-induced calcium release mechanism in cardiomyocytes. It can also be used to characterize mutations in genes crucial for this mechanism that frequently causes arrhythmia. Here we describe a high-throughput methodology of calcium imaging that records individual calcium transients in more than 10,000 human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) in less than 30 min. Key words hiPSC-derived cardiomyocytes, Drug screening, Calcium imaging, Arrhythmia, Highthroughput

1

Introduction Calcium is a ubiquitous intracellular messenger that passes through ion channels, pumps, and exchangers to create a calcium gradient. This gradient is crucial to cellular calcium signaling. Calcium signaling occurs in excitable cells such as cardiomyocytes as well as non-excitable cells such as endothelial cells and is involved in many pathways contributing to motility, proliferation, differentiation, and apoptosis [1]. Calcium signaling can be observed as an intracellular calcium transient which is described as a “wave” or by a localized increased concentration known as a spark. Excitation-contraction coupling (ECC) is the major mechanism that converts electrical excitation to cardiomyocyte contraction [2]. Calcium enters the cell through L-type calcium channels, increasing the intracellular concentration of calcium locally at the submembrane. This leads to ryanodine receptor activation on the sarcoplasmic reticulum (SR) and results in massive calcium release

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_10, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

241

242

Malorie Blancard et al.

from the SR to the cytosol. This mechanism is called calciuminduced calcium release (CICR) and can be observed in cells as a calcium transient [3]. Localized release of calcium through the SR can be observed as a spark. The increase in total intracellular calcium signals the cell to contract, relax, and cycle calcium out of the cytosol to both the extracellular space and the SR. Capturing this mechanism through calcium imaging has generated significant progress in understanding the contribution of calcium homeostasis dysfunction in arrhythmias. Calcium imaging remains an essential tool to study clinical phenotypes involving calcium dysfunction. Calcium signaling was first observed in living cells, specifically muscle fibers, in 1967 [4] and CICR was first imaged in 1980 [5]. The development of confocal microscopy has led to significant progress in calcium imaging, including the first recording of calcium sparks in cardiomyocytes in 1993 [6]. These analyses are performed by recording signals from one cell at a time, which is low throughput and not compatible with drug screens that require a high number of cell recordings and various drugs and concentrations. Recent high-throughput systems such as the FLIPR assay have been used to facilitate drug screenings but are only able to capture the whole cardiac flux generated by thousands of cells in an individual well. Here we describe a protocol that supports highthroughput experiments and provides single cell recording and analysis. Human-induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) are a powerful in vitro tool to study patientspecific clinical phenotypes and utilize in drug screening. Here we describe a high-throughput method for calcium imaging at the single cell level using patient-specific hiPSC-CMs to understand disease mechanisms and identify new treatments for patients.

2 2.1

Materials Cell Preparation

1. 96- or 384-well black cell culture microplates, polystyrene, F-bottom, μClear (Greiner). 2. Matrigel: Growth factor-reduced Matrigel diluted 1:800 in Dulbecco’s Modified Eagle Media (DMEM, Corning). 3. RBAI media for hiPSC-CMs: RPMI 1640 (Corning), 500 μg/ ml fatty acid-free albumin (GenDEPOT), 200 μg/ml L-ascorbic acid 2-phosphate (Wako), 5 μg/mL E. coli-derived recombinant human insulin (Gibco). Supplemented with 10% Cosmic Calf Serum for cell plating. 4. hiPSC-CMs: Generated using typical protocols, on d16 of differentiation, dissociated.

2.2 Recording Ca2+ Transients

1. Dye loading solution: Hanks’ Balanced Salt Solution (HBSS, Corning), 20 mM HEPES (1 M stock, Corning), 0.04%

High-Throughput Drug Screening using Calcium Imaging

243

Pluronic F-127 (5% stock in water, Sigma), 2.5 mM probenecid (500 mM stock, water-soluble, AAT Bioquest), 2 drops/ 10 mL NucBlue Live ReadyProbes Reagent (Invitrogen), 1 μM Cal-520 AM (1 mM stock in DMSO, AAT Bioquest). 2. Imaging solution: FluoroBrite DMEM (Gibco), 2.5 mM probenecid. 3. Vala Sciences IC200-Kinetic Imager Cytometer (KIC). 4. CyteSeer Scan software. 2.3

Analysis

1. CyteSeer Analysis software. 2. Microsoft Excel. 3. GraphPad Prism.

3 3.1

Methods Cell Preparation

1. Coat a 96- or 384-well black cell culture plate with diluted Matrigel (100 μL or 25 μL, respectively, per well). 2. Seed dissociated d16 hiPSC-CMs into a 96-well black cell culture plate (7.5  104 cells per well) or 384-well black cell culture plate (2.5  104 cells per well) in RBAI media supplemented with 10% CCS for 48 h. Keep cells at 37  C, 5% CO2. 3. Change media with RBAI without CCS every other day until drug addition. Add drug to hiPSC-CMs at the pre-determined dose and timepoint. 4. On the day of the experiment (typically d30), take 10 mL of HBSS supplemented with HEPES and add 80 μL of 5% Pluronic F-127, two drops of NucBlue Live, 50 μL of 500 mM probenecid, and 10 μL of 1 mM Cal-520 AM. Keep protected from light. Aspirate media and add 100 μL of the dye loading solution per well of a 96-well or 25 μL per well of a 384-well and incubate 1 h at 37  C, 5% CO2. After an hour of incubation, aspirate media and add 100 μL of the imaging solution to each well of a 96 well (see Notes 1 and 2).

3.2

Scan

1. Open CyteSeer Scan software, click “Eject” on the main screen, and firmly load the plate with lid into the holder. Press OK to close the drawer (see Note 3). 2. Click “Setup Scan” on the main menu to create a new configuration file (skip to step 3 if a configuration file has already been created). (a) File and Well Plate Settings. Create a name for the protocol to use for future experiments by entering a name in the “Configuration Name” field. Click “Eject Plate” if the plate has not yet been loaded. Use the “Well Plate Type” drop-down menu

244

Malorie Blancard et al.

to select a 96- or 384-well plate layout (see Note 4). Once the well plate type is loaded, select the wells to be scanned on the plate map and select the field of view(s) within the intrawell arrangement map (see Note 5). The current objective should be 20. This field will default to the objective that is currently installed in the machine. This cannot be edited once the protocol is saved so a separate protocol needs to be created to run the same protocol with a different objective (see Note 6). (b) Focus Settings. For the focus method select “Image Contrast and Surface Tracking.” For the fluorochrome, select “DAPI” (or appropriate channel for the nuclear stain). The automatic focus works better with the nuclear staining rather than the calcium flux (see Note 7). To begin, use 100 ms time exposure and 50% intensity and adjust as needed (see Note 8). The “Focus Position” section is useful for manually focusing and exploring wells. The Z step size can be changed from a more fine or coarse focus by adjusting between 100 μm, 10 μm, 1 μm, and 0.1 μm steps. The “Find Surface” button will locate the cells making it easier to manually focus. The focus position scale can be moved from 0 to 20,000. Be careful when scrolling manually up and down because the objective can knock the plate out of the holder. Use both the plate map and well plate position to move around plate and each well to determine which areas to image. Click “Advanced Settings.” Set focus stack settings to 20.0 range (μm), 1.0 step size (μm), lower cutoff of 100, and upper cutoff of 300. Adjust these settings accordingly depending on the autofocus range needed and the type of plate. Click “Image Trail Focus” to adjust different z positions and z step sizes as well as lower and upper cutoffs to see what works best for the focus of cells in a specific well. Once these parameters are established, click “Full Image Trial” to set the autofocus for the run. Once the configuration file is established, perform this step last when setting up a scan. Once the full image trial is complete, click “Done” to exit out of this screen and return to the focus setting screen (see Note 9). Click the light bulb icon to check the focus of the nuclear stain and switch to the FITC channel to ensure the calcium staining is also focused. (c) Channel Acquisition Settings. Establish the channel acquisition settings by using the “+” symbol in top right corner to add channels or the “x” symbol to delete channels. Two channels will need to be established with the following settings as a starting point and adjusted as needed:

High-Throughput Drug Screening using Calcium Imaging

245

Nuclear Channel – Channel Name: Nuclear Channel. Acquisition Type: Single Image. Fluorochrome: DAPI. Exposure: 100 ms. Intensity: 50%. Offset: 0 μm Calcium Channel – Channel Name: Calcium Channel. Acquisition Type: Kinetic Series. Fluorochrome: FITC. Exposure: 10 ms (this provides the fastest capture rate possible for this machine). Intensity: 100%. Offset: 0.1 μm. Within the Advanced Settings: KIC Acquisition Length: 10 s. Frame Rate: 99.75 (d) Scan Summary. Click the paper and pen icon to save the configuration, close out of the configuration setup, and return to the main menu. Once the scan configuration is saved, this template can be used for future experiments instead of repeating this process each time. 3. Click “Scan Editor” on the main screen (use this once a scan configuration is established). 4. Go to “File,” then “Load Scan Configuration,” and select the desired scan configuration. Navigate through the protocol by clicking on the sections in the menu on the left. (a) Well Plate: Select the well plate type as well as the wells and areas within the wells to be scanned. The objective set for this scan configuration cannot be changed (see Note 10). (b) Focus: Perform a full trial focus using the nuclear channel before each plate is scanned. (c) Acquisition: Keep exposure and intensity settings the same between different scans to be able to compare experiments more accurately. (d) At the end, click “Save and Scan Time” to go to the scan window. 5. Select scan properties window. (a) Scan Configuration File: “File Path” and “Configuration Name” will be the file path and configuration loaded through the scan editor. (b) Data Destination: Select where to save the data collected during the scan (see Notes 11 and 12). (c) Check the boxes next to “File Set” and “Scan Selection Set” to make the “Scan Now” button active and click it. (d) The scan will begin. The images can be viewed as they are acquired in real time. It is important to ensure that both channels are in focus. Otherwise, exit out of the scan, try refocusing in the scan editor, and start the scan again. If this does not fix the focus problem, make sure the plate is sitting properly in the holder.

246

3.3

Malorie Blancard et al.

Analysis

Open the CyteSeer Analysis software. Configure the “Source Folder” to the scan file to be analyzed. If this is the first analysis since opening the software and of this experiment, the “Result” folder will be set to “” by default. Select IC200 (normal) for the “Image Naming Convention” and then select “Plate.” Use the algorithm called “Cardiac Calcium Time Series.” Below are the different parameters to configure when setting up an analysis to run. 1. Defining Masking Parameters: In the “Assign Channels” column, check the boxes next to the “Nuclear Channel” and the “Calcium Channel.” The “Nuclear Channel” corresponds to the DAPI filter used to detect the nuclear stain and the “Calcium Channel” corresponds to the FITC filter used to detect the intracellular calcium transients. To adjust the masking of the cells, modify the “Sensitivity” and “Size Value.” The sensitivity may vary from 25% to 400% and the size value from 3 to 30 (see Note 13). Go to the “Setup Run” tab next. All the wells scanned are listed here. To determine the masking parameters, select only one or two wells. Click “Run” and then “OK” on the following “Configure Algorithm” window (see Note 14). After the run is complete, the software will generate a “Results Folder” that contains the masking results. To check the quality of the masking, go to “View” in the menu and select “Image Viewer.” Select a well to examine from the dropdown menu. Open the “Image Viewer Set Colors” and select “Cell Mask” and “Cytoplasm” in the “Edge” column as well as the color for each mask. Zoom in or out on the image to better view the masking for each channel. 2. Running Analysis: Once the appropriate mask for each channel is determined, the calcium transient analysis can be run. On the “Assign Channels” tab, enter the chosen masking parameters. Then, go to the “Setup Run” tab and select the wells to be analyzed. Click “Run.” The same “Configure Algorithm” window will open. Check the box “Make Measurements” to obtain the complete data rather than just the mask information. Check or uncheck the “Reanalyze Data Tables” box. In both cases, the program will reanalyze. For the “Starting Time Point,” drag the cursor to 0, and for the “Ending Time Point,” drag the cursor to the last point. Calculate the “msec Per Frame” value. This value is directly linked to the parameters set when the plate was scanned so this value must be calculated for each individual run. The default value is 33.333 and the software does not automatically calculate it (see Note 15). If there is significant photobleaching, check the box “Subtract Exponential Baseline.” With this setting, the software will correct for photobleaching (see Note 16). The default value for the “Smoothing Radius” is 2 frames. The default value for “Peak

High-Throughput Drug Screening using Calcium Imaging

247

Noise Threshold” is 0.1 (see Note 17). Check the box “Detect EADs” to include this parameter in the results. Check “Accept Only Full Peaks” (see Note 18). The default value for the “Max Baseline Asymmetry” is 0.1 (see Note 19). Always check “Trim Peaks Before Upswing.” Do not check “Align the Peaks Before Averaging.” 3. Exporting Results: Once the analysis is complete, the data need to be exported. If data are exported after closing the analysis software, be sure the “Source” and “Result” folders are set to the file path of the analysis to be exported. (a) Individual Cells in a Well On the main screen, go to the “View” menu and select “Data.” This menu will have a “Well Data” tab where the individual cell data per well can be viewed. Select a well to display the time series of individual cell transients and scan the results for quality. Display Time Series (Traces): To view the whole trace of the recording, select “Trimmed Cell on Cell Mask Average Pixel Intensity Function of Time” within the measurement list. This option corresponds to the normalized data. Every datapoint is divided by the minimum of the fluorescence (Fmin) and then subtracted by 1 (formula: F Fmin 1). To view the first peak of each cell, select “Pk#1 Cell on Cytoplasm Average Pixel Intensity Function of Time.” Click “Export” to export CSV file of these traces and plot in a program of choice. Display Table (Values): Click on the “All Measurement” drop-down menu and select “Means.” Click the “Uncheck All” button. Click all measurements (25 total) that begin with “Cell on Cytoplasm Average Pixel Intensity” and “Mean of Pk #1 Cell on Cytoplasm Average Pixel Intensity” (see Note 20). Click “Update.” Click “Export” to export a CSV file. (b) Well Average The “Data” menu also has a “Well Plate Data” tab which provides average well data. Display Time Series (Traces): To view the whole trace of the recording, select “Average of Trimmed Cell on Cell Mask Average Pixel Intensity Function of Time” within the measurement list (see Note 21). In the “Measurement” list, select “Average of Pk#1 Cell on Cytoplasm Average Pixel Intensity Function of Time.” This will display the average of the first peak of each well (see Note 22). Click “Export” to export CSV file of these traces and plot in a program of choice (see Fig. 1).

248

Malorie Blancard et al.

Fig. 1 Intracellular Ca2+ transients on Vala KIC. (a) Well average calcium transient traces of control hiPSC-CMs +/ drug treatment. Cells were treated for 72 h with media (control) or verapamil (10 μM), norverapamil (30 μM), gabapentin (30 μM), flecainide (30 μM), and (+)-cis-diltiazem (30 μM). (b) Whole trace of recording of 10 s of control hiPSC-CMs treated or not with isoproterenol (1 μM)

Display Table (Values): Click on the “All Measurements” drop-down menu and select “Means.” Click the “Uncheck All” button. Click all measurements (25 total) that begin with “Cell on Cytoplasm Average Pixel Intensity” and “Mean of Pk #1 Cell on Cytoplasm Average Pixel Intensity” (see Note 20). Click “Update.” Click “Export” to export a CSV file.

4

Notes 1. Do not forget the nuclear dye. The CyteSeer software algorithm will not be able to detect the cells without the nuclear staining. If the nuclear dye is not present at the time of the scan, simply prepare the imaging solution with two drops of NucBlue (see Subheading 3.1, step 4) and replace the media in the plate with this solution. The nuclei will be stained within 10 min. If the plate has been read, the only option is to analyze the data manually with ImageJ but this is time consuming for a full 96-well plate. 2. The imaging solution can be pre-warmed for 10 min in a water bath before adding it to the plate and/or the plate can be incubated for 10 min after adding the imaging solution in the IC200-KIC incubator at 37  C, 5% CO2, during the scan setup. 3. The built-in IC200-KIC incubator should be turned on at the time of dye loading to allow the incubator to reach the set temperature. The CO2 can be turned on at the time of the scan. 4. The IC200-KIC is compatible with most brand and size of plate. New plate-specific templates can be created by following the machine manual.

High-Throughput Drug Screening using Calcium Imaging

249

5. The wells and location within the wells to be scanned can be changed at a later point in the protocol if needed. 6. This machine is compatible with various objectives including 10, 20, and 40 objectives. To change the objective, go to main screen and click “Setting.” From the settings dialog screen, select “Change Objective” and select the desired objective. Once this objective is manually installed, select “OK.” The desired objective is now ready to use. 7. Other channels include DAPI, FITC, TRITC, Cy5, and Brightfield. 8. Intensity needs to be set to greater than 1 in order to visualize the cells. Otherwise, the screen will remain black until this setting is increased. 9. If this fails, the nuclear dye might not have been added, the wrong channel is selected, or the intensity is set to 0. To check for dye, click “Find Plate Surface” on the focus settings screen and manually look for the nuclear stain. Another method is to try the full image trial using the calcium dye channel but this may fail due to the continuous flux causing the machine to not focus properly. 10. If the configuration requires an objective that is not currently installed in the machine, an error message will notify the users and the objective will need to be changed before running the protocol. 11. Create a new folder and give it a descriptive name. The file name automatically generated is based on the configuration file name and date and there are no fields to enter descriptive information for each run. 12. Based on the numbers of wells and fields selected, the estimated space required will be calculated, and if the file is too large, a warning message will appear saying “unable to start scan: space required exceeds free space.” In order to start the run, the estimated space required must be less than the space available in the location where the file will be stored. For example, to scan 10 s at 10 ms exposure for a 96-well plate, ~1 TB is required. Therefore, it is important to have large amounts of storage available on the computer collecting the scan data. 13. To find an appropriate mask for the “Nuclear Channel,” begin with a size value of 3 and the sensitivity set to 100%. For the nuclear mask, if there are two masks for a single nucleus, the size value of the “Nuclear Channel” is too low. Increase this value until there is a single mask for each nucleus. On the contrary, if there is one mask for two nuclei, the size value of the “Nuclear Channel” is too high. By adjusting the sensitivity

250

Malorie Blancard et al.

Fig. 2 Schema representing how sensitivity and size value affect the masking of the cells

of the “Nuclear Channel,” the size of the mask will change. Therefore, if the current mask does not distinguish two different nuclei, lower the sensitivity value. To find an appropriate mask for the “Calcium Channel,” begin with a size value of 30 and the sensitivity set to 100%. Adjust these parameters as needed, following the above guidelines. For example, if the mask line is too far from the cell border, decrease the sensitivity value (see Fig. 2). 14. It can take several runs to test different masking parameters. Each run can take 5 min or more. To decrease the run time, first, uncheck all the boxes on the “Configure Algorithm” window. Second, choose 0 as a starting time point and 5–10 as the ending time point since this will be sufficient to determine the “Nuclear Channel” and “Calcium Channel” mask. 15. Failure to calculate this value for each scan will result in data that contains incorrect time course information and will require the analysis to be run again with the correct value. To calculate this value, divide the duration of the recording (e.g., 10,000 ms) by the number of frames (e.g., 249 frames). In this example, the value would be 40.16 ms so 40.16 would be entered into the field instead of 33.3333. There are two methods to determine the number of frames:

High-Throughput Drug Screening using Calcium Imaging

251

Fig. 3 Example of overcorrection. Whole trace of recording before and after correcting by selecting “Subtract Exponential Baseline” when no photobleaching was present during the experiment

(a) In the “Configure Algorithm,” move the cursor to the end in the time point section of the window. The number between the brackets after “Ending Time Point” is the number of frames for this run. (b) Go to the results folder stored on the computer, open the “Scan” folder, and within that folder, open any well folder. Then, open the folder named “Calcium.” The number of pictures in this folder is the number of frames. 16. If “Subtract Exponential Baseline” is selected but there is no photobleaching, the fluorescence at t0 will be greater than 0 (see Fig. 3). If this occurs, re-analyze without checking this box. 17. If the detection of the number of peaks is not accurate, the “Peak Noise Threshold” may need to be changed. 18. If the recording started when the cells were in the middle of calcium transient, the first peak will be incomplete. If the “Accept Only Full Peaks” is checked, the software will not count the peak selected in red (see Fig. 4). 19. When the value is 0.1, the software will accept a 10% difference between the baseline before the peak and after the peak. If the difference is higher than 10%, the software might count the two peaks as one or consider the next peak as an EAD. If the value of the “Max Baseline Asymmetry” is 0.1 and the blue arrow represents more than 10% of the red arrow, the software will count the last two peaks as one peak as if the third peak was an EAD after the second peak (see Fig. 5). 20. Here is a list of measurements we routinely use for each experiment: Max Downstroke Velocity, Max Upstroke Velocity, Peak Value, CTD10, CTD25, CTD40, CTD50, CTD70, CTD75, and Time from 75% to 25%. Relative velocity sigma provides information about noise.

252

Malorie Blancard et al.

Fig. 4 Whole trace of recording where the first and last peak are incomplete

Fig. 5 Baseline asymmetry detection

21. The average whole trace displays the average recordings of all the cells in a well over the full duration of the recording. If the cells are synchronously fluxing calcium, the individual trace for each cell will align to one trace. If the cells are not synchronously fluxing calcium, the individual trace for each cell will not align and appear disorganized. This will not correspond to any relevant calcium information. 22. It is important to check that the parameter selected in the “Configure Algorithm” window for the run analysis is showing the correct results. The detection of the first peak can be incorrect because the first two peaks can be detected as one (see Note 17). In that case, screen each well by eye and verify that there is only one peak. If the first peak has a small second

High-Throughput Drug Screening using Calcium Imaging

253

bump attached to it, it is possible that the software correctly detected the first peak for some cells but incorrectly detected two peaks instead of one for other cells. The average will display an “abnormal” peak with a second bump that can be incorrectly interpreted as a phenotype. To ensure this is not the case, look through the traces of the first peak at the individual cell level to confirm that there is only one peak. If the software has detected two or more peaks instead of one, re-run the analysis with new parameters.

Acknowledgement This work was supported by the National Institutes of Health (NIH) Distinguished Scholars Program (DSP) to LMR and the Division of Intramural Research (DIR) of the National Institute on Minority Health and Health Disparities (NIMHD) at NIH (1ZIAMD000016 and 1ZIAMD000018). This work was supported by the FAPA project (PVI0122029) from Universidad de Los Andes, granted to A.V.A. References 1. Knot HJ, Laher I, Sobie EA et al (2005) Twenty years of calcium imaging: cell physiology to dye for. Mol Interv 5:112–127 2. Bers DM (2002) Cardiac excitation-contraction coupling. Nature 415:198–205 3. Landstrom AP, Dobrev D, Wehrens XHT (2017) Calcium signaling and cardiac arrhythmias. Circ Res 120:1969–1993 4. Ridgway EB, Ashley CC (1967) Calcium transients in single muscle fibers. Biochem Biophys Res Commun 29:229–234

5. Wier WG (1980) Calcium transients during excitation-contraction coupling in mammalian heart: aequorin signals of canine purkinje fibers. Science 207:1085–1087 6. Cheng H, Lederer WJ, Cannell MB (1993) Calcium sparks: elementary events underlying excitation-contraction coupling in heart muscle. Science 262:740–744

Chapter 11 The Yin-Yang Dynamics in Cardiovascular Pharmacogenomics and Personalized Medicine Qing Yan Abstract Studies of genetic variants and systems biology have indicated that Yin-Yang dynamics are especially meaningful for cardiovascular pharmacogenomics and personalized therapeutic strategies. The comprehensive concepts of Yin-Yang can be used to characterize the dynamical factors in the adaptive microenvironments of the complex cardiovascular systems. The Yin-Yang imbalances in the complex adaptive systems (CAS) at different levels and stages are essential for cardiovascular diseases (CVDs), including atherosclerosis, hypertension, and heart failure (HF). At the molecular and cellular levels, Yin-Yang interconnections have been considered critical for genetic variants and various pathways, mitophagy, cell death, and cholesterol homeostasis. The significance of the adaptive and spatiotemporal factors in the nonlinear Yin-Yang interactions has been identified in different pathophysiological processes such as fibrosis. The Yin-Yang dynamical balances between proinflammatory and anti-inflammatory cytokines have vital roles in the complex reactions to stress and impairments to the heart. Procoagulant and anticoagulant lipids and lipoproteins in plasma have the Yin-Yang roles that increase or decrease thrombin productions and thrombosis. At the systems level, the Yin-Yang type of relationships has been suggested between atrial fibrillation (AF), diastolic dysfunction (DD), and HF. Based on such perceptions, systemic and personalized cardiovascular profiles can be constructed by embracing the features of CAS, especially the microenvironments and the adaptative pathophysiological stages. These features can be integrated into the comprehensive Yin-Yang dynamics framework to identify more accurate biomarkers for better prevention and treatments. The goal of reestablishing ubiquitous Yin-Yang dynamical balances may become the central theme for personalized and systems medicine for cardiovascular diseases. Key words Arrhythmia, Atherosclerosis, Cardiovascular, Complex adaptive systems (CAS), Heart failure, Hypertension, Pharmacogenomics, Stroke, Systems biology, Yin-Yang

1

Introduction: Yin-Yang Dynamics in Complex Cardiovascular Diseases Studies of genetic variants and systems biology have indicated that Yin-Yang dynamics are especially meaningful for cardiovascular pharmacogenomics and personalized therapeutic strategies. The comprehensive concepts of Yin-Yang can be used to characterize

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_11, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

255

256

Qing Yan

the dynamical interconnecting, counteracting, interdependent, and transforming factors in the adaptive microenvironments of complex cardiovascular systems. Such features of complex adaptive systems (CAS) are essential in various pathophysiological conditions and cardiovascular diseases (CVDs) [1, 2] (see Chap. 6). These conditions include vascular remodeling in restenosis, atherosclerosis, and cardiac hypertrophy related to heart failure [3, 4]. The ubiquitous YinYang dynamics in various networks at different systems levels and stages are fundamental for cardiovascular health and CVDs [1, 2]. For example, among the subsets of microRNAs (miRNAs) associated with vascular remodeling, opposite Yin-Yang functions have been recognized [3]. The two subsets miR-143/145 and miR-126 may have protective effects. However, miR-21 has been found to trigger the cellular reactions resulting in neointima formation [3]. Table 1 lists examples of the factors in the Yin-Yang interactions in various cardiovascular conditions. Similar to cancers (see Chap. 6), microenvironments and adaptation often have decisive roles in the Yin-Yang dynamics of many cardiovascular pathophysiological processes. For instance, the spatiotemporal microenvironment may decide the positions of the platelet-derived growth factor receptor A (PDGFRα+) mesenchymal progenitor cells in fibrosis in the Yin-Yang manner (see Table 1). The concentrations and temporal factors also determine the Yin-Yang roles of TNFα in heart failure (see Table 1). In another example, the microenvironment, including the conditions of superoxide, may decide the vasorelaxant or vasoconstrictive functions of estrogen (see Table 1). The following sections show that Yin-Yang dynamical balances have been recognized as critical in various CVDs during different stages, including atherosclerosis, hypertension, and heart failure. Such understanding would contribute to discovering and developing more effective and personalized therapeutics for CVDs.

2

Yin-Yang Dynamics in Atherosclerosis and Cardiovascular Pharmacogenomics Genetic variants’ investigations have demonstrated that Yin-Yang dynamics are critical for cardiovascular pharmacogenomics. In atherosclerosis, the allograft inflammatory factor-1 (AIF-1) and interferon responsive transcript-1 (IRT-1) proteins may have the YinYang type allocation and effects on vascular smooth muscle cell (VSMC) [5]. The expressions of the two variants are controlled by different nuclear factor of activated T-cells (NFAT) actions. The higher levels of AIF-1 may trigger the relocation and production of

The Yin-Yang Dynamics in Cardiovascular Pharmacogenomics and Personalized. . .

257

Table 1 Yin-Yang dynamical interactions in cardiovascular diseases Conditions

Factors

Yin-Yang interactions

References

Atherosclerosis

AIF-1 and IRT-1

Worsen or alleviate neointima

[5]

Atherosclerosis

LOX-1, LOXIN

LOXIN may inhibit LOX-1

[6]

Cardiovascular disorders

CETP

Proatherogenic vs. athero-protective

[7]

Cardiovascular disorders

PCSK9 and LDLR

Constant intracellular cholesterol vs. cardiovascular disorders

[8]

Metabolic, lipidrelated disorders

LXRS and FXR

Balanced control of body fat and cholesterol homeostasis

[9]

CAD, HF, atherosclerosis

Nitric oxide (NO)

The regulation of cardiovascular homeostasis

[10]

Chronic heart failure (HF)

TNFα

Short-term helpful vs. long-term harmful (cardiac ischemia)

[11]

Ischemic heart disorders

STAT1 (ST1) and STAT3 (ST3)

Cell death vs. cytoprotective signals

[12]

Fibrotic disease

CCN2, CCN3, and CCN5

Promotion vs. inhibition of fibrogenesis

[13, 14]

Injury, fibrosis

Tissue-resident mesenchymal progenitor cells

The inhibition of injury vs. the promotion of fibrosis

[15]

Ischemia, reperfusion

Mitochondria and mitophagy

Pro-survival vs. pro-death pathways

[16]

PPCM

Inflammatory biomarker CRP

Proinflammatory vs. antiinflammatory cytokines

[17]

Angiogenic sprouting

VEGF and Notch signaling

The regulation of vascular patterning

[18]

Cardiac functions

MuRF1/TRIM63

Damaging in skeletal muscles vs. cardioprotective in the heart

[19]

AF, HF

AF, HF, and DD

Interactions in AF, HF, and DD

[20]

Myocardial infarction

Ischemic, bleeding

Ischemic vs. bleeding complications

[21]

Thrombosis

Vascular thiol isomerases, PDI

Positive vs. negative control of vascular [22] homeostasis

Thrombosis

Plasma lipoproteins, APC

Procoagulant vs. anticoagulant networks

[23]

Clot and inflammation

Thrombin and APC

Coagulation vs. anticoagulant processes

[24] (continued)

258

Qing Yan

Table 1 (continued) Conditions

Factors

Yin-Yang interactions

References

Acute coronary syndromes

Antithrombotic treatments and PCI

Coronary thrombosis vs. bleeding complications

[25]

Hypertension

Estrogen

Vasorelaxant vs. vasoconstrictive effects

[26]

Vascular disorders

ET-1

Vasodilators vs. vasoconstrictors

[27]

Opposite effects on voltage-gated potassium (Kv) channels

[28]

Cardiac arrhythmias KCNE1, KCNE3 High blood pressure, stroke

Potassium and sodium

Cardiovascular impairments vs. vasculo-protective

[29]

Stroke

GABA, glutamate networks Early phases of stroke vs. later phases of recovery

[30]

human VSMC. However, IRT-1 may have the opposite functions [5] (see Table 1). The drugs and small interfering RNA (siRNA) targeting and suppressing NFAT may reduce the AIF-1/IRT-1 ratio and promote the anti-proliferative functions [5]. NFAT may be an inhibitor of the IRT-1 transcription and can be affected by interferon-γ. Activities of AIF-1 in carotid plaques have been related to the reduced extracellular matrix, proinflammatory features, and the possibility of plaque rupture [5]. However, IRT-1 has been related to less destructive results with reduced VSMCs at the plaque [5]. In addition, AIF-1 has been suggested to worsen intima hyperplasia, while IRT-1 may alleviate neointima. Upon such understanding, the suppression of the NFAT network through changing the AIF-1/IRT-1 ratio has been suggested to control the VSMC reactions and plaque solidity for the treatment of atherosclerosis [5]. In addition, oxidized low-density lipoproteins (OxLDL) may have essential roles in proatherogenic activities and atherosclerosis [6]. The functions of OxLDL have been associated with the lectinlike oxidized low-density lipoprotein receptor 1 (LOX-1), encoded by the OLR1 gene. LOX-1 usually refers to the full receptor. LOXIN is an isoform that is absent of some functional domains [6]. Variations of the OLR1 gene have been linked to different ratios of LOX-1/LOXIN. LOXIN may inhibit the negative functions of LOX-1, showing the Yin and Yang regulation of atherosclerosis [6]. Such elucidation is especially helpful for cardiovascular pharmacogenomics (see Table 1). Moreover, the functions of cholesteryl ester transfer protein (CETP) may lead to pro-atherogenic or athero-protective results in

The Yin-Yang Dynamics in Cardiovascular Pharmacogenomics and Personalized. . .

259

the Yin-Yang manner in cardiovascular disorders [7] (see Table 1). CETP may be involved in the transference of cholesteryl ester (CE) from high-density lipoprotein (HDL) elements to low-density lipoprotein (LDL)-very-low-density lipoprotein (VLDL) elements. It may also be involved in the transportation of cholesteryl ester from VLDL to LDL elements [7]. The Yin-Yang dynamical balance between proprotein convertase subtilisin/kexin type 9 (PCSK9) and low-density lipoprotein receptor (LDLR) controls the entrance of cholesterol into the cell [8]. Such a mechanism is essential for the maintenance of the constant intracellular cholesterol. Mutations with gain of function may reduce LDLR with higher levels of low-density lipoprotein cholesterol (LDL-C) and elevated risks of cardiovascular disorders [8]. On the other hand, mutations with loss of function may lead to higher levels of LDLR, reduced LDL-C, and lower risks of cardiovascular disorders [8]. The loss of function mutations of PCSK9 with lower levels of PCSK9 may be beneficial that has been suggested as potential treatments. Liver X receptors (LXRs) are the sensors for sterols. Farnesoid X receptors (FXR) are the sensors for bile acids. The interactions between these receptors and their ligands may help maintain the Yin-Yang balanced control of cholesterol and bile acid [9]. LXRs are involved in the accumulation of carbohydrate- and fat-originated energy. FXR has been related to reduced triglyceride and regulation of glucose metabolism [9]. Such Yin-Yang interconnections between these receptors have been essential for regulating body fat and cholesterol homeostasis. These interactions are crucial in the cardiovascular, immune, endocrine, renal, and central nervous systems [9]. This is why LXRs and FXR have been deemed candidates for the therapeutics of many metabolic and lipid-associated disorders (see Table 1).

3

Yin-Yang Dynamics at Various Systems Levels and Stages in Heart Failure The features of Yin-Yang dynamical balances have been found at various systems levels and stages in conditions associated with heart failure. For example, nitric oxide (NO) has the Yin-Yang function in maintaining cardiovascular homeostasis. It is critical for every stage of coronary artery disease (CAD) [10]. Vascular NO has an essential role in regulating coronary blood flow via suppressing smooth muscle tightening and platelet accumulation. It is important in angiogenesis and cytoprotection [10]. Impaired endothelial nitric oxide synthase (eNOS) has been associated with cardiovascular risks, hypertension, and insulin resistance (see Table 1). However, activities and excess inducible nitric oxide synthase (iNOS)

260

Qing Yan

may lead to metabolic insulin resistance, atherosclerosis, and heart failure [10]. Tumor necrosis factor α (TNFα) is proinflammatory with an essential role in chronic heart failure (CHF) development (see Table 1). However, TNFα antagonists in clinical trials have shown unsatisfactory effects. It has been suggested that the concentration and time-associated Yin-Yang modes of TNFα may contribute to such results [11]. In a short period, TNFα may be helpful [11]. However, in the long term, it may have damaging effects associated with glutathione shortages in cardiac ischemia and reperfusion. Glutathione precursors such as N-acetylcysteine (NAC) have been suggested to suppress the detrimental TNFα networks to provide possible treatments for CHF [11]. This example demonstrates the significance of the adaptive and temporal factors in the nonlinear YinYang transformation roles of TNFα. In addition, signal transducer and activator of transcription 1 (STAT1, ST1) and STAT3 (ST3) have the Yin-Yang styles in ischemic heart disorders, especially in the acute stage [12]. ST1 may be involved in cell death signals. ST3 has been associated with cytoprotective pathways. STAT1 may promote apoptosis and decrease cardioprotective autophagy, resulting in the damage of the non-replaceable cardiac myocytes [12]. On the other hand, the cardioprotective STAT3 has the preconditioning functions that provide short-term ischemia protection [12]. Insufficient STAT3 may result in more severe infarcts after ischemic injury (see Table 1). However, lengthy stimulation of STAT3 has been associated with post-infarction and heart failure [12]. Here, the temporal and adaptive factors have also shown their importance. Like Yin and Yang, different CCN (connective tissue growth factor) members may promote or inhibit each other’s functions, decided by the microenvironment [13]. A member of the CCN family, CCN2 has been associated with fibrogenesis and fibrotic disease via the interactions with cytokines, including transforming growth factor β (TGFβ). Another member, CCN5 (wisp2), has been identified as a dominant-negative molecule that may inhibit CCN2-associated fibrogenesis [13]. Unlike other CCN proteins, CCN5 is absent of the carboxy-terminal heparin-binding domain. In addition, CCN3 may also block the fibrogenesis functions of CCN2 [14]. The elucidation of such complex Yin-Yang interactions may help design more effective anti-fibrotic treatments (see Table 1). Depending on the spatiotemporal microenvironment, the tissue-resident PDGFRα+ mesenchymal progenitor cells have shown the Yin-Yang patterns in fibrosis and healing processes [15]. Although these cells have the innate functions to control

The Yin-Yang Dynamics in Cardiovascular Pharmacogenomics and Personalized. . .

261

injury enlargement, they may also lead to fibrosis when the environment is unfavorable (see Table 1). At the cellular level, mitochondrial autophagy is also called mitophagy. Mitophagy may lead to the removal of impaired mitochondria in cells [16]. Mitochondria and mitophagy are essential in stress adaptation and the maintenance of cardiovascular health (see Table 1). Mild stress can cause harm to some mitochondria with the consequences of autophagosomes [16]. More severe stress may lead to more severe mitochondrial impairments with higher reactive oxygen species (ROS) levels and proteins such as cytochrome c related to cell death networks. In stress conditions, including ischemia and reperfusion, the pro-survival and pro-death networks may interact in cardiac myocytes to achieve a dynamical Yin-Yang balance [16]. Therefore, during the initial cardioprotective reactions triggered by stress, mitophagy may promote the adaptative responses and homeostasis via eliminating impaired mitochondria. However, during the later stages, elevated levels of oxidative stress and apoptotic proteases may deactivate mitophagy and turn on the processes of cell death [16]. Such changes may lead to the damage of cardiac myocytes and the incidence of heart failure. The stages are crucial factors in these conditions. As illustrated in these examples, the cellular pathways and networks are essential in cardiovascular health and disorders. In acute peripartum cardiomyopathy (PPCM), the Yin-Yang dynamics between proinflammatory and anti-inflammatory cytokines have vital roles in the complex reactions to stress and impairments to the heart [17]. The elevated inflammatory biomarker C-reactive protein (CRP) levels may occur concurrently with the proinflammatory cytokine tumor necrosis factor (TNF). Studies have indicated that the reestablishment of the balances in the cytokine networks and the scheduling of treatments may be more effective strategies [17] (see Table 1). Moreover, the interactions between vascular endothelial growth factor (VEGF) and Notch networks have been considered the Yin and Yang of angiogenic sprouting [18]. Endothelial tip cells and nearby stalk cells affecting the vascular lumen may decide tubular sprouting in angiogenesis. The functions of VEGF-A may be crucial in the regulations of tip and stalk cell activities [18]. On the other hand, the delta-like (DII)4-Notch network may inhibit the penetrating tip cell activities triggered via VEGF-A (see Table 1). In addition, some endothelial cells promoted by VEGFA may not become tip cells. The dynamical interconnections between VEGF-A and DII4/Notch networks may be essential in the regulations of vascular patterning [18].

262

Qing Yan

The E3 ubiquitin ligase muscle RING-finger protein-1 (MuRF1)/TRIM63 (tripartite motif containing 63) is critical in skeletal muscle atrophy in catabolic conditions [19]. Such function has made MuRF1 a target for therapeutics against muscle wasting related to diseases including cancer and diabetes. However, although harmful in skeletal muscles, MuRF1 is cardioprotective with a positive role in the heart [19]. Such dual functions show the Yin-Yang roles of this protein in different environments (see Table 1). At the systems level, atrial fibrillation (AF) and heart failure (HF) often occur at the same time, with adverse results associated with thromboembolic problems and even death [20]. As a separate prognosticator, diastolic dysfunction (DD) is also often seen together with HF. The Yin-Yang type of interconnections has been suggested among AF, DD, and HF [20]. The presence of DD is a strong predictor of incident AF in patients with HF. Asymptomatic and undiagnosed paroxysmal atrial fibrillation (PAF) is also frequently seen among HF patients with potential clinical effects [20] (see Table 1). In the therapy of myocardial infarction, the dynamical balance is critical between the prevention of ischemic and bleeding complications [21]. The problems of ischemic and bleeding have opposite mechanisms, but the two incidents are interconnected like the Yin and Yang (see Table 1).

4

Dynamical Yin-Yang Balances in Thrombosis and Cardiovascular Homeostasis Many members of the protein disulfide isomerase (PDI) family, including endoplasmic reticulum resident protein 57 (ERp57) and ERp72, promote platelet accrual and coagulation [22]. The thioredoxin-related transmembrane protein 1 (TMX1) also belongs to the PDI family. It inhibits platelet activities and thrombosis, showing the opposite Yin-Yang modes of the PDI members in thrombosis (see Table 1). The thiol isomerases containing the redox networks may influence vascular homeostasis positively and negatively. TMX1 has been suggested critical in the balance of thiol/disulfide-associated responses [22]. Plasma lipoproteins have crucial roles in procoagulant or anticoagulant networks [23]. Procoagulant and anticoagulant lipids and lipoproteins in plasma have the Yin-Yang roles that increase or decrease thrombin productions and thrombosis (see Table 1). Procoagulant lipids and lipoproteins such as oxidized low-density lipoprotein (LDL) may promote prothrombin, factor X, and factor VII [23]. Anticoagulant lipids and lipoproteins that may inhibit factor Va through the activated protein C (APC) include cardiolipin, Gb3 ceramide (CD77), and high-density lipoprotein (HDL). The

The Yin-Yang Dynamics in Cardiovascular Pharmacogenomics and Personalized. . .

263

understanding of such Yin-Yang interactions may be helpful. For example, the therapeutics for hyperlipidemia may not just reduce lipids; they may also have antithrombotic functions [23]. The Yin-Yang connections have been identified between thrombin and activated protein C (APC) [24]. Such features are especially noticeable in the processes of coagulation and inflammation. Functional thrombin (IIa) may stimulate the production of APC and affect the levels of fibrin clot construction and inflammatory reactions [24]. The size of fibrin development has been associated with the thrombin activatable fibrinolysis inhibitor (TAFIa). The inflammatory reactions have been related to the network of endothelial protease-activated receptor-1 (PAR-1) [24]. The coagulation networks and the anticoagulant processes are opposite but interconnected in the dynamical axis, just like Yin and Yang (see Table 1). Antithrombotic methods have been found effective in relieving ischemic complications among those with acute coronary syndromes and those having percutaneous coronary intervention (PCI) [25]. However, the control of coronary thrombosis may often lead to elevated possibilities of bleeding complications at other places. On the other hand, decreasing bleeding complications may often enhance coronary thrombotic and ischemic occurrences [25]. The Yin-Yang dynamics between antithrombotic treatments and PCI methods requires personalized therapeutic strategies (see Table 1).

5

Yin-Yang Dynamics in Hypertension, Arrhythmias, Stroke, and Various Networks Estrogen has the Yin-Yang modes of controlling blood pressure through its vasorelaxant or vasoconstrictive functions [26]. Such patterns have been associated with relatively lower risks of hypertension in premenopausal females and higher risks in postmenopausal females. Patients with the polycystic ovarian syndrome (POS) and reduced estrogen levels have elevated risks of hypertension [26]. Estrogen also has cardioprotective functions against vascular inflammation, including atherosclerosis (see Table 1). The microenvironment decides estrogen’s different vasorelaxant or vasoconstrictive functions, especially the nitric oxide synthase-1 (nNOS) products, nitric oxide, or superoxide [26]. Aging and diabetes have been associated with O2 release and vasoconstriction. On the other hand, cofactors including Hsp90 and L-Arg may promote the reductase function of nNOS and NOS generations, resulting in vasorelaxation [26]. In addition, the vasoconstrictor endothelin-1 (ET-1) and the vasodilator nitric oxide (NO) have Yin-Yang-like effects on vascular functions [27]. ET-1 controls cell proliferation via ETA and ETB receptors. The production of ET-1 can be promoted by stress, thrombin, epinephrine, growth factors, and cytokines [27]. On

264

Qing Yan

the other hand, components including nitric oxide (NO), cyclic guanosine monophosphate (GMP), atrial natriuretic peptide, and prostacyclin may decrease the generation of ET-1. Lower levels of NO and elevated ET-1 productions may lead to malfunction of endothelium and vascular disorders later on [27]. Thus, the YinYang imbalance of these vasodilators and vasoconstrictors agents may result in hemodynamic diseases (see Table 1). The KCNE subunits potassium voltage-gated channel subfamily E member 1 (KCNE1) and KCNE3 are in charge of voltagegated potassium (Kv) channels with the opposite Yin-Yang patterns [28]. The IKs channel complexes from potassium voltage-gated channel subfamily Q member 1 (KCNQ1) and KCNE1 have essential roles in ventricular myocyte repolarization. On the other hand, the functioning KCNQ1-KCNE3 channels have critical parts in the intestine (see Table 1). Genetic variants in KCNE1 and KCNE3 may lead to cardiac arrhythmias via different processes [28]. KCNE1 and KCNE3 may have opposing influences on KCNQ1. Such effects have been suggested helpful for the regulation of channel gating as potential therapeutics in K(+) channel drugs [28]. Moreover, dietary potassium and sodium may have Yin-Yang effects on maintaining blood pressure (BP) [29]. BP has been related to the body’s sodium positively. However, the total body potassium has been negatively associated with BP as it may control BP and benefit cardiovascular health [29]. Higher levels of dietary salt may worsen hypertensive cardiovascular impairments and stroke (see Table 1). On the other hand, more elevated potassium may be vasculo-protective to suppress such damages with the control of vascular sensitivity to catecholamines for better endothelial health [29]. In stroke, Yin-Yang patterns have also been recognized during the brain excitability processes [30]. In the initial stages of stroke, the elevated levels of excitability are destructive. The changes may occur in the later phases of recovery with similar signaling networks that act as helpful for the patient. Such opposite functions and phases are linked through the gamma-aminobutyric acid (GABA) and glutamate complexes [30]. The perception of such Yin-Yang role exchanges in the networks during various stages of stroke development can be beneficial for personalized therapies (see Table 1).

6 Conclusion and Future Studies: Yin-Yang Dynamics in Personalized Medicine for CVDs As shown by the above examples, the imbalances in the Yin-Yang interactions at various systems levels and stages from miRNAs to

The Yin-Yang Dynamics in Cardiovascular Pharmacogenomics and Personalized. . .

265

mitochondria have significant roles in various cardiovascular disorders, from hypertension to heart failure. Understanding of the ubiquitous Yin-Yang interconnections is critical for the development of personalized medicine for CVDs. For instance, the Yin-Yang roles of KCNE1 and KCNE3 may be applied in K(+) channel drugs [28]. The Yin-Yang patterns between antithrombotic therapy and percutaneous coronary intervention (PCI) approaches can be used to develop personalized therapeutics [25]. In addition, the Yin-Yang links in LXRs and FXR are fundamental for the homeostasis in cholesterol levels and the cardiovascular and central nervous systems. These receptors have been suggested to manage various metabolic disorders [9]. Based on such perceptions, systemic and personalized cardiovascular profiles can be constructed by embracing the features of CAS, especially the microenvironments and adaptative pathophysiological stages [2]. These features can be integrated into the comprehensive Yin-Yang dynamics framework to recognize more inclusive and accurate biomarkers for better prevention and treatments (see Chap. 6). The goal of reestablishing ubiquitous Yin-Yang dynamical balances may become the central theme for personalized and systems medicine for cardiovascular diseases [1, 2]. References 1. Yan Q (2014) From pharmacogenomics and systems biology to personalized care: a framework of systems and dynamical medicine. Methods Mol Biol 1175:3–17 2. Yan Q (2018) Stress and systemic inflammation: Yin-Yang dynamics in health and diseases. Methods Mol Biol 1781:3–20 3. Wei Y, Schober A, Weber C (2013) Pathogenic arterial remodeling: the good and bad of microRNAs. Am J Physiol Heart Circ Physiol 304:H1050–H1059 4. Ferreira JCB, Brum PC, Mochly-Rosen D (2011) βIIPKC and εPKC isozymes as potential pharmacological targets in cardiac hypertrophy and heart failure. J Mol Cell Cardiol 51:479–484 5. Berglund LM, Kotova O, Osmark P et al (2012) NFAT regulates the expression of AIF-1 and IRT-1: Yin and Yang splice variants of neointima formation and atherosclerosis. Cardiovasc Res 93:414–423 6. Mango R, Predazzi IM, Romeo F, Novelli G (2011) LOX-1/LOXIN: the yin/yang of atheroscleorosis. Cardiovasc Drugs Ther 25: 489–494

7. Shah PK (2009) The Yin and Yang of cholesteryl ester transfer protein in cardiovascular disease. Circulation 120:2408–2410 8. Morales-Villegas E (2013) PCSK9 and LDLR the Yin-Yang in the cellular uptake of cholesterol. Curr Hypertens Rev 9:310–323 9. Kalaany NY, Mangelsdorf DJ (2006) LXRS and FXR: the yin and yang of cholesterol and fat metabolism. Annu Rev Physiol 68:159–191 10. Cook S (2006) Coronary artery disease, nitric oxide and oxidative stress: the “Yin-Yang” effect – a Chinese concept for a worldwide pandemic. Swiss Med Wkly 136:103–113 11. Adamy C, Le Corvoisier P, Candiani G et al (2005) Tumor necrosis factor alpha and glutathione interplay in chronic heart failure. Arch Mal Coeur Vaiss 98:906–912 12. Knight RA, Scarabelli TM, Stephanou A (2012) STAT transcription in the ischemic heart. JAKSTAT 1:111–117 13. Leask A (2010) Yin and Yang Part Deux: CCN5 inhibits the pro-fibrotic effects of CCN2. J Cell Commun Signal 4:155–156 14. Leask A (2009) Yin and Yang: CCN3 inhibits the pro-fibrotic effects of CCN2. J Cell Commun Signal 3:161–162

266

Qing Yan

15. Santini MP, Malide D, Hoffman G et al (2020) Tissue-resident PDGFRα+ progenitor cells contribute to fibrosis versus healing in a context- and spatiotemporally dependent manner. Cell Rep 30:555–570.e7 16. Kubli DA, Gustafsson ÅB (2012) Mitochondria and mitophagy: the Yin and Yang of cell death control. Circ Res 111:1208–1221 17. Fett JD, Ansari AA (2010) Inflammatory markers and cytokines in peripartum cardiomyopathy: a delicate balance. Expert Opin Ther Targets 14:895–898 18. Hellstro¨m M, Phng L-K, Gerhardt H (2007) VEGF and Notch signaling: the yin and yang of angiogenic sprouting. Cell Adhes Migr 1: 133–136 19. Peris-Moreno D, Taillandier D, Polge C (2020) MuRF1/TRIM63, master regulator of muscle mass. Int J Mol Sci 21:E6663 20. Caldwell JC, Mamas MA (2012) Heart failure, diastolic dysfunction and atrial fibrillation; mechanistic insight of a complex interrelationship. Heart Fail Rev 17:27–33 21. Brar SS (2016) The Yin and Yang of long-term dual antiplatelet therapy. J Am Coll Cardiol 67: 1155–1157 22. Wu Y, Essex DW (2020) Vascular thiol isomerases in thrombosis: the yin and yang. J Thromb Haemost 18:2790–2800

23. Griffin JH, Ferna´ndez JA, Deguchi H (2001) Plasma lipoproteins, hemostasis and thrombosis. Thromb Haemost 86:386–394 24. Dutt T, Toh CH (2008) The Yin-Yang of thrombin and activated protein C. Br J Haematol 140:505–515 25. Lardizabal JA, Joshi BK, Ambrose JA (2010) The balance between anti-ischemic efficacy and bleeding risk of antithrombotic therapy in percutaneous coronary intervention: a Yin-Yang paradigm. J Invasive Cardiol 22:284–292 26. Fardoun M, Dehaini H, Shaito A et al (2020) The hypertensive potential of estrogen: an untold story. Vasc Pharmacol 124:106600 27. Marasciulo FL, Montagnani M, Potenza MA (2006) Endothelin-1: the Yin and Yang on vascular function. Curr Med Chem 13: 1655–1665 28. Abbott GW (2016) KCNE1 and KCNE3: the Yin and Yang of voltage-gated K(+) channel regulation. Gene 576:1–13 29. Kanbay M, Bayram Y, Solak Y, Sanders PW (2013) Dietary potassium: a key mediator of the cardiovascular response to dietary sodium chloride. J Am Soc Hypertens 7:395–400 30. Carmichael ST (2012) Brain excitability in stroke. Arch Neurol 69:161–167

Chapter 12 GTPγS Assay for Measuring Agonist-Induced Desensitization of Two Human Polymorphic Alpha2B-Adrenoceptor Variants Jordana I. Borges, Alexandra M. Carbone, Natalie Cora, Anastasiya Sizova, and Anastasios Lymperopoulos Abstract α2-Adrenergic receptors (ARs) mediate many cellular actions of epinephrine and norepinephrine, including inhibition of their secretion (sympathetic inhibition) from adrenal chromaffin cells. Like many other G protein-coupled receptors (GPCRs), they undergo agonist-dependent phosphorylation and desensitization by GPCR kinases (GRKs), a phenomenon recently shown to play a major role in the sympathetic overdrive that accompanies and aggravates chronic heart failure. A three-glutamic acid deletion polymorphism in the human α2B-AR subtype gene (Glu301–303) causes impaired agonist-promoted receptor phosphorylation and desensitization, resulting in enhanced signaling to inhibition of cholinergic-induced catecholamine secretion in adrenal chromaffin cells. One of the various pharmacological assays that can be used to quantify and quantitatively compare the degrees of agonist-dependent desensitization, i.e., G protein decoupling, of these two polymorphic α2B-AR variants (or of any two GPCRs for that matter) is the guanosine-50 -O-3thiotriphosphate (GTPγS) assay that can directly quantify heterotrimeric G protein activation. Key words Polymorphic α2B-adrenergic receptor, Agonist-dependent desensitization, Heterotrimeric G protein, G protein-coupled receptor, Signal transduction, GTPγS assay

1

Introduction Three distinct α2-adrenergic receptor (α2-AR) subtypes (α2A, α2B, α2C) that mediate many of the physiological actions of the catecholamines (CAs) epinephrine (Epi) and norepinephrine (NE) have been described [1]. They belong to the family of G protein-coupled receptors (GPCRs) and they are linked to the inhibitory Gi/o proteins [1]. The α2B-AR is critically involved in cardiovascular regulation, as its gene disruption in mice affects blood pressure responses to α2-adrenoceptor agonists (e.g., clonidine) [2]. Its role in the central nervous system (CNS) remains largely elusive. It may be important in developmental processes, since homozygous

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_12, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

267

268

Jordana I. Borges et al.

α2B-KO mice do not breed well [2–12]. Like many other GPCRs, the α2B-AR undergoes agonist promoted desensitization [3] initiated by the phosphorylation of the receptor in its third intracellular loop by a well-characterized family of serine/threonine kinases termed G protein-coupled receptor kinases (GRKs), the most prominent member of which is the ubiquitously expressed GRK2 [4]. The phosphorylated receptor then interacts with a certain family of proteins termed β-arrestins, which physically uncouple the receptor from G proteins, thus terminating receptor signaling [4]. α2-ARs play a very important role in autocrine feedback regulation of catecholamine secretion from the chromaffin cells of the adrenal medulla [5]. By coupling to the Gi/o proteins, they inhibit further CA release upon their stimulation by the secreted CA, thereby participating in an autocrine negative feedback loop controlling adrenal CA secretion [5, 6]. The (patho)physiological importance of this α2-AR-mediated control of adrenal CA secretion has been well established, as deregulation of this signaling system in the adrenal chromaffin cells has been shown to underlie excessive sympathetic outflow and circulating CA levels that accompany and aggravate chronic heart failure (HF) [13– 41]. More specifically, upregulated GRK2 has been found to desensitize and downregulate chromaffin cell α2-ARs extensively in HF mouse and rat adrenal glands, thus rendering these receptors non-functional in HF. This allows for unopposed, continuous CA secretion, which contributes to the enhanced CA levels in chronic HF [7]. A common genetic variant of the α2B-AR subtype consisting of a deletion of three glutamic acid residues (residues 301–303) displays impaired agonist-promoted receptor phosphorylation and desensitization in various transfected cell lines [10–14]. Below, we describe a protocol for the guanosine-50 -O-3-thiotriphosphate (GTPγS) assay, which is an elegant and reliable assay for measuring heterotrimeric G protein activation levels by a GPCR in native cells in vitro or tissues in vivo. Therefore, it can also be used to determine levels of receptor desensitization, since GPCR desensitization essentially means inability of the receptor to activate G proteins [4]. This protocol was originally published in Ref. [14]. The assay’s principle is basically as follows: the agonistbound GPCR interacts with the C-terminus of the Gα subunit of a heterotrimeric G protein, inducing a conformational change in the Gα. This rather large conformational change forces the Gα subunit to release the bound GDP (guanosine diphosphate), which is then immediately exchanged for GTP (guanosine triphosphate, since the latter is much more abundant inside mammalian cells than GDP) [42]. Over time, the Gα subunit normally hydrolyzes the bound GTP to GDP thanks to its intrinsic GTPase activity, and this inactivates the Gα subunit (the G protein αβγ heterotrimer is re-formed) [43]. GTPγS is a non-hydrolysable GTP analog,

GTPγS Assay for Desensitization of Adrenoceptor Genetic Variants

269

whose agonist-induced binding to the plasma membrane can serve as an index of G protein activation. If the molecule is appropriately labeled, its binding can be measured and provide indirect measurement of G protein signaling activity. Indeed, the sulfur atom of the gamma phosphate of GTPγS, responsible for the resistance of the molecule to hydrolysis by the Gα subunit of the G protein, can be radiolabeled by using the 35S isotope of sulfur (that emits beta radiation, like tritium), resulting in the 35S-GTPγS radiochemical used in the GTPγS assay. Therefore, by measuring the beta radiation emitted from the membrane samples at the end of the assay, the level of G protein activation in the sample can be inferred. Of note, a non-hydrolysable radiolabeled GTP analog has to be used for this assay. A hydrolysable compound would be useless, because all the radioactivity would be lost from the membrane samples into the assay buffer, courtesy of the GTPase activity of the Gα subunit of the G protein, since the gamma phosphate (removed upon GTP hydrolysis) is the one that is radiolabeled.

2

Materials

2.1 Plasma Membrane Preparation

1. Polytron homogenizer. 2. Tabletop self-cooling Eppendorf tube centrifuge. 3. Beckman ultracentrifuge. 4. PBS: Phosphate buffer saline. 5. Hypotonic lysis buffer: 5 mM TrisCl, 5 mM EDTA, pH 7.4.

2.2

GTPγS Assay

1. Assay buffer: 20 mM HEPES pH 7.4; 100 mM NaCl, 10 μg/ ml saponin, 1 mM MgCl2. 2. Membranes diluted in assay buffer to give 250 μg/ml (2.5 μg/  10 μl); to be kept on ice (0 C). 3. 100 μM GDP solution in assay buffer (to be diluted to a final concentration of 10 μM). 4. ~25,000 dpm/10 μl GTPγ-35S, diluted in assay buffer. 5. Appropriate ligands to derive a dose-response curve. In the case of measuring α2B-AR desensitization, a full agonist (e.g., NE, Epi or brimonidine/UK14304) diluted in assay buffer at various concentrations across a logarithmic scale (e.g., 10 nM, 100 nM, 1 μM, 10 μM, 100 μM, 1 mM, 10 mM) is required. A concentration close to the EC50 (e.g., 10 μM for brimonidine) is also needed for pre-incubation (to desensitize the receptor). After a 30-min pre-exposure, the agonist is washed off and the membrane preparation is reloaded with a concentration of the same agonist for the GTPγS assay.

270

Jordana I. Borges et al.

6. Cold (non-radiolabeled) GTPγS (control for non-specific binding). 7. Appropriate microplate (with wells of appropriate volume capacities that fit the dimensions of the cell harvester used). 8. Glass fiber filters (e.g., UniFilter, GF/B). 9. Beta-scintillation fluid (or cocktail). 10. Cell harvester. 11. High-throughput beta-scintillation counter (e.g., TopCount® or MicroBeta®).

3

Methods

3.1 Plasma Membrane Preparation from Excised Adrenal Glands

1. Whole adrenal glands are excised from rats (or mice) and immediately placed in ice-cold PBS solution. 2. Surrounding fat gets trimmed as much as possible (see Note 1). 3. The rest of the tissue is then placed in 1 ml hypotonic lysis buffer solution inside a 15-ml Falcon tube kept on ice (see Note 2). 4. Homogenization of the tissue in a Polytron homogenizer in this buffer while kept on ice (0  C). 5. Short (25%), obesity (>70%), diabetes mellitus type 2 (>25%), hypercholesterolemia (40%), hypertriglyceridemia (20%), metabolic syndrome (20%), hepatobiliary disorder (15%), endocrine/metabolic disorders (>20%), cardiovascular disorder (40%), cerebrovascular disorder (60–90%), neuropsychiatric disorders (60–90%), and cancer (10%). For the past decades, pharmacological studies in search of potential treatments for AD focused on the following categories: neurotransmitter enhancers (11.38%), multitarget drugs (2.45%), anti-amyloid agents (13.30%), anti-tau agents (2.03%), natural products and derivatives (25.58%), novel synthetic drugs (8.13%), novel targets (5.66%), repository drugs (11.77%), anti-inflammatory drugs (1.20%), neuroprotective peptides (1.25%), stem cell therapy (1.85%), nanocarriers/nanotherapeutics (1.52%), and other compounds (60%) are carriers of over ten pathogenic genes. The genes that most frequently (>50%) accumulate pathogenic variants in the same AD case are A2M (54.38%), ACE (78.94%), BIN1 (57.89%), CLU (63.15%), CPZ (63.15%), LHFPL6 (52.63%), MS4A4E (50.87%), MS4A6A (63.15%), PICALM (54.38%), PRNP (80.7059), and PSEN1 (77.19%). There is also an accumulation of 15 to 26 defective pharmagenes in approximately 85% of AD patients. About 50% of AD patients are carriers of at least 20 mutant pharmagenes, and over 80% are deficient metabolizers for the most common drugs, which are metabolized via the CYP2D6, CYP2C9, CYP2C19, and CYP3A4/5 enzymes. The implementation of pharmacogenetics can help optimize drug development and the limited therapeutic resources available to treat AD, and personalize the use of anti-dementia drugs in combination with other medications for the treatment of concomitant disorders. Key words Alzheimer’s disease, Anti-dementia drugs, Cerebrovascular genomics, Concomitant disorders, Drug development, Neurodegenerative genomics, Pathogenic genes, Pharmacoepigenetics, Pharmacogenomics, Phenotypic profile

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_13, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

275

276

1

Ramo´n Cacabelos et al.

Introduction Alzheimer’s disease (AD) is a priority health problem in advanced societies and its prevalence and incidence is gradually increasing in many emerging and developing countries, with a high cost for public health services and for the economy of families. The cost of dementia exceeds US$800 billion dollars worldwide (>1% of GDP), with an average cost per patient/year that fluctuates between $30,000 and $60,000, depending on the country, social status, quality of medical care, and the stage of the disease [1, 2]. Approximately, 20% of direct costs are associated with pharmacological treatment, with poor cost-effectiveness. AD is the most frequent form of dementia (50–60%), followed by vascular dementia (30–40%), other forms of dementia (10–15%), and mixed dementia, which is the most prevalent form of dementia (>70%) in patients older than 75 years of age. AD is more frequent in females than in males [3]. AD results from the premature death of neurons caused by genomic, epigenomic, cerebrovascular, and multiple environmental factors, giving rise to a clinical phenotype characterized by progressive cognitive deterioration, behavioral changes, and functional decline [4–7]. By conventional criteria, AD is a continuum [8], which can be differentiated into early-onset AD (EOAD), associated with Mendelian mutations in specific genes with familial transmission (familial AD, FAD), and late-onset AD (LOAD) attributed to a more complex pathogenic profile in which multiple susceptibility genes (>600) are involved together with diverse environmental factors (sporadic AD, sAD) [5]. The neuropathological hallmarks of AD are represented by extracellular deposits of aggregated β-amyloid (Aβ) in senile plaques and vessels (amyloid angiopathy), and intracellular neurofibrillary tangles (NFTs), formed by hyperphosphorylation of tau protein in microtubules and neurofilaments. Aβ and tau may act independently or have synergistic effects on AD pathogenesis [9]. These classical neuropathological markers are accompanied by astrogliosis, microglia activation, dendritic dystrophy, and progressive neuronal loss in critical regions of the hippocampus and neocortex, compromising circuits involved in higher activities of the central nervous system (CNS). This neuropathological picture is accompanied by the phenotypic expression of epigenetic aberrations, neurotrophic dysfunction, neurotransmitter deficits (cholinergic, monoaminergic, glutamatergic, GABAergic, neuropeptidergic), neuroinflammation, lipid peroxidation due to oxidative stress reactions, and cerebrovascular (hypoperfusion) damage [4, 10–12]. The main challenges facing the scientific community, the medical services, and the pharmaceutical industry today are (i) to deepen into a better understanding of the primary causes of the

Pharmacogenomics of Alzheimer’s Disease

277

disease and its pathogenic mechanisms, (ii) the characterization and validation of reliable biomarkers that allow for an early diagnosis, (iii) the identification and development of new drugs and therapeutic strategies able to slow down or halt the course of the disease, and (iv) under optimal conditions, developing new preventive protocols capable of blocking the evolution of the disease in the population at risk at presymptomatic stages, taking into account that brain damage in AD begins several decades before the clinical manifestation of symptoms of dementia [13, 14]. The aging adult population accumulates many other pathologies concomitant with dementia that force the establishment of polypharmaceutical regimens, with the consequent increase in the risk of adverse drug reactions (ADRs) and dangerous drug-drug interactions (DDIs) [14, 15]. In fact, over 80% of dementia patients consume more than ten different medications daily. Presently, the most efficient way to reduce ADRs and DDIs is to implement pharmacogenetic protocols for the personalization of pharmacological treatment of patients with dementia [16]; several studies demonstrate that the therapeutic response to conventional drugs in AD is genotype-dependent [12, 17–19]. In the present chapter, we summarize (i) the phenotypic profile of patients with dementia, (ii) the concomitant pathologies susceptible to therapeutic intervention, in addition to conventional antidementia treatments, (iii) the prevalent lines of investigation for AD-related drug development, (iv) the major components of the pharmacogenetic machinery in AD, (v) the pharmacogenetics of anti-dementia drugs, and (vi) the pharmacogenetics and pharmacoepigenetics of multifactorial drug regimens for the treatment of AD.

2

Phenotypic Profile In a randomly selected cohort of patients diagnosed for neurocognitive disorder (NCD)-AD (331.0 (G30.9) (DSM-V/NINCDSADRDA criteria), we investigated sex-related common phenotypes including biochemistry, hematology, metabolism, hormones, neurotransmitters, and cardiovascular and cerebrovascular function, as well as cognition, mood, behavior, genomic, and pharmacogenomic profiles (Table 1). There are significant differences between females and males across many biological parameters, including those associated with biochemical, hormonal, hematological, cognitive, and behavioral markers (Table 1). Cognitive markers (MMSE, ADAS) indicate that females show a worse cognitive performance than males (Table 1). Depression and cognitive impairment are the first symptoms to appear in 98.5% and 99.1% of patients with LOAD and 9% and 80%, respectively, in EOAD cases [20].

278

Ramo´n Cacabelos et al.

Table 1 Phenotypic features of patients with Alzheimer’s disease Parameter (normal range)

Total

Females

Males

Differences

2701

1491 (55%)

1210 (45%)

Age (years) Range

67.63
0.19 50–96

68.26
0.27 50–96

66.86
0.28 50–94

Systolic blood pressure (mmHg) (120–160)

141.39
0.41 160: 21.84%

141.25
0.56 141.56
0.61 14.75% 12.37% 22.25% 21.33%

Diastolic blood pressure (mmHg) (70–85)

79.38
0.21 85: 28.53%

78.83
0.28 11.27% 26.84%

80.05
0.33 9.54% 30.74%

p < 0.001 p(χ 2) ¼ 0.24 p(χ 2) ¼ 0.10

Pulse (bpm) (60–100)

68.29
0.23 100: 2.07%

69.60
0.31 19.05% 2.35%

66.67
0.35 28.75% 1.74%

p < 0.001 p(χ 2) ¼ 0.003 p(χ 2) ¼ 0.34

Weight (kg)

72.23
0.27 (36–127)

66.41
0.32 (36–112)

79.41
0.37 (42–127)

P < 0.001

Height (cm)

160.71
0.21 (120–188)

154.96
0.21 167.79
0.26 (140–182) (120–188)

BMI (kg/m2) Underweight (15–18.5) Normal weight (18.5–25) Overweight (25–30) Obese class I (moderate) (30–35) Obese class II (severe) (35–40) Obese class III (very severe) (>40)

29.93
0.09 1.01% 25.03% 45.26% 21.25% 6.08% 1.33%

27.76
0.13 1.54% 29.15% 39.50% 20.85% 7.34% 1.62%

28.15
0.11 0.27% 19.86% 52.35% 21.75% 4.69% 1.08%

p < 0.002 p(χ 2) < 0.003 p(χ 2) < 0.001 p(χ 2) < 0.001 p(χ 2) ¼ 0.69 p(χ 2) < 0.01 P(χ 2) ¼ 0.34

Glucose (mg/dL) (70–105)

102.00
0.55 105: 26.55%

99.17
0.69 0.54% 21.63%

105.49
0.87 0.24% 32.70%

p < 0.001 p(χ 2) ¼ 0.38 p(χ 2) < 0.001

Cholesterol (mg/dL) (140–220)

219.02
0.90 220: 40.54%

228.61
1.22 207.20
1.28 5.79% 1.81% 21.16% 56.27%

p < 0.001 p(χ 2) < 0.001 p(χ 2) < 0.001

HDL-cholesterol (mg/dL) (35–75)

55.04
0.28 75: 10.11%

59.98
0.38 2.35% 15.49%

p < 0.001 p(χ 2) < 0.001 p(χ 2) < 0.001

LDL-cholesterol (mg/dL) (80–160)

140.85
0.77 160: 29.54%

146.08
1.03 134.41
1.12 3.69% 6.94% 33.67% 24.46%

p < 0.001 p(χ 2) < 0.001 p(χ 2) < 0.001

Triglycerides (mg/dL) (50–150)

113.83
1.31 150: 19.44%

108.47
1.50 120.44
2.26 5.12% 5.97% 22.89% 16.63%

p < 0.001 p(χ 2) ¼ 0.41 p(χ 2) < 0.001

N

48.95
0.34 9.09% 3.47%

p < 0.001 p ¼ 0.74 p(χ 2) ¼ 0.13 p(χ 2) ¼ 0.66

P < 0.001

(continued)

Pharmacogenomics of Alzheimer’s Disease

279

Table 1 (continued) Parameter (normal range)

Total

Females

Males

Differences

Urea (mg/dL) (15–30)

42.90
0.26 30: 88.49%

41.79
0.33 0.34% 86.45%

44.27
0.40 0.34% 90.99%

p < 0.001 p ¼ 0.75 p(χ 2) ¼ 0.37

Creatinine (mg/dL) (0.70–1.40)

0.91
0.006 1.40: 3.59%

0.81
0.004 23.47% 1.74%

1.03
0.01 2.40% 8.26%

p < 0.001 p(χ 2) < 0.001 p(χ 2) < 0.001

Uric acid (mg/dL) (3.4–7.0)

4.47
0.03 7.0: 6.03%

3.90
0.03 36.02% 2.48%

5.17
0.05 8.68% 10.41%

p < 0.001 p(χ 2) < 0.001 p(χ 2) < 0.001

Total protein (g/dL) (6.5–8.0)

6.93
0.03 8.0: 1.78%

6.90
0.01 12.54% 1.61%

6.97
0.07 15.29% 1.98%

p ¼ 0.17 p(χ 2) ¼ 0.08 p(χ 2) ¼ 0.56

Albumin (g/dL) (3.5–5.0)

4.20
0.007 5.0: 0.95%

4.18
0.008 1.50% 0.58%

4.21
0.01 1.60% 1.60%

p ¼ 0.02 p(χ 2) ¼ 0.99 p(χ 2) ¼ 0.03

Calcium (mg/dL) (8.1–10.4)

9.24
0.009 10.4: 2.04%

9.27
0.01 0.20% 2.35%

9.19
0.01 0.66% 1.65%

p < 0.001 p(χ 2) ¼ 0.12 p(χ 2) ¼ 0.26

Phosphorus (mg/dL) (2.5–5.0)

3.41
0.01 5.0: 0.67%

3.52
0.01 0.92% 0.81%

3.27
0.01 3.47% 0.49%

p < 0.001 p(χ 2) < 0.002 P(χ 2) ¼ 0.46

GOT/ASAT (IU/L) (10–40)

21.70
0.20 40: 3.85%

21.29
0.42 0.60% 3.35%

22.20
0.39 0.25% 4.46%

p ¼ 0.006 p(χ 2) ¼ 0.83 P(χ 2) ¼ 0.18

GPT/ALAT (IU/L) (9–43)

23.52
0.36 43: 7.07%

21.51
0.45 3.02% 4.96%

26.00
0.56 2.56% 9.67%

p < 0.001 p(χ 2) ¼ 0.56 p(χ 2) < 0.001

GGT (IU/L) (11–50)

30.55
0.79 50: 11.81%

26.37
1.07 11.54% 8.52%

35.69
1.17 2.65% 15.87%

p < 0.001 p(χ 2) < 0.001 P(χ 2) < 0.001

Alkaline phosphatase IU/L) (37–111)

77.05
0.62 111: 10.37%

79.41
0.85 1.74% 10.39%

74.14
0.82 2.48% 10.83%

p < 0.001 p(χ 2) ¼ 0.24 P(χ 2) ¼ 0.98

Bilirubin (mg/dL) (0.20–1.00)

0.75
0.02 1.00: 15.07%

0.71
0.04 0.34% 10.33%

0.80
0.01 0.24% 20.91%

p < 0.001 p(χ 2) ¼ 0.95 p(χ 2) < 0.001

CPK (IU/L) (38–174)

92.72
2.00 174: 7.29%

87.20
3.12 9.25% 5.03%

99.52
2.29 6.94% 10.08%

p < 0.001 p(χ 2) < 0.05 p(χ 2) < 0.001 (continued)

280

Ramo´n Cacabelos et al.

Table 1 (continued) Parameter (normal range)

Total

Females

Males

LDH (IU/L) (200–480)

277.29
1.52 480: 1.74%

289.77
2.06 261.88
2.16 15.87% 6.71% 1.16% 2.21%

p < 0.001 p(χ 2) < 0.008 p(χ 2) < 0.05

Na+ (mEq/L) (135–148)

141.75
0.04 148: 0.89%

141.86
0.05 141.62
0.06 0.36% 0.82% 0.80% 0.99%)

p < 0.002 p(χ 2) ¼ 0.65 p(χ 2) ¼ 0.76

K+ (mEq/L) (3.5–5.3)

4.33
0.006 5.3: 1.15%

4.29
0.009 1.07% 0.67%

p < 0.001 p(χ 2) ¼ 0.24 p(χ 2) < 0.01

Cl (mEq/L) (98–107)

104.06
0.07 107: 13.92%

104.20
0.13 103.88
0.07 1.40% 1.07% 12.96% 14.89%

Fe2+ (μg/dL) (35–160)

86.60
0.70 160: 2.61%

81.98
0.89 7.01% 1.64%

92.17
1.10 3.80% 4.20%

p < 0.001 p(χ 2) < 0.01 p(χ 2) < 0.001

Ferritin (ng/mL) (F: 11–307) (M: 24–336)

121.05
2.91 307: 7.65%

81.01
2.48 307: 2.09%

169.39
5.32 336: 15.3%

p < 0.001 p(χ 2) < 0.001 p(χ 2) < 0.001

Folate (ng/mL) (>3.00)

7.94
0.08 2.47: 8.91%

4.38
0.009 0.57% 1.74%

Differences

p < 0.04 p(χ 2) ¼ 0.55 P(χ 2) ¼ 0.29

p < 0.04 p(χ2) < 0.001 P(χ2) < 0.001

(continued)

Pharmacogenomics of Alzheimer’s Disease

281

Table 1 (continued) Parameter (normal range)

Total

FSH (mIU/mL) (F: 21.7–153) (M: 0.7–11.1) LH (mIU/mL) (F: 7.7–58.5) (M: 1.7–8.6)

Females

Males

Differences

67.07
1.57 41.37
1.36 153: 0.92%

9.39
0.47 11: 23.85%

p < 0.001 p(χ2) < 0.001 P(χ2) < 0.001

16.84
0.53 26.04
0.63 58.5: 2.77%

5.38
0.20 8.6: 12.93%

p < 0.001 p(χ2) ¼ 0.77 P(χ2) < 0.001

26.23
1.29 30: 16.31%

Estrogen (pg/mL) (20–30)

281.45
11.49 740: 1.87%

Testosterone (ng/dL) (193–740) α-Amylase (U/L) (28–100)

59.98
1.26 >100: 6.70%

57.94
1.64 5.70%

62.67
1.95 7.77%

p ¼ 0.05 P(χ2) ¼ 0.43

Lipase (U/L) (13–60)

43.62
0.71 >60: 10.58%

42.98
0.76 9.56%

44.46
1.29 11.94%

p ¼ 0.75 p(χ2) ¼ 0.42

AFP (ng/mL) (0–7)

3.17
0.10 >7: 5.44%

3.33
0.10 5.88%

2.97
0.19 4.85%

p < 0.001 P(χ2) ¼ 0.68

CEA (ng/mL) (0–3.8)

3.97
1.40 >3.8: 16.76%

2.32
0.09 12.99%

6.14
2.26 21.75%

p < 0.05 P(χ2) < 0.01

CA 19.9 (U/mL) (0–27)

14.41
2.05 >27: 8.94%

11.60
0.61 7.35%

18.14
4.70 11.04%

p ¼ 0.27 P(χ2) ¼ 0.15

CA 72.4 (U/mL) (0–6.9)

3.96
0.86 >6.9: 13.01%

4.86
1.46 14.71%

2.85
0.66 10.91%

p ¼ 0.18 P(χ2) ¼ 0.78

CA 125 (U/mL) (0–35)

16.38
1.67 >35: 5.97%

15.82
2.42 4.19%

17.16
2.14 8.50%

P ¼ 0.87 P(χ2) ¼ 0.16

CYFRA 21.1 (ng/mL) (0–3.3)

1.85
0.08 >3.3: 10.39%

1.76
0.06 9.52%

1.94
0.15 11.32%

P ¼ 0.83 P(χ2) ¼ 0.76

SCC (ng/mL) (0–2.3)

1.30
0.09 >2.3: 9.785

1.05
0.05 6.59%

1.55
0.18 12.96%

P < 0.001 P(χ2) ¼ 0.11

NSE (ng/mL) (0–16.3)

10.28
0.17 >16.3: 3.95%

10.77
0.24 5.36%

9.78
0.23 2.48%

p < 0.001 P(χ2) ¼ 0.31

2.31
0.16 >4: 13.45%

PSA (ng/mL) (26.4: 7.25%

CA 15.3 (U/mL) (0–26.4) RBC (x106/μL) (3.80–5.50)

4.62
0.008 5.50: 3.99%

4.47
0.01 4.09% 2.01%

4.81
0.01 1.16% 6.45%

p < 0.001 p(χ 2) < 0.001 P(χ 2) ¼ 0.001 (continued)

282

Ramo´n Cacabelos et al.

Table 1 (continued) Parameter (normal range)

Total

Females

Males

Differences

HCT (%) (40.0–50.0)

42.41
0.23 50.0: 3.07%

40.69
0.26 42.85% 0.54%

44.53
0.39 13.22% 6.19%

p < 0.001 p(χ 2) ¼ 0.001 p(χ 2) ¼ 0.001

Hb (g/dL) (13.5–17.0)

14.04
0.02 17.0: 2.29%

13.47
0.03 46.88% 0.13%

14.74
0.04 13.80% 4.95%

p < 0.001 p(χ 2) ¼ 0.001 p(χ 2) ¼ 0.001

MCV (fL) (80–100)

91.06
0.10 100: 3.59%

90.48
0.13 2.41% 2.55%

91.77
0.15 2.15% 4.88%

p < 0.001 p(χ 2) ¼ 0.75 P(χ 2) < 0.003

MCH (pg) (27.0–33.0)

30.41
0.03 33.0: 5.48%

30.19
0.05 4.02% 5.23%

30.69
0.05 3.14% 8.26%

p < 0.001 p(χ 2) ¼ 0.28 P(χ 2) < 0.004

MCHC (g/dL) (31.0–35.0)

33.37
0.01 35.0: 2.33%

33.32
0.02 0.80% 2.35%

33.43
0.02 0.41% 2.31%

P < 0.002 p(χ 2) ¼ 0.30 p(χ 2) ¼ 0.94

ADE (RDW) (%) (11.0–15.0)

12.95
0.02 15.0: 5.63%

13.01
0.03 1.61% 6.24%

12.88
0.03 2.07% 4.88%

p < 0.05 p(χ 2) ¼ 0.47 p(χ 2) ¼ 0.17

WBC (x103/μL) (4.0–11.0)

6.35
0.03 11.0: 2.55%

6.18
0.05 7.18% 2.28%

6.56
0.05 3.47% 2.89%

p < 0.001 p(χ 2) < 0.001 P(χ 2) ¼ 0.39

%Neu (45.0–70.0)

60.15
0.19 70.0: 17.92%

59.98
0.25 7.04% 15.16%

60.35
0.30 4.79% 16.86%

P ¼ 0.22 P(χ 2) < 0.02 P(χ 2) ¼ 0.33

%Lym (20.0–40.0)

29.60
0.22 40: 11.51%

30.22
0.22 11.80% 12.81%

28.83
0.42 15.45% 9.92%

P < 0.001 p(χ 2) < 0.02 P(χ 2) < 0.04

%Mon (3.0–10.0)

7.24
0.03 10.0: 9.48%

7.03
0.05 1.67% 7.31%

7.50
0.06 0.91% 12.15%

P < 0.001 P(χ 2) ¼ 0.12 P(χ 2) < 0.001

%Eos (1.0–5.0)

2.81
0.05 5.0: 23.92%

2.61
0.06 6.71% 24.21%

3.05
0.09 4.55% 23.55%

p < 0.001 p(χ 2) < 0.001 p(χ 2) < 0.001

%Bas (0.0–1.0)

0.69
0.07 >1.0: 14.09%

0.69
0.14 9.09%

0.69
0.01 19.60%

P < 0.001 P(χ 2) < 0.001

PLT (103/μL) (150–450)

227.11
1.25 450: 0.66%

239.85
1.67 211.29
1.80 11.24% 4.63% 0.85% 0.55%

p < 0.001 p(χ 2) < 0.001 P(χ 2) ¼ 0.49

MPV (fL) (6.0–10.0)

8.55
0.01 10.0: 9.03%

8.55
0.02 0.13% 8.58%

P ¼ 0.99 P(χ 2) ¼ 0.49 P(χ 2) ¼ 0.44

8.55
0.02 0.25% 9.58%

(continued)

Pharmacogenomics of Alzheimer’s Disease

283

Table 1 (continued) Parameter (normal range)

Total

Females

Males

Differences

MMSE Score (0–30) 70%), diabetes mellitus type 2 (26%), hypercholesterolemia (40%), hypertriglyceridemia (20%), hyperuricemia (6%), metabolic syndrome (20%), transaminitis (11%), hyperbilirubinemia (15%), endocrine disorders (5%), iron deficiency anemia (7%), folate deficit (17%), vitamin B12 deficit (10%), cardiovascular disorder (40%), cerebrovascular disorder (>90% in patients over 80 years of age), anxiety (60%), depression (65%), behavioral disorders (20–90%), and cancer (10%) (Table 1). Cardiovascular risk factors, such as hypertension, hypercholesterolemia/dyslipidemia, and ECG abnormalities, are more frequent in males than in females. Hypertension is present in 21% of the cases. Systolic blood pressure is similar in females and males, but diastolic blood pressure is significantly higher in males than in females ( p < 0.001) (Table 1). Cholesterol levels (total, LDL) are higher in men and HDL-cholesterol and triglyceride levels are more elevated in females (Table 1). Cardiovascular disorders and blood pressure changes, either hypertension or hypotension, in AD are currently reported in association with increased risk of brain damage and increased cognitive deterioration [35, 36]. Furthermore, APOE variants are associated with AD, cardiovascular disorders, atherosclerosis, and cerebrovascular damage in dementia [4, 11, 37–40]. Lipid metabolism disorder and the cerebrovascular component of AD have been extensively studied, and alterations in cholesterol, changes in cell membrane lipids, and arteriosclerosis

Pharmacogenomics of Alzheimer’s Disease

285

lead to ischemia and cerebral hypoperfusion that contributes to accelerating the premature death of neurons in patients with predisposition to AD [38, 41–44]. In contrast, the epidemiological link between diabetes and AD appears to be circumstantial, with no apparent pathogenic implications beyond the deleterious effects of hyperglycemia on brain function [45–47]. As a consequence of all these concomitant pathologies, patients with dementia consume a wide variety of drugs whose side effects contribute to accelerating the degenerative process and cognitive decline. Of special importance, quantitative and qualitative, are cardiovascular agents, statins, antidiabetics, antihypertensive drugs, analgesics, diuretics, bronchodilators, antirheumatics, and various categories of psychotropic drugs (neuroleptics, antidepressants, anxiolytics, hypnotics, sedatives). The correct administration of these drugs requires a personalized therapeutic intervention, together with conventional anti-dementia treatments [14, 16, 48]. Combination treatments applied under pharmacogenetic guidance indicate that biochemical, hematological, and metabolic differences may contribute to changes in drug efficacy and safety. Concerning cognitive function and neuropsychiatric disorders treated with multifactorial regimes, females and males respond differentially to treatment, showing a moderate improvement in cognition during the first year of treatment (with progressive cognitive decline thereafter) and significant improvements in anxiety and depression [14, 19, 49]. Pharmacogenetic studies show that APOE-3 carriers are the best responders and that APOE-4 carriers tend to be the worst responders to conventional treatments [49]. Among CYP2D6, CYP2C19, and CYP2C9 genophenotypes, normal metabolizers (NMs) and intermediate metabolizers (IMs) are significantly better responders than poor metabolizers (PMs) and ultra-rapid metabolizers (UMs) to therapeutic interventions that modify cognition and mood phenotypes in dementia [18, 19, 49–51].

4

Alzheimer’s Disease Therapeutics and Drug Development The major focus of pharmacological research in AD for the past 50 years has been the search for cognitive enhancers. The identification of a selective cholinergic dysfunction in the basal forebrain together with neuronal loss in neocortex and hippocampus led to the introduction of acetylcholinesterase inhibitors (AChEIs) as the first option to restore cholinergic neurotransmission in the early 1990s. Tacrine (9-amino-1,2,3,4-tetrahydroacridine) was introduced in 1993 as the first AChEI for the treatment of AD, and huperzine A was approved in China in 1994. These first AChEIs were followed by a new generation of AChEIs, years later, including

286

Ramo´n Cacabelos et al.

donepezil, galantamine, and rivastigmine, and memantine, an N-methyl-D-aspartate (NMDA) glutamate receptor partial inhibitor, was introduced in 2003 [16, 48]. No new FDA-approved drugs for AD have been reported for the past 18 years until the recent approval of the antibody aducanumab in 2021 [52]. In the last decades, pharmacological studies in search of potential treatments for AD focused on the following categories: neurotransmitter enhancers (11.38%), multitarget drugs (2.45%), antiamyloid agents (13.30%), anti-tau agents (2.03%), natural products and derivatives (25.58%), novel synthetic drugs (8.13%), novel targets (5.66%), other (old) drugs (11.77%), anti-inflammatory drugs (1.20%), neuroprotective peptides (1.25%), stem cell therapy (1.85%), nanocarriers/nanotherapeutics (1.52%), and other compounds and/or therapeutic strategies, including polyunsaturated fatty acids, nootropics, neurotrophic factors, hormones, epigenetic drugs, miRNAs, RNAi/gene silencing, gene therapy, and combination treatments (60 new products), Υ-secretase modulators (1.42%) (40 products), and diverse Aβ aggregation inhibitors (5.02%) (>220 new products), as well as a small number of α-secretase modulators, δ-secretase inhibitors, Notch inhibitors, APP modulators, β-sheet breakers, and Aβ scavengers (Table 3). However, after notorious failures with secretase inhibitors and immunotherapeutic procedures [52, 54] (Table 4), there was a progressive decline in the number of studies on anti-Aβ treatments since 2017, and studies on immunotherapy decreased by 35% until recently. Anti-tau treatments account for 2.03% of the studies; anti-tau immunotherapy increased by >40%; and anti-tau aggregation studies declined by 40% (Table 4). APOE, as a therapeutic target, only represented 0.30% of the studies; however, publications on the pathogenic role of APOE in AD increased substantially, positioning APOE as the most important risk factor for AD [52]. In 2020, there were 121 agents in clinical trials (29 in phase III trials, 65 in phase II trials, and 27 in phase I trials); 12 agents in trials target cognitive enhancement; 12 agents, neuropsychiatric and behavioral symptoms; and 97 agents, disease modification. In comparison to 2019, there was an increase in the number of disease-modifying agents targeting non-amyloid or tau pathways [55].

Pharmacogenomics of Alzheimer’s Disease

287

Table 2 Novel cholinesterase inhibitors with potential preclinical effects in AD models Selective, dual, and multitarget acetylcholinesterase inhibitors (3-[2-({4-[(Dimethylamino)methyl]-2-oxo-2H-chromen-7-yl}oxy)ethoxy]-6,7-dimethoxy-2Hchromen-2-one (Benz)imidazolopyridino tacrines (E)-2-(5,7-Dibromo-3,3-dimethyl-3,4-dhihydroacridin-1(2H)-ylidene)hydrazinecarboxamide (E)-2-(5,7-Dibromo-3,3-dimethyl-3,4-dihydroacridin-1(2H)-ylidene)hydrazinecarbothiomide [3-(2/3/4-Methoxyphenyl)-6-oxopyridazin-1(6H)-yl]methyl carbamate derivatives ([3-(2-methoxyphenyl)-6-oxopyridazin-1(6H)-yl]methyl heptylcarbamate), [3-(2-methoxyphenyl)6-oxopyridazin-1(6H)-yl]methyl (4-methylphenyl)carbamate) {4-[(3,5-Dichloro-2-hydroxyphenyl)sulfonyl]-1-piperazinyl}(2-furyl)methanone 1-(5-Amino-2-methyl-4-(1-methyl-1H-imidazol-2-yl)-6,7,8,9-tetrahydro-4H-pyrano[2,3-b]quinolin3-yl)ethan-1-one 1-(5-amino-4-(2-chloro-7-methoxyquinolin-3-yl)-2-methyl-6,7,8,9-tetrahydro-4H-pyrano [2,3-b] quinolin-3-yl)e´thanone 1, 3-Dihydroxyxanthone Mannich base derivatives 1,2,3,4,5,6-Hexahydroazepino[4,3-b]indole 1,2,3,4-tetrahydro-1-acridone analogs 1,2,3-Triazole-based acetylcholinesterase inhibitors 1,2,3-Triazole-chromenone carboxamide derivatives (N-(1-benzylpiperidin-4-yl)-7((1-(3,4-dimethylbenzyl)-1H-1,2,3-triazol-4-yl)methoxy)-2-oxo-2H-chromene-3-carboxamide) 1,2-Oxazine-based small molecules 1,3,5-Triazine-benzenesulfonamides 1,3-Dihydroxyxanthone derivatives 1,8-Cineole 1-Benzylamino-2-hydroxyalkyl derivatives 1-Benzyl-N-(1-methyl-3-oxo-2-phenyl-2,3-dihydro-1H-pyrazol-4-yl) piperidine-4-carboxamide 1-Benzyl-N-(5,6-dimethoxy-8H-indeno[1,2-d]thiazol-2-yl)piperidine-4-carboxamide 1-Butanoyl-3-arylthiourea derivatives 1H-Pyrazolo[3,4-b]pyridine derivatives 2-((4-(1,3-dioxoisoindolin-2-yl)benzyl)amino)-2-oxoethyl-2-(4-methoxyphenyl)acetate 2-(2-(4-(2-Oxo-2-phenylethyl)piperazin-1-yl) ethyl)Isoindoline-1,3-dione derivatives 2-(2-(4-Benzoylpiperazin-1-yl)ethyl)isoindoline-1,3-dione derivatives 2-(9-Acridinylamino)-2-oxoethyl piperazine/piperidine/morpholinecarbodithioate derivatives 2-(Benzylamino-2-hydroxyalkyl)Isoindoline-1,3-diones derivatives (12 (2-(5-(benzylamino)-4hydroxypentyl)isoindoline-1,3-dione) (continued)

288

Ramo´n Cacabelos et al.

Table 2 (continued) 2-(Naphthalen-1-yl)-7-(8-(pyrrolidin-1-yl)octyloxy)-4H-chromen-4-one 2-(Piperazin-1-yl)N-(1H-pyrazolo[3,4-b]pyridin-3-yl)acetamides 2,4,5-Trisubstituted imidazoles 2,5-Dihydroxyterephthalamide derivatives 2,6-Disubstituted pyridazinone derivatives 2-Acetyl-5-O-(amino-alkyl)phenol derivatives 2-Amino-4H-pyrans 2-Arylbenzofuran derivatives 2-Arylbenzofurans from Artocarpus lakoocha 2-Arylidene-4-methoxy (or hydroxy)-7-methyl-1-indanone derivatives 2-Benzofuran carboxamide-benzylpyridinum salts 2-Butenediamide and oxalamide derivatives. 2-Chloroquinolin-3-yl substituted PyranoTacrines (PTs) 2-Furoic piperazide derivatives (2-furoic piperazide (1; 1-(2-furoyl)piperazine), 3,5-dichloro-2hydroxybenzenesulfonyl chloride) 2-Furyl(1-piperazinyl)methanone 20 -Hydroxychalcones 2-Hydroxy-N-phenylbenzamides 2-Methylimidazole and alkyl/aryl halides 2-Phenoxy-indan-1-one derivatives (PIOs) 2-Pyrazoline derivatives 2-Pyridylpiperazine and 5-phenyl-1,3,4-oxadiazoles hybrids 2-Substituted benzo[d]oxazol-5-amine derivatives 2-Vinyl chromones 3-(4-(Aminoalkoxy)benzylidene)-chroman-4-ones 3,5-Dichloro-2-hydroxybenzenesulfonyl chloride 3,5-Dimethylorsellinic acid-based meroterpenoids 3-[3-(Amino)propoxy] benzenamines 3-Arylcoumarins 3-Phenylcoumarin-lipoic acid conjugates 4-(1,3-Dioxoisoindolin-2-yl)-N-phenyl benzamide derivatives 4-(2-Fluorophenyl)-3-(3-methoxybenzyl)-1H-1,2,4-triazol-5 (4H)-one 4,6-Diphenylpyrimidine derivatives (continued)

Pharmacogenomics of Alzheimer’s Disease

289

Table 2 (continued) 4-Aminoquinolines 4-Arylcoumarins derivatives (multitarget) 4-Arylthiazole-2-amine derivatives 4-Hydroxycoumarin derivatives containing substituted benzyl-1,2,3-triazole moiety 4-Isochromanone hybrids 4-Isochromanone hybrids bearing N-benzyl pyridinium 4-N-Phenylaminoquinoline derivatives 4-N-Phenylaminoquinolines 40 -OH-flurbiprofen Mannich base derivatives 5-(Aroylhydrazinocarbonyl)escitalopram 5,10,15,20-Tetrakis (4-sulfonatophenyl) porphyrinato iron(III) chloride (FeTPPS) 5,10,15,20-Tetrakis (4-sulfonatophenyl) porphyrinato iron(III) nitrosyl chloride (FeNOTPPS) 5,6-Dimethoxy-1-oxo-2,3-dihydro-1H-inden-2-yl 1-benzylpiperidine-4-carboxylate 5,6-Dimethoxybenzo[d]isothiazol-3(2H)-one-N-alkylbenzylamine derivatives 50 /60 -Isonucleosides and theobromine-containing N-isonucleosidyl derivatives 5-Amino-2-phenyl-4H-pyrano[2,3-b]quinoline-3-carboxylates 5-Methyl-2,4-dihydro-3H-1,2,4-triazole-3-one’s aryl Schiff base derivatives 5-Oxo-4,5-dihydropyrano[3,2-c]chromene derivatives 6-Chloro-pyridonepezils 6-Chlorotacrine 6H-Benzo[c]chromen-6-one, and 7,8,9,10-tetrahydro-benzo[c]chromen-6-one derivatives 6-Methyluracil derivatives 6-Substituted-3(2H)-pyridazinone-2-acetyl-2-(nonsubstituted/4-substituted benzenesulfonohydrazide) derivatives 7-Aminoalkyl-substituted flavonoid derivatives 7-Chloro-4-aminoquinoline Schiff bases 7H-Thiazolo-[3,2-b]-1,2,4-triazin-7-one derivatives 7-MEOTA (2 homodimers linked by 2 C4-C5 chains and 5 N-alkylated C4-C8 side chain derivatives) 7-MEOTA-tryptophan hybrids 7-Methoxy tacrine derivatives (7-MEOTA) 7-Methoxytacrine-adamantylamine heterodimers 9-Amino-1,2,3,4-tetrahydroacridine derivatives with iodobenzoic acid 9H-Carbazole derivatives containing the N-benzyl-1,2,3-triazole moiety (continued)

290

Ramo´n Cacabelos et al.

Table 2 (continued) 9-Methylfascaplysin) 9-Substituted acridine derivatives 9-Substituted acridine derivatives (9-aryl(heteroaryl)-N-methyl-9,10-dihydroacridines, 9-aryl (heteroaryl)-N-methyl-acridinium tetrafluoroborates, 9-heterocyclic amino-N-methyl-9,10dihydroacridine) Acalypha alnifolia Acetamide derivatives of chromen-2-ones Acetohydroxamate-chalcones Acridone-1,2,4-oxadiazole-1,2,3-triazole hybrids Acrine-ebselen hybrids Adamantyl derivatives Albumin-derived peptides Alkaloid extracts from kola nuts (Cola acuminata and Cola nitida) Alkaloids from Coptis chinensis (quaternary protoberberines) Alkaloids from Crinum, Habranthus, and Zephyranthes species (caranine, N-desmethylgalanthamine, lycoramine) Alkaloids from Hieronymiella species (Hieronymiella aurea, H. caletensis, H. clidanthoides, H. marginata, and H. speciosa) Alkaloids of Amaryllidaceae (extracts of Crinum jagus: galanthamine, galanthamine N-oxide, powelline) Alkyl bis(4-amino-5-cyanopyrimidine) derivatives Allium sativum essential oil Allobetulin-derived seco-oleananedicarboxylates Aminoalkyl-substituted coumarin derivatives Amino-alkyl-substituted fluoro-chalcones derivatives Andrographis paniculata and Spilanthes paniculata components (3,4-di-o-caffeoylquinic acid, apigenin, 7-o-methylwogonin) Anethole Annona cherimola Mill. derivatives (anonaine, glaucine, xylopine) Anthemis cotula L. extracts Anthocleista vogelii (Planch) extracts Anthraquinone from Rumex abyssinicus Jacq (Helminthosporin) Aporphines Aporphinoid alkaloids (dual) (liriodenine, cassythicine, laurotetatine clorhydrate, pachyconfine) Arachidonic acid carbamate derivatives (continued)

Pharmacogenomics of Alzheimer’s Disease

291

Table 2 (continued) Arylbenzofurans from the root bark of Morus alba Arylimides Aoaanthraquinone derivatives Azomethine-dihydroquinazolinone conjugates Bacopa monnieri derivatives (bacoside X, bacoside A, 3-beta-D-glucosylstigmasterol, daucosterol) Bacopa monnieri L. Wettst., Centella asiatica L. Urb. Benzamide and picolinamide derivatives Benzamide and picolinamide derivatives containing dimethylamine side chain Benzamide derivatives Benzimidazole derivatives Benzimidazole derivatives Benzimidazole-derived compounds Benzochromenoquinolinones-tacrine analogs (14-amino-13-(3-nitrophenyl)-2,3,4,13-tetrahydro-1Hbenzo[6,7]chromeno[2,3-b]quinoline-7,12-dione) Benzodiazepine-1,2,3-triazole hybrid derivatives (3,3-Dimethyl-11-(3-((1-(4-nitrobenzyl)-1H-1,2,3triazol-4-yl)methoxy)phenyl)-2,3,4,5,10,11-hexahydro-1H-dibenzo[b,e][1,4]diazepin-1-one) Benzofuran-based chalconoids Benzofuran-derived benzylpyridinium bromides Benzofuran-indole Benzofuran-tetrazole derivatives Benzohomoadamantane-chlorotacrine hybrids Benzoic acid-derived nitrones Benzothiazole/piperazine derivative (2-[(6-nitro-2-benzothiazolyl)amino]-2-oxoethyl4-[2-(N, N-dimethylamino)ethyl] piperazine-1 carbodithioate) Benzothiazole/piperazine derivatives (2-[(6-nitro-2-benzothiazolyl)amino]-2-oxoethyl4-[2-(N, N-dimethylamino)ethyl] piperazine-1 carbodithioate) Benzoxazine (indole-benzoxazinones and benzoxazine-arylpiperazine derivatives) Benzylidene-benzofuran-3-ones containing cyclic amine side chain Benzylpiperidine/piperazine-dantrolene Benzylpiperidine-linked 1,3-dimethylbenzimidazolinones (1,3-dimethylbenzimidazolinone derivatives) Bergenia ciliata (Haw.) Sternb. rhizomes (Saxifragaceae) Betula platyphylla var. japonica Betulinic acid bi-Aryl Pyrimidine heterocycles (continued)

292

Ramo´n Cacabelos et al.

Table 2 (continued) Binary conjugates of γ-carbolines Biphenyl-3-oxo-1,2,4-triazine linked piperazine derivatives Bis(heptyl)-cognitin (B7C) Bis(3)-tacrine Bis(7)-tacrine and cystamine-tacrine dimer Bis(9)-(-)-meptazinol Bis[5-(10 ,20 ,30 ,40 -tetrahydroacridin-90 -ylamino)pentyl]disulfide Bis-sulfone derivatives Bivalent β-carboline derivatives Boana pulchella (Anura: Hylidae) derivatives (Hp-1935) Brominated 2-phenitidine derivatives Brominated phenylacetic acid/tacrine hybrids Bromophenol derivatives Broussonin A Buchanania axillaris, Hemidesmus indicus, and Rhus mysorensis extracts BYZX, [(E)-2-(4-((diethylamino)methyl)benzylidene)-5,6-dimethoxy-2,3-dihydroinden-one] Caliphruria subedentata (Amaryllidaceae) derivatives Carbamate derivatives (bambuterol derivatives (BMC-3, BMC-16)) Carbamate-substituted thymol/carvacrol derivatives Carbazole-bearing oxazolones Carbazole-bearing oxazolones Carbazole-benzyl piperazine, carbazole-benzyl piperidine, carbazole-pyridine, carbazole-quinoline, carbazole-isoquinoline Carboxamide and propanamide derivatives Carvacrol Carvacrol-substituted amide derivatives (2-(5-Isopropyl-2-methylphenoxy)-N-(quinolin-8-yl) acetamide) Cativic acid derivatives Centaurea saligna (K.Koch) Wagenitz Ceratonia siliqua L. extract (Rutin) Chalcone Mannich base derivatives Chalcone, heterochalcone and bis-chalcone derivatives Chalcone-based carbamates (continued)

Pharmacogenomics of Alzheimer’s Disease

293

Table 2 (continued) Chalcone-O-alkylamine derivatives Chalcone-O-carbamate derivatives Chilean Rhodophiala (Amaryllidaceae) Chitosan (chitooligosaccharide) Chlorinated tacrine analogs (4-(chlorophenyl)tetrahydroquinoline derivatives Chlorophenoxy derivatives-histamine H3 receptor ligands Cholinesterase inhibitor 6-chlorotacrine Chroman-4-one linked to N-benzyl pyridinium derivatives Chromanone-dithiocarbamate hybrids Chromen-4-one based compounds Chromenone derivatives Chromones CID57390505 CID71605390 Cinnamic acid-tryptamine hybrids Citrus flavanones (hesperidin) Citrus Limon peel derivatives (neoeriocitrin, isonaringin, naringin, hesperidin, neohesperidin, limonin) Clinanthus microstephium (Amaryllidaceae) derivatives Conjugates of γ-carbolines, carbazoles, tetrahydrocarbazoles, phenothiazines, and aminoadamantanes Contilisant Corydalis cava (Papaveraceae) alkaloids ((+)-thalictricavine and (+)-canadine) Coumarin derivatives Coumarin derivatives Coumarin derivatives (3-(4-aminophenyl)-coumarin derivatives, Coumarin-3-carboxamide-Nmorpholine hybrids: N-[3-(morpholin-4-yl)propyl]-2-oxo-2H-chromene-3-carboxamide, 6-bromoN-[2-(morpholin-4-yl)ethyl]-2-oxo-2H-chromene-3-carboxamide; 3-aryl coumarin derivatives: 3-(3,4-dichlorophenyl)-7-[4-(diethylamino)butoxy]-2H-chromen-2-one Coumarin hybrids Coumarin linked thiourea derivatives Coumarin-3-carboxamides bearing N-benzylpiperidine moiety Coumarin-3-carboxamides bearing tryptamine moiety Coumarin-benzofuran hybrids Coumarin-chalcone hybrids Coumarin-dithiocarbamate hybrids (continued)

294

Ramo´n Cacabelos et al.

Table 2 (continued) Coumarin-dithiocarbamate hybrids Coumarin-piperazine derivatives Coumarins from Zosima absinthifolia Link (Apiaceae family) (coumarin-bergapten, imperatorin, pimpinellin, umbelliferone) Coumarinyl thiazoles and oxadiazoles Coumestrol and Puerarol from Pueraria lobate Cubebin (dibenzylbutyrolactone lignan from Piper cubeba) Cyclopentaquinoline analogs (6-chloro-N-[2-(2,3-dihydro-1H-cyclopenta[b]quinolin-9-ylamino)hexyl]]-nicotinamide hydrochloride) Cyclopentaquinoline hybrids Dehydroabietylamine derivatives Deoxyvasicinone-donepezil hybrids Dialkyl-3-cyanopropylphosphate derivatives Diaryl pyrazoline derivatives Dichocarpum auriculatum (Franch.) W.T. Wang & P. K. Hsiao derivatives (columbamine, palmatine, dauricine, jatrorrhizine, berberine) Difluoropyrido[4,3-b]indoles Dihydroactinidiolide Dihydroberberine (Coptis chinensis) Dihydroquinoline carbamate derivatives Dihydroxanthyletin-type coumarins from Angelica decursiva (decursidin) Dimethyl sulfoxide (DMSO) Dimethyl sulfoxide (DMSO) Diphenyl substituted pyridazinone derivatives (diphenyl-2-(2-(4-substitutedpiperazin-1-yl)ethyl) pyridazin-3(2H)-one derivatives) Dipropargyl substituted diphenylpyrimidines Dispiropyrrolodinyl-oxindole based alkaloids DL0410 Donepezil analogs Donepezil-butylated hydroxytoluene (BHT) hybrids Donepezil-chromone-melatonin hybrids Donepezil-flavonoid hybrids with sigma-1 affinity (N-(2-(1-benzylpiperidin-4-yl)ethyl)-6,7-dimethoxy4-oxo-4H-chromene-2-carboxamide) Donepezil-ginkgo ketoester (continued)

Pharmacogenomics of Alzheimer’s Disease

295

Table 2 (continued) Donepezil-hydrazinonicotinamide hybrids Donepezil-like 1,4-dihydropyridines Donepezil-related benzylpiperidine/benzylpiperazine derivatives (benzimidazole or benzofuran) EGb761 constituents Elatostema papillosum leaves Ellagic acid Embelin (3-undecyl-1,4-benzoquinone from Embelia ribes) Erucamide from radish leaf (Raphanus sativus L.) extracts Ester derivatives of cinnamic acid (p-coumaric acid 3,4-dihydroxyphenethyl ester, p-coumaric acid phenethyl ester) Ethyl 5-amino-4-(2-chloroquinolin-3-yl)-2-methyl-6,7,8,9-tetrahydro-4H-pyrano[2,3-b]quinoline-3carboxylate Ethyl acetate extract of Terminalia chebula (N-(N-benzyloxycarbonyl-beta-l-aspartyl)-beta-dglucosaminide) Ethynylphenyl carbonates and carbamates Fascaplysin Fasudil hydrochloride Ferulic acid derivatives Flavanone glycosides (naringenin, didymin, prunin, poncirin) Flavones, chromen-4-ones, and C-glucosyl derivatives (p-morpholinyl flavones, C-glucosylflavones) Flavonoid compounds (galangin, N,N-dibenzyl(N-methyl)amine hybrids) Flavonoid derivatives Flavonoids Flavonoids (myricetin, morin, rutin, quercetin, fisetin, kaempferol, apigenin, glycitein) Flavonoids from Atalantia monophylla (L.) DC (atalantraflavone, atalantoflavone, racemoflavone, 5,40 -dihydroxy-(300 ,400 -dihydro-300 ,400 -dihydroxy)-200 ,200 -dimethylpyrano-(500 ,600 :7,8)-flavone, lupalbigenin, anabellamide, citrusinine I, p-hydroxybenzaldehyde, frideline) Flavonols and 4-thioflavonols Flurbiprofen-chalcone hybrid Mannich base derivatives Forsythiaside Fused thiazolo[3,2-a]pyrimidines Galantamine derivatives Galanthamine-peptide derivatives Gelidiella acerosa (continued)

296

Ramo´n Cacabelos et al.

Table 2 (continued) Genipin derivatives Genistein-O-alkylamine derivatives Ginger (Zingiber officinale) extract ((E)-1,7-bis(4-hydroxy-3-methoxyphenyl)hept-4-en-3-on, 1-(3,4-dihydroxy-5-methoxyphenyl)-7-(4-hydroxy-3-ethoxyphenyl) heptane-3,5-diyl diacetate) Ginger root (Zingiber officinale Roscoe) extract Ginkgo biloba extract, EGb 761® Ginseng stem-leaf saponins Glycosides from Cassia obtusifolia (rubrofusarin, rubrofusarin 6-O-β-d-glucopyranoside, rubrofusarin 6-O-β-d-gentiobioside, nor-rubrofusarin 6-O-β-d-glucoside, isorubrofusarin 10-O-β-dgentiobioside, rubrofusarin 6-O-β-d-triglucoside) Glycyrrhizin Guanylhydrazones Guazuma ulmifolia Halogenated thiophene chalcones Hancornia speciosa Gomes (Brazil mangaba tree) Haplophyllum sahinii and H. vulcanicum extracts Harmaline Harmine Hemp seed protein-derived peptides Heracleum verticillatum, Heracleum sibiricum, Heracleum angustisectum, and Heracleum ternatum extracts Heterocyclic β-d-gluco- and β-d-galactoconjugates Heterometallic ruthenium(II)-Platinum(II) polypyridyl complexes Hibiscus sabdariffa L. (Sorrel) Calyx HuperTacrines Huperzine A derivatives Hydroethanolic extracts of Ecklonia maxima, Gelidium pristoides, Gracilaria gracilis, and Ulva lactuca Hydroxybenzoic acid derivatives Hydroxyoxindoles Hydroxypyridinone-benzofuran hybrids ((3-hydroxy-4-pyridinone)-benzofuran hybrids, O-benzylhydroxypyridinone hybrids) Hyperforin Hypericum perforatum L. prenylated β-diketones (2,6,9-trimethyl-8-decene-3,5-dione, 3,7,10trimethyl-9-undecene-4,6-dione) Hypsiboas cordobae and Pseudis minuta (Anura: Hylidae) (continued)

Pharmacogenomics of Alzheimer’s Disease

297

Table 2 (continued) Hyuganin C Imidazo pyranotacrines Imidazole analogs (2-substituted-4,5-diphenyl-1H-imidazole) Indazolylketones Indole-3-acetic acid (IAA)-tacrine hybrids Indoloquinoline alkaloid cryptolepine Indolotacrine analogs Iridoid glucosides from Anarrhinum pubescens (6-O-, 60 -O-di-trans-cinnamoyl-antirrhinoside, 5-O-, 6-O-difoliamenthoyl-antirrhinoside, 6-O-foliamenthoyl-(60 -O-cinnamoyl)-antirrhinoside) Irvingia gabonensis (Aubry-Lecomte ex O’Rorke) Baill bark Isoalloxazine derivatives Isoflavone analogs Isoindoline-1,3-dione derivatives Isotalatizidine hydrate Isothio- and isoselenochromanone derivatives bearing N-benzyl pyridinium moiety Isothiocyanates Kampo products (kihito (Gui-Pi-Tang), kamikihito (Jia-Wei-Gui-Pi-Tang)) KojoTacrines (11-amino-2-(hydroxymethyl)-12-(3-methoxyphenyl)-7,9,10,12-tetrahydropyrano [20 ,30 :5,6] pyrano[2,3-b]quinolin-4(8H)-one) Lactucopicrin Ligustrazine derivatives Limonium brasiliense Linarin (flavonoid glycoside in Flos chrysanthemi indici) L-Tetrahydropalmatine Macelignan (Myristica fragrans) Marine-derived Streptomyces sp. UTMC 1334 Matrine MCULE-5872671137-0-1 Memantine-6-chlorotacrine Metal chelator Metformin-sulfonamide derivative Methanesulfonate derivatives Methanosarcina species-derived phenazines (continued)

298

Ramo´n Cacabelos et al.

Table 2 (continued) Methyl(4-phenylbutyl){2-[2-(4-propylpiperazin-1-yl)-1,3-thiazol-5-yl]ethyl}amine Millettia pachycarpa Benth flavonoids MLC901 (Neuroaid II) Mono- and bis-spirooxindole-hexahydropyrrolidines Mono- and bis-spiro-pyrrolidines Mulberry diels-alder-type adducts from Morus alba (mulberrofuran C, mulberrofuran K, mulberrofuran G, isomulberrofuran G) Multitarget tacrine derivatives N-((1-(2-Chlorobenzyl)-1H-1,2,3-triazol-5-yl)methyl)-8-methoxy-2-oxo-2H-chromene-3carboxamide N-(4-Methoxyphenethyl)-N-(substituted)-4-methylbenzenesulfonamides N-(Alkyl/aralkyl)-N-(2,3-dihydro-1,4-benzodioxan-6-yl)-4-chlorobenzenesulfonamides (N-(2,3-dihydro-1,4-benzodioxan-6-yl)-4-chlorobenzenesulfonamide, N-2-phenethyl-N(2,3-dihydro-1,4-benzodioxin-6-yl)-4-chlorobenzenesulfonamide, N-(1-butyl)-N-(2,3-dihydro-1,4benzodioxin-6-yl)-4-chlorobenzenesulfonamide) N-{2-[4-(1H-Benzimidazole-2-yl)phenoxy]ethyl}substituted amine derivatives N1-Substituted derivatives of 2,3-dihydroquinazolin-4(1H)-one N-Allyl/propargyl tetrahydroquinolines Nanodandelion tin(IV) complex-carbacylamidophosphate (Sn(CH3)2Cl2}NC5H4C(O)NHP (O)[NHC6H11]2}2) Narcimatuline Narcissus poeticus cv. Pink Parasol N-Benzyl pyridinium N-Benzylpiperidine analogs N-Benzylpiperidine carboxamide derivatives N-Benzylpiperidine derivatives N-Benzyl-piperidinyl-aryl-acylhydrazone derivatives N-Benzylpyridinium-based benzoheterocycles Neonicotinoids Nepitrin from Rosmarinus officinalis N-Phenylcarbamates N-Phenylthiazol-2-amine derivatives N-Phosphorylated/N-tiophosphorylated tacrine N-Substituted aryl sulfonamides N-Substituted pyrazole-derived α-aminophosphonates (continued)

Pharmacogenomics of Alzheimer’s Disease

299

Table 2 (continued) N-Substituted rhodanines o/p-Propoxyphenylsubstituted-1H-benzimidazole derivatives Ochna obtusata Octohydroaminoacridine Oenothera biennis L. Olaparib analogs Olive biophenols from (Olea europaea L.) (rutin, verbascoside) Onion flavonols Onobrychis argyrea subsp. isaurica extracts Organoselenium compounds from purines (6-arylselanylpurines, 6-((4-fluorophenyl)selanyl)-9Hpurine) Origanum onites extracts Ovalbumin-derived peptides Oxazolidinone Oxoisoaporphine-tetrahydroisoquinoline hybrids Oxopyrrolidine derivatives (ethyl 2-(2-(2, 6-dimethylphenylcarbamoyl)-5-oxopyrrolidin-1-yl) acetate, 1-((4-(4-chlorophenyl) piperazin-1-yl) methyl)-N-(2,6-dimethylphenyl)-5-oxopyrrolidine-2carboxamide) p-Aminobenzoic acid derivatives Panax ginseng derivatives Papaver somniferum Papaverine Paullinia cupana Pavetta indica PC-37 and PC-48 (7-methoxytacrine-donepezil-like compounds) Peganum harmala Linn ingredients (deoxyvasicine) Pharbitis nil Choisy Phenanthroline-tetrahydroquinolines Phenolic extracts from Irvingia gabonensis (Aubry-Lecomte ex O’Rorke) Baill bark Phlorotannins from Ecklonia cava (eckol, dieckol, 8,80 -bieckol) Phosphazine and phosphazide derivatives Phthalimide-alkylamine derivatives Phthalimide-derived N-benzylpyridinium halides Phthalimide-dithiocarbamate hybrids (continued)

300

Ramo´n Cacabelos et al.

Table 2 (continued) Physostigma venenosum Balf. Pinoresinol Piperazine derivatives (1-(1,4-benzodioxane-2-carbonyl) piperazine, 4-(4-methyl)-benzenesulfonyl-1(1,4-benzodioxane-2-carbonyl) piperazine, 4-(4-Chloro)-benzenesulfonyl-1-(1,4-benzodioxane-2carbonyl) piperazine) Piperazinediones (3,6-diphenyl-1,4-bis(phenylsulfonyl)piperazine-2,5-dione derivatives) Piperidine and piperazine derivatives Piperidone grafted spiropyrrolidines Piperonyl-based 4-thiazolidinone derivatives Pistacia khinjuk essential oil Platanus orientalis L. extracts Poincianella pluviosa Polycyclic polyprenylated acylphloroglucinols from Hypericum perforatum Polyphenol-rich Boswellia serrata gum Polyphenols (quercetin, resveratrol, curcumin, gallocatechins, cinnamic acid, caffeine, caffeic acid) Polysaccharide extracts from the fungi Coprinus comatus and Coprinellus truncorum Prangos ferulacea extracts Propargylamine-derived compounds Propargylamine-modified 4-aminoalkyl imidazole substituted pyrimidinylthiourea derivatives Propargylamine-modified pyrimidinylthiourea derivatives Prunella vulgaris var. lilacina NAKAI Pterosin derivatives from Pteridium aquilinum ((2R)-Pterosin B, (2R,3 R)-Pteroside C) Pterostilbene β-amino alcohol derivatives Pulmonarin B Pyridazinone derivatives (6-substituted-3(2H)-pyridazinone-2-acetyl-2-(nonsubstituted/4-substituted benzenesulfonohydrazide) derivatives) Pyranopyrazolotacrines Pyranotacrines (Benzo)Chromeno-PyranoTacrines (11-amino-12-(3,4,5-trimethoxyphenyl)7,9,10,12-tetrahydro-8H-chromeno[2,3-b]quinolin-3-ol, 14-(3,4-dimethoxyphenyl)-9,11,12,14tetrahydro-10H-benzo[5,6]chromeno[2,3-b]quinolin-13-amine) Pyrazole and spiropyrazoline analogs Pyrazoline derivatives (1-aroyl-3-(5-(4-chlorophenyl)-1,2,4-triazole-3-thioneaminylthioureas, 4-amino-5-(4-chlorophenyl)-4H-1,2,4-triazole-3-thiol, 1-(3-(4-aminophenyl)-5-phenyl-4,5dihydro-1H-pyrazol-1-yl)-2-(4-isobutylphenyl)propan-1-one, 2-(4-isobutylphenyl) propanehydrazide) Pyrazolo[1,5-c][1,3]benzoxazin-5(5H)-one derivatives (continued)

Pharmacogenomics of Alzheimer’s Disease

301

Table 2 (continued) Pyridine amine derivative Pyridinium/isoquinolium derivatives Pyrido[2,3-b]pyrazine derivatives PyridoTacrines Pyrrolizine-based compounds Quercetin-metal complexes Quinoline-chalcone derivatives Quinoline-ferulic acid hybrids Quinoline-piperonal hybrids Quinolinone-dithiocarbamate derivatives Quinolone-benzylpiperidine derivatives Quinolone-triazole hybrids Quinoxaline derivatives QuinoxalineTacrine QT78 Racemic chromenotacrines (11-amino-12-aryl-8,9,10,12-tetrahydro-7H-chromeno[2,3-b]quinolin-3ols) Rivastigmine derivatives Rutaecarpine derivatives Rutin Rutin from Dianthus calocephalus Boiss S 47445 (α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid) Safflower (Carthamus tinctorius L.) seed oil cake ((
)-carthatins A–F, sinapyl alcohol, coniferyl alcohol, serotobenine, feruloylserotonin) Sailuotong Salicylamide derivatives Salvia miltiorrhiza components (miltirone, salvianolic acid A) Sargachromanol I and G (Sargassum siliquastrum) Saururus chinensis (Lour.) Baill. Schisandra chinensis (Chinese magnolia vine) (dibenzocyclooctadiene lignans, 6-O-benzoylgomisin, deoxyschisandrin, gomisin A, gomisin G, schisandrin, schisandrin C, schisanhenol, schisantherin A, and schisantherin B) Scrophularia lucida L. derivatives (rosmarinic acid, hesperidin) Semicarbazones ((E)-2-(5,7-dibromo-3,3-dimethyl-3,4-dihydroacridin-1(2H)-ylidene) hydrazinecarbothiomide, (E)-2-(5,7-Dibromo-3,3-dimethyl-3,4-dhihydroacridin-1(2H)-ylidene) hydrazinecarboxamide) (continued)

302

Ramo´n Cacabelos et al.

Table 2 (continued) Semisynthetic O-alkylcoumarin derivatives Sideritis albiflora and Sideritis leptoclada extracts Silymarin Spirocyclohexadienones Spiro-dioxolane-containing meroterpenoids from Aspergillus terreus Thom Spiropyrrolidine/spiroindolizino[6,7-b]indole heterocyclic hybrids Streblus asper leaf ethanolic extract Stryphnodendron adstringens Substituted 4-methyl-2-oxo-2H-chromen-7-yl phenyl carbamates Substituted aminobenzothiazole derivatives of tacrine Substituted chromones (3-cyanochromone, 7-amino-3-methylchromone) Sulfamate derivatives of menthol Synthetic chalcones Tacrine derivatives Tacrine (10)-hupyridone Tacrine-(β-carboline) hybrids Tacrine, phenolic acid, and ligustrazine hybrids Tacrine/acridine anticholinesterase inhibitors with piperazine and thiourea linkers Tacrine-1,2,3-triazole derivatives Tacrine-1,2,4,-thiadiazole derivatives Tacrine-4-oxo-4H-chromene hybrids Tacrine-acridine hybrids Tacrine-based pyrano[2,3-c]pyrazoles Tacrine-based pyrano[30 ,40 :5,6]pyrano[2,3-b]quinolinones Tacrine-benzimidazole-based human cannabinoid receptor subtype 2 agonist Tacrine-benzofuran hybrids Tacrine-benzofuran hybrids Tacrine-benzyl quinolone carboxylic acid hybrids Tacrine-bifendate conjugates Tacrine-cinnamic acid hybrids Tacrine-coumarin and tacrine-7-chloroquinoline hybrids with thiourea linkers Tacrine-coumarin hybrids Tacrine-coumarin hybrids linked to 1,2,3-triazole (continued)

Pharmacogenomics of Alzheimer’s Disease

303

Table 2 (continued) Tacrine-deferiprone hybrids Tacrine-ferulic acid hybrids Tacrine-hydroxyphenylbenzimidazole hybrids Tacrine-multialkoxybenzene hybrids Tacrine-phenylbenzothiazole Tacrine-propargylamine derivatives Tacrine-pyranopyrazole derivatives Tacrine-pyrazolo[3,4-b]pyridine hybrids Tacrine-scutellarin hybrids Tacrine-squaramide derivatives Tacrine-tianeptine hybrids Tacrine-trolox hybrids Tacrine-tryptophan hybrids (S-K1035) Tacrine-valmerin hybrids Tacripyrimidines (tacrine-dihydropyrimidine hybrids) (5-amino-4-aryl-3,4,6,7,8,9-hexahydropyrimido [4,5-b]quinoline-2(1H)-thiones) Tacripyrines (40 -metoxytacripyrine (S)-ITH122) Tadalafil derivatives Tannic acid Tanshinones and salvianolic acids from Salvia miltiorrhiza Tea polyphenols Termitomyces albuminosus sesquiterpenoids Terpenes and phenylpropanoids (methyl jasmonate; 1R-()-nopol; 1,4-cineole; allo-aromadendrene; nerolidol; β-ionone; and (R)-(+)-pulegone) Terpenoids from I. wightii (Bentham) H. Hara Tetraclinis articulata essential oil Tetrahydroacridine derivatives (thiocyanates, selenocyanates, ureas, selenoureas, thioureas, isothioureas, disulfides, diselenides, and several tacrine homo- and hetero-hybrids) Tetrahydroacridine derivatives with dichloronicotinic acid Tetrahydrocarbazole benzyl pyridine hybrids Tetrahydrochromeno[30 ,40 :5,6]pyrano[2,3-b]quinolin-6(7H)-one derivatives Tetrahydrocurcuminoid dihydropyrimidinone analogs Tetrahydropyranodiquinolin-8-amines Tetrahydroquinoline-isoxazole/isoxazoline hybrid compounds (continued)

304

Ramo´n Cacabelos et al.

Table 2 (continued) Tetraphenylporphinesulfonate (TPPS) Tetrasubstituted thiazoles Tetrazole-1,2,5,6-tetrahydronicotinonitriles Thiazole acetamide derivatives Thiazole-substituted benzoylpiperazine derivatives Thiazolylhydrazone derivatives Thymohydroquinone Trachyloban-19-oic acid from Syncephalastrum racemosum Trans-tephrostachin from Tephrosia purpurea (L.) Pers. Trianthema portulacastrum phenols Triazole derivatives Triazole-quinoline derivatives Triazolothiadiazole and triazolothiadiazine Trichilia catigua Tris-chalcones Tris-chalcones Uracil derivatives Ursolic and oleanolic acid derivatives Usnic acid (active dibenzofuran derivative) Vasicine enantiomers VB-037 Vilazodone-tacrine hybrids Vitis vinifera L. flavones Withania somnifera L. Dunal. Xanthene derivatives (spiro[indoline3,9-xanthene]-trione, hydroxy-spiro[indoline-3,9-xanthene]trione) Xanthohumol, naringenin, and acyl phloroglucinol derivatives from Humulus lupulus L. Yinhuang oral liquid ZT-1, a prodrug of huperzine A α- and β-asarone α-Pinene β-Amino di-carbonyl derivatives (continued)

Pharmacogenomics of Alzheimer’s Disease

305

Table 2 (continued) β-Crinane amaryllidaceae alkaloid haemanthamine derivatives (11-O-(2-methylbenzoyl)haemanthamine, 11-O-(4-nitrobenzoyl)-haemanthamine) δ-3-Carene δ-Sultone-fused pyrazoles Selective butyrylcholinesterase inhibitors (3-(4-Phenyl-piperazin-1-ylmethyl)-phenyl phenylcarbamate [3-(2/3/4-Methoxyphenyl)-6-oxopyridazin-1(6H)-yl]methyl carbamate derivatives ([3-(2-methoxyphenyl)-6-oxopyridazin-1(6H)-yl]methyl heptylcarbamate) [3-(2-Methoxyphenyl)-6-oxopyridazin-1(6H)-yl]methyl (4-methylphenyl)carbamate heterocyclic β-dgluco- and β-d-galactoconjugates 1,2,3,4,5,6-Hexahydroazepino[4,3-b]indole 2-Pheynlbenzofuran derivatives 2-Vinyl chromones Aryl-1,2,3-triazolyl benzylpiperidine inhibitors Benzamide derivatives Benzofuran appended benzothiazepine derivatives Carbazole-based BChEIs (carbazol-9-yl)methyl)-1-(4-chlorobenzyl)pyridin-1-ium chloride) Coumarin/1,2,4-oxadiazole hybrids Ferulic and dihydrocaffeic acids N-Phthaloylglycine Tricyclic pyrazolo[1,5-c][1,3]benzoxazin-5(5H)-one Tricyclic pyrazolo[1,5-d][1,4]benzoxazepin-5(6H)-one scaffold derivatives Tryptophan-derived inhibitors

5

Immunotherapy Since the pioneering studies of Schenk and coworkers in 1999 [56], several categories of vaccines and immunotherapeutic procedures have been developed for the treatment of AD based on the two most prevalent pathogenic hypotheses (amyloidosis vs. tauopathy) (Table 4). Two decades later, with millions of dollars invested in passive and active immunotherapy in experimental AD models and in clinical trials, and over 1000 papers published on AD immunotherapy (40–50 papers/year), about 140 (85%) immunization procedures against Aβ deposition and 25 (15%) against tau have been reported, until the FDA approval of aducanumab as an immunotherapeutic strategy in mild-AD (Table 4). However, most meta-

306

Ramo´n Cacabelos et al.

Table 3 Categories of drugs in development for the treatment of Alzheimer’s disease Anti-amyloid immunotherapy 4Aβ1–15-derived monoclonal antibody 5Aβ3–10 6Aβ15-T-Hc chimeric DNA vaccine (rCV02) AAB-003 (PF-05236812) ACC-001 (vanutide cridificar) Aducanumab Advax(CpG)-adjuvanted MultiTEP-based dual and single vaccines AFFITOPE® AD02 Anti-amyloid-β monoclonal antibody 4G8 Anti-Aβ31–35 antibody Antibody NT4X Anti-pyroglutamate-3 Aβ vaccine AOE1 vaccine AV-1955 AV-1959 Aβ B-cell epitope vaccine (rCV01) Aβ3–10-KLH vaccine AβpE3:CRM197 vaccine BAN2401 Bapineuzumab (AAB-001) and its derivative (AAB-003) CAD106 Crenezumab DNA Aβ42 vaccine DNA vaccine p(Aβ3–10)10-C3d-p28.3 EB101 vaccine Epitope vaccine (Lu AF20513) Fc-inactivated anti-β-amyloid monoclonal antibody GSK933776 Gammagard IVIg Gantenerumab Humanized anti-PrP antibody Ixekizumab (continued)

Pharmacogenomics of Alzheimer’s Disease

307

Table 3 (continued) mAb (3F5) MER5101, Aβ1–15:DT conjugate vaccine Multiple antigenic peptide (MAP)-based epitope vaccine Multivalent Aβ3–10 DNA vaccine NoV P particle-based chimeric protein vaccine Dual vaccine WSN-Aβ(1–10) Peptide epitope vaccine Aβ3–10-KLH Peptide vaccine targeting AβpE3-X Novel single-domain antibodies (VHHs or nanobodies) Aβ1–43 cDNA (rAAV/Aβ) Octagam IVIg PF-0436036 Ponezumab (PF-04360365) Pyroglutamate-3 Aβ (AβpE3-X) vaccine ScFv-h3D6 Solanezumab Tetravalent Aβ1–15 vaccine Tocilizumab UB-311 Vaccine vanutide cridificar (ACC-001) with QS-21 adjuvant β-Secretase (BACE) inhibitors (3S,4S)-4-Aminopyrrolidine-3-ol derivatives 1,2,4-Thiadiazole analogs 1,3-Oxazines 1,3,4,4a,5,10a-Hexahydropyrano[3,4-b]chromene analogs 1,4-Oxazine β-secretase 1 (BACE1) inhibitors 2,2,2-Trifluoroethyl-thiadiazines 2-Aminooxazoline 3-azaxanthenes 2-substituted-thio-N-(4-substituted-thiazol/1H-imidazol-2-yl)acetamides 3,4-Dihydro-1,3,5-triazin-2(1H)-ones 3,5-bis-N-(Aryl/heteroaryl) carbamoyl-4-aryl-1,4-dihydropyridines 3AA and XYT472B 3-Hydroxyhericenone F (continued)

308

Ramo´n Cacabelos et al.

Table 3 (continued) 3-Imino-1,2,4-thiadiazinane 1,1-dioxide derivative verubecestat (MK-8931) 4-Oxo-1,4-dihydro-quinoline-3-carboxamides 6,8-Dioxa-3-azabicyclo[3.2.1]-octane peptidomimetics 7,8-Dihydroxyflavone glycoside derivatives 7-Amino-1,4-dihydro-2H-isoquilin-3-one derivatives Acylguanidine Allylidene hydrazinecarboximidamide derivatives AMG-8718 Aminomethyl-derived beta-secretase (BACE1) inhibitors Aminopyridine derivatives Arylketo-containing P1-P3 linked macro-cyclic BACE-1 inhibitors Asperterpenes A and B AZ-4217 AZD3293 (LY3314814) AZD3839 Bergenin analogs Biphenylacetamide-derived inhibitors Breviscapine Cyclic sulfone hydroxyethylamines DNA aptamer E2609 Furo[2,3-d][1,3]thiazinamines Gastrodin GNE-892, (R)-2-amino-1,30 ,30 -trimethyl-70 -(pyrimidin-5-yl)-30 ,40 -dihydro-20 H-spiro[imidazole4,10 -naphthalen]-5(1H)-one GRL-8234 Heparan sulfate hexa- to dodecasaccharides Heparin oligosaccharides Hydroxyethylamine Imidazopyridines containing isoindoline-1,3-dione Indole acylguanidines KMI-1764 Lanabecestat (AZD3293) (continued)

Pharmacogenomics of Alzheimer’s Disease

309

Table 3 (continued) Leptin LY2886721 Macrocyclic prolinyl acyl guanidines Myo-inositol hexakisphosphate (IP6) N-(2-(Piperazin-1-yl)phenyl)arylamide derivatives N-(3-(2-Amino-6,6-difluoro-4,4a,5,6,7,7a-hexahydro-cyclopenta[e][1,3]oxazin-4-yl)-phenyl)-amides Naphthofurans NB-360 Compound VIa Peptide-mimetic BACE1 inhibitors Phenylimino-2H-chromen-3-carboxamide derivatives Quinazolinone-based hydrazones Salvianolic acid sAβPPα Spirocyclic pyranochromene Substituted 2-oxopiperazines Sulfonyl-amino-acetamides Triptolide Verubecestat (MK-8931) WY-25105 γ-Secretase inhibitors/modulators 2-Methylpyridine-based biaryl amides 5,6,7,8-Tetrahydro[1,2,4]triazolo[4,3-a]pyridine derivatives 5-Lipoxygenase Aminopiperidines Aminothiazole Anilinotriazoles AS2715348, cyclohexylamine-derived γ-secretase modulator, {(1R*,2S*,3R*)-3-[(cyclohexylmethyl) (3,3-dimethylbutyl)amino]-2-[4-(trifluoromethyl)phenyl]cyclohexyl}acetic acid Avagacestat Begacestat (GSI-953) Benzenesulfonamides Benzoazepinone (continued)

310

Ramo´n Cacabelos et al.

Table 3 (continued) BIIB042 BMS-869780 BMS-932481 BPN-15606 CHF5074 Cholestenoic acid Cyclopamine E2012-BPyne FRM-36143 Indole-derived pyridopyrazine-1,6-dione L685,458 LY450139 MH84, (2-((4,6-bis(4-(trifluoromethyl)phenethoxy)pyrimidin-2-yl)thio)hexanoic acid) MRK-560 NGP 555 Oxadiazine Phenylimidazole-type γ-secretase modulator Piperazine derivatives Pyridazine and pyridine-derived γ-secretase modulators Pyridopyrazine-1,6-dione Semagacestat SCH 697466 Soluble β-aminosulfone analogs of SCH 900229 SPI-1865 Spirocyclic sulfones Substituted 4-morpholine N-arylsulfonamides Triazolobenzazepinones Tricyclic gamma-secretase modulators Other anti-Aβ agents (D-Ser2) Oxm (E)-N-(Yyridin-2-ylmethylene)arylamine (LR) (R)-α-Trifluoromethylalanine containing short peptide (RI-OR2-TAT, Ac-rGffvlkGrrrrqrrkkrGy-NH(2)) (continued)

Pharmacogenomics of Alzheimer’s Disease

311

Table 3 (continued) (Z)-3-(2-(3-Chlorophenyl)hydrazono)-5,6-dihydroxyindolin-2-one (PT2) [Gly14]-Humanin 1-(3-(Benzyloxy)pyridin-2-yl)-3-(2-(piperazin-1-yl)ethyl)urea 1,2-(Dimethoxymethano)fullerene (DMF) 17β-Hydroxysteroid dehydrogenase type 10 1α,25-Dihydroxyvitamin D3 2-(4-(4-Substituted piperazin-1-yl)benzylidene)-1H-indene-1,3(2H)-diones 2,20 -Bipyridine derivatives 2,3-Dehydrosilybin A/B 2,4-Disubstituted quinazolines 2,6-Disubstituted pyridine derivatives 2-Arylethenylquinoline derivatives 2-Cyclopropylimino-3-methyl-1,3-thiazoline hydrochloride (KHG26377) 2-Phenylethynyl-butyltellurium 2-Piperidone derivatives 2-Pyridyl-benzimidazole iridium(III), ruthenium(II), and platinum(II) complexes 3-(4,5-di-Methylthiazol-2-yl)-2,5-diphenyltetrazolium bromide 3,5,40 -Trihydroxy-6,7,30 -trimethoxyflavone 3β-[N-(Dimethylaminoethane)carbamoyl]-cholesterol (DC-cholesterol) 4-Hydroxyl aurone derivatives 4-O-Methylhonokiol 5-Methoxyflavone 6-Methoxy-indanone derivatives 7-(4-Hydroxy-3-methoxyphenyl)-1-phenyl-4E-hepten-3-one 7-(4-Hydroxyphenyl)-1-phenyl-4E-hepten-3-one 8-Hydroxyquinolin derivatives substituted with (benzo[d][1,2]selenazol-3(2H)-one) AAV2/1 CD74 Ac-LPFFD-Th Acteoside Acylhydrazones Affibody-derived protein Aftins Aminostyrylbenzofuran (continued)

312

Ramo´n Cacabelos et al.

Table 3 (continued) Amyloid beta-GFP fusion protein Anti-amyloidogenic bis-styrylbenzene derivative, KMS80013 Anticalins Anti-necroptotic molecule necrostatin-1 (Nec-1) Arctigenin-4-yl carbamate derivatives Astaxanthin AZD2184 and thioflavin T Aβ aggregation modulator MRZ-99030 Aβ toxicity inhibitor SEN1500 Aβ16–20 (cyclo-[KLVFF]) Aβ1–6A2VTAT(D) peptide Aβ-degrading proteases (AβDPs) Aβ-fibrinogen interaction inhibitor BAN2401 Benzothiazole amphiphiles Betaine Bexarotene Bifunctional Aβ aggregation inhibitor Bis(indolyl)phenylmethane derivatives Bis(propyl)-cognitin Bis-chloroethylnitrosourea (BCNU or carmustine) Bismuth selenides exfoliated by hemin Bis-pyridylethenyl benzene Brain-derived neurotrophic factor-modulating peptide Camellikaempferoside B Carbenoxolone Carbonic anhydrase inhibitor Methazolamide Carboxylated Zn-phthalocyanine Chalcones Chicago Sky Blue 6B Chiral oxazino-indoles Chondroitin sulfate cis-Glycofused benzopyran compounds (continued)

Pharmacogenomics of Alzheimer’s Disease

313

Table 3 (continued) Cocaine- and amphetamine-regulated transcript (CART) Compound D737 Copper-64 labeled benzofuran derivatives Corticotrophin-releasing factor receptor 1 antagonists Covalent modifier-type aggregation inhibitor of Aβ, diazirine-equipped cyclo-KLVF(β-Ph)F CSP-1103 (formerly CHF5074) Curcumin/melatonin hybrid 5-(4-hydroxy-phenyl)-3-oxo-pentanoic acid [2-(5-methoxy-1H-indol-3yl)-ethyl]-amide Cyanidin 3-O-β-glucopyranoside Cyclic peptides Cyclodextrin derivatives Cyclophilin 40 Cyclophilin D inhibitor DBA-SLOH D-Enantiomeric peptide D3 Dextran sulfate sodium Diallyl disulfide (DADS) derivatives Diarylpropionitrile Dibenzyl imidazolidine and triazole acetamide derivatives Difluorophenylglycinols Di-O-Demethylcurcumin DL-3-n-butylphthalide-edaravone hybrids DMPD, N,N-dimethyl-p-phenylenediamine Dual peptide inhibitors Edaravone Endogenous human protein disaggregases: Hsp110, Hsp70, Hsp40, HtrA1, NMNAT2, Hsp90 Ephrin type-B receptor 2 (EphB2) inhibitors Epoxyeicosatrienoic acids (EETs) Exendin-4 Europium III chloride (Eu(3+)) Ferulic acid Fibrinogen-derived γ377–395 peptide Fisetin (continued)

314

Ramo´n Cacabelos et al.

Table 3 (continued) FK506 Flavonoid derivative 2-(40 benzyloxyphenyl)-3-hydroxy-chromen-4-one Formononetin Fusion protein NPT088 Gammagard liquid intravenous immunoglobulin (IGIV) Geldanamycin-activated P70S6K Gleevec Glucocerebrosidase Glucocorticoid receptor modulators Glycosaminoglycans Glycosphingolipid (GSL) synthesis inhibitors, synthetic ceramide analog D-PDMP (1-phenyl 2-decanoylamino-3-morpholino-1-propanol) and the iminosugar N-butyldeoxynojirimycin (NB-DNJ or miglustat) GM1 Ganglioside Graphene oxide-iron oxide Hematoxylin Histidine functionalized water-soluble perylene diimide HP-β-cyclodextrin Human innate immune peptide LL-37 Hydroxyalkylquinoline J2326 Hydroxypropyl-β-cyclodextrin Hydroxyquinoline-based derivatives Indole and 7-azaindole derivatives Indoleamine-2,3-dioxygenase Indolylquinoline derivatives Inhibitors of CD36-amyloid beta binding Isoliquiritigenin derivatives Isomeric 2,4-Diaminoquinazolines K114, (trans,trans)-1-bromo-2,5-bis(4-hydroxystyryl)benzene Lanthionine ketimine ([LK] 3,4-dihydro-2H-1,4-thiazine-3,5-dicarboxylic acid) Leptin Levetiracetam Lipocalin 2 (LCN2) (continued)

Pharmacogenomics of Alzheimer’s Disease

315

Table 3 (continued) Macrocyclic polyamines Memory-enhancing compound ISRIB Methionine 35 sulfoxide Methyl 2,4-dimethyl-5-oxo-5,6-dihydrobenzo[c][2,7]naphthyridine-1-carboxylate (BNC-1) Minocycline Monoterpenoid 2,3,4,4-tetramethyl-5-methylenecyclopent-2-enone MRZ-99030 N-(1,3-Benzodioxol-5-yl)-2-[5-chloro-2-methoxy(phenylsulfonyl)anilino]acetamide (LX2343) N(γ)-(2-Aminoethyl)-2,4-diaminobutanoic acid N-Adamantyl-4-methylthiazol-2-amine N-Alkyl carbazole derivatives Neural type I membrane protein alcadein α NiM@P hybrid particles Nonhemolytic 11-residue peptide, NF11 (NAVRWSLMRPF) NQTrp inhibitor Nucleoside-20 ,30 /30 ,50 -bis(thio)phosphate antioxidants NVP-BEZ235 (dactolisib) Oligo(p-phenylene ethynylene) electrolytes Oligothiophene p-FTAA Osmotin Oxabicycloheptene sulfonate P300/CBP-associated factor (PCAF) inhibitors p75NTR ectodomain Peroxiredoxin 5 Peroxovanadium complexes Phenothiazine-based theranostic compounds PICALM ortholog Poly(4-styrenesulfonate) Poly(LVFF-co-β-amino ester) Poly(propylene imine) glycodendrimers Polyfluorinated bis-styrylbenzenes Polyphenol-containing peptidomimetics Porphyrin derivatives (continued)

316

Ramo´n Cacabelos et al.

Table 3 (continued) Propafenone ProSP-C (prosurfactant protein C) BRICHOS Protein proSAAS Puromycin-sensitive aminopeptidase Pyrazinamide and D-cycloserine Pyroglutamylated amyloid-β peptide rAAV/ABAD-DP-6His Recombinant soluble neprilysin Resorcinarene Rhenium 2-arylbenzothiazoles Rifampicin RPS23RG1 Salvianolic acid Sarsasapogenin-AA13 Secretory chaperones 7B2 Selenium quinone active species Sialylated glycosylphosphatidylinositols SLM and SLOH, analogs of carbazole-based cyanine compounds Small-molecule ISRIB Small-molecule NPT-440-1 Sodium selenite Substituted dithiazole piperazine benzamides Sugar-based peptidomimetics Synthetic chalcone derivatives Synthetic tetrapeptides HAEE and RADD Tanshinone and analogs Taxifolin Tetra(ethylene glycol) derivative of benzothiazole aniline, BTA-EG4 Tetrahydroxystilbene glucoside Thalidomide Thrombospondin-1 Tolfenamic acid Tramiprosate (continued)

Pharmacogenomics of Alzheimer’s Disease

317

Table 3 (continued) Transition-metal-substituted polyoxometalate derivatives Transthyretin-derived peptides Tricyclic pyrone compounds Tripchlorolide Tripeptide GGH Triphenylmethane dye brilliant blue G Trolox-conjugated amyloid-β C-terminal peptides VVIACLPFFD (VCD10)@AuNP WGX-50 WIN 55,212–2 α-Mangostin β-Amyloid-derived peptide Aβ1–6A2V-TAT(D) γ-AApeptide

analytic studies conclude that no significant benefit in clinical terms has been obtained with immunotherapy in AD [57]. Assuming that mutations in the APP gene represent one of the primary causes of AD, the rationale of the initial active immunization studies was partially correct, demonstrating that presymptomatic immunization of PDAPP transgenic mice, which overexpress mutant human APP (Phe717Val), prevented Aβ-plaque formation, neuritic dystrophy, and astrogliosis; immunization with Aβ42, the most abundant isoform of pathogenic Aβ peptides, in animals that manifest AD neuropathology reduced the extent and progression of AD-like neurodegeneration markers [56]. Aβ immunization also improved cognition in the TgCRND8 murine model of AD without altering total Aβ brain levels [58]. These studies were replicated in over 400 experimental studies with different immunization procedures. However, the initial clinical trial with an active Aβ vaccine (AN1792) was halted due to the development of severe complications (i.e., acute meningoencephalitis, micro-hemorrhages) in a number of immunized patients. Some of these adverse reactions were associated with a T-cell-mediated pro-inflammatory response and other still unknown mechanisms; fortunately, most of these technical problems have been ameliorated or eliminated in recent times with improved immunization procedures. Second-generation Aβ-active immunotherapies, anti-Aβ monoclonal antibodies targeting different Aβ epitopes, and anti-tau immunotherapies have dominated the scenario of AD vaccines during the past decade [59–61]. Among anti-tau strategies, AADvac1 is an active

318

Ramo´n Cacabelos et al.

Table 4 Anti-Aβ and anti-tau immunization procedures and vaccine products Anti-Aβ products 4xAβ15 adenovirus vaccine 4Aβ1–15-derived monoclonal antibody 6Aβ15-TF chimeric vaccine 6Aβ15-THc-C chimeric vaccine (rCV02) 6copy-Aβ 1-6-PA-BLP AAB-003 (PF-05236812) AAV-CB-Aβ42 vaccine ABvac40 vaccine Ad-10  Aβ3–10-CpG AdCpG-(Aβ3–10)10 vaccine (Aβ1–42 Plp-Adenovirus [Ad]-X-CMV-(Aβ3–10)10-CpG) Adeno-associated virus (AAV) serotype 2 vector vaccine containing amyloid-β peptide (Aβ) 1–15 gene fragment (AAV-Aβ15) AdPEDI-(Aβ1–6)11 (DNA vaccine) (adenovirus vector encoding 11 tandem repeats of Aβ1–6 fused to the receptor-binding domain (Ia) of Pseudomonas exotoxin A) Aducanumab Advax(CpG)-adjuvanted MultiTEP-based dual and single vaccines AFFITOPE® AD01, AD02 AN1792 Anti-Aβ monoclonal antibody 266 Anti-Aβ monoclonal antibody 4G8 Anti-Aβ31–35 antibody Antibody NT4X Antigen-sensitized dendritic cell vaccination Anti-oligoAβ antibody AOE1 vaccine AOEP2-induced antibodies AV-1953R-Advax(CpG) (dual Aβ/tau vaccine) AV-1955 (p3Aβ 11-PADRE vaccine) (DNA epitope-based vaccine/Pan DR epitope (PADRE) AV-1959 (DNA epitope-based vaccine) AV-1959R-Advax(CpG) Aβ B-cell epitope vaccine (rCV01) Aβ IgG1 monoclonal antibody 3A1 (continued)

Pharmacogenomics of Alzheimer’s Disease

319

Table 4 (continued) Aβ vaccine with T-cell epitope of diphtheria-tetanus combined toxoid (DT) Aβ1–15-derived monoclonal antibody Aβ1–42 VAC Aβ1–43 cDNA (rAAV/Aβ) Aβ15-T-Hc chimeric DNA vaccines Aβ3–10-KLH vaccine Aβ33-41NP vaccine Aβ-HBc VLPs Aβ-Hsp70 vaccine AβpE3:CRM197 vaccine AβpE3-X peptide vaccine Abeta-VLP (papillomavirus-like particles) vaccine Bacillus Calmette-Gue´rin (BCG) (attenuated Mycobacterium bovis preparation) BAN2401 (humanized version of mAb158) Bapineuzumab (AAB-001) Bapinezumab derivative (AAB-003) C5a-peptide vaccines (AFF1, AFF2) CAD106 (Aβ1–6 coupled to the virus-like particle Qβ) Chimeric Aβ1–15-CMVs (Cucumber mosaic virus) Chimeric HPV16 L1 VLPs CMVTT-based vaccines Coimmunization vaccine Crenezumab CTB-Aβ15 silkworm pupae vaccine CTB-Aβ42 silkworm pupae-derived vaccine Dendrimeric Aβ1–15 (dAβ1–15) vaccine DNA Aβ42 vaccine DNA epitope chimeric vaccines (six copies of Aβ1–15 fused with PADRE or toxin-derived carriers) DNA epitope vaccine DepVac DNA vaccine p(Aβ3–10)10-C3d-p28.3 Dual vaccine WSN-Aβ(1–10) EB101 vaccine Edible vaccine Aβ rice (continued)

320

Ramo´n Cacabelos et al.

Table 4 (continued) Epitope vaccine Lu AF20513 Fc-inactivated anti-Aβ monoclonal antibody GSK933776 Gal4/UAS DNA A beta 42 trimer immunization Gantenerumab Gene-based vaccines Gene-gun-administered genetic immunization HSV amplicon-mediated Aβ vaccination Humanized antibody BAN2401 Humanized anti-PrP antibody IgL-Aβx4-Fc-IL-4 vaccine (DNA vaccine YM3711) Immunoglobulin M (IgM) class antibodies Immunoglobulin variable (IgV) domains (catalytic antibodies) Influenza-AD dual vaccines (flu-Aβ1–7, flu-Aβ1–10) Intravenous immunoglobulins (IVIG): IVIG immunotherapy (Octagam, Gammagard) Ixekizumab (anti-IL-17A monoclonal antibody) Keyhole limpet hemocyanin (KLH)-Aβ(37–42) Liposomal vaccine containing β-sheet aggregated lipopeptide (Palm1–15) LTB:RAGE vaccine mAb 3F5 mAb158 MAP4-Aβ conjugates and monoclonal antibody 1H7 MER5101 (Aβ1–15:DT conjugate vaccine) Multiple antigenic peptide (MAP)-based epitope vaccines (Aβ(1–11) B cell epitope fused to synthetic T cell epitope PADRE (Aβ(1–11)-PADRE)) Multivalent Aβ3–10 DNA vaccines Mutant Aβ-sensitized dendritic cell vaccine Nonviral Aβ DNA vaccine Norovirus P particle-based active Aβ immunotherapy (PP-3copy-Aβ1–6-loop123) Norovirus P particle-based Aβ chimeric protein vaccine Oligomeric-specific vaccine (polymer-encapsulated Aβ peptide fragments) p3Aβ(1–11)-PADRE-3C3d vaccine P4D6 (SDPM1 peptide-mimotope antibody) PADRE-Aβ(1–15) epitope vaccine (continued)

Pharmacogenomics of Alzheimer’s Disease

321

Table 4 (continued) p(Aβ3–10)10-C3d-p28.3 vaccine p(Aβ(3–10))(10)-IL-4 DNA vaccine p(Aβ3–10)10-MT pAβ42-IL-4 (chimeric Aβ-interleukin-4 minigene) PF-04360365 (humanized IgG(2)Δa anti-Aβ antibody Phage-displayed Aβ epitopes Plant-based vaccines pMDC-3Aβ(1–11)-PADRE Polyclonal anti-Abeta-KEK antibodies Ponezumab (PF-04360365) PSiP vaccine pV-Aβ pV-GE2 Pyroglutamate-3 amyloid-beta (pGlu-3 Aβ) immunoglobulin G1 (IgG1) monoclonal antibody, 07/1 Pyroglutamate-3 Aβ (AβpE3-X) vaccine RAGE/Aβ complex vaccine Recombinant 6Aβ15-THc-C chimeric vaccine (rCV02) Recombinant adeno-associated virus vector-Aβ1–43 cDNA (rAAV/Aβ) Recombinant chimeric vaccines with hexavalent foldable Aβ(1–15) (6Aβ15) fused to PADRE or toxinderived carrier proteins Recombinant DNA vaccine against neurite outgrowth inhibitors Recombinant Sendai virus vector carrying Aβ1–43/IL-10 cDNA Salmonella-based amyloid-beta derivative vaccine ScFv-h3D6 (bapineuzumab-derived anti-Aβ scFv) SDPM1 Single-chain variable fragment (scFv) antibody against Aβ42 oligomers Solanezumab (LY2062430) TAT-B6-C15 Tetravalent Aβ1–15 vaccine (adenovirus vaccine) Theranostic complex GNRs-APH-scFv (GAS) Tocilizumab (humanized anti-human IL-6R monoclonal antibody) UB-311 (UBITh® Aβ peptide vaccine) Vanutide cridificar (ACC-001)-QS-21 (aminoterminal Aβ1–7 peptide conjugate) (continued)

322

Ramo´n Cacabelos et al.

Table 4 (continued) Virosome-based active immunization Virus-like particle-based vaccines WSN-Aβ(1–10) dual vaccine ZSYM73-ABD Anti-tau products AADvac1 ACI-35 Anti-tau monoclonal antibody RO7105705 AV-1980R-Advax(CpG) Double phosphorylated chimeric peptide vaccine (28-mer linear chimeric peptide: double phosphorylated B cell epitope Tau229–237 (pThr231/pSer235) and immunomodulatory T cell epitope Ag85B241–237 of Mycobacterium tuberculosis, GPSL-linked) Human chimeric 4E6 (h4E6) antibody Humanized anti-tau antibody ABBV-8E12 Liposome-based amyloid vaccine with a synthetic phosphorylated peptide to mimic tau phospho-epitope pS396/pS404 Liposome-based vaccine against protein tau Monoclonal antibody BMS-986168 Monoclonal antibody C2N-8E12 Monoclonal antibody DC8E8 Monoclonal antibody HPT-101 Monoclonal antibody RG7345 MultiTEP platform-based DNA epitope vaccine AV-1980D Phospho-tau vaccine Phospho-tau monoclonal antibody ACI-5400 Sarkosyl insoluble PHF-tau T294-HBc VLP vaccination Tau oligomeric complex 1 (TOC1) Tau oligomer-specific monoclonal antibody (TOMA) Tau peptide vaccines composed of a double phosphorylated tau neoepitope (pSer202/pThr205, pThr212/pSer214, pThr231/pSer235), and an immunomodulatory T cell epitope from the tetanus toxin or tuberculosis antigen Ag85B VH5–51/VL4–1 anti-tau antibodies (CBTAU-27.1, CBTAU-28.1)

Pharmacogenomics of Alzheimer’s Disease

323

immunotherapy against tau pathology (phase II clinical trials). AADvac1-related antibodies target conformational epitopes in the microtubule-binding region of tau, preventing the spreading of tau aggregation and promoting tau clearance, with an apparently safe profile [60]. Some dual vaccines (EB101) [62], Aβ3–10-KLH vaccine [63], or active full-length DNA-Aβ42 trimer immunization [64] demonstrated a capacity to reduce both amyloid and tau aggregation and accumulation in transgenic animals, probably via the proteasome [65]; and some tau oligomer-specific monoclonal antibodies may also reduce Aβ load. On 7 June, aducanumab was formally approved by the FDA as the first drug with a putative disease-modifying mechanism for the treatment of AD, although not devoid of controversy in the scientific community [66–70]. Aducanumab is a human monoclonal antibody that selectively targets aggregated Aβ. It penetrates into the brain, binds parenchymal Aβ, and dose-dependently reduces soluble and insoluble Aβ. One year of monthly intravenous infusions of aducanumab in patients with incipient cognitive disorder or mild AD reduces brain Aβ in a time- and dose-dependent manner, with parallel slowing of cognitive decline [67–70]. Aducanumab and other injectable antibodies (gantenerumab, BAN2401), and a small molecule oral agent, ALZ-801, are amyloid-targeting drugs with variable efficacy and safety and differential effects in terms of selectivity for Aβ oligomers, plasma halflife, brain penetration, and time to peak brain exposure. Recent studies indicate that the degree of selectivity for Aβ oligomers and brain exposure drive the magnitude and onset of clinical efficacy; in contrast, the clearance of plaques with the highest doses of aducanumab and BAN2401 is associated with vasogenic brain edema, especially in APOE4 carriers [71, 72]. The selective anti-oligomer agent ALZ-801, an optimized oral prodrug of tramiprosate, shows efficacy in homozygous APOE4/4 AD subjects, without evidence of vasogenic edema [71].

6

Pharmacogenomics

6.1 The Pharmacogenomic Machinery in Alzheimer’s Disease

The pharmacogenomic machinery is composed of a network of gene clusters coding for proteins and enzymes responsible for drug targeting and processing, as well as critical components of the epigenetic machinery that regulates gene expression [73]. The pharmagenes involved in the pharmacogenomic response to drugs can be classified into five major categories: (i) pathogenic genes which are associated with disease pathogenesis (Tables 5 and 6; Figs. 1 and 2); (ii) mechanistic genes coding for components of enzymes, receptor subunits, transmitters, and messengers

324

Ramo´n Cacabelos et al.

associated with the mechanism of action of drugs; (iii) metabolic genes of different categories that encode phase I–II reaction enzymes responsible for drug metabolism (Figs. 3 and 4); (iv) transporter genes coding for drug transporters (Fig. 5); and (v) pleiotropic genes which encode proteins and enzymes involved in a great variety of metabolic cascades and metabolomic networks [10, 11, 73]. Most drugs may act as substrates, inducers, or inhibitors of pharmagenes (Fig. 6). Rare variants contribute to approximately 30–40% of functional variability in 146 pharmagenes with clinical relevance. Over 240 pharmagenes are potentially associated with ADRs, and over 400 genes and their products influence drug efficacy and safety [74, 75]. The distribution and frequency of pharmagene genophenotypes in the general population (Table 7; Fig. 7) and in patients with AD (Fig. 8) are similar, except in a few genes, with an uneven pattern in terms of defective function. Of the 60 genes integrated in the Smart Pharmacogenetic Card (PGx-60/4000) (60 genes; >300 SNPs; 4000 drugs), a bioinformatic device for personalized treatments at the International Center of Neuroscience and Genomic Medicine, Corunna, Spain, the most dysfunctional genes frequently found (frequency >20%) in AD are the following: CYP1A1 (31.47%), CYP1B1 (82.23%), MAOB (47.72%), CES1 (96.45%), CHAT (36.04%), COMT (30.96%), GSTM1 (54.82%), GSTT1 (25.38%), NAT2 (43.65%), SOD2 (67.51%), ABCB1 (44.16%), ABCG2 (90.36%), FABP2 (90.86%), SLCA2 (59.90%), SLC22A1 (34.52%), SLC30A8 (48.22%), ADRB2 (41.62%), AGT (42.13%), APOE (30.96%), CHRNA7 (42.13%), DRD2 (45.18%), GABRA1 (55.33%), HMGCR (73.60%), HTR1A (65.48%), HTR2C (79.19%), OPRM1 (69.54%), PPARG (81.73%), PRKCE (41.62%), and VKORC1 (65.99%) (Figs. 7 and 8). There is an accumulation of 15–26 defective pharmagenes in approximately 85% of AD patients. The maximum number of defective geno-phenotypes in AD patients is 17 genes in 14.50% of cases, 18 genes in 16.91% of cases, and 19 genes in 14.40% of AD cases (Fig. 9). The expression or repression of all these genes and their products are regulated in a redundant and promiscuous fashion by the epigenetic machinery (DNA methylation/demethylation, histone/ chromatin remodeling, miRNA regulation), configuring the pharmacoepigenetic apparatus [73, 76, 77]. The same enzyme/protein/transporter can process a multitude of drugs, and the same drug can be processed by different gene products in an orchestrated manner to operate as a security system against xenobiotic intruders [76]. 6.2 Pathogenic Genes

Over the past 30 years, more than 600 genes distributed throughout the human genome have been linked to the risk of AD [4, 78, 79]. Several pathogenic mutations in the amyloid precursor protein

Gene name

Alpha-2-macroglobulin

ATP-binding cassette subfamily A member 7

Angiotensin I converting enzyme

Apolipoprotein E

Bridging integrator 1

Chromosome 9 open reading frame 72

Clusterin

Carboxypeptidase Z

Gene

A2M

ABCA7

ACE

APOE

BIN1

C9ORF72

CLU

CPZ 603105

185430

614260

601248

107741

106180

605414

103950

OMIM

chr4:8650823

chr8:27607002

chr9:27543280

chr2:127137039

chr19:44908822

chr17:63477060

chr19:1046521

chr12:9067708

Location

Table 5 Frequencies of polymorphic variants in AD-related pathogenic genes

rs7436874

rs11136000

rs3849942

g.8649098C>T

c.247-478A>G

g.27543283T>C

g.127137039A>G

c.3932T>C

rs429358

rs744373

c.4070C>T

c.496-66T>C

c.1622+115T>G

c.2998A>G

Polymorphism

rs7412

rs4332

rs3764650

rs669

dbSNP

0.36 (C)

0.38 (A)

0.22 (T)

0.36 (G)

0.15 (C)

0.08 (T)

0.47 (T)

0.20 (G)

0.31 (G)

MAF

(continued)

CC: 17.54% CT: 47.37% TT: 35.09%

GG: 36.84% AG: 49.12% AA: 14.04%

CC: 59.65% TC: 36.84% TT: 3.51%

AA: 42.11% AG: 50.88% GG: 7.01%

*2*2: G

g.85868640T>C

c.894G>T

c.*149+175A>C

c.279-2443C>A

g.39298100G>A

c.2242-7030T>G

c.4946-54A>G

Polymorphism

0.73 (A)

0.31 (T)

0.18 (T)

0.45 (A)

0.38 (A)

0.35 (G)

0.03 (G)

0.25 (A)

MAF

GG: 19.30% AG: 40.35% AA: 40.35%

CC: 45.61% CT: 42.11% TT: 12.28%

GG: 50.88% GT: 35.09% TT: 14.03%

CC: 36.84% CA: 38.60% AA: 24.56%

CC: 49.12% CA: 42.11% AA: 8.77%

AA: 45.61% GA: 38.60% GG: 15.79%

TT: 94.74% TG: 5.26% GG: A

c.856+16G>T

0.09 (A)

0.43 (G)

GG: 73.68% GA: 22.81% AA: 3.51%

TT: 22.80% GT: 35.09% GG: 42.11%

Pharmacogenomics of Alzheimer’s Disease 327

Gene name

Angiotensin I converting enzyme

Angiotensinogen

Angiotensinogen

Apolipoprotein B

Apolipoprotein C-III

Apolipoprotein E

Cholesteryl ester transfer protein

Coagulation factor II. thrombin

Gene symbol

ACE

AGT

AGT

APOB

APOC3

APOE

CETP

F2

176930

118470

107741

107720

107730

106150

106150

106180

OMIM

chr11:46739505

rs1799963

rs708272

rs429358

chr19:44908684

chr16:56962376

rs7412

rs5128

rs693

rs699

rs4762

rs4332

dbSNP ID

chr19:44908822

chr11:116832924

chr2:21009323

chr1:230710048

chr1:230710231

chr17:63486920

Location

Table 6 Frequencies of polymorphic variants in cerebrovascular disease-related genes in Alzheimer’s disease

c.20210G>A

c.+279G>A

c.3932T>C

c.4070C>T

c.3175C>G

c.2488C>T

c.803T>C

c.620C>T

c.496-66T>C

Polymorphism

0.01 (A)

0.38 (A)

0.15 (C)

0.08 (T)

0.23 (C)

0.25 (T)

0.30 (T)

0.10 (T)

0.47 (T)

MAF

GG: 96.41% GA: 3.47% AA: 0.12%

GG: 37.39% GA: 49.42% AA: 13.19%

*2*2: 0.32% *2*3: 7.62% *2*4: 1.28% *3*3: 63.73% *3*4:23.88 % *4*4: 3.17%

CC: 78.94% CG: 17.60% GG: 3.46%

CC: 29.03% CT: 47.64% TT: 23.33%

TT: 21.97% TC: 56.48% CC: 21.61%

CC: 11.46% CT: 21.96% TT: 66.58%

CC: 18.70% CT: 39.81% TT: 41.49%

Genotype

328 Ramo´n Cacabelos et al.

Coagulation factor V

Interleukin 1 beta

Interleukin 6

Interleukin 6

Interleukin 6 receptor

Lipoprotein lipase

Methylenetetrahydrofolate reductase

Methylenetetrahydrofolate reductase

Nitric oxide synthase 3

Tumor necrosis factor

F5

IL1B

IL6

IL6

IL6R

LPL

MTHFR

MTHFR

NOS3

TNF

191160

163729

607093

607093

609708

147880

147620

147620

147720

227400

chr6:31575566

chr7:150991055

chr1:11794419

chr1:11796321

chr8:19962213

chr1:154454494

chr7:22726627

chr7:22727026

chr2:112832813

chr1:169549811

rs1800629

rs1799983

rs1801131

rs1801133

rs328

rs2228145

rs1800796

rs1800795

rs1143634

rs6025

c.-308G>A

c.894G>T

c.1286A>C

c.665C>T

c.1421C>G

c.1510A>C

c.-573G>C

c.-174G>C

c.3954T>C

c.1691G>A

0.09 (A)

0.18 (T)

0.25 (C)

0.25 (T)

0.09 (G)

0.36 (C)

0.31 (C)

0.14 (C)

0.13 (T)

0.01 (A)

GG: 73.79% GA: 22.86% AA: 3.35%

GG: 39.54% GT: 47.67% TT: 12.79%

AA: 50.36% AC: 39.90% CC: 9.74%

CC: 38.90% CT: 45.84% TT: 15.26%

CC: 76.02% CG: 20.00% GG: 3.98%

AA: 34.41% AC: 49.69% CC: 15.90%

GC: 15.90% GG: 81.12% CC: 2.98%

GG: 39.95% GC: 43.55% CC: 16.50%

TT: 4.59% TC: 31.39% CC: 64.02%

GG: 98.02% GA: 1.61% AA: 0.37%

Pharmacogenomics of Alzheimer’s Disease 329

330

Ramo´n Cacabelos et al.

Fig. 1 Risk gene variants in AD-related pathogenic genes and in genes associated with cerebrovascular disorders in dementia (for abbreviations, see Tables 5 and 6)

(APP) (>50 mutations), PSEN1 (>300 mutations), and PSEN2 genes (>40 mutations), present in less than 10% of AD cases, confer AD the condition of a brain amyloidopathy; mutations in the microtubule-associated protein tau (MAPT) gene (>100

Pharmacogenomics of Alzheimer’s Disease

331

Fig. 2 Accumulation of defective genes in patients with Alzheimer’s disease (for abbreviations, see Table 5)

Fig. 3 Distribution and frequency of sex-related CYP2D6, CYP2C9, CYP2C19, and CYP3A5 geno-phenotypes in patients with Alzheimer’s disease. EM extensive/normal metabolizers, IM intermediate metabolizers, PM poor metabolizers, UM ultra-rapid metabolizers

332

Ramo´n Cacabelos et al.

Prevalent CYP2D6-CYP2C9-CYP2C19 Trigenic Genotypes 25

Frequency (%) Genotype

20

15 10 5

1/1-2/3-1/1 4/4-1/1-1/1 1/4-1/2-1/2 1/1-2/2-1/1 1/1-1/1-1/17 1/4-1/3-1/1 1/1N-1/1-1/1 1/1-1/2-1/2 1/4-1/1-1/2 1/1-1/3-1/1 1/4-1/2-1/1 1/4-1/1-1/1 1/1-1/1-1/2 1/1-1/2-1/1 1/1-1/1-1/1

Total

T (%) 1,03 1,03 1,32 1,48 1,7 1,92 2,09 2,48 4,23 4,27 4,68 9,34 9,59 9,88 22,91

F (%)

M (%) 0,74 0,93 1,42 1,48 1,79 2,16 1,85 2,72 4,38 4,38 4,57 9,81 9,75 9,38 23,35

1,34 1,14 1,2 1,47 1,6 1,67 2,34 2,2 4,07 4,14 4,81 8,82 9,42 10,42 23,51

Females

Males

N=3117 Females: 1620 Males: 1497

0

Fig. 4 Frequency of the most prevalent trigenic (CYP2D6-CYP2C9-CYP2C19) genotypes in patients with Alzheimer’s disease

Fig. 5 Genotypes, haplotypes, and genotypes of the ABCB1 gene in patients with Alzheimer’s disease. LR low resistance, IR intermediate resistance, HR high resistance

mutations are associated with diverse tauopathies: frontotemporal dementia, Pick’s disease) link AD to other tauopathies, although MAPT variants are not specific to prototypical forms of AD [80– 82]. Amyloidopathy and tauopathy have been the two dominant hypotheses in the etiopathogenesis of AD for years [83, 84]. Missense APP mutations in EOAD cause AD, whereas the coding variant, APP A673T, reduces the risk for AD. AD riskassociated mutations in the APP gene increase total Aβ levels, Aβ42 levels, or Aβ fibrillogenesis, while protective alleles reduce Aβ levels [82].

Pharmacogenomics of Alzheimer’s Disease

333

Fig. 6 Mechanistic effects of substrates, inducers, and inhibitors of CYP enzymes

Presenilin is the catalytic site of γ-secretase and the most dominant mutations associated with familial EOAD occur either in the APP gene encoding the amyloid precursor protein substrate or in the PSEN1 and PSEN2 genes encoding the protease (presenilin) responsible for APP cleavage, leading to abnormal Aβ accumulation and deposition in senile plaques and vessels. Apolipoprotein E4 (APOE-4), the most important risk factor for AD in >40% of cases, impairs Aβ clearance from brain tissue. Immunotherapy with different Aβ antibodies (solanezumab, crenezumab, and aducanumab) attempts to reduce Aβ and slow down cognitive decline in presymptomatic and/or mild AD cases, as a novel line of therapeutic intervention [83]. In addition to these pathogenic genes, many other genes, as well as novel mutations in PSEN1, have been identified in association with AD in recent next-generation sequencing (NGS) and genome-wide association studies (GWAS) in different ethnic groups, indicating that many pathogenic genes can accumulate in each AD case [78, 79, 85–89]. A very important aspect in the analysis of any genetic study in polygenic and complex diseases is to weigh the pathogenic load that each gene has in an individual case. Using a panel with the 18 most influential genes in AD (Table 5; Fig. 1) and in cerebrovascular

Adrenoceptor beta 3

Angiotensinogen

Apolipoprotein E

ADRB3

AGT

APOE

rs699

rs4994

rs1042713 rs1042714

rs1800544

rs2231142

rs17222723 rs2273697 rs3740066 rs717620

rs1045642 rs1128503 rs2032582

dbSNP

19q13.32 rs429358 rs7412

1q42.2

8p11.23

5q32

4q22.1

Adrenoceptor beta 2

ATP-binding cassette subfamily G member 2 (Junior blood group)

ABCG2

10q24.2

ADRB2

ATP-binding cassette subfamily C member 2

ABCC2

7q21.12

10q25.2

ATP-binding cassette subfamily B member 1

ABCB1

Locus

ADRA2A Adrenoceptor alpha 2A

Gene name

Gene symbol

Table 7 Smart pharmacogenetic card (PGx-60/4000) geno-phenotypes

c.3932T>C; Cys112Arg c.4070C>T; Cys158Arg

c.803T>C; p.Met268Thr

c.190T>C; p.Trp64Arg

c.46A>G; p.Arg16Gly c.79C>G; p.Gln27Glu

c.-1252G>C

c.421C>A; p.Gln141Lys

c.3563T>A c.1249G>A c.3972C>T c.-24C>T

c.3435T>C; p.Ile1145¼ c.1236T>C; p.Gly412¼ c.2677C>A/T; p. Ser893Thr/Ala

Polymorphism

Phenotype

HR: 85.44% IR: 11.69% LR: 20.87%

HR: 13.84% IR: 44.87% LR: 41.29%

0.15 (C) 0.08 (T)

0.30 (T)

AR: 14.58% DR: 21.95% NR: 63.47%

AR: 35.80% DR: 37.47% NR: 26.63%

0.08 (G) AR: 1.19% DR: 13.13% NR: 85.68%

0.39 (A) AR: 41.77% 0.47 (G) DR: 40.81% NR: 17.42%

0.26 (G) AR: 9.31% DR: 35.32% NR: 55.37%

0.09 (A)

0.04 (A) 0.19 (A) 0.29 (T) 0.13 (T)

0.48 (C) HR: 45.35% 0.43 (T) IR: 20.05% 0.07 LR: 34.60% (T/A)

Allele

334 Ramo´n Cacabelos et al.

Cytochrome P450 family 1 subfamily A member 1

Cytochrome P450 family 1 subfamily B member 1

Cytochrome P450 family 2 subfamily A member 6

CYP1A2

CYP1B1

CYP2A6

19q13.2

2p22.2

15q24.1

rs28399433

rs1056836

rs2069514 rs35694136 rs762551

rs1378942

Cytochrome P450 family 1 subfamily A member 1

CYP1A1 15q24.1

22q11.21 rs4680

Catechol O-methyltransferase

rs6494223

10q11.23 rs2177369

rs71647871

COMT

Choline O-acetyltransferase

CHAT

16q12.2

rs1803274

15q13.3

Carboxylesterase 1

CES1

3q26.1

CHRNA7 Cholinergic receptor nicotinic alpha 7 subunit

Butyrylcholinesterase

BCHE

g.-48T>G; *9

c.1294C>G; p.Leu432Val

g.-3860G>A; *1C g.-2467delT; *1D g.-163C>A: *1F

c.-66+2306C>A

c.472G>A; p.Val158Met

c.240+2907C>T

c.757+7898G>A

c.432G>A; p.Gly144Glu

c.1699G>A; p.Ala567Thr

NM:30.79% IM: 0.48% PM: 0.23% UM: 68.50%

NR: 38.19% DR: 31.26% AR: 30.55%

AR: 43.44% DR: 0.00% NR: 56.56%

AR: 40.10% DR: 39.85% NR: 20.05%

AR: 35.32% DR: 27.92% NR: 36.76%

AR: 95.47% DR: 3.58% NR: 0.95%

AR: 11.93% DR: 21.72% NR: 66.35%

0.07 (C)

(continued)

NM: 94.27% IM: 5.73% PM: 0.00%

0.49 (G) NM: 21.24% RM: 78.76%

0.02 (A) 0.08 (delT) 0.21 (C)

0.39 (C)

0.50 (A)

0.39 (T)

0.38 (A)

0.01 (A)

0.19 (A)

Pharmacogenomics of Alzheimer’s Disease 335

Cytochrome P450 family 2 subfamily E member 1

CYP2E1

10q26.3

rs3813867 rs6413420

Indel Indel rs28371725 rs35742686 rs3892097 rs5030655

Cytochrome P450 family 2 subfamily D member 6

CYP2D6 22q13.2

10q23.33 rs1057910 rs1799853 rs28371685 rs28371686 rs7900194 rs9332131

Cytochrome P450 family 2 subfamily C member 9

rs3745274

CYP2C9

19q13.2

dbSNP

10q23.33 rs12248560 rs4244285

Cytochrome P450 family 2 subfamily B member 6

CYP2B6

Locus

CYP2C19 Cytochrome P450 family 2 subfamily C member 19

Gene name

Gene symbol

Table 7 (continued)

Allele

g.-1293G>C; *5 g.-71G>T; *7

Insertion; *1xN Deletion; *5 g.2988G>A; *41 g.2549delA; *3A g.1846G>A; *4A g.1707T>del; *6A

c.1075A>C; *3A c.430C>T; *2A c.1003C>T; *11 c.1080C>A; *5 c.449G>T; *8 c.817delA; *6

g.-806C>T; *17 c.681G>A; *2A

NM: 59.42% IM: 18.14% PM: 3.34% UM: 19.10%

NM: 65.87% IM: 27.45% PM: 6.68%

Phenotype

0.04 (C) 0.07 (T)

0.11 (T) 0.02 (delA) 0.19 (T) 0.02 (delT)

NM: 84.96% IM: 14.06% PM: 0.72%

NM: 50.12% IM: 31.50% PM: 13.13% UM: 5.25%

0.07 (A) NM: 70.64% 0.12 (T) IM: 24.58% 0.01 (T) PM: 4.78% 0.00 (G) 0.01 (A) 0.00 (delA)

0.22 (T) 0.14 (A)

c.516G>T; p.Gln172His; *6 0.24 (T)

Polymorphism

336 Ramo´n Cacabelos et al.

Cytochrome P450 family 3 subfamily A member 4

Cytochrome P450 family 3 subfamily A member 5

Cytochrome P450 family 4 subfamily F member 2

Dihydropyrimidine dehydrogenase

Dopamine receptor D2

Fatty acid binding protein 2

Glucose-6-phosphate dehydrogenase

Gamma-aminobutyric acid type A receptor alpha1 subunit

Glutathione S-transferase mu 1

Glutathione S-transferase pi 1

Glutathione S-transferase theta 1

CYP3A4

CYP3A5

CYP4F2

DPYD

DRD2

FABP2

G6PD

GABRA1

GSTM1

GSTP1

GSTT1

rs776746

rs2242480 rs35599367

rs1138272 rs1695

Indel

rs2279020

rs1050828 rs5030868

rs1799883

rs1076560

rs3918290 rs55886062 rs67376798

22q11.23 Indel

11q13.2

1p13.3

5q34

Xq28

4q26

11q23.2

1p21.3

19p13.12 rs2108622

7q22.1

7q22.1

Deletion; *0

c.338C>T; Ala114Val; *1B c.313A>G; Ile105Val; *1C

Deletion; *0

c.1059+15G>A

c.202G>A; p.Val68Met c.563C>T; Ser188Phe

c.163G>A; p.Ala54Thr

c.811-83G>T

c.1905+1G>A/C; *2A c.1679T>G; p.Ile560Ser; *13 c.2846A>T; p.Asp949Val; 949V

c.1297G>A

g.6986A>G; *3A

g.20230G>A; *1G g.20493C>T; *22

NA:100% MD: 0.00% SD: 0.00%

NR: 9.07% AR: 90.93%

NR: 56.56%% AR: 43.44%

NM: 99.04% IM: 0.96% PM: 0.00%

NM: 57.76% IM: 31.74% PM: 10.50%

NM: 85.20% RM: 14.08%

NM: 79.24% IM: 17.66% PM: 3.10%

(continued)

NA: 24.35% PD: 28.16% TD: 47.49 %

0.05 (T) NM: 57.52% 0.27 (G) IM: 31.98% PM: 10.50%

NM: 56.80% IM: 19.57% PM: 23.63%

0.01 (G) NR: 41.53% AR: 58.47%

0.00 (T) 0.00 (A)

0.33 (T)

0.15 (A)

0.01 (T) 0.01 (C) 0.01 (A)

0.24 (T)

0.06 (T)

0.08 (T) 0.05 (A)

Pharmacogenomics of Alzheimer’s Disease 337

Gene name

3-Hydroxy-3-methylglutaryl-CoA reductase

5-Hydroxytryptamine receptor 2A

5-Hydroxytryptamine receptor 2C

Interferon lambda 3

Monoamine oxidase B

Methylenetetrahydrofolate reductase

N-Acetyltransferase 2

Gene symbol

HMGCR

HTR2A

HTR2C

IFNL3

MAOB

MTHFR

NAT2

Table 7 (continued)

8p22

1p36.22

Xp11.3

19q13.2

Xq23

13q14.2

5q13.3

Locus

rs1041983 rs1208 rs1799929 rs1799930 rs1799931 rs1801279 rs1801280

rs1801131

rs1799836

rs8099917

rs3813929

rs6313

rs3846662

dbSNP

c.282C>T; p.Tyr94Tyr; *13 c.803G>A; p.Arg268Lys; *12 c.481C>T; p.Leu161Leu; *11 c.590G>A; p.Arg197Gln; *6 c.857G>A; p.Gly286Glu; *7 g.191G>A; *14 c.341T>C; p.Ile114Thr; *5

c.1286A>C; p.Glu429Ala

c.1300-36A>G

g.39743165T>G

c.-759C>T

c.102C>T

c.1722+45A>G

Polymorphism

AR: 81.86% NR: 18.14%

AR: 67.78% NR: 32.22%

AR: 73.51% NR: 26.47%

Phenotype

AR: 17.18% DR: 43.91% NR: 38.91%

AR: 49.40% NR: 50.60%

0.31 (T) SA: 42.96% 0.44 (G) IA: 44.15% 0.44 (T) RA: 12.89% 0.28 (A) 0.02 (A) 0.00 (A) 0.45 (C)

0.25 (C)

0.45 (C)

0.17 (G) AR: 5.49% NR: 94.51%

0.10 (T)

0.44 (T)

0.38 (T)

Allele

338 Ramo´n Cacabelos et al.

Neurobeachin

mu-type opioid receptor

Peroxisome proliferator activated receptor gamma

Protein kinase C epsilon

Prostaglandin-endoperoxide synthase 2

Ryanodine receptor 1

Solute carrier family 22 member 1

Solute carrier family 2 member 2

Solute carrier family 2 member 9

Solute carrier family 30 member 8

NBEA

OPRM1

PPARG

PRKCE

PTGS2

RYR1

SLC22A1

SLC2A2

SLC2A9

SLC30A8

8q24.11

4p16.1

3q26.2

6q25.3

19q13.2

1q31.1

2p21

3p25.2

6q25.2

13q13.3

rs13266634

rs16890979

rs5400

rs622342

0.06 (T)

0.16 (C)

AR: 72.32% DR: 17.90% NR: 9.78%

AR:10.50% DR: 22.19% NR: 67.31%

c.*427T>C

c.349-45927T>C

c.826C>T; Arg276Trp

c.844G>A; p.Val253Ile

c.329C>T; p.Thr110Ile

c.810-2964C>A

0.28 (T)

0.21 (T)

0.14 (A)

0.41 (A)

0.01 (T)

0.40 (C)

0.01 (T)

(continued)

AR: 52.51% DR: 31.98% NR: 15.51%

AR: 8.11% DR: 31.27% NR: 60.62%

AR: 53.94% DR: 35.80% NR: 10.26%

AR: 28.64% DR: 34.13% NR: 37.23%

AR: 6.21% NR: 93.79%

AR: 18.85% DR: 28.16% NR: 52.99%

AR: 42.72% DR: 41.05% NR: 16.23%

c.-2-28078C>G; p.Pro12Ala 0.12 (G) AR: 80.43% DR: 17.66% NR: 1.91%

c.397A>G; p.Asn40Asp

g.34376390C>T

rs118192172 c.1840C>T; p.Arg614Cys

rs5275

rs6720975

rs1801282

rs1799971

rs17798800

Pharmacogenomics of Alzheimer’s Disease 339

Solute carrier family 6 member 2

Solute carrier family 6 member 3

Solute carrier family 6 member 4

Solute carrier organic anion transporter family member 12p12.1 1B1

Superoxide dismutase 2

Superoxide dismutase 3

Thiopurine S-methyltransferase

SLC6A2

SLC6A3

SLC6A4

SLCO1B1

SOD2

SOD3

TPMT

6p22.3

4p15.2

6q25.3

17q11.2

5p15.33

16q12.2

4q24

Solute carrier family 39 member 8

SLC39A8

Locus

Gene name

Gene symbol

Table 7 (continued)

rs1142345 rs1800460 rs1800462 rs1800584

rs1799895

rs4880

rs4149056

rs2020936

rs460000

rs5569

rs13107325

dbSNP

c.674A>G; *3C c.460G>A; *3B c.238G>C; *2 c.626-1G>A; *4

c.691C>G; p.Arg231Gly

c.47T>C; p.Val16Ala

c.521T>C; p.Val174Ala; *5

c.-220-881C>T

g.77063C>A

c.1287G>A

c.1171G>A

Polymorphism

AR: 7.16% DR: 15.27% NR: 77.57%

AR: 5.97% DR: 30.31% NR: 63.72%

AR: 8.11% DR: 25.78% NR: 66.11%

AR: 9.55% DR: 36.28% NR: 54.17%

AR: 0.95% DR: 8.11% NR: 90.94%

Phenotype

0.04 (G) NM: 93.08% 0.01 (A) IM: 4.77% 0.01 (G) PM: 2.15% 0.01 (T)

0.01 (G) AR: 2.39% NR: 97.61%

0.46 (G) AR: 69.45% NR: 30.55%

0.16 (C)

0.17 (C)

0.25 (T)

0.36 (A)

0.02 (T)

Allele

340 Ramo´n Cacabelos et al.

Vitamin K epoxide reductase complex subunit 1

VKORC1 16p11.2

2q37.1

rs9923231

rs4148323

c.-1639G>A

c.211G>A; *6

0.39 (T)

0.01 (A)

HS: 62.05% NS: 37.95%

NM: 97.14% IM: 0.72% PM: 2.14%

AR abnormal response, DR deficient response, HR high resistance, HS high sensitivity, IA intermediate acetylator, IM intermediate metabolizer, IR intermediate resistance, LR low resistance, MD moderate deficiency, NA normal activity, NM normal metabolizer, NR normal response, NS normal sensitivity, PD partial deletion, PM poor metabolizer, RA rapid acetylator, RM rapid metabolizer, SA slow acetylator, SD severe deficiency, TD total deletion, UM ultra-rapid metabolizer

UDP glucuronosyltransferase family 1 member A1

UGT1A1

Pharmacogenomics of Alzheimer’s Disease 341

342

Ramo´n Cacabelos et al.

Distribuon and Frequency of Pharmagene GenoPhenotypes in the General Populaon (Smart Pharmacogenec Card) Abnormal

Intermediate

Normal

100 90

80 70

60 50

40 30 20

0

ABCB1 ABCC2 ABCG2 ADRA2A ADRB2 ADRB3 AGT APOE BCHE CES1 CHAT CHRNA7 COMT CYP1A1 CYP1A2 CYP1B1 CYP2A6 CYP2B6 CYP2C19 CYP2C9 CYP2D6 CYP2E1 CYP3A4 CYP3A5 CYP4F2 DPYD DRD2 FABP2 G6PD GABRA1 GSTM1 GSTP1 GSTT1 HMGCR HTR2A HTR2C IFNL3 MAOB MTHFR NAT2 NBEA OPRM1 PPARG PRKCE PTGS2 RYR1 SLC22A1 SLC2A2 SLC2A9 SLC30A8 SLC39A8 SLC6A2 SLC6A3 SLC6A4 SLCO1B1 SOD2 SOD3 TPMT UGT1A1 VKORC1

10

Fig. 7 Distribution and frequency of pharmagene geno-phenotypes in the general population. (Source: EuroEspes Smart Pharmacogenetic Card. CIBE PGx Database. International Center of Neuroscience and Genomic Medicine, Corunna, Spain (for abbreviations, see Table 7))

Frequency of defecve genophenotypes of pharmagenes in paents with Alzheimer’s disease 100 Frequency (%) 90 80 70 60 50 40 30 20 10

ABCB1 ABCC2 ABCG2 ADRA2A ADRB2 ADRB3 AGT APOE BCHE CES1 CHAT CHRNA7 COMT CYP1A1 CYP1A2 CYP1B1 CYP2A6 CYP2B6 CYP2C19 CYP2C9 CYP2D6 CYP2E1 CYP3A4 CYP3A5 CYP4F2 DPYD DRD2 FABP2 G6PD GABRA1 GSTM1 GSTP1 GSTT1 HMGCR HTR2A HTR2C IFNL3 MAOB MTHFR NAT2 NBEA OPRM1 PPARG PRKCE PTGS2 RYR1 SLC22A1 SLC2A2 SLC2A9 SLC30A8 SLC39A8 SLC6A2 SLC6A3 SLC6A4 SLCO1B1 SOD2 SOD3 TPMT UGT1A1 VKORC1

0

Fig. 8 Frequency of defective geno-phenotypes of pharmagenes in patients with Alzheimer’s disease. (Source: EuroEspes Smart Pharmacogenetic Card. CIBE PGx Database. International Center of Neuroscience and Genomic Medicine, Corunna, Spain (for abbreviations, see Table 7))

Pharmacogenomics of Alzheimer’s Disease

343

Cumulave frequency of defecve pharmagene variants in Alzheimer’s disease 18 F% 16

Linear (F%) Log (%)

14

Frequency (%)

12 10

8 6 4 2 0 0

5

10 15 20 Number of defecve pharmagene variants per paent

25

30

Fig. 9 Cumulative frequency of defective pharmagene variants in patients with Alzheimer’s disease

disorders associated with dementia (Table 6; Fig. 1), in our AD cohort we found that (i) no patient is a carrier of a single pathogenic gene, (ii) most patients (>60%) are carriers of several pathogenic genes (maximum frequency: 10 pathogenic variants per patient), (iii) a considerable number of cerebrovascular risk variants are present in the genotype of AD patients, and (iv) the genes that most frequently (>50%) accumulate pathogenic variants in the same case of AD are A2M (54.38%), ACE (78.94%), BIN1 (57.89%), CLU (63.15%), CPZ (63.15%), LHFPL6 (52.63%), MS4A4E (50.87%), MS4A6A (63.15%), PICALM (54.38%), PRNP (80.7059), and PSEN1 (77.19%) (Fig. 2). In relation to the pathogenic load that the APOE-4 allele may represent in the clinical expression of AD and in its neuropathological phenotype, the pathogenic influence of the APOE-4 allele, from a quantitative point of view, does not affect more than 35–40% of AD cases. However, the pathogenic role of the genotypes APOE-2/ 4, APOE-3/4, and especially APOE-4/4 is highly relevant (Tables 5 and 6). From multiple studies designed to characterize APOErelated AD phenotypes for the past 30 years, several conclusions can be drawn: (i) the age at onset is 5–10 years earlier in approximately 80% of AD cases harboring the APOE-4/4 genotype; (ii) APOE serum levels are lowest in APOE-4/4, intermediate in APOE-3/3 and APOE-3/4, and highest in APOE-2/3 and APOE2/4 cases; (iii) serum cholesterol levels are higher in APOE-4/4

344

Ramo´n Cacabelos et al.

than in the other genotypes; (iv) HDL-cholesterol levels are lower in APOE-3 homozygotes than in APOE-4 allele carriers; (v) LDL-cholesterol levels are systematically higher in APOE-4/4 than in any other genotype; (vi) triglyceride levels are significantly lower in APOE-4/4; (vii) nitric oxide levels are slightly lower in APOE-4/4; (viii) serum and CSF Aβ levels differ between APOE-4/ 4 and the other most frequent genotypes (APOE-3/3, APOE-3/4); (ix) blood histamine levels are dramatically reduced in APOE-4/4 compared with the other genotypes; (x) brain atrophy and AD neuropathology are markedly increased in APOE-4/4>APOE-3/ 4>APOE-3/3 carriers; (xi) brain mapping activity shows a significant increase in slow wave activity in APOE-4/4 from early stages of the disease; (xii) brain hemodynamics, reflected by reduced brain blood flow velocity and increased pulsatility and resistance indices, is significantly worse in APOE-4 than in APOE-3 carriers; brain hypoperfusion and neocortical oxygenation are also more deficient in APOE-4 carriers; (xiii) lymphocyte apoptosis is markedly enhanced in APOE-4 carriers; (xiv) cognitive deterioration is faster in APOE-4/4 patients than in carriers of any other APOE genotype; (xv) in approximately 3–8% of the AD cases, some dementia-related metabolic dysfunctions accumulate more in APOE-4 carriers than in APOE-3 carriers; (xvi) some behavioral disturbances, alterations in circadian rhythm patterns, and mood disorders are slightly more frequent in APOE-4 carriers; (xvii) aortic and systemic atherosclerosis are also more frequent in APOE-4 carriers; (xviii) liver metabolism and transaminase activity differ in APOE-4/4 carriers than in other genotypes; (xix) hypertension and other cardiovascular risk factors also accumulate in APOE-4; and (xx) APOE-4/4 carriers are the poorest responders to conventional drugs. These 20 major phenotypic features clearly illustrate the biological disadvantage of APOE-4 homozygotes and the potential consequences that these patients may experience when they receive pharmacological treatment for AD and/or concomitant pathologies [4, 10–12, 17, 38, 39, 49–51, 79, 90–96]. 6.3 Mechanistic Genes Involved in Cholinergic Neurotransmission

Mechanistic genes encode proteins, enzymes, and receptor subunits related to the mechanism of drug action. Key elements in the cholinergic neurotransmission include ACh precursors (choline, acetyl-CoA), ACh synthesis (choline acetyltransferase), and degradation enzymes (acetylcholinesterase, butyrylcholinesterase), choline transporter, vesicular ACh transporter, and cholinergic receptors (nicotinic, muscarinic). Cholinergic neurons of the basal forebrain (basocortical cholinergic pathway, septohippocampal cholinergic pathway), where the nucleus basalis of Meynert is located, and cortical cholinergic projections are the brain territories primarily affected in AD, with 60–80% depletion of cholinergic markers in severe cases [48].

Pharmacogenomics of Alzheimer’s Disease

345

Choline acetyltransferase (ChAT; EC 2.3.1.6), the enzyme responsible for the biosynthesis of ACh from choline and acetylCoA, is encoded in the ChAT gene at 10q11.23. The first intron of the ChAT gene encompasses the open reading frame encoding the vesicular acetylcholine transporter (VAChT), which is responsible for the transportation of ACh from the cytoplasm into the synaptic cleft. Mutations in ChAT and/or VAChT (SLC18A3) represent potential susceptibility to AD [48]. Acetylcholinesterase (AChE; EC 3.1.1.7) is a serine hydrolase that hydrolyzes ACh to produce choline and acetate in the synaptic cleft; residual choline is recycled by the choline transporter and available at the presynaptic level for de novo synthesis of ACh. AChE variants with catalytic activity include synaptic AChE (AChE-S) or tailed AChE (AChE-T) (exon 6), the most frequent variant in the brain; erythrocyte AChE (AChE-E) or hydrophobic AChE (AChE-H) (exon 5); and read-through AChE (AChE-R) (intron 4-exon 5). The Yt erythrocyte blood group antigen system is inserted in the AChE molecule (AChE, His322Asn). A 4-bp deletion located 17 kb upstream of the transcription start site that abolishes 1 of 2 adjacent hepatocyte nuclear factor-3 (HNF3) binding sites causes hypersensitivity to AChEIs and severe CNS symptoms under low-dose exposure to pyridostigmine. AChE activity is decreased in AD brains and APP is involved in the regulation of AChE [48, 97]. Human serum cholinesterase (acylcholine acylhydrolase) or butyrylcholinesterase (BuChE) (EC 3.1.1.8) is a serine hydrolase that catalyzes the hydrolysis of ACh and choline esters such as the muscle relaxants succinylcholine and mivacurium. BuChE is a 574-amino acid protein encoded by the butyrylcholinesterase (BCHE) gene (4 exons, 64 kb) at 3q26.1. Over 30 genetic variants of BCHE have been described. In AD cases, the allelic frequency of the K-variant is >0.2 (vs. 0.09 in controls), and the risk for AD in carriers of the K-variant increases in the presence of the APOE-4 allele in most studies. In mild cognitive impairment (MCI) cases, BCHE-K and APOE-4 accelerate cognitive decline, hippocampal volumetric loss, and progression to AD [98]. The choline transporter (CHT, CHT1) is encoded by the SLC5A7 (solute carrier family 5 (choline transporter), member 7) gene (2q12.3), with 9 exons spanning 25 kb. CHT is a Na+- and Cl-dependent high-affinity 580-amino acid protein (63.2 kD) with 12 transmembrane (TM) domains responsible for the uptake of choline for ACh synthesis in cholinergic neurons. An A-to-G transition at nucleotide 265, resulting in an Ile89-to-Val substitution (I89V) within the third TM domain, reduces the maximum rate of choline uptake by about 40%. The high-affinity choline transporter (CHT1, SLC5A7) expressed in cholinergic neurons represents a rate-limiting step for ACh synthesis. Disruption of CHT function lowers choline uptake and ACh synthesis, with the

346

Ramo´n Cacabelos et al.

consequent impairment in cholinergic neurotransmission. CHT1 dysfunction may contribute to AD pathology. Aβ decreases choline uptake activity and cell surface CHT protein levels. CHT trafficking is different in wild-type APP (APPwt) and in Swedish mutant APP (APPSwe) SH-SY5Y human neuroblastoma cells. APP-CHT interaction is decreased in APPSwe [99]. Mutations in PSEN1 may regulate cholinergic signaling via CHT1. Cortical neurons express active CHT1, and CHT1mediated choline uptake activity is reduced in PSEN1 M146V mutant knock-in mice [100]. In a mouse model of scopolamineinduced amnesia, scopolamine decreases ChAT, CHT, vesicular ACh transporter (VAChT), and M1 muscarinic ACh receptor (M1R) expression in the septum and hippocampus [48]. The vesicular acetylcholine transporter (VAChT) is encoded by the SLC18A3 (solute carrier family 18 (vesicular acetylcholine), member 3) gene (10q11.23). This gene encodes a transmembrane protein that transports ACh into presynaptic secretory vesicles to be released at cholinergic terminals in the CNS and peripheral nervous system. The SLC18A3 gene is located within the first intron of the ChAT gene. Mutations in the SLC18A3 gene cause presynaptic congenital myasthenic syndrome-21 (CMS21). Nitrosylation of VAChT is increased in the frontal cortex and hippocampus of APP/PS1 mice [101]. B6.eGFPChAT congenic mice with multiple gene copies of VAChT exhibit high VAChT protein expression in the hippocampal formation, accompanied by enhanced ACh release [102]. Mice with a targeted mutation in the SLC18A3 gene show a 40% reduction in transporter expression, with memory deficits that can be reversed with AChEIs. Decreased expression of the splicing regulator hnRNPA2/B1BACE1 causes abnormal splicing in BACE1, increased APP processing, and accumulation of soluble Aβ1–42 together with increases in GSK3, tau hyperphosphorylation, caspase-3, and neuronal death. In human brains, there is correlation between reduced levels of VAChT and hnRNPA2/B1 levels with increased tau hyperphosphorylation [103]. There is a selective loss of cholinergic terminals in the neocortex and hippocampus of double transgenic (APP-K670N/M671L + PS1-M146L) mice, with relevant alterations in VAChT. The levels of ChAT, AChE, and BuChE are similar in the hippocampus of young apoE4 and apoE3 mice. ChAT levels are lower in apoE4 than in apoE3 mice. The levels of muscarinic receptors are higher in apoE4 mice. ACh release from hippocampal slices is reduced in old apoE4 mice in parallel with reduced VAChT levels [104]. VAChT distribution in early AD decreases by about 47–62% in the cingulate cortex and parahippocampal-amygdaloid complex. The number of ChAT and VAChT neurons correlates with the severity of dementia and shows no relationship with APOE status. Cholinergic basocortical and septohippocampal pathways are

Pharmacogenomics of Alzheimer’s Disease

347

particularly damaged in AD as reflected by PET studies of the VAChT [105]. Nicotinic and muscarinic ACh receptors are the final effectors of cholinergic neurotransmission. Nicotinic ACh receptors (nAChRs) play an important role in the prefrontal cortex where cortical and subcortical inputs are integrated to execute higher activities of the CNS (learning, attention, working memory planning, decision-making, perception of reality). Mutations in the CHRNB2 or CHRNA7 genes that encode the β2 and α7 nicotinic receptor subunits lead to brain disorders, including AD [106]. α7nAChR, encoded by the CHRNA7 gene, is involved in AD pathogenesis via hypocholinergic neurotransmission and Aβ deposition. Carriers of the CHRNA7 rs7179008 variant showed decreased risk of dementia. Single-nucleotide polymorphisms (SNPs) in the CHRNA7 gene or in the fusion gene containing CHRNA7 partial duplication (CHRFAM7A) may represent a susceptibility trait to AD. CHRFAM7A-2-bp deletion or CHRNA7 SNPs (rs1514246, rs2337506, rs8027814) might be protective for AD [48, 107]. Loss of basal forebrain cholinergic neurons correlates with cognitive decline in AD. Exposure to Aβ upregulates neuronal α7nAChRs and increases neuronal excitability; α7-nAChRs mediate, in part, Aβ-induced neurotoxicity, which is prevented by the α7-nAChR antagonist methyllycaconitine or by α7 subunit gene deletion. In contrast, it appears that α7nAChR selective agonists (e.g., PHA-543613) and galantamine ameliorate Aβ-impaired working and reference memory, suggesting that α7 nAChR activation reduces Aβ-induced cognitive deficits, whereas receptor blockage increases Aβ toxicity and cognitive impairment [108]. 6.4

Metabolic Genes

Metabolic genes encode enzymes involved in phase I–II reactions in the liver and other tissues. Phase I reaction enzymes include alcohol dehydrogenases, aldehyde dehydrogenases, aldo-keto reductases, amine oxidases, carbonyl reductases, cytidine deaminases, cytochrome P450 family (CYPs) of monooxygenases (Figs. 3 and 4), cytochrome b5 reductase, dihydropyrimidine dehydrogenase, esterases, epoxidases, flavin-containing monooxygenases, glutathione reductase/peroxidases, peptidases, prostaglandin endoperoxide synthases, short-chain dehydrogenases, reductases, superoxide dismutases, and xanthine dehydrogenase. The most relevant phase II reaction enzymes include the following: amino acid transferases, dehydrogenases, esterases, glucuronosyltransferases, glutathione transferases, methyl transferases, N-acetyl transferases, thioltransferase, and sulfotransferases [73]. Metabolic genes are essential in drug biotransformation, and epigenetic changes in metabolic genes contribute to interindividual differences in drug response [90, 109]. Drug-metabolizing enzymes (DMEs) exhibit dramatic inter- and intra-individual

348

Ramo´n Cacabelos et al.

variability in expression and activity, in part due to DNA methylation. Highly variable DNA methylation was observed in 37 DME genes, 7 of which showed significant inverse correlations between DNA methylation and mRNA expression. Some DMEs may act as tumor suppressor or housekeeper based on their unique DNA methylation features [110]. Interethnic variability is an important issue in pharmacoepigenetics. Ethnicity-associated CpG sites may trigger gene expression and influence drug absorption, distribution, metabolism, and excretion, with tumor-specific patterns of methylation and gene regulation [111]. CYPs are a major source of variability in drug pharmacokinetics and pharmacodynamics. The highest expressed forms in liver are CYPs 3A4, 2C9, 2C8, 2E1, and 1A2, while 2A6, 2D6, 2B6, 2C19, and 3A5 are less abundant; CYPs 2J2, 1A1, and 1B1 are mainly expressed extrahepatically. Expression of each CYP is influenced by (i) genetic polymorphisms, (ii) ethnicity, (iii) induction by xenobiotics, (iv) regulation by endogenous factors, (v) health/disease states, (vi) age, (vii) sex, (viii) nutrition, and (ix) epigenetic regulation. Multiallelic genetic polymorphisms define CYP function and pharmacogenetic phenotypes (extensive, intermediate, poor, and ultra-rapid metabolizers) [90, 112–114]. The most relevant CYPs involved in the metabolism of common drugs are CYP2D6, CYP2C9, CYP2C19, and CYP3A4/5. Over 80% of the Caucasian population are deficient metabolizers for the CYP2D6/2C19/2C9 trigenic cluster (Fig. 4); and for the CYP2D6/2C19/2C9/3A4 tetragenic cluster, more than 90% of the subjects may exhibit a deficient metabolizer geno-phenotype [11]. These four CYP genes encode enzymes responsible for the metabolism of 60–80% of drugs of current use, showing ontogenic, age-, sex-, circadian- and ethnic-related differences [11, 51, 114, 115]. The CYP2D6 gene encodes a 55.73-kDa (497 aa) monooxygenase that metabolizes antiarrhythmic drugs (sparteine and propafenone), amitriptyline and debrisoquine, catalyzes the demethylation of 7-ethoxy-coumarin, and is implicated in electron transfer, detoxification, and oxidative metabolism of drugs and xenobiotics. At least 141 allelic variants have been identified, 10 of which are clinically relevant (-100C>T, -1023C>T, -1659G>A, -1707delT, -1846G>A, -2549delA, -26132615delAGA, -2850C>T, -2988G>A, -3183G>A). The most frequent genotypes in the Caucasian population are CYP2D6*1/ *1, *1/*3, *1/*4, *1/*5, *1/*6, *1/*41, *3/*3, *3/*4, *3/*6, *4/*4, *4/*5, *4/*6, *4/*41, *5/*5, *5/*6, *41/*41, *1/*1xN, *3/*1xN, *4/*1xN, *5/*1xN, and *41/*1xN (Fig. 3), corresponding to the following allelic variants: 2D6*1 (wild type), 2D6*3 (rs35742686), 2D6*4 (rs3892097), 2D6*5 (deletion), 2D6*6 (rs5030655), 2D6*41 (rs28371725), and 2D6*1xN (duplication). These genotypes confer the condition

Pharmacogenomics of Alzheimer’s Disease

349

of extensive metabolizer (EM) (*1/*1, *3/*1xN, *4/1xN, *5/*1xN) (58.17%), intermediate metabolizer (IM) (*1/*3, *1/*4, *1/*5, *1/*6, *1/*41) (31.65%), poor metabolizer (PM) (*3/*3, *3/*4, *4/*4, *4/*5, *4/*6, *4/*41, *5/*5, *41/*41) (4.88%), and ultra-rapid metabolizer (UM) (*1/*1xN) (5.29%) phenotypes in the Caucasian population. In the Iberian population (N ¼ 3175; 1650 females; 1525 males) (Fig. 3), there are significant sex-related differences in the CYP2D6*4/*41 ( p < 0.04), *5/*5 ( p < 0.05), *5/*6 ( p < 0.03), *41/*41 ( p < 0.05), and *1/*1xN genotypes ( p < 0.05). The number of CYP2D6-UMs is higher in males (6.16%) than in females (4.48%) ( p < 0.05) (Fig. 3) [16]. The distribution and frequency of PMs show great ethnic and geographic variability. PMs account for 7.86% in Europeans, 7.27 in Polynesians, 6.73% in Africans, 2.89% in Americans, 2.09% in Asians, 0.94% in Orientals, and 2.27% in Mid-Eastern populations [116]. Only 1% of the Turkish population is CYP2D6-PM (>4% in Spain; >10% in Africa) [117]. CYP2D6 variants are associated with 217 diseases, and 995 drugs are CYP2D6-related (218 major substrates, 174 minor substrates, 75 strong inhibitors, 183 moderate inhibitors, 32 weak inhibitors, and 18 inducers) [11, 16, 118]. The CYP2C9 gene encodes a microsomal 55.65-kDa (490 aa) monooxygenase that oxidizes a variety of structurally unrelated compounds including steroids and fatty acids, hydroxylates many xenobiotics (phenytoin, S-warfarin, and tolbutamide), and is involved in an NADPH-dependent electron transport pathway. Over 480 allelic variants have been identified. The most relevant alleles are 2C9*1 (wild type), 2C9*2 (rs1799853), 2C9*3 (rs1057910), and 2C9*5 (rs28371686). Major CYP2C9 genophenotypes in the Caucasian population are CYP2C9*1/*1 (61.08%), *1/*2-IM (23.77%), *1/*3-IM (10.43%), *2/*2-PM (2.37%), *2/*3-PM (1.80%), and *3/*3-PM (0.55%) (Fig. 3). The frequencies of the CYP2C9 phenotypes are 61.08% EMs, 34.20% IMs, and 4.72% PMs, with no sex-related differences between females and males in either genotypes or phenotypes. According to the World Guide for Drug Use and Pharmacogenomics [118], CYP2C9 variants are associated with 264 diseases, and 710 drugs are CYP2C9-related (180 major substrates, 141 minor substrates, 105 strong inhibitors, 183 moderate inhibitors, 95 weak inhibitors, and 49 inducers) [11, 16, 118]. The CYP2C19 gene encodes a 55.93-kDa (490 aa) monooxygenase that hydroxylates mephenytoin and many other xenobiotics (i.e., omeprazole). This gene exhibits 541 allelic variants. Important alleles are CYP2C19*1 (wild type), 2C19*2 (rs4244285), and 2C19*17 (rs12248560). In the Caucasian population, the most prevalent geno-phenotypes are the following: CYP2C19*1/*1-EM (69.63%), *1/*2-IM (23.68%), *2/*2-PM

350

Ramo´n Cacabelos et al.

(1.44%), *1/*17-UM (4.13%), *2/*17*-EM (0.67%), and *17/ *17-UM (0.45%). The frequencies of major phenotypes are 70.27% EMs, 23.77% IMs, 1.41% PMs, and 4.45% UMs, with no significant differences between females and males [11, 16, 118]. CYP2C19 variants are associated with 230 diseases, and 588 drugs are CYP2C19-related (150 major substrates, 139 minor substrates, 70 strong inhibitors, 131 moderate inhibitors, 132 weak inhibitors, and 26 inducers) [118]. Another important CYP gene with clinical relevance is CYP3A4/5 on 7q21.1 which encodes a 57.34-kDa (503 aa) monooxygenase that catalyzes the metabolism of various exogenous and endogenous chemicals. Human CYP3A4 is the most abundant hepatic and intestinal phase I enzyme that metabolizes approximately 50% of marketed drugs. There are 347 allelic variants for the CYP3A4 gene. Major CYP3A5 geno-phenotypes are CYP3A*1/*1-RM (rapid metabolizer) (1.32%), *1/*3-IM (15.69%), and *3/*3-EM (82.99%), with no sex-related differences [11, 16] (Fig. 3). A loss-of-function allele (CYP3A4*20) is present in 1.2–3.8% of the Spanish population (founder effect) [119]. CYP3A4/5 variants are associated with 447 diseases, and a number of important drugs (>1950) have been identified as substrates (908 major substrates, 144 minor substrates), inhibitors (143 strong, 443 moderate, and 127 weak inhibitors), or inducers (250 drugs) [11, 118]. The ability of drugs to act as inducers, inhibitors, or substrates for CYP3A is predictive of whether concurrent administration of these compounds with a known CYP3A substrate may alter drug disposition, efficacy, or toxicity (Fig. 6). CYP3A4 substrates considerably overlap with those of ABCB1 (Fig. 5). To date, the identified clinically important CYP3A4 inhibitors mainly include macrolide antibiotics (e.g., clarithromycin and erythromycin), anti-HIV agents (e.g., ritonavir and delavirdine), antidepressants (e.g., fluoxetine and fluvoxamine), calcium channel blockers (e.g., verapamil and diltiazem), steroids and their modulators (e.g., gestodene and mifepristone), and several herbal and dietary components. Many of these drugs are also mechanismbased inhibitors of CYP3A4, which involve formation of reactive metabolites, binding to CYP3A4, and irreversible enzyme inactivation [112, 118]. Other CYPs with high impact in disease pathogenesis and drug metabolism are CYP1A1 (465 diseases; 684 drugs: 100 major substrates, 62 minor substrates, 35 strong inhibitors, 44 moderate inhibitors, 6 weak inhibitors, and 114 inducers), CYP1A2 (396 diseases; 1097 drugs: 198 major substrates, 159 minor substrates, 119 strong inhibitors, 145 moderate inhibitors, 102 weak inhibitors, and 154 inducers), CYP2B6 (297 diseases; 416 drugs: 92 major substrates, 71 minor substrates, 44 strong inhibitors,

Pharmacogenomics of Alzheimer’s Disease

351

49 moderate inhibitors, 21 weak inhibitors, and 76 inducers), and CYP2E1 (419 diseases; 623 drugs: 111 major substrates, 78 minor substrates, 63 strong inhibitors, 49 moderate inhibitors, 62 weak inhibitors, and 71 inducers) [118]. CYP2C8 is involved in the metabolism of over 100 drugs (87 are major substrates and 52 are minor substrates; 31 dugs act as strong inhibitors, 45 are moderate inhibitors, and 27 are weak inhibitors; 30 drugs are inducers of CYP2C8 enzymes) [118]. Amodiaquine, cerivastatin, dasabuvir, enzalutamide, imatinib, loperamide, montelukast, paclitaxel, pioglitazone, repaglinide, and rosiglitazone are typical substrates of CYP2C8. Gemfibrozil is a strong, irreversible inhibitor of CYP2C8, and the acyl-β-glucuronides of gemfibrozil and clopidogrel cause metabolism-dependent inactivation of CYP2C8, with a high risk for drug interactions. Other glucuronide metabolites frequently interact with CYP2C8 as substrates or inhibitors. Some CYP2C8 genetic variants are associated with altered CYP2C8 activity, showing interethnic differences. CYP2C8 variants contribute to frequent ADRs and drug-drug interactions with clinical relevance [120]. In certain ethnic groups, such as Ashkenazi Jews, CYP2A6, CYP2C9, NAT2, and VKORC1 variants may cause different ADRs in response to many common drug substrates [121]. In Papua New Guinea, the frequencies of some common SNPs are higher than anywhere, worldwide. The frequencies of CYP2B6*6, CYP2C19*2, and CYP2C19*3 are over 50%, 40%, and 25%, respectively. In contrast, CYP2A6*9, 2B6*2, 2B6*3, 2B6*4, and 2B6*18, and 2C8*3 are much lower than in Caucasians [122]. The distribution and frequency of CYP variants in Africa are substantially different to that of Asian and Caucasian populations, especially regarding CYP2B6*6, CYP2C8*2, CYP2D6*3, CYP2D6*17, CYP2D6*29, CYP3A5*6, and CYP3A5*7 [123]. In Portugal, regional differences were found in CYP2C9 rs1057910, CYP2D6 rs3892097, MTHFR rs1801133, F5 rs6025, ADH1B rs2066702, ADH1B rs1229984, NAT2 rs1799931, and VKORC1 rs9923231. In this cohort, 18.9% of the population is IM or PM for at least three drugs and 84.6% show a few ADRs [124]. 6.5 Transporter Genes

Transporter genes encode proteins that regulate the influx-efflux of xenobiotics through cell membranes and biological barriers (blood-brain barrier, placental membranes, tumor barriers, etc.). Transporter proteins are classified into four major categories, including: (i) ATPases such as P-type (ATPs), V-type (vacuolar H+-ATPase subunit), and F-type; (ii) ATP-binding cassette transporters subfamily A (ABCAs) (ABCA1–13), subfamily B (MDR/TAP) (ABCBs) (ABCB1, TAP1–2, ABCB4–11), subfamily C (CFTR/MRP) (ABCCs) (ABCC1–6, CFTR, ABCC8–13), subfamily D (ALD) (ABCDs) (ABCD1–4), subfamily E (OABP) (ABCEs) (ABCE1), subfamily F (GCN20) (ABCFs) (ABCF1–3),

352

Ramo´n Cacabelos et al.

and subfamily G (WHITE) (ABCGs) (ABCG1–5, ABCG8); (iii) solute carriers such as high-affinity glutamate and neutral amino acid transporter family (SLC1) (SLC1A1–7), facilitative GLUT transporter family (SLC2) (SLC2A1–14), heavy subunits of the heteromeric amino acid transporters (SLC3) (SLC3A1–2), bicarbonate transporter family (SLC4) (SLC4A1–11), sodium glucose cotransporter family (SLC5) (SLC5A1–12), sodium- and chloridedependent neurotransmitter transporter family (SLC6) (SLC6A1–20), cationic amino acid transporter/glycoprotein-associated family (SLC7A1–14), Na+/Ca2+ exchanger family (SLC8) (SLC8A1–3, SLC8B1), Na+/H+ exchanger family (SLC9) (SLC9A1–10, SLC9B1–2), sodium bile salt cotransport family (SLC10) (SLC10A1–7), proton-coupled metal ion transporter family (SLC11) (SLC11A1–2), electroneutral cation-coupled Cl cotransporter family (SLC12) (SLC12A1–9), human Na+-sulfate/carboxylate cotransporter family (SLC13) (SLC13A1–5), urea transporter family (SLC14) (SLC14A1–2), proton oligopeptide cotransporter family (SLC15) (SLC15A1–4), monocarboxylate transporter family (SLC16) (SLC16A1–14), vesicular glutamate transporter family (SLC17) (SLC17A1–9), vesicular amine transporter family (SLC18) (SLC18A1–3, SLC18B1), folate/thiamine transporter family(SLC19) (SLC19A1–3), Type III Na+-phosphate cotransporter family (SLC20) (SLC20A1–2), organic anion transporter family (SLCO/SLC21) (SLCO1A2, SLCO1B1, SLCO1B3, SLCO1C1, SLCO2A1, SLCO2B1, SLCO3A1, SLCO4A1, SLCO4C1, SLCO5A1, SLCO6A1), organic cation/anion/zwitterion transporter family (SLC22) (SLC22A1–25), Na+-dependent ascorbic acid transporter family (SLC23) (SLC23A1–3), Na+/(Ca2+K+) exchanger family (SLC24) (SLC24A1, SLC24A2–5), mitochondrial carrier family (SLC25) (SLC25A1–6, UCP1–3, SLC25A10–53), multifunctional anion exchanger family (SLC26) (SLC26A1–11), fatty acid transporter family (SLC27) (SLC27A1–6), Na+-coupled nucleoside transport family (SLC28) (SLC28A1–3), facilitative nucleoside transporter family (SLC29) (SLC29A1–4), zinc efflux family (SLC30) (SLC30A1–10), copper transporter family (SLC31) (SLC31A1–2), vesicular inhibitory amino acid transporter family (SLC32) (SLC32A1), acetyl-CoA transporter family (SLC33) (SLC33A1), Type II Na+-phosphate cotransporter family (SLC34) (SLC34A1–3), nucleoside-sugar transporter family (SLC35) (SLC35A1–5, SLC35B1–4, SLC35C1–2, SLC35D1–3, SLC35E1–4, SLC35F1–6, SLC35G1–6), proton-coupled amino acid transporter family (SLC36) (SLC36A1–4), glycerol-3-phosphate transporter family (SLC37) (SLC37A1–4), system A and system N sodium-coupled neutral amino acid transporter family (SLC38) (SLC38A1–11), metal ion transporter family (SLC39) (SLC39A1–14), basolateral iron transporter family (SLC40) (SLC40A1), MgtE-like magnesium transporter family (SLC41)

Pharmacogenomics of Alzheimer’s Disease

353

(SLC41A1–3), Rh ammonium transporter family (SLC42) (RHAG, RHBG, RHCG), Na+-independent, system-L-like amino acid transporter family (SLC43) (SLC43A1–3), choline-like transporter family (SLC44) (SLC44A1–5), putative sugar transporter family (SLC45) (SLC45A1–4), folate transporter family (SLC46) (SLC46A1–3), multidrug and toxin extrusion (MATE) family (SLC47) (SLC47A1–2), heme transporter family (SLC48) (SLC48A1), FLVCR-related transporter family (SLC49) (FLVCR1–2), sugar efflux transporters (SLC50) (SLC50A1), transporters of steroid-derived molecules (SLC51) (SLC51A–B), and riboflavin transporter family (RFVT/SLC52) (SLC52A1–3); and (iv) a miscellaneous group of transporters represented by aquaporins (AQP1, AQP7, AQP9), major vault protein (MVP), and metallothioneins [73, 118]. Genetic variation in transporter genes affects drug metabolism, brain penetrance, and drug resistance [125]. Mutations in ABC transporters are associated with AD pathogenesis [126]. ABCB1 (Fig. 5) is probably the most important drug transporter in the CNS. ABCB1 variants are associated with 445 diseases, and 1230 drugs are ABCB1-related (478 substrates, 618 inhibitors, and 183 inducers) [11, 16, 118]. Genetic variation in ABCB1 contributes to the accumulation and progression of Aβ deposits in the AD brain [126]. Conversely, the cholesterol transporter ABCA1 neutralizes Aβ aggregation and facilitates Aβ elimination from brain tissues [127]. Other ABCs show association with AD [126, 128], such as the G allele of the ABCA7 rs115550680 SNP or variants in the ABCG2 gene (C421A; rs2231142) (ABCG2 C/C genotype) [13, 129]. P-glycoprotein (P-gp, ABCB1) is an ATP-binding cassette transporter highly expressed on the luminal side of the bloodbrain barrier, with a fundamental role in facilitating Aβ clearance from the brain and in the PGx process of over 1400 drugs, including AChEIs and memantine [127, 130]. For example, the organophosphate pesticide chlorpyrifos is a substrate of P-gp [131]. The absence of P-gp causes a remarkable decrease in Aβ removal from the brain and increased intraparenchymal cerebral amyloid angiopathy [132]. Some ABCB1 genetic variants (rs1128503 (C1236T), rs2032582 (G2677T/A), rs1045642 (C3435T)) are associated with AD, especially the rs1045642 polymorphism (C/C genotype) [129]. ABCB1 is one of the most clinically relevant ABC transporters. The ABCB1 gene encodes a large transmembrane protein (P-glycoprotein-1/multiple drug resistance-1; P-gp) (141.48 kDa; 1280 aa), which is an integral part of the blood-brain barrier and functions as a drug-transport pump transporting a variety of drugs from the brain back into the blood. This ABC transporter (traffic ATPase) is an energy-dependent efflux pump responsible for decreased drug accumulation in multidrug-resistant cells. About

354

Ramo´n Cacabelos et al.

1630 SNPs and over 60 haplotypes have been identified. In the Iberian population (N ¼ 819; 398 females; 421 males), 37 haplotypes were detected, among which the most frequently found were CGC/TTT-1/2-IR (30.04%), CGC/CGC-1/1-HR (20.88%), TTT/TTT-2/2-LR (9.16%), TTT/cGT-2-LR (5.74%), and TTT/TGC-2-LR (3.54%) (Fig. 5). These haplotypes yield six different haplo-phenotypes: H1-HR (high resistance) (20.15%), H2-LR (low resistance) (14.53%), H1/1-HR (20.88%), H1/2IR (intermediate resistance), H2/2-LR (low resistance), and 11 novel haplotypes (Hx) with a potential LR phenotype (5.25%) (Fig. 5). Interestingly, significant differences between females and males were found among H1/2-IR ( p < 0.05) and H2/2-LR carriers ( p < 0.005). About 12% of males harbor an H2/2-LR haplo-phenotype (females: 6%); in contrast, 33% of females are carriers of an H1/2-IR haplo-phenotype (males: 26%) (Fig. 5). Globally, ABCB1 phenotypes are 28.94% LR, 30.04% IR, and 41.03% HR, with males showing a lower frequency of IR phenotypes (26.37%) than females (33.92%) ( p < 0.05) [16] (Fig. 5). AD patients suffer abnormalities in neuronal glucose metabolism, the main source of cerebral choline. Since cell membranes are impermeable to glucose, glucose is transferred across membranes by specific sodium-independent glucose transporters (GLUTs), encoded by SLC2A genes, and sodium-dependent glucose transporters (SGLTs), encoded by SLC5 genes. GLUT1 and GLUT3 levels are decreased in the neocortex of AD patients [133]. EAAT2 (excitatory amino acid transporter 2) is the predominant astrocyte glutamate transporter responsible for synaptic glutamate reuptake in the CNS; epigenetic dysregulation of EAAT2 may be involved in AD pathogenesis and pharmacoepigenetics [134, 135]. Alterations in MCT2 (monocarboxylate transporter 2), encoded by the SLC16A7 gene and involved in brain energy metabolism, may also affect AD pathogenesis and PGx [136]. Other transporters with recent evidence supporting a potential role in AD pathogenesis include the following: astrocytic transporters [137]; the urate transporter GLUT9 (SLC2A9 gene) [138]; the sarco-/ endoplasmic reticulum (SR/ER) calcium (Ca2+)-ATPase (SERCA) pump, responsible for preserving low cytosolic Ca2+ levels [139]; components (CHMP6, CHMP2A, CHMP2B) of the endosomal sorting complexes required for the transport (ESCRT) pathway involved in tau propagation in frontotemporal dementia [140]; endolysosomal transporters (e.g., PSD2, TCIRG1, RIN3, and RUFY1) [141]; mitochondrial Na+/Ca2+ exchanger (NCLX, Slc8b1 gene) whose deletion accelerates memory decline and increases amyloidosis and tau pathology in 3xTgAD mice [142]; Na+/Ca2+ exchangers (NCXs) (NCX1, NCX2, and NCX3) [143]; the potassium channels (Kv1.3 and Kir2.1) [144]; the translocator protein (TSPO), with multifunctional activity at the outer mitochondrial membrane and the voltage-

Pharmacogenomics of Alzheimer’s Disease

355

dependent anion channel (VDAC), which mediates the flux of ions, Ca2+, nucleotides, and metabolites across the mitochondrial membrane [145]; the voltage-dependent anion channel 1 (VDAC1), a hub protein that interacts with over 150 other proteins and AD-related pathogenic elements (e.g., phosphorylated tau, Aβ, γ-secretase) [146]; components of the translocase of the outer (TOMM) and inner mitochondrial membrane (TIMM) complexes (TIMM8A (DDP) and DNAJC19 (TIMM14), TOMM40) [147]; the adaptor protein Alix which is a co-factor for the interaction between the E3-ubiquitin ligase NEDD4-1 and the ABC transporter targets ABCG1 and ABCG4 [148]; soluble Nethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins and vesicle-associated SNARE (v-SNARE) protein Sec22 (SEC22 mutations associate with atherosclerosis and AD-related phenotypes) [149]; 5-hydroxytryptamine transporter gene-linked polymorphic region (5-HTTLPR), a variable number of tandem repeats in the promoter region of the serotonin transporter encoding gene that affects transcription and delusions in AD [150]; and several transporters with SNPs associated with AD, including ABCA1 [151], ABCA2 [152, 153], ABCA7 [154– 158], ABCB1 [159], ABCC9 [160], ABCG1, ABCG2, and ABCG4 [129, 161]; copper-transporting P-type adenosine triphosphatase (ATP7B) [162]; uncoupling protein-4 (UCP4/ SLC25A27) [163]; SLC2A14 [164]; SLC6A3 (DAT1, dopamine transporter) [165, 166]; SLC6A4 (serotonin transporter) [167]; SLC17A4 [168], aquaporins [169, 170]; neuronal ceroid lipofuscinosis genes linked to the intracellular trafficking pathway retromer (CLN3, CLN5, CTSD) [171]; folate transporter [172]; proton myo-inositol cotransporter (SLC2A13) [173]; and glutamate transporters EAAC1 (SLC1A1), GLT1 (SLC1A2), GLAST (SLC1A3), EAAT4 (SLC1A6), and EAAT5 (SLC1A7) [48, 174]. In some of them (ABCB1, ABCC5, ABCG2, pregnane X receptor (PXR), constitutive androstane receptor (CAR)), Aβ1–42 alters their activity and expression, contributing to altered amyloid processing and transport [175]. Many of these defective transporters are attractive candidates for therapeutic intervention in AD and their mutations should be taken into account in PGx strategies [176]. Some small compounds are effective in regulating APOE-related lipid metabolism and transporter function in animal models. For instance, the chrysanthemic ester 82879 increases ApoE and stimulates liver x receptor (LXR) target genes such as ABCA1, LXRα, and inducible degrader of low-density lipoprotein receptor (IDOL) [177]. N-(2,2,2-Trifluoroethyl)-N-[4-[2,2,2-trifluoro-1-hydroxy-1-(trifluoromethyl)ethyl]phenyl]-benzenesulfonamide (T0901317) is a potent activator of PXR, the nuclear receptor that controls P-gp expression. T0901317 and a series of N-triazolyl-methylene-linked benzenesulfonamides induce P-gp (ABCB1) [48, 178].

356

Ramo´n Cacabelos et al.

6.6 Pharmacogenetics of Acetylcholinesterase Inhibitors 6.6.1

Donepezil

Donepezil is the most prescribed AChEI for the treatment of AD [48, 49, 179–181] (Table 8). Donepezil is a selective AChEI with a long elimination half-life (t1/2) of 70 h and is metabolized in the liver [182]. Donepezil clearance is 7.3 l h1 with gender- and interindividual variability (30%) [183]. Donepezil increases brain ACh levels by 35% and decreases AChE activity by 40–90%, with no effect on ChAT, vesicular ACh transporter, CHT, or muscarinic receptors. Donepezil may also exert beneficial effects against Aβ1–40-induced neurotoxicity. It is a major substrate of CYP2D6, CYP3A4, AChE, and UGTs, inhibits AChE and BCHE, and is transported by ABCB1 [118, 182, 183] (Table 8). Individual variation in metabolic genes (CYP2D6) and pathogenic genes (APOE) modulates the response to donepezil treatment [184]. Several CYP2D6 variants may modify donepezil efficacy and safety in AD [139, 140]; and APOE and CYP2D6 variants are determinant of the effects of donepezil [179, 180, 184]. APOE-4 carriers tend to be the worst responders and APOE-3 carriers are the best responders to donepezil in either monotherapy or drug combination regimes; CYP2D6-EMs are the best responders and CYP2D6-PMs are the worst responders [10, 11, 48, 50, 51, 118, 179]. ABCA1 regulates cholesterol transport and APOE metabolism. AD patients with the ABCA1 rs2230806 G/G genotype respond better to donepezil than carriers of the A/A and A/G genotypes; and ABCA1 rs2230806 G/G-APOE3 non-carriers show a better clinical response to donepezil. Lower plasma donepezil concentration-to-dose ratios and better clinical response to donepezil have been reported in patients homozygous for the T/T/T genotype in the ABCB1 haplotypes 1236C/2677G/3435C (46%) and 1236T/2677T/3435T (41%). Donepezil may inhibit ABCB1 [48]. APOE-ε4/BCHE-K* carriers show an earlier age of onset, an accelerated cognitive decline, and a differential response to donepezil therapy. In patients with MCI, donepezil accelerates cognitive decline in homozygous BCHE-K and APOE-4 carriers. The BCHE-K variant is associated with lower AChE hydrolyzing activity, and BuChE activity increases in parallel with disease progression. These results suggest that BCHE-K and APOE4 carriers should not be prescribed donepezil for MCI and, consequently, donepezil is not recommended in AD patients with the BCHE-K and/or APOE-4 variants [185]. Donepezil may induce upregulation of α7nAChR protein levels, potentially protecting neurons against neurodegeneration. CHRNA7 rs8024987 (C/G) and rs6494223 (C/T) respond better to donepezil. Donepezil-induced α7nAChR upregulation is higher in T/T carriers (7–15%) than in C/C or C/T carriers [186].

Pharmacogenomics of Alzheimer’s Disease

357

Table 8 Pharmacological properties and pharmacogenetics of conventional anti-dementia drugs Drug

Properties

Pharmacogenetics

Name: Donepezil hydrochloride, Aricept, 120011-70-3, Donepezil HCl, BNAG, E-2020, E2020 IUPAC name: 2-[(1-Benzylpiperidin-4yl)methyl]-5,6dimethoxy-2,3dihydroinden-1-one; hydrochloride Molecular formula: C24H30ClNO3 Molecular weight: 415.9529 g/mol Category: Cholinesterase inhibitor Mechanism: Centrally active, reversible acetylcholinesterase inhibitor; increases the acetylcholine available for synaptic transmission in the CNS Effect: Nootropic agent, cholinesterase inhibitor, parasympathomimetic effect

Pathogenic genes: APP, APOE, CHAT Mechanistic genes: ACHE, BCHE, CHAT, CHRNA7 Drug metabolismrelated genes: Substrate: CYP2D6 (major), CYP3A4 (major), UGTs, ACHE Inhibitor: ABCB1, ACHE, BCHE, hERG Transporter genes: ABCB1, ABCA1, ABCG2, SCN1A Pleiotropic genes: APOE, PLP, MAG, MBP, CNPase, MOG

Name: Galantamine hydrobromide, galanthamine hydrobromide, 1953-044, Nivalin, Razadyne, UNII-MJ4PTD2VVW, Nivaline IUPAC name: (1S,12S,14R)-9methoxy-4-methyl-11oxa-4-azatetracyclo [8.6.1.0^{1,12}.0^ {6,17}]heptadeca-6,8,10 (17),15-tetraen-14-ol Molecular formula: C17H22BrNO3 Molecular weight: 368.26548 g/mol Category: Cholinesterase inhibitor Mechanism: Reversible and competitive acetylcholinesterase inhibition leading to an

Pathogenic genes: APOE, APP Mechanistic genes: ACHE, BCHE, CHRNA4, CHRNA7, CHRNB2, SLC18A3 Drug metabolismrelated genes: Substrate: ABCB1, CYP2D6 (major), CYP3A4 (major), UGT1A1 Inhibitor: ACHE, BCHE Transporter

(continued)

358

Ramo´n Cacabelos et al.

Table 8 (continued) Drug

Properties increased concentration of acetylcholine at cholinergic synapses; modulates nicotinic acetylcholine receptor; may increase glutamate and serotonin levels Effect: Nootropic agent, cholinesterase inhibitor, parasympathomimetic effect

Pharmacogenetics genes: ABCB1, SLC18A3

Name: Memantine Pathogenic hydrochloride, 41100genes: APOE, 52-1, Namenda, MAPT, PSEN1 Memantine HCL, Axura, Mechanistic 3,5-Dimethyl-1genes: adamantanamine CHRFAM7A, hydrochloride, DLGAP1, FOS, 3,5-dimethyladamantanGRIN2A, 1-amine hydrochloride GRIN2B, IUPAC name: GRIN3A, 3,5-DimethyladamantanHOMER1, 1-amine;hydrochloride HTR3A Molecular formula: Drug C12H22ClN metabolismMolecular weight: related genes: 215.76278 g/mol Inhibitor: Category: N-Methyl-DCYP1A2 aspartate receptor (weak), antagonist CYP2A6 Mechanism: Binds (weak), preferentially to NMDA CYP2B6 receptor-operated cation (strong), channels; may act by CYP2C9 blocking actions of (weak), glutamate, mediated in CYP2C19 part by NMDA receptors (weak), Effect: Dopamine agent, CYP2D6 antiparkinson agent, (strong), excitatory amino acid CYP2E1 antagonist, antidyskinetic (weak), CYP3A4 (weak), NR1I2 Transporter genes: NR1I2 Pleiotropic genes: APOE, MAPT, MT-TK, PSEN1

(continued)

Table 8 (continued) Drug

Properties

Pharmacogenetics

Name: Rivastigmine tartrate, 129101-54-8, SDZ-ENA 713, rivastigmine hydrogentartrate, rivastigmine hydrogen tartrate, ENA 713, ENA-713 IUPAC name: (2R,3R)2,3Dihydroxybutanedioic acid;[3-[(1S)-1(dimethylamino)ethyl] phenyl] N-ethyl-Nmethylcarbamate Molecular formula: C18H28N2O8 Molecular weight: 400.42352 g/mol Category: Cholinesterase inhibitor Mechanism: Increases acetylcholine in CNS through reversible inhibition of its hydrolysis by acetylcholinesterase Effect: Neuroprotective agent, cholinesterase inhibitor, cholinergic agent

Pathogenic genes: APOE, APP, CHAT Mechanistic genes: ACHE, BCHE, CHAT, CHRNA4, CHRNB2, SLC18A3 Drug metabolismrelated genes: Substrate: UGT1A9, UGT2B7 Inhibitor: ACHE, BCHE Transporter genes: SLC18A3 Pleiotropic genes: APOE, MAPT

Name: Tacrine hydrochloride, Tacrine HCl, 1684-40-8, hydroaminacrine, tacrine HCl, 9-amino-1,2,3,4tetrahydroacridine hydrochloride, Tenakrin IUPAC name: 1,2,3,4tetrahydroacridin-9amine;hydrochloride Molecular formula: C13H15ClN2 Molecular weight: 234.7246 g/mol Category: Cholinesterase inhibitor Mechanism: Elevates acetylcholine in cerebral cortex by slowing degradation of acetylcholine Effect: Nootropic agent, cholinesterase inhibitor, parasympathomimetic effect

Pathogenic genes: APOE Mechanistic genes: ACHE, BCHE, CHRNA4, CHRNB2 Drug metabolismrelated genes: Substrate: CYP1A2 (major), CYP2D6 (minor), CYP3A4 (major), CES1, GSTM1, GSTT1 Inhibitor: ACHE, BCHE, CYP1A2 (weak) Transporter genes: ABCB4, SCN1A

(continued)

360

Ramo´n Cacabelos et al.

Table 8 (continued) Drug

Properties

Pharmacogenetics Pleiotropic genes: APOE, LEPR, MTHFR

Name: ()-Huperzine A, huperzine A; huperzineA; 102518-79-6; ()huperzine A; (+/)huperzine A IUPAC name: (1R,9R,13E)-1-amino13-ethylidene-11methyl-6-azatricyclo [7.3.1.02,7]trideca-2 (7),3,10-trien-5-one Molecular formula: C15H18N2O Molecular weight: 242.32 g/mol Category: Neuroprotectant, cholinesterase inhibitor Mechanism: Increases acetylcholine in the brain by inhibiting acetylcholinesterase and slowing acetylcholine hydrolysis Effect: Neuroprotective, acetylcholinesterase inhibitor, cognitive enhancer, antiepileptic

Pathogenic genes: APP, APOE Mechanistic genes: ACHE Drug metabolismrelated genes: Substrate: ABCB1, CYP1A2, CYP3A1, CYP3A2, CYP2C11, CYP2E1, CES1, CES2 Inhibitor: ACHE Inducer: CYP1A2 Transporter genes: ABCB1, ABCG2 Pleiotropic genes: APOE, BDNF

ABCA1 ATP-binding cassette, subfamily A, member 1, ABCB1 ATP-binding cassette, subfamily B, member 1, ABCB4 ATP-binding cassette, subfamily B, member 4, ABCG2 ATP-binding cassette, subfamily G, member 2, ACHE Acetylcholinesterase, APOE Apolipoprotein E, APP Amyloid precursor protein, BCHE Butyrylcholinesterase, BDNF Brainderived neurotrophic factor, CES1 Carboxylesterase 1, CES2 Carboxylesterase 2, CHAT Choline acetyltransferase, CHRNA4 Cholinergic receptor, neuronal nicotinic, alpha polypeptide 4, CHRNA7 Cholinergic receptor, neuronal nicotinic, alpha polypeptide 7, CHRNB2 Cholinergic receptor nicotinic beta 2 subunit, CNPase Cyclic nucleotide phosphodiesterase, CYP1A2 Cytochrome P450, family 1, subfamily A, polypeptide 2, CYP2A6 Cytochrome P450, family 2, subfamily A, polypeptide 6, CYP2B6 Cytochrome P450, family 2, subfamily B, polypeptide 6, CYP2C1 Cytochrome P450, family 2, subfamily C, polypeptide 1, CYP2C9 Cytochrome P450, family 2, subfamily C, polypeptide 9, CYP2C11 Cytochrome P450, family 2, subfamily C, polypeptide 11, CYP2C19 Cytochrome P450, family 2, subfamily C, polypeptide 19, CYP2D6 Cytochrome P450, family 2, subfamily D, polypeptide 6, CYP2E1 Cytochrome P450, family 2, subfamily E, polypeptide 1, CYP3A1 Cytochrome P450, family 3, subfamily A, polypeptide 1, CYP3A2 Cytochrome P450, family 3, subfamily A, polypeptide 2, CYP3A4 Cytochrome P450, family 3, subfamily A, polypeptide 4, DLGAP1 discs, large (Drosophila) homolog-associated protein 1, FOS FBJ murine osteosarcoma viral oncogene homolog, GRIN2A Glutamate receptor, ionotropic, n-methyl-d-aspartate, subunit 2A, GRIN2B Glutamate receptor, ionotropic, N-methyl-D-aspartate, subunit 2B, GRIN3A Glutamate receptor, ionotropic, N-methyl-D-aspartate, subunit 3A, GSTM1 Glutathione S-transferase mu 1, HOMER1 Homer homolog 1 (Drosophila), HTR3 5-Hydroxytryptamine receptor 3, LEPT Leptin receptor, MAPT Microtubule-associated protein tau, MBP Myelin basic protein, MOG Myelin-oligodendrocyte glycoprotein, MTHFR 5,10-Methylenetetrahydrofolate reductase, NR1I2 Nuclear receptor subfamily 1, group I, member 2, PLP Proteolipid protein, PSEN1 Presenilin 1, SCN1A Sodium voltage-gated channel, alpha subunit 1, SLC18A3 Solute carrier family 18 (vesicular acetylcholine), member 3, UGT1A1 UDP glucuronosyltransferase 1 family, polypeptide A1, UGT1A9 UDP glucuronosyltransferase 1 family, polypeptide A9, UGT2B7 UDP glucuronosyltransferase 2 family, polypeptide B7

Pharmacogenomics of Alzheimer’s Disease

361

Donepezil treatment tends to increase APP forms in APOE-4 non-carriers and interacts with many drugs, causing cardiotoxicity [48]. 6.6.2

Galantamine

Galantamine is a reversible, competitive AChEI and an allosteric modulator of nicotinic ACh receptors. This drug is rapidly absorbed (Cmax ¼ 1 h), with low protein binding (28.3–33.8%), a steady-state volume of distribution (Vss) of 193 L, and an elimination half-life of 7–8 h (20–25% is excreted unchanged in urine) [182]. Median clearance in male and female patients with AD is 14.8 and 12.4 L/h, respectively, probably due to body weight differences rather than a real gender effect. Metabolic clearance is reduced by 60% in patients with liver dysfunction. Galantamine increases the levels of brain VAChT. Galantamine is a major substrate of CYP2D6, CYP3A4, ABCB1, and UGT1A1 and an inhibitor of AChE and BCHE. APOE, APP, AChE, BCHE, CHRNA4, CHRNA7, and CHRNB2 variants may also affect galantamine efficacy and safety [48, 118, 182] (Table 8). Galantamine is mainly metabolized by CYP2D6 and CYP3A4 enzymes. Major metabolic pathways are glucuronidation, O-demethylation, N-demethylation, N-oxidation, and epimerization. CYP2D6 variants are major determinants of galantamine pharmacokinetics, with CYP2D6PMs presenting 45% and 61% higher dose-adjusted galantamine plasma concentrations than heterozygous and homozygous CYP2D6-EMs [182, 187]; however, these pharmacokinetic changes might not substantially affect pharmacodynamics. There is no linear correlation between galantamine concentration and cognitive response in AD patients. Galantamine bioavailability and its therapeutic effects may be modified by interaction with foods and nutritional components [48]. Several studies indicate that galantamine in AD may show better results in APOE-4 non-carriers. Patients with MCI treated with galantamine for 1 year also showed a lower rate of whole brain atrophy, preferentially among APOE ϵ4 carriers. Others suggest no major influence of APOE variants in the effects of galantamine in AD. CHRNA7 rs8024987 variants may also affect galantamine in females [48].

6.6.3

Rivastigmine

Rivastigmine (ENA 713, carbamoylatine) is a dual AChEI with brain-region selectivity (>40% AChE inhibition) and a long-lasting effect. Rivastigmine also inhibits peripheral BuChE (>10%). Saturable first-pass metabolism leads to 35% bioavailability of the administered dose and nonlinear short half-life pharmacokinetics, with renal elimination. Rivastigmine is a pseudo-irreversible dual inhibitor of AChE and BCHE with a very short t1/2 (1–2 h) and longer duration of action due to blockade of AChE and BCHE for around 8.5 and 3.5 h, respectively.

362

Ramo´n Cacabelos et al.

Rivastigmine increases VAChT and ChAT expression in frontal cortex, hippocampus, striatum, and cerebellum, providing additional effects on cholinergic neurotransmission. Rivastigmine metabolism is mediated by esterases in the liver and in the intestine [118, 182, 188]. APOE, APP, ChAT, AChE, BCHE, CHRNA4, CHRNB2, and MAPT variants may affect rivastigmine pharmacokinetics and pharmacodynamics (Table 8). Rivastigmine is more effective in APOE-4 non-carriers in different ethnic groups: APOE-3 carriers are the best responders and APOE-4 carriers are the worst responders. CYP enzymes are not involved in the metabolism of rivastigmine. UGT2B7-PMs show higher rivastigmine levels with a poor response to treatment. In combination treatments with memantine, carriers of CYP2D6*3, UGT2B7, and UGT1A9*5 variants show differential responses to treatment. Two SNPs in the intronic region of ChAT (rs2177370 and rs3793790) and CHRNA7 variants may influence the response to AChEIs. The BCHE-K-variant (rs1803274) causes reduced enzyme activity and lower response to rivastigmine. Other SNPs outside the coding sequence (50 UTR (rs1126680) and/or intron 2 (rs55781031)) of the BCHE gene, in addition to the K-variant (p.A539T), may also be responsible for reduced enzyme activity. Carriers of these deleterious SNPs should receive lower doses of rivastigmine or start treatment with a different AChEI. Females with the BCHE-wt/wt show a better benefit with rivastigmine than males, and BCHE-K* male carriers show a faster cognitive decline than females. Rivastigmine may attenuate the progression of cognitive decline in male BuChE-K and in female BuChE wt/wt. BCHEK-APOE-4 carriers show poor cognitive responses to rivastigmine patch or memantine add-on therapy [48]. 6.6.4

Huperzine A

Huperzine A, a natural Lycopodium sesquiterpene alkaloid extracted from the Chinese medicinal plant Huperzia serrata, is a reversible and highly selective second-generation AChEI for AD approved in China in 1994 [189]. Huperzine A shows different pharmacokinetic features in elderly and young healthy subjects. In elderly subjects, the plasma concentration-time profile of huperzine A follows a one-compartment model with first-order absorption and elimination. Age is a covariate with significant influence on huperzine A clearance. In rat liver microsomes, huperzine A metabolism is mediated primarily by CYP1A2, with a secondary contribution of CYP3A1/ 2 and negligible involvement of CYP2C11 and 2E1. Huperzine A is excreted unchanged by the kidney rather than metabolized by human liver, with no apparent involvement of CYP enzymes (CYP1A2, 2A6, 2C9, 2C19, 2D6, 2E1, and 3A4) [190] (Table 8). At a toxicological dose administered to rats, huperzine

Pharmacogenomics of Alzheimer’s Disease

363

A may induce CYP1A2 by transcription enhancement. Carboxylesterases (CESs) (CES1 and CES2) are enzymes that catalyze the hydrolysis of ester, amide, and carbamate chemicals. CESs may be tangentially involved in huperzine A metabolism. Huperzine A penetrates the brain and interacts with ABCB1 and ABCG2 efflux transporters and is a substrate of ABCB1. In Abcb1a/ mice, the brain to plasma concentration ratio of huperzine A is higher than in wild-type animals [48]. 6.7 Pharmacogenetics of Memantine

Memantine is a non-competitive low-affinity NMDA receptor antagonist, which binds preferentially to NMDA receptor-operated cation channels. Its long t1/2 is about 70 h and is eliminated unchanged via the kidneys; however, several genes influence its efficacy and safety [182]. Memantine inhibits the actions of glutamate via NMDA receptors and antagonizes GRIN2A, GRIN2B, GRIN3A, HTR3A, and CHRFAM7A. APOE, PSEN1, and MAPT are pathogenic genes which might influence the effects of memantine in AD; and variants in some mechanistic genes (GRIN2A, GRIN2B, GRIN3A, HTR3A, CHRFAM7A, c-Fos, Homer1b, and PSD-95) may also modify its therapeutic effects. Memantine strongly inhibits CYP2B6 and CYP2D6. In contrast, CYP1A2, CYP2A6, CYP2C9, CYP2C19, CYP2E1, and CYP3A4 are weakly inhibited [118, 182, 191] (Table 8). Studies in human liver microsomes show that memantine inhibits CYP2B6 and CYP2D6, decreases CYP2A6 and CYP2C19, and has no effect on CYP1A2, CYP2E1, CYP2C9, or CYP3A4. Its co-administration with CYP2B6 substrates decreases memantine metabolism by 65%. NR1I2 rs1523130 is the only genetic covariate for memantine clearance in clinical studies. NR1I2 rs1523130 CT/TT carriers show a slower memantine elimination than carriers of the CC genotype [191]. Memantine transport across the blood-brain barrier (BBB) might be facilitated by proton-coupled organic cation antiporters. Memantine can be used alone or in combination with AChEIs in AD. Proteomic studies in the hippocampus and the cerebral cortex of AD-related transgenic mice (3  Tg-AD) treated with memantine revealed alterations in the expression of 233 and 342 proteins, respectively. In APP23 transgenic mice with cerebral amyloid angiopathy, memantine reduces cerebrovascular Aβ and hemosiderin deposits by enhancing Aβ-cleaving insulin-degrading enzyme (IDE) expression. Memantine increases histamine neuron activity, as reflected by a 60% increase in brain tele-methylhistamine levels and an increase in hypothalamic H3 autoreceptors, where histamine neurons are located [48].

364

Ramo´n Cacabelos et al.

6.8 Pharmacogenetics of Aducanumab

Up to now, no reliable information is available on the pharmacogenetics of aducanumab and no specific studies have been reported in this regard. Based on its potential mechanism of action, as a scavenger of Aβ, it can be inferred that different mutations in the APP, PSEN1, and PSEN2 genes, related to the intensity of the Aβ load in neuritic plaques, affect the efficacy of this monoclonal antibody. It also seems clear that the APOE genotype affects the safety and efficacy of aducanumab. APOE-4/4 carriers do not only respond poorly to aducanumab, but worsen and develop severe side effects, neuroinflammatory reactions, perivascular microedema, and white matter lesions [71, 72]. In tgAPPPS1–21 mice chronically treated with aducanumab for 4 months with weekly dosing (10 mg/kg), 17 regulated proteins, associated with (i) metabolism and mitochondria (ACAT2, ATP5J, ETFA, EXOG, HK1, NDUFA4, NDUFS7, PLCB1, PPP2R4), (ii) neuronal cytoskeleton (ADD1, CAPZB, DPYSL3, MAG), (iii) stress response (HIST1H1C/HIST1H1D, HSPA12A), and (iv) AβPP trafficking/processing (CD81, GDI2), showed significant changes. If some of these pathways and their respective proteins that are potentially implicated in AD pathogenesis are affected in humans, then it is likely that mutations in the corresponding genes may affect aducanumab efficacy and safety [192]. By analogy with other monoclonal antibodies, such as abciximab, a Fab antibody fragment of the chimeric human-murine monoclonal antibody 7E3 that inhibits platelet aggregation by binding to IIb/IIIa receptors, or trastuzumab, an immunoglobulin G1 (human-mouse monoclonal rhuMab HER2γ1-chain antihuman p185c-erbB2 receptor), which binds to the extracellular domain of human epidermal growth factor receptor 2 protein (HER-2) in breast neoplastic cells overexpressing HER-2, it cannot be excluded that aducanumab might be processed via phase I enzymes of the CYP family, phase II enzymes, and CNS transporters (ABC and SLC families) and that mutagenic variants that affect genes responsible for neuroimmune cascades (IL1B, IL6, TNF) might modify its therapeutic response and toxicity [118].

6.9 Pharmacogenetics of Multifactorial Treatments

Most studies in which AD patients are treated with multifactorial combinations reveal that APOE-3/3 carriers are the best responders and APOE-4/4 carriers are the worst responders (Figs. 10, 11, and 12). Concerning CYP-related pharmacogenetic outcomes, CYP2D6-EMs are the best responders, CYP2D6-PMs are the worst responders, and CYP2D6-IMs and UMs show an intermediate response [10–14, 17–19, 38, 48–51, 76, 90, 93, 179, 180] (Fig. 12). APOE-TOMM40 interactions affect the risk of AD and the response to drugs. TOMM40 poly T-S/S carriers are the best responders, VL/VL and S/VL carriers are intermediate responders, and L/L carriers are the worst responders to treatment (Fig. 10).

Pharmacogenomics of Alzheimer’s Disease

365

Fig. 10 APOE- and TOMM40-poly-T-related therapeutic response to a drug combination regime in patients with Alzheimer’s disease (for details, see Cacabelos et al. [17])

Fig. 11 Distribution and frequency of APOE and SIRT2 variants and APOE-SIRT2 bigenic genotypes in patients with Alzheimer’s disease (for details, see Cacabelos et al. [12])

TOMM40-L/L and S/L carriers in haplotypes with APOE-4 are the worst responders to treatment. TOMM40-S/S carriers, and to a lesser extent TOMM40-S/VL and TOMM40-VL/VL carriers, in haplotypes with APOE-3, are the best responders to treatment. The TOMM40-L/L genotype is exclusively associated with the APOE-4/4 genotype in 100% of the cases, and this haplotype (4/ 4-L/L) might be responsible for premature neurodegeneration and

366

Ramo´n Cacabelos et al.

Fig. 12 SIRT2-, APOE-SIRT2-, CYP2D6-, and SIRT2-CYP2D6 bigenic genotype-related therapeutic response to a drug combination regime in patients with Alzheimer’s disease (for details, see Cacabelos et al. [12])

consequent early onset of the disease, faster cognitive deterioration, and a limited response to conventional treatments [17]. 6.10 Pharmacoepigenetics

Epigenetic factors are important in AD pathogenesis and response to treatment [73, 193, 194]. Various epigenetic aberrations are associated with AD pathogenesis, including hypomethylation/hypermethylation in the promoters of pathogenic genes, alterations in histones, and changes in the linear and three-dimensional structure of nuclear chromatin, as well as profound alterations in microRNAs that regulate gene expression in the cytoplasm. Some of these epigenetic alterations are potential biomarkers of AD [122, 195–202]. The major problems with the use of epigenetic biomarkers in AD are their variability and lack of specificity. Changes in global DNA methylation are very sensitive and appear to be diminished in a large number of CNS diseases, such as AD ( p < 0.001), Parkinson’s disease ( p < 0.004), cerebrovascular disorders and stroke ( p < 0.01), major depression ( p < 0.05), migraine ( p < 0.03), epilepsy ( p < 0.05), and intellectual organic disability (OID) ( p < 0.001) and to a lesser extent schizophrenia (Fig. 13, upper panel). These values are very sensitive to therapeutic interventions but are unreliable as predictive or diagnostic values. The low diagnostic value of DNA methylation is compensated by the exquisite sensitivity of this

Pharmacogenomics of Alzheimer’s Disease

367

biomarker that responds in a highly sensitive manner to the therapeutic response of each patient. In addition, what appears to be important in AD is that global DNA methylation shows an APOEdependent pattern. APOE-4 carriers show a more severe DNA hypomethylation pattern than patients carrying the APOE-3 allele, which is aggravated in parallel with the degree of cognitive impairment (Fig. 13, lower panel). Epigenetic biomarkers also help in the personalization of antiAD treatments (pharmacoepigenetics) [73, 193, 194] and serve as a guide in the search for epigenetic drugs with prophylactic and/or therapeutic action in the treatment of AD [194, 196]. Sirtuin variants may alter the epigenetic machinery, contributing to AD pathogenesis. The SIRT2-C/T genotype (rs10410544) (50.92%) has been associated with AD susceptibility in the APOEε4-negative population (SIRT2-C/C, 34.72%; SIRT2-T/T 14.36%). SIRT2-APOE bigenic clusters yield 18 haplotypes that influence the PGx outcome. APOE-3/4 and APOE-4/4 genotypes accumulate in SIRT2-T/T > SIRT2-C/T > SIRT2-C/C carriers, and SIRT2-T/T and SIRT2-C/T genotypes accumulate in APOE-4/ 4 carriers. SIRT2-C/T carriers are the best responders, SIRT2-T/T carriers are intermediate responders, and SIRT2-C/C carriers are the worst responders to a multifactorial treatment (Figs. 11 and 12). PGx outcomes related to APOE-SIRT2 bigenic clusters show that 33CC carriers respond better than 33TT and 34CT carriers, whereas 24CC and 44CC carriers are the worst responders. SIRT2C/T-CYP2D6-EMs are the best responders [12] (Fig. 12). 6.11 Pharmacogenomics of Mood Disorders and Anxiety

Conventionally, the main neurotransmitter affected in AD is acetylcholine and the mechanisms that regulate cholinergic neurotransmission; however, premature neuronal death alters the levels of many other essential neurotransmitters for the normal functioning of the CNS [203–205] (Fig. 14). Deficits of noradrenaline, dopamine, serotonin, histamine, GABA, glutamate, and various neuropeptides (GRF, CRF, somatostatin, vasopressin) are particularly relevant, the alteration of which can lead to AD-related neuropsychiatric disorders. However, none of these biomarkers are sufficiently sensitive or specific to AD (Fig. 14), although their quantification in CSF or blood is useful for monitoring brain damage and/or the efficacy or ineffectiveness of the treatments the patient receives. Noradrenaline levels in the blood increase significantly in most neurodegenerative diseases, including ataxic syndromes ( p < 0.002), AD ( p < 0.001), amyotrophic lateral sclerosis (ALS) ( p < 0.007), and Parkinson’s disease (PD) ( p < 0.003), as well as in vascular encephalopathies ( p < 0.001) and in cases of organic intellectual disability (OID) (Fig. 14). In contrast, serotonin levels tend to decrease in anxiety ( p < 0.02), AD ( p < 0.05), depression ( p < 0.05), and ALS ( p < 0.05), showing high levels in ataxic syndromes ( p < 0.002), in OID ( p < 0.05), and in xenoestrogenic

368

Ramo´n Cacabelos et al.

Fig. 13 Global DNA methylation in patients with central nervous system disorders (upper panel) and APOErelated DNA methylation in patients with Alzheimer’s disease (lower panel). C Control, AD Alzheimer’s disease, PD Parkinson’s disease, CVD cerebrovascular disorder, MD major depression, SCZ schizophrenia and psychotic syndromes, MIG migraine, EPI epilepsy, OID organic intellectual disability

Pharmacogenomics of Alzheimer’s Disease

369

Fig. 14 Noradrenaline, dopamine, serotonin, and histamine levels in central nervous system disorders. C control, ANS anxiety, STR stroke, BTU brain tumors, ATX ataxia, SCZ schizophrenia and psychosis, MIG migraine, EPI epilepsy, AD Alzheimer’s disease, DEP depression, ALS amyotrophic lateral sclerosis, VEN vascular encephalopathy, MS multiple sclerosis, PD Parkinson’s disease, OID organic intellectual disability, XES xenoestrogenic syndrome, BTR: brain rauma

syndrome (XES) ( p < 0.001), a novel clinical entity present in women with chronic use of contraceptives or hormone replacement therapy (HRT) (Fig. 14). Dopamine levels are drastically reduced in PD ( p < 0.001) (Fig. 14). Histamine levels are increased in schizophrenia ( p < 0.05), epilepsy ( p < 0.05), ALS ( p < 0.05), and XES ( p < 0.05) and tend to be reduced in AD, PD, and brain trauma (Fig. 14). Neurotransmitter dysfunction may contribute to AD-related neuropsychiatric symptoms. APOE-4/4 carriers tend to show nonsignificant higher levels of catecholamines (noradrenaline, adrenaline, dopamine) than APOE-4 non-carriers. APOE-4/4 carriers are more likely to develop behavioral changes and respond worse to psychotropic treatment [18, 19, 49] (Fig. 16). In contrast, serotonin levels are significantly reduced in APOE-4 carriers and are highly increased in APOE-2 carriers (Fig. 15). The treatment of depression and anxiety in cases of dementia is always inserted into multifactorial therapeutic regimens with antidementia agents and complementary treatments for the control of concomitant pathologies. The addition of antidepressants and anxiolytics, in small doses, to these therapeutic regimens is usually well tolerated. In personalized protocols, a similar favorable response is observed in men and women. APOE-4 carriers respond

Ramo´n Cacabelos et al.

370

800

2500

600

120

100

250 80

400

200

1500 0

2/2

1000

2/3

500

2/4

3/3

3/4

4/4

200

60

40

150

20

0

2/2

100

0

2/4

3/3

3/4

4/4

3/4-AD APOE-3/4 4/4-AD APOE-4/4

2/4-AD APOE-2/4 3/3-AD APOE-3/3

2/2-AD APOE-2/2 173-AD APOE-2/3

0

0

10

20

30

40

50

60

70

80

90

100

110

120

0

10

20

30

40

Cases

90

100

110

120

p25%), obesity (>70%), diabetes mellitus type 2 (>25%), hypercholesterolemia (40%), hypertriglyceridemia (20%), metabolic syndrome (20%), hepatobiliary disorder (15%), endocrine/metabolic disorders (>20%), cardiovascular disorder (40%), cerebrovascular disorder (60–90%), neuropsychiatric disorders (60–90%), and cancer (10%). Cognitive markers indicate that females show a worse cognitive performance than males. Likewise, depression and anxiety are also more prevalent in women than in men. ECG is abnormal in 40% of the patients (38% F; 43% M). No sex-related differences are found in MRI anomalies, which are present in over 70% of the cases. The genomic screening of pathogenic genes in AD patients revealed that (i) no patient is a carrier of a single pathogenic gene, (ii) most patients (>60%) are carriers of over 10 pathogenic genes, (iii) several cerebrovascular risk variants are present in the AD genotype, and (iv) the genes that most frequently (>50%)

376

Ramo´n Cacabelos et al.

accumulate pathogenic variants in the same AD case are A2M (54.38%), ACE (78.94%), BIN1 (57.89%), CLU (63.15%), CPZ (63.15%), LHFPL6 (52.63%), MS4A4E (50.87%), MS4A6A (63.15%), PICALM (54.38%), PRNP (80.7059), and PSEN1 (77.19%). Over 80% of AD patients in the Caucasian population are deficient metabolizers for the most common drugs which are metabolized via CYP2D6, CYP2C9, CYP2C19, and CYP3A4/5 enzymes. Depending on the clinical stage of the disease, 60% to >90% of the patients require multifactorial treatments with risk of ADRs and DDIs. There is also an accumulation of 15–26 defective pharmagenes in approximately 85% of AD patients. The maximum number of defective geno-phenotypes in AD patients is 17 genes in 14.50% of cases, 18 genes in 16.91% of cases, and 19 genes in 14.40% of AD cases. The most dysfunctional genes frequently found (F >20%) in AD are the following: CYP1A1 (31.47%), CYP1B1 (82.23%), MAOB (47.72%), CES1 (96.45%), CHAT (36.04%), COMT (30.96%), GSTM1 (54.82%), GSTT1 (25.38%), NAT2 (43.65%), SOD2 (67.51%), ABCB1 (44.16%), ABCG2 (90.36%), FABP2 (90.86%), SLCA2 (59.90%), SLC22A1 (34.52%), SLC30A8 (48.22%), ADRB2 (41.62%), AGT (42.13%), APOE (30.96%), CHRNA7 (42.13%), DRD2 (45.18%), GABRA1 (55.33%), HMGCR (73.60%), HTR1A (65.48%), HTR2C (79.19%), OPRM1 (69.54%), PPARG (81.73%), PRKCE (41.62%), and VKORC1 (65.99%). The implementation of pharmacogenetics can help optimize drug development and the limited therapeutic resources available to treat AD and personalize the use of anti-dementia drugs in combination with other medications for the treatment of concomitant disorders. Major limitations for the routine use of pharmacogenetics are the lack of education and training in physicians and pharmacists, poor characterization of drug-related pharmacogenetics, non-specific biomarkers of drug efficacy and toxicity, costeffectiveness, administrative problems in health organizations, and insufficient regulation for the generalized use of pharmacogenetics in the clinical setting. The implementation of pharmacogenetics requires (i) education of physicians and all other parties involved in the use and benefits of pharmacogenetics; (ii) prospective studies to demonstrate the benefits of pharmacogenetic genotyping; (iii) standardization of pharmacogenetics procedures and development of clinical guidelines; (iv) biochips and microarrays to cover genes with high pharmacogenetic potential; and (v) new regulations for pharmacogenetics-related drug development and drug labeling.

Pharmacogenomics of Alzheimer’s Disease

377

Acknowledgments The authors would like to thank their collaborators at the International Center of Neuroscience and Genomic Medicine for technical assistance. Declaration of Interest RC is President and stockholder of EuroEspes (Biomedical Research Center), EuroEspes Biotechnology, IABRA, and EuroEspes Publishing Co. NC is a shareholder of EuroEspes, S.A. The authors have no other relevant affiliations or financial involvement with any other organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed apart from those disclosed.

Funding This article was funded by EuroEspes, S.A., and IABRA (International Agency for Brain Research and Aging). References 1. Cantarero-Prieto D, Leon PL, BlazquezFernandez C, Sanchez-Juan P, Sarabia-Cobo C (2020) The economic cost of dementia: a systematic review. Dementia 19:2637–2657 2. Sado M, Ninomiya A, Shikimoto R, Ikeda B, Baba T, Yoshimura K, Mimura M (2018) The estimated cost of dementia in Japan, the most aged society in the world. PLoS One 13: e0206508 3. Bachman DL, Wolf PA, Linn R, Knoefel JE, Belanger A, D’Agostino RB, White LR (1992) Prevalence of dementia and probable senile dementia of the Alzheimer type in the Framingham study. Neurology 42:115–119 L, 4. Cacabelos R, Ferna´ndez-Novoa Lombardi V, Kubota Y, Takeda M (2005) Molecular genetics of Alzheimer’s disease and aging. Methods Find Exp Clin Pharmacol 27:1–573 5. Masters CL, Bateman R, Blennow K, Rowe CC, Sperling RA, Cummings JL (2015) Alzheimer’s disease. Nat Rev Dis Primers 1: 15056 6. Arvanitakis Z, Shah RC, Bennett DA (2019) Diagnosis and management of dementia: review. JAMA 322:1589–1599 7. Hersi M, Irvine B, Gupta P, Gomes J, Birkett N, Krewski D (2017) Risk factors associated with the onset and progression of

Alzheimer’s disease: a systematic review of the evidence. Neurotoxicology 61:143–187 8. Dubois B, Hampel H, Feldman HH, Scheltens P, Aisen P, Andrieu S, Bakardjian H et al (2016) Preclinical Alzheimer’s disease: definition, natural history, and diagnostic criteria. Alzheimers Dement 12(3):292–323 9. Busche MA, Hyman BT (2020) Synergy between amyloid-β and tau in Alzheimer’s disease. Nat Neurosci 23(10):1183–1193 10. Cacabelos R, Cacabelos P, Torrellas C, Tellado I, Carril JC (2014) Pharmacogenomics of Alzheimer’s disease: novel therapeutic strategies for drug development. Methods Mol Biol 1175:323–556 11. Cacabelos R, Carril JC, Cacabelos P, Teijido O, Goldgaber D (2016) Pharmacogenomics of Alzheimer’s disease: genetic determinants of phenotypic variation and therapeutic outcome. J Genom Med Pharmacogenom 1:151–209 12. Cacabelos R, Carril JC, Cacabelos N, Kazantsev AG, Vostrov AV, Corzo L et al (2019) Sirtuins in Alzheimer’s disease: SIRT2-related GenoPhenotypes and implications for PharmacoEpiGenetics. Int J Mol Sci 20:1249 13. Cacabelos R (2018) Population-level pharmacogenomics for precision drug development

378

Ramo´n Cacabelos et al.

in dementia. Expert Rev Precision Med Drug Dev 3:163–188 14. Cacabelos R (2020) Pharmacogenomics of cognitive dysfunction and neuropsychiatric disorders in dementia. Int J Mol Sci 21:3059 15. Matej R, Tesar A, Rusina R (2019) Alzheimer’s disease and other neurodegenerative dementias in comorbidity: a clinical and neuropathological overview. Clin Biochem 73: 26–31 16. Cacabelos R, Cacabelos N, Carril JC (2019) The role of pharmacogenomics in adverse drug reactions. Expert Rev Clin Pharmacol 12:407–442 17. Cacabelos R, Goldgaber D, Vostrov A, Matsuki H, Torrellas C, Corzo D et al (2014) APOE-TOMM40 in the pharmacogenomics of dementia. J Pharmacogenom Pharmacoproteom 5:135 18. Cacabelos R (2020) Pharmacogenomic of drugs to treat brain disorders. Expert Rev Precision Med Drug Develop 5:181–234 19. Cacabelos R, Carril JC, Corzo L, FernandezNovoa L, Pego R, Cacabelos N et al (2021) Influence of pathogenic and metabolic genes on the pharmacogenetics of mood disorders in Alzheimer’s disease. Pharmaceuticals 14: 366 20. Bature F, Guinn BA, Pang D, Pappas Y (2017) Signs and symptoms preceding the diagnosis of Alzheimer’s disease: a systematic scoping review of literature from 1937 to 2016. BMJ Open 7:015746 21. Linnemann C, Lang UE (2020) Pathways connecting late-life depression and dementia. Front Pharmacol 11:279 22. Bennett S, Thomas AJ (2014) Depression and dementia: cause, consequence or coincidence? Maturitas 79:184–190 23. Gutzmann H, Qazi A (2015) Depression associated with dementia. Z Gerontol Geriatr 48(4):305–311 24. Baruch N, Burgess J, Pillai M, Allan CL (2019) Treatment for depression comorbid with dementia. Evid Based Ment Health 22: 167–171 25. Bingham KS, Flint AJ, Mulsant BH (2019) Management of late-life depression in the context of cognitive impairment: a review of the recent literature. Curr Psychiatry Rep 21: 74 26. Kratz T (2017) The diagnosis and treatment of behavioral disorders in dementia. Dtsch Arztebl Int 114:447–454 27. Aarsland D (2020) Epidemiology and pathophysiology of dementia-related psychosis. J Clin Psychiatry 81:AD19038BR1C

28. Matsuda H (2016) MRI morphometry in Alzheimer’s disease. Ageing Res Rev 30: 17–24 29. Chen C, Homma A, Mok VC, Krishnamoorthy E, Alladi S, Meguro K et al (2016) Alzheimer’s disease with cerebrovascular disease: current status in the Asia-Pacific region. J Intern Med 280:359–374 30. Chandra A, Dervenoulas G, Politis M, Alzheimer’s Disease Neuroimaging Initiative (2019) Magnetic resonance imaging in Alzheimer’s disease and mild cognitive impairment. J Neurol 266:1293–1302 31. Whitwell JL (2018) Alzheimer’s disease neuroimaging. Curr Opin Neurol 31:396–404 32. Barbe C, Jolly D, Morrone I, Wolak-ThierryA, Drame´ M, Novella JL et al (2018) Factors associated with quality of life in patients with Alzheimer’s disease. BMC Geriatr 18:159 33. Ettcheto M, Olloquequi J, Sa´nchez-Lo´pez E, Busquets O, Cano A, Manzine PR et al (2020) Benzodiazepines and related drugs as a risk factor in Alzheimer’s disease dementia. Front Aging Neurosci 11:344 34. Vidoni ED, Kamat A, Gahan WP, Ourso V, Woodard K, Kerwin DR et al (2020) Baseline prevalence of polypharmacy in older hypertensive study subjects with elevated dementia risk: findings from the Risk Reduction for Alzheimer’s Disease Study (rrAD). J Alzheimers Dis 77:175–182 35. Andin U, Passant U, Gustafson L, Englund E (2007) Alzheimer’s disease (AD) with and without white matter pathology-clinical identification of concurrent cardiovascular disorders. Arch Gerontol Geriatr 44:277–286 36. Ibrahim B, Suppiah S, Piersson AD, Razali RM, Mohamad M, Abu Hassan H et al (2021) Cardiovascular risk factors of Alzheimer’s disease and other neurocognitive disorders in Malaysia. Med J Malaysia 76(3): 291–297 37. Cacabelos R (2004) Genomic characterization of Alzheimer’s disease and genotyperelated phenotypic analysis of biological markers in dementia. Pharmacogenomics 5: 1049–1105 38. Cacabelos R, Meyyazhagan A, Carril JC, Cacabelos P, Teijido O (2018) Pharmacogenetics of vascular risk factors in Alzheimer’s disease. J Personalized Med 8:3 39. Cacabelos R, Ferna´ndez-Novoa L, Corzo L, Pichel V, Lombardi V, Kubota Y (2004) Genomics and phenotypic profiles in dementia: implications for pharmacological treatment. Methods Find Exp Clin Pharmacol 26:421–444

Pharmacogenomics of Alzheimer’s Disease 40. Contois JH, Anamani DE, Tsongalis GJ (1996) The underlying molecular mechanism of apolipoprotein E polymorphism: relationships to lipid disorders, cardiovascular disease, and Alzheimer’s disease. Clin Lab Med 16: 105–123 41. Cacabelos R, Ferna´ndez-Novoa L, Lombardi V, Corzo L, Pichel V, Kubota Y (2003) Cerebrovascular risk factors in Alzheimer’s disease: brain hemodynamics and pharmacogenomic implications. Neurol Res 25: 567–580 42. Loera-Valencia R, Goikolea J, ParradoFernandez C, Merino-Serrais P, Maioli S (2019) Alterations in cholesterol metabolism as a risk factor for developing Alzheimer’s disease: potential novel targets for treatment. J Steroid Biochem Mol Biol 190:104–114 43. Jeong W, Lee H, Cho S, Seo J (2019) ApoE4induced cholesterol dysregulation and its brain cell type-specific implications in the pathogenesis of Alzheimer’s disease. Mol Cells 42:739–746 44. Di Paolo G, Kim TW (2011) Linking lipids to Alzheimer’s disease: cholesterol and beyond. Nat Rev Neurosci 12:284–296 45. Pugazhenthi S, Qin L, Reddy PH (2017) Common neurodegenerative pathways in obesity, diabetes, and Alzheimer’s disease. Biochim Biophys Acta Mol basis Dis 1863: 1037–1045 46. Chornenkyy Y, Wang WX, Wei A, Nelson PT (2019) Alzheimer’s disease and type 2 diabetes mellitus are distinct diseases with potential overlapping metabolic dysfunction upstream of observed cognitive decline. Brain Pathol 29:3–17 47. Baglietto-Vargas D, Shi J, Yaeger DM, Ager R, LaFerla FM (2016) Diabetes and Alzheimer’s disease crosstalk. Neurosci Biobehav Rev 64:272–287 48. Cacabelos R (2020) Pharmacogenetic considerations when prescribing cholinesterase inhibitors for the treatment of Alzheimer’s disease. Expert Opin Drug Metab Toxicol 16:673–701 49. Cacabelos R (2007) Molecular pathology and pharmacogenomics in Alzheimer’s disease: polygenic-related effects of multifactorial treatments on cognition, anxiety, and depression. Methods Find Exp Clin Pharmacol 29: 1–91 50. Cacabelos R (2008) Pharmacogenomics in Alzheimer’s disease. Methods Mol Biol 448: 213–357 51. Cacabelos R, Ferna´ndez-Novoa L, Martı´nezBouza R, McKay A, Carril JC, Lombardi V et al (2010) Future trends in the

379

pharmacogenomics of brain disorders and dementia: influence of APOE and CYP2D6 variants. Pharmaceuticals 3:3040–3100 52. Cacabelos R (2018) Have there been improvement in Alzheimer’s disease drug discovery over the past 5 years? Expert Opin Drug Discovery 13:523–538 53. Maramai S, Benchekroun M, Gabr MT, Yahiaoui (2020) Multitarget therapeutic strategies for Alzheimer’s disease: review on emerging target combinations. Biomed Res Int 2020:5120230 54. Cacabelos R (2020) How plausible is an Alzheimer’s disease vaccine? Expert Opin Drug Discovery 15:1–6 55. Cummings J, Lee G, Ritter A, Sabbagh M, Zhong K (2020) Alzheimer’s disease drug development pipeline: 2020. Alzheimers Dement 6:e12050 56. Schenk D, Barbour R, Dunn W, Gordon G, Grajeda H, Guido T et al (1999) Immunization with amyloid-beta attenuates Alzheimerdisease-like pathology in the PDAPP mouse. Nature 400(6740):173–177 57. Foroutan N, Hopkins RB, Tarride JE, Florez ID, Levine M (2019) Safety and efficacy of active and passive immunotherapy in mild-tomoderate Alzheimer’s disease: a systematic review and network meta-analysis. Clin Invest Med 42:53–65 58. Janus C, Pearson J, McLaurin J, Mathews PM, Jiang Y, Schmidt SD et al (2000) A beta peptide immunization reduces behavioural impairment and plaques in a model of Alzheimer’s disease. Nature 408:979–982 59. Herline K, Drummond E, Wisniewski T (2018) Recent advancements toward therapeutic vaccines against Alzheimer’s disease. Expert Rev Vaccines 17:707–721 60. Novak P, Zilka N, Zilkova M, Kovacech B, Skrabana R, Ondrus M et al (2019) AADvac1, an active immunotherapy for Alzheimer’s disease and non Alzheimer tauopathies: an overview of preclinical and clinical development. J Prev Alzheimers Dis 6:63–69 61. Hoskin JL, Sabbagh MN, Al-Hasan Y, Decourt B (2019) Tau immunotherapies for Alzheimer’s disease. Expert Opin Investig Drugs 28:545–554 62. Carrera I, Fernandez-Novoa L, Aliev G, Vigo C, Cacabelos R (2016) Validating immunotherapy in Alzheimer’s disease: the EB101 vaccine. Curr Pharm Des 22:849–858 63. Zhang HY, Zhu K, Meng Y, Ding L, Wang JC, Yin WC et al (2018) Reduction of amyloid beta by Aβ3-10-KLH vaccine also decreases tau pathology in 3Tg-AD mice. Brain Res Bull 142:233–240

380

Ramo´n Cacabelos et al.

64. Rosenberg RN, Fu M, Lambracht-Washington D (2018) Active full-length DNA Aβ42 immunization in 3xTg-AD mice reduces not only amyloid deposition but also tau pathology. Alzheimers Res Ther 10:115 65. Oddo S, Billings L, Kesslak JP, Cribbs DH, LaFerla FM (2004) Abeta immunotherapy leads to clearance of early, but not late, hyperphosphorylated tau aggregates via the proteasome. Neuron 43:321–332 66. Lalli G, Schott JM, Hardy J, De Strooper B (2021) Aducanumab: a new phase in therapeutic development for Alzheimer’s disease? EMBO Mol Med 13:e14781 67. Mukhopadhyay S, Banerjee D (2021) A primer on the evolution of aducanumab: the first antibody approved for treatment of Alzheimer’s disease. J Alzheimers 83:1537–1552 68. Knopman DS, Jones DT, Greicius MD (2021) Failure to demonstrate efficacy of aducanumab: An analysis of the EMERGE and ENGAGE trials as reported by Biogen, December 2019. Alzheimers Dement 17: 696–701 69. Sevigny J, Chiao P, Bussie`re T, Weinreb PH, Williams L, Maier M et al (2016) The antibody aducanumab reduces Aβ plaques in Alzheimer’s disease. Nature 537:50–56 70. Ferrero J, Williams L, Stella H, Leitermann K, Mikulskis A, O’Gorman J et al (2016) Firstin-human, double-blind, placebo-controlled, single-dose escalation study of aducanumab (BIIB037) in mild-to-moderate Alzheimer’s disease. Alzheimers Dement 2:169–176 71. Tolar M, Abushakra S, Hey JA, Porsteinsson A, Sabbagh M (2020) Aducanumab, gantenerumab, BAN2401, and ALZ-801-the first wave of amyloid-targeting drugs for Alzheimer’s disease with potential for near term approval. Alzheimers Res Ther 12:95 72. VandeVrede L, Gibbs DM, Koestler M, La Joie R, Ljubenkov PA, Provost K et al (2020) Symptomatic amyloid-related imaging abnormalities in an APOE ε4/ε4 patient treated with aducanumab. Alzheimers Dement 12:e12101 73. Cacabelos R, Carril JC, Sanmartı´n A, Cacabelos P. Pharmacoepigenetic processors: epigenetic drugs, drug resistance, toxicoepigenetics, and nutriepigenetics. In: Cacabelos R (ed) (2019) Pharmacoepigenetics. Academic Press/Elsevier, Oxford, UK, pp. 191–424 74. Kozyra M, Ingelman-Sundberg M, Lauschke VM (2017) Rare genetic variants in cellular

transporters, metabolic enzymes, and nuclear receptors can be important determinants of interindividual differences in drug response. Genet Med 19:20–29 75. Zhou ZW, Chen XW, Sneed KB, Yang YX, Zhang X, He ZH et al (2015) Clinical association between pharmacogenomics and adverse drug reactions. Drugs 75:589–631 76. Cacabelos R (2018) Pleiotropy and promiscuity in pharmacogenomics for the treatment of Alzheimer’s disease and related risk factors. Future Neurol 13:71–86 77. Cacabelos R, Teijido O (2018) Epigenetic drug discovery for Alzheimer’s disease. In: Moskalev A, Vaiserman A (eds) Epigenetics of aging and longevity. Elsevier/Academic Press, London, pp 453–495 78. Cacabelos R, Martı´nez-Bouza R, Carril JC, Fernandez-Novoa L, Lombardi V, Carrera I et al (2012) Genomics and pharmacogenomics of brain disorders. Curr Pharm Biotechnol 13:674–725 79. Cacabelos R, Martı´nez-Bouza R (2011) Genomics and pharmacogenomics of dementia. CNS Neurosci Ther 17:566–576 80. Cai Y, An SS, Kim S (2015) Mutations in presenilin 2 and its implications in Alzheimer’s disease and other dementia-associated disorders. Clin Interv Aging 10:1163–1172 81. Kelleher RJ, Shen J (2017) Presenilin-1 mutations and Alzheimer’s disease. Proc Natl Acad Sci U S A 114:629–631 82. Tcw J, Goate AM (2017) Genetics of β-amyloid precursor protein in Alzheimer’s disease. Cold Spring Harb Perspect Med 7: a024539 83. Selkoe DJ, Hardy J (2016) The amyloid hypothesis of Alzheimer’s disease at 25 years. EMBO Mol Med 8:595–608 84. Tolar M, Abushakra S, Sabbagh M (2020) The path forward in Alzheimer’s disease therapeutics: reevaluating the amyloid cascade hypothesis. Alzheimers Dement 16: 1553–1560 85. Giau VV, Bagyinszky E, Yang Y, Youn YC, An SSA, Kim SY (2019) Genetic analyses of earlyonset Alzheimer’s disease using next generation sequencing. Sci Rep 9:8368 86. Van Cauwenberghe C, Van Broeckhoven C, Sleegers K (2016) The genetic landscape of Alzheimer disease: clinical implications and perspectives. Genet Med 18:421–430 87. Bertram L (2016) Next generation sequencing in Alzheimer’s disease. Methods Mol Biol 1303:281–297

Pharmacogenomics of Alzheimer’s Disease 88. Zhu JB, Tan CC, Tan L, Yu JT (2017) State of play in Alzheimer’s disease genetics. J Alzheimers Dis 58:631–659 89. Karch CM, Goate AM (2015) Alzheimer’s disease risk genes and mechanisms of disease pathogenesis. Biol Psychiatry 77:43–51 90. Cacabelos R, Torrellas C, Teijido O, Carril JC (2016) Pharmacogenetic considerations in the treatment of Alzheimer’s disease. Pharmacogenomics 17:1041–1074 91. Cacabelos R (2012) Pharmacogenomics of central nervous system (CNS) drugs. Drug Dev Res 73:461–476 92. Cacabelos R, Takeda M (2006) Pharmacogenomics, nutrigenomics and future therapeutics in Alzheimer’s disease. Drugs Future 3: 5–146 93. Cacabelos R (2001) Psychogeriatric research: a conceptual introduction to aging and geriatric neuroscience. Psychogeriatrics 1: 158–188 94. Sabbagh MN, Malek-Ahmadi M, Dugger BN, Lee K, Sue LI, Serrano G et al (2013) The influence of Apolipoprotein E genotype on regional pathology in Alzheimer’s disease. BMC Neurol 13:44 95. Kennedy RE, Cutter GR, Schneider LS (2014) Effect of APOE genotype status on targeted clinical trials outcomes and efficiency in dementia and mild cognitive impairment resulting from Alzheimer’s disease. Alzheimers Dement 10:349–359 96. Reiman EM, Arboleda-Velasquez JF, Quiroz YT, Huentelman MJ, Beach TG, Caselli RJ et al (2020) Exceptionally low likelihood of Alzheimer’s dementia in APOE2 homozygotes from a 5,000-person neuropathological study. Nat Commun 11:667 97. Shapira M, Tur-Kaspa I, Bosgraaf L, Livni N, Grant AD, Grisaru D et al (2000) A transcription-activating polymorphism in the ACHE promoter associated with acute sensitivity to anti-acetylcholinesterases. Hum Mol Genet 9:1273–1281 98. Lane R, Feldman HH, Meyer J, He Y, Ferris SH, Nordberg A et al (2008) Synergistic effect of apolipoprotein E epsilon4 and butyrylcholinesterase K-variant on progression from mild cognitive impairment to Alzheimer’s disease. Pharmacogenet Genomics 18: 289–928 99. Cuddy LK, Seah C, Pasternak SH, Rylett RJ (2017) Amino-terminal β-amyloid antibody blocks β-amyloid-mediated inhibition of the high-affinity choline transporter CHT. Front Mol Neurosci 10:361

381

100. Payette DJ, Xie J, Guo Q (2007) Reduction in CHT1-mediated choline uptake in primary neurons from presenilin-1 M146V mutant knock-in mice. Brain Res 1135:12–21 101. Wang Y, Zhou Z, Tan H, Zhu S, Wang Y, Sun Y et al (2017) Nitrosylation of vesicular transporters in brain of amyloid precursor protein/ presenilin 1 double transgenic mice. J Alzheimers Dis 55:1683–1692 102. Nagy PM, Aubert I (2012) Overexpression of the vesicular acetylcholine transporter increased acetylcholine release in the hippocampus. Neuroscience 218:1–11 103. Kolisnyk B, Al-Onaizi MA, Xu J, Parfitt GM, Ostapchenko VG, Hanin G et al (2016) Cholinergic regulation of hnRNPA2/B1 translation by M1 muscarinic receptors. J Neurosci 36:6287–6296 104. Dolejsˇ´ı E, Liraz O, Rudajev V, Zimcik P, Dolezal V, Michaelson DM (2016) Apolipoprotein E4 reduces evoked hippocampal acetylcholine release in adult mice. J Neurochem 136:503–509 105. Albin RL, Bohnen NI, Muller MLTM, Dauer W, Sarter M, Frey KA et al (2018) Regional vesicular acetylcholine transporter distribution in human brain: a [18 F]fluoroethoxybenzovesamicol positron emission tomography study. J Comp Neurol 526: 2884–2897 106. Wallace TL, Bertrand D (2013) Importance of the nicotinic acetylcholine receptor system in the prefrontal cortex. Biochem Pharmacol 85:1713–1720 107. Ma KG, Qian YH (2019) Alpha 7 nicotinic acetylcholine receptor and its effects on Alzheimer’s disease. Neuropeptides 73:96–106 108. Sadigh-Eteghad S, Talebi M, Mahmoudi J, Babri S, Shanehbandi D (2015) Selective activation of α7 nicotinic acetylcholine receptor by PHA-543613 improves Aβ25-35mediated cognitive deficits in mice. Neuroscience 298:81–93 109. Cacabelos R, Teijido O, Carril JC (2019) Pharmacoepigenetic processors: epigenetic drugs, drug resistance, toxicoepigenetics and nutriepigenenetics. In: Cacabelos R (ed) Pharmacoepigenetics. Elsevier/Academic Press, Oxford 110. Tang X, Chen S (2015) Epigenetic regulation of cytochrome P450 enzymes and clinical implication. Curr Drug Metab 16:86–96 111. Habano W, Kawamura K, Iizuka N, Terashima J, Sugai T, Ozawa S (2015) Analysis of DNA methylation landscape reveals the roles of DNA methylation in the regulation of

382

Ramo´n Cacabelos et al.

drug metabolizing enzymes. Clin Epigenetics 7:105 112. Cacabelos R, Torrellas C, Carrera I (2015) Opportunities in pharmacogenomics for the treatment of Alzheimer’s disease. Future Neurol 10:229–252 113. Chu SK, Yang HC (2017) Interethnic DNA methylation difference and its implications in pharmacoepigenetics. Epigenomics 9: 1437–1454 114. Lee IS, Kim D (2011) Polymorphic metabolism by functional alterations of human cytochrome P450 enzymes. Arch Pharm Res 34: 1799–1816 115. Xie HG, Kim RB, Wood AJ, Stein CM (2001) Molecular basis of ethnic differences in drug disposition and response. Annu Rev Pharmacol Toxicol 41:815–850 116. Isaza C, Henao J, Ramı´rez E (2005) Polymorphic variants of the beta-2-adrenergic receptor (ADRB2) gene and ADRB2-related propanolol-induced dyslipidemia in the Colombian population. Methods Find Exp Clin Pharmacol 27:237–244 117. Taskin B, Percin FE, Ergun MA (2016) Investigation of CYP2D6 gene polymorphisms in Turkish population. Psychopharmacol Bull 46:67–72 118. Cacabelos R (ed) (2012) World guide for drug use and pharmacogenomics. EuroEspes Publishing, A Corunna 119. Apella´niz-Ruiz M, Inglada-Pe´rez L, Naranjo ME, Sa´nchez L, Mancikova V, Curra´s-Freire et al (2015) High frequency and founder effect of the CYP3A4*20 loss-of-function allele in the Spanish population classifies CYP3A4 as a polymorphic enzyme. Pharmacogenomics J 15:288–292 120. Backman JT, Filppula AM, Niemi M, Neuvonen PJ (2016) Role of cytochrome P450 2C8 in drug metabolism and interactions. Pharmacol Rev 68:168–241 121. Zhou Y, Lauschke VM (2018) Comprehensive overview of the pharmacogenetic diversity in Ashkenazi Jews. J Med Genet 55: 617–627 122. Tucci JD, Pumuye PP, Helsby NA, Barrat DT, Pokeya PP, Hombhanje D et al (2018) Pharmacogenomics in Papua New Guineans: unique profiles and implications for enhancing drug efficacy while improving drug safety. Pharmacogenet Genomics 28:153–164 123. Rajman I, Knapp L, Morgan T, Masimirembwa C (2017) African genetic diversity: implications for cytochrome P450-mediated drug metabolism and drug development. EBioMedicine 17:67–74

124. Gaio V, Picanc¸o I, Nunes B, Fernandes A, Mendonca F, Horta-Correia F et al (2015) Pharmacogenetic profile of a South Portuguese population: results from the pilot study of the European Health Examination Survey in Portugal. Public Health Genomics 18:139–150 125. Marquez B, Van Bambeke F (2011) ABC multidrug transporters: target for modulation of drug pharmacokinetics and drug-drug interactions. Curr Drug Targets 12:600–620 ˜ oz F (2014) The 126. Cacabelos R, Lo´pez-Mun ABCB1 transporter in Alzheimer’s disease. Clin Exp Pharmacol 4:e128 127. Elali A, Rivest S (2013) The role of ABCB1 and ABCA1 in beta-amyloid clearance at the neurovascular unit in Alzheimer’s disease. Front Physiol 4:45 128. Cascorbi I, Flu¨h C, Remmler C, Haenisch S, Faltraco F, Grumbt M et al (2013) Association of ATP-binding cassette transporter variants with the risk of Alzheimer’s disease. Pharmacogenomics 14:485–494 ´ , Juha´sz A, La´szlo´ A, Pa´ka´ski M, 129. Fehe´r A Ka´lma´n J, Janka Z (2013) Association between the ABCG2 C421A polymorphism and Alzheimer’s disease. Neurosci Lett 550: 51–54 130. Chai AB, Leung GKF, Callaghan R, Gelissen IC (2020) P-glycoprotein: a role in the export of amyloid-β in Alzheimer’s disease? FEBS J 287:612–625 131. Alharbi HA, Alcorn J, Al-Mousa A, Giesy JP (2017) Toxicokinetics and toxicodynamics of chlorpyrifos is altered in embryos of Japanese medaka exposed to oil sands process-affected water: evidence for inhibition of P-glycoprotein. J Appl Toxicol 37:591–601 132. Bruckmann S, Brenn A, Grube M, Niedrig K, Holtfreter S, von Bohlen, Halbach O et al (2017) Lack of P-glycoprotein results in impairment of removal of beta-amyloid and increased intraparenchymal cerebral amyloid angiopathy after active immunization in a transgenic mouse model of Alzheimer’s disease. Curr Alzheimer Res 14:656–667 133. Szablewski L (2017) Glucose transporters in brain: in health and in Alzheimer’s disease. J Alzheimers Dis 55:1307–1320 134. Alam MA, Datta PK (2019) Epigenetic regulation of excitatory amino acid transporter 2 in neurological disorders. Front Pharmacol 10:1510 135. Sharma A, Kazim SF, Larson CS, Ramakrishnan A, Gray JD, McEwen BS et al (2019) Divergent roles of astrocytic versus neuronal EAAT2 deficiency on cognition

Pharmacogenomics of Alzheimer’s Disease and overlap with aging and Alzheimer’s molecular signatures. Proc Natl Acad Sci U S A 116:21800–21811 136. Pang R, Wang X, Du Z, Pei F, Li Z, Sun L et al (2020) The distribution and density of monocarboxylate transporter 2 in cerebral cortex, hippocampus and cerebellum of wild-type mice. J Anat 236:370–377 137. Ugbode C, Hu Y, Whalley B, Peers C, Rattray M, Dallas ML (2017) Astrocytic transporters in Alzheimer’s disease. Biochem J 474:333–355 138. Mandal AK, Mount DB (2019) Interaction between ITM2B and GLUT9 links urate transport to neurodegenerative disorders. Front Physiol 10:1323 139. Britzolaki A, Saurine J, Klocke B, Pitychoutis PM (2020) A role for SERCA pumps in the neurobiology of neuropsychiatric and neurodegenerative disorders. Adv Exp Med Biol 1131:131–161 140. Chen JJ, Nathaniel DL, Raghavan P, Nelson M, Tian R, Tse E et al (2019) Compromised function of the ESCRT pathway promotes endolysosomal escape of tau seeds and propagation of tau aggregation. J Biol Chem 294:18952–18966 141. Kunkle BW, Vardarajan BN, Naj AC, Whitehead PL, Rolati S, Slifer S et al (2017) Earlyonset Alzheimer disease and candidate risk genes involved in endolysosomal transport. JAMA Neurol 74:1113–1122 142. Jadiya P, Kolmetzky DW, Tomar D, Di Meco A, Lombardi AA, Lambert JP et al (2019) Impaired mitochondrial calcium efflux contributes to disease progression in models of Alzheimer’s disease. Nat Commun 10:3885 143. Moriguchi S, Kita S, Fukaya M, Osanai M, Inagaki R, Sasaki Y et al (2018) Reduced expression of Na+/Ca2+ exchangers is associated with cognitive deficits seen in Alzheimer’s disease model mice. Neuropharmacology 131:291–303 144. Maezawa I, Nguyen HM, Di Lucente J, Jenkins DP, Singh V, Hilt S et al (2018) Kv1.3 inhibition as a potential microglia-targeted therapy for Alzheimer’s disease: preclinical proof of concept. Brain 141:596–612 145. Shoshan-Barmatz V, Pittala S, Mizrachi D (2019) VDAC1 and the TSPO: expression, interactions, and associated functions in health and disease states. Int J Mol Sci 20: E3348 146. Shoshan-Barmatz V, Nahon-Crystal E, Shteinfer-Kuzmine A, Gupta R (2018)

383

VDAC1, mitochondrial dysfunction, and Alzheimer’s disease. Pharmacol Res 131:87–101 147. Heinemeyer T, Stemmet M, Bardien S, Neethling A (2019) Underappreciated roles of the translocase of the outer and inner mitochondrial membrane protein complexes in human disease. DNA Cell Biol 38:23–40 148. Alrosan A, Aleidi SM, Yang A, Brown AJ, Gelissen IC (2019) The adaptor protein Alix is involved in the interaction between the ubiquitin ligase NEDD4-1 and its targets, ABCG1 and ABCG4. Int J Mol Sci 20:E2714 149. Adnan M, Islam W, Zhang J, Zheng W, Lu GD (2019) Diverse role of SNARE protein Sec22 in vesicle trafficking, membrane fusion, and autophagy. Cell 8:E337 150. D’Onofrio G, Panza F, Sancarlo D, Lauriola M, Dagostino MP, Paroni G et al (2019) Hydroxytryptamine transporter gene-linked polymorphic region (5HTTLPR) is associated with delusions in Alzheimer’s disease. Transl Neurodegen 8:4 ´ , Giricz Z, Juha´sz A, Pakaski M, 151. Fehe´r A Janka Z, Kalman J (2018) ABCA1 rs2230805 and rs2230806 common gene variants are associated with Alzheimer’s disease. Neurosci Lett 664:79–83 152. Hu W, Lin X, Zhang H, Zhao N (2017) ATP binding cassette subfamily a member 2 (ABCA2) expression and methylation are associated with Alzheimer’s disease. Med Sci Monit 23:5851–5861 153. Davis W, Tew KD (2018) ATP-binding cassette transporter-2 (ABCA2) as a therapeutic target. Biochem Pharmacol 151:188–200 154. Bhatia S, Fu Y, Hsiao JT, Halliday GM, Kim WS (2017) Deletion of Alzheimer’s disease risk gene ABCA7 alters White adipose tissue development and leptin levels. J Alzheimers Dis Rep 1:237–247 155. Lamartinie`re Y, Boucau MC, Dehouck L, Krohn M, Pahnke J, Candela P et al (2018) ABCA7 downregulation modifies cellular cholesterol homeostasis and decreases amyloid-β peptide efflux in an in vitro model of the blood-brain barrier. J Alzheimers Dis 64: 1195–1211 156. De Roeck A, Duchateau L, Van Dongen J, Cacace R, Bjerke M, van den Bossche T (2018) An intronic VNTR affects splicing of ABCA7 and increases risk of Alzheimer’s disease. Acta Neuropathol 135:827–837 157. Yamazaki K, Yoshino Y, Mori T, Yoshida T, Ozaki Y, Sao T et al (2017) Gene expression and methylation analysis of ABCA7 in patients with Alzheimer’s disease. J Alzheimers Dis 57:171–181

384

Ramo´n Cacabelos et al.

158. Liao YC, Lee WJ, Hwang JP, Wang YF, Tsai CF, Wang PN et al (2014) ABCA7 gene and the risk of Alzheimer’s disease in Han Chinese in Taiwan. Neurobiol Aging 35:2423. e7–2423.e13 159. Zhong X, Liu MY, Sun XH, Wei MJ (2016) Association between ABCB1 polymorphisms and haplotypes and Alzheimer’s disease: a meta-analysis. Sci Rep 6:32708 160. Nelson PT, Estus S, Abner EL, Parikh I, Malik M, Neltner JH et al (2014) ABCC9 gene polymorphism is associated with hippocampal sclerosis of aging pathology. Acta Neuropathol 127:825–843 161. Sano O, Tsujita M, Shimizu Y, Kato R, Kobayashi A, Kioka N et al (2016) ABCG1 and ABCG4 suppress γ-secretase activity and amyloid β production. PLoS One 11: e0155400 162. Liu HP, Lin WY, Wang WF, Tsai CH, Wu WC, Chiou MT et al (2013) Genetic variability in copper-transporting P-type adenosine triphosphatase (ATP7B) is associated with Alzheimer’s disease in a Chinese population. J Biol Regul Homeost Agents 27:319–327 163. Montesanto A, Crocco P, Dato S, Geracitano S, Frangipane F, Colao R et al (2018) Uncoupling protein 4 (UCP4) gene variability in neurodegenerative disorders: further evidence of association in Frontotemporal dementia. Aging (Albany NY) 10: 3283–3293 164. Amir Shaghaghi M, Murphy B, Eck P (2016) The SLC2A14 gene: genomic locus, tissue expression, splice variants, and subcellular localization of the protein. Biochem Cell Biol 94:331–335 165. Roussotte FF, Gutman BA, Hibar DP, Madsen SK, Narr KL, Thompson PM et al (2015) Carriers of a common variant in the dopamine transporter gene have greater dementia risk, cognitive decline, and faster ventricular expansion. Alzheimers Dement 11: 1153–1162 ´ , Juha´sz A, Pa´ka´ski M, Kalman J, 166. Fehe´r A Janka Z (2014) Association between the 9 repeat allele of the dopamine transporter 40bp variable tandem repeat polymorphism and Alzheimer’s disease. Psychiatry Res 220: 730–731 167. Yamazaki K, Yoshino Y, Mori T, Okita M, Yoshida T, Mori Y et al (2016) Association study and meta-analysis of polymorphisms, methylation profiles, and peripheral mRNA expression of the serotonin transporter gene in patients with Alzheimer’s disease. Dement Geriatr Cogn Disord 41:334–347

168. Teumer A, Chaker L, Groeneweg S, Li Y, Munno CD, Barbieri C et al (2018) Genome-wide analyses identify a role for SLC17A4 and AADAT in thyroid hormone regulation. Nat Commun 9:4455 169. Xu M, Xiao M, Li S, Yang B (2017) Aquaporins in nervous system. Adv Exp Med 969: 81–103 170. Zeppenfeld DM, Simon M, Haswell JD, D’Abreo D, Association of perivascular localization of aquaporin-4 with cognition and Alzheimer disease in aging brainsMurchison C, Quinn JF et al (2017) . JAMA Neurol 74: 91–99 171. Qureshi YH, Patel VM, Berman DE, Kothiya MJ, Neufeld JL, Vardarajan B et al (2018) An Alzheimer’s disease-linked loss-of-function CLN5 variant impairs cathepsin D maturation, consistent with a retromer trafficking defect. Mol Cell Biol 38:e00011–e00018 172. Mohanty V, Siddiqui MR, Tomita T, Mayanil CS (2017) Folate receptor alpha is more than just a folate transporter. Neurogenesis (Austin) 4:e1263717 173. Teranishi Y, Inoue M, Yamamoto NG, Kihara T, Weihager B, Ishikawa T et al (2015) Proton myo-inositol cotransporter is a novel γ-secretase associated protein that regulates Aβ production without affecting Notch cleavage. FEBS J 282:3438–3451 174. Kanai Y, Cle´menc¸on B, Simonin A, Leuenberger M, Lochner M, Weisstanner M et al (2013) The SLC1 high-affinity glutamate and neutral amino acid transporter family. Mol Asp Med 34:108–120 175. Shubbar MH, Penny JI (2018) Effect of amyloid beta on ATP-binding cassette transporter expression and activity in porcine brain microvascular endothelial cells. Biochim Biophys Acta Gen Subj 1862:2314–2322 176. Vauthier V, Housset C, Falguie`res T (2017) Targeted pharmacotherapies for defective ABC transporters. Biochem Pharmacol 136: 1–11 177. Fan J, Zareyan S, Zhao W, Shimizu Y, Pfeifer TA, Tak JH et al (2016) Identification of a chrysanthemic ester as an apolipoprotein E inducer in astrocytes. PLoS One 11: e0162384 178. Padala AK, Wani A, Vishwakarma RA, Kumar A, Bharate SB (2016) Functional induction of P-glycoprotein efflux pump by phenyl benzenesulfonamides: synthesis and biological evaluation of T0901317 analogs. Eur J Med Chem 122:744–755 179. Cacabelos R, Llovo R, Fraile C, Ferna´ndezNovoa L (2007) Pharmacogenetic aspects of

Pharmacogenomics of Alzheimer’s Disease therapy with cholinesterase inhibitors: the role of CYP2D6 in Alzheimer’s disease pharmacogenetics. Curr Alzheimer Res 4: 479–500 180. Cacabelos R (2007) Donepezil in Alzheimer’s disease: from conventional trials to pharmacogenetics. Neuropsychiatr Dis Treat 3: 303–333 181. Brewster JT, Dell’Acqua S, Thach DQ, Sessler JL (2019) Classics in chemical neuroscience: donepezil. ACS Chem Neurosci 10:155–167 182. Noetzli M, Eap CB (2013) Pharmacodynamic, pharmacokinetic and pharmacogenetic aspects of drugs used in the treatment of Alzheimer’s disease. Clin Pharmacokinet 52:225–241 183. Noetzli M, Guidi M, Ebbing K, Eyer S, Wilhelm L, Michon A et al (2014) Population pharmacokinetic approach to evaluate the effect of CYP2D6, CYP3A, ABCB1, POR and NR1I2 genotypes on donepezil clearance. Br J Clin Pharmacol 78:135–144 184. Xiao T, Jiao B, Zhang W, Tang B, Shen L (2016) Effect of the CYP2D6 and APOE polymorphisms on the efficacy of donepezil in patients with Alzheimer’s disease: a systematic review and meta-analysis. CNS Drugs 30: 899–907 185. Sokolow S, Li X, Chen L, Taylor KD, Rotter JI, Rissman RA et al (2017) Deleterious effect of butyrylcholinesterase K-variant in donepezil treatment of mild cognitive impairment. J Alzheimers Dis 56:229–237 186. Russo P, Kisialiou A, Moroni R, Prinzi G, Fini M (2017) Effect of genetic polymorphisms (SNPs) in CHRNA7 gene on response to acetylcholinesterase inhibitors (AChEI) in patients with Alzheimer’s disease. Curr Drug Targets 18:1179–1190 187. Noetzli M, Guidi M, Ebbing K, Eyer S, Zumbach S, Giannakopoulos P (2013) Relationship of CYP2D6, CYP3A, POR, and ABCB1 genotypes with galantamine plasma concentrations. Ther Drug Monit 35: 270–275 188. Birks JS, Grimley Evans J (2015) Rivastigmine for Alzheimer’s disease. Cochrane Database Syst Rev 10:CD001191 189. Gul A, Bakht J, Mehmood F (2019) Huperzine-A response to cognitive impairment and task switching deficits in patients with Alzheimer’s disease. J Chin Med Assoc 82:40–43 190. Lin PP, Li XN, Yuan F, Chen WL, Yang MJ, Xu HR (2016) Evaluation of the in vitro and in vivo metabolic pathway and cytochrome P450 inhibition/induction profile of

385

Huperzine A. Biochem Biophys Res Commun 480:248–253 191. Noetzli M, Guidi M, Ebbing K, Eyer S, Wilhelm L, Michon A et al (2013) Population pharmacokinetic study of memantine: effects of clinical and genetic factors. Clin Pharmacokinet 52:211–223 192. Bastrup J, Hansen KH, Poulsen TBG, Kastaniegaard K, Asuni AA, Christensen S et al (2021) Anti-Aβ antibody aducanumab regulates the proteome of senile plaques and closely surrounding tissue in a transgenic mouse model of Alzheimer’s disease. J Alzheimers Dis 79:249–265 193. Cacabelos R, Cacabelos P, Carril JC (2019) Epigenetics and pharmacoepigenetics of age-related neurodegenerative disorders. In: Cacabelos R (ed) Pharmacoepigenetics. Academic Press/Elsevier, Oxford, UK, pp 903–950 194. Cacabelos R (2014) Epigenomic networking in drug development: from pathogenic mechanisms to pharmacogenomics. Drug Dev Res 75:348–365 ˜ oz F 195. Cacabelos R, Torrellas C, Lo´pez-Mun (2014) Epigenomics of Alzheimer’s disease. J Exp Med 6:75–82 196. Cacabelos R, Torrellas C (2014) Epigenetic drug discovery for Alzheimer’s disease. Expert Opin Drug Discovery 9:1059–1086 197. Cacabelos R (2019) Pathoepigenetics: the role of epigenetic biomarkers in disease pathogenesis. In: Cacabelos R (ed) Pharmacoepigenetics. Academic Press/ Elsevier, Oxford, UK, pp 139–189 198. Qazi TJ, Quan Z, Mir A, Qing H (2018) Epigenetics in Alzheimer’s disease: perspective of DNA methylation. Mol Neurobiol 55:1026–1044 199. Ciceri F, Rotllant D, Maes T (2017) Understanding epigenetic alterations in Alzheimer’s and Parkinson’s disease: towards targeted biomarkers and therapies. Curr Pharm Des 23: 839–857 200. Swarbrick S, Wragg N, Ghosh S, Stolzing A (2019) Systematic review of miRNA as biomarkers in Alzheimer’s disease. Mol Neurobiol 56:6156–6167 201. Gupta P, Bhattacharjee S, Sharma AR, Sharma G, Lee SS, Chakraborty C (2017) miRNAs in Alzheimer disease – a therapeutic perspective. Curr Alzheimer Res 14: 1198–1206 202. Silvestro S, Bramanti P, Mazzon E (2019) Role of miRNAs in Alzheimer’s disease and possible fields of application. Int J Mol Sci 20: 3979

386

Ramo´n Cacabelos et al.

203. Kandimalla R, Reddy PH (2017) Therapeutics of neurotransmitters in Alzheimer’s disease. J Alzheimers Dis 57:1049–1069 204. Chu LW (2012) Alzheimer’s disease: early diagnosis and treatment. Hong Kong Med J 18:228–237 205. Ferreira-Vieira TH, Guimaraes IM, Silva FR, Ribeiro FM (2016) Alzheimer’s disease: targeting the cholinergic system. Curr Neuropharmacol 14:101–115 206. Serretti A (2017) Genetics and pharmacogenetics of mood disorders. Psychiatr Pol 51: 197–203 207. Wyska E (2019) Pharmacokinetic considerations for current state-of-the-art antidepressants. Expert Opin Drug Metab Toxicol 15: 831–847 208. Tiwari AK, Souza RP, Mu¨ller DJ (2009) Pharmacogenetics of anxiolytic drugs. J Neural Transm 116:667–677 209. Torrellas C, Risso A, Carril JC (2015) Optimizacio´n del uso de antidepresivos con estrategias farmacogene´ticas. Gen-T 10:17–26 210. Osanlou O, Pirmohamed M, Daly AK (2018) Pharmacogenetics of adverse drug reactions. Adv Pharmacol 83:155–190 211. Ross CJ, Carleton B, Warn DG, Stenton SB, Rassekh SR, Hayden MR (2007) Genotypic approaches to therapy in children: a national active surveillance network (GATC) to study the pharmacogenomics of severe adverse drug reactions in children. Ann N Y Acad Sci 1110: 177–192 212. Eissenberg JC, Aurora R (2019) Pharmacogenomics: what the doctor ordered? Mo Med 116:217–225 213. Becquemont L (2009) Pharmacogenomics of adverse drug reactions: practical applications and perspectives. Pharmacogenomics 10: 961–969 214. Elzagallaai AA, Greff M, Rieder MJ (2017) Adverse drug reactions in children: the double-edged sword of therapeutics. Clin Pharmacol Ther 101:725–735 215. Malki MA, Pearson ER (2020) Drug-druggene interactions and adverse drug reactions. Pharmacogenomics J 20:355–366 216. Pirmohamed M (2014) Personalized pharmacogenomics: predicting efficacy and adverse drug reactions. Annu Rev Genomics Hum Genet 15:349–370 217. Aagaard L, Hansen EH (2009) Information about ADRs explored by pharmacovigilance approaches: a qualitative review of studies on antibiotics, SSRIs and NSAIDs. BMC Clin Pharmacol 9:4

218. Magro L, Moretti U, Leone R (2012) Epidemiology and characteristics of adverse drug reactions caused by drug-drug interactions. Expert Opin Drug Saf 11:83–94 219. Weinshilboum RM, Wang L (2017) Pharmacogenomics: precision medicine and drug response. Mayo Clin Proc 92:1711–1722 220. Roden DM, McLeod HL, Relling MV, Williams MS, Mensah GA, Peterson JF et al (2019) Pharmacogenomics. Lancet 394: 521–532 221. Borroni RG (2015) Role of dermatology in pharmacogenomics: drug-induced skin injury. Pharmacogenomics 16:401–412 222. Davies EC, Green CF, Mottram DR, Pirmohamed M (2007) Adverse drug reactions in hospitals: a narrative review. Curr Drug Saf 2: 79–87 223. Chan SL, Ang X, Sani LL, Ng HY, Winther MD, Liu JJ et al (2016) Prevalence and characteristics of adverse drug reactions at admission to hospital: a prospective observational study. Br J Clin Pharmacol 82:1636–1646 224. Empey PE (2010) Genetic predisposition to adverse drug reactions in the intensive care unit. Crit Care Med 38:S106–S116 225. Alessandrini M, Chaudhry M, Dodgen TM, Pepper MS (2016) Pharmacogenomics and global precision medicine in the context of adverse drug reactions: top 10 opportunities and challenges for the next decade. OMICS 20:593–603 226. Kim GJ, Lee SY, Park JH, Ryu BY, Kim JH (2017) Role of preemptive genotyping in preventing serious adverse drug events in south Korean patients. Drug Saf 40:65–80 227. Schildcrout JS, Denny JC, Bowton E, Pulley JM, Basford MA, Cowan JD et al (2012) Optimizing drug outcomes through pharmacogenetics: a case for preemptive genotyping. Clin Pharmacol Ther 92:235–242 228. Dunnenberger HM, Crews KR, Hoffman JM, Caudle KE, Broeckel U, Howard SC et al (2015) Preemptive clinical pharmacogenetics implementation: current programs in five US medical centers. Annu Rev Pharmacol Toxicol 55:89–106 229. Roden DM, Van Driest SL, Mosley JD, Wells QS, Robinson JR, Denny JC et al (2018) Benefit of preemptive pharmacogenetic information on clinical outcome. Clin Pharmacol Ther 103:787–794 230. Aagaard L, Hansen EH (2013) Adverse drug reactions reported by consumers for nervous system medications in Europe 2007 to 2011. BMC Pharmacol Toxicol 14:30

Pharmacogenomics of Alzheimer’s Disease 231. Cacabelos R (2019) Epigenetics and pharmacoepigenetics of neurodevelopmental and neuropsychiatric disorders. In: Cacabelos R (ed) Pharmacoepigenetics. Academic Press/ Elsevier, Oxford, UK, pp 609–709 232. Kam H, Jeong H (2020) Pharmacogenomic biomarkers and their applications in psychiatry. Genes 11:1445 233. Cacabelos R, Torrellas C (2015) Pharmacogenomics of antidepressants. HSOA J Psych Dep Anx 1:001 234. Cacabelos R (2020) Pharmacogenomics of Alzheimer’s and Parkinson’s diseases. Neurosci Lett 726:133807 235. Cacabelos R (2009) Pharmacogenomic biomarkers in neuropsychiatry: the path to personalized medicine in mental disorders. In: Ritsner MS (ed) The handbook of neuropsychiatric biomarkers, endophenotypes and genes. Vol. IV: Molecular genetic and genomic markers. Springer, Amsterdam, pp 3–63 236. Caroli A, Frisoni GB, Alzheimer’s Disease Neuroimaging Initiative (2010) The dynamics of Alzheimer’s disease biomarkers in the Alzheimer’s disease neuroimaging initiative cohort. Neurobiol Aging 31:1263–1274 237. Jack CR Jr, Knopman DS, Jagust WJ, Petersen RC, Weiner MW, Aisen PS et al (2013) Tracking pathophysiological processes in Alzheimer’s disease: an updated hypothetical model of dynamic biomarkers. Lancet Neurol 12:207–216 238. Blennow K, Zetterberg H (2018) Biomarkers for Alzheimer’s disease: current status and prospects for the future. J Intern Med 284: 643–663 239. Zetterberg H, Bendlin BB (2021) Biomarkers for Alzheimer’s disease: preparing for a new era of disease-modifying therapies. Mol Psychiatry 26:296–308 240. Ma´rquez F, Yassa MA (2019) Neuroimaging biomarkers for Alzheimer’s disease. Mol Neurodegener 14:21

387

241. Cacabelos R, Lombardi V, Ferna´ndez-Novoa L (2004) A functional genomics approach to the analysis of biological markers in Alzheimer disease. In: Takeda M, Tanaka T, Cacabelos R (eds) Molecular neurobiology of Alzheimer’s Disease and related disorders. Karger, Basel, pp 236–285 242. Ruths S, Straand J, Nygaard HA (2001) Psychotropic drug use in nursing homes: diagnostic indications and variations between institutions. Eur J Clin Pharmacol 57: 523–528 243. Brimelow RE, Wollin JA, Byrne GJ, Dissanayaka NN (2019) Prescribing of psychotropic drugs and indicators for use in residential aged care and residents with dementia. Int Psychogeriatr 31:837–847 244. Holmquist IB, Svensson B, Ho¨glund P (2005) Perceived anxiety, depression, and sleeping problems in relation to psychotropic drug use among elderly in assisted-living facilities. Eur J Clin Pharmacol 61:215–224 245. Zahirovic I, Torisson G, Wattmo C, Londos E (2018) Psychotropic and anti-dementia treatment in elderly persons with clinical signs of dementia with Lewy bodies: a cross-sectional study in 40 nursing homes in Sweden. BMC Geriatr 18:50 246. Gulla C, Selbaek G, Flo E, Kjome R, Kirkevold O, Husebo BS (2016) Multipsychotropic drug prescription and the association to neuropsychiatric symptoms in three Norwegian nursing home cohorts between 2004 and 2011. BMC Geriatr 16:115 247. Janus SI, van Manen JG, IJzerman MJ, Zuidema SU (2016) Psychotropic drug prescriptions in Western European nursing homes. Int Psychogeriatr 28:1775–1790 248. Helvik AS, Sˇaltyte˙ Benth J, Wu B, Engedal K, Selbaek G (2017) Persistent use of psychotropic drugs in nursing home residents in Norway. BMC Geriatr 17:52

Chapter 14 Pharmacogenetics of Antipsychotic Treatment in Schizophrenia Samar S. M. Elsheikh, Daniel J. Mu¨ller, and Jennie G. Pouget Abstract Antipsychotics are the mainstay treatment for schizophrenia. There is large variability between individuals in their response to antipsychotics, both in efficacy and adverse effects of treatment. While the source of interindividual variability in antipsychotic response is not completely understood, genetics is a major contributing factor. The identification of pharmacogenetic markers that predict antipsychotic efficacy and adverse reactions is a growing area of research and holds the potential to replace the current trial-and-error approach to treatment selection in schizophrenia with a personalized medicine approach. In this chapter, we provide an overview of the current state of pharmacogenetics in schizophrenia treatment. The most promising pharmacogenetic findings are presented for both antipsychotic response and commonly studied adverse reactions. The application of pharmacogenetics to schizophrenia treatment is discussed, with an emphasis on the clinical utility of pharmacogenetic testing and directions for future research. Key words Antipsychotics, Pharmacogenetics, Genetics, Schizophrenia, Response, Side effects

1

Introduction Schizophrenia is a debilitating disorder affecting 1% of the global population. It is a pervasive disease, affecting many aspects of mental function. Positive, negative, affective, and cognitive symptom clusters characterize schizophrenia (Table 1). Antipsychotics are the current standard of care in schizophrenia management. There are two classes of antipsychotics: typical or first-generation antipsychotics (FGAs) and atypical or secondgeneration antipsychotics (SGAs). Both classes of antipsychotics block dopamine D2 receptors, and this dopaminergic antagonism is considered necessary and sufficient for antipsychotic action [1]. While there are no major differences in efficacy of FGAs and SGAs, their tolerability profiles are diverse. FGAs are more likely to cause extrapyramidal side effects (e.g., acute motor side effects, tardive dyskinesia) and increased prolactin secretion, while SGAs

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_14, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

389

390

Samar S. M. Elsheikh et al.

Table 1 Symptom clusters in schizophrenia Symptom cluster

Clinical examples

Positive symptoms

Delusions, hallucinations, and disorganized speech

Negative symptoms

Avolition, anhedonia, and poverty of thought

Affective symptoms

Depression, anxiety, and affective flattening

Cognitive symptoms

Attention memory executive function

Adapted with permission from Ref. [186]

are generally more likely to cause marked weight gain. These differences in adverse drug reactions are most likely because FGAs dissociate more slowly from D2 receptors, resulting in distortion of physiological dopaminergic transmission, and because SGAs modulate a variety of additional neurotransmitter systems (Table 2) [1]. The selection of appropriate antipsychotic treatment is often a “trial-and-error” procedure, with multiple failed trials required before achieving an acceptable balance between symptom management and tolerability of adverse effects. This is problematic because it increases the risk of adverse drug reactions and delays symptom management, worsening long-term treatment outcomes [2]. There is large variability between individuals in their response to and tolerability of antipsychotic treatment. Some patients enter complete remission following treatment with a particular antipsychotic, while others treated with the same antipsychotic show no response or experience common adverse effects such as tardive dyskinesia or weight gain. While environmental factors such as lifestyle habits (smoking, diet), demographics (sex, ethnicity), and health status (concurrent medications, illness onset and duration, medical comorbidities) contribute to the variability in both response and tolerability of antipsychotic treatment, there is also a clear genetic contribution to this variability [3–7]. Given the underlying genetic regulation of response to antipsychotic treatment, pharmacogenetics holds the potential to provide a robust rationale for treatment optimization. The goal of pharmacogenetics in schizophrenia is to replace the current trial-and-error treatment paradigm with a personalized medicine approach, allowing clinicians to map the right dose of the right drug to first-episode schizophrenia patients based on their genetic profile [8]. In recent years, much progress has been made in identifying genetic variants associated with antipsychotic response and adverse reactions to treatment. Initially, candidate gene studies were conducted to explore single nucleotide polymorphisms (SNPs) in pharmacokinetic genes affecting the bioavailability of antipsychotics (through absorption, distribution, metabolism, and excretion), with a particular emphasis on the cytochrome P450 enzymes.

Pharmacogenetics of Schizophrenia

391

Table 2 Differences in pharmacology and side effect profiles of FGAs and SGAs First generation

Second generation

Metabolism CYP2D6 Chlorpromazinea Fluphenazinea Haloperidola Perphenazinea Thioridazinea CYP3A4 Haloperidola CYP1A2 Chlorpromazinea Loxapinea Perphenazinea Thioridazinea Thiothixenea Trifluoperazinea

CYP2D6 Aripiprazolea Brexpiprazole Cariprazine Clozapineb Iloperidonea Olanzapineb Paliperidone Pimozide Risperidonea CYP3A4 Aripiprazolea Clozapinea Iloperidonea Lurasidoneb Quetiapineb Risperidonea Ziprasidonea CYP1A2 Clozapinea Olanzapinea CYP2C19 Clozapine

Site of action

DRD2: Higher-affinity antagonists. Variable DRD2: Lower-affinity antagonists. HTR2: effects on other receptors: serotonergic, Higher-affinity antagonists. Variable adrenergic, histaminic, and muscarinic effects on other receptors: adrenergic, histaminic, and muscarinic

Acute motor side effects Common Tardive dyskinesia adverse reactions Hyperprolactinemia

Weight gain and metabolic disturbances Sedation Agranulocytosisc Spatial working memory

a

Primary metabolism Secondary metabolism c This adverse reaction is associated primarily with clozapine treatment and occurs in a minority of patients b

Considerable attention has also been given to studying pharmacodynamic genes affecting the mechanism of antipsychotic drug action (through neurotransmitter transporters or receptors). Genome-wide association studies (GWAS) have identified variants in previously uninvestigated genes that are associated with antipsychotic response and adverse reactions [9–13]. The scope of the present chapter is to provide an overview of the current state of pharmacogenetics in schizophrenia treatment. The most promising pharmacogenetic findings will be presented for both antipsychotic response and the most commonly studied antipsychotic side effects. The application of pharmacogenetics to

392

Samar S. M. Elsheikh et al.

schizophrenia treatment will also be discussed, with an emphasis on the clinical utility of pharmacogenetic testing and directions for future research.

2

Pharmacogenetics of Antipsychotic Response Response to antipsychotics is a complex phenotype likely involving numerous different genes, making pharmacogenetic studies in this area challenging. Despite these methodological challenges, a number of genetic variants have been consistently observed in association with antipsychotic treatment response, and preliminary efforts to integrate these pharmacogenetic findings into treatment selection for schizophrenia patients are now underway.

2.1 Pharmacokinetic Candidates in Antipsychotic Response

As most antipsychotic medications undergo extensive first-pass metabolism, drug metabolizing enzymes (DME) play an important role in patient response to antipsychotic treatment by determining drug bioavailability (the fraction of the antipsychotic that reaches the systemic circulation and is available to act on its targets in the brain). Some antipsychotics, such as clozapine and risperidone, also undergo bioactivation by DME resulting in the generation of active metabolites of the parent drug, which may have toxic or therapeutic effects. DMEs are also important in antipsychotic clearance (the ability of the body to eliminate the drug). The cytochrome P450 (CYP) enzymes are the major enzymes influencing antipsychotic bioavailability, bioactivation, and clearance [14]. The genes encoding CYP enzymes are typically polymorphic, and their variation leads to decreased or elevated catalytic activity. Individual CYP genotypes contribute to various combinations of haplotypes commonly classified as “star (*) alleles,” which are considered to be functionally “active,” “decreased activity,” or “inactive” in terms of catalytic activity. An individual’s phenotype for a particular CYP enzyme is commonly classified as “poor metabolizer” (two inactive alleles), “intermediate metabolizer” (one inactive allele + one active or decreased activity allele or two decreased activity alleles), “extensive/normal metabolizer” (two active alleles), or “ultrarapid metabolizer” (gene duplication with no inactive or decreased activity alleles). Therefore, genetic variation in CYPs affects their catalytic activity, contributing to differences in the bioavailability, bioactivation, and clearance of antipsychotic drugs between patients and likely influencing their drug plasma levels or metabolite ratios and, consequently, their response and adverse effect profiles. In addition to the CYP enzymes, drug transporters in the blood–brain barrier play an important role in the pharmacokinetics of antipsychotics by regulating their accumulation in the brain. Several antipsychotics are substrates of the P-glycoprotein

Pharmacogenetics of Schizophrenia

393

transporter [15]. P-glycoprotein is expressed ubiquitously by cells of the blood–brain barrier and acts as an efflux pump to remove antipsychotics from brain tissues. By influencing the transport efficacy of P-glycoprotein, polymorphisms in the P-glycoprotein gene ABCB1 may lead to variability among schizophrenia patients in the accumulation of antipsychotics in their brains, thereby leading to differences in treatment response. As antipsychotic drugs are metabolized primarily by CYP1A2, CYP2D6, and CYP3A4, with CYP2C19 partially involved in clozapine metabolism [16], the most promising findings in relation to these enzymes, in addition to the P-glycoprotein gene ABCB1, will be reviewed here (see Table 3 for a summary of pharmacokinetic findings). As a supplement to the findings presented here, the interested reader is referred to the comprehensive review by Yoshida et al. [17]. 2.1.1

CYP2C19

A recent study used a sample of 108 individuals with treatmentrefractory schizophrenia (TRS) and superrefractory schizophrenia (SRS) to assess the association between clozapine treatment response and CYP2C19 variants [18]. The study found that two variants, CYP2C19*2 and CYP2C19*17, were significantly associated with clozapine response measured using the Brief Psychiatric Rating Scale (BPRS), in both TRS (BPRS scores 0. Normal metabolizers had exposures approximately 11.4-fold greater than poor metabolizers following this single dose [26]. While older pharmacokinetic studies of atomoxetine in adults focused primarily on normal vs. poor metabolizers, two studies in Chinese [27] and Japanese [28] populations and one study conducted in South Korea [29] have examined the impact of the CYP2D6 *10 allele on atomoxetine pharmacokinetics. The *10 allele is found at higher frequencies in East Asian populations and is decreased in function as compared to normal function alleles (e.g., *1, *2) [30, 31]. Cui et al. conducted a single- (40 mg) and multiple-dose (80 mg) pharmacokinetic study of atomoxetine in Chinese adults and compared pharmacokinetic parameters between CYP2D6 genotype groups of *10/*10 (n ¼ 7) and individuals who were *1/*10 (n ¼ 7) or *1/*1 (n ¼ 2). When comparing individuals who were homozygous for the *10 allele vs. those with one or no copies of the *10 allele after a single 40 mg dose, exposure was approximately 2.2-fold higher (4962 ng/ml*h vs. 2242 ng/ml*h), half-life was 1.6-fold longer (5.4 h vs. 3.4 h), and clearance about half (0.13 l/h/kg vs. 0.29 l/h/kg), all of which were statistically significant [27]. Matsui et al. compared Japanese individuals’ atomoxetine pharmacokinetic parameters through a single 10 mg or 120 mg dose with CYP2D6 genotypes of *1/*1 or *1/*2 (n ¼ 5), *1/*10 or *2/*10 (n ¼ 5), and *10/*10 (n ¼ 4). Ratios of AUC and Cmax between the two doses were similar, with individuals homozygous for the *10 allele having approximately 2.2- and 1.8-fold higher AUC0-inf and 1.5- and 1.3-fold higher Cmax as compared to with the *1/*10 or *2/*10 and *1/*1 or *1/*2 genotypes. Individuals with the *10/*10 genotype also had significantly lower clearances as compared to the other two genotype groups [28].

Atomoxetine Half-Life (h)

Pharmacogenetics and Pharmacokinetics of ADHD Medications

433

20 15 10 5

Sauer et al

*wt/*wt

*wt/*10

*10/*10

*10/*10

Non-Poor Metabolizers

Brown et al

AS=0

AS=0.5

AS=1

AS=2

Non-Poor Metabolizers

Poor Metabolizers

0

Byeon et al

Cui et al

CYP2D6 Activty Fig. 2 Atomoxetine half-life across different levels of CYP2D6 activity

Byeon et al. conducted a similar study with a larger sample size in which they compared atomoxetine pharmacokinetics in individuals who were CYP2D6 homozygous wild-type (n ¼ 22), wildtype/*10 (n ¼ 22), and *10/*10 (n ¼ 18). Those with the CYP2D6 *10/*10 genotype had significantly greater Cmax, halflife, and AUC0-inf and lower apparent oral clearance values as compared to both the *wild-type/*wild-type and *wild-type/*10 groups. When comparing the *10/*10 genotype to the homozygous wild-type groups, Cmax was 1.7-fold greater, half-life was 1.8fold greater, AUC0-inf was 3.4-fold greater, and apparent oral clearance was 3.3-fold lower [29]. The impact of CYP2D6 variation on the pharmacokinetic disposition of atomoxetine is well described in both children and adults, particularly as it pertains to the influence of the *10 allele (Fig. 2). Additionally, dosing guidance is available for when CYP2D6 genotypes are available prior to atomoxetine initiation, allowing for an approach that may reduce the overall variability in exposure. 3.3 Mixed Amphetamine Salts and CYP2D6

Limited published data exists regarding the pharmacogenetic impact on the pharmacokinetics of mixed amphetamine salts. The package insert for Adderall states that CYP2D6 contributes to the active 4-hydroxy-amphetamine metabolite although little other

434

Jacob T. Brown

data is available [32]. The FDA Table of Pharmacogenetic Associations does recommend considering a lower starting dose or the use of an alternative agent in CYP2D6 poor metabolizers for mixed amphetamine salts due to effects on systemic concentrations and risk of adverse reactions [31, 33, 34]. 3.4 Guanfacine and CYP3A4/5

Guanfacine is a non-stimulant medication used as adjunct therapy to stimulants in the treatment of ADHD. Known to be a CYP3A4/ 5 substrate, there are no studies to date on how genetic variation in CYP3A4/5 impacts the pharmacokinetic disposition of guanfacine.

3.5 Clonidine and CYP2D6

Similar to guanfacine, clonidine extended release is also a non-stimulant used in the treatment of ADHD either as monotherapy or adjunct to stimulants. Clonidine appears to be a substrate for CYP2D6 in vitro [33, 35]. The Dutch Pharmacogenetic Working Guidelines describe clonidine and CYP2D6 as not being a gene-drug interaction [21], and currently, no other pharmacogenetic guidelines exist for this gene-drug pair.

4

Notes 1. The impact of CYP2D6 variation on atomoxetine pharmacokinetics and exposure is considerable, as reflected by the summarized pharmacokinetic studies and dosing recommendations for CYP2D6 poor metabolizers in the package insert, the recently published CPIC guideline for CYP2D6 and atomoxetine, and recommendations by the DPWG. Pharmacogenetic testing for CYP2D6 is readily available to clinicians and patients, making the guidance for how to use this information when prescribing atomoxetine important. While atomoxetine’s pharmacokinetic disposition as it relates to CYP2D6 genotype is reasonably well described in individuals who are normal, intermediate, and poor metabolizers, future studies focusing on CYP2D6 ultrarapid metabolizers will help to characterize individuals with increased CYP2D6 activity. 2. The impact of the CES1 variant rs71647871 on methylphenidate pharmacokinetics appears to be significant although the clinical impact of this remains to be seen. Despite this, pharmacogenetic testing companies are starting to include CES1 as part of the genes included in their reports. As variation in the CES1 gene is better understood and characterized, it will enhance the current level of understanding of which variants are most clinically relevant. To date, limited published data is available for the influence of pharmacogenetics on the pharmacokinetics of mixed amphetamine salts, guanfacine, and clonidine.

Pharmacogenetics and Pharmacokinetics of ADHD Medications

435

3. Provider, parent, and patient interest in having a predictive pharmacogenetic test to help guide medication dosing and selection for the treatment of ADHD is considerable. As additional research helps describe the relationship between exposure and clinical response for these medications, the application of pharmacogenetic testing will become more pertinent to individual treatment. Although the focus of this chapter is limited to the pharmacogenetic impact on pharmacokinetics of ADHD medications, it is important to note that the pharmacodynamic genes relevant to these medications may also play a role in predictive testing in the future. As discussed in Elsayed et al., it is likely that future clinical application of pharmacogenetic testing in ADHD medications will include multiple genes and/or polygenic risk scores to guide ADHD pharmacotherapy [4]. References 1. Volkow ND, Swanson JM (2013) Clinical practice: adult attention deficit-hyperactivity disorder. N Engl J Med 369:1935–1944 2. Posner J, Polanczyk GV, Sonuga-Barke E (2020) Attention-deficit hyperactivity disorder. Lancet Lond Engl 395:450–462 3. Stevens T, Sangkuhl K, Brown JT et al (2019) PharmGKB summary: methylphenidate pathway, pharmacokinetics/pharmacodynamics. Pharmacogenet Genomics 29:136–154 4. Elsayed NA, Yamamoto KM, Froehlich TE (2020) Genetic influence on efficacy of pharmacotherapy for pediatric attention-deficit/ hyperactivity disorder: overview and current status of research. CNS Drugs 34:389–414 5. Brown JT, Bishop JR, Sangkuhl K et al (2019) Clinical pharmacogenetics implementation consortium guideline for cytochrome P450 (CYP)2D6 genotype and atomoxetine therapy. Clin Pharmacol Ther 106:94–102 6. Markowitz JS, Straughn AB, Patrick KS (2003) Advances in the pharmacotherapy of attentiondeficit-hyperactivity disorder: focus on methylphenidate formulations. Pharmacotherapy 23: 1281–1299 7. Merali Z, Ross S, Pare´ G (2014) The pharmacogenetics of carboxylesterases: CES1 and CES2 genetic variants and their clinical effect. Drug Metabol Drug Interact 29:143–151 8. Her L, Zhu H-J (2020) Carboxylesterase 1 and precision pharmacotherapy: pharmacogenetics and nongenetic regulators. Drug Metab Dispos Biol Fate Chem 48:230–244 9. Wang X, Rida N, Shi J et al (2017) A comprefunctional assessment of hensive

carboxylesterase 1 nonsynonymous polymorphisms. Drug Metab Dispos Biol Fate Chem 45:1149–1155 10. Zhu H-J, Patrick KS, Yuan H-J et al (2008) Two CES1 gene mutations lead to dysfunctional carboxylesterase 1 activity in man: clinical significance and molecular basis. Am J Hum Genet 82:1241–1248 11. Tarkiainen EK, Backman JT, Neuvonen M et al (2012) Carboxylesterase 1 polymorphism impairs oseltamivir bioactivation in humans. Clin Pharmacol Ther 92:68–71 12. Tarkiainen EK, Tornio A, Holmberg MT et al (2015) Effect of carboxylesterase 1 c.428G >A single nucleotide variation on the pharmacokinetics of quinapril and enalapril. Br J Clin Pharmacol 80:1131–1138 13. Lewis JP, Horenstein RB, Ryan K et al (2013) The functional G143E variant of carboxylesterase 1 is associated with increased clopidogrel active metabolite levels and greater clopidogrel response. Pharmacogenet Genomics 23:1–8 14. Pare´ G, Eriksson N, Lehr T et al (2013) Genetic determinants of dabigatran plasma levels and their relation to bleeding. Circulation 127:1404–1412 15. Stage C, Ju¨rgens G, Guski LS et al (2017) The impact of CES1 genotypes on the pharmacokinetics of methylphenidate in healthy Danish subjects. Br J Clin Pharmacol 83:1506–1514 16. Lyauk YK, Stage C, Bergmann TK et al (2016) Population pharmacokinetics of methylphenidate in healthy adults emphasizing novel and known effects of several carboxylesterase 1 (CES1) variants. Clin Transl Sci 9:337–345

436

Jacob T. Brown

17. Stage C, Dalhoff K, Rasmussen HB et al (2019) The impact of human CES1 genetic variation on enzyme activity assessed by ritalinic acid/methylphenidate ratios. Basic Clin Pharmacol Toxicol 125:54–61 18. ClinicalTrials.gov [Internet]. [cited 15 Sep 2022]. Available from: https://clinicaltrials. gov/ 19. Brown JT, Bishop JR (2015) Atomoxetine pharmacogenetics: associations with pharmacokinetics, treatment response and tolerability. Pharmacogenomics 16:1513–1520 20. Strattera [Package insert]. Eli Lilly and Company, Indianapolis 21. Swen JJ, Nijenhuis M, de Boer A et al (2011) Pharmacogenetics: from bench to byte – an update of guidelines. Clin Pharmacol Ther 89:662–673 22. Zerbe RL, Rowe H, Enas GG et al (1985) Clinical pharmacology of tomoxetine, a potential antidepressant. J Pharmacol Exp Ther 232: 139–143 23. Farid NA, Bergstrom RF, Ziege EA et al (1985) Single-dose and steady-state pharmacokinetics of tomoxetine in normal subjects. J Clin Pharmacol 25:296–301 24. Sauer J-M, Ponsler GD, Mattiuz EL et al (2003) Disposition and metabolic fate of atomoxetine hydrochloride: the role of CYP2D6 in human disposition and metabolism. Drug Metab Dispos Biol Fate Chem 31:98–107 25. Witcher JW, Long A, Smith B et al (2003) Atomoxetine pharmacokinetics in children and adolescents with attention deficit hyperactivity disorder. J Child Adolesc Psychopharmacol 13:53–63 26. Brown JT, Abdel-Rahman SM, van Haandel L et al (2016) Single dose, CYP2D6 genotypestratified pharmacokinetic study of

atomoxetine in children with ADHD. Clin Pharmacol Ther 99:642–650 27. Cui YM, Teng CH, Pan AX et al (2007) Atomoxetine pharmacokinetics in healthy Chinese subjects and effect of the CYP2D6*10 allele. Br J Clin Pharmacol 64:445–449 28. Matsui A, Azuma J, Witcher JW et al (2012) Pharmacokinetics, safety, and tolerability of atomoxetine and effect of CYP2D6*10/*10 genotype in healthy Japanese men. J Clin Pharmacol 52:388–403 29. Byeon J-Y, Kim Y-H, Na H-S et al (2015) Effects of the CYP2D6*10 allele on the pharmacokinetics of atomoxetine and its metabolites. Arch Pharm Res 38:2083–2091 30. Bradford LD (2002) CYP2D6 allele frequency in European Caucasians, Asians, Africans and their descendants. Pharmacogenomics 3: 229–243 31. Sistonen J, Sajantila A, Lao O et al (2007) CYP2D6 worldwide genetic variation shows high frequency of altered activity variants and no continental structure. Pharmacogenet Genomics 17:93–101 32. Adderall XR [Package insert]. Shire LLC, Wayne 33. Claessens AJ, Risler LJ, Eyal S et al (2010) CYP2D6 mediates 4-hydroxylation of clonidine in vitro: implication for pregnancyinduced changes in clonidine clearance. Drug Metab Dispos Biol Fate Chem 38:1393–1396 34. Table of Pharmacogenetic Associations [Internet]. [cited 15 Sep 2021]. Available from: https://www.fda.gov/medical-devices/preci sion-medicine/table-phar macogeneticassociations 35. Li X-Y, Hu X-X, Yang F et al (2019) Effects of 24 CYP2D6 variants found in Chinese population on the metabolism of clonidine in vitro. Chem Biol Interact 313:108840

Chapter 16 Pharmacogenetics of Addiction Therapy David P. Graham, Mark J. Harding, and David A. Nielsen Abstract Drug addiction is a serious relapsing disease that has high costs to society and to the individual addicts. Treatment of these addictions is still in its nascency, with only a few examples of successful therapies. Therapeutic response depends upon genetic, biological, social, and environmental components. A role for genetic makeup in the response to treatment has been shown for several addiction pharmacotherapies with response to treatment based on individual genetic makeup. In this chapter, we will discuss the role of genetics in pharmacotherapies, specifically for cocaine, alcohol, and opioid dependences. The continued elucidation of the role of genetics should aid in the development of new treatments and increase the efficacy of existing treatments. Key words Gene, Alcohol, Cocaine, Opioid, Addiction, Dependence, Abuse, Drug, Therapy, Polymorphism

1

Introduction Addiction to illicit drugs such as cocaine and opioids, as well as to legal drugs such as alcohol and nicotine, can lead to both physiological and social morbidities and have significant economic impact. It is estimated that in the United States alone, over $400 billion per year is lost in productivity and time overall, of which $120 billion is incurred in health-care settings either providing treatment for drug-related behaviors ($35 billion) or providing treatment of the injuries, infections, and illnesses associated with drug-related behaviors ($85 billion) [1]. In addition, drug addiction may lead to other risky behaviors, making comorbid drug use and addiction, emergency care, and HIV infection more likely [2–5]. Cigarette smoking is the leading cause of preventable death nationally, with more than 16 million people having at least one smoking-caused disease, amounting to $170 billion in direct medical costs [6]. Similarly, excessive alcohol use contributes to 88,000 deaths nationally each year, including 1 in 10 deaths among working-age adults

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_16, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

437

438

David P. Graham et al.

[7, 8], with binge drinking being responsible for over half the deaths and three-quarters of the $249 billion economic cost due to excessive use [9]. In 2019, among Americans aged 12 or older, 12.1 million were addicted to alcohol only, 5.9 million to illicit drugs only (for marijuana, cocaine, heroin, hallucinogens, inhalants, methamphetamine, or prescription drugs misuse) and 2.4 million with codependency to both alcohol and illicit drugs [10]. Approximately 1 million cocaine users were classified as having dependence or abuse. Although the use of cocaine is lower than that of alcohol or opioids, the numbers are still considerable. In 2002, there were 1.5 million cocaine users (0.6% of population) in the United States, decreasing to 1 million users (0.4% of population) in 2019 [10]. Alcohol, a legal drug in most of the world, is used by 50.8% of people aged 12 or older in the United States. However, of those using alcohol, 23.9% participated in binge drinking of five or more drinks on the same occasion in the last month, with 54.3% of young adults aged 18–25 engaging in binge drinking [10]. Alcoholism in the United States is slowly declining, with 14.5 million persons aged 12 or older having an alcohol use disorder compared to 18.1 million in 2002. However, these numbers still translate to 5.3% of the US population, a considerable proportion. Among other drugs, marijuana has shown a significant increase of use in 2019 compared to 2018, with 17.5% of the American population reporting use (48.2 million). Rates for other substances include psychotherapeutic drugs 5.9% (16.3 million), hallucinogens 2.2% (6.0 million), cocaine 2.0% (5.5 million), inhalants 0.8% (2.1 million), methamphetamines 0.7% (2.0 million), and heroin 0.3% (745,000). Heroin alone underestimates the actual risk of opioid misuse. Overall, 3.7% of the total American population (10.1 million) have been diagnosed with opioid misuse [10]. Opioid addiction in the United States has been increasing at alarming rates. Just over ten million persons aged 12 or older have misused opioids in the past year [10]. The majority, 96.6% (9.7 million), is prescription pain reliever misusers, while the 745,000 heroin users make up 7.4% of opioid misusers. Of those using heroin, 404,000 misuse both heroin and a prescription pain reliever (4.0% of overall opioid misusers). The main prescription pain relievers used in 2019 were hydrocodone (5.1 million), oxycodone (3.2 million), and fentanyl (269,000) [10]. Half of the pain relievers (50.8%) were obtained from family and friends (37.0% given freely, 9.2% were bought from, and 4.6% were taken without permission), 35.7% were prescribed by a doctor, 1.1% were obtained from more than one doctor, 0.8% were stolen from a health-care provider, 6.2% were bought from a drug dealer or stranger, and 5.5% were obtained in some other way. In the treatment of chronic pain, opioid addiction develops in only about 3% of those treated in general medical practice [11].

Addiction Pharmacogenetics

439

Addiction develops in several stages: initiation of drug use, intermittent to regular use, and, finally, addiction and relapse [12, 13]. Features of addiction are the development of dependence to the drug such that there is a physiological need for the drug for the individual to function properly, the development of tolerance whereby larger doses of the drug are required to achieve the same effect, and the development of withdrawal symptoms that occur when a drug is discontinued. Addiction pharmacotherapy interventions try to reduce or reverse some of these features to attenuate craving and drug use and to prevent possible withdrawal and relapse after drug use is terminated. The intent of this review is to present the evidence to date of the role of genetic variation in the pharmacological treatment of cocaine, alcohol, and opioid addiction. We will communicate these findings by first presenting the major neurochemical systems, dopaminergic, serotonergic, and opioidergic, and then by discussing the genes that function in other physiological pathways. Within these sections, we will discuss the genes for which genetic variants have been found to be associated with pharmacotherapeutic response for addiction treatment, with respect to three addictions: alcohol, cocaine, and opioids. As poignantly noted by Kanzler and colleagues, other than having direct access to relevant tissue samples for in vitro testing, the largest difference among oncology and cardiovascular pharmacogenetic biomarkers compared to addictive disorders is funding. The author’s conclusion was that increased research funding in personalized addiction treatments may yield large benefits for public health and reductions in treatment costs [14].

2

Pharmacogenetics Pharmacogenetics is the study of genetic variation that affects treatment response, where “treatment response” may be defined in terms of treatment outcome or adverse effects. New technologies that consider the full genome have given rise to a newer term, pharmacogenomics. Only recently has pharmacogenetics been applied to the study of addiction treatment, and only a few examples of FDA-approved treatment regimens exist that are successful for alcoholism and opioid addiction. For cocaine and methamphetamine addiction, no FDA-approved therapies currently exist. A central goal of pharmacogenetic research focuses on drugs that target craving in order to promote abstinence and on understanding the mechanistic differences of drug addiction affecting individuals. Response to addiction pharmacotherapy is complex, depending upon genetic, biological, environmental, and social components. A substantial portion of the success rate of a therapy may depend upon the genetic makeup of those receiving the treatment. For psychiatric diseases in general, pharmacotherapies

440

David P. Graham et al.

succeed in only 60–70% of patients [15]. In a recent review of the genetic epidemiology of substance use disorder [16], the authors noted several key takeaways. First, there are multiple genetic influences on substance use affecting multiple neurotransmitter pathways (notably the dopaminergic system), drug processing, and drug metabolism. Second, gene-environment studies have emphasized the importance of the environmental context for substance use. Third, multiple substances have been shown to produce drugspecific epigenetic changes in gene expression, and these changes are expected to differ by the stage of substance use (e.g., initiation versus chronic use). Evidence from a number of medical specialties demonstrates that the consideration of a patient’s genetic makeup can improve the initial selection of medications [17–19]. The use of genetic information may increase both compliance and positive therapeutic response, as well as avoid dangerous side effects due to toxicity [17, 18, 20–24], but will require that many biological and technical challenges be addressed to realize the goal of geneticinformed personalized prescribing [25].

3

Genetics The human genome consists of 3.2 billion nucleotides of DNA composed of four nucleobases [adenine (A), guanine (G), cytosine (C), and thymine (T)], phosphate groups, and deoxyribose sugar molecules. The order of these nucleotides encodes genomic information the same way the order of letters creates words and words create sentences. The protein-coding regions of the human genome make up only 2% of our total genome. Individual amino acids are coded by 3-nucleotide triplet codons, with the sequence of these codons specifying the order of the amino acids that are assembled to form proteins. Every cell contains two copies of each chromosome, receiving one set from the father and one from the mother. A variant form of a gene is called an allele, a term that can describe either an individual gene or a larger region of a chromosome (or the entire chromosome). The paternal chromosome can be considered one allele with the maternal chromosome being the other allele. For each allele, the nucleotide at a specific location may be the same or different. Combining the nucleotide from each allele forms the genotype for that specific locus. If the paternal allele has a guanine (G) nucleotide at a specific position and the maternal allele has a thymine (T) nucleotide at the same position, the resulting genotype would be referred to as GT. This creates three possible genotypes at each locus, the two homozygous genotypes GG or TT and the heterozygous GT genotype. Allele frequency is determined by counting the number of each allele in a population and dividing by the total number of alleles. If the frequency is less than 0.5, that allele is

Addiction Pharmacogenetics

441

called the minor allele, with the major allele having a frequency greater than 0.5. This means that there will always be more participants with the homozygous major/major allele genotype than in rarer minor/minor allele genotype. Because of this difference, studies with small subject cohorts will often group the homozygous minor/minor participants with the heterozygous major/minor group for comparison against the homozygous major/major allele genotype group. This group of minor/minor allele and major/ minor allele genotypes is referred to as the “minor allele carrier” group, consisting of individuals with either one or two copies of the minor allele. In the genome, there are approximately 21,000 proteinencoding genes (~2% of genome), 9600 long noncoding RNA “genes” (>200 nucleotides long), 8800 small RNAs “genes,” and 11,224 pseudogenes, which previously had been considered to be remnants of “dead” genes but, in fact, may be active in specific cells [26, 27]. These regions account for approximately 80% of the genome. Over 38 million single nucleotide polymorphisms (SNPs), 1.4 million short insertions and deletions, and more than 14,000 large structural variants have been mapped [28]. The average person has from 250 to 300 loss-of-function variants in genes that have been annotated [29]. Of these, 50–100 previously have been found to be involved in the development of inherited diseases. In addition, in every person, various regions of the genome may be deleted or duplicated. These regions can range from one to millions of nucleotides. It is estimated that approximately 0.4% of the genome is different between any two unrelated individuals with respect to copy number, that is, the number of repeated or deleted regions. Additionally, functional variation also may be introduced by mRNA splicing or via epigenetic modification processes, such as DNA methylation or histone modifications. The majority of these types of functional variation are based on the differential codification of genes and not on the quantity of genes [26]. Genetic variation can regulate how a gene is expressed at several levels, including transcriptional regulation, mRNA splicing and stability, and protein translation, stability, and function (such as enzymatic activity or binding affinity). A variant as small as a single nucleotide can affect gene expression by changing the regulation of transcription or altering the splicing and stability of mRNA. Genetic variation also may lead to downstream effects by changing the amino acid sequence of the translated protein, which in turn could affect folding and structure of the protein or affect biological processes such as enzymatic activity, binding affinity, or stability. Variants (polymorphisms) in gene coding for components of pathways involved in substance use disorders may be responsible for the variation found a patient’s response to pharmacotherapy for an addiction. These responses are likely to be dependent upon many genes (e.g., polygenic) but also may be oligogenic, where only a few

442

David P. Graham et al.

genes play a major role. Data from studies on the genetic basis of response to pharmacotherapy may allow the effective tailoring of therapy to the needs of the individual based on his or her genetic makeup, as exemplified in pain management [23]. Overall, optimistic arguments remain justified regarding the transformational potential of genomics on medicine despite the identification of associated challenges [30]. Genome-wide association studies (GWAS) have shown many complex disorders are polygenic [31]. Rather than focusing on individual mutations, polygenic risk scores (PRS) represent a weighted count of the possible thousands to hundreds of thousands of genetic risk variants an individual may carry [32]. PRS use the variants and weights identified in GWAS studies. Within an individual, the number of risk alleles for each variant (0, 1, or 2) are summed and then weighted by the variants effect size (log odds ratio for binary traits, beta coefficients for continuous traits) resulting in a single outcome score [31]. While it is unlikely PRSs will be fully diagnostic given the PRS only accounts for a portion of a disease’s genetic profile, it is likely that PRSs will be helpful in risk stratification, prognosis, and treatment response prediction [33]. The clinical utility of PRS still remains to be established as the discriminative ability is low in the general population, but PRSs may be differentially informative at different points in the target disease trajectory: (1) risk prediction from birth, (2) early symptoms/prodromal phase, (3) support of diagnosis, (4) treatment decision-making, or (5) prediction of disease course or outcome [31]. In psychiatric disorders, PRSs predominately to date have been evaluated in depression, schizophrenia, and Alzheimer’s disease [31]. There has been very little work published on attention deficit hyperactivity disorder [34, 35] or generalized anxiety disorder [36]. Regarding addictions, work remains in the fledgling stages for smoking [37], cocaine use [38], and cannabis use [39]. For a thorough review of the genetic epidemiology of SUD, see Prom-Wormley et al. [16].

4

Drugs of Abuse and Dopamine System Central to the development of addiction is the reward pathway of the brain, which is mediated by the catecholamine neurotransmitter dopamine (DA) and involves “liking” or the hedonic impact on the limbic system and “wanting” or incentive salience making cues attractive and triggering cravings [40]. Natural reinforcers, such as copulation and food, produce a surge of dopamine release in the nucleus accumbens (NAc) of the brain from neurons originating in the ventral tegmental area (VTA) [41, 42]. Most drugs of abuse act either directly or indirectly on the midbrain dopaminergic reward system, increasing levels of dopamine in the NAc [43, 44]. In the

Addiction Pharmacogenetics

443

long term, drugs of abuse hijack the reward system so that the individual requires the drug of abuse for activation of their reward system [45], with distinct circuits mediating rewards and aversion [46]. For an overview of the behavioral and neural circuitry in the context of commonly used substances, please refer to Crummy and colleagues [47]. Cocaine’s mechanism of action is achieved by blockage of the dopamine, norepinephrine, and serotonin transporters [47, 48], thereby reducing the reuptake of dopamine, norepinephrine, and serotonin, respectively, from the synapse. The catecholamine neurotransmitter dopamine facilitates the reward pathway in the brain [49]. It is not clear how alcohol increases dopamine levels, but it is believed to be through μ-opioid receptors in the mesolimbic system [47, 50, 51]. Alcohol most likely produces its reinforcing effects by stimulating the release of endogenous opioids that increase extracellular mesolimbic dopamine levels in the NAc [47, 52–55]. Binding of the endogenous opioid peptide β-endorphin disinhibits the GABAergic interneurons in the VTA promoting the release of dopamine in the NAc [44, 47, 56]. Thus, both opioid receptor and dopamine receptor antagonists play critical roles in the investigation of moderating alcohol craving and stimulation. Opioids, on the other hand, directly bind to the μ-opioid receptors, causing disinhibition of the GABAergic interneurons and the subsequent release of dopamine [47, 57]. Most drugs of abuse act on the dopaminergic system to simulate these natural reinforcers and subsequently increase dopamine levels in the NAc [43, 47] and play an important role in the development of craving and the loss of control over use of these substances [58]. As drug dependence develops, the abused drug hijacks the reward system making its use necessary to activate the reward pathway at all [45, 46]. Reward is a complicated process and as such has been separated into five phases: (1) anticipation, (2) evaluation of cost and benefits of expected reward, (3) actions obtaining reward, (4) pleasure in response to reward, and (5) reward learning [49]. Recent research has identified that dopaminergic drugs have different effects on different phases of reward, the relationship between dopamine and reward is not likely linear, and the ability to detect the effects of dopaminergic drugs depends upon which measures are used (subjective, behavioral, imaging) [49]. In this chapter, we will discuss the genetics of several pharmacotherapies for the addictions. Some of these are FDA-approved therapies, while others are under investigation. The pharmacotherapies we will explore are listed in Table 1, and the genetic variants are listed in Table 2.

444

David P. Graham et al.

Table 1 Drug therapies discussed in pharmacogenetic studies Pharmacotherapy Addiction

Notes

Acamprosate

Alcohol

Acamprosate is a drug with a chemical structure similar to that of γ-aminobutyric acid (GABA) and acts as a partial agonist of N-methyl-D-aspartate (NMDA) receptors in the brain. Acamprosate has been used to treat alcoholism since 1989 in Europe and since 2004 in the United States. Alcohol exposure is thought to depress glutamatergic signaling, which then rebounds after the cessation of alcohol use resulting in hyperstimulation. Acamprosate’s effects include increasing taurine as well as in binding NMDA receptors, therefore inhibiting the excitatory effects of alcohol withdrawal [343].

Baclofen

Alcohol

Baclofen, β-(4-chlorophenyl)-γ-aminobutyric acid, currently is the only FDA-approved GABAB agonist, typically used to treat spasticity. Off label, baclofen has been used for the treatment of alcohol abuse and drug addiction [344].

Bromocriptine

Alcohol

Bromocriptine is an ergot alkaloid and a dopamine receptor D2 agonist that inhibits prolactin release from the pituitary gland [345]. Bromocriptine is typically used to treat Parkinsonian syndrome as well as hyperprolactinemia and growth hormone- and prolactin-related disorders such as menstrual disorders, infertility, and hypogonadism.

Buprenorphine

Opioid

Buprenorphine is a synthetic μ-opioid partial agonist synthesized in 1967 [346] and initially utilized as an analgesic [347]. It was not used as a maintenance treatment for opioid addiction until the mid-1980s. Studies have showed that buprenorphine’s effects were longer acting and that it had a lower potential for abuse than did morphine [348]. Suboxone is a combination of buprenorphine plus naloxone formulated to prevent misuse. For a 2020 review of the pharmacogenomics of buprenorphine, see Seguı´ et al. [349].

Cocaine vaccine

Cocaine

The cocaine vaccine consists of a cocaine derivative conjugated to cholera toxin [139]. Following a series of vaccinations with the vaccine, participants produce anti-cocaine antibodies. The hypothesis is that immunization of treatment-seeking patients vaccinated with this vaccine will stimulate the production of anticocaine antibodies. When participants who are abstinent relapse and take cocaine, the anti-cocaine antibodies will sequester the cocaine in the blood, thereby preventing a rapid surge of cocaine into the brain.

Disulfiram

Alcohol and cocaine

Disulfiram was initially synthesized as a reagent to vulcanize rubber. Disulfiram is approved for use in treating alcoholism [350]. A major metabolite of disulfiram is diethyldithiocarbamate, a copper chelator. The aldehyde dehydrogenase isozyme ALDH2 is an enzyme active in the ethanol metabolic pathway that uses copper as a cofactor and is inhibited by disulfiram treatment [351]. When inhibited by disulfiram, the reduction in aldehyde dehydrogenase activity causes the accumulation of acetaldehyde, thus inducing nausea, vertigo, flushing, and other unpleasant effects after the (continued)

Addiction Pharmacogenetics

445

Table 1 (continued) Pharmacotherapy Addiction

Notes consumption of alcohol. Disulfiram also inhibits other coppercontaining enzymes including dopamine β-hydroxylase (DβH), which is the enzyme that converts dopamine to norepinephrine. Therefore, disulfiram treatment increases dopamine levels due to an inhibition of DβH activity, with a concomitant decrease in central and peripheral norepinephrine levels [350–353]. In addition to treating alcoholism, disulfiram has been shown to reduce cocaine use as well [354–356]. Its action may be through the inhibition of DβH as well as by the inhibition of plasma and microsomal carboxylesterases and plasma cholinesterase, enzymes that inactivate cocaine systemically [357]. In addition, disulfiram reduced subjective measures of IV-administered, cocaineproduced craving [358, 359]. In SLC6A3 10,10-repeat genotype, compared to the 9-repeat carrier group, disulfiram reduced the number of cocaine-positive urines [90]. Studies have found that in treatment of individuals with disulfiram who were dependent both on cocaine and alcohol, the use of both illicit drugs were reduced [360, 361].

Doxazosin

Cocaine

Doxazosin is a competitive α1-adrenergic postsynaptic receptor antagonist. It has FDA approval for use in benign prostatic hyperplasia and in hypertension and has been used off-label in the treatment of ureteral stones and PTSD-associated nightmares [362]. Doxazosin appears more effective in reducing cocaine usage in those persons with genetically lower DβH levels (rs1611115 CT/TT as compared to CC genotype) [92].

Gabapentin

Alcohol

Gabapentin is a structural analog to gamma-aminobutyric acid (GABA) and has been approved for use in partial seizures in patients 12 years old and higher and in mixed seizure disorders and refractory partial seizures in children. It is also used off-label for different types of neuropathic pain including chemotherapeuticinduced pain [363, 364], trigeminal neuralgia [365], diabetic neuropathy [366, 367], and postherpetic neuralgia [368, 369]. Gabapentin acts in both the central nervous system (spinal and supraspinal areas) and in peripheral regions (DRG neurons). Gabapentin’s primary molecular target is the accessory α2δ-1 subunit of voltage-gated calcium channels. Gabapentin also has been found to decrease GABAergic activity in the locus coeruleus, leading to an increased noradrenaline release [370]. A targeted review of gabapentin’s use in the management of alcohol withdrawal and dependence noted limited data suggesting benefit in mild alcohol dependence [371].

Methadone

Opioid

Methadone was synthesized in the late 1937 by Bockmu¨hl and Ehrhart [342] and first utilized experimentally to relieve opioid withdrawal in 1948 [372]. It is a synthetic μ-opioid receptor agonist that binds with high affinity, reduces opioid cravings, and can block the binding of other superimposed opioids [101]. Methadone is a synthetic opioid that is used in the (continued)

446

David P. Graham et al.

Table 1 (continued) Pharmacotherapy Addiction

Notes pharmacotherapy of the addition to short-acting opiates such as heroin [373]. Due to its long half-life of approximately 22 h and efficacy, methadone is also used in the management of chronic pain. Methadone is a μ-opioid receptor agonist and a weak NMDA receptor antagonist. It is effective at reducing opioid withdrawal symptoms and at blocking the euphoric effects of heroin, morphine, and other opioids. Methadone is a racemic mixture of levomethadone and dextromethadone. Levomethadone is the selective μ-opioid receptor agonist, while dextromethadone is a glutamatergic NMDA receptor antagonist [100]. Since glutamate is an excitatory neurotransmitter, NMDA antagonism may be partly involved in methadone’s efficacy.

Naltrexone

Alcohol and opioid

Naltrexone is a μ-opioid receptor antagonist first synthesized in the 1960s. Naltrexone was approved by the FDA for the treatment of opioid addiction treatment in 1984 and alcohol addiction in 1994 [374]. Naltrexone blocks the euphoric effects of opioids by binding competitively to opioid receptors but does little to curb craving for opioids. Because naltrexone is an opioid antagonist, there is little risk of abuse or dependence given that it does not have intrinsic opiate effects and therefore isn’t reinforcing [375].

Olanzapine

Alcohol

Olanzapine is a second-generation antipsychotic used to treat schizophrenia and mania related to bipolar disorder. Olanzapine binds neurotransmitter receptors of several classes including dopaminergic, adrenergic, and serotonergic receptors [376, 377].

Ondansetron

Alcohol

Ondansetron is a serotonin 5-HT3 receptor antagonist, with low affinity for α1-adrenergic, 5-HT1B, 5-HT1C, and μ-opioid receptors [378]. It is used primarily to treat nausea and vomiting (antiemetic) following chemotherapy.

Tiapride

Alcohol

Tiapride is a dopamine receptor D2 and D3 antagonist. It is used to treat alcohol withdrawal syndrome where it has anxiolytic effects. Tiapride has been shown to reduce psychological stress, decrease drinking, and improve reintegration into society [379].

Topiramate

Alcohol

Topiramate is an anticonvulsant [380], a migraine prophylactic [381], and a weight loss medication when in combination with phentermine [382]. Topiramate modulates synaptic transmission and neuronal excitability [383].

5

Dopaminergic System Genes

5.1 Dopamine Receptor D2 (DRD2)/ Ankyrin Repeat and Kinase DomainContaining 1 (ANKK1) Genes

The DRD2 and ANKK1 genes are located approximately ten thousand nucleotides apart on chromosome 11q22–23. Variants in both genes have been found to be associated with several psychiatric diseases such as schizophrenia, as well as with substance abuse disorders, including alcohol, heroin, nicotine, cocaine, opioid, gambling, methamphetamine, and polysubstance addiction [59– 66].

Addiction Pharmacogenetics

447

Table 2 Genetic variants involved in pharmacotherapy for the addictions Gene

Product

Variant

Addiction

Pharmacotherapy

ABCB1

ATP-binding cassette, subfamily B, member 1

rs1045642, rs1128503, and rs2032582

Opioids

Methadone

ADRA1A α1A-adrenoceptor

rs1048101

Cocaine

Disulfiram

ANKK1

Ankyrin repeat and kinase domain containing 1

rs1800497 (TaqIA)

Cocaine, alcohol, and opioids

Disulfiram, naltrexone, and methadone

ARRB2

Arrestin β-2

rs2036657, rs3786047, and rs1045280

Opioids

Methadone

BDNF

Brain-derived neurotrophic rs988748, factor rs1967554, rs2030324, rs2239622, rs7127507, rs11030118, and rs11030119

Opioids

Methadone

COMT

Catechol-Omethyltransferase

rs4680 (Val158Met) Methamphetamine Modafinil

CYP2B6

Cytochrome P450, family 2, subfamily B, polypeptide 6

rs2279343 and rs3745274

Opioids

Methadone

CYP2D6

Cytochrome P450, family 2, subfamily D, polypeptide 6

Multiple (see [384]) Opioids

Methadone

DBH

Dopamine β-hydroxylase

rs1611115 (C-1021T)

Cocaine

Disulfiram and cocaine vaccine

DRD2

Dopamine receptor D2

rs6277, rs6275, and rs1799978

Cocaine, alcohol, and opioids

Acamprosate, bromocriptine, disulfiram, and methadone

DRD4

Dopamine receptor D4

Exon 3 VNTR

Alcohol

Olanzapine

GABBR1

γ-aminobutyric acid β-1

rs29220

Alcohol

Baclofen

GABRB2

γ-aminobutyric acid β-2

rs3219151 (C +1412T)

Alcohol

Acamprosate and naltrexone

GABRA6

γ-aminobutyric acid α-6

rs3219151 (T +1519C)

Alcohol

Acamprosate and naltrexone

GATA4

GATA-binding protein 4

rs13273672

Alcohol

Acamprosate

GRIK1

Glutamate ionotropic receptor kainate type subunit 1

rs2832407

Alcohol

Topiramate

(continued)

448

David P. Graham et al.

Table 2 (continued) Gene

Product

Variant

Addiction

Pharmacotherapy

KCNJ6

Potassium inwardly rectifying, channel subfamily J, member 6

rs2070995

Opioids

Methadone

MTHFR

Methylenetetrahydrofolate rs1801133 (C677T) Cocaine reductase

Disulfiram

MYOCD

Myocardin

rs1714984

Opioids

Methadone

NECTIN4 Nectin cell adhesion molecule 4

rs3820097, rs4656978, and rs11265549

Opioids

Methadone

OPRD1

δ-opioid receptor

rs529520 rs581111 rs678849

Cocaine Cocaine Opioids

Buprenorphine Buprenorphine Methadone, suboxone, and buprenorphine

OPRK1

κ-opioid receptor

rs6473797

Cocaine

Cocaine vaccine

OPRM1

μ-opioid receptor

Opioids rs558025 rs1799971 (A118G) Alcohol Opioids rs2075572

Methadone Naltrexone Methadone

SLC6A3

Dopamine transporter

rs28363170

Cocaine

Disulfiram

SLC6A4

Serotonin transporter

5-HTTLPR VNTR

Alcohol

Ondansetron and sertraline

TPH2

Tryptophan hydroxylase 2

rs4290270

Cocaine

Disulfiram

The DRD2 gene encodes the G-protein-coupled dopamine receptor D2, which is central to dopaminergic signaling in the brain. This gene can be alternatively spliced to produce two protein isoforms, designated the long and short forms of the receptor protein (D2L and D2S, respectively). The variant rs2283265 is a G!T transversion in an intron of the DRD2 gene. The T-allele of this variant has been shown to alter the ratio of D2L to D2S protein isoforms and is overrepresented in cocaine-addicted populations [67, 68]. The DRD2 variant rs6277 is a synonymous (does not alter the amino acid coding) C to T transition in DRD2. The T-allele has been shown to be associated with enhanced D2 receptor availability, altered mRNA folding, and reduced mRNA stability [69, 70]. Several variants in DRD2 have been associated with psychiatric disorders such as schizophrenia and alcoholism [71– 74], and a multi-locus genetic composite (MGC) score including at least six key polymorphisms (rs7118900, rs1554929, rs907094, rs12364283, rs6278, and rs107656) revealed reward-related

Addiction Pharmacogenetics

449

activation in the ventral striatum, and ventral tegmental area was modulated by the MGC and sex [75]. Specifically, in those with a low MGC (representing lower dopamine transmission), only females displayed a downregulation of mesolimbic system activation during reward rejection. The ANKK1 gene (also known as receptor interacting protein 5 or RIP5) encodes the ankyrin repeat and kinase domain-containing 1 protein. The RIP serine/threonine kinase family is involved in activation of various cellular signaling pathways, including NF-KB, JNK, and apoptotic signaling [76]. The ANKK1/DRD2 TaqIA variant, also known as rs1800497, is a functional variant located in the final exon of ANKK1 that codes for a non-synonymous Glu!Lys (C!T) amino acid change in the C-terminus of the ANKK1 protein. Initially associated with alcohol addiction in 1990 by Blum et al. (1990), the ANKK1/DRD2 TaqIA variant has been one of the most examined variants with regard to substance addiction [77]. The T (TaqIA1) allele of ANKK1/DRD2 also has been found to be associated with reduced dopamine receptor D2 density [78] and with reduced opioid receptor binding [79]. The brains of cocaine-, opioid-, and alcohol-addicted individuals have shown reduced D2 receptor availability, therefore providing a potential mechanism through which the gene variant may affect addictive behavior [80–83]. Additionally, it has been suggested the ANKK1 Taq1A gene and its protein product may be a variant involved in altering both brain structure and dopaminergic function, increasing an individual’s risk to developing addiction [84]. However, a recent meta-analysis reassessed the proposed association of DRD2/ANKK1 rs1800497 with alcohol use disorder and determined the association was attributable to low-control allele frequencies and not to allele function, giving rise to an odds ratio of 1.23 (95% CI: 1.14–1.31) [85]. Further, rs1800497 also has been noted to be associated with higher scores on the StateTrait Anxiety Inventory (STAI) as well as the NEO Five-Factor Inventory scales of neuroticism, extraversion, and agreeability [86]. Another variant in ANKK1 is rs7118900 that codes for an alanine to threonine (Ala239Thr) substitution creating a predicted phosphorylation site and is found to be in strong linkage disequilibrium (LD) with ANKK1/DRD2 TaqIA [87]. Cells transfected with the ANKK1 rs7118900 Thr239 variant constructs tagged with green fluorescent protein (GFP) expressed greater levels than did constructs containing the Ala239 variant. The Thr239 constructs decreased expression when treated with the dopamine agonist apomorphine, while the Ala239 constructs increased expression. This finding provides a potential functional link for the ANKK1 gene product to the dopaminergic system. Disulfiram (Table 1) was tested in a cohort of cocaine- and opioid-codependent individuals as a pharmacotherapy for cocaine addiction in a placebo-controlled clinical trial [88]. Patients were

450

David P. Graham et al.

stabilized on methadone, and the treatment group received 250 mg of disulfiram daily. Disulfiram pharmacotherapy decreased cocaine use as measured by urine cocaine metabolites, with no concomitant difference in the amount of opioid use. Genetic analysis revealed that ANKK1 rs1800497 T-allele (TaqIA1) carriers showed reduced cocaine-positive urines during pharmacotherapy with disulfiram, while CC homozygous individuals, those carrying two copies of the same allele, showed no treatment response [89]. Additionally, DRD2 rs2283265 T-allele carriers showed a large reduction in cocaine-positive urines with disulfiram pharmacotherapy, while GG homozygous individuals displayed less reduction. Disulfiram also has been examined in relation to the dopamine transporter SLC6A3 (rs28363107) regarding cocaine dependence [90]. After stabilization on methadone for 2 weeks, cocaine- and opioid-codependent patients were randomized into disulfiram or placebo groups for 12 weeks of treatment. Among the 10,10-repeat genotype group, the cocaine-positive urines dropped from 78% to 48% and from 80% to 75% among the 9-repeat carrier group. No differences were identified in the placebo group. The conclusion was that patients with genetically higher levels of DAT had better outcomes with disulfiram therapy when compared to those persons with lower DAT levels. The same group of patients were tested for influence upon the δ-opioid receptor (OPRD1) rs678849 variant. Patients with the rs678849 CC genotype was associated with a drop in cocaine-positive urines compared to T-allele carriers, but neither groups showed any differences in opioid-positive urines [91]. Doxazosin (Table 1), an α1-adrenergic antagonist, was examined in a cohort of 76 cocaine use-dependent patients regarding the dopamine β-hydroxylase (DBH) (C-1021T, rs1611115) polymorphism. The T-allele is associated with reduced conversion by DβH of dopamine to norepinephrine. The findings included lower rates of cocaine use among participants with a T-allele (CT/TT, lower DβH levels) than those with the CC genotype, with the T-allele carriers showing a 41% decrease in positive urines compared to no net reduction in those with the CC genotype. These results suggest that the DBH polymorphism rs1611115 may play an important role in cocaine use-dependent patient’s response to doxazosin treatment [92]. Neuroendocrine studies have suggested that alcoholics have reduced DRD2 receptor sensitivity after months to years of abstinence from alcohol [93–95]. This may be reflective of why the dopamine receptor agonists bromocriptine and the D2 antagonist tiapride (Table 1) are efficacious in the treatment of alcoholism and alcohol withdrawal syndrome, respectively. The dopamine receptor D2 agonist bromocriptine has been shown to decrease alcohol craving and anxiety in TaqIA1-carrier alcoholics [96]. Lucht et al. (2001) showed that the AA genotype of rs71653615 in DRD2 was

Addiction Pharmacogenetics

451

found to be associated with higher doses of the selective dopamine receptor D2 antagonist tiapride that was required to treat alcohol withdrawal symptoms of alcohol-dependent Caucasians [97]. The pharmacogenetics of acamprosate, an NMDA receptor partial agonist, and naltrexone, a μ-opioid receptor antagonist (Table 1), were evaluated in a cohort of Dutch alcoholics [98]. Acamprosate was shown to have a greater effect on cue-induced craving than did naltrexone in alcoholic participants homozygous for the ANKK1/DRD2 rs1800497 TaqIA1 allele. However, naltrexone was more effective in participants homozygous for the TaqIA2 allele. In heterozygous participants (those carrying one copy of each allele), naltrexone and acamprosate were equally effective. The greater effectiveness of naltrexone in the TaqIA2 homozygous participants may be related to the finding that individuals homozygous for the TaqIA2 allele exhibited higher [3H]naloxone binding in the caudate nucleus indicative of having greater opioid receptor density [79, 99]. A common and well-established opioid addiction treatment is the methadone maintenance therapy program or MMTP. Methadone, a mixture of levomethadone, a selective μ-opioid receptor agonist, and dextromethadone, a glutamatergic NMDA receptor antagonist [100] is an effective therapy for opioid addiction [101, 102]. In methadone maintenance therapy for opioid addiction, carriers of the DRD2 rs6275 T-allele were found to require higher methadone doses than did noncarriers and required longer periods of time to reach maximum methadone maintenance dose [103]. Another study illustrated that participants carrying at least one copy of the ANKK1/ DRD2 rs1800497 TaqIA1 allele in the ANKK1 gene were more likely to be in the methadone maintenance treatment “poor treatment” outcome group, which was composed of individuals who withdrew from the study or who continued use of heroin at least once weekly, compared to being in the “successful treatment” outcome group [66]. Additionally, TaqIA1 allele carrier participants used twice the amount of heroin in the previous year than did A2 homozygous participants. Similarly, participants with the DRD2 rs6277 CC genotype were more likely to be nonresponders to methadone maintenance therapy, and their duration of opioid-free urines were shorter than did participants who were T-allele carriers but with no difference in the frequency of the ANKK1/DRD2 TaqIA variant with these measures [104, 105]. Hung et al. (2011) examined a different DRD2 variant in Han Chinese participants and found that participants carrying the rs1799978 G (-214A>G) allele required a lower methadone dose than did noncarriers [105, 106]. In a Malaysian population, multiple DRD2 polymorphisms have not been found informative regarding pain responses, opioid withdrawal severity, or sleep quality in persons on methadone maintenance therapy [107].

452

David P. Graham et al.

5.2 Dopamine Receptor D4 (DRD4) Gene

The DRD4 gene has a variable number tandem repeat (VNTR) located in exon 3, with the common alleles of two, four, and seven repeats [108]. The seven repeat appears to be the critical allele relevant to pharmacotherapy and alcohol consumption, as well as to behavioral responses to the food retail environment [109]. Functionally, the DRD4 seven tandem repeat allele codes for a dopamine receptor D4 that blunts intracellular forskolin-stimulated cyclic AMP (cAMP) response to dopamine relative to the receptor encoded by the two or four tandem repeat alleles [110]. The exon 3 DRD4 VNTR has been shown to moderate the influence of alcohol on craving and related responses. Participants who were carriers of the seven or longer repeat had a greater “urge to drink” and a lower “subjective high” following alcohol consumption than did those without this allele [111]. Further, a recent meta-analysis identified that carriers of the long alleles had an increased number of drinking days, binge drinking days, and severity of alcohol use disorder than did persons homozygous for the short allele [112]. Similarly, smoking status was noted to be associated with carriers of the seven or longer repeat as well [113]. Hence, the exon 3 DRD4 VNTR long allele carrier may be a contributing factor in the neurobiological mechanisms underlying drug abuse, eating disorders, and related comorbid conditions [114]. Although both genotype groups had similar decreases in craving at baseline for alcohol following olanzapine treatment, olanzapine only reduced craving after exposure to alcohol in participants carrying a 7 or longer VNTR allele [115]. However, unlike the μ-opioid receptor antagonist naltrexone (see below), olanzapine had little to no effect on moderating alcohol’s reinforcing effects [116].

5.3 Dopamine βHydroxylase (DBH) Gene

Dopamine β-hydroxylase (DβH) is the enzyme that metabolizes dopamine into norepinephrine (reviewed in [117, 118]). Norepinephrine modulates many behavioral, cognitive, and physiological functions [119–123]. The catecholamine neurotransmitters dopamine and norepinephrine are stored in synaptic vesicles prior to release from the cell. It is within these vesicles that DβH is localized. Although most of the DβH is bound to the membrane of the vesicles, some DβH is free and is co-released with the catecholamines during synaptic transmission from neurons and into the blood from the neurosecretory cells of the adrenal medulla [124]. The levels found in serum is highly correlated between sibs but varies between unrelated participants [125]. This variation has been found to be heritable in family and twin studies in both serum and CSF [126]. Although a number of polymorphisms have been studied in the DBH gene and have been found to be associated with DβH levels [127–130], one appears to be the primary functional variant [131– 133]. In a study examining 11 variants spanning the DBH locus,

Addiction Pharmacogenetics

453

the C-1021T (rs1611115) variant was found to be most highly associated with DβH plasma levels [131]. The C-1021T variant has been shown to be associated with decreased enzyme activity in plasma in several populations, including European Americans, African Americans, Japanese, and an Eastern Indian population, with the T-allele being associated with decreased activity in all populations examined [132, 133]. This polymorphism accounted for 35–52% of the variation in DβH levels. The rs1611115 variant has been shown to be associated with alcohol dependence in females [134], as well as with both progression of heroin selfadministration [135] and memory impairment after heroin dependence [136], and may be associated with Alzheimer’s disease [137] but not Parkinson’s disease [138]. In the aforementioned trial of disulfiram for the reduction of cocaine use with cocaine- and methadone-codependent participants, the role of this DBH variant was studied [88]. When the disulfiram treatment group was stratified by DBH rs1611115 genotype, the CC genotype group with normal DβH expression reduced their cocaine use when treated with disulfiram, while those patients who were carriers of the low-expressing T-allele did not. It was suggested that the reduction in norepinephrine neurotransmission by disulfiram may not reduce the use of cocaine in those with the low DβH levels, which may have caused an upregulation of dopamine receptors. Vaccines have been tested as a potential pharmacotherapy for cocaine addiction. A cocaine vaccine (Table 1) of succinyl norcocaine conjugated to the cholera toxin (TA-CD) was administered to cocaine-dependent participants [139]. Individuals who produced adequate antibody levels showed reduced cocaine use. Genetic analysis showed that T-allele carriers of the rs1611115 variant (that has been shown to be associated with low DβH expression) reduced cocaine use with the vaccine, while participants with the CC genotype did not [140]. This may be related to the increase incidence of paranoia while using cocaine in participants with the low-expressing DBH genotype [130]. 5.4 Dopamine Transporter (SLC6A3) Gene

As noted above in Sect. 5.1, disulfiram has been examined in relation to the dopamine transporter SLC6A3 variant rs28363107 in relation to cocaine dependence [90]. After stabilization on methadone for 2 weeks, cocaine- and opioid-codependent patients were randomized into disulfiram or placebo groups for 12 weeks of treatment. Among the 10,10-repeat genotype group, the cocainepositive urines dropped from 78% to 48% and from 80% to 75% among the 9-repeat carrier group. No differences were identified in the placebo group. The conclusion was that patients with genetically higher levels of DAT had better outcomes with disulfiram therapy when compared to those persons with lower DAT levels.

454

6

David P. Graham et al.

Opioidergic System Genes

6.1 μ-Opioid Receptor (OPRM1) Gene

The G protein-coupled μ-opioid receptor (MOP-r) mediates most opioid antagonists and is the receptor for morphine, methadone, and endogenous opioids, such as endomorphin and β-endorphin. The opioid receptor family functions in nociception but also indirectly in the mesolimbic dopaminergic pathway, partly mediating the addictive process [141]. Opioids bind MOP-r and attenuate GABAergic inhibition of dopaminergic neurons [57], producing an increase in dopamine release at nerve terminals in the ventral striatum and medial prefrontal cortex by VTA neurons [142, 143]. OPRM1, the gene encoding MOP-r, is located on chromosome 6q25.2 and contains several functional variants that have been studied extensively in association with not only substance abuse and dependence [100, 144, 145] but also a variety of conditions, such as major depression, schizophrenia, and pain sensitivity (reviewed in [12]). The most studied variant in OPRM1, rs1799971, is located in the coding region at nucleotide 118. At this location, there is an A to G non-synonymous transition that codes for an aspartic acid (Asp) instead of an asparagine (Asn) at position 40 in the N-terminus of the receptor [146, 147] and also is referred to as the A118G variant. This substitution removes one of five highly conserved putative N-glycosylation sites from the N-terminal extracellular domain of the receptor. Approximately 30% of Europeans and 60% of Asians carry one or two copies of this allele (http://www.ncbi.nlm.nih.gov/snp). However, this variant is nearly absent in African American (AA) individuals (http:// www.ncbi.nlm.nih.gov/snp). At the molecular level, the 118G allele of OPRM1 rs1799971 encodes a receptor that binds the endogenous opioid peptide β-endorphin with three times the affinity than does the variant receptor encoded by the 118A allele [147]. However, the G allele leads to reduced mRNA and protein levels resulting in a net functional loss of OPRM1 gene expression. In postmortem autopsy brain tissue of 118A/118G heterozygous individuals, the Asn40 mRNA encoded by the118A allele was about 1.5 times more prevalent than was the Asp40 mRNA encoded by the 118G allele [148]. In vitro cellular expression assays have shown that the 118G receptor allele produced lower cell-surface binding site availability than did the 118A receptor allele [149]. G-allele carriers of OPRM1 have increased hypothalamic-pituitary-adrenal (HPA) axis response relative to those homozygous for the A allele under opioid receptor antagonism [150], enhanced cortisol response, and a reduced agonist effect of morphine-6-glucoronide [151]. Genetically homozygous (118A/118A or 118G/118G)-induced inhibitory neuronal cells from EA have shown that the 118G/118G cells displayed stronger suppression (versus N40) of spontaneous

Addiction Pharmacogenetics

455

inhibitory postsynaptic currents (sIPSCs) compared to 118A/ 118A cells [152]. When buprenorphine (a partial opioid agonist buprenorphine)-maintained heroin-dependent patients were challenged with metyrapone, a glucocorticoid synthesis inhibitor, the 118G carrier patients had a reduced HPA axis response compared those with the 118A/118A genotype, normalizing the hyperactive HPA axis in opioid dependence [153]. In a sample of Swedish participants, the odds ratio of being alcohol dependent was twofold greater in OPRM1 118G-allele carriers [154]; however, this finding was not replicated in two recent reports using European populations [155, 156]. Studies on the subjective effects of alcohol have shown that OPRM1 118G-allele carriers experience greater subjective feelings of intoxication, euphoria, and sensitivity to both the reinforcing and sedative effects of alcohol and had a threefold increase in family history of alcoholism [157]. OPRM1 118A-allele carriers have enhanced euphoria, feelings of intoxication, and the rewarding effects of alcohol compared with individuals who are homozygous for the 118G-allele experience [157]. The OPRM1 118G variant has been found to be associated with heroin addiction in Asians [158]. In addition, the rare OPRM1 G allele rs62638690 is associated with heroin and cocaine addiction in a cohort of European Americans (EA) [159], not in AA [160]. Hence, both common and rare genetic variants may influence susceptibility to drug addiction [161]. Neonatal abstinence syndrome, NAS, occasionally occurs in neonates born of opioid-dependent mothers. Neonates who were born to opioid-dependent mothers and had NAS and who carried one or two 118G alleles had less severe withdrawal than those with the AA genotype [162, 163]. In fact, NAS infants of G-allele carrier mothers had a reduced need for treatment. In a study that examined 123 variants in genes involved in psychiatric disorders, an association was found between the rs558025 variant in the 30 flanking region of OPRM1 and MMTP treatment response [164]. The methadone dose required to maintain abstinence in the G-allele carriers was 132 mg, while in TT participants, it was 147 mg. Additionally, it has been observed that OPRM1 rs1799971 G-allele and rs2075572 C-allele carriers required higher methadone doses for the treatment of heroin addiction than those without these alleles [165]. Similar results were found with a variant in the 30 untranslated region of OPRM1 [166]. However, these associations did not survive corrects for the testing of multiple variants. The association of methadone dose requirements and several genes, including OPRM1 in opioid-dependent participants patients [106], found that the A118G variant along with variants in genes coding for the protein kinase PKK2, the dopamine receptor 2, and cytochrome P450 were associated with requiring higher methadone doses.

456

David P. Graham et al.

Buprenorphine is another treatment regime for opioid dependence. It is a partial agonist of the μ-opioid (MOP-r) and δ-opioid (DOP-r) receptors, an antagonist of the κ-opioid (KOP-r) receptor, and an agonist of the nociception opioid receptor-like (NOP) receptor [167]. In a cellular assay utilizing stably transfected OPRM1 constructs, buprenorphine inhibited adenylyl cyclase and activated ERK1/2 less effectively in CHO cells expressing the 118G MOP-r than cells transfected with the 118A construct [153]. This reduced signaling of the 118G MOP-r may partially explain the reduced efficacy of buprenorphine in the treatment of opioid dependence in patients carrying the 118G allele [168]. Alcohol craving develops through repeated alcohol administration and intensifies over time. Alcohol cues, such as the sight and smell of alcohol and the consumption of small priming doses of alcohol, elicit craving. Craving, in part, acts by prompting dopamine release as an incentive to continue drinking. A functional neuroimaging study found a greater hemodynamic (blood flow) response in mesocorticolimbic structures, including the VTA, following alcohol tastes in alcoholics compared to healthy volunteers [169]. A positron emission tomography (PET) study found that OPRM1 118G-allele carriers had stronger striatal dopamine response to intravenous alcohol administration compared to 118A-allele homozygous participants as measured by the displacement of the D2 receptor ligand [11C]-raclopride [170]. These studies demonstrate the interconnectivity between the opioidergic and dopaminergic systems in response to alcohol consumption. Naltrexone (Table 1), a μ-opioid receptor antagonist, targets the dopaminergic pathway by inhibiting μ-opioid receptors and disrupting the neurocascade that leads to striatal dopamine release [171]. After animal studies demonstrated the involvement of endogenous opioid system on the effect of alcohol, naltrexone was selected to be tested on alcohol-dependent participants in the hope of improving psychosocial rehabilitation [172, 173]. These studies showed the beneficial effects of naltrexone and were quickly replicated (e.g., [174]), which led to FDA approving this medication for treatment of alcoholism in 1983. Pharmacotherapeutic trials have tested the effectiveness of naltrexone for treatment of alcoholism and found positive therapeutic results. Naltrexone has been shown to reduce the frequency of heavy drinking days, increase the time before first relapse, produce lower relapse rates, reduce the number of total drinking days, and lower the number of drinks per drinking episode in alcoholics and to decrease the time lapse between first and second drinks among social drinkers [173–184]. However, other studies were unable to establish naltrexone’s efficacy as a moderator of alcohol consumption [185, 186].

Addiction Pharmacogenetics

457

Although its method of action is not fully understood, naltrexone is widely accepted as one of the safest and most effective pharmacotherapies for alcohol dependence. Naltrexone blunts alcohol’s reinforcing effects, including the “high” and the subjective positive stimulation following alcohol consumption, and, in general, restricts the euphoria produced by alcohol [187– 189]. This makes alcohol intoxication less satisfying and impedes the progression of drinking when delivered in combination with behavioral intervention [190], causing alcoholics to be less likely to resist relapse into heavy drinking. Naltrexone has been shown to activate the HPA axis by increasing proopiomelanocortin (POMC) synthesis [191] and cortisol levels, which have been correlated with decreased craving for alcohol [192]. When the OPRM1 A118G rs1799971 polymorphism was examined, 118G-allele carriers had higher cortisol concentrations both at baseline and after naloxone, a μ-opioid antagonist, treatment than did 118A/118A homozygous participants [151]. 118G-allele carriers experienced more intense “highs” and greater positive stimulation following alcohol consumption and experienced greater blunting of these subjective effects when treated with naltrexone than did AA homozygous participants [193]. When given naltrexone prior to drinking, G-allele carrier participants had lower levels of alcohol craving and more intense alcohol “highs” as their blood alcohol content increased. On the other hand, when 118A/118A homozygous participants were given naltrexone, alcohol cues produced greater craving with no effect in 118G-allele carriers [194]. Pharmacogenetic trials of naltrexone that examined the OPRM1 A118G rs1799971 polymorphism demonstrated that G-allele carriers experience better clinical response and lower relapse rates than did AA homozygous patients when treated with naltrexone [195, 196]. Oslin et al. (2003) examined the C17T rs1799972 and A118G polymorphisms of OPRM1 and their association with treatment outcome of naltrexone in alcohol-dependent patients [195]. They found that individuals of European descent with at least one copy of the A118G G allele had better results (e.g., lower rates of relapse and a longer time before relapse into heavy drinking) when treated with naltrexone than did those participants who were homozygous for the A allele. However, no difference in long-term abstinence rates between genotype groups was found. The C17T variant had no effect on treatment response to naltrexone. Ray et al. (2012) showed that naltrexone blunted alcohol craving but increased subjective intoxication in OPRM1 rs1799971 G-allele carriers compared to AA homozygous participants or placebo in a cohort of Asian Americans [197]. Setiawan et al. (2011) found that naltrexone blunted alcohol-induced euphoria both in women and in individuals with the G allele of OPRM1 A118G in a cohort of social drinkers [198]. Similarly, in a Korean cohort, the A118G G carriers

458

David P. Graham et al.

had longer time until relapse as compared to AA homozygous participants while on naltrexone [199]. However, studies have shown no pharmacogenetics effect of naltrexone, such as Coller et al., 2011, where naltrexone was found to be effective in reducing craving and alcohol use, but the A118G variant of OPRM1 was not found to be a predictor of these effects [200]. Oroszi et al., 2009, showed that specific OPRM1 haplotypes are found to be associated with good clinical outcome on naltrexone [201]. Specifically, 90% of individuals with the diplotypes (haplotype combination) AACCC/AGCCC of the variants rs1074287, rs1799971, rs510769, rs524731, and rs1381376, respectively, had “good clinical outcome” as compared to participants with other diplotypes that did not carry the rs1799971 G allele. Clinical trials have established the efficacy of naltrexone as a pharmacotherapy in conjunction with standard treatment protocols. Anton et al., 2008, studied the clinical outcomes of patients treated with naltrexone or placebo [196]. All participants in that study received either standard medical management (MM) alone or along with combined behavioral intervention (CBI). In patients who received both MM and CBI, no gene by medication interactions were found. However, in the participants who received MM without CBI, the OPRM1 118G-allele carriers treated with naltrexone had an increased percentage of days abstinent and an overall decrease in the percentage of drinking days relative to individuals receiving placebo. Within the MM without CBI group, patients who were carriers of the 118G allele had better results on naltrexone than did those patients homozygous for the A allele. However, within the participants who were treated with placebo, those homozygous for the A allele had better results than did those patients who were G-allele carriers. The association in a haplotype-based approach between genetic variants of OPRM1 and response to naltrexone in alcohol-dependent participants demonstrated that those with haplotypes containing the rs1799971 Asp40 (A118A) variant responded better to naltrexone than did those with the Asn40Asn (118G) genotype [201]. In another study of the OPRM1 A118G and response to naltrexone, Asp40 allele carriers had reduced alcohol craving when treated with naltrexone compared to Asn40Asn carriers [202]. Furthermore, a cohort of Alaskans with alcohol dependence found that naltrexone was more efficacious in those with the homozygous for the OPRM1 Asn40 allele [203]. However, this study had a low number of Asp40 carriers. Two studies fail to find an effect of OPRM1 on naltrexone treatment response [204, 205]. In the Tidey et al. clinical trial [204], a 3-week naltrexone treatment was associated to a reduction of drinking days, in particular in women and those with an early onset of alcoholism. In Gelernter et al.’s study [205], naltrexone treatment reduced heavy drinking.

Addiction Pharmacogenetics

6.2 δ-Opioid Receptor (OPRD1) Gene

459

The OPRD1 gene is located on chromosome 1p36.1 and encodes the G-protein-coupled δ-opioid receptor (DOP-r). Although MOP-r is the primary opioid receptor thought to function in the reinforcing effects of drugs of abuse, DOP-r can form heterodimers with the MOP-r to result in dopamine release in the nucleus accumbens [206–208]. OPRD1 genetic variants have been shown to be associated with substance abuse and dependence. Variants and haplotypes of OPRD1 were found to be associated with opioid, alcohol, and cocaine dependence in a case-control study [209]. Other studies have found other genetic variation in the OPRD1 gene in association with heroin addiction vulnerability in Germans [210] and European Americans [211, 212] and with cocaine addiction susceptibility in African Americans [213]. In a large case-control study in Australian heroin users, the GA haplotype composed of two intronic OPRD1 variants (rs2236857 and rs581111) was found to be related with vulnerability to develop heroin addiction [214]. However, this GA haplotype of these variants is rare T (rs4290270) in exon 9 of TPH2

Addiction Pharmacogenetics

463

and an intron 7 variant of TPH1 was found to be associated with heroin addiction [269]. The TPH2 variant 1125A>T (rs4290270) has been demonstrated to be a marker for allelic expression imbalance with the T-allele being expressed at twice the level of the A allele [270]. Hence, individuals with a TT genotype may produce more serotonin than do A-allele carrier participants. TPH2 (rs2129575) was included among the most relevant polymorphisms in development of a polygenic risk score associated with heroin use and craving scores in China [271]. In the disulfiram pharmacotherapy study for cocaine addiction, genetic analysis showed that individuals carrying the TPH2 rs4290270 A allele responded better to disulfiram compared to placebo than did TT homozygous individuals. Additionally, A carriers responded even better to disulfiram if they were also carrying an 5-HTTLPR S0 allele [251]. Hence, individuals with the low-expressing TPH2 and low-expressing 5-HTTLPR variants responded better, presumably in response to disulfiram’s effect of increasing serotonin levels. Thus, it appears that participants with low serotonergic metabolism respond to disulfiram, while those with normal serotonergic metabolism do not. 7.3 Tetraspanin 5 (TSPAN5) Gene

8

TSPAN5 is one of 33 members of the tetraspanin gene family associated with four transmembrane domain proteins of up to 350 amino acids [272] and has been reported to be associated with plasma concentrations of serotonin [273]. The TSPAN5 rs11947402 variant is located on chromosome four and associated with changes in serotonin at both 4 and at 8 weeks [273] in both whites and African Americans in major depressive disorder treated with SSRIs, while both knockdown and overexpression of TSPAN5 altered the expression of serotonin pathway genes [TPH1, TPH2, dopa decarboxylase (DDC), and monoamine oxidase A (MAOA)]. This was followed up in 2020 showing downregulation of TSPAN5 occurred with exposure to both ethanol or acamprosate and demonstrated that a TSPAN5 diplotype (rs11940430 AA, rs4699354 AA, and rs10029405 AA) was associated with acamprosate treatment outcomes in AUD, in particular to both abstinence length until heavy drinking and complete abstinence during 3 months of treatment while the three T-carrier diplotypes were associated with greater risk of relapse [274].

Other Genes with Variants Associated with Pharmacotherapeutic Response

8.1 ATP-Binding Cassette, Subfamily B (MDR/TAP), Member 1 (ABCB1) Gene

Methadone is transported across the blood-brain barrier by the P-glycoprotein 170 (P-gp) methadone transporter [275, 276] encoded by the ABCB1 gene located on chromosome 7q21.12 and is a member of the ATP-binding cassette (ABC) transporters. The ABCB1 gene is a member of the MDR/TAP subfamily whose

464

David P. Graham et al.

gene products are involved in multidrug resistance. Israeli individuals with the TT genotype of the ABCB1 rs1128503 (C1236T) variant and those with the TT-TT-TT genotype pattern of the ABCB1 variants rs1045642, rs2032582, and rs1128503, respectively, require higher doses of methadone to achieve stabilization [277]. However, this finding was not seen in a Han Chinese subject population [278]. Additionally, several additional variants and haplotypes in ABCB1 moderate methadone dosing [279]. Homozygous carriers of the AGCGC haplotype of five variants [A61G, G1199A, rs1128503 (C1236T), rs2032582 (G2677T), and rs1045642 (C3435T)] required higher doses of methadone than did noncarriers of this haplotype. Noncarriers of another haplotype, AGCTT, required higher methadone doses to achieve stabilization than did carriers of this haplotype. In contrast, a study of Han Chinese reported that participants who were carriers of the T variant allele of rs1045642 (C3435T) had a higher likelihood of requiring a larger methadone dose than did noncarrier participants [106]. The TT genotype of rs1045642 also is associated with a higher ratio of methadone in the brain compared to blood in individuals with methadone-related deaths, suggesting a role of this variant in methadone toxicity [280]. In a multigene analysis, OPRM1 118A/118A homozygous participants who also were homozygous for the ABCB1 AGCGC (“wild-type”) haplotype (defined above) or who were homozygous for the AGTTT haplotype required lower methadone doses and had higher plasma methadone concentrations (Ctrough) to suppress withdrawal than did AGCGC/AGTTT diplotype participants [281]. Conversely, those participants with ABCB1 AGCGC/AGTTT diplotype who were also OPRM1 AA homozygous required a lower methadone dose and had lower plasma methadone concentrations than did OPRM1 G-allele carriers. 8.2 α1AAdrenoceptor (ADRA1A) Gene

α-1-adrenoceptors are members of the G-coupled protein receptor (GCPR) superfamily and regulate proliferation and growth through the activation of the phosphatidylinositol-calcium second messenger system. Previous studies have shown that the adrenergic system is involved in cocaine addiction, treatment, and the development of cocaine-induced paranoia [88, 130, 140, 282]. In the disulfiram study for cocaine addiction, genetic analysis showed that T carriers of the rs1048101 (Arg347Cys) variant in exon 1 of the ADRA1A gene on chromosome 8q21.2 had a reduced number of cocaine-positive urines on disulfiram, while individuals with the CC genotype showed no treatment effect [283]. It is likely that the Arg allele of ADRA1A encodes an α1A-adrenoceptor with reduced signaling efficiency, and this may partly explain its role in the efficacy of disulfiram. Additionally, A-allele carriers of ADRA1A, rs1048101, which codes for serine, is associated with a shorter

Addiction Pharmacogenetics

465

duration of time between first use of heroin and addiction [284]. This study also examined another ADRA1A variant, rs3808585, and showed that C-allele carriers have increased memory impairment after prolonged heroin use. 8.3 Brain-Derived Neurotrophic Factor (BDNF) and Nerve Growth Factor (Beta Polypeptide) (NGF) Genes

Brain-derived neurotrophic factor (BDNF) is a member of the neurotrophin family of growth factors and is required for neuronal growth and differentiation. It is encoded by the BDNF gene located on chromosome 11p14.1. Variants in BDNF have been found to be associated with synaptic plasticity, hippocampal volume, TBI severity, and a number of psychiatric diseases, including schizophrenia [285–287]. The BDNF Val66Met variant (rs6265) has been associated with substance abuse in males and drug-seeking behavior in heroin-dependent individuals [288, 289]. Met-allele carriers have a higher frequency and an earlier onset of substance abuse than Val/Val homozygotes. In a study of the response of opioid addicts to methadone treatment, carriers of the BDNF haplotype CCGCCG (rs7127507, rs1967554, rs11030118, rs988748, rs2030324, and rs11030119) had poorer response to methadone maintenance treatment compared to individuals with the other haplotypes [290]. The nerve growth factor (beta polypeptide) gene (NGF) located on chromosome 1p13.1 encodes a neurotrophic factor that is important in the differentiation and maintenance of several types of sympathetic and sensory neurons and is critical to the sensation of pain [291]. Individuals homozygous for the rs2239622 A allele of NGF were found to require a lower mean daily methadone dose than did those individuals with the other genotypes [292].

8.4 β-Arrestin 2 (ARRB2) Gene

Arrestin/beta-arrestin protein family members are involved in the agonist-mediated desensitization of G-protein-coupled receptors (GPCRs). GPCRs are a large family of receptors that signal ligand binding through the cell membrane. A member of the arrestin/ beta-arrestin protein family is β-arrestin2, encoded by ARRB2 on chromosome 17p13.2. ARRB2 is expressed in many tissues, with high expression in the brain. The function of β-arrestin2 is to promote the desensitization and internalization of GPCRs, including the opioid receptors [293]. A study on the role ARRB2 variants influence response to methadone treatment was conducted on a Swiss cohort. Heroin-dependent individuals homozygous for the ARRB2 gene variant who had either the rs3786047 AA, rs1045280 CC, or rs2036657 GG genotypes had poor response to methadone maintenance therapy [294]. A study looking at African American heroin users found that the C allele of rs1045280 actually has a protective effect, suggesting an interaction between ancestral differences and genetic polymorphisms in the ARRB2 gene [295].

466

David P. Graham et al.

8.5 Cytochrome P450, Family 2, Subfamily D, Polypeptide 6 (CYP2D6) and Subfamily B, Polypeptide 6 (CYP2B6) Genes

The cytochrome (CYP) P450 superfamily of enzymes is involved in the oxidation of many organic substances, including methadone. Methadone is primarily metabolized in the liver by CYP3A4 by N-demethylation to the inactive metabolite EDDP (2-ethylidene1,5-dimethyl-3,3-diphenylpyrrolidine) [296]. Several other cytochrome P450s, including CYP2B6 and CYP2D6, are also involved in the metabolism of methadone. The product of the cytochrome P450, family 2, subfamily D, polypeptide 6 (CYP2D6) gene is an enzyme that metabolizes specific opioid drugs into their active state, such as codeine and methadone [297]. CYP2D6 is a highly polymorphic gene located on chromosome 22q13.2 with at least 75 different alleles. Individuals can be categorized as poor (PM), extensive (EM), and ultrarapid (UM) metabolizers of drugs based on their number of functional CYP2D6 alleles. Using this classification based on the CYP2D6 genotyping, heroin-dependent patient satisfaction of MMTP was assessed in a Caucasian cohort [298]. Patients that were classified as PM or EM scored higher on the Verona Service Satisfaction Scale for methadone-treated, opioid-dependent patients (VSSS-MT) than did the patients classified as UM. In addition, UM males were less satisfied than were UM females with MMTP on the Basic Intervention VSSS-MT subscale. Other studies have shown the involvement of specific variants in CYP3A, as well as CYP2D6, on methadone plasma levels [299]. A study of an Italian cohort of patients with chronic lower back pain treated with the opioids codeine or oxycodone showed that individuals with low CYP2D6 activity (corresponding to the PM group) had a lack of therapeutic effect, and individuals in the UM group showed increased risk of side effects [300]. For a 2021 review of pharmacogenomics and pharmacokinetics for oxycodone, please see [301]. The cytochrome P450, family 2, subfamily B, polypeptide 6 gene (CYP2B6) located on chromosome 19q13.2 encodes a protein that is localized to the liver and metabolizes specific drugs, including methadone. Israeli former heroin addicts in MMTP who were homozygous for the CYP2B6 6*6 genotype, which is defined by the variant alleles of variants 516G>T (rs3745274) and 785A>G (rs2279343), required lower mean methadone doses than did heterozygous or noncarrier individuals of the 6*6 genotype [302]. Related to this finding is another study that found an association of CYP2B6 6*6 homozygous participants with having higher methadone plasma levels [303]. Resequencing of the CYP2B6 gene identified seven variants, including rs3745274 and rs2279343, that were found to be associated with (S)-methadone plasma levels, suggesting an association with reduced CYP2B6 activity [304]. Additionally, individuals with the T-allele of rs374274 (516 A->T) had a threefold higher chance of requiring lower methadone doses than did individuals homozygous for the G allele [106]. The previously mentioned variant rs3745274 and two

Addiction Pharmacogenetics

467

other variants, rs8192719 and rs321137, also are associated with increased rates of methadone fatalities [305]. This underlines the importance of identifying these variants in individuals receiving methadone maintenance therapy, as receiving too high of a dose coupled with a reduced ability to metabolize the methadone can have fatal consequences. 8.6 GammaAminobutyric Acid Type B Receptor Subunit 1 (GABBR1) Gene

In a single double-blind, randomized, placebo-controlled trial of baclofen [306], the variant rs29220 in the GABAB receptor subunit 1 gene (GABBR1) showed a medication x genotype interaction for time to relapse and for follow-up alcohol consumption. The relapse hazard ratio for the CC genotype was 0.32 and for the G-allele carriers (CG/GG) was 1.07. The authors summarized these findings noting possible clinical applications regarding the CC genotype having displayed a beneficial response to baclofen.

8.7 GATA Binding Protein 4 (GATA4) Gene

The transcription factor GATA-binding protein 4 encoded by the GATA4 gene located on chromosome 8p23.1 is a zinc-finger transcription factor, which binds a GATA motif found in the promoter region of a variety of genes. GATA4 has been shown to regulate the expression of atrial natriuretic peptide (ANP) [307], which may play a major role in alcohol withdrawal and dependence [308]. A variant in GATA4, rs13273672, has been found to be associated with ANP plasma concentration in alcohol-dependent participants, with the AA genotype group having higher levels of ANP than did the G-allele carrier genotype groups [309]. One study found an association with nominal significance of alcohol dependence with several GATA4 variants but not with the rs13273672 variant [310]. A genome-wide association study of alcohol dependence conducted with German alcoholics identified the GATA4 rs13273672 as being one of 16 variants showing an association with nominal significance [311]. A follow-up study was conducted on this variant in a trial of acamprosate, naltrexone, or placebo for the treatment of alcohol dependence [309]. The GATA4 rs13273672 variant was shown to be associated with relapse risk during the 90-day trial, with the GG genotype group relapsing sooner than did the AG genotype group, which relapsed sooner than did the AA group. Post hoc analysis demonstrated this finding was due to those participants treated with acamprosate. It was hypothesized that the GATA4 rs13273672 variant would regulate ANP activity, which subsequently would interact with the glutamate system, since GABA receptors are a target of acamprosate’s action.

8.8 Glutamate Ionotropic Receptor Kainate-Type Subunit 1 (GRIK1) Gene

The glutamate ionotropic receptor kainate-type subunit 1 (GRIK1) gene encodes the GluK1 subunit of the excitatory kainite receptor. Prior work has shown CC homozygous individuals for rs2832407 were associated with a reduced rate of heavy drinking when treated with topiramate as compared to the A-allele carriers [312], an effect

468

David P. Graham et al.

that persisted for 6 months [313]. In a more recent study involving electrophysiology with pluripotent stem cells, no effect of GRIK1 genotype on mRNA expression was identified, but CC genotype was noted to have a greater expression of the GRIK1 antisense2 expression as compared to those with the AA genotype. Differential effects in synaptic activity when exposed to 5 μM topiramate was identified, with a smaller reduction in excitatory event frequency in CC genotype compared to the AA genotype [314]. 8.9 Methylenetetrahydrofolate Reductase (MTHFR) Gene

Methylenetetrahydrofolate reductase (MTHFR) is encoded by the MTHFR gene located on chromosome 1p36.22. It is the ratelimiting enzyme in the folic acid metabolic cycle that is critical for the production of metabolites for downstream DNA and protein methylation. The minor T-allele of rs1801133 codes for an alanine to valine substitution that results in a thermolabile form of the MTHFR enzyme having 50–60% the activity of the non-variant enzyme [315]. This variant has been found to be associated with vulnerability to develop spinal bifida, a condition in which the neural tube fails to close leaving neural tissue open to the environment [316, 317]. The MTHFR enzyme also plays an important role in homocysteine metabolism, with T-allele carriers having increased plasma homocysteine levels due to reduced MTHFR activity [315]. Increased homocysteine concentrations have been associated with alcohol use disorder [318], suggesting a link between the rs1801133 C677T variant and alcohol use disorder, although the results of these studies have been inconsistent [319– 321]. In the disulfiram pharmacogenetics of cocaine addiction study, T-allele carriers of the MTHFR rs1801133 C677T variant were shown to respond to disulfiram treatment, with no change in cocaine-free urines in the placebo group with these genotypes [322]. Those participants with two copies of the C allele had a poorer response to disulfiram than did those with one or two T-alleles. Both disulfiram and cocaine have been shown to change the epigenetic landscape of the genome [323, 324]. Since MTHFR is a critical enzyme in the production of metabolites for DNA and protein methylation, it is possible that there is an interaction of these variants with epigenetics. The resulting effects on gene expression may cause the pharmacogenetic effect observed with disulfiram and MTHFR.

8.10 Myocardin (MYOCD) and Glutamate Receptor Metabotropic 6 (GRM6) Genes

Myocardin (encoded by the MYOCD gene on chromosome 17p12) is a transcription factor that functions in the cardiovascular system. It is expressed primarily early in development in cardiac muscle cells, where it is involved in the chromatin remodeling of SRF target genes [325]. The GRM6 gene located on chromosome 5q35 encodes the metabotropic glutamate receptor 6 (mGuR6), a G-protein-coupled receptor involved in glutamatergic signaling in

Addiction Pharmacogenetics

469

the central nervous system. An intronic variant in MYOCD, rs1714984, was previously found as the top hit with the smallest p-value based on genotype frequency to be associated with heroin addiction vulnerability in a genome-wide association study of former heroin addicts and controls, with the AA genotype group at greatest risk [326]. In that same study, GRM6 rs953741was identified as one of the top candidate gene variants being in association with heroin addiction, with the A allele conferring increased risk. A subsequent study was conducted using these variants to assess response to treatment in methadone-treated participants [327]. Patients carrying the A-allele at MYCOD rs1714984 had an increased risk of being nonresponders to methadone maintenance treatment if they were also carriers of the AG genotype at GRM6 rs953741. 8.11 Nectin Cell Adhesion Molecule 4 (NECTIN4) Gene

The nectin cell adhesion molecule 4 gene (NECTIN4) encodes two immunoglobulin-like (Ig-like) C2-type domains and one Ig-like V-type domain [328]. The molecule is a single-pass type I membrane protein and is involved in cell adhesion through transhomophilic and trans-heterophilic interactions. In a study of MMTP Taiwanese participants, three highly linked NECTIN4 polymorphisms (rs11265549, rs3820097, rs4656978) were associated with methadone dosage, R,S-methadone plasma concentration, and TNF-α [329].

8.12 Potassium Inwardly Rectifying Channel, Subfamily J, Member 6 (KCNJ6) Gene

The potassium inwardly rectifying channel, subfamily J, member 6 gene (KCNJ6, also known as GIRK2) is located on chromosome 21q22.13. It encodes a potassium channel that is regulated by Gprotein-coupled receptors and is important in opioid receptor signaling and analgesia [330]. KCNJ6 is expressed in the substantia nigra (SN), as well as in the VTA [331]. Genetic variation in KCNJ6 has been shown to be associated in alcoholism. Specifically, the rs2836016 G allele was found to be associated with alcohol dependence in an adult and adolescent cohort. The GG homozygous participants in the adolescent group had an increase in hazardous alcohol use but only in those participants who experienced early life stress [332]. Activation of μ-, δ-, or κ-opioid or dopamine D2 receptors can open GIRK channels, inhibiting voltage-gated calcium channels and adenylyl cyclases [333, 334], as well as inhibiting neuronal activity [335]. Genetic variation in KCNJ6 was examined in relation to the requirements for postoperative analgesics [336]. In patients who had undergone major open abdominal surgery, those with the AA genotype of rs2070995 were found to require a higher dose of equivalent oral morphine and had more frequent administration of analgesics. Another study reported on the involvement of this variant in MMTP of former heroin addicts, chronic pain patients, and healthy volunteers [337]. The average and daily methadone

470

David P. Graham et al.

dose during the first year of treatment was greater in former heroin participants with the KCNJ6 rs2070995 AA genotype than did in individuals with other genotypes as was a higher opioid dosing if chronic pain patients. Another KCNJ6 variant, rs702859, has been shown to influence event-related oscillations (EROs) measured during EEGs of individual offspring from families with high incidences of alcoholism among family members [338]. The A allele of rs702859 contributes to lower theta power at central and parietal regions of the brain, with AA homozygotes having lower theta power than AG heterozygotes. Lower theta power is thought to indicate lower efficiency of cognitive processing and is also correlated with alcohol use disorder in general [339, 340].

9

Conclusions In this chapter, we have shown that the response to pharmacotherapy for addiction is genetically complex; that is, therapeutic response is a result of small influences of many genetic variants, as well as that of the environment. The research reviewed herein suggests that these influences may be additive and that the identification of more predictive genetic factors may help to tailor more effective pharmacotherapies. Most of the genetic associations we have reviewed have yet to be replicated, which will be required to confirm and extend their findings. A number of confounding factors may be influencing the results. Differences in ethnic composition may alter the findings, and ethnic stratification should be properly controlled. Many of these studies were conducted in cohorts with small sample sizes, which may have influenced their conclusions. Large sample sizes will be required to validate the role these variants have in pharmacotherapeutic response, especially for variants with low effect sizes or low allele frequencies. In addition, some genetic studies may have not been adequately powered to detect a genetic effect, and therefore, the results have appeared as a negative finding. The use of polygenic risk scores may be of particular benefit as it allows for the simultaneous evaluation of a large number of variants with a single score representing each participant’s genetic loading for either a disease or a continuous trait [31]. Given the heterogeneity of drug abuse and dependence, the “one-size-fits-all” strategy typically utilized in the treatment of these disorders has seen limited success. As demonstrated with the examples presented in this chapter, a significant proportion of the variability seen in drug addiction and dependence pharmacotherapy is due to genetic heterogeneity. Similar to the effects that genetic variation has on the vulnerability to develop an addiction [341], genetic variation affects response to treatments for drug addiction, including reward and positive stimulation resulting from drug use.

Addiction Pharmacogenetics

471

A goal of personalized medicine is to match the patient to the most effective pharmacotherapy for their disease based on that person’s genetic background. Further studies will need to be conducted to identify genetic differences in the pathways involved in the development, maintenance, and relapse of addiction to optimize treatment outcomes. Personalized medical treatment already has been incorporated into the care of cardiology and oncology patients to improve their efficacy. This success may be replicated with the creation of personalized treatments for cocaine, opioid, and alcohol addiction based on an individual’s genetics, history, current physical condition, and other elements unique to that person [342]. Since the initial success of treatment plays a substantial role in compliance and retention, personalizing treatment based on genetic background should increase treatment efficacy, therefore improving compliance as well. The findings from pharmacogenetic studies of treatments of drug addiction will, in the near future, better equip medical professionals with the knowledge to assign personalized treatment strategies to patients with substance use disorders, therefore effecting better outcomes and greater treatment success.

Acknowledgments Supported by NIH/NIDA 5 P50 DA018197-05 (TK, DN) through MD Anderson’s Cancer Center Support Grant DA026120 NIH/NIDA DA026120 and the Toomim Family Fund. This material is the result of work supported with resources and the use of facilities at the Michael E. DeBakey VA Medical Center, Houston, TX. References 1. Goplerud E, Hodge S, Benham T (2017) A substance use cost calculator for US employers with an emphasis on prescription pain medication misuse. J Occup Environ Med 59(11):1063–1071 2. Wolf BC, Lavezzi WA, Sullivan LM, Flannagan LM (2004) Methadone-related deaths in Palm Beach County. J Forensic Sci 49(2): 375–378 3. Zvosec DL, Smith SW, Porrata T, Strobl AQ, Dyer JE (2011) Case series of 226 γ-hydroxybutyrate-associated deaths: lethal toxicity and trauma. Am J Emerg Med 29(3):319–332 4. MacGowan RJ, Fichtner RR, Swanson N, Collier C, Kroliczak A, Cole G (1997) Factors associated with client-reported HIV infection

among clients entering methadone treatment. AIDS Educ Prev 9:205–217 5. Meandzija B, O’Connor PG, Fitzgerald B, Rounsaville BJ, Kosten TR (1994) HIV infection and cocaine use in methadone maintained and untreated intravenous drug users. Drug Alcohol Depend 36:109–113 6. The health consequences of smoking—50 years of progress: a report of the surgeon general (2014) US Dept. of Health and Human Services, Centers for Disease Control and Prevention, Atlanta. Accessed 10 Feb 2021 7. Centers for Disease Control and Prevention (CDC). Alcohol-Related Disease Impact (ARDI). CDC, Atlanta. April 2022

472

David P. Graham et al.

8. Stahre M, Roeber J, Kanny D, Brewer RD, Zhang X (2014) Contribution of excessive alcohol consumption to deaths and years of potential life lost in the United States. Prev Chronic Dis 11:E109 9. Kanny D, Naimi TS, Liu Y, Lu H, Brewer RD (2018) Annual Total binge drinks consumed by U.S. adults, 2015. Am J Prev Med 54(4): 486–496 10. Substance Abuse and Mental Health Services Administration (2020) Key substance use and mental health indicators in the United States: Results from the 2019 National Survey on Drug Use and Health (HHS Publication No. PEP20-07-01-001, NSDUH Series H-55). Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration, Rockville 11. Fishbain DA, Cole B, Lewis J, Rosomoff HL, Rosomoff RS (2008) What percentage of chronic nonmalignant pain patients exposed to chronic opioid analgesic therapy develop abuse/addiction and/or aberrant drugrelated behaviors? A structured evidencebased review. Pain Med 9(4):444–459 12. Kreek MJ, Nielsen DA, Butelman ER, LaForge KS (2005) Genetic influences on impulsivity, risk taking, stress responsivity and vulnerability to drug abuse and addiction. Nat Neurosci 8:1450–1457 13. Volkow ND, Michaelides M, Baler R (2019) The neuroscience of drug reward and addiction. Physiol Rev 99(4):2115–2140 14. Kanzler HR, Smith RV, Schnoll R, Moustafa A, Greenstreet-Akman E (2017) Precision medicine and pharmacogenetics: what does oncology have that addiction medicine doesn’t? Addiction 112(12):2086–2094 15. Haile CN, Kosten TR, Kosten TA (2009) Pharmacogenetic treatments for drug addiction: cocaine, amphetamine and methamphetamine. Am J Drug Alcohol Abuse 35: 161–177 16. Prom-Wormley EC, Ebejer J, Dick DM, Bowers MS (2017) The genetic epidemiology of substance use disorder: a review. Drug Alcohol Depend 180:241–259. https://doi. org/10.1016/j.drugalcdep.2017.06.040 17. Malhotra AK, Murphy GM Jr, Kennedy JL (2004) Pharmacogenetics of psychotropic drug response. Am J Psychiatry 161:780–796 18. Murphy GM Jr, Kremer C, Rodrigues HE, Schatzberg AF (2003) Pharmacogenetics of antidepressant medication intolerance. Am J Psychiatry 160:1830–1835 19. Gupta S, Jain S, Brahmachari SK, Kukreti R (2006) Pharmacogenomics: a path to

predictive medicine for schizophrenia. Pharmacogenomics 7:31–47 20. deLeon J, Armstrong SC, Cozza KL (2006) Clinical guidelines for psychiatrists for the use of pharmacogenetic testing for CYP450 and CYP450 2C19. Psychosomatics 47:75–85 21. Spina E, de Leon J (2015) Clinical applications of CYP genotyping in psychiatry. J Neural Transm (Vienna) 122(1):5–28 22. Rogers JF, Nafziger AN, Bertino JS Jr (2002) Pharmacogenetics affects dosing, efficacy, and toxicity of cytochrome P450-metabolized drugs. Am J Med 113:746–750 23. Yoshida K, Nishizawa D, Ide S, Ichinohe T, Fukuda KI, Ikeda K (2018) A pharmacogenetics approach to pain management. Neuropsychopharmacol Rep 38(1):2–8 24. Ortiz JF, Wirth M, Eskander N, Cozar JC, Fatade O, Rathod B (2020) The genetic foundations of serotonin syndrome, neuroleptic malignant syndrome, and malignant hyperthermia: is there a genetic association between these disorders? Cureus 12(9):e10635 25. Jarvis JP, Peter AP, Shaman JA (2019) Consequences of CYP2D6 copy-number variation for pharmacogenomics in psychiatry. Front Psych 10:432 26. Moraes F, Go´es A (2016) A decade of human genome project conclusion: scientific diffusion about our genome knowledge. Biochem Mol Biol Educ 44(3):215–223 27. ENCODE Project Consortium, Bernstein BE, Birney E, Dunham I, Green ED, Gunter C, Snyder M (2012) An integrated encyclopedia of DNA elements in the human genome. Nature 489:57–74 28. Genomes Project Consortium, Abecasis GR, Auton A, Brooks LD, DePristo MA, Durbin RM, Handsaker RE, Kang HM, Marth GT, McVean GA (2012) An integrated map of genetic variation from 1,092 human genomes. Nature 491:56–65 29. Genomes Project Consortium, Abecasis GR, Altshuler D, Auton A, Brooks LD, Durbin RM, Gibbs RA, Hurles ME, McVean GA (2010) A map of human genome variation from population-scale sequencing. Nature 467:1061–1073 30. Shendure J, Findlay GM, Snyder MW (2019) Genomic medicine – progress, pitfalls, and promise. Cell 177(1):45–57 31. Lewis CM, Vassos E (2020) Polygenic risk scores: from research tools to clinical instruments. Genome Med 12(1):44 32. Wray NR, Lin T, Austin J, McGrath JJ, Hickie IB, Murray GK, Visscher JM (2021) From Basic Science to Clinical Application of

Addiction Pharmacogenetics Polygenic Risk Scores A Primer. JAMA Psychiatry 78(1):101–109 33. Folkersen L, Pain O, Ingason A, Werge T, Lewis CM, Austin J (2020) Impute.me: an open-source, non-profit tool for using data from direct-to-consumer genetic testing to calculate and interpret polygenic risk scores. Front Genet 11:578 34. Palladino VS, McNeill R, Reif A, KittelSchneider S (2019) Genetic risk factors and gene-environment interactions in adult and childhood attention-deficit/hyperactivity disorder. Psychiatr Genet 29(3):63–78 35. Zhong Y, Yang B, Qian Y, Cao Q, Chang S, Wang Y, Yang L (2020) The association with quantitative response to attention-deficit/ hyperactivity disorder medication of the previously identified neurodevelopmental network genes. J Child Adolesc Psychopharmacol 30(6):348–354 36. Tomasi J, Lisoway AJ, Zai CC, Harripaul R, Mu¨ller DJ, Zai GCM, McCabe RE, Richter MA, Kennedy JL, Tiwari AK (2019) Towards precision medicine in generalized anxiety disorder: review of genetics and pharmaco(epi)genetics. J Psychiatr Res 119:33–47 37. Lerman CE, Schnoll RA, Munafo` MR (2007) Genetics and smoking cessation improving outcomes in smokers at risk. Am J Prev Med 33(6 Suppl):S398–S405 38. Meoller SJ, Parvaz MA, Shumay E, Wu S, Beebe-Wang N, Konova AB, Misyrlis M, Alia-Klein N, Goldstein RZ (2014) Monoamine polygenic liability in health and cocaine dependence: imaging genetics study of aversive processing and associations with depression symptomatology. Drug Alcohol Depend 140:17–24 39. Gerra MC, Manfredini M, Cortese E, Antonioni MC, Leonardi C, Magnelli F, Somaini L, Jayanthi S, Cadet JL, Donnini C (2019) Genetic and environmental risk factors for cannabis use: preliminary results for the role of parental care perception. Subst Use Misuse 54(4):670–680 40. Nguyen D, Naffziger EE, Berridge KC (2021) Positive affect: nature and brain bases of liking and wanting. Curr Opin Behav Sci 39:72–78 41. Pfaus JG, Damsma G, Nomikos GG, Wenkstern DG, Blaha CD, Phillips AG, Fibiger HC (1990) Sexual behavior enhances central dopamine transmission in the male rat. Brain Res 530(2):345–348 42. Small DM, Jones-Gotman M, Dagher A (2003) Feeding-induced dopamine release in dorsal striatum correlates with meal

473

pleasantness ratings in healthy human volunteers. NeuroImage 19(4):1709–1715 43. Di Chiara G, Imperato A (1988) Drugs abused by humans preferentially increase synaptic dopamine concentrations in the mesolimbic system of freely moving rats. Proc Natl Acad Sci USA 85(14):5274–5278 44. Mollick JA, Chang LJ, Krishnan A, Hazy TE, Krueger KA, Frank GKW, Wager TD, O’Reilly RC (2021) The neural correlates of cued reward omission. Front Hum Neurosci 15:615313 45. Dackis C, O‘Brien C (2005) Neurobiology of addiction: treatment and public policy ramifications. Nat Neurosci 8(11):1431–1436 46. Garcia AF, Crummy EA, Webb IG, Nooney MN, Ferguson SM (2021) Distinct populations of cortical pyramidal neurons mediate drug reward and aversion. Nat Commun 12(1):182 47. Crummy EA, O’Neal TJ, Baskin BM, Ferguson SM (2020) One is not enough: understanding and modeling polysubstance use. Front Neurosci 14:569 48. Han D, Gu H (2006) Comparison of the monoamine transporters from human and mouse in their sensitivities to psychostimulant drugs. BMC Pharmacol 6(1):6 49. Webber HE, Lopez-Gamundi P, Stamatovich SN, de Wit H, Wardle MC (2021) Using pharmacological manipulations to study the role of dopamine in human reward functioning: A review of studies in healthy adults. Neurosci Biobehav Rev 120:123–158 50. Heilig M, Goldman D, Berrettini W, O’Brien CP (2011) Pharmacogenetic approaches to the treatment of alcohol addiction. Nat Rev Neurosci 12:670–684 51. Ma H, Zhu G (2014) The dopamine system and alcohol dependence. Shanghai Arch Psychiatry 26(2):61–68. https://doi.org/10. 3969/j.issn.1002-0829.2014.02.002 52. Koob GF (1992) Drugs of abuse: anatomy, pharmacology and function of reward pathways. Trends Pharmacol Sci 13:177–184 53. Weiss F, Lorang MT, Bloom FE, Koob GF (1993) Oral alcohol self-administration stimulates dopamine release in the rat nucleus accumbens: genetic and motivational determinants. J Pharmacol Exp Ther 267:250–258 54. Imperato A, Di Chiara G (1986) Preferential stimulation of dopamine release in the nucleus accumbens of freely moving rats by ethanol. J Pharmacol Exp Ther 239:219–228 55. Vena AA, Zandy SL, Cofresı´ RU, Gonzales RA (2020) Behavioral, neurobiological, and neurochemical mechanisms of ethanol self-

474

David P. Graham et al.

administration: a translational review. Pharmacol Ther 212:107573 56. Bourdy R, Barrot M (2012) A new control center for dopaminergic systems: pulling the VTA by the tail. Trends Neurosci 35:681–690 57. Johnson SW, North RA (1992) Opioids excite dopamine neurons by hyperpolarization of local interneurons. J Neurosci 12: 483–488 58. Berridge KC, Robinson TE (1998) What is the role of dopamine in reward: hedonic impact, reward learning, or incentive salience? Brain Res Brain Res Rev 28:309–369 59. Comings DE, Comings BG, Muhleman D, Dietz G, Shahbahrami B, Tast D, Knell E, Kocsis P, Baumgarten R, Kovacs BW et al (1991) The dopamine D2 receptor locus as a modifying gene in neuropsychiatric disorders. JAMA 266(13):1793–1800 60. Golimbet VE, Aksenova MG, Nosikov VV, Orlova VA, Kaleda VG (2003) Analysis of the linkage of the Taq1A and Taq1B loci of the dopamine D2 receptor gene with schizophrenia in patients and their siblings. Neurosci Behav Physiol 33:223–225 61. Comings DE, Rosenthal RJ, Lesieur HR, Rugle LJ, Muhleman D, Chiu C, Dietz G, Gade R (1996) A study of the dopamine D2 receptor gene in pathological gambling. Pharmacogenetics 6:223–234 62. Blum K, Noble EP, Sheridan PJ, Finley O, Montgomery A, Ritchie T, Ozkaragoz T, Fitch RJ, Sadlack F, Sheffield D et al (1991) Association of the A1 allele of the D2 dopamine receptor gene with severe alcoholism. Alcohol 8:409–416 63. Persico AM, Bird G, Gabbay FH, Uhl GR (1996) D2 dopamine receptor gene TaqI A1 and B1 restriction fragment length polymorphisms: enhanced frequencies in psychostimulant-preferring polysubstance abusers. Biol Psychiatry 40:776–784 64. Noble EP (1994) Polymorphisms of the D2 dopamine receptor gene and alcoholism and other substance use disorders. Alcohol Alcohol Suppl 2:35–43 65. Munafo M, Clark T, Johnstone E, Murphy M, Walton R (2004) The genetic basis for smoking behavior: a systematic review and metaanalysis. Nicotine Tob Res 6:583–597 66. Lawford BR, Young RM, Noble EP, Sargent J, Rowell J, Shadforth S, Zhang X, Ritchie T (2000) The D(2) dopamine receptor A(1) allele and opioid dependence: association with heroin use and response to

methadone treatment. Am J Med Genet 96(5):592–598 67. Zhang Y, Bertolino A, Fazio L, Blasi G, Rampino A, Romano R, Lee ML, Xiao T, Papp A, Wang D, Sadee W (2007) Polymorphisms in human dopamine D2 receptor gene affect gene expression, splicing, and neuronal activity during working memory. Proc Natl Acad Sci U S A 104:20552–20557. do 68. Moyer RA, Wang D, Papp AC, Smith RM, Duque L, Mash DC, Sadee W (2011) Intronic polymorphisms affecting alternative splicing of human dopamine D2 receptor are associated with cocaine abuse. Neuropsychopharmacology 36:753–762 69. Hirvonen MM, Lumme V, Hirvonen J, Pesonen U, Nagren K, Vahlberg T, Scheinin H, Hietala J (2009) C957T polymorphism of the human dopamine D2 receptor gene predicts extrastriatal dopamine receptor availability in vivo. Prog NeuroPsychopharmacol Biol Psychiatry 33: 630–636 70. Duan J, Wainwright MS, Comeron JM, Saitou N, Sanders AR, Gelernter J, Gejman PV (2003) Synonymous mutations in the human dopamine receptor D2 (DRD2) affect mRNA stability and synthesis of the receptor. Hum Mol Genet 12(3):205–216 71. Gupta M, Chauhan C, Bhatnagar P, Gupta S, Grover S, Singh PK, Purushottam M, Mukherjee O, Jain S, Brahmachari SK, Kukreti R (2009) Genetic susceptibility to schizophrenia: role of dopaminergic pathway gene polymorphisms. Pharmacogenomics 10(2): 277–291 72. Meyers JL, Nyman E, Loukola A, Rose RJ, Kaprio J, Dick DM (2013) The association between DRD2/ANKK1 and genetically informed measures of alcohol use and problems. Addict Biol 18:523–536 73. He H, Wu H, Yang L, Gao F, Fan Y, Feng J, Ma X (2016) Associations between dopamine D2 receptor gene polymorphisms and schizophrenia risk: a PRISMA compliant meta-analysis. Neuropsychiatr Dis Treat 12: 3129–3144 74. Gonzalez-Castro TB, Hernendez-Diaz Y, Juarez-Rojop IE, Lopez-Narvaez ML, Tovilla-Zarate CA, Genis-Mendoza A, Alpuin-Reyes M (2016) The role of C957T, TaqI and Ser311Cys polymorphisms of the DRD2 gene in schizophrenia: systematic review and meta-analysis. Behav Brain Funct 12(1):29 75. Diekhof EK, Richter A, Brodmann K, Gruber O (2021) Dopamine multilocus genetic

Addiction Pharmacogenetics profiles predict sex differences in reactivity of the human reward system. Brain Struct Funct 226(4):1099–1114 76. Zha J, Zhou Q, Xu LG, Chen D, Li L, Zhai Z, Shu HB (2004) RIP5 is a RIP-homologous inducer of cell death. Biochem Biophys Res Commun 319(2):298–303 77. Blum K, Noble EP, Sheridan PJ, Montgomery A, Ritchie T, Jagadeeswaran P, Nogami H, Briggs AH, Cohn JB (1990) Allelic association of human dopamine D2 receptor gene in alcoholism. JAMA 263(15): 2055–2060 78. Jo¨nsson EG, No¨then MM, Gru¨nhage F, Farde L, Nakashima Y, Propping P, Sedvall GC (1999) Polymorphisms in the dopamine D2 receptor gene and their relationships to striatal dopamine receptor density of healthy volunteers. Mol Psychiatry 4(3):290–296 79. Ritchie T, Noble EP (1996) [3H]naloxone binding in the human brain: alcoholism and the TaqI A D2 dopamine receptor polymorphism. Brain Res 718(1–2):193–197 80. Volkow ND, Fowler JS, Wolf AP, Schlyer D, Shiue CY, Alpert R, Dewey SL, Logan J, Bendriem B, Christman D et al (1990) Effects of chronic cocaine abuse on postsynaptic dopamine receptors. Am J Psychiatry 147(6):719–724 81. Chen R, Ferris MJ, Wang S (2020) Dopamine D2 autoreceptor interactome: targeting the receptor complex as strategy for treatment of substance use disorder. Pharmacol Ther 213: 107583 82. Volkow ND, Wang GJ, Fowler JS, Logan J, Hitzemann R, Ding YS, Pappas N, Shea C, Piscani K (1996) Decreases in dopamine receptors but not in dopamine transporters in alcoholics. Alcohol Clin Exp Res 20(9): 1594–1598 83. Wang GJ, Volkow ND, Fowler JS, Logan J, Abumrad NN, Hitzemann RJ, Pappas NS, Pascani K (1997) Dopamine D2 receptor availability in opiate-dependent subjects before and after naloxone-precipitated withdrawal. Neuropsychopharmacology 16(2): 174–182 84. Koeneke A, Ponce G, Troya-Valesca J, Palomo T, Hoenicka J (2020) Ankyrin repeat and kinase domain containing 1 gene, and addiction vulnerability. Int J Mol Sci 21(7): 2516 85. Jung Y, Montel RA, Shen PH, Mash DC, Goldman D (2019) Assessment of the association of D2 dopamine receptor gene and reported allele frequencies with alcohol use

475

disorders: a systematic review and metaanalysis. JAMA Netw Open 2(11):e1914940 86. Grzywacz A, Chmielowiec J, Chmielowiec K, Mroczek B, Masiak J, Suchanecka A, SipakSzmigiel O, Szumilas K, Trybek G (2019) The ankyrin repeat and kinase domain containing 1 gene polymorphism (ANKK1 Taq1A) and personality traits in addicted subjects. Int J Environ Res Public Health 16(15): 2687 87. Garrido E, Palomo T, Ponce G, GarciaConsuegra I, Jimenez-Arriero MA, Hoenicka J (2011) The ANKK1 protein associated with addictions has nuclear and cytoplasmic localization and shows a differential response of Ala239Thr to apomorphine. Neurotox Res 20(1):32–39 88. Kosten TR, Wu G, Huang W, Harding MJ, Hamon SC, Lappalainen J, Nielsen DA (2013) Pharmacogenetic randomized trial for cocaine abuse: disulfiram and dopamine beta-hydroxylase. Biol Psychiatry 73(3): 219–224 89. Spellicy CJ, Kosten TR, Hamon SC, Harding MJ, Nielsen DA (2013) ANKK1 and DRD2 pharmacogenetics of disulfiram treatment for cocaine abuse. Pharmacogenet Genomics 23(7):333–340 90. Kampangkaew JP, Spellicy CJ, Nielsen EM, Harding MJ, Ye A, Hamon SC, Kosten TR, Nielsen DA (2019) Pharmacogenetic roles of dopamine transporter (SLC6A3) variation on response to disulfiram treatment for cocaine addiction. Am J Addict 28(4):311–317 91. Thomas PS Jr, Nielsen EM, Spellicy CJ, Harding MJ, Ye A, Patriquin M, Hamon SC, Kosten TR, Nielsen DA (2021) The OPRD1 rs678849 variant influences outcome of disulfiram treatment for cocaine dependency in methadone-maintained patients. Psychiatr Genet 31(3):88–94 92. Zhang X, Nielsen DA, Domingo CB, Shorter DI, Nielsen EM, Kosten TR (2019) Pharmacogenetics of dopamine beta-hydroxylase in cocaine dependence therapy with doxazosin. Addict Biol 24(3):531–538 93. Balldin JI, Berggren UC, Lindstedt G (1992) Neuroendocrine evidence for reduced dopamine receptor sensitivity in alcoholism. Alcohol Clin Exp Res 16(1):71–74 94. Patriquin MA, Bauer IE, Soares JC, Graham DP, Nielsen DA (2015) Addiction pharmacogenetics: a systematic review of the genetic variation of the dopaminergic system. Psychiatr Genet 25(5):181–193 95. Balldin J, Berggren U, Lindstedt G, Sundkler A (1993) Further neuroendocrine evidence

476

David P. Graham et al.

for reduced D2 dopamine receptor function in alcoholism. Drug Alcohol Depend 32(2): 159–162 96. Lawford BR, Young RM, Rowell JA, Qualichefski J, Fletcher BH, Syndulko K, Ritchie T, Noble EP (1995) Bromocriptine in the treatment of alcoholics with the D2 dopamine receptor A1 allele. Nat Med 1(4): 337–341 97. Lucht MJ, Kuehn KU, Schroeder W, Armbruster J, Abraham G, Schattenberg A, Gaensicke M, Barnow S, Tretzel H, Herrmann FH, Freyberger HJ (2001) Influence of the dopamine D2 receptor (DRD2) exon 8 genotype on efficacy of tiapride and clinical outcome of alcohol withdrawal. Pharmacogenetics 11:647–653 98. Ooteman W, Naassila M, Koeter MW, Verheul R, Schippers GM, Houchi H, Daoust M, van den Brink W (2009) Predicting the effect of naltrexone and acamprosate in alcohol-dependent patients using genetic indicators. Addict Biol 14:328–337 99. Zastrozhin MS, Skryabin VY, Mrioshkin SS, Bryun EA, Sychev DA (2019) Pharmacogenetics of alcohol addiction: current perspectives. Appl Clin Genet 12:131–140 100. Choi DW, Viseskul V (1988) Opioids and non-opioid enantiomers selectively attenuate N-methyl-D-aspartate neurotoxicity on cortical neurons. Eur J Pharmacol 155:27–35 101. Dole VP, Nyswander M (1965) A medical treatment for diacetylmorphine (heroin) addiction. A clinical trial with methadone hydrochloride. JAMA 193:646–650 102. Kreek MJ, Bart G, Lilly C, LaForge KS, Nielsen DA (2005) Pharmacogenetics and human molecular genetics of opiate and cocaine addictions and their treatments. Pharmacol Rev 57:1–26 103. Doehring A, Hentig N, Graff J, Salamat S, Schmidt M, Geisslinger G, Harder S, Lotsch J (2009) Genetic variants altering dopamine D2 receptor expression or function modulate the risk of opiate addiction and the dosage requirements of methadone substitution. Pharmacogenet Genomics 19:407–414 104. Crettol S, Besson J, Croquette-Krokar M, Hammig R, Gothuey I, Monnat M, Deglon JJ, Preisig M, Eap CB (2008) Association of dopamine and opioid receptor genetic polymorphisms with response to methadone maintenance treatment. Prog NeuroPsychopharmacol Biol Psychiatry 32: 1722–1727 105. Burns JA, Kroll DS, Feldman DE, Liu CK, Manza P, Wiers CE, Volkow ND, Wang GJ

(2019) Molecular imaging of opioid and dopamine systems: insights into the pharmacogenetics of opioid use disorders. Front Psych 10:626 106. Hung CC, Chiou MH, Huang BH, Hsieh YW, Hsieh TJ, Huang CL, Lane HY (2011) Impact of genetic polymorphisms in ABCB1, CYP2B6, OPRM1, ANKK1 and DRD2 genes on methadone therapy in Han Chinese patients. Pharmacogenomics 12(11): 1525–1533 107. Zahari Z, Lee CS, Ibrahim MA, Musa N, Yasin MAM, Lee YY, Tan SC, Mohamad N, Ismail R (2020) Influence of DRD2 polymorphisms on the clinical outcomes of opioid-dependent patients on methadone maintenance therapy. J Pharm Bioallied Sci 12(Suppl 2):S787–S803. https://doi.org/ 10.4103/jpbs.JPBS_248_19 108. Van Tol HH, Wu CM, Guan HC, Ohara K, Bunzow JR, Civelli O, Kennedy J, Seeman P, Niznik HB, Jovanovic V (1992) Multiple dopamine D4 receptor variants in the human population. Nature 358:149–152 109. Paquet C, Portella AK, Moore S, Ma Y, Dagher A, Meaney MJ, Kennedy JL, Levitan RD, Silveira PP, Dube L (2021) Dopamine D4 receptor gene polymorphism (DRD4 VNTR) moderates real-world behavioural response to the food retail environment in children. BMC Public Health 21(1):145 110. Asghari V, Sanyal S, Buchwaldt S, Paterson A, Jovanovic V, Van Tol HH (1995) Modulation of intracellular cyclic AMP levels by different human dopamine D4 receptor variants. J Neurochem 65:1157–1165 111. Hutchison KE, McGeary J, Smolen A, Bryan A, Swift RM (2002) The DRD4 VNTR polymorphism moderates craving after alcohol consumption. Health Psychol 21:139–146 112. Daurio AM, Dechaine SL, Modabbernia A, Leggio L (2020) Parsing out the role of dopamine D4 receptor gene (DRD4) on alcoholrelated phenotypes: a meta-analysis and systematic review. Addict Biol 25(3):e12770 € Nursal AF, Yedikule 113. Uysal MA, Sever U, Smoking Cessation Study Group, Pehllvan S (2019) Dopamine D4 receptor gene exon III VNTR variant influences smoking status in Turkish population. Noro Psikiyatr Ars 56(4):248–252 114. Botticelli L, Di Bonaventura EM, Del Bello F, Giorgioni G, Piergentili A, Romano A, Quaglia W, Cifani C, Di Bonaventura MVM (2020) Underlying susceptibility to eating disorders and drug abuse: genetic and

Addiction Pharmacogenetics pharmacological aspects of dopamine D4 receptors. Nutrients 12(8):2288 115. Hutchison KE, Wooden A, Swift RM, Smolen A, McGeary J, Adler L, Paris L (2003) Olanzapine reduces craving for alcohol: a DRD4 VNTR polymorphism by pharmacotherapy interaction. Neuropsychopharmacology 28:1882–1888 116. Hutchison KE, Swift R, Rohsenow DJ, Monti PM, Davidson D, Almeida A (2001) Olanzapine reduces urge to drink after drinking cues and a priming dose of alcohol. Psychopharmacology 155:27–34 117. Kaufman S, Friedman S (1965) DopamineBeta-Hydroxylase. Pharmacol Rev 17: 71–100 118. Weinshilboum RM (1978) Serum dopamine beta-hydroxylase. Pharmacol Rev 30: 133–166 119. Arnsten AF (2000) Stress impairs prefrontal cortical function in rats and monkeys: role of dopamine D1 and norepinephrine alpha-1 receptor mechanisms. Prog Brain Res 126: 183–192 120. Wang J, John Y, Barbas H (2021) Pathways for contextual memory: the primate hippocampal pathway to anterior cingulate cortex. Cereb Cortex 31(3):1807–1826 121. Arnsten AF (2000) Through the looking glass: differential noradrenergic modulation of prefrontal cortical function. Neural Plast 7:133–146 122. Borderies N, Bornert P, Gilardeau S, Bouret S (2020) Pharmacological evidence for the implication of noradrenaline in effort. PLoS Biol 18(10):e3000793 123. Gong P, Liu J, Li S, Zhou X (2014) Dopamine beta-hydroxylase gene modulates individuals’ empathic ability. Soc Cogn Affect Neurosci 9(9):1341–1345 124. Stewart LC, Klinman JP (1988) Dopamine beta-hydroxylase of adrenal chromaffin granules: structure and function. Annu Rev Biochem 57:551–592 125. Weinshilboum RM, Raymond FA, Elveback LR, Weidman WH (1973) Serum dopaminebeta-hydroxylase activity: sibling-sibling correlation. Science 181:943–945 126. Oxenstierna G, Edman G, Iselius L, Oreland L, Ross SB, Sedvall G (1986) Concentrations of monoamine metabolites in the cerebrospinal fluid of twins and unrelated individuals--a genetic study. J Psychiatr Res 20:19–29 127. Wei J, Ramchand CN, Hemmings GP (1997) Possible control of dopamine betahydroxylase via a codominant mechanism

477

associated with the polymorphic (GT)n repeat at its gene locus in healthy individuals. Hum Genet 99(1):52–55 128. Wei J, Xu HM, Ramchand CN, Hemmings GP (1997) Is the polymorphic microsatellite repeat of the dopamine beta-hydroxylase gene associated with biochemical variability of the catecholamine pathway in schizophrenia? Biol Psychiatry 41(7):762–767 129. Cubells JF, van Kammen DP, Kelley ME, Anderson GM, O’Connor DT, Price LH, Malison R, Rao PA, Kobayashi K, Nagatsu T, Gelernter J (1998) Dopamine beta-hydroxylase: two polymorphisms in linkage disequilibrium at the structural gene DBH associate with biochemical phenotypic variation. Hum Genet 102(5):533–540 130. Cubells JF, Kranzler HR, McCance-Katz E, Anderson GM, Malison RT, Price LH, Gelernter J (2000) A haplotype at the DBH locus, associated with low plasma dopamine beta-hydroxylase activity, also associates with cocaine-induced paranoia. Mol Psychiatry 5(1):56–63 131. Zabetian CP, Buxbaum SG, Elston RC, Kohnke MD, Anderson GM, Gelernter J, Cubells JF (2003) The structure of linkage disequilibrium at the DBH locus strongly influences the magnitude of association between diallelic markers and plasma dopamine beta-hydroxylase activity. Am J Hum Genet 72(6):1389–1400 132. Zabetian CP, Anderson GM, Buxbaum SG, Elston RC, Ichinose H, Nagatsu T, Kim KS, Kim CH, Malison RT, Gelernter J, Cubells JF (2001) A quantitative-trait analysis of human plasma-dopamine beta-hydroxylase activity: evidence for a major functional polymorphism at the DBH locus. Am J Hum Genet 68(2):515–522 133. Bhaduri N, Mukhopadhyay K (2008) Correlation of plasma dopamine beta-hydroxylase activity with polymorphisms in DBH gene: A study on Eastern Indian population. Cell Mol Neurobiol 28(3):343–350 134. Preuss UW, Wurst FM, Ridinger M, Rujescu D, Fehr C, Koller G, Bondy B, Wodarz N, Soyka M, Zill P (2013) Association of functional DBH genetic variants with alcohol dependence risk and related depression and suicide attempt phenotypes: results from a large multicenter association study. Drug Alcohol Depend 133(2):459–467 135. Xie X, Xu L, Liu H, Chen W, Zhuang D, Zhang J, Duan S, Zhou W (2013) Positive association between--1021TT genotype of dopamine beta hydroxylase gene and

478

David P. Graham et al.

progressive behavior of injection heroin users. Neurosci Lett 541:258–262 136. Li Y, Zhu Y, Lai J, Shi X, Chen Y, Zhang J, Wei S (2020) Association studies of dopamine synthesis and metabolism genes with multiple phenotypes of heroin dependence. BMC Med Genet 21(1):157 137. Leko MB, Perkovic´ MN, Klepac N, Sˇtrac DSˇ, Borovecˇki F, Pivac N, Hof PR, Sˇimic´ G (2020) Relationships of cerebrospinal fluid Alzheimer’s disease biomarkers and COMT, DBH, and MAOB single nucleotide polymorphisms. J Alzheimers Dis 73(1):135–145 138. Kang S, Bi M, Du S, Jiao Q, Jiang H (2018) Association of the rs1611115 polymorphism in DBH gene with Parkinson’s disease: a meta-analysis. Neurol Sci 39(12):2085–2089 139. Martell BA, Mitchell E, Poling J, Gonsai K, Kosten TR (2005) Vaccine pharmacotherapy for the treatment of cocaine dependence. Biol Psychiatry 58:158–164 140. Kosten TR, Domingo CB, Hamon SC, Nielsen DA (2013) DBH gene as predictor of response in a cocaine vaccine clinical trial. Neurosci Lett 541:29–33 141. Spanagel R, Herz A, Shippenberg TS (1992) Opposing tonically active endogenous opioid systems modulate the mesolimbic dopaminergic pathway. Proc Natl Acad Sci U S A 89: 2046–2050 142. Di Chiara G, Imperato A (1988) Opposite effects of mu and kappa opiate agonists on dopamine release in the nucleus accumbens and in the dorsal caudate of freely moving rats. J Pharmacol Exp Ther 244:1067–1080 143. Spanagel R, Herz A, Shippenberg TS (1990) The effects of opioid peptides on dopamine release in the nucleus accumbens: an in vivo microdialysis study. J Neurochem 55: 1734–1740 144. Bart G, Heilig M, LaForge KS, Pollak L, Leal SM, Ott J, Kreek MJ (2004) Substantial attributable risk related to a functional mu-opioid receptor gene polymorphism in association with heroin addiction in Central Sweden. Mol Psychiatry 9:547–549 145. Deb I, Chakraborty J, Gangopadhyay PK, Choudhury SR, Das S (2010) Singlenucleotide polymorphism (A118G) in exon 1 of OPRM1 gene causes alteration in downstream signaling by mu-opioid receptor and may contribute to the genetic risk for addiction. J Neurochem 112:486–496 146. Bergen AW, Kokoszka J, Peterson R, Long JC, Virkkunen M, Linnoila M, Goldman D (1997) Mu opioid receptor gene variants: lack

of association with alcohol dependence. Mol Psychiatry 2:490–494 147. Bond C, LaForge KS, Tian M, Melia D, Zhang S, Borg L, Gong J, Schluger J, Strong JA, Leal SM, Tischfield JA, Kreek MJ, Yu L (1998) Single-nucleotide polymorphism in the human mu opioid receptor gene alters beta-endorphin binding and activity: possible implications for opiate addiction. Proc Natl Acad Sci USA 95:9608–9613 148. Zhang Y, Wang D, Johnson AD, Papp AC, Sadee W (2005) Allelic expression imbalance of human mu opioid receptor (OPRM1) caused by variant A118G. J Biol Chem 280: 32618–32624 149. Kroslak T, Laforge KS, Gianotti RJ, Ho A, Nielsen DA, Kreek MJ (2007) The single nucleotide polymorphism A118G alters functional properties of the human mu opioid receptor. J Neurochem 103:77–87 150. Wand GS, McCaul M, Yang X, Reynolds J, Gotjen D, Lee S, Ali A (2002) The mu-opioid receptor gene polymorphism (A118G) alters HPA axis activation induced by opioid receptor blockade. Neuropsychopharmacology 26: 106–114 151. Hernandez-Avila CA, Wand G, Luo X, Gelernter J, Kranzler HR (2003) Association between the cortisol response to opioid blockade and the Asn40Asp polymorphism at the mu-opioid receptor locus (OPRM1). Am J Med Genet B Neuropsychiatr Genet 118B:60–65 152. Halikere A, Popova D, Scarnati MS, Hamod A, Swerdel MR, Moore JC, Tischfield JA, Hart RP, Pang ZP (2020) Addiction associated N40D mu-opioid receptor variant modulates synaptic function in human neurons. Mol Psychiatry 25(7):1406–1419 153. Kakko J, von Wachenfeldt J, Svanborg JD, Lidstrom J, Barr CS, Heilig M (2008) Mood and neuroendocrine response to a chemical stressor, metyrapone, in buprenorphinemaintained heroin dependence. Biol Psychiatry 63(2):172–177 154. Bart G, Kreek MJ, Ott J, LaForge KS, Proudnikov D, Pollak L, Heilig M (2005) Increased attributable risk related to a functional mu-opioid receptor gene polymorphism in association with alcohol dependence in Central Sweden. Neuropsychopharmacology 30:417–422 155. Rouvinen-Lagerstrom N, Lahti J, Alho H, Kovanen L, Aalto M, Partonen T, Silander K, Sinclair D, Raikkonen K, Eriksson JG, Palotie A, Koskinen S, Saarikoski ST (2013) Mu-opioid receptor gene (OPRM1)

Addiction Pharmacogenetics polymorphism A118G: lack of association in finnish populations with alcohol dependence or alcohol consumption. Alcohol Alcohol 48: 519–525 156. Cupic B, Stefulj J, Zapletal E, Matosic A, Bordukalo-Niksic T, Cicin-Sain L, Gabrilovac J (2013) Opioid system genes in alcoholism: A case-control study in Croatian population. Neuropeptides 47(5):315–319 157. Ray LA, Hutchison KE (2004) A polymorphism of the mu-opioid receptor gene (OPRM1) and sensitivity to the effects of alcohol in humans. Alcohol Clin Exp Res 28: 1789–1795 158. Tan EC, Tan CH, Karupathivan U, Yap EP (2003) Mu opioid receptor gene polymorphisms and heroin dependence in Asian populations. Neuroreport 14:569–572 159. Zhou H, Rentsch CT, Cheng Z, Kember RL, Nunez YZ, Sherva RM, Tate JP, Dao C, Xu K, Polimanti R, Farrer LA, Justice AC, Kranzler HR, Gelernter J, Veterans Affairs Million Veteran Program (2020) Association of OPRM1 functional coding variant with opioid use disorder: a genome-wide association study. JAMA Psychiat 77(10):1072–1080 160. Clarke TK, Crist RC, Kampman KM, Dackis CA, Pettinati HM, O’Brien CP, Oslin DW, Ferraro TN, Lohoff FW, Berrettini WH (2013) Low frequency genetic variants in the μ-opioid receptor (OPRM1) affect risk for addiction to heroin and cocaine. Neurosci Lett 542:71–75 161. Arias AJ, Armeli S, Gelernter J, Covault J, Kallio A, Karhuvaara S, Koivisto T, Makela R, Kranzler HR (2008) Effects of opioid receptor gene variation on targeted nalmefene treatment in heavy drinkers. Alcohol Clin Exp Res 32(7):1159–1166 162. Wachman EM, Hayes MJ, Brown MS, Paul J, Harvey-Wilkes K, Terrin N, Huggins GS, Aranda JV, Davis JM (2013) Association of OPRM1 and COMT single-nucleotide polymorphisms with hospital length of stay and treatment of neonatal abstinence syndrome. JAMA 309(17):1821–1827 163. Wachman EM, Hayes MJ, Sherva R, Brown MS, Davis JM, Farrer LA, Nielsen DA (2015) Variations in opioid receptor genes in neonatal abstinence syndrome. Drug Alcohol Depend 155:253–259 164. Levran O, Peles E, Randesi M, Shu X, Ott J, Shen PH, Adelson M, Kreek MJ (2013) Association of genetic variation in pharmacodynamic factors with methadone dose required for effective treatment of opioid addiction. Pharmacogenomics 14(7):755–768

479

165. Wang SC, Tsou HH, Chen CH, Chen YT, Ho IK, Hsiao CF, Chou SY, Lin YF, Fang KC, Huang CL, Su LW, Fang YC, Liu ML, Wu HY, Lin KM, Liu SC, Kuo HW, Chiang IC, Chen AC, Tian JN, Liu YL (2012) Genetic polymorphisms in the opioid receptor mu1 gene are associated with changes in libido and insomnia in methadone maintenance patients. Eur Neuropsychopharmacol 22(10):695–703 166. Crist RC, Doyle GA, Nelson EC, Degenhardt L, Martin NG, Montgomery GW, Saxon AJ, Ling W, Berrettini WH (2018) A polymorphism in the OPRM1 30 -untranslated region is associated with methadone efficacy in treating opioid dependence. Pharmacogenomics J 18(1):173–179 167. Lutfy K, Cowan A (2004) Buprenorphine: a unique drug with complex pharmacology. Curr Neuropharmacol 2(4):395–402 168. Muriel J, Margarit C, Planelles B, Serralta MJ, Puga C, Inda MD, Cutillas E, Morales D, Horga JF, Peiro AM (2018) OPRM1 influence on and effectiveness of an individualized treatment plan for prescription opioid use disorder patients. Ann N Y Acad Sci 1425(1):82–93 169. Filbey FM, Claus E, Audette AR, Niculescu M, Banich MT, Tanabe J, Du YP, Hutchison KE (2008) Exposure to the taste of alcohol elicits activation of the mesocorticolimbic neurocircuitry. Neuropsychopharmacology 33:1391–1401 170. Ramchandani VA, Umhau J, Pavon FJ, RuizVelasco V, Margas W, Sun H, Damadzic R, Eskay R, Schoor M, Thorsell A, Schwandt ML, Sommer WH, George DT, Parsons LH, Herscovitch P, Hommer D, Heilig M (2011) A genetic determinant of the striatal dopamine response to alcohol in men. Mol Psychiatry 16:809–817 171. Baron SA, Testa FM, Gintzler AR (1985) Simultaneous quantitation of norepinephrine, dopamine and serotonin in brain during and following chronic naltrexone administration. Brain Res 340:192–198 172. Volpicelli JR, O’Brien CP, Alterman AI, Hayashida M (1990) Naltrexone and the treatment of alcohol dependence: initial observations. In: Reid LD (ed) Opioids, bulimia, and alcohol abuse & addiction. Springe, New York, pp 195–214 173. Volpicelli JR, Alterman AI, Hayashida M, O’Brien CP (1992) Naltrexone in the treatment of alcohol dependence. Arch Gen Psychiatry 49:876–880 174. O’Malley SS, Jaffe AJ, Chang G, Schottenfeld RS, Meyer RE, Rounsaville B (1992)

480

David P. Graham et al.

Naltrexone and coping skills therapy for alcohol dependence. A controlled study. Arch Gen Psychiatry 49:881–887 175. Davidson D, Swift R, Fitz E (1996) Naltrexone increases the latency to drink alcohol in social drinkers. Alcohol Clin Exp Res 20: 732–739 176. Oslin D, Liberto JG, O’Brien J, Krois S, Norbeck J (1997) Naltrexone as an adjunctive treatment for older patients with alcohol dependence. Am J Geriatr Psychiatry 5: 324–332 177. Anton RF, Moak DH, Waid LR, Latham PK, Malcolm RJ, Dias JK (1999) Naltrexone and cognitive behavioral therapy for the treatment of outpatient alcoholics: results of a placebocontrolled trial. Am J Psychiatry 156: 1758–1764 178. Chick J, Anton R, Checinski K, Croop R, Drummond DC, Farmer R, Labriola D, Marshall J, Moncrieff J, Morgan MY, Peters T, Ritson B (2000) A multicentre, randomized, double-blind, placebo-controlled trial of naltrexone in the treatment of alcohol dependence or abuse. Alcohol Alcohol 35: 587–593 179. Heinala P, Alho H, Kiianmaa K, Lonnqvist J, Kuoppasalmi K, Sinclair JD (2001) Targeted use of naltrexone without prior detoxification in the treatment of alcohol dependence: a factorial double-blind, placebo-controlled trial. J Clin Psychopharmacol 21:287–292 180. Monti PM, Rohsenow DJ, Swift RM, Gulliver SB, Colby SM, Mueller TI, Brown RA, Gordon A, Abrams DB, Niaura RS, Asher MK (2001) Naltrexone and cue exposure with coping and communication skills training for alcoholics: treatment process and 1-year outcomes. Alcohol Clin Exp Res 25: 1634–1647 181. Guardia J, Caso C, Arias F, Gual A, Sanahuja J, Ramirez M, Mengual I, Gonzalvo B, Segura L, Trujols J, Casas M (2002) A double-blind, placebo-controlled study of naltrexone in the treatment of alcohol-dependence disorder: results from a multicenter clinical trial. Alcohol Clin Exp Res 26:1381–1387 182. Latt NC, Jurd S, Houseman J, Wutzke SE (2002) Naltrexone in alcohol dependence: a randomised controlled trial of effectiveness in a standard clinical setting. Med J Aust 176: 530–534 183. Balldin J, Berglund M, Borg S, Mansson M, Bendtsen P, Franck J, Gustafsson L, Halldin J, Nilsson LH, Stolt G, Willander A (2003) A 6-month controlled naltrexone study: combined effect with cognitive behavioral therapy

in outpatient treatment of alcohol dependence. Alcohol Clin Exp Res 27:1142–1149 184. Kiefer F, Jahn H, Tarnaske T, Helwig H, Briken P, Holzbach R, Kampf P, Stracke R, Baehr M, Naber D, Wiedemann K (2003) Comparing and combining naltrexone and acamprosate in relapse prevention of alcoholism: a double-blind, placebo-controlled study. Arch Gen Psychiatry 60:92–99 185. Kranzler HR, Modesto-Lowe V, Van Kirk J (2000) Naltrexone vs. nefazodone for treatment of alcohol dependence. A placebocontrolled trial. Neuropsychopharmacology 22:493–503 186. Krystal JH, Cramer JA, Krol WF, Kirk GF, Rosenheck RA, Veterans Affairs Naltrexone Cooperative Study Group (2001) Naltrexone in the treatment of alcohol dependence. N Engl J Med 345:1734–1739 187. Swift RM, Whelihan W, Kuznetsov O, Buongiorno G, Hsuing H (1994) Naltrexone-induced alterations in human ethanol intoxication. Am J Psychiatry 151: 1463–1467 188. Volpicelli JR, Watson NT, King AC, Sherman CE, O’Brien CP (1995) Effect of naltrexone on alcohol "high" in alcoholics. Am J Psychiatry 152:613–615 189. Drobes DJ, Anton RF, Thomas SE, Voronin K (2004) Effects of naltrexone and nalmefene on subjective response to alcohol among nontreatment-seeking alcoholics and social drinkers. Alcohol Clin Exp Res 28:1362–1370 190. Anton RF, Drobes DJ, Voronin K, DurazoAvizu R, Moak D (2004) Naltrexone effects on alcohol consumption in a clinical laboratory paradigm: temporal effects of drinking. Psychopharmacology 173:32–40 191. Jaffe SB, Sobieszczyk S, Wardlaw SL (1994) Effect of opioid antagonism on betaendorphin processing and proopiomelanocortin-peptide release in the hypothalamus. Brain Res 648:24–31 192. O’Malley S, Krishnan-Sarin S, Farren C, Sinha R, Kreek MJ (2002) Naltrexone decreases craving and alcohol selfadministration in alcohol-dependent subjects and activates the hypothalamo-pituitary-adrenocortical axis. Psychopharmacology 160: 19–29 193. Ray LA, Hutchison KE (2007) Effects of naltrexone on alcohol sensitivity and genetic moderators of medication response: a double-blind placebo-controlled study. Arch Gen Psychiatry 64:1069–1077 194. McGeary JE, Monti PM, Rohsenow DJ, Tidey J, Swift R, Miranda R Jr (2006) Genetic

Addiction Pharmacogenetics moderators of naltrexone’s effects on alcohol cue reactivity. Alcohol Clin Exp Res 30: 1288–1296 195. Oslin DW, Berrettini W, Kranzler HR, Pettinati H, Gelernter J, Volpicelli JR, O’Brien CP (2003) A functional polymorphism of the mu-opioid receptor gene is associated with naltrexone response in alcoholdependent patients. Neuropsychopharmacology 28:1546–1552 196. Anton RF, Oroszi G, O’Malley S, Couper D, Swift R, Pettinati H, Goldman D (2008) An evaluation of mu-opioid receptor (OPRM1) as a predictor of naltrexone response in the treatment of alcohol dependence: results from the Combined Pharmacotherapies and Behavioral Interventions for Alcohol Dependence (COMBINE) study. Arch Gen Psychiatry 65:135–144 197. Ray LA, Bujarski S, Chin PF, Miotto K (2012) Pharmacogenetics of naltrexone in Asian Americans: a randomized placebocontrolled laboratory study. Neuropsychopharmacology 37:445–455 198. Setiawan E, Pihl RO, Cox SM, Gianoulakis C, Palmour RM, Benkelfat C, Leyton M (2011) The effect of naltrexone on alcohol’s stimulant properties and self-administration behavior in social drinkers: influence of gender and genotype. Alcohol Clin Exp Res 35: 1134–1141 199. Kim SG, Kim CM, Choi SW, Jae YM, Lee HG, Son BK, Kim JG, Choi YS, Kim HO, Kim SY, Oslin DW (2009) A micro opioid receptor gene polymorphism (A118G) and naltrexone treatment response in adherent Korean alcohol-dependent patients. Psychopharmacology 201:611–618 200. Coller JK, Cahill S, Edmonds C, Farquharson AL, Longo M, Minniti R, Sullivan T, Somogyi AA, White JM (2011) OPRM1 A118G genotype fails to predict the effectiveness of naltrexone treatment for alcohol dependence. Pharmacogenet Genomics 21:902–905 201. Oroszi G, Anton RF, O’Malley S, Swift R, Pettinati H, Couper D, Yuan Q, Goldman D (2009) OPRM1 Asn40Asp predicts response to naltrexone treatment: a haplotype-based approach. Alcohol Clin Exp Res 33:383–393 202. Ray LA, Bujarski S, Chin PF, Miotto K (2011) Pharmacogenetics of naltrexone in Asian Americans: a randomized placebocontrolled laboratory study. Neuropsychopharmacology 37(2):445–455 203. O’Malley SS, Robin RW, Levenson AL, GreyWolf I, Chance LE, Hodgkinson CA, Romano D, Robinson J, Meandzija B, Stillner

481

V (2008) Naltrexone alone and with sertraline for the treatment of alcohol dependence in Alaska natives and non-natives residing in rural settings: a randomized controlled trial. Alcohol Clin Exp Res 32(7):1271–1283 204. Tidey JW, Monti PM, Rohsenow DJ, Gwaltney CJ, Miranda R, McGeary JE, MacKillop J, Swift RM, Abrams DB, Shiffman S (2008) Moderators of naltrexone’s effects on drinking, urge, and alcohol effects in nontreatment-seeking heavy drinkers in the natural environment. Alcohol Clin Exp Res 32(1): 58–66 205. Gelernter J, Gueorguieva R, Kranzler HR, Zhang H, Cramer J, Rosenheck R, Krystal JH, V.A.C.S.S. Group (2007) Opioid receptor gene (OPRM1, OPRK1, and OPRD1) variants and response to naltrexone treatment for alcohol dependence: results from the VA Cooperative Study. Alcohol Clin Exp Res 31(4):555–563 206. Gomes I, Jordan BA, Gupta A, Trapaidze N, Nagy V, Devi LA (2000) Heterodimerization of mu and delta opioid receptors: A role in opiate synergy. J Neurosci 20:RC110 207. Hirose N, Murakawa K, Takada K, Oi Y, Suzuki T, Nagase H, Cools AR, Koshikawa N (2005) Interactions among mu- and deltaopioid receptors, especially putative delta1and delta2-opioid receptors, promote dopamine release in the nucleus accumbens. Neuroscience 135:213–225 208. Gupta A, Mulder J, Gomes I, Rozenfeld R, Bushlin I, Ong E, Lim M, Maillet E, Junek M, Cahill CM, Harkany T, Devi LA (2010) Increased abundance of opioid receptor heteromers after chronic morphine administration. Sci Signal 3:ra54 209. Zhang H, Kranzler HR, Yang BZ, Luo X, Gelernter J (2008) The OPRD1 and OPRK1 loci in alcohol or drug dependence: OPRD1 variation modulates substance dependence risk. Mol Psychiatry 13:531–543 210. Mayer P, Rochlitz H, Rauch E, Rommelspacher H, Hasse HE, Schmidt S, Hollt V (1997) Association between a delta opioid receptor gene polymorphism and heroin dependence in man. Neuroreport 8: 2547–2550 211. Levran O, Londono D, O’Hara K, Nielsen DA, Peles E, Rotrosen J, Casadonte P, Linzy S, Randesi M, Ott J, Adelson M, Kreek MJ (2008) Genetic susceptibility to heroin addiction: a candidate gene association study. Genes Brain Behav 7:720–729 212. Levran O, Randesi M, Adelson M, Kreek MJ (2021) OPRD1 SNPs associated with opioid

482

David P. Graham et al.

addiction are cis-eQTLs for the phosphatase and actin regulator 4 gene, PHACTR4, a mediator of cytoskeletal dynamics. Transl Psychiatry 11(1):316 213. Crist RC, Clarke TK, Ang A, Ambrose-Lanci LM, Lohoff FW, Saxon AJ, Ling W, Hillhouse MP, Bruce RD, Woody G, Berrettini WH (2013) An Intronic variant in OPRD1 predicts treatment outcome for opioid dependence in African-Americans. Neuropsychopharmacology 38(10): 2003–2010 214. Nelson EC, Lynskey MT, Heath AC, Wray N, Agrawal A, Shand FL, Henders AK, Wallace L, Todorov AA, Schrage AJ, Madden PA, Degenhardt L, Martin NG, Montgomery GW (2014) Association of OPRD1 polymorphisms with heroin dependence in a large case-control series. Addict Biol 19(1): 111–121 215. Filliol D, Ghozland S, Chluba J, Martin M, Matthes HW, Simonin F, Befort K, Gaveriaux-Ruff C, Dierich A, LeMeur M, Valverde O, Maldonado R, Kieffer BL (2000) Mice deficient for delta- and mu-opioid receptors exhibit opposing alterations of emotional responses. Nat Genet 25: 195–200 216. Zhu Y, King MA, Schuller AG, Nitsche JF, Reidl M, Elde RP, Unterwald E, Pasternak GW, Pintar JE (1999) Retention of supraspinal delta-like analgesia and loss of morphine tolerance in delta opioid receptor knockout mice. Neuron 24:243–252 217. Nitsche JF, Schuller AG, King MA, Zengh M, Pasternak GW, Pintar JE (2002) Genetic dissociation of opiate tolerance and physical dependence in delta-opioid receptor-1 and preproenkephalin knock-out mice. J Neurosci 22:10906–10913 218. Unterwald EM, Cox BM, Kreek MJ, Cote TE, Izenwasser S (1993) Chronic repeated cocaine administration alters basal and opioid-regulated adenylyl cyclase activity. Synapse 15:33–38 219. Crist RC, Phillips KA, Furnari MA, Moran LM, Doyle GA, McNicholas LF, Cornish JW, Kampman KM, Preston KL, Berrettini WH (2019) Replication of the pharmacogenetic effect of rs678849 on buprenorphine efficacy in African-Americans with opioid use disorder. Pharmacogenomics J 19(3): 260–268 220. Kabli N, Fan T, O’Dowd BF, George SR (2014) Mu-delta opioid receptor heteromerspecific signaling in the striatum and hippocampus. Biochem Biophys Res Commun 450(1):906–911

221. Tejeda HA, Shippenberg TS, Henriksson R (2012) The dynorphin/kappa-opioid receptor system and its role in psychiatric disorders. Cell Mol Life Sci 69:857–896 222. Claye LH, Maisonneuve IM, Yu J, Ho A, Kreek MJ (1997) Local perfusion of dynorphin A 1-17 reduces extracellular dopamine levels in the nucleus accumbens. NIDA Res Monogr 174:113 223. Zhang Y, Butelman ER, Schlussman SD, Ho A, Kreek MJ (2004) Effect of the kappa opioid agonist R-84760 on cocaine-induced increases in striatal dopamine levels and cocaine-induced place preference in C57BL/ 6J mice. Psychopharmacology 173:146–152 224. Unterwald EM, Rubenfeld JM, Kreek MJ (1994) Repeated cocaine administration upregulates kappa and mu, but not delta, opioid receptors. Neuroreport 5:1613–1616 225. Simonin F, Valverde O, Smadja C, Slowe S, Kitchen I, Dierich A, Le Meur M, Roques BP, Maldonado R, Kieffer BL (1998) Disruption of the kappa-opioid receptor gene in mice enhances sensitivity to chemical visceral pain, impairs pharmacological actions of the selective kappa-agonist U-50,488H and attenuates morphine withdrawal. EMBO J 17: 886–897 226. McLaughlin JP, Marton-Popovici M, Chavkin C (2003) Kappa opioid receptor antagonism and prodynorphin gene disruption block stress-induced behavioral responses. J Neurosci 23:5674–5683 227. Haile CN, Kosten TA, Kosten TR (2008) Pharmacogenetic treatments for drug addiction: alcohol and opiates. Am J Drug Alcohol Abuse 34:355–381 228. Xuei X, Dick D, Flury-Wetherill L, Tian HJ, Agrawal A, Bierut L, Goate A, Bucholz K, Schuckit M, Nurnberger J Jr, Tischfield J, Kuperman S, Porjesz B, Begleiter H, Foroud T, Edenberg HJ (2006) Association of the kappa-opioid system with alcohol dependence. Mol Psychiatry 11:1016–1024 229. Wang SC, Tsou HH, Chung RH, Chang YS, Fang CP, Chen CH, Ho K, Kuo HW, Liu SC, Shih YH (2014) The association of genetic polymorphisms in the κ-opioid receptor 1 gene with body weight, alcohol use, and withdrawal symptoms in patients with methadone maintenance. J Clin Psychopharmacol 34(2):205–211 230. Zhang H, Kranzler HR, Yang BZ, Luo X, Gelernter J (2008) The OPRD1 and OPRK1 loci in alcohol or drug dependence: OPRD1 variation modulates substance dependence risk. Mol Psychiatry 13(5): 531–543

Addiction Pharmacogenetics 231. Xu K, Seo D, Hodgkinson C, Hu Y, Goldman D, Sinha R (2013) A variant on the kappa opioid receptor gene (OPRK1) is associated with stress response and related drug craving, limbic brain activation and cocaine relapse risk. Transl Psychiatry 3(8): e292 232. Nielsen DA, Hamon S, Kosten TR (2013) The K-opioid receptor gene as a predictor of response in a cocaine vaccine trial. Psychiatr Genet 23:225–232 233. Lesch KP, Bengel D, Heils A, Sabol SZ, Greenberg BD, Petri S, Benjamin J, Muller CR, Hamer DH, Murphy DL (1996) Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science 274(5292): 1527–1531 234. Heils A, Teufel A, Petri S, Stober G, Riederer P, Bengel D, Lesch KP (1996) Allelic variation of human serotonin transporter gene expression. J Neurochem 66(6): 2621–2624 235. Hu XZ, Lipsky RH, Zhu G, Akhtar LA, Taubman J, Greenberg BD, Xu K, Arnold PD, Richter MA, Kennedy JL, Murphy DL, Goldman D (2006) Serotonin transporter promoter gain-of-function genotypes are linked to obsessive-compulsive disorder. Am J Hum Genet 78(5):815–826 236. Praschak-Rieder N, Kennedy J, Wilson AA, Hussey D, Boovariwala A, Willeit M, Ginovart N, Tharmalingam S, Masellis M, Houle S, Meyer JH (2007) Novel 5-HTTLPR allele associates with higher serotonin transporter binding in putamen: a [(11) C] DASB positron emission tomography study. Biol Psychiatry 62(4):327–331 237. Fox E, Zougkou K, Ridgewell A, Garner K (2011) The serotonin transporter gene alters sensitivity to attention bias modification: evidence for a plasticity gene. Biol Psychiatry 70(11):1049–1054 238. Wankerl M, Miller R, Kirschbaum C, Hennig J, Stalder T, Alexander N (2014) Effects of genetic and early environmental risk factors for depression on serotonin transporter expression and methylation profiles. Transl Psychiatry 4(6):e402 239. Graham DP, Helmer DA, Harding MJ, Kosten TR, Petersen NJ, Nielsen DA (2013) Serotonin transporter genotype and mild traumatic brain injury independently influence resilience and perception of limitations in veterans. J Psychiatr Res 47(6):835–842 240. Gonza´lez-Giraldo Y, Forero DA (2020) Association between resilience and a functional polymorphism in the serotonin transporter

483

(SCL6A4) gene: A meta-analysis. EXCLI J 19:1174–1183 241. Caspi A, Moffitt TE (2006) Geneenvironment interactions in psychiatry: joining forces with neuroscience. Nat Rev Neurosci 7(7):583–590 242. Zhou F, Ren J, Lu X, Ma S, Wu C (2021) Gene-environment interaction: a variable selection perspective. Methods Mol Biol 2212:191–223 243. Hariri AR (2009) The neurobiology of individual differences in complex behavioral traits. Annu Rev Neurosci 32:225–247 244. Uher R (2009) The role of genetic variation in the causation of mental illness: an evolution-informed framework. Mol Psychiatry 14(12):1072–1082 245. Alexander N, Wankerl M, Hennig J, Miller R, Zankert S, Steudte-Schmiedgen S, Stalder T, Kirschbaum C (2014) DNA methylation profiles within the serotonin transporter gene moderate the association of 5-HTTLPR and cortisol stress reactivity. Transl Psychiatry 4(9):e443 246. Rauch SAM, King AP, Liberzon I, Sripada RK (2017) Changes in salivary cortisol during psychotherapy for posttraumatic stress disorder: a pilot study in 30 veterans. J Clin Psychiatry 78(5):599–603 247. Liu PZ, Nusslock R (2018) How stress gets under the skin: early life adversity and glucocorticoid receptor epigenetic regulation. Curr Genomics 19(8):653–664 248. Spies M, Nasser A, Ozenne B, Jensen PS, Knudsen GM, Fisher PM (2020) Common HTR2A variants and 5-HTTLPR are not associated with human in vivo serotonin 2A receptor levels. Hum Brain Mapp 41(16): 4518–4528 249. Torrens M, Fonseca F, Mateu G, Farre M (2005) Efficacy of antidepressants in substance use disorders with and without comorbid depression. A systematic review and metaanalysis. Drug Alcohol Depend 78(1):1–22 250. Mannelli P, Patkar AA, Murray HW, Certa K, Peindl K, Mattila-Evenden M, Berrettini WH (2005) Polymorphism in the serotonin transporter gene and response to treatment in African American cocaine and alcohol-abusing individuals. Addict Biol 10(3):261–268 251. Nielsen DA, Harding MJ, Hamon SC, Huang W, Kosten TR (2012) Modifying the role of serotonergic 5-HTTLPR and TPH2 variants on disulfiram treatment of cocaine addiction: a preliminary study. Genes Brain Behav 11(8):1001–1008

484

David P. Graham et al.

252. Nagendra SN, Shetty KT, Subhash MN, Udaya HB, Pradhan N (1993) Effect of disulfiram administration on brain tryptophan, serotonin and peripheral tryptophan content. Neurochem Int 22(1):31–36 253. Nilsson GE, Tottmar O (1989) Effects of disulfiram and coprine on rat brain tryptophan hydroxylation in vivo. Neurochem Res 14(6):537–540 254. Fukumori R, Minegishi A, Satoh T, Kitagawa H, Yanaura S (1980) Changes in the serotonin and 5-hydroxyindoleacetic acid contents in rat brain after ethanol and disulfiram treatments. Eur J Pharmacol 61:199–202 255. Bauer IE, Graham DP, Soares JC, Nielsen DA (2015) Serotonergic gene variation in substance use pharmacotherapy: a systematic review. Pharmacogenomics 16(11): 1307–1314 256. Johnson BA, Roache JD, Javors MA, DiClemente CC, Cloninger CR, Prihoda TJ, Bordnick PS, Ait-Daoud N, Hensler J (2000) Ondansetron for reduction of drinking among biologically predisposed alcoholic patients: A randomized controlled trial. JAMA 284(8):963–971 257. Johnson BA, Ait-Daoud N, Seneviratne C, Roache JD, Javors MA, Wang XQ, Liu L, Penberthy JK, DiClemente CC, Li MD (2011) Pharmacogenetic approach at the serotonin transporter gene as a method of reducing the severity of alcohol drinking. Am J Psychiatry 168(3):265–275 258. Seneviratne C, Johnson BA (2012) Serotonergic genomic biomarker for quantitative assessment of ondansetron treatment response in alcoholics. Front Psych 3:23 259. Kranzler HR, Armeli S, Tennen H, Covault J, Feinn R, Arias AJ, Pettinati H, Oncken C (2011) A double-blind, randomized trial of sertraline for alcohol dependence: moderation by age of onset [corrected] and 5-hydroxytryptamine transporter-linked promoter region genotype. J Clin Psychopharmacol 31(1):22–30 260. Kranzler HR, Armeli S, Tennem H (2012) Post-treatment outcomes in a double-blind, randomized trial of sertraline for alcohol dependence. Alcohol Clin Exp Res 36(4): 739–744 261. Kranzler HR, McKay JR (2012) Personalized treatment of alcohol dependence. Curr Psychiatry Rep 14(5):486–493 262. Kenna GA, Zywiak WH, McGeary JE, Leggio L, McGeary C, Wang S, Grenga A, Swift RM (2009) A within-group design of

nontreatment seeking 5-HTTLPR genotyped alcohol-dependent subjects receiving ondansetron and sertraline. Alcohol Clin Exp Res 33(2):315–323 263. Crist RC, Li J, Doyle GA, Gilbert A, Dechairo BM, Berrettini WH (2018) Pharmacogenetic analysis of opioid dependence treatment dose and dropout rate. Am J Drug Alcohol Abuse 44(4):431–440 264. Cooper JR, Melcer I (1961) The enzymatic oxidation of tryptophan to 5-hydroxytryptophan in the biosynthesis of serotonin. J Pharmacol Exp Ther 132: 265–268 265. Zill P, Bu¨ttner A, Eisenmenger W, Moller HJ, Ackenheil M, Bondy B (2007) Analysis of tryptophan hydroxylase I and II mRNA expression in the human brain: A postmortem study. J Psychiatr Res 41(1–2): 168–173 266. Park S, Kim Y, Lee J, Lee JY, Kim H, Lee S, Oh CM (2021) A systems biology approach to investigating the interaction between serotonin synthesis by tryptophan hydroxylase and the metabolic homeostasis. Int J Mol Sci 22(5):2452 267. Waider J, Araragi N, Gutknecht L, Lesch KP (2011) Tryptophan hydroxylase-2 (TPH2) in disorders of cognitive control and emotion regulation: a perspective. Psychoneuroendocrinology 36(3):393–405 268. Deneris E, Gaspar P (2018) Serotonin neuron development: shaping molecular and structural identities. Wiley Interdiscip Rev Dev Biol 7(1). https://doi.org/10.1002/ wdev.301 269. Nielsen DA, Barral S, Proudnikov D, Kellogg S, Ho A, Ott J, Kreek MJ (2008) TPH2 and TPH1: association of variants and interactions with heroin addiction. Behav Genet 38(2):133–150 270. Lim JE, Pinsonneault J, Sadee W, Saffen D (2007) Tryptophan hydroxylase 2 (TPH2) haplotypes predict levels of TPH2 mRNA expression in human pons. Mol Psychiatry 12(5):491–501 271. Huang CL, Chen PH, Lane HY, Ho IK, Chung CM (2021) Risk assessment for heroin use and craving score using polygenic risk score. J Pers Med 11(4):259 272. Berditchevski F, Odintsova E (2007) Tetraspanins as regulators of protein trafficking. Traffic 8(2):89–96 273. Gupta M, Neavin D, Liu D, Biernacka J, HallFlavin D, Bobo WV, Frye MA, Skime M, Jenkins GD, Batzler A, Kalari K, Matson W,

Addiction Pharmacogenetics Bhasin SS, Zhu H, Mushiroda T, Nakamura Y, Kubo M, Wang L, KaddurahDaouk R, Weinshilboum RM (2016) TSPAN5, ERICH3 and selective serotonin reuptake inhibitors in major depressive disorder: pharmacometabolomics-informed pharmacogenomics. Mol Psychiatry 21(12): 1714–1725 274. Ho MF, Zhang C, Zhang L, Wei L, Zhou Y, Moon I, Geske JR, Choi DS, Biernacka J, Frye M, Wen Z, Karpyak VM, Li H, Weinshilboum R (2020) TSPAN5 influences serotonin and kynurenine: pharmacogenomic mechanisms related to alcohol use disorder and acamprosate treatment response. Mol Psychiatr y. https://doi.org/10.1038/ s41380-020-0855-9 275. Crettol S, Digon P, Powell Golay K, Brawand M, Eap CB (2007) In vitro Pglycoprotein-mediated transport of (R)-, (S)-, (R,S)-methadone, LAAM and their main metabolites. Pharmacology 80:304–311 276. Dagenais C, Graff CL, Pollack GM (2004) Variable modulation of opioid brain uptake by P-glycoprotein in mice. Biochem Pharmacol 67:269–276 277. Levran O, O’Hara K, Peles E, Li D, Barral S, Ray B, Borg L, Ott J, Adelson M, Kreek MJ (2008) ABCB1 (MDR1) genetic variants are associated with methadone doses required for effective treatment of heroin dependence. Hum Mol Genet 17:2219–2227 278. Peng S, Jiang H, Du J, Lin S, Pan S, Yu S, Zhao M (2018) Methadone dosage and plasma levels, SNPs of OPRM1 gene and age of first drug use were associated with outcomes of methadone maintenance treatment. Front Genet 29:450 279. Coller JK, Barratt DT, Dahlen K, Loennechen MH, Somogyi AA (2006) ABCB1 genetic variability and methadone dosage requirements in opioid-dependent individuals. Clin Pharmacol Ther 80:682–690 280. Iwerson-Bergmann S, Plattner S, Hischke S, Mu¨ller A, Andresen-Streichert H, Jungen H, Erb R, Beer-Sandner B (2021) Brain/blood ratios of methadone and ABCB1 polymorphisms in methadone related deaths. Int J Legal Med 135:473–482 281. Barratt DT, Coller JK, Hallinan R, Byrne A, White JM, Foster DJ, Somogyi AA (2012) ABCB1 haplotype and OPRM1 118A > G genotype interaction in methadone maintenance treatment pharmacogenetics. Pharmacogenomics Pers Med 5:53–62 282. Guindalini C, Laranjeira R, Collier D, Messas G, Vallada H, Breen G (2008)

485

Dopamine-beta hydroxylase polymorphism and cocaine addiction. Behav Brain Funct 4:1 283. Shorter D, Nielsen DA, Huang W, Harding MJ, Hamon SC, Kosten TR (2013) Pharmacogenetic randomized trial for cocaine abuse: disulfiram and alpha-adrenoceptor gene variation. Eur Neuropsychopharmacol 23(11): 1401–1407 284. Deji C, Li Y, Chen Y, Jianghua L, Shuguang W (2021) Association study of catechol-omethyltransferase and Alpha-1-adrenergic receptor gene polymorphisms with multiple phenotypes of heroin use disorder. Neurosci Lett 748:135677 285. Lu B, Nagappan G, Guan X, Nathan PJ, Wren P (2013) BDNF-based synaptic repair as a disease-modifying strategy for neurodegenerative diseases. Nat Rev Neurosci 14:401–416 286. Rostami E, Krueger F, Zoubak S, Dal Monte O, Raymont V, Pardini M, Hodgkinson CA, Goldman D, Risling M, Grafman J (2011) BDNF polymorphism predicts general intelligence after penetrating traumatic brain injury. PLoS One 6:e27389 287. Li W, Zhou N, Yu Q, Li X, Yu Y, Sun S, Kou C, Chen da C, Xiu MH, Kosten TR, Zhang XY (2013) Association of BDNF gene polymorphisms with schizophrenia and clinical symptoms in a Chinese population. Am J Med Genet B Neuropsychiatr Genet 162:538–545 288. Cheng CY, Hong CJ, Yu YWY, Chen TJ, Wu HC, Tsai SJ (2005) Brain-derived neurotropic factor (Val66Met) genetic polymorphism is associated with substance abuse in males. Brain Res Mol 140:86–90 289. Greenwald MK, Steinmiller CL, Sliwerska E, Lundahl L, Burmeister M (2013) BDNF Val (66)met genotype is associated with drugseeking phenotypes in heroin-dependent individuals; a pilot study. Addict Biol 18: 836–845 290. de Cid R, Fonseca F, Gratacos M, Gutierrez F, Martin-Santos R, Estivill X, Torrens M (2008) BDNF variability in opioid addicts and response to methadone treatment: preliminary findings. Genes Brain Behav 7:515–522 291. Pezet S, McMahon SB (2006) Neurotrophins: mediators and modulators of pain. Annu Rev Neurosci 29:507–538 292. Levran O, Peles E, Hamon S, Randesi M, Zhao C, Zhang B, Adelson M, Kreek MJ (2012) Nerve growth factor beta polypeptide (NGFB) genetic variability: association with the methadone dose required for effective

486

David P. Graham et al.

maintenance treatment. Pharmacogenomics J 12:319–327 293. Gainetdinov RR, Premont RT, Bohn LM, Lefkowitz RJ, Caron MG (2004) Desensitization of G protein-coupled receptors and neuronal functions. Annu Rev Neurosci 27: 107–144 294. Oneda B, Crettol S, Bochud M, Besson J, Croquette-Krokar M, Hammig R, Monnat M, Preisig M, Eap CB (2011) β-Arrestin2 influences the response to methadone in opioid-dependent patients. Pharmacogenomics J 11:258–266 295. Karavidha KK, Burmeister M, Greenwald MK (2021) β-Arrestin 2 (ARRB2) polymorphism is associated with adverse consequences of chronic heroin use. Am J Addict 30:351–357 296. Ferrari A, Coccia CP, Bertolini A, Sternieri E (2004) Methadone--metabolism, pharmacokinetics and interactions. Pharmacol Res 50: 551–559 297. Reynolds KK, Ramey-Hartung B, Jortani SA (2008) The value of CYP2D6 and OPRM1 pharmacogenetic testing for opioid therapy. Clin Lab Med 28:581–598 298. Perez de los Cobos J, Sinol N, Trujols J, del Rio E, Banuls E, Luquero E, Menoyo A, Queralto JM, Baiget M, Alvarez E (2007) Association of CYP2D6 ultrarapid metabolizer genotype with deficient patient satisfaction regarding methadone maintenance treatment. Drug Alcohol Depend 89: 190–194 299. Crettol S, Deglon JJ, Besson J, CroquetteKrokar M, Hammig R, Gothuey I, Monnat M, Eap CB (2006) ABCB1 and cytochrome P450 genotypes and phenotypes: influence on methadone plasma levels and response to treatment. Clin Pharmacol Ther 80:668–681 300. Dagostino C, Allegri M, Napolioni V, D’Agnelli S, Bignami E, Mutti A, van Schalk RHN (2018) CYP2D6 genotype can help to predict effectiveness and safety during opioid treatment for chronic lower back pain: results from a retrospective study in an Italian cohort. Pharmacogenomics Pers Med 11: 179–191 301. Umukoro NN, Aruldhas BW, Rossos R, Pawale D, Renschler JS, Sadhasivam S (2021) Pharmacogenomics of oxycodone: a narrative literature review. Pharmacogenomics 22(5):275–290 302. Levran O, Peles E, Hamon S, Randesi M, Adelson M, Kreek MJ (2013) CYP2B6 SNPs are associated with methadone dose required

for effective treatment of opioid addiction. Addict Biol 18:709–716 303. Crettol S, Deglon JJ, Besson J, CroquetteKrokkar M, Gothuey I, Hammig R, Monnat M, Huttemann H, Baumann P, Eap CB (2005) Methadone enantiomer plasma levels, CYP2B6, CYP2C19, and CYP2C9 genotypes, and response to treatment. Clin Pharmacol Ther 78:593–604 304. Dobrinas M, Crettol S, Oneda B, Lahyani R, Rotger M, Choong E, Lubomirov R, Csajka C, Eap CB (2013) Contribution of CYP2B6 alleles in explaining extreme (S)methadone plasma levels: a CYP2B6 gene resequencing study. Pharmacogenet Genomics 23:84–93 305. Ahmad T, Sabet S, Primerano D, RichardsWaugh L, Rankin GO (2017) Tell-tale SNPs: the role of CYP2B6 in methadone fatalities. J Anal Toxicol 41(4):325–333 306. Morley KC, Luquin N, Baillie A, Fraser I, Trent RJ, Dore G, Phung N, Haber PS (2018) Moderation of baclofen response by a BABAB receptor polymorphism: results from the BacALD randomized controlled trial. Addiction 113(12):2205–2213 307. He Q, Mendez M, LaPointe MC (2002) Regulation of the human brain natriuretic peptide gene by GATA-4. Am J Physiol Endocrinol Metab 283:E50–E57 308. Kovacs GL (2003) Natriuretic peptides in alcohol withdrawal: central and peripheral mechanisms. Curr Med Chem 10:2559–2576 309. Kiefer F, Witt SH, Frank J, Richter A, Treutlein J, Lemenager T, Nothen MM, Cichon S, Batra A, Berner M, Wodarz N, Zimmermann US, Spanagel R, Wiedemann K, Smolka MN, Heinz A, Rietschel M, Mann K (2011) Involvement of the atrial natriuretic peptide transcription factor GATA4 in alcohol dependence, relapse risk and treatment response to acamprosate. Pharmacogenomics J 11:368–374 310. Karpyak VM, Winham SJ, Biernacka JM, Cunningham JM, Lewis KA, Geske JR, Colby CL, Abulseoud OA, Hall-Flavin DK, Loukianova LL, Schneekloth TD, Frye MA, Heit JA, Mrazek DA (2012) Association of GATA4 sequence variation with alcohol dependence. Addict Biol 19(2):312–315 311. Treutlein J, Cichon S, Ridinger M, Wodarz N, Soyka M, Zill P, Maier W, Moessner R, Gaebel W, Dahmen N, Fehr C, Scherbaum N, Steffens M, Ludwig KU, Frank J, Wichmann HE, Schreiber S, Dragano N, Sommer WH, LeonardiEssmann F, Lourdusamy A, Gebicke-HaerterP, Wienker TF, Sullivan PF, Nothen MM,

Addiction Pharmacogenetics Kiefer F, Spanagel R, Mann K, Rietschel M (2009) Genome-wide association study of alcohol dependence. Arch Gen Psychiatry 66:773–784 312. Kranzler HR, Covault J, Feinn R, Armeli S, Tennen H, Arias AJ, Gelernter J, Pond T, Oncken C, Kampman KM (2014) Topiramate treatment for heavy drinkers: moderation by a GRIK1 polymorphism. Am J Psychiatry 171:445–452 313. Kranzler HR, Wetherill R, Feinn R, Pond T, Gelernter J, Covault J (2014) Posttreatment effects of topiramate treatment for heavy drinking. Alcohol Clin Exp Res 38: 3017–3023 314. Lieberman R, Jensen KP, Clinton K, Levine ES, Kranzler HR, Covault J (2020) Molecular correlates of topiramate and GRIK1 rs2832407 genotype in pluripotent stem cell-derived neural cultures. Alcohol Clin Exp Res 44(8):1561–1570 315. Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, Boers GJ, den Heijer M, Kluijtmans LA, van den Heuvel LP, Rozen R (1995) A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet 10:111–113 316. van der Put NM, Steegers-Theunissen RP, Frosst P, Trijbels FJ, Eskes TK, van den Heuvel LP, Mariman EC, den Heyer M, Rozen R, Blom HJ (1995) Mutated methylenetetrahydrofolate reductase as a risk factor for spina bifida. Lancet 346:1070–1071 317. Harisha PN, Devi BI, Christopher R, Kruthika-Vinod TP (2010) Impact of 5,10methylenetetrahydrofolate reductase gene polymorphism on neural tube defects. J Neurosurg Pediatr 6:364–367 318. Bleich S, Carl M, Bayerlein K, Reulbach U, Biermann T, Hillemacher T, Bo¨nsch D, Kornhuber J (2005) Evidence of increased homocysteine levels in alcoholism: the Franconian alcoholism research studies (FARS). Alcohol Clin Exp Res 29(3):334–336 319. Lutz UC, Batra A, Kolb W, Machicao F, Maurer S, Ko¨hnke MD (2006) Methylenetetrahydrofolate reductase C677Tpolymorphism and its association with alcohol withdrawal seizure. Alcohol Clin Exp Res 30(12):1966–1971 320. Saffroy R, Benyamina A, Pham P, Marill C, Karila L, Reffas M, Debuire B, Reynaud M, Lemoine A (2008) Protective effect against alcohol dependence of the thermolabile variant of MTHFR. Drug Alcohol Depend 96(1–2):30–36

487

321. Benyamina A, Saffroy R, Blecha L, Pham P, Karila L, Debuire B, Lemoine A, Reynaud M (2009) Association between MTHFR 677CT polymorphism and alcohol dependence according to Lesch and Babor typology. Addict Biol 14(4):503–505 322. Spellicy CJ, Kosten TR, Hamon SC, Harding MJ, Nielsen DA (2012) The MTHFR C677T variant is associated with responsiveness to disulfiram treatment for cocaine dependency. Front Psych 3:109 323. Lin J, Haffner MC, Zhang Y, Lee BH, Brennen WN, Britton J, Kachhap SK, Shim JS, Liu JO, Nelson WG, Yegnasubramanian S, Carducci MA (2011) Disulfiram is a DNA demethylating agent and inhibits prostate cancer cell growth. Prostate 71:333–343 324. Anier K, Malinovskaja K, Aonurm-Helm A, Zharkovsky A, Kalda A (2010) DNA methylation regulates cocaine-induced behavioral sensitization in mice. Neuropsychopharmacology 35:2450–2461 325. Pipes GC, Creemers EE, Olson EN (2006) The myocardin family of transcriptional coactivators: versatile regulators of cell growth, migration, and myogenesis. Genes Dev 20: 1545–1556 326. Nielsen DA, Ji F, Yuferov V, Ho A, Chen A, Levran O, Ott J, Kreek MJ (2008) Genotype patterns that contribute to increased risk for or protection from developing heroin addiction. Mol Psychiatry 13:417–428 327. Fonseca F, Gratacos M, Escaramis G, De Cid R, Martin-Santos R, Fernandez-Espejo E, Estivill X, Torrens M (2010) Response to methadone maintenance treatment is associated with the MYOCD and GRM6 genes. Mol Diagn Ther 14:171–178 328. Miyoshi J, Takai Y (2007) Nectin and nectinlike molecules: biology and pathology. Am J Nephrol 27(6):590–604 329. Fang CP, Liu TH, Chung RH, Tsou HH, Kuo HW, Wang SC, Liu CC, Liu SC, Chen ACH, Liu YL (2020) Genetic variants in NECTIN4 encoding an adhesion molecule are associated with continued opioid use. PLoS One 15(6):e0234549 330. Mitrovic I, Margeta-Mitrovic M, Bader S, Stoffel M, Jan LY, Basbaum AI (2003) Contribution of GIRK2-mediated postsynaptic signaling to opiate and alpha 2-adrenergic analgesia and analgesic sex differences. Proc Natl Acad Sci U S A 100:271–276 331. Reyes S, Fu Y, Double K, Thompson L, Kirik D, Paxinos G, Halliday GM (2012) GIRK2 expression in dopamine neurons of

488

David P. Graham et al.

the substantia nigra and ventral tegmental area. J Comp Neurol 520:2591–2607 332. Clarke TK, Laucht M, Ridinger M, Wodarz N, Rietschel M, Maier W, Lathrop M, Lourdusamy A, Zimmermann US, Desrivieres S, Schumann G (2011) KCNJ6 is associated with adult alcohol dependence and involved in gene x early life stress interactions in adolescent alcohol drinking. Neuropsychopharmacology 36: 1142–1148 333. Webb CK, McCudden CR, Willard FS, Kimple RJ, Siderovski DP, Oxford GS (2005) D2 dopamine receptor activation of potassium channels is selectively decoupled by Galphaspecific GoLoco motif peptides. J Neurochem 92:1408–1418 334. Minami M, Satoh M (1995) Molecular biology of the opioid receptors: structures, functions and distributions. Neurosci Res 23: 121–145 335. Ikeda K, Kobayashi T, Kumanishi T, Niki H, Yano R (2000) Involvement of G-proteinactivated inwardly rectifying K (GIRK) channels in opioid-induced analgesia. Neurosci Res 38:113–116 336. Nishizawa D, Nagashima M, Katoh R, Satoh Y, Tagami M, Kasai S, Ogai Y, Han W, Hasegawa J, Shimoyama N, Sora I, Hayashida M, Ikeda K (2009) Association between KCNJ6 (GIRK2) gene polymorphisms and postoperative analgesic requirements after major abdominal surgery. PLoS One 4:e7060 337. Lotsch J, Pruss H, Veh RW, Doehring A (2010) A KCNJ6 (Kir3.2, GIRK2) gene polymorphism modulates opioid effects on analgesia and addiction but not on pupil size. Pharmacogenet Genomics 20:291–297 338. Kamarajan C, Pandey AK, Chorlian DB, Manz N, Stimus AT, Edenberg HJ, Wetherill L, Schuckit M, Wang JC, Kuperman S, Kramer J, Tischfield JA, Porjesz B (2017) A KCNJ6 gene polymorphism modulates theta oscillations during reward processing. Int J Psychol 115:13–23 339. Klimesch W (1999) EEG alpha and theta oscillations reflect cognitive and memory performance: a review and analysis. Brain Res Brain Res Rev 29:169–195 340. Basar E, Schurmann M, Sakowitz O (2001) The selectively distributed theta system: functions. Int J Psychophysiol 39:197–212 341. Kreek MJ, Nielsen DA, LaForge KS (2004) Genes associated with addiction: alcoholism, opiate, and cocaine addiction. NeuroMolecular Med 5:85–108

342. Kleiderer E, Rice J, Conquest V, Williams J (1945) Pharmaceutical activities at the I.G. Farbenindustrie plant, Hochst am Main. (Commerce, O. o. t. P. B. D. o., Ed.), Washington, DC 343. Mason BJ, Heyser CJ (2010) Acamprosate: a prototypic neuromodulator in the treatment of alcohol dependence. CNS Neurol Disord Drug Targets 9:23–32 344. Kent CN, Park C, Lindsley CW (2020) Classics in chemical neuroscience: baclofen. ACS Chem Neurosci 11(12):1740–1755 345. Jarvik ME, Caskey NH, Wirshing WC, Madsen DC, Iwamoto-Schaap PN, Elins JL, Eisenberger NI, Olmstead RE (2000) Bromocriptine reduces cigarette smoking. Addiction 95:1173–1183 346. Bentley KW, Hardy DG (1967) Novel analgesics and molecular rearrangements in the morphine-thebaine group. 3. Alcohols of the 6,14-endo-ethenotetrahydrooripavine series and derived analogs of N-allylnormorphine and -norcodeine. J Am Chem Soc 89: 3281–3292 347. Rolly G, Versichelen L (1976) Buprenorphine as postoperative analgesic. Acta Anaesthesiol Belg 27:183–186 348. Jasinski DR, Pevnick JS, Griffith JD (1978) Human pharmacology and abuse potential of the analgesic buprenorphine: a potential agent for treating narcotic addiction. Arch Gen Psychiatry 35:501–516 ˜ ones DS, Duconge J 349. Seguı´ HA, Melin K, Quin (2020) A review of the pharmacogenomics of buprenorphine for the treatment of opioid use disorder. J Transl Genet Genom 4: 263–277 350. Barth KS, Malcolm RJ (2010) Disulfiram: an old therapeutic with new applications. CNS Neurol Disord Drug Targets 9:5–12 351. Shorter D, Kosten TR (2011) Novel pharmacotherapeutic treatments for cocaine addiction. BMC Med 9:119 352. Bourdelat-Parks BN, Anderson GM, Donaldson ZR, Weiss JM, Bonsall RW, Emery MS, Liles LC, Weinshenker D (2005) Effects of dopamine beta-hydroxylase genotype and disulfiram inhibition on catecholamine homeostasis in mice. Psychopharmacology 183:72–80 353. Schank JR, Ventura R, Puglisi-Allegra S, Alcaro A, Cole CD, Liles LC, Seeman P, Weinshenker D (2006) Dopamine betahydroxylase knockout mice have alterations in dopamine signaling and are hypersensitive to cocaine. Neuropsychopharmacology 31: 2221–2230

Addiction Pharmacogenetics 354. Carroll KM, Nich C, Ball SA, McCance E, Rounsavile BJ (1998) Treatment of cocaine and alcohol dependence with psychotherapy and disulfiram. Addiction 93:713–727 355. Carroll KM, Fenton LR, Ball SA, Nich C, Frankforter TL, Shi J, Rounsaville BJ (2004) Efficacy of disulfiram and cognitive behavior therapy in cocaine-dependent outpatients: a randomized placebo-controlled trial. Arch Gen Psychiatry 61:264–272 356. Petrakis IL, Carroll KM, Nich C, Gordon LT, McCance-Katz EF, Frankforter T, Rounsaville BJ (2000) Disulfiram treatment for cocaine dependence in methadonemaintained opioid addicts. Addiction 95: 219–228 357. McCance-Katz EF, Kosten TR, Jatlow P (1998) Disulfiram effects on acute cocaine administration. Drug Alcohol Depend 52: 27–39 358. Hameedi FA, Rosen MI, McCance-Katz EF, McMahon TJ, Price LH, Jatlow PI, Woods SW, Kosten TR (1995) Behavioral, physiological, and pharmacological interaction of cocaine and disulfiram in humans. Biol Psychiatry 37:560–563 359. Baker JR, Jatlow P, McCance-Katz EF (2007) Disulfiram effects on responses to intravenous cocaine administration. Drug Alcohol Depend 87:202–209 360. Carroll KM, Ziedonis D, O’Malley S, McCance-Katz E, Gordon L, Rounsaville B (1993) Pharmacologic interventions for alcohol- and cocaine-abusing individuals. Am J Addict 2:77–79 361. Higgins ST, Budney AJ, Bickel WK, Hughes JR, Foerg F (1993) Disulfiram therapy in patients abusing cocaine and alcohol. Am J Psychiatry 150:675–676 362. Kim J, Nguyen H (2021) Doxazosin. In: StatPearls [Internet]. StatPearls Publishing, Treasure Island. Bookshelf ID: NBK557459 363. Tsavaris N, Kopterides P, Kosmas C, Efthymiou A, Skopelitis H, Dimitrakopoulos A, Pagouni E, Pikazis D, Zis PV, Koufos C (2008) Gabapentin monotherapy for the treatment of chemotherapyinduced neuropathic pain: a pilot study. Pain Med 9(8):1209–1216 364. Arai YC, Matsubara T, Shimo K, Suetomi K, Nishihara M, Ushida T, Kobayashi K, Suzuki C, Kinoshita A, Kondo M, Matsubara S, Hayashi R, Tohyama Y, Nishida K, Arakawa M (2010) Low-dose gabapentin as useful adjuvant to opioids for neuropathic cancer pain when combined with

489

low-dose imipramine. J Anesth 24(3): 407–410 365. Lemos L, Fontes R, Flores S, Oliveira P, Almeida A (2010) Effectiveness of the association between carbamazepine and peripheral analgesic block with ropivacaine for the treatment of trigeminal neuralgia. J Pain Res 3: 201–212 366. Vranken JH (2009) Mechanisms and treatment of neuropathic pain. Cent Nerv Syst Agents Med Chem 9(1):71–78 367. Paradowski B, Bilinska M (2003) Gabapentin in the treatment of neuropathic pain in patients with type 2 diabetes mellitus. Pol Merkur Lekarski 15(85):61–64 368. Jean WH, Wu CC, Mok MS, Sun WZ (2005) Starting dose of gabapentin for patients with post-herpetic neuralgia—a dose response study. Acta Anaesthesiol Taiwanica 43(2): 73–77 369. Rice AS, Maton S (2001) Gabapentin in postherpetic neuralgia: A randomised, double blind, placebo controlled study. Pain 94(2): 215–224. https://doi.org/10.1016/S03043959(01)00407-9 370. Kukkar A, Bali A, Singh N, Jaggi AS (2013) Implications and mechanism of action of gabapentin in neuropathic pain. Arch Pharm Res 36(3):237–251 371. Leung JG, Hall-Flavin D, Nelson S, Schmidt KA, Schak KM (2015) The role of gabapentin in the management of alcohol withdrawal and dependence. Ann Pharmacother 49(8): 897–906 372. Isbell H, Wilker A, Eisenman AJ, Daingerfield M, Frank K (1948) Liability of addiction to 6-dimethylamino-4-4-diphenyl3-heptanone (methadon, amidone or 10820) in man; experimental addiction to methadon. Arch Intern Med (Chicago, Ill.: 1908) 82: 362–392 373. Dole VP, Nyswander ME, Kreek MJ (1966) Narcotic blockade--a medical technique for stopping heroin use by addicts. Trans Assoc Am Phys 79:122–136 374. Thorsell A (2013) The mu-opioid receptor and treatment response to naltrexone. Alcohol Alcohol 48:402–408 375. Bart G (2012) Maintenance medication for opiate addiction: the foundation of recovery. J Addict Dis 31:207–225 376. Bymaster FP, Calligaro DO, Falcone JF, Marsh RD, Moore NA, Tye NC, Seeman P, Wong DT (1996) Radioreceptor binding

490

David P. Graham et al.

profile of the atypical antipsychotic olanzapine. Neuropsychopharmacology 14:87–96 377. Bymaster FP, Rasmussen K, Calligaro DO, Nelson DL, DeLapp NW, Wong DT, Moore NA (1997) In vitro and in vivo biochemistry of olanzapine: a novel, atypical antipsychotic drug. J Clin Psychiatry 58(Suppl 10):28–36 378. Constenla M (2004) 5-HT3 receptor antagonists for prevention of late acute-onset emesis. Ann Pharmacother 38:1683–1691 379. Peters DH, Faulds D (1994) Tiapride. A review of its pharmacology and therapeutic potential in the management of alcohol dependence syndrome. Drugs 47:1010–1032 380. Lyseng-Williamson KA, Yang LP (2007) Topiramate: a review of its use in the treatment of epilepsy. Drugs 67:2231–2256

381. Parikh SK, Silberstein SD (2019) Current status of antiepileptic drugs as preventive migraine therapy. Curr Treat Options Neurol 21:16 382. Hurt RT, Edakkanambeth Varayil J, Ebbert JO (2014) New pharmacological treatments for the management of obesity. Curr Gastroenterol Rep 16:394 383. Shank RP, Gardocki JF, Streeter AJ, Maryanoff BE (2000) An overview of the preclinical aspects of topiramate: pharmacology, pharmacokinetics, and mechanism of action. Epilepsia 41(Suppl 1):S3–S9 384. Sachse C, Brockmoller J, Bauer S, Roots I (1997) Cytochrome P450 2D6 variants in a Caucasian population: allele frequencies and phenotypic consequences. Am J Hum Genet 60:284–295. April 2022

Chapter 17 Pharmacogenomics of Opioid Treatment for Pain Management Sarahbeth Howes, Alexandra R. Cloutet, Jaeyeon Kweon, Taylor L. Powell, Daniel Raza, Elyse M. Cornett, and Alan D. Kaye Abstract Pain affects approximately 100 million Americans. Pain harms quality of life and costs patients billions of dollars per year. Clinically, nonpharmacologic and pharmacologic therapies can alleviate acute and chronic pain suffering. Opioids are one type of medication used to manage pain. However, opioids can potentially create dependence and substance abuse, and the effects are not consistent in all patients. Pharmacogenomics is the study of the genome to understand the effects of drugs on individual patients based on their genetic information. Through pharmacogenomics, researchers can investigate genetic polymorphisms related to pain that maximize individual patient drug responses and minimize toxicity. This chapter discusses the pharmacogenomics of opioids to treat pain, including individual genetic differences to opioid treatments, opioid pharmacokinetics and pharmacodynamics, and the genetic polymorphisms associated with individual opioid medications. Key words Pharmacogenomics, Opioids, Pain management, Pharmacogenetics, Precision medicine, Acute pain, Chronic pain

1

Introduction One in three Americans suffers from untreated pain [1]. Pain is one of the most commonly reported physical symptoms and costs the United States up to $600 billion annually if left untreated [1– 3]. For example, the most common cause of disability worldwide is low back pain, with a prevalence of 9.4%, which increases in frequency as patients age [2]. Management and treatment of pain depend on the severity, source, and nature of its manifestation, which can be inflammatory, neuropathic, mechanical, or muscular [2, 3]. The complexities of pain transmission and its various mechanisms are even more challenging to treat because of differences in individual perceptions of pain [1]. Thus, pain management can be difficult for healthcare providers to treat.

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_17, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

491

492

Sarahbeth Howes et al.

In clinical practice, nonpharmacologic and pharmacologic therapies can alleviate acute and or chronic pain suffering. Opioids are widely used to treat pain [2]. However, their addiction potential and limited efficacy in some patients have spurred much debate on how they should be used. Pharmacogenomics, the study of how genes affect drug responses, helps identify which patients can have positive outcomes with opioid treatments. Pharmacogenomics has allowed researchers to investigate genetic polymorphisms associated with pain to maximize drug responses and minimize toxicity [4]. Subtle changes in DNA can explain individual variations in responses to pain, including race, ethnicity, gender, and sex, among many other qualities [3, 4]. This chapter discusses the pharmacogenomics of opioids to treat pain, including individual genetic differences to opioid treatments, opioid pharmacokinetics and pharmacodynamics, and the genetic polymorphisms associated with individual opioid medications.

2

Individual Genetic Differences to Opioids Within the general population, there is an inter-individual variation in opioid responsiveness. Each medication has a different level of opioid non-responsiveness. For instance, oral morphine does not effectively control pain in up to 30% of patients owing to insufficient analgesia or too many side effects [5]. Age, sex, ethnicity, and genetics are several clinical variables that have a role in variations in opioid responsiveness [3, 4]. Furthermore, individuals can be categorized into poor, intermediate, or efficient opioid metabolizers based on their alleles [1]. Because opioid-related pharmacokinetic and pharmacodynamic genes are highly polymorphic, several genetic variables have been revealed to play a crucial role in responses to opioids. Human genetic studies have elucidated interindividual heterogeneity in pain sensitivity. The analgesic response and protein expression can change because of various genetic differences. An individual’s genetic information can be used to categorize their metabolic capacity and, as a result, determine which medications may provide an optimal response for that individual. This is because specific enzymes primarily metabolize particular drugs, and specific analgesics bind more readily to certain transporters and receptors.

3

Opioid Pharmacodynamics Opioids exert their effects on the human body through a variety of receptors. The principal receptors include Mu (MOR), Kappa (KOR), and Delta (DOR). MOR, KOR, and DOR are G-protein

Pharmacogenomics of Opioid Treatment

493

coupled receptors that couple to inhibitory heterotrimeric G-proteins. An extracellular N-terminus has seven transmembrane helical twists, three extracellular and intracellular loops, and an intracellular C-terminus. This structure communicates with cell-signaling molecules to modify calcium influx and potassium efflux ion channels to block the release of pain-related neurotransmitters like substance P, glutamate, and calcitonin gene-related peptide (CGRP) [6]. Genetic variations in these signaling pathways are being identified and studied as a possible explanation for the individual differences we see in the clinical effects of various exogenous opioids. Additionally, some genes have been identified for their ability to upregulate levels of these receptors in tissues [6]. Mu receptors are divided into subtypes Mu 1 and Mu 2. They are principally found in the brainstem and medical thalamus. Mu 1 receptors produce analgesic and euphoric effects, while Mu 2 stimulation produces respiratory depression, pruritus, prolactin release, dependence, anorexia, and sedation. The endogenous ligands that preferentially bind these receptors are endorphins, enkephalins, and dynorphins. It has been established that there are single-nucleotide polymorphisms (SNPs) in the OPRM1 gene (codes for the MOR) that may change its functional properties and account for individual differences we see in the way human bodies respond to both endogenous and exogenous opioids [7]. A wellstudied example of an SNP in the OPRM1 gene that modifies the response to exogenous opioids is the 118 A > G substitution. The G allele causes an asparagine to aspartic acid substitution at residue 40. It has been shown that cancer patients receiving morphine for pain control required more of the drug to achieve the desired effect if they were homozygous for the 118 A > G substitution [8]. Another study found that this specific substitution was responsible for higher pain scores, more self-administered morphine, and increased incidence of nausea in post-cesarean delivery patients receiving intrathecal morphine [9]. Kappa receptors are found in the limbic system, brainstem, and spinal cord [10]. The primary endogenous ligands for this receptor are dynorphins. Similar to MOR, these receptors can produce potent downstream analgesic effects. However, the clinical utility of pure KOR agonists is limited due to their ability to produce adverse effects such as dysphoria [11]. Some studies have found that genetic variations in the OPRK1 gene may contribute to differences in analgesic properties among individuals, but the evidence is not conclusive for the OPRM1 gene [12]. Delta receptors are found mostly in the brain, and the primary endogenous receptors are endorphins and enkephalins. This family of receptors is not as well studied as the others, but new evidence suggests that they may play a role in managing chronic pain [13]. Some studies have also suggested that certain SNPs in OPRK1 may influence cold and heat pain sensitivity [14].

494

4

Sarahbeth Howes et al.

Opioid Pharmacokinetics Opioid metabolism is a highly complex process that differs considerably among individuals. It occurs mainly in the liver, where the most important enzymes for this process are made. All opioid drugs undergo metabolism to some extent, chiefly by the cytochrome P450 (CYP) system and to a lesser extent by the UDP-glucuronosyltransferases (UGTs). Both active metabolites and inactive metabolites are produced. It can be divided into phases 1 and 2. Phase 1 metabolism is a series of modification reactions chiefly carried out by CYP enzymes. Phase 2 metabolism refers to conjugation reactions, where opioid metabolites are conjugated to hydrophilic substances to facilitate excretion. The most important conjugation reaction is glucuronidation. Phase 1 of opioid metabolism mostly involves the two enzymes CYP3A4 and CYP2D6. CYP3A4 metabolizes over 50% of the drugs and is the primary metabolizer of both oxycodone and fentanyl. CYP2D6 is solely responsible for the metabolism of hydrocodone, codeine, and dihydrocodeine [15]. Research suggests considerable variability in certain individuals’ ability to metabolize these drugs based on which gene alleles of CYP3A4 and CYPD6 they possess. Specifically, over 100 variants of CYP2D6 have been identified [16]. Approximately, 5–10% of Caucasians have inactivating mutations in both alleles of the CYP2D6 gene, making them poor metabolizers of certain opioids. Conversely, 1–7% of Caucasians have alleles of the gene that make them rapid metabolizers. It is important to monitor patients like these for decreased drug efficacy and heightened toxicity [17]. Many well-studied genes code for enzymes that facilitate the clearance of opioid analgesics. UGT2B7 is the predominant enzyme that catalyzes morphine glucuronidation and forms its two major metabolites: morphine-6-glucuronide (M6G) and morphine-3-glucuronide (M3G). M6G is a potent analgesic, while M3G inhibits analgesia. Two allelic variants (the UGT2B7-840G and
79 alleles) have been connected to a significant reduction in the rate of morphine glucuronidation, with resulting accumulation of morphine and reduction in metabolite formation [15]. Genetic variations in opioid transport proteins have been described as potential causes of variability. The ATP binding cassette (ABC) superfamily of efflux transporters has been identified as locations for SNPs that affect the length of analgesia. The ABCB1 efflux transporter works at the blood-brain barrier and prevents the uptake of opioids into the central nervous system. Morphine and oxycodone are known substrates of this efflux transporter. The single-nucleotide polymorphism (SNP) in exon 26 of the MDR1 gene, C3435T, has been correlated with lower levels of P-glycoprotein expression. This could result in a population of individuals who clear morphine and oxycodone at a slower rate and therefore experience increased systemic analgesia [18].

Pharmacogenomics of Opioid Treatment

5

495

Opioids to Treat Pain

5.1

Codeine

Codeine is an opioid and prodrug of morphine used to treat pain, coughing, and diarrhea. It is found naturally in the sap of Papaver somniferum, the opium poppy. Codeine undergoes O-demethylation, mediated by CYP2D6, into morphine, accounting for about 10% of the overall elimination pathway of codeine. CYP2D6 is a highly polymorphic gene responsible for the metabolism of many commonly prescribed medications [19]. Because of the polymorphisms, CYP2D6 enzymatic activity is different among individuals with differing variants. Those with two loss-of-function variants are referred to as poor metabolizers. Poor metabolizers produce minimal plasma concentrations of morphine, and they are less likely to clinically benefit from the usage of codeine. In contrast to CYP2D6 poor metabolizers, studies show that increased conversion of codeine to morphine in CYP2D6 ultrarapid and normal metabolizers can lead to a toxic level of morphine concentration even at relatively low codeine doses [20]. There is a high degree of variability within the patients who are genotyped as normal metabolizers. These patients may develop symptoms similar to patients who are genotyped as ultrarapid metabolizers [21]. Also, it is shown that CYP2D6-guided care was shown to be clinically relevant improvement in pain control among the patients with poor and intermediate metabolizers [22]; therefore, CYP2D6 genotype testing may be a helpful tool for a physician’s chronic pain management plan.

5.2

Fentanyl

Fentanyl is a very potent opioid (approximately 75–100 times more potent than morphine) that has been used for anesthesia and pain relief for decades. After sublingual administration, fentanyl is absorbed throughout the body, transported to the target tissues by the P-glycoprotein, and encoded by ABCB1, an ATP-dependent efflux transporter. Fentanyl is predominantly metabolized in the liver by two isoforms of the cytochrome P450 (CYP) complex, CYP3A4 and CYP3A5, which mediate N-dealkylation to the inactive metabolite norfentanyl. Less than 1% is metabolized by alkyl hydroxylation, N-dealkylation, or amide hydrolysis to produce the inactive metabolites hydroxyfentanyl, hydroxynorfentanyl, and desproprionylfentanyl [23]. Since CYP3A4 is the main enzyme for the metabolism of fentanyl, variants in its coding gene have been studied. A study showed that CYP3A5*3 SNP reduced activity of the CYP3A5 enzyme, which was correlated with increased exposure to fentanyl [24]. However, a different study concluded that CYP3A genotypes (CYP3A4*22 and CYP3A5*3) did not affect plasma levels of fentanyl [25]. Other SNPs in the gene encoding for CYP3A4 and the transporter ABCB1 and SCLOB1B1 were studied, but there was no significant relationship

496

Sarahbeth Howes et al.

with exposure to fentanyl [26, 27]. Neither the metabolic pathways for fentanyl metabolism nor the polymorphism variations in its effects are well-defined. Further pharmacogenetic research regarding fentanyl treatment is needed to conclude the role of pharmacogenetics in the dosing of fentanyl. 5.3

Hydrocodone

Hydrocodone is an opioid receptor agonist and produces its effects by activating μ opioid receptors [28]. Hydrocodone activates κ and δ opioid receptors as its plasma concentration increases [29]. Hydrocodone is metabolized primarily in the liver via the cytochrome P450 enzymes CYP2D6 and CYP3A4. CYP3A4 metabolizes hydrocodone to the inactive metabolite norhydrocodone [30]. Like codeine metabolism, polymorphisms in CYP2D6 metabolism in different patients can lead to variations in plasma hydrocodone concentrations from the same dosage. Genetic variants in hydrocodone metabolism include ultra-rapid, extensive, and weak metabolizer phenotypes [31]. To develop a hydrocodone phenotype-specific dosing strategy, pharmacogenetics must be integrated with clinical pharmacokinetics for patient safety.

5.4

Hydromorphone

Hydromorphone is a hydrogenated ketone analog of morphine, and it is widely used as an alternative to morphine to relieve acute and chronic pain. Hydromorphone is an opioid agonist that binds to several opioid receptors, and its analgesic actions are accomplished mostly through its effect on the μ-opioid receptors. While many opioids are metabolized via CYP450 enzymes, hydromorphone is not metabolized via CYP450 enzymes. Like morphine, genetic polymorphisms of CYP450 should have little to no effect on the metabolism or clearance of hydromorphone.

5.5

Meperidine

Meperidine is a synthetic opioid that binds to μ opioid receptors in the CNS to produce analgesia. Meperidine was a more commonly prescribed analgesic in the United states in the past, but serious adverse side effects have led to restrictions in its use [32]. Like most opioids, meperidine is largely metabolized by the liver, and the adverse effect of meperidine is related to its metabolism in the liver. The major metabolic pathways of meperidine are hydrolysis by human liver carboxylesterase (hCE-1) to form the inactivate metabolite meperidine acid and N-demethylation via hepatic cytochrome P450 enzymes to form normeperidine. It is understood that CYP2B6, CYP3A4, and CYP2C19 are the primary P450 enzymes responsible for normeperidine formation in the liver [33]. Enzymes CYP2B6 and CYP2C19 are highly polymorphic; however, whether CYP2B6 and CYP2C19 polymorphism influences meperidine N-demethylation is not known. Further research is necessary to examine the impact of P450 polymorphisms and other factors on meperidine metabolism and normeperidine generation [34].

Pharmacogenomics of Opioid Treatment

497

5.6

Methadone

Methadone is a highly lipophilic μ-opioid agonist that is highly protein-bound. Opioid receptor mu 1 (OPRM1), coding for the μ-opioid receptor, has been the most widely studied gene [35], and OPRM1 polymorphism is found to influence the pharmacodynamics of methadone [36]. A study by Levran et al. focused on 11 genes encoding components of the opioidergic (OPRM1, POMC, and ARRB2), the dopaminergic (ANKK1 and DRD2), and the glutamatergic pathways (GRIN1 and GRIN2A); many variants of these SNPs were associated with methadone dose [37]. Moreover, because CYP enzymes are the most important contributors to methadone metabolism, their genetic polymorphism remains a highlight of the methadone pharmacogenetics study [38]. CYP2B6 is the most important CYP enzyme for methadone metabolism, and its expression is highly variable among individuals [39].

5.7

Morphine

Morphine is a natural alkaloid derived from the opium poppy Papaver somniferum, and it is the most widely used opioid to treat moderate to severe pain. Morphine is a full agonist at the μ-opioid receptor, and genetic variation of the μ-opioid receptors may contribute to interindividual differences in morphine consumption [40]. Numerous studies have studied the association between OPRM1 polymorphisms and response to morphine and its active metabolite, morphine-6-glucuronide. Additionally, CYP enzyme polymorphism is associated with the metabolization process of morphine [40]; however, morphine is not greatly metabolized by CYP450 enzymes in the liver. Therefore, genetic polymorphisms of CYP450 should have little effect on the metabolism or clearance of morphine.

5.8

Oxycodone

Oxycodone is a semisynthetic opioid with agonistic properties on mu, kappa, and delta-type opioid receptors, with the strongest affinity on the μ-opioid receptors [41]. Andreassen et al. reported in a cross-sectional study of 450 cancer patients genotyped for CYP2D6 that poor metabolizers, extensive metabolizers, and ultra-rapid metabolizers did not have differing oxycodone and noroxymorphone plasma concentration [42]. It is believed that CYP2D6 genotypes cause significant differences in pharmacokinetics without producing major pharmacodynamic differences or clinical consequences [43, 44].

5.9

Oliceridine

Oliceridine is a μ-opioid receptor-biased agonist developed by Trevena. The main therapeutic action of oliceridine is analgesia. CYP3A4 and CYP2D6 primarily metabolize Oliceridine. There are minor contributions from CYP2C9 and CYP2C19 into inactive metabolites that do not have major activity at the μ-opioid receptor. CYP2D6 poor metabolizers have shown reduced mean clearance compared to non-poor CYP2D6 metabolizers [45]. Further study on the pharmacogenetic effects of oliceridine is necessary (see Table 1).

498

Sarahbeth Howes et al.

Table 1 Compilation of clinical studies of opioids and their pharmacogenetic considerations Author (year)

Opioid studied

Subjects (n)

Results and findings

Conclusion

Smith et al. (2019) [22]

Codeine

375

The pain intensity among intermediate and poor metabolizers (IMs and PMs) prescribed with codeine had greater improvement in the CYP2D6-guided versus usual care arm. In contrast, there was no difference in normal metabolizers prescribed codeine between CYP2D6 guided and usual care

The implementation of CYP2D6-guided care was shown to be feasible and yielded clinically relevant improvements in pain control among the subset of patients most expected to benefit. These results indicate CYP2D6 genotype testing may be a helpful addition to the primary care physician’s armamentarium in chronic pain management

SaizRodrı´guez et al. (2019) [46]

Fentanyl

35

The CYP3A4*22 allele carriers showed a higher area under the concentration-time (AUC) curve and lower clearance (CI) values. Carriers of the ABCB1 C1236T T/T genotype presented a lower AUC and higher CI. CYP3A5*3, ABCB1 C3435T, and ABCB1 G2677T/A were not associated with fentanyl’s pharmacokinetics

Fentanyl pharmacokinetics is affected by multiple genetic polymorphisms and differences. The study showed whether pharmacogenetics explains some of the variability in response to fentanyl by studying several genes related to fentanyl receptors, transporters, and metabolic enzymes

Linares et al. (2015) [31]

Hydrocodone

11

CYP2D6 phenotypespecific hydrocodone clinical pharmacokinetics were studied. Results showed significant phenotype-dependent effects of hydrocodone metabolism within different individuals

The results show that pharmacogenetics may provide an opportunity to individualize hydrocodone dosing

(continued)

Pharmacogenomics of Opioid Treatment

499

Table 1 (continued) Author (year)

Opioid studied

Subjects (n)

Results and findings

Conclusion

Zahari et al. (2016) [47]

Methadone

148

Results revealed associations of ABCB1 polymorphisms and serum methadone concentration over the 24-h dosing interval. Patients with CGC/TTT diplotype had 32.9% higher dose-adjusted serum methadone concentration over the 24-h dosing interval when compared to those without the diplotype

Genotyping of ABCB1 among opioiddependent patients on methadone maintenance therapy may help individualize and optimize methadone substitution treatment

Kharasch et al. (2015) [48]

Methadone

64

Several variant alleles of the CYP2B6 gene were studied. Average S-methadone apparent oral clearance was 35 and 45% lower in CYP2B6*1/*6 and CYP2B6*6/*6 genotypes, respectively, compared with CYP2B6*1/*1, and R-methadone apparent oral clearance was 25 and 30% lower. R- and S-methadone apparent oral clearance was 3- and 4-fold greater in CYP2B6*4 carriers. Intravenous and oral R- and S-methadone metabolism were significantly lower in CYP2B6*6 carriers than CYP2B6*1 homozygotes, and greater in CYP2B6*4 carriers

CYP2B6 polymorphisms influence methadone plasma concentration due to altered methadone metabolism and thus clearance. CYP2B6 pharmacogenetics partly explain individual variability in methadone elimination

Sadhasivam et al. (2012) [49]

Morphine

140

Increased morphine clearance is partly because of morphine3-clucuronide

Race is an important factor in perioperative intravenous (continued)

500

Sarahbeth Howes et al.

Table 1 (continued) Author (year)

Opioid studied

Subjects (n)

Results and findings formation, rather than the formation of morphine-6glucuronide. Common uridine diphosphate glucuronosyl transferase 2B7 genetic variation were not associated with racial differences in morphine’s clearance

Conclusion morphine’s clearance, and its potential role in personalizing analgesia with morphine needs further investigation

Olesen et al. (2015) [50]

Oxycodone

50

Genetic variation across selected mu- and delta-opioid receptor genes (OPRM1 and OPRD1) influenced the analgesic response to oxycodone in healthy controls

Variation in opioid receptor genes may partly explain responder characteristics to oxycodone

Balyan et al. (2017) [51]

Oxycodone

30

Compared with poor metabolizer (PM) or intermediate metabolizer (IM) subjects, significantly greater oxymorphone exposure was seen in extensive metabolizer (EM) subjects. Higher conversion of oxycodone to oxymorphone was observed in EM subjects compared to PM or IM subjects

CYP2D6 phenotypes explain metabolism of oxycodone in children, and oxymorphone exposure is higher in CYP2D6 EM phenotype

6

Conclusion The discovery of opioids for pain management is of the utmost importance in medicine and has prompted further investigation into its efficacy and the variation in effects seen in patients. Research into the pharmacogenomics of opioid treatment is necessary to investigate the ranges of opioid metabolism to further understand the role of genetic polymorphisms and drug efficacy.

Pharmacogenomics of Opioid Treatment

501

Individual genetic differences in opioids contribute to the variation in pain perception and management. Furthermore, individuals can be categorized into poor, intermediate, or efficient opioid metabolizers based on their alleles [1]. An example of this is CYP2D6, which metabolizes most opioid medications, and shows significant variance in races and ethnicities and should continue to be investigated so that treatment can be tailored to be the most effective in treatment management. The purpose and intent of these investigations are to help reduce confounding factors that could compromise the efficiency of opioid medications. Opioids have multiple receptors that bind to exert their analgesic effects and are located throughout the human body. The receptors are G-protein coupled receptors (GPCR) and are classified as Mu (found in the brainstem and thalamus), Kappa (limbic system, brainstem, and spinal cord), and Delta (also found in the brain). Mu receptors modulate the calcium influx and the potassium efflux to block pain transmission to higher cortical centers in the brain via neurotransmitters such as substance P, glutamate, and CGRP [3]. As previously stated, the pharmacogenomics of each patient is critical in determining their response to either the effectiveness or metabolism of opioids. An example is an SNP in the OPRM1 gene that encodes for the Mu opioid receptor that can determine the response to exogenous opioids. [5] However, opioid metabolism has the greatest variation among patients. Phase 1 metabolism of opioids is carried out largely by CYP3A4 (oxycodone and fentanyl) and CYP2D6 (hydrocodone, codeine, and dihydrocodeine). An example of Phase 2 metabolism of an opioid (morphine) is carried out via UGT2B7, where the drug is catalyzed into M6G and M3G. There are two variants connected to a reduction in the rate of glucuronidation, causing an accumulation of morphine and a decrease in the formation of metabolites [12]. Healthcare professionals must consider the variation in the ability for different responses and alleles to determine the optimal drug not only for analgesia but in the context of efficacy and metabolism. Pharmacogenetic effects in the metabolism of opioid analgesics described in this chapter illustrate the importance of individualized medicine. The wide range of use for opioids and their effectiveness have made them a necessary medical advancement and has been beneficial for many patients who experience severe and chronic pain. However, further consideration must be taken when prescribing these medications to ensure efficacy in the analgesia for a wide variety of patients.

502

Sarahbeth Howes et al.

References 1. Peiro´ AM, Planelles B, Juhasz G, Bagdy G, Libert F, Eschalier A, Busserolles J, Sperlagh B, Llerena A (2016) Pharmacogenomics in pain treatment. Drug Metab Pers Ther 31:131–142 2. Kaye AD, Garcia AJ, Hall OM, Jeha GM, Cramer KD, Granier AL, Kallurkar A, Cornett EM, Urman RD (2019) Update on the pharmacogenomics of pain management. Pharmacogenomics Pers Med 12:125–143 3. Cornett EM, Carroll Turpin MA, Pinner A, Thakur P, Sekaran TSG, Siddaiah H, Rivas J, Yates A, Huang GJ, Senthil A, Khurmi N, Miller JL, Stark CW, Urman RD, Kaye AD (2020) Pharmacogenomics of pain management: the impact of specific biological polymorphisms on drugs and metabolism. Curr Oncol Rep 22:18 4. Jimenez N, Galinkin JL (2015) Personalizing pediatric pain medicine: using populationspecific pharmacogenetics, genomics, and other – omics approaches to predict response. Anesth Analg 121:183–187 5. Mandala` M, Moro C, Labianca R, Cremonesi M, Barni S (2006) Optimizing use of opiates in the management of cancer pain. Ther Clin Risk Manag 2:447–453 6. Paice JA (2007) Pharmacokinetics, pharmacodynamics, and pharmacogenomics of opioids. Pain Manag Nurs 8:S2–S5 7. Han W, Ide S, Sora I, Yamamoto H, Ikeda K (2004) A possible genetic mechanism underlying individual and interstrain differences in opioid actions: focus on the mu opioid receptor gene. Ann N Y Acad Sci 1025:370–375 8. Klepstad P, Rakva˚g TT, Kaasa S, Holthe M, Dale O, Borchgrevink PC, Baar C, Vikan T, Krokan HE, Skorpen F (2004) The 118 A > G polymorphism in the human mu-opioid receptor gene may increase morphine requirements in patients with pain caused by malignant disease. Acta Anaesthesiol Scand 48: 1232–1239 9. Sia AT, Lim Y, Lim ECP, Goh RWC, Law HY, Landau R, Teo Y-Y, Tan EC (2008) A118G single nucleotide polymorphism of human mu-opioid receptor gene influences pain perception and patient-controlled intravenous morphine consumption after intrathecal morphine for postcesarean analgesia. Anesthesiology 109:520–526 10. Trescot AM, Datta S, Lee M, Hansen H (2008) Opioid pharmacology. Pain Physician 11: S133–S153

11. Land BB, Bruchas MR, Lemos JC, Xu M, Melief EJ, Chavkin C (2008) The dysphoric component of stress is encoded by activation of the Dynorphin κ-opioid system. J Neurosci 28:407–414 12. Chatti I, Woillard J-B, Mili A, Creveaux I, Ben Charfeddine I, Feki J, Langlais S, Ben Fatma L, Saad A, Gribaa M, Libert F (2017) Genetic analysis of mu and kappa opioid receptor and COMT enzyme in cancer pain Tunisian patients under opioid treatment. Iran J Public Health 46:1704–1711 13. Abdallah K, Gendron L (2018) The Delta opioid receptor in pain control. Handb Exp Pharmacol 247:147–177 14. Gave´riaux-Ruff C, Kieffer BL (2011) Delta opioid receptor analgesia: recent contributions from pharmacology and molecular approaches. Behav Pharmacol 22:405–414 15. Smith HS (2009) Opioid Metabolism. Mayo Clin Proc 84:613–624 16. Opioid Metabolism and Effects of Cytochrome P450 | Pain Medicine | Oxford Academic. https://academic.oup.com/painmedicine/arti cle/10/suppl_1/S20/1914905?login¼true. Retrieved 9 Aug 2021 17. Smith HS (2008) Variations in opioid responsiveness. Pain Physician 11:237–248 18. Singh A, Zai C, Mohiuddin AG, Kennedy JL (2020) The pharmacogenetics of opioid treatment for pain management. J Psychopharmacol (Oxf) 34:1200–1209 19. Zhou S-F (2009) Polymorphism of human cytochrome P450 2D6 and its clinical significance: part I. Clin Pharmacokinet 48:689–723 20. Kirchheiner J, Schmidt H, Tzvetkov M, Keulen J-T, Lo¨tsch J, Roots I, Brockmo¨ller J (2007) Pharmacokinetics of codeine and its metabolite morphine in ultra-rapid metabolizers due to CYP2D6 duplication. Pharmacogenomics J 7: 257–265 21. Kelly LE, Rieder M, van den Anker J, Malkin B, Ross C, Neely MN, Carleton B, Hayden MR, Madadi P, Koren G (2012) More codeine fatalities after tonsillectomy in north American children. Pediatrics 129:e1343–e1347 22. Smith DM, Weitzel KW, Elsey AR, Langaee T, Gong Y, Wake DT, Duong BQ, Hagen M, Harle CA, Mercado E, Nagoshi Y, Newsom K, Wright A, Rosenberg EI, Starostik P, Clare-Salzler MJ, Schmidt SO, Fillingim RB, Johnson JA, Cavallari LH (2019) CYP2D6-guided opioid therapy improves pain control in CYP2D6 intermediate and

Pharmacogenomics of Opioid Treatment poor metabolizers: a pragmatic clinical trial. Genet Med 21:1842–1850 23. Koolen SL, Van der Rijt CC (2017) Is there a role for pharmacogenetics in the dosing of fentanyl? Pharmacogenomics 18:417–419 24. Takashina Y, Naito T, Mino Y, Yagi T, Ohnishi K, Kawakami J (2012) Impact of CYP3A5 and ABCB1 gene polymorphisms on fentanyl pharmacokinetics and clinical responses in cancer patients undergoing conversion to a transdermal system. Drug Metab Pharmacokinet 27:414–421 25. Barratt DT, Bandak B, Klepstad P, Dale O, Kaasa S, Christrup LL, Tuke J, Somogyi AA (2014) Genetic, pathological and physiological determinants of transdermal fentanyl pharmacokinetics in 620 cancer patients of the EPOS study. Pharmacogenet Genomics 24:185–194 26. Kuip EJM, Zandvliet ML, Koolen SLW, Mathijssen RHJ, van der Rijt CCD (2017) A review of factors explaining variability in fentanyl pharmacokinetics; focus on implications for cancer patients: variability in fentanyl pharmacokinetics. Br J Clin Pharmacol 83:294–313 27. Nielsen LM, Olesen AE, Branford R, Christrup LL, Sato H, Drewes AM (2015) Association between human pain-related genotypes and variability in opioid analgesia: an updated review. Pain Pract 15:580–594 28. Rosenblum A, Marsch LA, Joseph H, Portenoy RK (2008) Opioids and the treatment of chronic pain: controversies, current status, and future directions. Exp Clin Psychopharmacol 16:405–416 29. Cardia L, Calapai G, Quattrone D, Mondello C, Arcoraci V, Calapai F, Mannucci C, Mondello E (2018) Preclinical and clinical pharmacology of hydrocodone for chronic pain: a mini review. Front Pharmacol 9: 1122 30. Hutchinson MR, Menelaou A, Foster DJR, Coller JK, Somogyi AA (2004) CYP2D6 and CYP3A4 involvement in the primary oxidative metabolism of hydrocodone by human liver microsomes. Br J Clin Pharmacol 57:287–297 31. Linares OA, Fudin J, Daly AL, Boston RC (2015) Individualized hydrocodone therapy based on phenotype, pharmacogenetics, and pharmacokinetic dosing. Clin J Pain 31: 1026–1035 32. Latta KS, Ginsberg B, Barkin RL (2002) Meperidine: a critical review. Am J Ther 9: 53–68 33. Ramı´rez J, Innocenti F, Schuetz EG, Flockhart DA, Relling MV, Santucci R, Ratain MJ (2004) CYP2B6, CYP3A4, and CYP2C19 are responsible for the in vitro N-demethylation of

503

meperidine in human liver microsomes. Drug Metab Dispos Biol Fate Chem 32:930–936 34. Murray JL, Mercer SL, Jackson KD (2020) Impact of cytochrome P450 variation on meperidine N -demethylation to the neurotoxic metabolite normeperidine. Xenobiotica 50:132–145 35. Packiasabapathy S, Aruldhas BW, Horn N, Overholser BR, Quinney SK, Renschler JS, Sadhasivam S (2020) Pharmacogenomics of methadone: a narrative review of the literature. Pharmacogenomics 21:871–887 36. Walter C, Lo¨tsch J (2009) Meta-analysis of the relevance of the OPRM1 118A>G genetic variant for pain treatment. Pain 146:270–275 37. Levran O, Peles E, Randesi M, Shu X, Ott J, Shen P-H, Adelson M, Kreek MJ (2013) Association of genetic variation in pharmacodynamic factors with methadone dose required for effective treatment of opioid addiction. Pharmacogenomics 14:755–768 38. Eap CB, Buclin T, Baumann P (2002) Interindividual variability of the clinical pharmacokinetics of methadone: implications for the treatment of opioid dependence. Clin Pharmacokinet 41:1153–1193 39. Thorn CF, Lamba JK, Lamba V, Klein TE, Altman RB (2010) PharmGKB summary: very important pharmacogene information for CYP2B6. Pharmacogenet Genomics 20: 520–523 40. Klepstad P, Dale O, Skorpen F, Borchgrevink PC, Kaasa S (2005) Genetic variability and clinical efficacy of morphine. Acta Anaesthesiol Scand 49:902–908 41. Sadiq NM, Dice TJ, Mead T (2021) OxycodoneStatPearls. StatPearls Publishing, Treasure Island 42. Andreassen TN, Eftedal I, Klepstad P, Davies A, Bjordal K, Lundstro¨m S, Kaasa S, Dale O (2012) Do CYP2D6 genotypes reflect oxycodone requirements for cancer patients treated for cancer pain? A cross-sectional multicentre study. Eur J Clin Pharmacol 68:55–64 43. So¨derberg Lo¨fdal KC, Andersson ML, Gustafsson LL (2013) Cytochrome P450-mediated changes in oxycodone pharmacokinetics/pharmacodynamics and their clinical implications. Drugs 73:533–543 44. Zwisler ST, Enggaard TP, Mikkelsen S, Brosen K, Sindrup SH (2010) Impact of the CYP2D6 genotype on post-operative intravenous oxycodone analgesia. Acta Anaesthesiol Scand 54:232–240 45. Trevena Inc (2020) Olinvyk (oliceridine) [prescribing information]

504

Sarahbeth Howes et al.

46. Saiz-Rodrı´guez M, Ochoa D, Herrador C, Belmonte C, Roma´n M, Alday E, Koller D, Zubiaur P, Mejı´a G, Herna´ndez-Martı´nez M, Abad-Santos F (2019) Polymorphisms associated with fentanyl pharmacokinetics, pharmacodynamics and adverse effects. Basic Clin Pharmacol Toxicol 124:321–329 47. Zahari Z, Lee CS, Ibrahim MA, Musa N, Mohd Yasin MA, Lee YY, Tan SC, Mohamad N, Ismail R (2016) Relationship between ABCB1 polymorphisms and serum methadone concentration in patients undergoing methadone maintenance therapy (MMT). Am J Drug Alcohol Abuse 42: 587–596 48. Kharasch ED, Regina KJ, Blood J, Friedel C (2015) Methadone pharmacogenetics. Anesthesiology 123:1142–1153

49. Sadhasivam S, Krekels EHJ, Chidambaran V, Esslinger HR, Ngamprasertwong P, Zhang K, Fukuda T, Vinks AA (2012) Morphine clearance in children: does race or genetics matter? J Opioid Manag 8:217–226 50. Olesen AE, Sato H, Nielsen LM, Staahl C, Droney J, Gretton S, Branford R, Drewes AM, Arendt-Nielsen L, Riley J, Ross J (2015) The genetic influences on oxycodone response characteristics in human experimental pain. Fundam Clin Pharmacol 29:417–425 51. Balyan R, Mecoli M, Venkatasubramanian R, Chidambaran V, Kamos N, Clay S, Moore DL, Mavi J, Glover CD, Szmuk P, Vinks A, Sadhasivam S (2017) CYP2D6 pharmacogenetic and oxycodone pharmacokinetic association study in pediatric surgical patients. Pharmacogenomics 18:337–348

Chapter 18 The Role of Pharmacogenomics in Postoperative Pain Management E. Paylor Sachtleben, Kelsey Rooney, Hannah Haddad, Victoria L. Lassiegne, Megan Boudreaux, Elyse M. Cornett, and Alan D. Kaye Abstract Pharmacogenomics can improve pain management by considering individual variations in pain perception and susceptibility and sensitivity to medicines related to genetic diversity. Due to the subjective nature of pain and the fact that people respond differently to medicines, it can be challenging to develop a consistent and successful regimen for pain disorders. Numerous factors influence the outcome of pain treatment programs, but two stand out: altered perception of pain and varying responsiveness to analgesic medicines. Numerous polymorphisms in genes such as CYP2D6, OPRM1, and ABCB1 have been identified, culminating in a heterogeneous response to pain medication in people who have these genetic polymorphisms. Improved treatment regimens that factor in pharmacogenetic differences in patients would help reduce the risk of opioid dependency and help effectively treat postoperative pain. Key words Pharmacogenomics, Pain management, Postoperative pain, Pharmacogenetics, Precision medicine, Acute pain, Chronic pain

1

Introduction Pharmacogenomics is the study of genetic variants that alter the pharmacokinetics and pharmacodynamics of drugs. At present, pharmacogenomics is being incorporated into pain treatment regimens to better manage pain based on individualized differences in perception of pain and susceptibility and sensitivity to drugs based on genetic variation [1]. Since pain is subjective and patients respond to medications differently, it can be difficult to produce a consistent and effective regimen for pain conditions [2]. Many variables affect the outcome of pain treatment plans, including alteration in the perception of pain and varied response to analgesic medications. Many polymorphisms have been found for genes such as CYP2D6, OPRM1, and ABCB1, resulting in a heterogeneous

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_18, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

505

506

E. Paylor Sachtleben et al.

response to pain medication in these patients [3]. CYP2D6 is a specific cytochrome P450 enzyme gene that alters the metabolism of various medications, OPRM1 encodes the opioid μ1 receptor responsible for the perception of pain, and ABCB1 is an adenosine triphosphate-binding cascade that is responsible for medication transport and effects within cells [4]. These are only a few of many genes that can play a role in the pain pathway and medication effectiveness. A better understanding of the dynamics of these various genetic polymorphisms would help develop a more efficacious approach to pain control. In addition, better treatment regimens would help decrease the chance for potential opioid dependence and mistreatment secondary to chronic pain.

2

Pharmacogenetic Factors That Affect Postoperative Pain Pharmacogenetics encompasses a group of inherited factors that can potentially affect drug pharmacokinetics and pharmacodynamics [2]. Identifying these pharmacogenetic factors can help predict how patients respond to a drug regimen and help tailor treatments based on an individual’s genetic and social factors [2]. Five main pharmacogenetic factors result in differences in perceived pain among individuals: environmental factors, biological factors, psychological factors, genetic factors, and ethnic factors [2]. Detecting and employing these individual variances is the beginning of the precision medicine era [5].

2.1 Environmental Factors

Socioeconomic qualities and stress are two identified environmental factors that have an impact on pain [2]. It has been described that as socioeconomic status decreases, chronic pain prevalence increases [2, 6]. Furthermore, this increase in chronic pain identified in less prosperous areas has been linked to a rise in analgesic use and amplified morbidity for chronic noninflammatory musculoskeletal pain [2, 7]. Additionally, chronic stress may also play a role in pain perception and perpetuate pain in fibromyalgia patients [2, 8]. Environmental factors such as smoking, diet, drug use, and alcohol can also potentially affect a patient’s medication response [5]. For example, preoperatively patients who used benzodiazepines or serotonin reuptake inhibitors have been linked to requiring morphine postoperatively [2, 9].

2.2 Biological Factors

Biological factors such as age and gender also play a role in pain perception [2]. A person’s age accounts for many drug distribution, metabolism, and elimination [2]. A linear relationship has been established between age and sensitivity to the analgesic effects of morphine [2, 9]. On the other hand, long-term musculoskeletal pain is increasing in adolescents [2, 10]. This variation may be

Pharmacogenomics in Postoperative Pain Management

507

explained by increasing perceived stress levels leading to increased pain sensitivity [2, 10]. Additionally, gender differences have been studied as a source for pain discrepancies. It has been demonstrated that men require an increased frequency of morphine than women in the postoperative setting [2, 11]. There are also biological polymorphisms involved in pain perception [2]. Interleukin 6 (IL-6) is a pro-inflammatory cytokine that has been linked to causing increased neuronal responses to temperature changes and an increased pain reported postoperatively after total knee arthroplasty [2, 12, 13]. Moreover, cytokines and chemokines are involved in the initiation of pain and the persistence of pain [13]. 2.3 Psychological Factors

There is a correlation between depression and a higher pain prevalence, and the two have a synergistic effect when coexisting together [2, 14]. For example, patients with both pain and depression are more likely to have pain grievances, greater impairment, severe depressive symptoms, and treatment-resistant depression than if they existed separately [14]. Depression and anxiety are psychological risk factors for developing certain pain syndromes, such as tension headaches [2, 15]. However, this relationship between psychological factors and pain can be advantageous. This is one of the only pharmacogenetic factors that is not fixed and can be treated to help improve postoperative pain. For instance, managing a patient’s depression has been shown to improve pain and weakens the relationship between socioeconomic status and chronic pain, as discussed earlier [2, 16].

2.4

Humans’ DNA only differs by about 0.5% between each individual [5]. Genetic and epigenetic factors have been identified as contributing to 24–60% of the variances observed in pain perception [2, 17]. The polymorphisms in drug-metabolizing enzymes, cytochrome p450 (CYP450) and UDP glucuronosyltransferase (UGT), account for 10–10,000-fold variations in drug activity [5]. For example, homozygous patients for the UGT2B7 802C enzyme need less morphine for adequate pain management in the postoperative period [5]. Employing a discovery like this could potentiate a tremendously positive effect on the opioid epidemic by allowing less of the drug to be used with the same effect [2]. Along with the same theme, the epigenetic factors, caspase 9 and interleukin 16, on their respective genes, have been identified as influences that increase self-reported pain without worsening disease severity [2, 17]. Additionally, the increased pain perception in fibromyalgia patients is attributed to the polymorphisms involving the serotonin 5-HT2A receptor, serotonin transporter, and dopamine-4 receptor [2, 17]. Val(158)Met is also a polymorphism identified on the Catechol O-methyl transferase (COMT) enzyme gene that causes hypo-functioning of the enzyme resulting in increased pain

Genetic Factors

508

E. Paylor Sachtleben et al.

sensitivity in fibromyalgia and perioperative pain and increased sensitivity to opioids [2]. Similarly, the A118G allele located on chromosome 6q25.2 encodes for a Mu-opioid receptor that results in an increased requirement for opioids to achieve satisfactory pain management [5]. 2.5

3

Ethnic Factors

African Americans have a lower threshold for temperature differences, pressure, ischemia, and pain specifically related to glaucoma, AIDS, migraines, jaw pain, headaches, postoperative pain, angina, joint pain, and arthritis than Caucasians [2, 18, 19]. Additionally, 10% more African Americans and 11% more Hispanics over 50 years old report having more severe pain most of the time when compared to non-Hispanic white individuals [2, 20]. Certain ethnic factors can correlate to genetic factors as well. For instance, the variant rs8007267 within the GCH1 gene that leads to the increased production of tetrahydrobiopterin (BH4) is protective from chronic pain observed in African Americans with sickle cell disease [2]. Unfortunately, some of the pain differences observed between different ethnic groups can also be attributed to healthcare disparities in assessment and treatment between ethnic groups for all types of pain, including acute, chronic, cancer, nonmalignant, and experimental pain [19]. Identifying and utilizing individual patient’s pharmacogenetics is the future of precision medicine [5]. Developing medications tailored to the patient’s genetics and related factors instead of their height and weight will increase treatment effectiveness and decrease adverse events while diminishing the opioid crisis [5]. Additionally, awareness of these inherent differences in pain perception can hopefully manifest into fewer biases and disparities in the healthcare system.

Pharmacogenomics and Postoperative Pain Management Nonsteroidal anti-inflammatory drugs (NSAIDs) are frequently employed for postoperative pain management because of their negligible addictive potential. Regularly used NSAIDs include ibuprofen, naproxen, celecoxib, diclofenac, and indomethacin. The mechanism of these anti-inflammatory drugs is disruption of prostaglandin synthesis from arachidonic acid by inhibiting cyclooxygenase (COX) enzymes COX-1 and/or COX-2, which are encoded by the prostaglandin-endoperoxidase synthase 1 and 2 genes (PTGS1 and PTGS2), respectively [4]. The PTGS1 and PTGS2 genes are considered highly polymorphic, with variations that are thought to potentially affect the functioning of COX enzymes about NSAIDs. Many studies have failed to identify a link between PTGS polymorphisms and adverse reactions of NSAIDs [4, 21]. Other studies have acknowledged an improved pain

Pharmacogenomics in Postoperative Pain Management

509

response with selective COX-1 and COX-2 agents in individuals with variations resulting in greater expression of PTGS1 and PTGS2, respectively [22]. Further investigation is needed to fully elucidate the influence of PTGS genetic variants on the pharmacogenomics of NSAIDs. Cytochrome P450 (CYP) genes encode enzymes that contribute largely to the phase 1 metabolism of many types of NSAIDs [4]. The functionality of CYP encoded enzymes is influenced by many factors, including genetic variation. Specifically, CYP2C9 (hepatic cytochrome P450 2D6) variations influence metabolic clearance impacting systemic levels of NSAIDs. The CYP2C9 genotype is highly polymorphic, consisting of 61 variant alleles and numerous sub-alleles [23]. Some evidence indicates that the genetic variation of CYP2C9 is linked to increased adverse events in those taking NSAIDs. Patients carrying the poor metabolizer (PM) phenotype of CYP2C9 (alleles CYP2C9*2 and CYP2C9*3) may exhibit significantly reduced drug metabolism, which results in an increased drug half-life, higher plasma concentrations, and greater potential of toxicity of NSAIDs [4, 22, 23]. The Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines advise implementing non-CYP2C9-metabolized NSAIDs for patients with decreased function alleles in the CYP2C9 gene [23].

4

Acetaminophen Postoperatively, acetaminophen is often used for mild pain relief or as an opioid-sparing analgesic to minimize opioid dosing in cases of severe pain. The pharmacogenomics of acetaminophen is a crucial topic, as the most common cause of acute liver failure in the United States is acetaminophen overdose [24]. The metabolization of acetaminophen predominantly occurs in the liver with phase 1 oxidation mediated by CYP encoded enzymes, primarily CYP2E1 and CYP2D6, with support from CYP1A2, CYP2A6, and CYP3A4 [4]. After a toxic dose, excess acetaminophen metabolization saturates phase 2 glucuronidation and sulfation pathways (mediated by uridine-50 -diphospho-glucuronosyltransferase (UGT)1A1/1A6/ 1A9/2B15 and sulfotransferase (SULT)1A1/1A3/1A4/2A1), resulting in excess oxidation, primarily by CYP2E1, producing the reactive metabolite N-acetyl-p-benzoquinone imine (NAPQI) [25, 26]. NAPQI is the toxic compound liable for acetaminophen-induced hepatotoxicity, a growing concern due to frequent worldwide acetaminophen poisoning. Current literature does not support a link between altered acetaminophen metabolism and genetic variants [4]. However, some evidence exists regarding increased tolerance to high doses of acetaminophen through consequential upregulation of genes encoding adenosine triphosphatebinding cassette (ABC) transporters [26]. ABC transporters are

510

E. Paylor Sachtleben et al.

responsible for substrate efflux, and ABCC2, ABCC3, and ABCG2 can play a role in acetaminophen excretion [4]. Patients who experienced prior acetaminophen overdose have shown an increased expression in ABC transporter encoding messenger-ribonucleic acids (mRNAs). It is hypothesized that a greater concentration of ABC transporters may result in increased toxic metabolite excretion, leading to reduced accumulation and lessened hepatic injury [26]. More studies are needed to fully elucidate the potential of hepatic protection from acetaminophen through ABC gene upregulation.

5

Lidocaine Incorporating systemic lidocaine into multimodal pain therapy is expanding due to growing clinical evidence of its antinociceptive and anti-hyperalgesic effect on acute postoperative and chronic neuropathic pain states [27]. The full extent of lidocaine’s analgesic benefits cannot be explained by its modulatory effects on sodium channels alone. However, other responsible mechanisms have yet to be described. CYP1A2 and CYP3A4 mainly catalyze the metabolization of lidocaine in the liver encoded enzymes [28]. Previous studies have identified 46 distinctive types of CYP3A4 allelic variants [29]. Fang et al. recently reported on the functional assessment of 22 CYP3A4 variants in the Chinese Han population. The study identified seven allelic variants (CYP3A4*2, *5, *9, *16, *17, *24, or *30) that result in the poor metabolization of lidocaine [29]. Special care should be considered when dosing lidocaine to patients carrying these variants to avoid adverse events. Additionally, lidocaine increases susceptibility to oxidative damage in those with deficiency of the gene encoding glucose-6-phosphate dehydrogenase (G6PD) [30].

6

Codeine Intravenous opioids are the most common postoperative analgesics to produce drug-gene interactions [30]. Genetic effects on systemic levels of opioids, such as codeine and morphine, may lead to serious adverse events. Codeine is converted into morphine through CYP2D6 encoded enzymes [4]. Four phenotypes of CYP2D6 polymorphisms on opioid conversion have been identified, including normal metabolizer (NM), PM, intermediate metabolizer (IM), and ultrarapid metabolizer (UM) [4]. UM alleles (CYP2D6*1*1, CYP2D6*1/*2) develop more side effects to codeine and have an increased risk of respiratory depression; PM alleles (CYP2D6*4/4, CYP2D6*4/5, CYP2D6*4/6) develop fewer side effects to codeine and cause reduced analgesic effect

Pharmacogenomics in Postoperative Pain Management

511

[22, 30]. CPIC guidelines recommend avoiding codeine in UMs and PMs due to prospective adverse reactions and reduced efficacy, respectively. Considerations are also given to avoid postoperative codeine in those 18 years or younger following tonsillectomy or adenoidectomy [31]. Additionally, CPIC guidelines recommend breastfeeding mothers avoid codeine. Codeine metabolites are secreted into breast milk, and fatal doses of opioids may be fed to an infant if the mother has a UM phenotype [31]. ABCB1 functions as an efflux transporter for various substrates, including endogenous and exogenous opioids [22]. Studies have reported ABCB1 single-nucleotide polymorphism (SNP) expression may impact opioid bioactivity. The most common ABCB1 variants influencing opioid expression are rs1045642, rs2032582, and rs1128503 [32]. Allele A of ABCB1 (rs1045642) is correlated with fewer side effects and a significantly reduced postoperative opioid dosage than allele G [33]. Furthermore, breastfed infants of a codeine-treated mother are at increased risk of opioid-induced central nervous system depression with an ABCB1 (rs1045642, rs2032582, and rs1128503) AA or AG genotype or with a mother with an ABCB1(rs1128503) AA genotype [34]. This becomes important in the setting of postoperative pain management of a breastfeeding mother [32].

7

Morphine Conversion of morphine into morphine-6-glucuronide (M6G) and morphine-3-glucuronide (M3G) occurs primarily via UGT2B7 with some contribution from UGT1A1 [22]. M6G is an active metabolite with greater antinociceptive activity on opioid receptors than M3G [35]. There is contradicting evidence at present supporting an effect of UGT2B7 genetic variants on the response to postoperative morphine treatment. In a trial of healthy individuals, UGT2B7 and ABCB1 variants were not shown to influence the antinociceptive and pharmacokinetic properties of morphine [36]. Contrarily, women with the UGT2B7 CC genotypic variant required less postoperative morphine than the TT genotype following hysterectomy [37]. Similarly, cancer patients with the UGT2B7 CC genotypic variant showed reduced pain scores and increased plasma concentration of morphine compared to the CT or TT genotypes following morphine treatment [38]. Common SNPs of UGT2B7 have also been shown to regioselectively influence the intrinsic clearance of M6G and M3G, potentially resulting in altered morphine efficacy [35]. Additional investigation is needed to fully understand the effects of UGT2B7 polymorphisms on postoperative morphine dosing.

512

E. Paylor Sachtleben et al.

Inconsistent evidence exists on the effect of opioid receptor mu 1 (OPRM1) variants on morphine activity. Some studies have found OPRM1 polymorphisms adjust postoperative morphine dose requirements. Patients with the OPRM1 (rs1799971) G allele may require a small increased postoperative morphine dose compared to the wild-type A allele [31, 37]. CPIC guidelines do not currently recommend morphine dosing based on OPRM1 genotypes [31].

8

Tramadol Tramadol is converted into the active metabolite O-desmethyltramadol (ODT) and the inactive metabolite N-desmethyltramadol (NDT) by CYP2D6 and CYP2B6, respectively [22, 39]. The effects of CYP2D6 genetic variation on tramadol expression have been widely studied. As previously mentioned, the four phenotypes of CYP2D6 polymorphisms on opioids are NM, PM, IM, and UM [4]. Studies have identified PMs receiving tramadol for postoperative analgesia exhibit reduced ODT plasma concentrations, increased NDT plasma concentration, and diminished analgesic response [31, 39, 40]. Contrarily, UMs taking tramadol have been shown to experience increased ODT plasma concentration, enhanced analgesia, and greater side effects, such as miosis and nausea [31, 40, 41]. Similar to codeine, the CPIC guidelines recommend avoiding tramadol in UMs and PMs due to prospective adverse reactions and reduced efficacy, respectively. Considerations are also given to avoid postoperative tramadol in those 18 years or younger following tonsillectomy or adenoidectomy [31]. Interestingly, current clinical CYP2D6 phenotype predictions have well-documented inaccuracies in specific populations. A pathway-driven predictive model assessing polymorphisms of multiple metabolic enzymes of tramadol, specifically including UGT2B7 genotypes, may improve predictive accuracy to 90% [42]. Limited studies exist on the potential pharmacogenetic effect of CYP2B6 variants on tramadol expression. Few clinical trials have demonstrated an influence of CYP2B6 polymorphisms on the pharmacokinetics of tramadol, particularly in the form of reduced enzyme function [39, 43]. Other studies have noted no effect from CYP2B6 genotypes on tramadol expression [44]. Additional variations that have controversial effects on tramadol include ABCB1, organic cation transporter 1 (OCT1), and OPRM1 genes. Patients carrying the ABCB1 (rs2032582) CC genotype may require increased tramadol dosing following total joint arthroplasty for adequate pain control [32]. Reduced postoperative tramadol usage and increased plasma ODT have been linked to patients carrying a loss-of-function allele in OCT1 [45]. Lastly, patients carrying the OPRM1 (rs1799971) GG genotype have been linked

Pharmacogenomics in Postoperative Pain Management

513

to increased postoperative tramadol requirements compared to AA and AG genotypes [46].

9

Hydrocodone Hydrocodone is converted into the active metabolite hydromorphone by CYP2D6. CYP3A5 and CYP3A4 catalyze the conversion of hydrocodone to the minor metabolite norhydrocodone [4, 47]. The parent compound of hydrocodone provides analgesic activity regardless of conversion to hydromorphone or norhydrocodone [31, 48]. CYP3A5 and CYP3A4 SNPs were not shown to influence plasma metabolite concentration, analgesic response, or dosage of hydrocodone [47]. Limited evidence supports a clinically relevant effect of CYP2D6 UMs on hydrocodone expression. However, more established evidence supports a pharmacokinetic influence of CYP2D6 PMs on hydrocodone [31, 48, 49]. CYP2D6 PMs may require higher doses of postoperative hydrocodone for effective pain relief [49]. However, current CPIC guidelines determined inadequate evidence to make special recommendations for UMs, PMs, and IMs taking hydrocodone [31]. OPRM1 variations may have a role in hydrocodone expression. Patients homozygous for the 118A allele of OPRM1 have been shown to have a relationship between postoperative pain relief and total hydrocodone dose. Additionally, those with the AG or GG genotype have exhibited increased hydrocodone side effects, such as constipation, dry mouth, and respiratory depression [50]. However, due to insufficient evidence, current CPIC guidelines do not recommend hydrocodone dosing based on OPRM1 genotypes [31]. See Table 1.

10

Oxycodone Oxycodone is a μ- and κ-opioid receptor agonist with a prolonged duration of action compared to morphine [51]. The drug is metabolized in the liver primarily by cytochrome P450 enzymes CYP3A4 and CYP2D6. CYP3A4 converts oxycodone into noroxycodone, while CYP2D6 is responsible for converting oxycodone into its pharmacologically active metabolite, oxymorphone. Depending on the genetic variant of the CYP2D6 enzyme a patient has, he/she can be classified as a poor (PM), intermediate (IM), extensive (EM), or ultrarapid metabolizer (UM) of oxycodone into oxymorphone [52]. Those who are ultrarapid metabolizers will thus have higher levels of the potent metabolite, oxymorphone, while those who are poor metabolizers will have lower levels of oxymorphone. One may theorize that ultrarapid metabolizers are thus at a greater risk for oxycodone toxicity, while poor

514

E. Paylor Sachtleben et al.

Table 1 Clinical efficacy

Author (year)

Groups studied and intervention

Results and findings

Conclusions

Neskovic et al. [73] (2021)

Postoperative patients (n ¼ No difference was seen in CYP2D6 polymorphisms greatly influence tramadol concentration 47) admitted to the ICU tramadol metabolism. between metabolic following tramadol Consideration should phenotypes. UM NDT administration were be given when dosing levels were below the tested for CYP2D6 tramadol to limit of quantification polymorphisms to assess postoperative patients until the second its effect on tramadol, tramadol dose was given. ODT, and NDT plasma PMs exhibited higher concentrations, and NDT levels versus EMs postoperative pain scores and IMs. EMs showed higher ODT levels versus IMs and PMs. Pain scores confirmed that pain was adequately controlled in all patients following second dose of tramadol, except for PM patients

Smith et al. [74] (2019)

Patients with chronic pain IMs and PMs treated with This data supports utilizing personalized CYP2D6-guided care were treated with medicine for opioid experienced greater opioids (tramadol or dosing, specifically improvement of pain codeine) following usual CYP2D6-guided pain intensity versus those care (n ¼ 135) or management for managed with usual care. CYP2D6-guided care codeine and tramadol NMs exhibited no (n ¼ 235) for 3 months use difference in pain intensity between CYP2D6-guided and usual care

Hosseinnejad et al. [47] (2019)

Females 18–34 years who CYP3A4/A5 variants did These results demonstrate that CYP3A4/A5 not exhibit a significant underwent cesarean phenotypes do not difference in pain section and received influence pain relief or outcomes or plasma hydrocodone for adverse effects of hydromorphone postoperative pain relief hydrocodone, but concentrations. had their blood tested CYP2D6 phenotypes CYP2D6 phenotypes for genotyping and are predictors for exhibited a significant assessment of plasma postoperative plasma effect on a total dosage drug concentrations and hydromorphone of hydrocodone and had pain scores concentration and pain a predictive effect on relief plasma hydromorphone concentrations

Saiz-Rodriguez et al. [43] (2020)

CYP2D6 IMs exhibited Healthy individuals (n ¼ greater plasma 24) given an oral dose of concentrations and less tramadol were analyzed

CYP2B6 is a less studied variant, and the connection between (continued)

Pharmacogenomics in Postoperative Pain Management

515

Table 1 (continued)

Author (year)

Groups studied and intervention for 18 enzymatic polymorphisms to examine if any polymorphisms affect the pharmacokinetics, pharmacodynamics, and safety of tramadol

Results and findings clearance of tramadol versus NMs and UMs. CYP2B6 G516T T/T genotype exhibited higher plasma levels of tramadol. None of the other polymorphism examined (genes: CYP2D6, CYP2B6, CYP3A, COMT, ABCB1, OPRM1) showed influence on tramadol pharmacokinetics

Conclusions the CYP2B6 genotype and altered tramadol concentration is novel. The power of this study is limited by small sample size; therefore, the results should be interpreted with caution

Wang et al. [46] (2019)

OPRM1 (rs1799971) Patients with OPRM1 Females (n ¼ 266) who genetic variations may (rs1799971) GG underwent cesarean influence postoperative genotype needed a section and given opioid requirements. greater dosage of opioids fentanyl and tramadol Additional studies are at 24 and 48 hours for postoperative needed to confirm this postoperatively versus analgesia were screened effect the AG and AA for COMT (rs4680) and genotypes. The AG OPRM1 (rs1799971) genotype required a polymorphisms to higher dose than the AA compare pain scores and genotype. COMT analgesic consumption (rs4680) genotypes did not exhibit a significant difference in postoperative opioid dosage

Haage et al. [39] (2018)

This study confirmed CYP2D6 PMs exhibited Healthy individuals (n ¼ CYP2D6 PMs altered large AUC of NDT and 19) given an oral dose of metabolic profile as low ODT. CYP2D6 IMs tramadol were assessed compared to IMs and exhibited a greater AUC for CYP2D6, CYP2B6, EMs. Additionally, the of NDT vs ODT. and CYP3A4 genotypes findings associated with Patients homozygous for to examine their effects CYP2B6 are important CYP2B6 *5 and *6 on tramadol because it is less widely alleles indicate reduced pharmacokinetics studied. However, enzyme function with these findings need to very low AUC of NDT be confirmed with a larger study population

Tanaka et al. [44] Japanese cancer patients No effect on the plasma (2018) (n ¼ 70) treated with concentration of tramadol were evaluated tramadol was seen in for CYP polymorphisms CYP2D6 genotypes.

This study confirmed CYP2D6 genotypes influence plasma concentrations of ODT (continued)

516

E. Paylor Sachtleben et al.

Table 1 (continued)

Author (year)

Groups studied and intervention to assess their effects on plasma concentrations of tramadol and its metabolites

Results and findings The concentration of ODT/tramadol ratio was reduced in CYP2D6 IMs and PMs versus NMs. The concentration of NDT/tramadol ratio was greater in CYP2D6 IMs and PMs versus NMs. No effect on the plasma concentration of tramadol or its metabolites was seen with CYP2B6*6 and CYP3A5*3 alleles

Conclusions and NDT. However, the less widely studied genotypes of CYP2B6 and CYP3A5 were not shown to affect tramadol metabolism

NM normal metabolizer, PM poor metabolizer, EM extensive metabolizer, IM intermediate metabolizer, UM ultrarapid metabolizer, ODT O-desmethyltramadol, NDT N-desmethyltramadol, CYP Cytochrome P450, AUC area under the curve, OPRM opioid mu-receptor, ABC adenosine triphosphate-binding cassette, COMT catechol-O-methyl transferase

metabolizers are at risk of inadequate analgesia postoperatively. Klepstad et al. reported that PMs had lower oxymorphone serum concentrations in comparison with EMs and UMs (P < 0.001) and a lower ratio of oxymorphone to oxycodone (P < 0.000), but ultimately there were no statistically significant differences between the phenotypes and the dosages required for analgesia and pain scores [53]. Zwisler et al. performed a study that analyzed genetic singlenucleotide polymorphisms A118G in OPRM1 (opioid receptor mu 1) and C3435T and G2677T/A in ABCB1 (encodes ABC1 transport protein for opioids) and the analgesic effect of intravenous oxycodone in postoperative pain. No significant clinical association was found between the tested genetic variants in OPRM1 and ABCB1 and the analgesic effect of oxycodone [54].

11

Diamorphine Diamorphine, also known as heroin, is a widely used and abused synthetic opioid receptor agonist. Diamorphine has exhibited use in controlling postoperative pain and pain associated with childbirth and in the treatment of opioid use disorder. This drug is metabolized via deacetylation into 6-monoacetylmorphine and morphine by carboxylesterase 1 (CES1) in the liver [55]. The dopaminergic system in the central nervous system is implicated in the abuse of several different substances, such as cocaine and

Pharmacogenomics in Postoperative Pain Management

517

heroin. One study suggests that a dopamine receptor (DRD2) genetic variant, called TaqIA, results in a lower density of the receptors it codes for. Thus these patients exhibit decreased dopaminergic activity in response to a stimulus, such as diamorphine [56]. The implications of decreased dopaminergic activity in patients with this genetic variant vary. Still, one may postulate that this may be linked to addiction. These patients require higher doses of diamorphine to produce a response similar to those who may not have lower dopaminergic activity.

12

Fentanyl Fentanyl is a synthetic opioid analgesic 50–100 times stronger than morphine and is a useful drug for managing postoperative pain and cancer-related pain, as it is available in a transdermal patch formulation. Fentanyl undergoes N-dealkylation to norfentanyl by CYP3A4 and CYP3A5. Patients display a wide variation in response to fentanyl, as they do with other opioids. In particular, patients homozygous for the genetic polymorphism CYP3A5*3 (CYP3A5*3/*3) exhibited an increased rate of absorption of transdermal fentanyl, with plasma levels two times higher than heterozygous patients (*1/*3) or wild type (*1/*1) [57]. Takashina et al. investigated the effect of polymorphisms in the ABCB1 transporters (C1236*T, C3435T, G2677A/T) on response to and toxicity of fentanyl in patients with cancer-related pain. Patients homozygous for C1236*T had decreased breakthrough pain medication requirements and, thus, a decreased need for rescue medications [57]. Barrat et al. investigated whether genetic variation in immune system activation and signaling pathways would have any differential effect of transdermal fentanyl concentration-response relationships to suitable pain control, cognitive dysfunction, and adverse effects because opioids have been known to trigger a response by the innate immune system. It was discovered that people with a polymorphism in a toll-like receptor gene, MYD88 rs6853, had a decreased risk of cognitive dysfunction in response to fentanyl [58].

13

Buprenorphine Buprenorphine acts as a partial agonist at the mu-opioid receptor (OPRM1) and the delta-opioid receptor (OPRD1) but acts as a full antagonist at the kappa-opioid receptor. This medication is useful in treating patients with opioid use disorder, but it displays a wide efficacy variability. Buprenorphine is metabolized to its active metabolite norbuprenorphine by CYP3A4 and CYP2C8. One study conducted on genetic variations of the OPRD1 among patients being treated with buprenorphine for opioid use disorder

518

E. Paylor Sachtleben et al.

with the OPRD1 rs678849 polymorphism showed increased relapse rates as evidenced by an increased likelihood of illicit opioids in urine drug screens of African American patients [59]. This study also found a positive correlation between treatment failure rates and the OPRD1 rs581111 and rs529250 polymorphisms in European American women [59].

14

Ketamine Ketamine is a dissociative anesthetic that exhibits its action by antagonizing NMDA receptors and inhibiting excitatory glutamate neurotransmission. At sub-anesthetic doses, ketamine also plays a significant role in the management of depression and pain. This drug is also of distinct interest due to its “opioid-sparing ability, by counteracting the central nervous system sensitization seen in opioid-induced hyperalgesia” [60]. Ketamine is metabolized to its active metabolite, norketamine, by the cytochrome P450 family of enzymes. Zhet et al. conducted a study that compared the enzymatic activity in 39 different CYP2C9 alleles and their effect on ketamine metabolism. This study found that patients with CYP2C9*40, *49, and *51 alleles had dramatically increased intrinsic clearance compared with CYP2C9*1 [60, 61]. A separate study conducted by Li et al. evaluated genetic variations in CYP2B6 genotypes and their effects on plasma clearance of ketamine. This study found that those who were homozygous for CYP2B6*6/*6 had a pointedly lower plasma clearance of ketamine compared with the CYP2B6*6/*1 CYP2B*1/*1 genotypes [62]. Decreased clearance of ketamine in patients with the CYP2B6*6/*6 genotype introduces the concern that these patients may be at higher risk of adverse effects such as hallucinations, dysphoria, dizziness, and diplopia [63].

15

Remifentanil Remifentanil was developed by pharmaceutical companies as a “soft drug” and acts much like fentanyl by binding to and activating the mu-opioid receptor but has a much shorter action duration than fentanyl. The term “soft drug” means that remifentanil is metabolically fragile and quickly removed from the body, thus allowing anesthesiologists to manipulate drug concentrations as needed [64]. Catechol-O-methyltransferase (COMT) is an enzyme that is involved in metabolizing the neurotransmitters dopamine and noradrenaline, and studies have revealed that patients with the Val158 genotype of the COMT gene have a higher activity of this enzyme, and this is associated with enhanced dopaminergic neurotransmission, thus conferring an advantage in the processing of aversive

Pharmacogenomics in Postoperative Pain Management

519

stimuli [65]. In contrast, the Met 158 genotype may be associated with increased pain sensitivity. Studies have shown decreased mu-opioid system activation with remifentanil, but only after repeated administration of painful stimuli and not after a single painful stimulus [65]. This suggests that the initial pain response is not influenced by the COMT gene but is apparent after the endogenous pain system is challenged with repeated painful stimuli. Landau et al. suggest that the amplified pain sensitivity in Met158 patients post-remifentanil administration may result from reduced efficacy of endogenous pain modulation or predisposition to opioid-induced hyperalgesia [65].

16

Escitalopram Escitalopram is an SSRI traditionally used in the treatment of mood disorders such as depression and anxiety. It can also be utilized for treating neuropathic pain. One study genotyped 34 patients in a placebo-controlled trial of escitalopram used to treat neuropathic pain, assessing for polymorphisms in five different genes. These genes included the serotonin receptor 2A (HTR2A) gene, serotonin receptor 2C (HTR2C) gene, ABCB1 gene encoding for the P-glycoprotein transporter, CYP2C19 gene, and the serotonin transporter gene (SLC6A4) [66]. The polymorphism rs6318 (cys23ser) in HTR2C exhibited a significant association with treatment response in both men and women, and this translated to improved pain relief during treatment with escitalopram [66]. See Table 2.

17

Conclusion Pharmacogenomics is an exciting, rapidly advancing field that potentially provides healthcare professionals with the opportunity to treat a patient’s pain in a holistic and more precise way that is uniquely adapted to the individual. As we begin to understand more about the interplay between genetic polymorphisms and the effects that an individual’s DNA has on their ability to metabolize drugs and process pain, we will be better able to tailor pain management regimens in a meaningful way that can result in clinically relevant improvements in outcome in postoperative pain management. One of the most compelling reasons for expanding research into the field of pharmacogenomics is to improve the quality of postoperative pain relief while also potentially limiting the possibility of drug dependence in certain patients. Genetics could play a role in helping to predict the response of certain individuals or patient populations to specific opioids. It could also be used to

520

E. Paylor Sachtleben et al.

Table 2 Pharmacogenomics of drugs used for pain relief Author (year)

Groups studied and intervention

Huddart et al. [2] (2018)

This article reviewed the pathways and enzymes involved in converting oxycodone to its active metabolite, noroxycodone

Klepstad et al. [3] (2011)

PMs had lower oxymorphone There were no statistically This study aimed to test significant differences serum concentrations in whether there is a between the phenotypes comparison with EMs and significant link between (PMs vs. EMs) and the UMs (P < 0.001) and a genetic variations and dosages required for lower ratio of morphine efficacy and to analgesia and pain scores oxymorphone to further expand the analysis oxycodone (P < 0.000) to opioids other than morphine

Zwisler et al. [4] (2012)

While this study exhibited The purpose of this study was No association was found differences in the rate of between the tested genetic to seek for an association metabolism between variants in OPRM1 and between the SNPs A118G different genetic variants, ABCB1 and changes in the in OPRM1 and C3435T no clinical significance was analgesic effect of and G2677T/A in the determined oxycodone ABCB1 transporter gene and the pain-relieving effect of oxycodone in postoperative pain

Lawford et al. [5] (2000)

Results suggest that a This study reviewed the dopamine receptor association between the (DRD2) genetic variant, presence of the TaqIA allele called TaqIA, results in a that codes for the DRD2 lower density of the receptor and the rate of receptors it codes for, and substance abuse compared thus these patients exhibit to those without this decreased dopaminergic genetic polymorphism activity in response to a stimulus, such as diamorphine

Results and findings

Conclusions

CYP3A4 and CYP2D6 metabolize oxycodone. Individuals can be divided into PMs, IMs, EMs, and UMs, depending on the metabolic activity of their CYP2D6 enzymes

Ultrarapid metabolizers may be at a greater risk for oxycodone toxicity. In contrast, those who are poor metabolizers are at risk of inadequate analgesia postoperatively, but this is still yet to be determined. More in-depth studies are required to determine if these different genotypes are clinically relevant

The results show that DRD2 variations are predictors of heroin use and possibly subsequent methadone treatment outcomes. This even proposes that there is a possible pharmacogenetic approach to the treatment of opioid use disorders

Takashina The study assessed the impact Patients homozygous for the Genetic polymorphisms in of CYP3A5 and ABCB1 genetic polymorphism the CYP3A5 enzyme and et al. [6] gene variations on CYP3A5*3 (CYP3A5*3/ ABCB1 transporter may (2012) transdermal fentanyl *3) exhibited an increased predict clinical responses (continued)

Pharmacogenomics in Postoperative Pain Management

521

Table 2 (continued) Author (year)

Groups studied and intervention pharmacokinetics and clinical responses in cancer patients

Barrat et al. [7] (2014)

Results and findings rate of absorption of transdermal fentanyl compared to those heterozygous (*1/*3) or wild type (*1/*1). Also, patients homozygous for the C1236*T allele of the ABC1B transporter had decreased breakthrough pain medication requirements

Conclusions to transdermal fentanyl in patients suffering from cancer pain

It was discovered that people This result demonstrates the This study proposed that relationship between with a polymorphism in a genetic variation in opioids and immune toll-like receptor gene, a immune system signaling system pathways while signaling molecule for the pathways would have an showing that genetic innate immune system, effect on transdermal variation in such signaling MYD88 rs6853, had a fentanyl response and its proteins influence decreased risk of cognitive relationship to suitable pain patients’ responses to dysfunction in response to control, cognitive fentanyl fentanyl dysfunction, and adverse effects

The findings of this study This study revealed that Crist et al. This study analyzed genetic suggest that genotyping variants in OPRD1, on the African American patients [8] individuals with opioid prevalence of opioid being treated for opioid (2013) use disorders looking for detection in UDS dependence with genetic variations in buprenorphine who carried different opioid receptors the OPRD1 rs678849 allele may assist providers in had a higher prevalence of choosing the best opioid detection on UDS. treatment plan for these This study also found a patients positive correlation between treatment failure rates and the OPRD1 rs581111 and rs529250 polymorphisms in European American women Zheng et al. [10] (2017)

This study that compared the This study found that patients It is not yet known what the clinical significance of this with CYP2C9*40, *49, and enzymatic activity different finding is yet, but the fact *51 alleles had dramatically CYP2C9 alleles and their increased intrinsic clearance that ketamine clearance effect on ketamine compared with CYP2C9*1 varies with different CYP metabolism polymorphisms may (continued)

522

E. Paylor Sachtleben et al.

Table 2 (continued) Author (year)

Groups studied and intervention

Results and findings

Conclusions impact the frequency of adverse effects, among other clinical effects of the drug

Li et al. [11] (2015)

This study evaluated genetic Those who were homozygous The finding that genetic variants of CYP2C9, variations in CYP2B6 for CYP2B6*6/*6 had a which metabolizes genotypes and their effects pointedly lower plasma ketamine, influences on plasma clearance of clearance of ketamine and plasma clearance of the ketamine when compared with the drug may be of clinical CYP2B6*6/*1 and significance in patients CYP2B*1/*1 genotypes being treated for pain, as decreased plasma clearance may lead to increased adverse effects. There is still further investigation to be done to determine the clinical significance of this

Landau et al. [14] (2012)

The article suggests that the This review article Met 158 genotype of the investigated the impact of COMT gene may be some genetic associated with increased polymorphisms of genes pain sensitivity, but only involved in pain and after repeated analgesic responses in administration of painful patients treated with stimuli and not after a remifentanil, the single painful stimulus prototypical “soft drug” in anesthesia

Brasch-

Andersen et al. [15] (2011)

This

finding suggests that variants in serotonin receptors influence patient’s perception of pain

The amplified pain sensitivity in Met158 patients post-remifentanil administration may be the result of reduced efficacy of endogenous pain modulation or predisposition to opioidinduced hyperalgesia

This study genotyped patients The polymorphism rs6318 (cys23ser) in HTR2C, a in a placebo-controlled trial serotonin receptor, of escitalopram used to exhibited a significant treat neuropathic pain, association with treatment assessing for response in both men and polymorphisms in five women and this translated different genes to improved pain relief during treatment with escitalopram

Pharmacogenomics in Postoperative Pain Management

523

predict the potential for the development of drug tolerance in a given patient [67]. While opioids remain a key form of pain management in the postoperative setting, it is of paramount importance that measures be taken to prevent dependence and death from opioid abuse. Since 2000, there has been a 200% increase in the rate of opioid-related deaths [68]. Additionally, the total amount of money spent on the treatment of opioid abuse and overdose is estimated to be approximately $78.5 billion annually [69]. The sheer cost of opioid abuse, both tangible and intangible, is undeniable, and the field of pharmacogenomics provides us with a key weapon in our arsenal for the fight against the opioid epidemic. Additionally, pharmacogenomics could be useful for increased precision in postoperative pain management. The subjective nature of postoperative pain makes it difficult to manage. However, using genetics effectively to determine which patients will respond to which medication regimens most effectively will take out some of the guess work involved in clinical pain management and help decrease the variations in response to pain medications that are largely subjective. Previous studies have established that differences in ethnicity [70] and gender [71] correspond to differences in response to pain. Further elucidating the genetic polymorphisms that correlate with pain response will help to improve pain relief for these different groups. In short, the use of pharmacogenomics in the management of postoperative pain is a moral and ethical imperative for healthcare providers and researchers. Having the tools to discover the secrets of the human genome and using them to develop effective treatment plans could save many lives and ease the pain of millions of people. Particularly, in terms of postoperative pain management, the use of pharmacogenomics would help to lessen the pain of the approximately 51 million Americans who undergo inpatient surgical procedures every year [72]. The application of pharmacogenomics in the clinical setting can provide us with the means to create unique and personalized treatment plans that are safer and more effective for the patient. References 1. Cornett EM, Carroll Turpin MA, Pinner A et al (2020) Pharmacogenomics of pain management: the impact of specific biological polymorphisms on drugs and metabolism. Curr Oncol Rep 22:18 2. Kaye AD, Garcia AJ, Hall OM et al (2019) Update on the pharmacogenomics of pain management. PGPM 12:125–143

3. Chaturvedi R, Alexander B, A’Court AM et al (2020) Genomics testing and personalized medicine in the preoperative setting: can it change outcomes in postoperative pain management? Best Pract Res Clin Anaesthesiol 34: 283–295 4. Aroke EN, Kittelsrud JM (2020) Pharmacogenetics of postoperative pain management: a review. AANA J 88:229–236

524

E. Paylor Sachtleben et al.

5. Nurse Anesthesiology (2020) A primer to pharmacogenetics of postoperative pain mana g e m e n t . A v a i l a b l e f r o m : h t t p s : // nurseanesthesiology.aana.com/a-primer-topharmacogenetics-of-postoperative-painmanagement 6. Nahin RL (2015) Estimates of pain prevalence and severity in adults: United States, 2012. J Pain 16:769–780 7. Brekke M, Hjortdahl P, Kvien TK (2002) Severity of musculoskeletal pain: relations to socioeconomic inequality. Soc Sci Med 54: 221–228 8. Fischer S, Doerr JM, Strahler J et al (2016) Stress exacerbates pain in the everyday lives of women with fibromyalgia syndrome – the role of cortisol and alpha-amylase. Psychoneuroendocrinology 63:68–77 9. Coulbault L, Beaussier M, Verstuyft C et al (2006) Environmental and genetic factors associated with morphine response in the postoperative period. Clin Pharmacol Ther 79: 316–324 10. Østera˚s B, Sigmundsson H, Haga M (2015) Perceived stress and musculoskeletal pain are prevalent and significantly associated in adolescents: an epidemiological cross-sectional study. BMC Public Health 15:1081 11. Periasamy S, Poovathai R, Pondiyadanar S (2014) Influences of gender on postoperative morphine consumption. J Clin Diagn Res 8: GC04–GC07 12. Si H, Yang T, Zeng Y et al (2017) Correlations between inflammatory cytokines, muscle damage markers and acute postoperative pain following primary total knee arthroplasty. BMC Musculoskelet Disord 18:265 13. Zhang J-M, An J (2007) Cytokines, inflammation, and pain. Int Anesthesiol Clin 45:27–37 14. Bair MJ, Robinson RL, Katon W et al (2003) Depression and pain comorbidity: a literature review. Arch Intern Med 163:2433–2445 15. Song T-J, Cho S-J, Kim W-J et al (2016) Anxiety and depression in tension-type headache: a population-based study. PLoS One 11: e0165316 16. Davies KA, Silman AJ, Macfarlane GJ et al (2009) The association between neighbourhood socio-economic status and the onset of chronic widespread pain: results from the EPIFUND study. Eur J Pain 13:635–640 17. Nielsen CS, Stubhaug A, Price DD et al (2008) Individual differences in pain sensitivity: genetic and environmental contributions. Pain 136:21–29

18. Mossey JM (2011) Defining racial and ethnic disparities in pain management. Clin Orthop Relat Res 469:1859–1870 19. Green CR, Anderson KO, Baker TA et al (2003) The unequal burden of pain: confronting racial and ethnic disparities in pain. Pain Med 4:277–294 20. Reyes-Gibby CC, Aday LA, Todd KH et al (2007) Pain in aging community-dwelling adults in the United States: non-hispanic whites, non-hispanic blacks, and hispanics. J Pain 8:75–84 21. Agu´ndez JAG, Blanca M, Cornejo-Garcı´a JA et al (2015) Pharmacogenomics of cyclooxygenases. Pharmacogenomics 16:501–522 22. Kaye AD, Garcia AJ, Hall OM et al (2019) Update on the pharmacogenomics of pain management. Pharmgenomics Pers Med 12: 125–143 23. Theken KN, Lee CR, Gong L et al (2020) Clinical Pharmacogenetics Implementation Consortium Guideline (CPIC) for CYP2C9 and nonsteroidal anti-inflammatory drugs. Clin Pharmacol Ther 108:191–200 24. Ramachandran A, Jaeschke H (2018) Acetaminophen toxicity: novel insights into mechanisms and future perspectives. Gene Expr 18:19–30 25. James LP, McGill MR, Roberts DW et al (2020) Advances in biomarker development in acetaminophen toxicity. Adv Clin Chem 98:35–50 26. Mazaleuskaya LL, Sangkuhl K, Thorn CF et al (2015) PharmGKB summary: pathways of acetaminophen metabolism at the therapeutic versus toxic doses. Pharmacogenet Genomics 25:416–426 27. Hermanns H, Hollmann MW, Stevens MF et al (2019) Molecular mechanisms of action of systemic lidocaine in acute and chronic pain: a narrative review. Br J Anaesth 123:335–349 28. Wang JS, Backman JT, Taavitsainen P et al (2000) Involvement of CYP1A2 and CYP3A4 in lidocaine N-deethylation and 3-hydroxylation in humans. Drug Metab Dispos 28:959–965 29. Fang P, Tang P-F, Xu R-A et al (2017) Functional assessment of CYP3A4 allelic variants on lidocaine metabolism in vitro. Drug Des Devel Ther 11:3503–3510 30. Zarei S, Costas Y, Orozco G et al (2020) A web-based pharmacogenomics search tool for precision medicine in perioperative care. J Pers Med 10:E65 31. Crews KR, Monte AA, Huddart R et al (2021) Clinical Pharmacogenetics Implementation

Pharmacogenomics in Postoperative Pain Management Consortium Guideline for CYP2D6, OPRM1, and COMT genotypes and select opioid therapy. Clin Pharmacol Ther 110(4):888–896 32. Awad ME, Padela MT, Sayeed Z et al (2019) Pharmacogenomic testing for postoperative pain optimization before total joint arthroplasty: a focus on drug-drug-gene interaction with commonly prescribed drugs and prior opioid use. JBJS Rev 7:e2 33. Lo¨tsch J, von Hentig N, Freynhagen R et al (2009) Cross-sectional analysis of the influence of currently known pharmacogenetic modulators on opioid therapy in outpatient pain centers. Pharmacogenet Genomics 19:429–436 34. Sistonen J, Madadi P, Ross CJ et al (2012) Prediction of codeine toxicity in infants and their mothers using a novel combination of maternal genetic markers. Clin Pharmacol Ther 91:692–699 35. Yang Z-Z, Li L, Wang L et al (2017) The regioselective glucuronidation of morphine by dimerized human UGT2B7, 1A1, 1A9 and their allelic variants. Acta Pharmacol Sin 38: 1184–1194 36. Nielsen LM, Sverrisdo´ttir E, Stage TB et al (2017) Lack of genetic association between OCT1, ABCB1, and UGT2B7 variants and morphine pharmacokinetics. Eur J Pharm Sci 99:337–342 37. Bastami S, Gupta A, Zackrisson A-L et al (2014) Influence of UGT2B7, OPRM1 and ABCB1 gene polymorphisms on postoperative morphine consumption. Basic Clin Pharmacol Toxicol 115:423–431 38. Ning M, Tao Y, Hu X et al (2019) Roles of UGT2B7 C802T gene polymorphism on the efficacy of morphine treatment on cancer pain among the Chinese han population. Niger J Clin Pract 22:1319–1323 39. Haage P, Kronstrand R, Josefsson M et al (2018) Enantioselective pharmacokinetics of tramadol and its three main metabolites; impact of CYP2D6, CYP2B6, and CYP3A4 genotype. Pharmacol Res Perspect 6:e00419 40. Rollason V, Lloret-Linares C, Lorenzini KI et al (2020) Evaluation of phenotypic and genotypic variations of drug metabolising enzymes and transporters in chronic pain patients facing adverse drug reactions or non-response to analgesics: a retrospective study. J Pers Med 10:198 41. Lee J, Yoo H-D, Bae J-W et al (2019) Population pharmacokinetic analysis of tramadol and O-desmethyltramadol with genetic polymorphism of CYP2D6. Drug Des Devel Ther 13: 1751–1761

525

42. Wendt FR, Novroski NMM, Rahikainen A-L et al (2019) A pathway-driven predictive model of tramadol pharmacogenetics. Eur J Hum Genet 27:1143–1156 43. Saiz-Rodrı´guez M, Ochoa D, Roma´n M et al (2020) Involvement of CYP2D6 and CYP2B6 on tramadol pharmacokinetics. Pharmacogenomics 21:663–675 44. Tanaka H, Naito T, Sato H et al (2018) Impact of CYP genotype and inflammatory markers on the plasma concentrations of tramadol and its demethylated metabolites and drug tolerability in cancer patients. Eur J Clin Pharmacol 74: 1461–1469 45. Stamer UM, Musshoff F, Stu¨ber F et al (2016) Loss-of-function polymorphisms in the organic cation transporter OCT1 are associated with reduced postoperative tramadol consumption. Pain 157:2467–2475 46. Wang L, Wei C, Xiao F et al (2019) Influences of COMT rs4680 and OPRM1 rs1799971 polymorphisms on chronic postsurgical pain, acute pain, and analgesic consumption after elective cesarean delivery. Clin J Pain 35:31–36 47. Hosseinnejad K, Yin T, Gaskins JT et al (2019) Lack of influence by CYP3A4 and CYP3A5 genotypes on pain relief by hydrocodone in postoperative cesarean section pain management. J Appl Lab Med 3:954–964 48. Smith DM, Weitzel KW, Cavallari LH et al (2018) Clinical application of pharmacogenetics in pain management. Per Med 15: 117–126 49. Stauble ME, Moore AW, Langman LJ et al (2014) Hydrocodone in postoperative personalized pain management: pro-drug or drug? Clin Chim Acta 429:26–29 50. Boswell MV, Stauble ME, Loyd GE et al (2013) The role of hydromorphone and OPRM1 in postoperative pain relief with hydrocodone. Pain Physician 16:E227–E235 51. Umukoro NN, Aruldhas BW, Rossos R et al (2021) Pharmacogenomics of oxycodone: a narrative literature review. Pharmacogenomics 22:275–290 52. Huddart R, Clarke M, Altman RB et al (2018) PharmGKB summary. Pharmacogenet Genomics 28:230–237 53. Klepstad P, Fladvad T, Skorpen F et al (2011) Influence from genetic variability on opioid use for cancer pain: a European genetic association study of 2294 cancer pain patients. Pain 152: 1139–1145 54. Zwisler ST, Enggaard TP, Mikkelsen S et al (2012) Lack of association of OPRM1 and ABCB1 single-nucleotide polymorphisms to

526

E. Paylor Sachtleben et al.

oxycodone response in postoperative pain. J Clin Pharmacol 52:234–242 55. Qian Y, Gilliland TK, Markowitz JS (2020) The influence of carboxylesterase 1 polymorphism and cannabidiol on the hepatic metabolism of heroin. Chem Biol Interact 316: 108914 56. Lawford BR, Young RM, Noble EP et al (2000) The D2 dopamine receptor A1 allele and opioid dependence: association with heroin use and response to methadone treatment. Am J Med Genet 96:592–598 57. Takashina Y, Naito T, Mino Y et al (2012) Impact of CYP3A5 and ABCB1 gene polymorphisms on fentanyl pharmacokinetics and clinical responses in cancer patients undergoing conversion to a transdermal system. Drug Metab Pharmacokinet 27:414–421 58. Barratt DT, Bandak B, Klepstad P et al (2014) Genetic, pathological and physiological determinants of transdermal fentanyl pharmacokinetics in 620 cancer patients of the EPOS study. Pharmacogenet Genomics 24:185–194 59. Crist RC, Clarke T-K, Ang A et al (2013) An intronic variant in OPRD1 predicts treatment outcome for opioid dependence in AfricanAmericans. Neuropsychopharmacology 38: 2003–2010 60. Culp C, Kim HK, Abdi S (2021) Ketamine use for cancer and chronic pain management. Front Pharmacol 11:599721 61. Zheng X, Fang P, Bao S et al (2017) Function of 38 variants CYP2C9 polymorphism on ketamine metabolism in vitro. J Pharmacol Sci 135: 8–13 62. Li Y, Jackson KA, Slon B et al (2015) CYP2B6 * 6 allele and age substantially reduce steadystate ketamine clearance in chronic pain patients: impact on adverse effects. Br J Clin Pharmacol 80:276–284 63. Stoker AD, Rosenfeld DM, Buras MR et al (2019) Evaluation of clinical factors associated with adverse drug events in patients receiving sub-anesthetic ketamine infusions. J Pain Res 12:3413–3421 64. Birgenheier NM, Stuart AR, Egan TD (2020) Soft drugs in anesthesia: remifentanil as

prototype to modern anesthetic drug development. Curr Opin Anaesthesiol 33:499–505 65. Landau R (2012) Pharmacogenomic considerations in opioid analgesia. Pharmgenomics Pers Med 5:73–87 66. Brasch-Andersen C, Møller MU, Christiansen L et al (2011) A candidate gene study of serotonergic pathway genes and pain relief during treatment with escitalopram in patients with neuropathic pain shows significant association to serotonin receptor2C (HTR2C). Eur J Clin Pharmacol 67:1131–1137 67. Trescot AM, Faynboym S (2014) A review of the role of genetic testing in pain medicine. Pain Physician 17:425–445 68. Rudd RA, Aleshire N, Zibbell JE et al (2016) Increases in drug and opioid overdose deaths – United States, 2000-2014. MMWR Morb Mortal Wkly Rep 64:1378–1382 69. Florence C, Luo F, Xu L et al (2016) The economic burden of prescription opioid overdose, abuse and dependence in the United States, 2013. Med Care 54:901–906 70. Faucett J, Gordon N, Levine J (1994) Differences in postoperative pain severity among four ethnic groups. J Pain Symptom Manag 9: 383–389 71. Fillingim RB (2000) Sex, gender, and pain: women and men really are different. Curr Rev Pain 4:24–30 72. Hah JM, Bateman BT, Ratliff J et al (2017) Chronic opioid use after surgery: implications for perioperative management in the face of the opioid epidemic. Anesth Analg 125: 1733–1740 73. Neskovic N, Mandic D, Marczi S et al (2021) Different pharmacokinetics of tramadol, O-demethyltramadol and N-demethyltramadol in postoperative surgical patients from those observed in medical patients. Front Pharmacol 12:656748 74. Smith DM, Weitzel KW, Elsey AR et al (2019) CYP2D6-guided opioid therapy improves pain control in CYP2D6 intermediate and poor metabolizers: a pragmatic clinical trial. Genet Med 21:1842–1850

Chapter 19 Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis Atinuke Aluko and Prabha Ranganathan Abstract Rheumatoid arthritis (RA) is a chronic systemic inflammatory disorder that can lead to severe joint damage and is often associated with a high morbidity and disability. Disease-modifying anti-rheumatic drugs (DMARDs) are the mainstay of treatment in RA. DMARDs not only relieve the clinical signs and symptoms of RA but also inhibit the radiographic progression of disease and reduce the effects of chronic systemic inflammation. Since the introduction of biologic DMARDs in the late 1990s, the therapeutic range of options for the management of RA has significantly expanded. However, patients’ response to these agents is not uniform with considerable variability in both efficacy and toxicity. There are no reliable means of predicting an individual patient’s response to a given DMARD prior to initiation of therapy. In this chapter, the current published literature on the pharmacogenetics of traditional DMARDS and the newer biologic DMARDs in RA is highlighted. Pharmacogenetics may help individualize drug therapy in patients with RA by providing reliable biomarkers to predict medication toxicity and efficacy. Key words Pharmacogenetics, Polymorphisms, Rheumatoid arthritis, Methotrexate, Azathioprine, Sulfasalazine, Tumor necrosis factor antagonists, Rituximab, Tocilizumab

1

Introduction Rheumatoid arthritis (RA) is a chronic systemic inflammatory arthritis that occurs in about 1% of the population worldwide. Untreated, RA is associated with joint destruction, disability, and increased mortality. Early detection and therapy with diseasemodifying anti-rheumatic drugs (DMARDs) is critical in preventing these sequelae of RA. With the advent of biologic DMARDs, which are effective albeit expensive therapies for RA, there has been a focus on developing methods including those based on pharmacogenetics to predict a priori, an individual patient’s response to a given DMARD. This chapter will highlight some of the recent, major publications in the field of pharmacogenetics in RA and describe the implications of this field for future research and clinical care. The

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_19, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

527

528

Atinuke Aluko and Prabha Ranganathan

pharmacogenetics of three major nonbiologic DMARDs (methotrexate, azathioprine, and sulfasalazine) as well as one class of biologic DMARDs (tumor necrosis factor antagonists) and two other biologic DMARDs, rituximab and tocilizumab, in RA will be reviewed.

2

Pharmacogenetics of Drugs in RA

2.1 Pharmacogenetics of Methotrexate

For the past two decades, methotrexate (MTX) has remained the cornerstone DMARD used in RA because of its well established efficacy [1]. However, the response to MTX has a wide variation ranging from 46% to 65% [2], and about 30% of patients develop toxicities that result in medication discontinuation within the first 12 months. A prospective study performed in the United Kingdom evaluated MTX failure rates and associated factors. MTX failure was defined as discontinuation of MTX due to patient reported adverse effects or inefficacy. Of 431 patients enrolled, all DMARD naı¨ve, 143 discontinued MTX due to inefficacy and 67 due to adverse effects with 28 of these being gastrointestinal-related adverse effects. Female gender, younger age, rheumatoid factor seropositivity, and higher disease activity were associated with earlier discontinuation due to inefficacy, while patients with a positive rheumatoid factor were less likely to discontinue MTX due to adverse effects [3]. The exact mechanism of action of MTX in RA remains unclear; it is likely that MTX’s anti-inflammatory effects in RA occur due to its effects on the intracellular folate and adenosine pathway in immune mediator cells. MTX is actively transported into the cell by solute carrier 19A1 (SLC 19A1), also called reduced folate carrier 1 (RFC1) (Fig. 1). MTX is pumped out of the cell by members of the ATP binding cassette (ABC) family of transporters, also known as multidrugresistant transporters (MDRs), and multidrug resistance-associated proteins (MRPs) [4]. Intracellular MTX is polyglutamated by the enzyme folylpolyglutamyl synthase (FPGS). This process can be reversed by gamma glutamyl hydrolase (GGH). Polyglutamation of MTX (MTXPGn) helps retain MTX within the cell preventing drug efflux by the ABC transporters. Decreased FPGS activity occurs in MTX-resistant cells [5]. MTXPGs inhibit dihydrofolate reductase (DHFR), which catalyzes the reduction of dihydrofolate to tetrahydrofolate (THF) [6]. THF is converted to 5, 10 methylene tetrahydrofolate (5, 10-CH2-THF) and subsequently to 5-methyl THF (5-CH3-THF) by methylene tetrahydrofolate reductase (MTHFR). 5-Methyl THF is a biologically active folate cofactor that functions as a one-carbon donor for many important cellular reactions, including the conversion of homocysteine to methionine. MTXPGs also inhibit thymidylate synthase (TYMS),

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis adenosine

cell membrane

Methotrexate

ATP adenosine

ADP

ADA inosine

Methotrexate

AMP de novo purine synthesis

Methotrexate

ATIC

MTX-PG

DNA

529

IMP

AMP deaminase

FAICAR AICAR

FH4

dTMP

MTR de novo pyrimidine synthesis

dUMP

5,10-CH2-THF FH 2

MTHFR 5-CH3-THF

Fig. 1 Cellular pathway of methotrexate. RFC1 reduced folate carrier 1, ABCB1 ABCC1-4, ABC transporters, GGH γ-glutamyl hydrolase, FPGS folylpolyglutamyl synthase, MTX-PG methotrexate polyglutamate, TYMS thymidylate synthase, d UMP deoxyuridine monophosphate, d TMP deoxythymidine monophosphate, DHFR dihydrofolate reductase, FH2 dihydrofolate, 5-CH3-THF 5-methyl tetrahydrofolate, MTHFR methylene tetrahydrofolate reductase, 5, 10-CH2-THF 5, 10-methylene tetrahydrofolate, MTR methyl tetrahydrofolate reductase, AICAR aminoimidazole carboxamide ribonucleotide, FAICAR 10-formyl AICAR, ATIC AICAR transformylase, IMP inosine monophosphate, AMP adenosine monophosphate, ADP adenosine diphosphate, ATP adenosine triphosphate, ADA adenosine deaminase. Italicized genes have been targets of pharmacogenetic analyses in studies published so far. (Reproduced from Ranganathan and McLeod [161] by permission of John Wiley & Sons, Inc)

which converts deoxyuridylate to deoxythymidylate in the de novo pyrimidine synthetic pathway [7]. MTX also has several effects on the purine synthetic pathway. MTXPGs inhibit the enzyme aminoimidazole carboxamide ribonucleotide (AICAR) transformylase which leads to intracellular accumulation of its substrate, AICAR. AICAR and its metabolites in turn inhibit two enzymes in the adenosine pathway, adenosine deaminase (ADA) and adenosine monophosphate (AMP) deaminase, which leads to intracellular accumulation of adenosine and adenine nucleotides. Subsequent dephosphorylation of these nucleotides results in increased extracellular concentrations of adenosine which is a powerful anti-inflammatory agent [8].

530

Atinuke Aluko and Prabha Ranganathan

Polymorphisms in genes encoding MTX transporters and enzymes in the folate and adenosine pathways inhibited by MTX can lead to differences in response to the medication. Due to the clinical implications for patient management and achieving remission or low disease activity in RA, this has been an increasing area of research focus. 2.1.1 Transporter Genes: RFC1/GGH/ABCB1

RFC1 transports MTX into the cell. Polymorphisms that inactivate the RFC1 enzyme or change the function of transcription factors leading to loss of RFC1 gene expression can alter MTX transport [9, 10]. The RFC1 gene is a 27 kb gene located on chromosome 21 (21q22.3). An 80 G>A polymorphism leading to substitution of arginine for histidine at codon 27 in the first transmembrane domain (TMD1) of the RFC1 protein and a 61 base pair (bp) repeat polymorphism in the RFC1 promoter, causing increased transcriptional activity of the promoter, have been described [11]. In a study by Dervieux et al., the effect of the G80A single-nucleotide polymorphism (SNP) on response to MTX in RA patients was examined. Patients homozygous for the RFC SNP 80A/A had a greater response to MTX compared to patients carrying the wild-type 80G/G SNP. Thus the RFC 80AA SNP may be a marker of increased response to MTX in RA [12]. SNPs exist in the GGH promoter that influence GGH expression and MTX polyglutamation [12]. A 452C>T SNP leading to decreased GGH activity and accumulation of intracellular longchain MTX polyglutamates and a 401C>T promoter polymorphism, also altering intracellular MTXPG levels, have been described [12]. There is a GGH 16T>C polymorphism whose functional effects are unknown. A Japanese study of patients with RA demonstrated that the presence of the RFC 80AA and GGH– 401TT genotypes independently predicted MTXPG levels. Patients carrying the RFC 80AA genotype were 3.4-fold more likely to have MTXPG levels above the group median compared to patients with the 80GG or 80GA genotype (odds ratio (OR) 3.4, 95% CI 1.4–8.4; p ¼ 0.007). In contrast, patients with the GGH–401TT genotype were 4.8-fold (OR 4.8, 95% CI 1.8–13.0; p ¼ 0.002) more likely to have MTXPG below the study median compared to those who carried the GGH–401CC or CT genotype. Thus both the GGH 401 C>T and RFC 80G>A SNPs influenced intracellular MTXPG levels and thereby may predict MTX response in RA [12, 13].

2.1.2

The ABCB1 gene is a 209 kb gene located on chromosome 7 (7q21.1). P-Glycoprotein (P-gp) the product of the ABCB1 gene is a membrane transporter important in the transport of several drugs, including MTX. SNPs in the ABCB1 gene have been identified and their effects on P-gp expression studied [14]. The 3435C>T SNP is a synonymous SNP in exon 26 of the

ABCB1

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

531

ABCB1 gene. It is often linked to a 2677G>T SNP in exon 21 which results in substitution of alanine at position 893 by serine or threonine [15]. It is unclear whether variations in ABCB1 and/or P-gp expression impact MTX efflux from the cell. Although there is lack of published data to support that ABCB1 SNPs influence MTX cellular transport directly, some studies suggest that higher P-gp expression may mediate MTX resistance [16], while other studies do not support this [17]. Considering the linkage of the two SNPs, haplotype analyses may be more helpful in examining the role of these genetic variants in influencing MTX’s effects. 92 RA patients and 97 healthy controls were genotyped for the 3435C>T polymorphism in one study. Patients who had active RA (n ¼ 62) after 6 months of treatment with MTX (7.5–15 mg/ week) and prednisone (5–10 mg daily) and those who responded after 6 months of the same treatment (n ¼ 30) were classified as two groups and studied. The ABCB1 genotypes were distributed similarly among the RA patients and controls. Patients with the 3435CC and 3435CT genotypes were more likely to have active RA compared to patients with the 3435TT genotype (OR 2.89, CI 0.87–9.7; p < 0.05). Thus, the presence of the 3435T allele seemed to be protective in that patients homozygous for this allele had less severe RA that was more responsive to MTX and prednisone [18]. A study of 194 patients with RA on MTX monotherapy analyzed the effects of three SNPs in the ABCB1 gene in addition to those previously mentioned above. Although there was no significant association of these SNPs with response to MTX, there was a higher risk of overall toxicity with the ABCB1 rs868755 G>T genotype. The risk was higher in homozygous T/T patients (65.5%) than in heterozygous (40.2%) and wild-type patients (36.8%) ( p ¼ 0.025) [19]. 2.1.3 MTHFR/DHFR/ FPGS/TYMS

The MTHFR gene is a 19 kb gene located on chromosome 1 (1p36.3). Of the several MTHFR polymorphisms that have been identified [20], two polymorphisms, the 677C>T and 1298A>C polymorphisms, have been well studied for their influence on MTX’s clinical effects. The 677C>T polymorphism leads to a substitution of alanine with valine in the codon at nucleotide 677 of the MTHFR gene [21]. This change results in the production of a thermolabile variant of MTHFR with resultant reduction in enzymatic activity. The 1298A>C polymorphism causes a substitution of glutamine with alanine in the codon at nucleotide 1298 and also leads to reduced enzymatic activity of MTFHR [22]. As MTHFR is important in the generation of 5-methyl THF (Fig. 1), which acts as the carbon donor for the remethylation of homocysteine to methionine, these two SNPs by reducing MTHFR activity can increase plasma homocysteine levels [23].

532

Atinuke Aluko and Prabha Ranganathan

Elevated plasma homocysteine levels mediated by these two SNPs may exacerbate MTX’s toxic effects. Several studies have evaluated the effects of these SNPs on MTX response. One study examined 105 patients with RA, 35 of whom were treated with MTX (7.5–15 mg/week), 34 with sulfasalazine (SSZ) (2–3 g/day), and 36 with MTX and SSZ. All patients were genotyped for the 677C>T SNP and their plasma homocysteine levels measured. Patients on MTX had higher plasma homocysteine levels than those on SSZ alone, but those on both MTX and SSZ had the highest levels. Patients heterozygous for the 677C>T SNP had higher plasma homocysteine levels after 1 year than patients without the SNP. Patients homozygous for the SNP had a higher plasma homocysteine level at baseline which did not change significantly. Elevated plasma homocysteine levels (17%, p < 0.05) were found in patients with gastrointestinal (GI) side effects from MTX, such as nausea, abdominal pain, and discomfort, compared to patients without side effects. Patients on MTX and SSZ had the highest homocysteine levels and the highest incidence of GI side effects. However, the presence of the C677T SNP was not directly associated with the occurrence of GI events. This study suggests that plasma homocysteine levels (exacerbated by the presence of the MTHFR 677C>T SNP) may influence the GI toxicity of MTX [24]. In another study 236 patients with RA on MTX were genotyped for the 677C>T SNP. MTX was initiated at 7.5 mg/week and titrated to a maximum dose of 25 mg/week. Patients were assessed for MTX toxicity and disease activity periodically. 122 of 236 patients (52%) did not have the SNP, 19 patients (8%) were homozygous, and 95 patients (40%) heterozygous for the polymorphism. Patients who were homozygous and heterozygous for the 677C>T SNP had an increased risk of discontinuing MTX due to adverse events (relative risk (RR) 2.01, CI 1.09–3.70) particularly hepatotoxicity (RR 2.38, CI 1.06–5.34). This effect of the genotype on MTX toxicity was also evident in patients on folate supplementation in the study. However, there was no effect of the 677C>T genotype on MTX efficacy [25]. In a cross-sectional study, 93 RA patients treated with MTX (average dose 11.9 mg/week) and 377 healthy controls were genotyped for the 677C>T and 1298A>C polymorphisms and assessed for RA disease activity as well as MTX toxicity. Serum folate and plasma homocysteine levels were measured. More RA patients carried the 1298CC genotype (24.7%) than the controls (12.8%), and this was statistically significant ( p < 0.001). There were interesting effects of the 1298CC genotype on MTX toxicity but not efficacy. Homozygotes for the 1298C SNP appeared to be protected from MTX toxicity; 33% did not experience toxicity, only 9.1% had adverse reactions ( p ¼ 0.035). In contrast, patients with the AA

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

533

genotype were five times more likely to develop toxicities than those with the CC genotype (OR 5.24, CI 1.38–20). Also, patients carrying the CC genotype had higher plasma homocysteine levels than patients with AA or AC genotype, and this was not influenced by serum folate levels. The 677C>T polymorphism had no effects on MTX toxicity or efficacy in this study. This study suggests that homozygosity for the C1298 allele increases susceptibility to RA but also protects from MTX toxicity presumably via a homocysteine-dependent mechanism [26]. Three studies demonstrated an effect of the 677C>T polymorphism on MTX efficacy. However, results from these studies were conflicting with the T allele being a marker of both decreased and increased MTX efficacy in US [27] and Polish [28] cohorts, respectively, and the C allele a marker of increased MTX efficacy in a Dutch cohort [29]. Eight studies showed an effect of the T allele on MTX toxicity. Four of these studies examined Asian patients which included Japanese [30], Korean [31], and Chinese [32]; others included Dutch [25], US [33, 34], and Spanish cohorts [35]. A few meta-analyses have also yielded disparate results. One meta-analysis found an association between the 677C>T polymorphism and methotrexate toxicity, but no such association for the 1298A>C variant [36]. However, two meta-analyses (each with over 1000 patients with RA) found no association between either of these polymorphisms, methotrexate efficacy, and toxicity [37, 38]. An analysis of the effect of SNPs in FPGS, performed in 194 patients with RA on MTX monotherapy, revealed a reduction in disease activity (using Disease Activity Score 28, DAS 28, at onset and 6 months after MTX initiation) in patients with 2 SNPs in the FPGS gene, rs10987742 G>A ( p ¼ 0.033) and rs10106 T>C ( p ¼ 0.041). In concordance with this, patients with the rs10106 T>C genotype, but not rs10987742 G>A, remained on MTX monotherapy for a significantly longer duration of time (77–144 months, in those with the TT genotype than those with the C allele). There was no significant reduction in disease activity based on the effect of these SNPs on additional clinical variables such as CRP. Patients with the rs10106 T>C polymorphism in the FPGS gene had more side effects; those with the TT genotype had more side effects than carriers of the C allele (55% vs 45%, p ¼ 0.021) [19]. TYMS is an important enzyme in the de novo synthesis of pyrimidines. It converts deoxyuridine monophosphate (dUMP) to deoxythymidine monophosphate (dTMP) and is a direct target of polyglutamated MTX. The TYMS gene is a 15 kb gene located on chromosome 18 (18p11.32). A polymorphic tandem 28 bp repeat sequence has been described in the 50 untranslated region (TSER) of the TYMS gene with a variable number of repeat

534

Atinuke Aluko and Prabha Ranganathan

elements [39]. In vitro studies have shown increase in the enzymatic activity of TYMS with increasing number of bp repeats due to the presence of a putative enhancer box (E-box) sequence at the bp repeat region [39–41]. Patients homozygous for the triple repeat allele (TSER*3/*3 or 3R) have higher TYMS mRNA expression compared to patients homozygous for the double repeat allele (TSER*2/*2 or 2R) [41, 42]. An additional G to C substitution within the 3R allele further diversifies 3R into 3RC and 3RG; the C allele abolishes a critical residue of the upstream stimulatory factor (USF)-binding site lowering TYMS activity in 3RC carriers [43]. Deletion of a 6 bp sequence at nucleotide 1494 in the 30 untranslated region (30 UTR) of TYMS has been described and may be associated with decreased TYMS mRNA stability and expression [44]. 167 patients with RA, 115 of whom were treated with MTX, were genotyped for the following polymorphisms – TYMS 50 UTR enhancer repeat (TSER), 30 UTR deletion, MTHFR 677C>T, and 1298A>C. The mean weekly MTX dose in this study was 5.7  2.3 mg. Information on MTX toxicity data was collected retrospectively. Both MTX treated and untreated groups displayed similar frequencies of these SNPs. The TYMS and MTHFR polymorphisms were not associated with toxicity, although a significant percentage (45%) of patients on MTX experienced adverse effects. Weekly MTX dose (rather than standardized disease activity measures) was used as a marker of efficacy in this study. A dose of >6 mg per week was considered indicative of less efficacy, and T and 1298 A>C) genes on MTX efficacy and adverse effects. There was no relationship between the presence of these SNPs and MTX response. However, in this population, there was an increased risk of adverse events, most commonly gastrointestinal, in patients with MTHFR 677C>T [higher risk with TT genotype compared with CT or CC ( p ¼ 0.04, OR ¼ 2.20, 95% CI: 1.01–4.77)] [46]. DHFR reduces dihydrofolate to tetrahydrofolate in the intracellular folate pathway. It is directly inhibited by polyglutamated

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

535

MTX and encoded by the DHFR gene which is a 28 kb gene located on chromosome 5 (5q11.2-q13.2). DHFR gene polymorphisms were studied in 205 MTX-treated RA patients. MTX was started at 7.5 mg per week and increased to 15 mg per week (with folic acid supplementation) after 4 weeks based on response to the drug. MTX efficacy and toxicity (gastrointestinal side effects, elevated liver enzyme levels, skin and mucosal disorders, pneumonitis, and overall adverse drug events) were assessed periodically. Genotyping for the MTHFR 677C>T, MTHFR 1298A>C, DHFR 473G>A, DHFR 35289G>A, and RFC 80 G>A SNPs was performed. At 6 months, patients carrying the MTHFR 1298AA and MTHFR 677CC (wild type) genotypes showed a greater response to MTX compared to patients carrying the heterozygous or homogenous genotype (OR 2.3, CI 1.18–4.41 and OR 2.73, CI 1.03–7.26, respectively). Haplotype analysis for the MTHFR 1298A and 677C SNPs revealed that patients with two copies of the haplotype had greater improvement than those with one or no copies of the haplotype (OR 3.0, CI 1.4–6.4). Patients homozygous and heterozygous for the MTHFR 1298 SNP (MTHFR 1298AC+CC) had an increased number of overall adverse drug events at 3 and 6 months (OR 2.55, CI 1.20–5.41 and OR 2.5, CI 1.32–4.72, respectively) compared to those with other genotypes. The RFC and DHFR SNPs were not associated with MTX toxicity or efficacy. Thus patients with the wild-type MTHFR alleles (MTHFR 1298AA and 677CC) responded better to MTX, while those with the 1298C allele had an increased risk for MTX toxicity [29]. In general, based on the literature cited above (Table 1), the 677C>T SNP in MTHFR appears to have effects on MTX toxicity, presumably through its effects on homocysteine metabolism [24, 25] and on MTX efficacy [29]. The effects of the 1298A>C polymorphism on MTX are less clear, with data suggesting that it may impact MTX efficacy and toxicity [26, 29]. The seemingly inconsistent results of these studies may stem from the fact that these SNPs may have other effects, outside of homocysteine metabolism, which may influence response to MTX. Some of these studies were retrospective which may also have led to inaccuracies in the assessment of MTX effects, particularly adverse effects. Although one of the studies [45]concluded that MTHFR SNPs did not affect MTX efficacy or toxicity, it is worth pointing out that the doses of MTX used in this study were small (6 mg/week) which may have masked the differences in MTX response between patient groups and the development of adverse events. Also, MTX efficacy was not assessed using standardized measures of disease activity in this study; rather MTX dose was used as a surrogate marker of MTX efficacy. Race may also play an important role in the effects of these polymorphisms in different populations [34].

536

Atinuke Aluko and Prabha Ranganathan

Table 1 Pharmacogenetics of methotrexate in RA Postulated effect of Role in MTX pathway Polymorphism polymorphism

Clinical effects

RFC1

Active transport of MTX into cell

80G>A

Increased transcriptional activity of RFC1 gene; increased MTX entry into cell

Increased response to MTX

ABCB1

MTX efflux from the cell

3435C>T

Increased MTX entry into cell Increased response to MTX

MTHFR Important in regeneration of reduced folate; indirectly inhibited by MTX

677C>T

Increased GI side Thermolabile MTHFR with effects decreased activity; increased plasma homocysteine Increased hepatic toxicity, GI toxicity, alopecia, stomatitis, and rash No effect on toxicity No effect on efficacy or toxicity

MTHFR

1298A>C

Decreased MTHFR activity; increased plasma homocysteine

Increased MTX efficacy Increased susceptibility to RA; decreased MTX toxicity No effect on efficacy or toxicity

473G>A 35289G>A

Decreased DHFR activity

No effect on MTX efficacy or toxicity

Decreased ATIC activity; increased AICAR accumulation; increased adenosine

Increased MTX efficacy

Gene

DHFR

Reduction of dihydrofolate to tetrahydrofolate

ATIC

Conversion of AICAR 347C>G to 10-formyl AICAR; directly inhibited by MTX

TYMS

Conversion of dUMP to dTMP; directly inhibited by MTX

50 UTR 28-bp repeat (TSER) 30 UTR 6-bp deletion

Increased TYMS enzyme activity Decreased TYMS mRNA stability and expression

Decreased MTX efficacy Increased MTX efficacy

GGH

Reversal of polyglutamation of MTX

452C>T

Decreased GGH activity, decreased accumulation of intracellular long-chain MTX polyglutamates Increased GGH activity

Increased MTX efficacy

3 G/3 G genotype

Decreased response to MTX and increased bone marrow toxicity

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis 2.1.4

CCND1/ATIC

537

G1/S-specific cyclin-D1 is a protein that in humans is encoded by the CCND1 gene, its activity is required for cell cycle G1/S transition [47]. The 870A>G substitution resides at the mRNA splicing site [48]. The A allele preferentially encodes the transcript lacking the exon 5 leading to an increased level and a longer half-life of cyclin D1 [49]. Elevated levels or expression of cyclin D1 protein (coded by the A allele) have been found in a variety of cancers, including breast cancer, head and neck cancer, non-small cell lung cancer, and mantle cell lymphomas [50]. Hochhauser et al., using a human fibrosarcoma HT1080 cell line, showed that CCND1 overexpression affected sensitivity to MTX [50]. 184 RA patients treated with MTX were genotyped for selected polymorphisms in the GGH (354G>T and 452C>T), CCND1 (870A>G), and TYMS (variable number of tandem repeats [VNTR] and G to C substitution of the triple repeat, 3R allele) genes. Based on the European League Against Rheumatism (EULAR, a standardized measure of disease activity in RA) response criteria, 146 RA patients (79.3%) were classified as responders and 38 (20.7%) as non-responders after 6 months of MTX therapy. There was no difference in the frequency of polymorphisms in the GGH and CCND1 genes or TYMS VNTRs between MTX responders and non-responders. However, when the TYMS gene was analyzed with respect to VNTRs and the 3R G to C substitution, a higher frequency of the 3 G/3 G genotype was found in MTX non-responders when compared to other genotypes; p ¼ 0.02, OR 5.4, 95% CI 1.0–21.1 [51]. 8/184 patients developed bone marrow toxicity, and all 8 patients carried the GGH-354GG genotype. The 354 T allele has been shown to correlate with increased GGH gene expression. Thus the authors postulated that the -354GG genotype possibly results in reduced GGH levels and higher cellular concentrations of MTXPG, which may lead to increased toxicity and explain the observed association [51]. AICAR transformylase (ATIC) converts AICAR to 10-formyl AICAR and is directly inhibited by MTX (Fig. 1). This leads to accumulation of AICAR and adenosine, a purine with antiinflammatory properties. Adenosine may be an important mediator of the anti-inflammatory effects of MTX [8]. The ATIC gene is a 37 kb gene located on chromosome 2 (2q35). The ATIC 347C>G SNP leads to a threonine to serine substitution in codon 2, which may cause a decrease in ATIC’s enzymatic activity and affect AICAR accumulation and adenosine release. A study examined the combined effects of the ATIC 347C>G SNP, TSER*2, and RFC 80G>A polymorphism on MTX efficacy. 108 RA patients on MTX at a dose of 14 mg/week (range 5–25 mg/week) were examined. Red blood cell long-chain MTX polyglutamate (MTXPG) concentrations were measured, and a pharmacogenetic index was calculated from the sum of homozygous variant

538

Atinuke Aluko and Prabha Ranganathan

genotypes (RFC1 80AA, ATIC 347GG, TSER*2/*2). The allelic frequency for the ATIC 347G variant was 37%. Patients were categorized as MTX responders or MTX non-responders using a visual analog scale (VAS). Eighteen patients who were carriers of the ATIC 347GG genotype had fewer swollen joints (1.9  0.6 versus 4.5  0.6 [p ¼ 0.06]) and a lower score for physician’s assessment of patient’s response to MTX (1.8  0.3 versus 2.8  0.2 [p ¼ 0.02]) compared to 90 patients who were carriers of the C allele (ATIC 347CC; n ¼ 47) or ATIC 347CG (n ¼ 43) genotype [52]. Among other studies, only one study demonstrated the ATIC 347C allele to be associated with MTX efficacy in a Dutch cohort [29]. Four studies showed an association of the ATIC 347G allele with MTX toxicity in US [27, 33], Dutch [29], and Slovenian cohorts [53]. The variable response of patients with RA to MTX led Wessels et al. to propose the development of a clinical pharmacogenetic model (CP-MTX) that predicts the response to MTX using four clinical variables (disease activity, sex, presence of rheumatoid factor, and smoking status) and four SNPs in genes involved in the folate and adenosine pathways (MTHFR 1958G>A, AMPD1 34C>T, ITPA 94C>A, and ATIC 347C>G) [54]. (The gene ITPA and its function are described below in the section on azathioprine). Lopez-Rodriguez applied data from components of CP-MTX to 720 patients with RA to evaluate its utility in a larger cohort of patients on MTX monotherapy (including some patients on steroids) and its accuracy in predicting MTX non-responders. Half of the 66.7% non-responders were accurately predicted supporting the credibility of this model [55]. 2.2 Pharmacogenetics of Azathioprine

Azathioprine (AZA) is used in the treatment of several rheumatic diseases, including systemic lupus erythematous (SLE) and RA. About 10–30% of patients with RA discontinue AZA due to side effects [56]. AZA is a prodrug that after oral intake is converted into 6-mercaptopurine (6-MP), an active purine antimetabolite, which affects the purine de novo synthetic and salvage pathways (Fig. 2). 6-MP is converted by hypoxanthine phosphoribosyl transferase (HPRT) to 6-thio-inosine monophosphate (6-TIMP) which is in turn converted to 6-thioxanthosine monophosphate (6-TXMP) by inosine monophosphate dehydrogenase (IMPDH) before eventual conversion to cytotoxic thioguanine nucleotides (6-TGN). Inosine monophosphate is phosphorylated to inosine triphosphate (ITP), a toxic metabolite, and this process can be reversed by inosine triphosphate pyrophosphatase (ITPase, encoded by ITPA). ITPase deficiency results in the accumulation of toxic ITP. ITPase-deficient individuals treated with AZA can develop toxicity because of accumulation of thio-ITP [57]. By a parallel pathway, 6-MP can be inactivated by thiopurine methyltransferase (TPMT) to 6-methylmercaptopurine (6-MMP) or by

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

6 – TIMP

539

6 - TGN

HPRT

AZA

6 – MP

XO

Thiouric Acid

TPMT 6 - MMP

Fig. 2 Scheme of thiopurine drug metabolism. TPMT thiopurine methyltransferase, XO xanthine oxidase, HPRT hypoxanthine phosphoribosyl transferase, 6-TIMP 6-thiosine monophosphate, 6-MMP 6-methylmercaptopurine, 6-TGN 6-thioguanine nucleotides

xanthine oxidase (XO) to thiouric acid (TU). Thus, a relative deficiency of TPMT leads to accumulation of cytotoxic TGN and significantly increased AZA toxicity. Studies have shown no link between IMPDH activity, both in erythrocytes and monocytes, and 6-TGN concentrations calling into question the role of this enzyme in 6-TGN production [58, 59]. However, lower red blood cell IMPDH activity could result in higher levels of cytotoxic 6-MMP and higher risk of toxicity, especially hepatotoxicity [60]. Common, significant toxicities of AZA are hematologic and gastrointestinal. The TPMT gene is a 26 kb gene located on chromosome 6 (6p22.3). Allelic variants of this gene determine the level of TPMT activity in erythrocytes. TPMT activity in erythrocytes can be classified into high, intermediate, and low or no activity. Population studies have shown that approximately 90% of the population has high activity, 10% has intermediate activity, and 0.3% has little to no activity [61]. Standard doses of AZA when given to patients with low TPMT activity can lead to severe hematologic toxicity which may be fatal. 80–95% of patients with low TPMT activity usually possess one of the three common allelic variants of the TPMT gene, TPMT*2, TPMT*3A, or TPMT*3C [62–64]. The frequencies of these allelic variants vary in different populations worldwide; thus ethnicity influences the occurrence of these variants [65, 66]. Patients with low TPMT activity require lower AZA doses to avoid toxicities [67]. 68 patients with rheumatic disease on AZA (2–3 mg/kg/day) were genotyped for TPMT*2 and TPMT*3A alleles. All patients were assessed for side effects from AZA such as leucopenia, liver function abnormalities, and gastrointestinal intolerance. Six (9%) patients were heterozygous for TPMT*3A, five of whom discontinued AZA within 4 weeks of starting the medication due to hematologic toxicity [68]. In another study 40 patients with RA on AZA (0.7–1.4 mg/kg/day) were genotyped for the TPMT

540

Atinuke Aluko and Prabha Ranganathan

alleles. 6 out of the 40 patients discontinued AZA due to toxicity. Three of the six patients with severe gastrointestinal toxicity were heterozygous for the TPMT*3A allele, while the remainder possessed the wild-type TPMT allele. The association between the TPMT allele and AZA toxicity was significant ( p ¼ 0.018). Based on the results of this study, the positive predictive value for toxicity in a TPMT polymorphism carrier was 60% [69]. Boonsrirat et al. reported the case of a patient with SLE who presented with pancytopenia, sepsis, typhlitis, and disseminated intravascular coagulopathy after a short period of AZA therapy. The patient had low TPMT activity due to the TPMT*3C genotype [70]. A recent meta-analysis of 67 studies examined whether patients with intermediate TPMT activity were at increased risk of myelosuppression when taking thiopurine medications. This metaanalysis included all primary studies of patients on a thiopurine medication that included either genotypic or phenotypic testing for TPMT activity and reported cases of hematological adverse reactions. The search was not limited to a specific disease or condition. Patients with two TPMT variant alleles who were TPMT deficient had a substantial increase in their risk of myelotoxicity (86% of deficient patients developed myelosuppression). Heterozygous patients (i.e., with one variant allele) with any of the TPMT variant alleles that led to intermediate TPMT activity were also at high risk for drug-induced myelosuppression compared to those with wild-type alleles (OR 4.19, 95% CI 3.20–5.48) [71]. In a prospective study of patients with Crohn’s disease on AZA, dropouts during the first 2 weeks of AZA therapy due to adverse events (AEs) were significantly more frequent in carriers of the ITPA 94C>A allele [72]. In another study of patients with inflammatory bowel disease, 40 out of 160 patients on AZA were found to have decreased ITPA activity [below the lower limit of the reference range, A missense mutation in ITPA, TPMT*2, and TPMT*3 in 73 patients with inflammatory bowel disease who had side effects from AZA and 74 patients who had tolerated AZA without adverse events. There was no significant difference in the frequency of the ITPA allele between patients who had experienced an adverse event and those who had not (16/146 versus 16/148, p ¼ 0.56). There was no association between the ITPA 94C>A polymorphism and development of adverse events such as rash, pancreatitis, or flu-like symptoms [74]. The conversion of AZA to 6-MP was previously thought to be a non-enzymatic reaction. Subsequent studies showed that glutathione S transferases (GSTs) may be involved in this conversion [75, 76]. A study performed in patients with inflammatory bowel disease with the GST-M1 null genotype demonstrated that such patients required a lower dose of AZA to generate 6-TGN compared to those with the GST-M1 wild-type genotype who required higher doses of AZA. Other GST genotypes did not show a significant effect on AZA metabolism [77]. Other studies have examined the association between the activity of TPMT, other enzymes in the purine pathway, and AZA toxicity. One study evaluated the activity of TPMT, HPRT, 50 -nucleotidase, and purine nucleoside phosphorylase in red blood cells (RBC) of 33 RA patients on AZA (dose of approximately 2 mg/kg/day), and 66 controls were measured. 14 patients with low ( p ¼ 0.004) and 7 patients with intermediate TPMT activity (RR 3.1) developed AZA toxicity compared to patients with normal TPMT activity [56]. Another study measured TPMT activity in 3 patients with RA who had experienced AZA-induced hematologic toxicity and 16 patients with RA without AZA toxicity. Two patients with AZA-induced hematologic toxicity were TPMT deficient, one partial and the other complete [78]. Patients were not genotyped in either of these studies. Thus, both TPMT genotyping and measurement of TPMT activity in RBC have been studied as tools to predict and prevent AZA toxicity. Clearly, large, prospective studies are needed to validate the observations from the smaller studies described above (Table 2). There is some evidence for ITPA variants influencing AZA toxicity in inflammatory bowel disease, but there are no studies to date on the effect of this polymorphism on AZA toxicity when used in RA. Of note, TPMT genotyping is available to clinicians to screen patients prior to initiation of AZA, for genotype-directed dose adjustment, and is the first commercially available assay for pharmacogenetic testing in rheumatology. Through genome-wide association studies (GWAS), variants in a new enzyme involved in thiopurine metabolism have been identified. Nucleoside diphosphate linked moiety X (nudix)-type motif 15 (NUDT15) catalyzes the conversion of cytotoxic thioguanine

542

Atinuke Aluko and Prabha Ranganathan

Table 2 Pharmacogenetics of AZA in RA Amino acid change in Polymorphism enzyme

Population Effect of Biochemical prevalence of polymorphism on effect of polymorphism enzyme activity polymorphism

TPMT*2 238G>C

Alanine to proline

0.2–0.5%

TPMT*3A 460G>A, 71A>9G

TPMT*3C 719A>G

Clinical effects

“Low to Decreased intermediate” methylation due to enhanced of AZA to degradation of inactive enzyme compounds

Hematologic and GI toxicity

3.2–5.7% Alanine to threonine and tyrosine to cystine respectively

As above

As above

Hematologic toxicity

Tyrosine to cystine

As above

As above

Hematologic toxicity

0.2–0.8%

triphosphate metabolites to less toxic thioguanine monophosphate. Defects in this process allows for accumulation of cytotoxic TGNs and incorporation into DNA (DNA-TG) resulting in DNA damage and apoptosis. A clinically significant variant of NUDT15 rs116855232 (415C>T) leads to near-complete loss of enzymatic activity and protein stability in vitro. Patients expressing this allele have excessive DNA-TG and experience severe myelosuppression. Although additional variant alleles have been identified, the 415C>T variant has been most extensively studied and has the strongest evidence for clinical use and is linked to most cases of AZA-related myelosuppression in Asians and Hispanics. A recent study of 86 Chinese patients revealed a significant association between TPMT*3C and NUDT15*3 genotypes and the development of AZA-induced leukopenia. Additionally, NUDT15*3 was significantly associated with gastrointestinal effects, erythropenia, and thrombocytopenia, while TPMT*3C was associated with the development of alopecia [79]. NUDT15 genotyping is not currently widely performed prior to AZA initiation but might be useful especially in at-risk populations [80]. 2.3 Pharmacogenetics of Sulfasalazine

Sulfasalazine (SSZ) is another DMARD often used in the treatment of RA. It is estimated that 20–30% of patients with RA on SSZ report adverse drug reactions. Adverse drug events of SSZ are gastrointestinal and hematologic. SSZ is a combination of sulfapyridine (SP) and 5-aminosalicylic acid (5-ASA). After ingestion, colonic bacteria split SSZ into these two compounds. 5-ASA remains in the large bowel, while most of sulfapyridine is completely absorbed and undergoes acetylation,

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

543

hydroxylation, and glucoronidation in the liver. Acetylation of sulfapyridine is carried out by the enzyme N-acetyltransferase 2 (NAT2) which acetylates sulfapyridine into N-acetyl sulfapyridine. The NAT2 gene is a 9 kb gene located on chromosome 8 (8p22) and can be polymorphic. NAT2 gene polymorphisms may alter the acetylator phenotype of an individual (slow, intermediate, or fast acetylator) and thus have effects on an individual’s susceptibility to SSZ toxicity. Slow acetylators have been shown to be more prone to SSZ toxicity such as abdominal discomfort, nausea, rash, and headaches compared with fast acetylators [81, 82]. The fecal route is the main route of elimination of absorbed SSZ, mediated by ATP-binding cassette protein G2 (ABCG2), a transmembrane protein that transports SSZ into the intestinal lumen. Following SSZ administration, ABCG2 knockout mice had 111-fold plasma concentration of SSZ compared with wild-type mice [83]. Similarly, human studies revealed that healthy subjects with ABCG2 421C>A polymorphism (rs2231142) leading to a loss of function A allele is associated with reduced protein expression and activity leading to elevated plasma levels of SSZ [84]. When studied in patients with early RA in whom SSZ was part of triple DMARD combination therapy (SSZ at a dose of 2 g daily, increased to 3 g daily to achieve low disease activity, along with MTX 10 mg weekly and hydroxychloroquine 400 mg daily), sensitivity analysis revealed that patients with loss of function alleles (ABCG2 C421A AC or AA genotype) were more likely to achieve remission after 12 months [OR 3.34% (95% CI 1.17–9.64), p ¼ 0.025]. ABCG2 genotype however was not associated with SSZ discontinuation suggesting that the higher plasma concentration of SSZ is associated with a higher efficacy but not a higher risk of adverse effects. This is in contrast to NAT2 polymorphisms which result in higher sulfapyridine concentrations which may be responsible for SSZ toxicity. The effect of ABCG2 function on SSZ is less marked when SSZ is given concomitantly with MTX and hydroxychloroquine because the latter are substrates for multiple ABC transport proteins [85]. A few studies have evaluated the effects of NAT2 polymorphisms on SSZ toxicity in patients with RA. One retrospective study assessed 144 patients with RA on SSZ at a dose ranging between 500 and 1500 mg per day. NAT2 genotyping was carried out in all patients. Slow acetylators lacking the wild-type NAT2*4 allele experienced adverse reactions more frequently (63%) compared with fast acetylators who had at least one NAT2*4 allele (8%). This association between the NAT2 genotype and SSZ toxicity was clinically significant (OR 7.73, CI 3.54–16.86, p < 0.001). In fact, 25% of the slow acetylators had to be hospitalized for their toxicities [86]. In a second study, 114 patients with inflammatory bowel or joint disease treated with SSZ (mean dose of 2 gm per day) were studied. Patients were genotyped for five allelic variants,

544

Atinuke Aluko and Prabha Ranganathan

NAT2*5A, NAT2*5B, NAT2*5C, NAT2*6, and NAT2*7 (encoding slow acetylator status) and the wild-type NAT2*4 allele (encoding rapid acetylator status). 27 of 39 patients (69%) who developed agranulocytosis within 3 months of starting treatment with SSZ were slow acetylators compared with 34 of 75 patients (45%) who received SSZ without developing an adverse hematologic event (OR 2.7; p ¼ 0.002). Patients with SSZ-induced agranulocytosis had higher frequencies of the NAT2*6 alleles among other allelic variants (36%) compared with those without agranulocytosis (23%) ( p ¼ 0.033) [87]. In another study, the authors performed two bioavailability studies under comparable conditions with 24 healthy subjects of both genders equally distributed. Plasma levels of sulfapyridine and acetyl sulfapyridine were determined after oral intake of entericcoated formulations of sulfasalazine (500 mg and 1000 mg, respectively). NAT2 genotype was analyzed in parallel for all subjects. Compared to those with the rapid acetylator genotype, the apparent terminal elimination half-life of sulfapyridine as well as of acetylsulfapyridine was prolonged in those with the slow acetylator genotype. The strongest functional effect on enzymatic activity was noticed in slow acetylators carrying the 341T > C polymorphism in NAT2 [88]. A meta-analysis of 9 studies which included 1077 patients showed a similar association between the NAT2 genotype and risk of developing adverse effects to SSZ, especially in Asians. This supports a role for performing NAT2 genotyping prior to SSZ initiation to predict toxicity [89]. Based on the limited data published thus far, it appears that the acetylator status of a patient, as determined by the NAT2 genotype, is an important determinant for the development of SSZ toxicity. (Table 3) While more studies and data are clearly needed, this suggests that prospective screening of patients for the NAT2

Table 3 Pharmacogenetics of SSZ in RA

Polymorphism Effect of polymorphism

Biochemical changes associated with polymorphism

Clinical effects

NAT2*5A

Decreased activity of NAT2 enzyme Increased concentrations of SSZ intermediates due to leading to slow acetylation (slow slow acetylation acetylator)

Agranulocytosis Fever, rash

NAT2*5B

As above

As above

As above

NAT2*5C

As above

As above

As above

NAT2*6

As above

As above

As above

NAT2*7

As above

As above

As above

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

545

genotype prior to initiation of SSZ may be a useful tool to prevent SSZ toxicity. 2.4 Pharmacogenetics of Biologics 2.4.1 Tumor Necrosis Factor Antagonists

TNF Gene Polymorphisms

The tumor necrosis factor alpha (TNF-α) antagonists, a class of biologic DMARDs (bDMARDs), have revitalized the treatment of RA in recent years. These agents not only ameliorate the signs and symptoms of RA, but more importantly are highly effective in slowing the radiographic progression of disease [2, 90]. In spite of their well-established efficacy, about 20–40% of RA patients do not respond adequately to these agents [91, 92]. Five TNF antagonists are currently approved for the treatment of RA, etanercept (ETN), infliximab (INF), adalimumab (ADA), golimumab (GOL), and certolizumab pegol (CZP). ETN, a fusion protein of two identical chains of the recombinant human TNF receptor, p75, fused with the Fc portion of human IgG1, binds to soluble TNFα in vivo. INF, ADA, and GOL are all monoclonal antibodies to TNF-α; INF is chimeric while ADA and GOL are fully humanized. CZP is a human anti-TNF Fab fragment that is conjugated with polyethylene glycol. Over the past few years, several studies have attempted to define pharmacogenetic markers to predict response to anti-TNFα therapy (Table 4). Some studies have used a candidate gene approach and have looked at genes in the TNF pathway, including genes for TNF-α, TNF receptors, and polymorphisms affecting signaling pathways downstream of the TNF receptors. Fc receptor polymorphisms and risk variants for RA have also been studied. Other studies have identified predictive variants from GWAS in large cohorts of patients treated with anti-TNF therapy. The TNF family, consisting of TNF-α and lymphotoxins A (LTA) and B (LTB), has vital functions in immune regulation. The TNF-α gene is located on chromosome 6 and lies within the human MHC III region (Fig. 3). The TNF locus is a 7 kb region where the TNF, LTA, and LTB genes are arranged in tandem and lie near the HLA B and MHC III DR regions. The most studied of the TNF polymorphisms is the TNF-308 A>G SNP which is in the promoter region. The TNF-308A allele is associated with increased transcription and synthesis of TNF-α as compared to the TNF-308G allele. In one study of 59 RA patients treated with INF, patients without the A allele had improved DAS, a standardized measure of disease activity in RA, with use of INF (81%) compared to patients with the A allele (42%) ( p ¼ 0.0009) [93]. Cuchacovich et al. proposed an interesting explanation for the findings of the above study based on the results of their own study. In the study by Cuchacovich et al., 132 patients with RA were genotyped for the TNF -308 promoter polymorphism. From these 132 patients, 10 patients with the TNF 308 G/A and 10 with the TNF 308 G/G polymorphism were selected and

546

Atinuke Aluko and Prabha Ranganathan

Table 4 Pharmacogenetics of TNF antagonists in RA

Genes/polymorphisms

Postulated effect of gene/ polymorphism

Clinical effects

TNF promoter 308A>G

May increase transcription of TNF-α gene May increase circulating TNF-α levels

Increased response to INF No effect on response to INF

TNF promoter 238G>A

May increase transcription of TNF-α gene

No effect on response to ETN

TNF +489 G/G

Intronic polymorphism – function unknown

No effect on response to ETN

TNFRSF1A 609, 580, 383

May affect ligand binding

No effect on response to ETN

TNFRSF1B 196 T>G

May affect receptor shedding and ligand Increased response to INF, ETN No effect on response to ETN binding; may increase synthesis of IL-6

TNFRSF1B 676T>G

May affect membrane receptor shedding Increased response to anti TNF

TNF microsatellites a, b, May influence production of TNF by c, d, and e PBMC, linked to TNF -308 SNP, increased susceptibility to RA

Specific TNFa/b haplotype associated with response to INF No effect on response to ETN

HLA DR, DQ alleles

May affect response to TNF blockade and increase susceptibility to and severity of RA because of close proximity to TNF locus

No effect on response to INF Specific individual HLA DRB1 alleles and haplotypes markers of increased response to ETN

MHC class I chainrelated gene A transmembrane polymorphism

As above

No effect on response to INF

HLA microsatellites BAT2, D6S273, D6S2223

Haplotype may carry “response gene”

BAT2-D6S273 haplotype associated with increased response to INF

FcγRIIIA 158FF

Low affinity for IgG, affects antibody clearance

Increased response to anti-TNF therapy

FCGγIIA-131RR

Low affinity for IgG, affects antibody clearance

Increased response to IFN

PTPRC rs10919563

RA susceptibility marker

Associated with good response to anti-TNF therapy

Polymorphisms in MAP Affect signaling pathways in RA kinase pathways MAP3K1 rs96844, MAP3K14 rs4792847, MAP2K6 rs11656130

Associated with good response to anti-TNF therapy

(continued)

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

547

Table 4 (continued)

Genes/polymorphisms

Postulated effect of gene/ polymorphism

Clinical effects

AFF3 SNPs rs10865035 RA susceptibility markers and rs1160542

Associated with good response to anti-TNF therapy

CD226 SNP rs763361

T and NK cell cytotoxicity

Associated with good response to anti-TNF therapy

IL10 promoter microsatellite polymorphism R3

Associated with IL10 secretion

Good response to ETN

–1087G>A in IL10 in combination with 308 in TNF

Associated with IL10 secretion

Good response to ETN

IL6 -174GG

Associated with IL6 levels

Good response to ETN

Regions within the human MHC Class Ch6

HLA

II DP

DQ

III DR

C2 C4

HSP

I TNF

B

C

A

G

Ch6

TNFc TNFa TNFb LTB

TNF

LTA

+489 -238 -308

Fig. 3 TNF locus with some of the polymorphic sites known within the TNF locus. C2, C4 complement C2, C4, Ch chromosome, HLA human leukocyte antigen, HSP heat shock protein, LTA lymphotoxin alpha, LTB lymphotoxin beta, MHC major histocompatibility complex, TNF tumor necrosis factor. (Reproduced from Ranganathan [162] by permission of Future Medicine Ltd)

received INF. Although both groups showed a similar response and demonstrated an increase in TNF-α levels after INF treatment, the increase in TNF-α levels correlated with the ACR50 response only in patients with the G/A polymorphism ( p < 0.03). The authors postulated that the TNF -308 polymorphism influences response to INF through its effects on circulating TNF-α levels [94]. Over the past decade, more than a dozen studies have looked at the effects of this polymorphism on anti-TNF therapy in RA. A few meta-analyses analyzed these studies but yielded mixed results. A

548

Atinuke Aluko and Prabha Ranganathan

meta-analyses by Pavy et al. did not find a significant association between the TNF -308 polymorphism and response to anti-TNF therapy in RA [95]. Zeng et al. in their meta-analysis of 15 studies with a total of 2127 patients with RA did however report an improved response to anti-TNF therapy in patients with the 308G allele [96]. The 238G>A SNP is also in the promoter region and has possible effects on TNF-α production. Maxwell et al. found a poor response to IFN but not ETN with the GA genotype at TNF-238 rs361525 ( p ¼ 0.028, n ¼ 40) [97]. Other TNF gene polymorphisms including the intronic SNP +489 and the promoter region SNP 857C >T are associated with severe RA, but no clear association has emerged between these SNPs and response to anti-TNF therapy [98]. TNF Receptor Polymorphisms

Polymorphisms in the TNF-α receptors also appear to be important in determining response to anti-TNF therapy. Soluble TNF-α binds to two transmembrane receptors: p55, also known as CD 120a or TNF receptor type 1 (TNFRSF1A), and p75, also known as CD 120b or TNF receptor type 2 (TNFRSF1B). Local production of soluble TNFRs and their upregulation is important in the modulation of TNF-α activity in RA joints. The TNFRSF1B gene is located on chromosome 1 and has 10 exons and 9 introns. A SNP has been described in exon 6 of the TNFRSF1B gene, a single base substitution at codon 196 (T to G) that leads to a nonconservative amino acid substitution, methionine to arginine [99]. The TNFRSF1B 196T>G polymorphism was studied in 175 RA patients for its effects on response to anti-TNF therapy. Of the 175 patients, 66 were treated with either ETN or INF and their response to treatment assessed using the DAS. Of the 66 patients on TNF antagonist therapy, 38 had the TT, 22 had the TG, and 6 had the GG genotypes. Patients with severe RA carried the GG genotype more often (6.4%) than those in the mild to moderate group (3.1%). Patients carrying the TT genotype had a better response to therapy over 24 weeks compared to the patients with the TG or GG genotype as measured by the DAS, with the greatest difference seen at 12 weeks (OR 5.1, CI 1.3–19.96, p ¼ 0.03) [98]. Another polymorphism in the TNFRSF1B gene 676T>G in exon 6 that results in an amino acid change from methionine to arginine has been described. The 676TT genotype is associated with a better response to anti-TNF therapy as compared to the 676TG genotype at 3 months (OR 3.78, 95% CI 1.07–13.31) and 12 months (OR 4.30, 95% CI 1.16–15.99) in RA [100]. Some studies could only verify this association in patients who were anti-CCP positive [100, 101], while a study of 457 patients with RA by Criswell et al. could not confirm this association at all [102].

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis MHC Gene Polymorphisms, TNF, and MHC Microsatellites

549

The close proximity of the “TNF locus” to the HLA B and HLA DR genes (MHC genes) on chromosome 6 and the fact that there is a strong link between specific HLA DRB1 alleles (also called the shared epitope alleles) and susceptibility to RA and its severity [103] make it likely that MHC gene polymorphisms may influence response to anti-TNF agents. Some microsatellite haplotypes have been previously associated with susceptibility to RA or linked to TNF promoter region SNPs [104, 105]. In one study, 78 RA patients treated with INF were genotyped for HLA-DRB1, HLA-DQA1, HLA-DQB1, MHC class I chain-related gene A (MICA) transmembrane polymorphism alleles, microsatellites TNFa-e, D6S273, HLA-B-associated transcript 2 (BATS2), and D6S2223. None of the alleles influenced response to INF including the TNFa/b microsatellites (linked to the TNF -308 promoter polymorphism), implying that this TNF promoter variant may not be important in determining response to INF. However, there were some interesting associations observed between certain microsatellite haplotypes and response. Among the microsatellite haplotypes, the D6S273_4/BAT2_2 pair was a marker of the INF responder group, both among patients and when compared with controls (46% versus 11% in non-responders; p ¼ 0.001; 46% in responders versus 17% in controls; p ¼ 0.00002) indicating that this microsatellite pair may occur on the haplotype that carries the “response gene” or each microsatellite allele could serve as a marker of a “response gene” in proximity. The frequency of one of the TNFa/b haplotypes was increased in responders compared to non-responders (41% versus 16% in non-responders; p ¼ 0.01). Thus, some microsatellite haplotypes were associated with response to INF in this study; single alleles did not reveal similar associations [106]. In a second study, patients were genotyped for specific HLA-DRB1 alleles, i.e., the shared epitope (SE) alleles and categorized as having 0, 1, or 2 copies of the SE. 457 patients with early active RA (duration of 3 years) treated with MTX or ETN were genotyped and response to therapy measured by ACR50 response rates after a year of treatment. SNPs at positions 308, 238, and +488 of the TNF gene and +249, +365, and +720 of the LTA gene were also examined. (These 6 LTA-TNF SNPs mark haplotypes spanning the “TNF locus” region.) Five TNF microsatellites (TNF a through e); SNPs in TNFRSF1A at positions 609, 580, and 383; and the 196T/G polymorphism in TNFRSF1B were also examined. As the Fc receptor (FcR) pathway appears important in the degradation of ETN-TNF complexes, three FcR polymorphisms were also examined. The number of HLA-DRB1 (SE) alleles correlated with response to treatment. Patients with two copies of the SE alleles had a better response to ETN compared to those with zero or one copy of the allele (OR 4.3, 95% CI 1.8–10.3). Haplotypes defined by the 6 LTA-TNF SNPs and DRB1 alleles were

550

Atinuke Aluko and Prabha Ranganathan

deduced for the 16 most common DRB1 alleles in a subset of 224 Caucasian patients. Among 448 haplotypes thus examined, two haplotypes HLA-DRB1 *0101-GGGAGG and HLA-DRB1 *0404-GGAAGG strongly correlated with response (76 and 61% ACR50 response at 12 months, respectively). Thus the number of copies of HLA-DRB1 SE alleles inherited and specific haplotypes spanning the HLA-DRB1 region and SNPs in the LTA-TNF region may be associated with response to ETN, at least in the Caucasian population [102]. RA Risk Variants

There are more than 30 risk alleles associated with susceptibility to RA [107]. Among the RA risk variants studied so far, protein tyrosine phosphatase receptor type C (PTPRC) variants have the strongest association with response to anti-TNF therapy. The PTPRC gene encodes a CD45 antigen and is a known RA susceptibility marker. In a large cohort of 1283 patients with RA, those who had the rs10919563 SNP in the PTPRC gene had a better EULAR response to TNF inhibitors as compared to patients without the polymorphism ( p ¼ 0.0001) [108]. Plant et al. reported similar findings in a large cohort of 1115 RA patients from the United Kingdom who were tested for 29 SNPs known to be RA susceptibility variants. The rs10919563 SNP in the PTPRC gene was associated with improved response to anti-TNF therapy (OR 0.62, 95% CI 0.40–0.95; p ¼ 0.03) [109]. Despite the strength of the associations found in these studies, a subsequent study by Krintel et al. was unable to replicate these findings [110]. Mitogen-activated protein kinases (MAPKs) are crucial to several signaling pathways in RA. Bowes et al. reported two SNPs rs96844 in MAP3K1 and rs4792847 in MAP3K14 to be nominally associated with response to anti-TNF therapy in a UK cohort of 642 RA patients ( p < 0.05). However, these associations could not be confirmed in a validation cohort of 428 patients with RA in the same study [111]. In another large cohort of 1102 RA patients, polymorphisms in 11 genes in the MAPK pathway were investigated. Seven polymorphisms in five genes in the MAPK signaling pathways were associated with an improved DAS28 response to IFN and ADA therapy, but not to ETN. One polymorphism rs11656130 in MAP2K6 was associated with a good EULAR response [112]. Tan et al. genotyped 1012 RA patients and identified two more SNPs in susceptibility genes associated with response to anti-TNF therapy. The AFF3 gene which encodes nuclear transcription activators in lymphoid tissue has been identified as a RA susceptibility marker [113]. The G allele of two SNPs, rs10865035 and rs1160542, in AFF3 gene was associated with response to antiTNF therapy [113]. CD226 is a membrane protein on the surface of hematopoietic cells which is involved in T and NK cell cytotoxicity. The SNP rs763361 in the CD226 gene results in a glycine to

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

551

serine substitution and is a RA risk variant [114]. This SNP was also associated with response to anti-TNF therapy [115]. Other genes involved with susceptibility to RA that play a role in response to anti-TNF-α therapy include signal transducer and activator of transcription 4 (STAT4), protein tyrosine phosphatase non-receptor 2 (PTPN2), TRAF3 interacting protein 2 (TRAF3IP2), and Psoriasis susceptibility1 candidate 1 (PSORS1C1). A study analyzing the effects of polymorphisms in these four genes with response to anti-TNF-α therapy revealed associations that appeared to be drug specific. In ETN-treated patients, there was no EULAR response at 2 years in patients carrying the STAT4 rs7574865 allele, and patients with the PSORS1C1 rs2233945 allele did not achieve low disease activity at 6 months. These effects were not seen in ADA-treated patients who on the contrary had no EULAR response at 2 years in those with polymorphisms involving TRAF3IP2 rs33980500 and after 6 months in those with PTPN2 rs7234029 polymorphisms [116]. Cytokines, Toll-Like Receptors, and Signaling Pathways

Cytokines play an important role in the pathogenesis of RA. Polymorphisms in interleukin (IL)1β, IL1 receptor antagonist (ILRA), IL6, IL10, and tumor growth factor (TGF) β have been studied in relation to anti-TNF therapy. While no polymorphisms in individual cytokine genes have been significantly associated with response [117], one study found that a SNP –1087G>A in the IL10 gene in combination with the 308 SNP in the TNF gene was associated with a good response to ETN. The same study reported that the combination of the A2 allele in intron 2 of the ILRA gene and a rare +915C allele in codon 25 of the TGFB1 gene correlated with a poor response to ETN therapy. The +915C allele which is a rare allele was found significantly more often in combination with the A2 allele in the ILRA gene in non-responders to ETN ( p < 0.05) [118]. The IL10 promoter microsatellite polymorphism R3 allele and the R3-G9 haplotype were associated with a good response, whereas the G13 allele and the haplotype R2-G13 were associated with moderate to no response to ETN [119]. In a small study, 77 patients with RA were genotyped for the 174 G>C polymorphism in the IL6 gene. The 174 G>C polymorphism influences IL6 levels. After 12 months of ETN therapy, more patients with the IL6 -174GG (95.7%) genotype had an improvement in disease activity by DAS as compared to those with the GC (75.6%) or CC (44.4%) genotype ( p ¼ 0.006) [120]. Genetic variants in Toll-like receptor (TLR) and nuclear factor (NF)-kB signaling pathways have shown an association with response to anti-TNF-α agents in the treatment of RA [121, 122]. A study by Potter et al. looked at genotypes in the TLR and NFkB pathways to predict response to anti-TNF therapy in RA. A total of 187 SNPs in these pathways were studied in

552

Atinuke Aluko and Prabha Ranganathan

909 patients with RA. Twelve SNPs in nine genes showed a nominal association with treatment response in patients on anti-TNF therapy [123]. A validation study revealed a similar association between the variant rs11541076 in IRAK 3 (encodes IL1 receptor associated kinase 3, IRAK-3, which regulates TLR signaling such that there is inhibition of downstream signaling and activation of transcription factors involved in inflammation) and a positive treatment response but no association with the other previously reported SNPs in this pathway [124]. The function of IRAK3 rs11541076 is however unknown; it is possible that the polymorphism is a surrogate for other clinically significant genetic markers. Larger sample sizes are required, and consistent replication of results is necessary to show the significance of these variants. IL 33, a recently discovered member of the IL 1 family, functions as an endogenous danger signal and as a nuclear factor involved in regulation of gene transcription. It is predominantly involved in the Th2-mediated immune response by inducing the production of other cytokines such as IL4, IL5, and IL10 and activates cells such as Th2 T cells, mast cells, and dendritic cells. It has also been implicated in Th1 responses through its involvement in the activation of Th1 cells, CD8+ T cells, NK cells, neutrophils, macrophages, and B cells as well as enhancement of the Th17 response. The frequency of three genetic variants within the IL 33 gene, IL33 rs10975519, rs16924159, and rs7044343, was similar between patients with RA and healthy controls. However, the IL33 rs 16924159 AA genotype was more frequently observed among patients with RA with a decreased response to anti-TNF-α agents, as evidenced by higher DAS28 scores, compared with other genotypes (OR ¼ 1.97, 95% CI ¼ 1.05–3.72, p ¼ 0.046). No differences were observed between the presence of the other two IL33 polymorphisms and clinical response to anti-TNF-α therapy [125]. Fcγ Receptor Variants

Most anti-TNF agents are antibodies and therefore have a Fc component to them (except CZB which is a pegylated Fab fragment). The Fc component of these antibodies attaches to the Fc gamma receptor (FcγR) on various cells. Polymorphisms in the FcγR gene affect the avidity and strength of Fc binding. Since binding of the Fc fragment to FcγR is a mechanism for antibody clearance, polymorphisms in the FcγR gene may potentially influence the efficacy of the anti-TNF drugs. Two polymorphisms, FcγRIIIA F158V and FcγRIIA R131H, have been studied in this respect. The FcγRIIIA 158FF genotype has a lower affinity for IgG1. In a small study of 30 patients with RA, the 158 FF variant was found in greater frequency among patients who had a very good response to anti-TNF agents [126]. In a study by Canete et al. of 98 individuals with RA, those who were homozygous for the low-affinity FCγRIIIA 158FF genotype had a better ACR50

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

553

(American College of Rheumatology 50, another standardized measure of disease activity in RA) and EULAR responses at week 6 of IFN therapy. Other individuals with the low affinity variant FCGγIIA 131RR also had better ACR20 responses at week 30 of INF treatment [127]. Similar results were seen in a study of 33 Japanese patients with RA [128]. However, this finding could not be replicated in a subsequent larger study of 282 Swedish RA patients treated with INF or ETN [129]. GWAS Studies

Several genome-wide association studies (GWAS) have been done to assess markers of response to anti-TNF therapy in RA. The first GWAS study in this regard by Liu et al. looked at a small sample of 89 patients and reported 16 SNPs in nine genes that were associated with response to anti-TNF therapy [130]. This finding was not confirmed by subsequent studies. Plant et al. looked at a larger cohort of 566 patients and found seven loci by multistage GWAS that were associated with response to anti-TNF therapy. The strongest association was for SNP rs17301249, mapping to the eyes absent homolog 4 (EYA4) gene on chromosome 6. EYA4 is a co-transcription factor associated with expression of interferon β and CXCL10. Another SNP rs1532269 mapped to the PDZ domain-containing protein 2 (PDZD2) gene which is associated with insulin secretion. (PDZ is an acronym combining the first letters of three proteins – Post synaptic density protein, Drosophila disc large tumor suppressor, and Zonula occludens-1 protein – which were first discovered to share a domain). Five SNPs mapped within intergenic loci on chromosomes 1, 4, 11, and 12 [130]. Krintel et al. studied 196 Danish patients with moderate to severe RA, treated with IFN, ETN, or ADA, and analyzed 486,450 SNPs for association with response to anti-TNF therapy. The findings of the earlier GWAS by Lui et al. and Plant et al. could not be confirmed in this study, and no SNPs achieved significance despite the wide array studied [110]. Another multistage GWAS of 882 patients with RA from the Dutch Rheumatoid Arthritis Monitoring (DREAM) registry evaluated 2,557,253 SNPs for response to anti-TNF therapy. Although no single SNP reached significance, three SNPs (rs1568885, rs1813443, and rs4411591) showed directional consistency and eight genetic loci were suggestive of association with response in this cohort. However, none of the associations found in earlier studies could be confirmed in this study as well [131]. Cui et al. reported a GWAS meta-analysis, looking at two million common variants in 2706 patients with RA from 13 different cohorts. The SNP (rs6427528) was associated with higher CD84 gene expression in peripheral blood mononuclear cells and a better response by DAS scores in patients treated with ETN. CD84 is a cell surface receptor found on immune cells including T cells, B cells, monocytes, and platelets. It is thought to play a role

554

Atinuke Aluko and Prabha Ranganathan

in signaling T cell activation and IFNγ secretion [132, 133]. The variant, rs6427528, was not associated with response to therapy with either INF or ADA [134]. Acosta-Colman et al. described a SNP, rs3794271, in the fourth intron of the SLCO1C1 gene in their cohort of 315 patients which was associated with a good EULAR response to anti-TNF therapy. SLCO1C1 is a cell membrane transporter important in drug metabolism. The GWAS study by Krintel et al. described above had reported an intergenic SNP rs11045392, located between the 50 end of SLCO1C1 and 30 end of the PDE3A gene (which codes for a phosphodiesterase) as a putative marker of response to anti-TNF therapy. The authors speculated these two SNPs, rs3794271 and rs11045392, may be in linkage disequilibrium, and therefore the PDE3A- SLCO1C1 locus may be an important determinant of response to anti-TNF therapy [135]. A GWAS study performed to determine anti-TNF-α response in 444 Japanese patients with RA starting ETN, INF, or ADA (94% were naı¨ve to anti-TNF-α drugs; 80% were on MTX) revealed a borderline significant association at three non-correlated regions with repeated DAS28 measurements at 3 monthly intervals. The associated SNPs mapped close to four genes: mitogenactivated protein kinase kinase kinase 7 (MAP3K7), basic leucine zipper transcription factor 2 (BACH2), WD repeat containing protein 27 (WDR27), and glial cell line-derived neurotrophic factor family receptor alpha 1 (GFRA1) genes – the last of which has been previously reported as having a significant association with anti-TNF-α response in Caucasians [136]. A replication study evaluating 12 TNF-α GWAS published SNPs in 788 patients with RA did not show any significant association between the 12 SNPs and change in disease activity using the DAS 28 score 6 months after initiating therapy. In addition, the CD84 SNP rs6427528 that was associated with response to ETN in the above study by Cui et al. was not replicated in this current study suggesting that multiple loci may have small effects that may require even larger studies to identify [137]. A biomarker profile that could predict treatment response would be useful in predicting and thus optimizing therapy by avoiding therapies that would be ineffective in certain patients and those that would be associated with certain side effects. Although various polymorphisms have been discovered and studied, results are not very replicable among studies. A systematic review and meta-analysis performed to evaluate the associations between all reported genetic variants and response to anti-TNF-α therapy in RA found a total of 25 SNPs, 19 from GWAS and 6 from the author’s meta-analysis. Most of these polymorphisms mapped to genes involved in T cell function, NFkB and TLR signaling pathways. Exploratory analysis of an available cohort to determine the predictive power of some of the polymorphisms in making clinical treatment decisions however had moderate positive and

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

555

negative predictive values [138]. It is unclear if these polymorphisms are functional themselves or perhaps linked to functional polymorphisms. The EULAR criteria for TNF response employ a 4 composite score (4C-DAS28) which is made of 28 swollen and tender joint count, a visual analogue score (VAS ¼ 0–100 mm) of patient global assessment of arthritis activity, and a marker of systemic inflammation such as CRP. Due to the complexity surrounding some items of the score such as VAS, which could be affected by factors other than active RA such as depression, the use of 4C-DAS28 as an outcome measure could distract from the impact of genetic markers on the response to anti-TNF agents. As such, use of a two composite score (2C-DAS28), based on 28 swollen joint count and CRP, was validated in a previous study [139]. When applied to a cohort of 1812 patients, more SNPs demonstrated therapeutic response at 6 months when the 2C-DAS28 (5SNPs: ALPL rs885814, CERS6 rs13393173, EYA4 rs17301249, and two intergenic SNPs rs12081765 and rs1350948) score was used compared with the 4C-DAS28 score (1SNP intergenic rs12081765). The mean absolute change in 2C-DAS28 and 4C-DAS28 at 6 months was 2.13 and 2.46, respectively [140]. This suggests that using the 2C-DAS28 score as a disease outcome measure in pharmacogenetic RA studies may be a more accurate indicator of joint inflammation and may reveal masked genetic associations. Machine learning is another possible prediction tool that shows some promise in predicting response to anti-TNF -α therapy in RA and may help in guiding treatment choices in the future. Further studies using this tool in a larger and more diverse cohort is required to validate its clinical utility and potentially lead to the development of better prediction models [141]. 2.4.2

Rituximab

Rituximab is an anti-CD20 chimeric antibody with proven efficacy in RA. Rituximab binds to FcγR on mononuclear cells then leads to B cell depletion through different mechanisms, including complement- and antibody-dependent cytotoxicity or phagocytosis. Several candidate gene studies have looked at pharmacogenetic associations affecting response to rituximab therapy. In a small study by Daien et al. [142], 63 patients were analyzed for 13 SNPs in nine genes including IL10, LTA, TGFβ1, TNF-α, TNFRSF1B, TRAF1-C5, STAT4, TNFAIP3, and PTPN22. The following SNPs, PTPN22 rs2476601, STAT4 rs7574865, TRAF1C5 rs1081848, and TNFAIP3 rs6920220 have been associated with RA susceptibility in prior studies [143, 144]. Two SNPs in TGFβ1, rs1800470 in codon 10 and rs1800471 in codon 25, were associated with a good response to rituximab. At codon 10, the CT genotype was more prevalent in responders compared to the TT genotype (OR 1.6, 95% CI 1.2–2.3; p ¼ 0.002), while the CC genotype was equally present in responders and non-responders. At TGFβ1 codon 25, all patients with the

556

Atinuke Aluko and Prabha Ranganathan

GC genotype were responders, while 63% of patients with the GG genotype were responders (OR 1.6, 95% CI 1.3–1.9; p ¼ 0.025). Patients with both SNPs had an even better response to rituximab (OR 2.6, p ¼ 0.008) [142]. SNPs in codon 10 and 25 have been associated with lower TGFβ1 production which may explain the better response to rituximab therapy [145]. B lymphocyte stimulating factor (BLyS) is a B lymphocyte survival factor. BlyS levels increase after treatment with rituximab. Polymorphisms in the BlyS gene promoter may affect serum BlyS levels and B cell repopulation after rituximab therapy. The polymorphism – 871 C>T in the BlyS promoter has been studied in this respect. In a study of 115 patients with RA, the 871CC genotype was associated with a better EULAR response to rituximab than the TT genotype (OR 6.9, 95% CI 1.6–29.6; p ¼ 0.03) [146]. This association of the 871 C>T SNP with rituximab response could not be confirmed in another study of 152 Italian patients which examined four polymorphisms –2841 T>C, 2704 T>C, 2701 T>A, and–871 C>T in the BLySS promoter that are in linkage disequilibrium. The BLyS haplotype TTTT was associated with a good response to rituximab (OR 14.4, 95% CI 1.77–117.39; p ¼ 0.003) only in patients who were seropositive and had a prior poor response to anti-TNF agents. This finding was replicated in an additional 115 patients in the same study. However, no association was found with the 871 C>T SNP [147]. Some studies looking at the F158V polymorphism in the FCGRIIIA gene (rs396991) found the V variant to be significantly associated with a good response to rituximab therapy. In one study with 111 patients, V allele carriage was associated with a higher response rate (91% of responders vs 70% of non-responders (OR 4.6, 95% CI 1.5–13.6; p ¼ 0.006)) [148]. Similar results were seen in 212 RA patients where 89.5% of patients with the VV genotype had a good EULAR response at 6 months vs. 66% with the VF genotype and 66.2% with the FF genotype ( p ¼ 0.01) [149]. Kastbom et al. however reported that heterozygotes (158 V/F) had a better response than homozygotes (158 VV or 158FF) [150]. Sarsour et al. did not find a difference in response to either rituximab or TNF inhibitors in patients with FCGRIIIA polymorphisms [151]. A small study including 52 Hungarian patients with RA, who had received at least one TNF inhibitor prior to initiation of rituximab, showed a significant reduction in posttreatment DAS28 in patients with 3 FCGR3A genotypes; VV (1.98  0.54, p ¼ 0.008), VF (2.07  0.23, p < 0.001), and FF (1.59 0.52, p ¼ 0.014). A significant difference in posttreatment DAS28 was found between the VF and FF groups ( p ¼ 0.032) as well as between heterozygotes and all homozygotes (VV and FF) ( p ¼ 0.017) [152]. This suggests that carriers of at least one V allele had a better response to rituximab, where the higher binding

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

557

affinity of the V158 isoform may play a role in more effective peripheral B cell depletion. It is also possible that FCGR expression differs between both isoforms, but this was not confirmed in this study. IL6 is another B cell survival factor. Fabris et al. studied the 174G>C polymorphism in IL-6 and the D358A polymorphism in IL-6-receptor alpha (IL-6Ra) genes; these polymorphisms are associated with expression of IL6 and IL6Ra. The IL-6 -174CC genotype (rs1800795) was associated with a poor response to rituximab by EULAR criteria (39.1%) as compared to the GC/CC genotypes (18.5%) (OR 2.83, 95% CI 1.10–7.27; p ¼ 0.031). No association was found between the D358A polymorphism in IL-6-Ra and response to rituximab therapy [153]. The type I interferon (IFN) pathway has been identified as playing a role in response to rituximab as studies have shown that the only difference between rituximab responders and non-responders were changes in expression of genes regulated by type 1 IFN [154]. Several genes encoding proteins that are key components of the type 1 IFN system were analyzed for efficacy following rituximab therapy in 224 patients with RA. Three distinct variants, IRF5 rs2004640 (genotype G/T or G/G), SPP1 rs9138 (genotype A/A), and TNFSF13B rs 9514828 (genotype C/C), were strongly associated with EULAR response to rituximab at week 24. This allelic combination is 2.8 times more likely to be associated with response to rituximab compared with what is seen with TNFSF13B rs9514828 alone suggesting the synergistic effect of the type 1 IFN genetic variants [155]. SPP1 rs9138 has been reported as an RA susceptibility variant seen in anti-citrullinated autoantibody (ACPA) negative individuals with more severe joint damage [156]. Further studies are required to validate these results and establish the individual contribution of these polymorphisms. 2.4.3

Tocilizumab

Tocilizumab is a humanized monoclonal IL6 receptor antagonist. Recently, Wang et al. reported the first GWAS demonstrating genetic variants associated with response to tocilizimab. This study pooled data from six studies with a total of 1683 patients. 253 variants showed an association with tocilizumab therapy, of which 4 SNPs, rs11886534, rs850246, rs13302591, and rs12110787, reached genome-wide significance. Seven of these 253 variants (rs11052877, rs4910008, rs9594987, rs10108210, rs703297, rs703505, and rs1560011) achieved significance on conformational analysis. Of these, rs11052877 is located in the 30 -untranslated region of CD69, and rs1560011 is an intronic SNP in CLEC2D which blocks osteoclast function [157]. 77 Caucasians with RA receiving subcutaneous or intravenous tocilizumab (32.5% are naı¨ve to bDMARD therapy, 88.4% received concomitant MTX or leflunomide, and 92.2% received glucocorticoids) had

558

Atinuke Aluko and Prabha Ranganathan

genomic DNA extracted from saliva samples to study four polymorphisms of IL6R including rs12083537, rs2228145, rs4329505, and rs11265618. In multivariate analysis, none of these polymorphisms was associated with EULAR response. The only variable independently associated with EULAR response at 12 months was lower number of previous bDMARDs used (OR ¼ 0.61, 95% CI 0.42–0.88, p ¼ 0.008). Non-bDMARDnaı¨ve patients had used an average of 2.2 (1–3) bDMARDs during 43.5 (16.5–81.0) months. None of the gene polymorphisms were associated with remission rate either. Similarly, higher remission rates were seen with lower number of previous bDMARDs used (OR ¼ 0.51, 95% CI 0.35–0.722, p ¼ 0.001). Patients with the AA genotype for rs12083537 (OR ¼ 13.0, 95% CI 2.31–72.91, p ¼ 0.004) and CC for rs11265618 (OR ¼ 12.15, 95% IC 2.18–67.81, p ¼ 0.004) had a better response to tocilizumab in terms of low disease activity at 12 months (DAS28 < 3.6); no effect was seen with the other two SNPs [158]. Despite the rs11265618 polymorphism being associated with a better response or low disease activity in this study, it has been associated with susceptibility to RA in an Asian population [159]. A previous study analyzing the effect of IL6R polymorphisms had conversely revealed an association between the A allele of rs12083537 and the C allele of rs4329505 with lack of improvement in swollen joint counts [160].

3

Conclusions and Future Directions There is a growing body of literature on the pharmacogenetics of therapies used in RA. Clearly, inherited differences in drugmetabolizing enzymes, drug receptors, and drug targets are important in determining an individual’s response to a given drug. Nonetheless, several caveats need to be considered before pharmacogenetics can be translated into clinical medicine. In several of the studies reviewed above, the strength of the association between genotype and phenotype can be interrogated for several different reasons, not the least of which is lack of reproducibility. Many of the studies had small sample sizes and were likely underpowered to detect meaningful genetic associations. Some of the studies were retrospective leading to varying biases such as information and selection bias. In addition, race likely has a strong influence on pharmacogenetic associations, and the populations examined in most of the above studies were racially homogenous. Our study examining the frequencies of SNPs in the MTX pathway in different racial groups demonstrated significant differences in the allele frequencies of several SNPs between Caucasians and African-Americans with RA [102]. Hence genotype-phenotype associations may differ significantly in ethnically diverse

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis

559

populations. For example, in the study by Criswell et al., although certain MHC/TNF haplotypes were predictive of response to ETN in a Caucasian population, it is unclear whether these results will apply to other populations [102]. As many of the drugs used in the treatment of RA (such as MTX, TNF antagonists) work through several different cellular (and genetic) pathways, examination of SNPs in different metabolic pathways rather than a single pathway may be more predictive of response [12]. It is also worth noting that if a variant is only weakly associated with response, this may be due to the fact that this variant may occur in tandem or close proximity to the gene which is the actual marker of response. For reasons described above, haplotype analyses may be more useful than single SNP analyses in predicting response [102, 106]. Genome-wide association studies in pharmacogenetics are expanding with some promising results. Finally, the cost-effectiveness of pharmacogenetic testing is an important point to consider before pharmacogenetics can be incorporated into daily clinical practice. Drugs with narrow therapeutic indices, severe side effects, well-established associations between a specific genotype and phenotype (usually toxicity), and for which the frequency of the genetic variant of interest is high are the ideal candidates for pharmacogenetic testing. Notwithstanding these caveats, as genotyping becomes more readily available and less expensive, and major funding agencies display an increasing commitment to pharmacogenetic research (International HapMap Consortium (www.hapmap.org) and the Pharmacogenetics Research Network (http://www.nigms.nih. gov/pharmacogenetics/) by the National Institutes of Health), it is quite likely that genotype-guided therapy of patients with RA will happen in the near future. References 1. Weinblatt ME, Coblyn JS, Fox DA et al (1985) Efficacy of low-dose methotrexate in rheumatoid arthritis. N Engl J Med 312(13): 818–822 2. Bathon JM, Martin RW, Fleischmann RM et al (2000) A comparison of etanercept and methotrexate in patients with early rheumatoid arthritis. N Engl J Med 343(22): 1586–1593 3. Bluett J, Sergeant JC, MacGregor AJ et al (2018) Risk factors for oral methotrexate failure in patients with inflammatory polyarthritis: results from a UK prospective cohort study. Arthritis Res Ther 20(1):50 4. Hooijberg JH, Broxterman HJ, Kool M et al (1999) Antifolate resistance mediated by the

multidrug resistance proteins MRP1 and MRP2. Cancer Res 59(11):2532–2535 5. Barredo JC, Synold TW, Laver J et al (1994) Differences in constitutive and postmethotrexate folylpolyglutamate synthetase activity in B-lineage and T-lineage leukemia. Blood 84(2):564–569 6. Galivan J (1980) Evidence for the cytotoxic activity of polyglutamate derivatives of methotrexate. Mol Pharmacol 17(1):105–110 7. Szeto DW, Cheng YC, Rosowsky A et al (1979) Human thymidylate synthetase--III. Effects of methotrexate and folate analogs. Biochem Pharmacol 28(17):2633–2637 8. Chan ES, Cronstein BN (2002) Molecular action of methotrexate in inflammatory diseases. Arthritis Res 4(4):266–273

560

Atinuke Aluko and Prabha Ranganathan

9. Rothem L, Aronheim A, Assaraf YG (2003) Alterations in the expression of transcription factors and the reduced folate carrier as a novel mechanism of antifolate resistance in human leukemia cells. J Biol Chem 278(11): 8935–8941 10. Rothem L, Stark M, Kaufman Y et al (2004) Reduced folate carrier gene silencing in multiple antifolate-resistant tumor cell lines is due to a simultaneous loss of function of multiple transcription factors but not promoter methylation. J Biol Chem 279(1):374–384 11. Whetstine JR, Witt TL, Matherly LH (2002) The human reduced folate carrier gene is regulated by the AP2 and sp1 transcription factor families and a functional 61-base pair polymorphism. J Biol Chem 277(46): 43873–43880 12. Dervieux T, Kremer J, Lein DO et al (2004) Contribution of common polymorphisms in reduced folate carrier and gammaglutamylhydrolase to methotrexate polyglutamate levels in patients with rheumatoid arthritis. Pharmacogenetics 14(11):733–739 13. Hayashi H, Fujimaki C, Daimon T et al (2009) Genetic polymorphisms in folate pathway enzymes as a possible marker for predicting the outcome of methotrexate therapy in Japanese patients with rheumatoid arthritis. J Clin Pharm Ther 34(3):355–361 14. Hoffmeyer S, Burk O, von Richter O et al (2000) Functional polymorphisms of the human multidrug-resistance gene: multiple sequence variations and correlation of one allele with P-glycoprotein expression and activity in vivo. Proc Natl Acad Sci U S A 97(7):3473–3478 15. Kim RB, Leake BF, Choo EF et al (2001) Identification of functionally variant MDR1 alleles among European Americans and African Americans. Clin Pharmacol Ther 70(2): 189–199 16. Norris MD, De Graaf D, Haber M et al (1996) Involvement of MDR1 P-glycoprotein in multifactorial resistance to methotrexate. Int J Cancer 65(5):613–619 17. Mickisch GH, Merlino GT, Galski H et al (1991) Transgenic mice that express the human multidrug-resistance gene in bone marrow enable a rapid identification of agents that reverse drug resistance. Proc Natl Acad Sci U S A 88(2):547–551 18. Pawlik A, Wrzesniewska J, FiedorowiczFabrycy I et al (2004) The MDR1 3435 polymorphism in patients with rheumatoid arthritis. Int J Clin Pharmacol Ther 42(9):496–503

19. Moya P, Salazar J, Arranz MJ et al (2016) Methotrexate pharmacokinetic genetic variants are associated with outcome in rheumatoid arthritis patients. Pharmacogenomics 17(1):25–29 20. Rozen R (1996) Molecular genetics of methylenetetrahydrofolate reductase deficiency. J Inherit Metab Dis 19(5):589–594 21. Frosst P, Blom HJ, Milos R et al (1995) A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase. Nat Genet 10(1): 111–113 22. van der Put NM, Gabreels F, Stevens EM et al (1998) A second common mutation in the methylenetetrahydrofolate reductase gene: an additional risk factor for neural-tube defects? Am J Hum Genet 62(5):1044–1051 23. Kang SS, Zhou J, Wong PW et al (1988) Intermediate homocysteinemia: a thermolabile variant of methylenetetrahydrofolate reductase. Am J Hum Genet 43(4):414–421 24. Haagsma CJ, Blom HJ, van Riel PL et al (1999) Influence of sulphasalazine, methotrexate, and the combination of both on plasma homocysteine concentrations in patients with rheumatoid arthritis. Ann Rheum Dis 58(2):79–84 25. van Ede AE, Laan RF, Blom HJ et al (2001) The C677T mutation in the methylenetetrahydrofolate reductase gene: a genetic risk factor for methotrexate-related elevation of liver enzymes in rheumatoid arthritis patients. Arthritis Rheum 44(11):2525–2530 26. Berkun Y, Levartovsky D, Rubinow A et al (2004) Methotrexate related adverse effects in patients with rheumatoid arthritis are associated with the A1298C polymorphism of the MTHFR gene. Ann Rheum Dis 63(10): 1227–1231 27. Dervieux T, Greenstein N, Kremer J (2006) Pharmacogenomic and metabolic biomarkers in the folate pathway and their association with methotrexate effects during dosage escalation in rheumatoid arthritis. Arthritis Rheum 54(10):3095–3103 28. Kurzawski M, Pawlik A, Safranow K et al (2007) 677C>T and 1298A>C MTHFR polymorphisms affect methotrexate treatment outcome in rheumatoid arthritis. Pharmacogenomics 8(11):1551–1559 29. Wessels JA, de Vries-Bouwstra JK, Heijmans BT et al (2006) Efficacy and toxicity of methotrexate in early rheumatoid arthritis are associated with single-nucleotide polymorphisms in genes coding for folate pathway enzymes. Arthritis Rheum 54(4):1087–1095

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis 30. Taniguchi A, Urano W, Tanaka E et al (2007) Validation of the associations between single nucleotide polymorphisms or haplotypes and responses to disease-modifying antirheumatic drugs in patients with rheumatoid arthritis: a proposal for prospective pharmacogenomic study in clinical practice. Pharmacogenet Genomics 17(6):383–390 31. Kim SK, Jun JB, El-Sohemy A et al (2006) Cost-effectiveness analysis of MTHFR polymorphism screening by polymerase chain reaction in Korean patients with rheumatoid arthritis receiving methotrexate. J Rheumatol 33(7):1266–1274 32. Xiao H, Xu J, Zhou X et al (2010) Associations between the genetic polymorphisms of MTHFR and outcomes of methotrexate treatment in rheumatoid arthritis. Clin Exp Rheumatol 28(5):728–733 33. Weisman MH, Furst DE, Park GS et al (2006) Risk genotypes in folate-dependent enzymes and their association with methotrexaterelated side effects in rheumatoid arthritis. Arthritis Rheum 54(2):607–612 34. Ranganathan P, Culverhouse R, Marsh S et al (2008) Methotrexate (MTX) pathway gene polymorphisms and their effects on MTX toxicity in Caucasian and African American patients with rheumatoid arthritis. J Rheumatol 35(4):572–579 35. Caliz R, del Amo J, Balsa A et al (2012) The C677T polymorphism in the MTHFR gene is associated with the toxicity of methotrexate in a Spanish rheumatoid arthritis population. Scand J Rheumatol 41(1):10–14 36. Fisher MC, Cronstein BN (2009) Metaanalysis of methylenetetrahydrofolate reductase (MTHFR) polymorphisms affecting methotrexate toxicity. J Rheumatol 36(3):539–545 37. Lee YH, Song GG (2010) Associations between the C677T and A1298C polymorphisms of MTHFR and the efficacy and toxicity of methotrexate in rheumatoid arthritis: a meta-analysis. Clin Drug Investig 30(2): 101–108 38. Owen SA, Lunt M, Bowes J et al (2013) MTHFR gene polymorphisms and outcome of methotrexate treatment in patients with rheumatoid arthritis: analysis of key polymorphisms and meta-analysis of C677T and A1298C polymorphisms. Pharmacogenomics J 13(2):137–147 39. Horie N, Aiba H, Oguro K et al (1995) Functional analysis and DNA polymorphism of the tandemly repeated sequences in the 50 -terminal regulatory region of the human gene for thymidylate synthase. Cell Struct Funct 20(3):191–197

561

40. Kawakami K, Omura K, Kanehira E et al (1999) Polymorphic tandem repeats in the thymidylate synthase gene is associated with its protein expression in human gastrointestinal cancers. Anticancer Res 19(4B):3249–3252 41. DiPaolo A, Chu E (2004) The role of thymidylate synthase as a molecular biomarker. Clin Cancer Res 10(2):411–412 42. Pullarkat ST, Stoehlmacher J, Ghaderi V et al (2001) Thymidylate synthase gene polymorphism determines response and toxicity of 5-FU chemotherapy. Pharmacogenomics J 1(1):65–70 43. Kawakami K, Watanabe G (2003) Identification and functional analysis of single nucleotide polymorphism in the tandem repeat sequence of thymidylate synthase gene. Cancer Res 63(18):6004–6007 44. Ulrich CM, Bigler J, Velicer CM et al (2000) Searching expressed sequence tag databases: discovery and confirmation of a common polymorphism in the thymidylate synthase gene. Cancer Epidemiol Biomark Prev 9(12):1381–1385 45. Kumagai K, Hiyama K, Oyama T et al (2003) Polymorphisms in the thymidylate synthase and methylenetetrahydrofolate reductase genes and sensitivity to the low-dose methotrexate therapy in patients with rheumatoid arthritis. Int J Mol Med 11(5):593–600 46. Lv S, Fan H, Li J et al (2018) Genetic polymorphisms of TYMS, MTHFR, ATIC, MTR, and MTRR are related to the outcome of methotrexate therapy for rheumatoid arthritis in a Chinese population. Front Pharmacol 9: 1390 47. Sherr CJ (1994) G1 phase progression: cycling on cue. Cell 79(4):551–555 48. Betticher DC, Thatcher N, Altermatt HJ et al (1995) Alternate splicing produces a novel cyclin D1 transcript. Oncogene 11(5): 1005–1011 49. Lu F, Gladden AB, Diehl JA (2003) An alternatively spliced cyclin D1 isoform, cyclin D1b, is a nuclear oncogene. Cancer Res 63(21):7056–7061 50. Hochhauser D, Schnieders B, Ercikan-Abali E et al (1996) Effect of cyclin D1 overexpression on drug sensitivity in a human fibrosarcoma cell line. J Natl Cancer Inst 88(18): 1269–1275 51. Jekic B, Lukovic L, Bunjevacki V et al (2013) Association of the TYMS 3G/3G genotype with poor response and GGH 354GG genotype with the bone marrow toxicity of the

562

Atinuke Aluko and Prabha Ranganathan

methotrexate in RA patients. Eur J Clin Pharmacol 69(3):377–383 52. Dervieux T, Furst D, Lein DO et al (2004) Polyglutamation of methotrexate with common polymorphisms in reduced folate carrier, aminoimidazole carboxamide ribonucleotide transformylase, and thymidylate synthase are associated with methotrexate effects in rheumatoid arthritis. Arthritis Rheum 50(9): 2766–2774 53. Grabar PB, Rojko S, Logar D et al (2010) Genetic determinants of methotrexate treatment in rheumatoid arthritis patients: a study of polymorphisms in the adenosine pathway. Ann Rheum Dis 69(5):931–932 54. Wessels JA, van der Kooij SM, le Cessie S et al (2007) A clinical pharmacogenetic model to predict the efficacy of methotrexate monotherapy in recent-onset rheumatoid arthritis. Arthritis Rheum 56(6):1765–1775 55. Lopez-Rodriguez R, Ferreiro-Iglesias A, Lima A et al (2018) Evaluation of a clinical pharmacogenetics model to predict methotrexate response in patients with rheumatoid arthritis. Pharmacogenomics J 18(4): 539–545 56. Stolk JN, Boerbooms AM, de Abreu RA et al (1998) Reduced thiopurine methyltransferase activity and development of side effects of azathioprine treatment in patients with rheumatoid arthritis. Arthritis Rheum 41(10): 1858–1866 57. Marinaki AM, Ansari A, Duley JA et al (2004) Adverse drug reactions to azathioprine therapy are associated with polymorphism in the gene encoding inosine triphosphate pyrophosphatase (ITPase). Pharmacogenetics 14(3): 181–187 58. Haglund S, Taipalensuu J, Peterson C et al (2008) IMPDH activity in thiopurine-treated patients with inflammatory bowel disease relation to TPMT activity and metabolite concentrations. Br J Clin Pharmacol 65(1):69–77 59. Haglund S, Vikingsson S, Soderman J et al (2011) The role of inosine-50 -monophosphate dehydrogenase in thiopurine metabolism in patients with inflammatory bowel disease. Ther Drug Monit 33(2): 200–208 60. Citterio-Quentin A, El Mahmoudi A, Perret T, at al. (2019) Red blood cell IMPDH activity in adults and children with or without azathioprine: relationship between thiopurine metabolites, ITPA and TPMT activities. Basic Clin Pharmacol Toxicol 124(5):600–606

61. Krynetski EY, Tai HL, Yates CR et al (1996) Genetic polymorphism of thiopurine S-methyltransferase: clinical importance and molecular mechanisms. Pharmacogenetics 6(4):279–290 62. Tai HL, Krynetski EY, Yates CR et al (1996) Thiopurine S-methyltransferase deficiency: two nucleotide transitions define the most prevalent mutant allele associated with loss of catalytic activity in Caucasians. Am J Hum Genet 58(4):694–702 63. Tai HL, Krynetski EY, Schuetz EG et al (1997) Enhanced proteolysis of thiopurine S-methyltransferase (TPMT) encoded by mutant alleles in humans (TPMT*3A, TPMT*2): mechanisms for the genetic polymorphism of TPMT activity. Proc Natl Acad Sci U S A 94(12):6444–6449 64. Yates CR, Krynetski EY, Loennechen T et al (1997) Molecular diagnosis of thiopurine S-methyltransferase deficiency: genetic basis for azathioprine and mercaptopurine intolerance. Ann Intern Med 126(8):608–614 65. Ameyaw MM, Collie-Duguid ES, Powrie RH et al (1999) Thiopurine methyltransferase alleles in British and Ghanaian populations. Hum Mol Genet 8(2):367–370 66. Hon YY, Fessing MY, Pui CH et al (1999) Polymorphism of the thiopurine S-methyltransferase gene in AfricanAmericans. Hum Mol Genet 8(2):371–376 67. Evans WE, Hon YY, Bomgaars L et al (2001) Preponderance of thiopurine S-methyltransferase deficiency and heterozygosity among patients intolerant to mercaptopurine or azathioprine. J Clin Oncol 19(8): 2293–2301 68. Black AJ, McLeod HL, Capell HA et al (1998) Thiopurine methyltransferase genotype predicts therapy-limiting severe toxicity from azathioprine. Ann Intern Med 129(9): 716–718 69. Corominas H, Domenech M, Laiz A et al (2003) Is thiopurine methyltransferase genetic polymorphism a major factor for withdrawal of azathioprine in rheumatoid arthritis patients? Rheumatology (Oxford) 42(1):40–45 70. Boonsrirat U, Angsuthum S, Vannaprasaht S et al (2008) Azathioprine-induced fatal myelosuppression in systemic lupus erythematosus patient carrying TPMT*3C polymorphism. Lupus 17(2):132–134 71. Higgs JE, Payne K, Roberts C et al (2010) Are patients with intermediate TPMT activity at increased risk of myelosuppression when

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis taking thiopurine medications? Pharmacogenomics 11(2):177–188 72. von Ahsen N, Armstrong VW, Behrens C et al (2005) Association of inosine triphosphatase 94C>A and thiopurine S-methyltransferase deficiency with adverse events and study drop-outs under azathioprine therapy in a prospective Crohn disease study. Clin Chem 51(12):2282–2288 73. Shipkova M, Franz J, Abe M et al (2011) Association between adverse effects under azathioprine therapy and inosine triphosphate pyrophosphatase activity in patients with chronic inflammatory bowel disease. Ther Drug Monit 33(3):321–328 74. Gearry RB, Roberts RL, Barclay ML et al (2004) Lack of association between the ITPA 94C>A polymorphism and adverse effects from azathioprine. Pharmacogenetics 14(11):779–781 75. Eklund BI, Moberg M, Bergquist J et al (2006) Divergent activities of human glutathione transferases in the bioactivation of azathioprine. Mol Pharmacol 70(2):747–754 76. Stocco G, Martelossi S, Barabino A et al (2007) Glutathione-S-transferase genotypes and the adverse effects of azathioprine in young patients with inflammatory bowel disease. Inflamm Bowel Dis 13(1):57–64 77. Stocco G, Cuzzoni E, De Iudicibus S et al (2014) Deletion of glutathione-s-transferase m1 reduces azathioprine metabolite concentrations in young patients with inflammatory bowel disease. J Clin Gastroenterol 48(1): 43–51 78. Kerstens PJ, Stolk JN, De Abreu RA et al (1995) Azathioprine-related bone marrow toxicity and low activities of purine enzymes in patients with rheumatoid arthritis. Arthritis Rheum 38(1):142–145 79. Yang J, Wang P, Qin Z et al (2019) NUDT15 and TPMT genetic polymorphisms are related to azathioprine intolerance in Chinese patients with rheumatic diseases. Genet Test Mol Biomarkers 23(10):751–757 80. Relling MV, Schwab M, Whirl-Carrillo M et al (2019) Clinical Pharmacogenetics implementation consortium guideline for thiopurine dosing based on TPMT and NUDT15 genotypes: 2018 update. Clin Pharmacol Ther 105(5):1095–1105 81. Das KM, Eastwood MA, McManus JP et al (1973) Adverse reactions during salicylazosulfapyridine therapy and the relation with drug metabolism and acetylator phenotype. N Engl J Med 289(10):491–495

563

82. Pullar T, Capell HA (2006) Variables affecting efficacy and toxicity of sulphasalazine in rheumatoid arthritis. A review. Drugs 32(Suppl 1): 54–57 83. Zaher H, Khan AA, Palandra J et al (2006) Breast cancer resistance protein (Bcrp/abcg2) is a major determinant of sulfasalazine absorption and elimination in the mouse. Mol Pharm 3(1):55–61 84. Yamasaki Y, Ieiri I, Kusuhara H et al (2008) Pharmacogenetic characterization of sulfasalazine disposition based on NAT2 and ABCG2 (BCRP) gene polymorphisms in humans. Clin Pharmacol Ther 84(1):95–103 85. Wiese MD, Alotaibi N, O’Doherty C et al (2014) Pharmacogenomics of NAT2 and ABCG2 influence the toxicity and efficacy of sulphasalazine containing DMARD regimens in early rheumatoid arthritis. Pharmacogenomics J 14(4):350–355 86. Tanaka E, Taniguchi A, Urano W et al (2002) Adverse effects of sulfasalazine in patients with rheumatoid arthritis are associated with diplotype configuration at the N-acetyltransferase 2 gene. J Rheumatol 29(12):2492–2499 87. Wadelius M, Stjernberg E, Wiholm BE et al (2000) Polymorphisms of NAT2 in relation to sulphasalazine-induced agranulocytosis. Pharmacogenetics 10(1):35–41 88. Kuhn UD, Anschutz M, Schmucker K et al (2010) Phenotyping with sulfasalazine - time dependence and relation to NAT2 pharmacogenetics. Int J Clin Pharmacol Ther 48(1): 1–10 89. Yee J, Kim SM, Han JM et al (2020) The association between NAT2 acetylator status and adverse drug reactions of sulfasalazine: a systematic review and meta-analysis. Sci Rep 10(1):3658 90. Genovese MC, Bathon JM, Martin RW et al (2002) Etanercept versus methotrexate in patients with early rheumatoid arthritis: two-year radiographic and clinical outcomes. Arthritis Rheum 46(6):1443–1450 91. Maini RN, Breedveld FC, Kalden JR et al (1998) Therapeutic efficacy of multiple intravenous infusions of anti-tumor necrosis factor alpha monoclonal antibody combined with low-dose weekly methotrexate in rheumatoid arthritis. Arthritis Rheum 41(9):1552–1563 92. Keystone EC, Kavanaugh AF, Sharp JT et al (2004) Radiographic, clinical, and functional outcomes of treatment with adalimumab (a human anti-tumor necrosis factor monoclonal antibody) in patients with active rheumatoid arthritis receiving concomitant methotrexate therapy: a randomized,

564

Atinuke Aluko and Prabha Ranganathan

placebo-controlled, 52-week trial. Arthritis Rheum 50(5):1400–1411 93. Seitz M, Wirthmuller U, Moller B et al (2007) The 308 tumour necrosis factor-alpha gene polymorphism predicts therapeutic response to TNFalpha-blockers in rheumatoid arthritis and spondyloarthritis patients. Rheumatology (Oxford) 46(1):93–96 94. Cuchacovich M, Ferreira L, Aliste M et al (2004) Tumour necrosis factor-alpha (TNF-alpha) levels and influence of 308 TNF-alpha promoter polymorphism on the responsiveness to infliximab in patients with rheumatoid arthritis. Scand J Rheumatol 33(4):228–232 95. Pavy S, Toonen EJ, Miceli-Richard C et al (2010) Tumour necrosis factor alpha 308G->A polymorphism is not associated with response to TNFalpha blockers in Caucasian patients with rheumatoid arthritis: systematic review and meta-analysis. Ann Rheum Dis 69(6):1022–1028 96. Zeng Z, Duan Z, Zhang T et al (2013) Association between tumor necrosis factor-alpha (TNF-alpha) promoter 308 G/A and response to TNF-alpha blockers in rheumatoid arthritis: a meta-analysis. Mod Rheumatol 23(3):489–495 97. Maxwell JR, Potter C, Hyrich KL et al (2008) Association of the tumour necrosis factor-308 variant with differential response to anti-TNF agents in the treatment of rheumatoid arthritis. Hum Mol Genet 17(22):3532–3538 98. Fabris M, Di Poi E, D’Elia A et al (2002) Tumor necrosis factor-alpha gene polymorphism in severe and mild-moderate rheumatoid arthritis. J Rheumatol 29(1):29–33 99. Morita C, Horiuchi T, Tsukamoto H et al (2001) Association of tumor necrosis factor receptor type II polymorphism 196R with systemic lupus erythematosus in the Japanese: molecular and functional analysis. Arthritis Rheum 44(12):2819–2827 100. Ongaro A, De Mattei M, Pellati A et al (2008) Can tumor necrosis factor receptor II gene 676T>G polymorphism predict the response grading to anti-TNFalpha therapy in rheumatoid arthritis? Rheumatol Int 28(9):901–908 101. Vasilopoulos Y, Bagiatis V, Stamatopoulou D et al (2011) Association of anti-CCP positivity and carriage of TNFRII susceptibility variant with anti-TNF-alpha response in rheumatoid arthritis. Clin Exp Rheumatol 29(4):701–704 102. Criswell LA, Lum RF, Turner KN et al (2004) The influence of genetic variation in the HLA-DRB1 and LTA-TNF regions on the response to treatment of early rheumatoid

arthritis with methotrexate or etanercept. Arthritis Rheum 50(9):2750–2756 103. Gregersen PK, Silver J, Winchester RJ (1987) The shared epitope hypothesis. An approach to understanding the molecular genetics of susceptibility to rheumatoid arthritis. Arthritis Rheum 30(11):1205–1213 104. Waldron-Lynch F, Adams C, Amos C et al (2001) Tumour necrosis factor 50 promoter single nucleotide polymorphisms influence susceptibility to rheumatoid arthritis (RA) in immunogenetically defined multiplex RA families. Genes Immun 2(2):82–87 105. Mulcahy B, Waldron-Lynch F, McDermott MF et al (1996) Genetic variability in the tumor necrosis factor-lymphotoxin region influences susceptibility to rheumatoid arthritis. Am J Hum Genet 59(3):676–683 106. Martinez A, Salido M, Bonilla G et al (2004) Association of the major histocompatibility complex with response to infliximab therapy in rheumatoid arthritis patients. Arthritis Rheum 50(4):1077–1082 107. Stahl EA, Raychaudhuri S, Remmers EF et al (2010) Genome-wide association study metaanalysis identifies seven new rheumatoid arthritis risk loci. Nat Genet 42(6):508–514 108. Cui J, Saevarsdottir S, Thomson B et al (2010) Rheumatoid arthritis risk allele PTPRC is also associated with response to anti-tumor necrosis factor alpha therapy. Arthritis Rheum 62(7):1849–1861 109. Plant D, Prajapati R, Hyrich KL et al (2012) Replication of association of the PTPRC gene with response to anti-tumor necrosis factor therapy in a large UK cohort. Arthritis Rheum 64(3):665–670 110. Krintel SB, Palermo G, Johansen JS et al (2012) Investigation of single nucleotide polymorphisms and biological pathways associated with response to TNFalpha inhibitors in patients with rheumatoid arthritis. Pharmacogenet Genomics 22(8):577–589 111. Bowes JD, Potter C, Gibbons LJ et al (2009) Investigation of genetic variants within candidate genes of the TNFRSF1B signalling pathway on the response to anti-TNF agents in a UK cohort of rheumatoid arthritis patients. Pharmacogenet Genomics 19(4):319–323 112. Coulthard LR, Taylor JC, Eyre S et al (2011) Genetic variants within the MAP kinase signalling network and anti-TNF treatment response in rheumatoid arthritis patients. Ann Rheum Dis 70(1):98–103 113. Barton A, Eyre S, Ke X et al (2009) Identification of AF4/FMR2 family, member 3 (AFF3) as a novel rheumatoid arthritis susceptibility locus and confirmation of two

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis further pan-autoimmune susceptibility genes. Hum Mol Genet 18(13):2518–2522 114. Hafler JP, Maier LM, Cooper JD et al (2009) CD226 Gly307Ser association with multiple autoimmune diseases. Genes Immun 10(1): 5–10 115. Tan RJ, Gibbons LJ, Potter C et al (2010) Investigation of rheumatoid arthritis susceptibility genes identifies association of AFF3 and CD226 variants with response to antitumour necrosis factor treatment. Ann Rheum Dis 69(6):1029–1035 116. Conigliaro P, Ciccacci C, Politi C et al (2017) in STAT4, PTPN2, Polymorphisms PSORS1C1 and TRAF3IP2 genes are associated with the response to TNF inhibitors in patients with rheumatoid arthritis. PLoS One 12(1):e0169956 117. Marotte H, Pallot-Prades B, Grange L et al (2006) The shared epitope is a marker of severity associated with selection for, but not with response to, infliximab in a large rheumatoid arthritis population. Ann Rheum Dis 65(3):342–347 118. Padyukov L, Lampa J, Heimburger M et al (2003) Genetic markers for the efficacy of tumour necrosis factor blocking therapy in rheumatoid arthritis. Ann Rheum Dis 62(6): 526–529 119. Schotte H, Schluter B, Drynda S et al (2005) Interleukin 10 promoter microsatellite polymorphisms are associated with response to long term treatment with etanercept in patients with rheumatoid arthritis. Ann Rheum Dis 64(4):575–581 120. Jancic I, Arsenovic-Ranin N, Sefik-Bukilica M et al (2013) 174G/C interleukin-6 gene promoter polymorphism predicts therapeutic response to etanercept in rheumatoid arthritis. Rheumatol Int 33(6):1481–1486 121. Sode J, Vogel U, Bank S et al (2014) AntiTNF treatment response in rheumatoid arthritis patients is associated with genetic variation in the NLRP3-inflammasome. PLoS One 9(6):e100361 122. Sode J, Vogel U, Bank S et al (2015) Genetic variations in pattern recognition receptor loci are associated with anti-TNF response in patients with rheumatoid arthritis. PLoS One 10(10):e0139781 123. Potter C, Cordell HJ, Barton A et al (2010) Association between anti-tumour necrosis factor treatment response and genetic variants within the TLR and NF{kappa}B signalling pathways. Ann Rheum Dis 69(7):1315–1320 124. Sode J, Vogel U, Bank S et al (2018) Confirmation of an IRAK3 polymorphism as a genetic marker predicting response to anti-

565

TNF treatment in rheumatoid arthritis. Pharmacogenomics J 18(1):81–86 125. Iwaszko M, Wielinska J, Swierkot J et al (2021) IL-33 gene polymorphisms as potential biomarkers of disease susceptibility and response to TNF inhibitors in rheumatoid arthritis, ankylosing spondylitis, and psoriatic arthritis patients. Front Immunol 12:631603 126. Tutuncu Z, Kavanaugh A, Zvaifler N et al (2005) Fcgamma receptor type IIIA polymorphisms influence treatment outcomes in patients with inflammatory arthritis treated with tumor necrosis factor alpha-blocking agents. Arthritis Rheum 52(9):2693–2696 127. Canete JD, Suarez B, Hernandez MV et al (2009) Influence of variants of Fc gamma receptors IIA and IIIA on the American College of Rheumatology and European League Against Rheumatism responses to antitumour necrosis factor alpha therapy in rheumatoid arthritis. Ann Rheum Dis 68(10): 1547–1552 128. Tsukahara S, Ikari K, Sato E et al (2008) A polymorphism in the gene encoding the Fcgamma IIIA receptor is a possible genetic marker to predict the primary response to infliximab in Japanese patients with rheumatoid arthritis. Ann Rheum Dis 67(12): 1791–1792 129. Kastbom A, Bratt J, Ernestam S et al (2007) Fcgamma receptor type IIIA genotype and response to tumor necrosis factor alphablocking agents in patients with rheumatoid arthritis. Arthritis Rheum 56(2):448–452 130. Liu C, Batliwalla F, Li W et al (2008) Genome-wide association scan identifies candidate polymorphisms associated with differential response to anti-TNF treatment in rheumatoid arthritis. Mol Med 14(9–10): 575–581 131. Umicevic Mirkov M, Cui J, Vermeulen SH et al (2013) Genome-wide association analysis of anti-TNF drug response in patients with rheumatoid arthritis. Ann Rheum Dis 72(8): 1375–1381 132. Martin M, Romero X, de la Fuente MA et al (2001) CD84 functions as a homophilic adhesion molecule and enhances IFN-gamma secretion: adhesion is mediated by Ig-like domain 1. J Immunol 167(7): 3668–3676 133. Tangye SG, Nichols KE, Hare NJ et al (2003) Functional requirements for interactions between CD84 and Src homology 2 domain-containing proteins and their contribution to human T cell activation. J Immunol 171(5):2485–2495 134. Cui J, Stahl EA, Saevarsdottir S et al (2013) Genome-wide association study and gene

566

Atinuke Aluko and Prabha Ranganathan

expression analysis identifies CD84 as a predictor of response to etanercept therapy in rheumatoid arthritis. PLoS Genet 9(3): e1003394 135. Acosta-Colman I, Palau N, Tornero J et al (2013) GWAS replication study confirms the association of PDE3A-SLCO1C1 with antiTNF therapy response in rheumatoid arthritis. Pharmacogenomics 14(7):727–734 136. Honne K, Hallgrimsdottir I, Wu C et al (2016) A longitudinal genome-wide association study of anti-tumor necrosis factor response among Japanese patients with rheumatoid arthritis. Arthritis Res Ther 18:12 137. Ferreiro-Iglesias A, Montes A, Perez-Pampin E et al (2019) Evaluation of 12 GWAS-drawn SNPs as biomarkers of rheumatoid arthritis response to TNF inhibitors. A potential SNP association with response to etanercept. PLoS One 14(2):e0213073 138. Bek S, Bojesen AB, Nielsen JV et al (2017) Systematic review and meta-analysis: pharmacogenetics of anti-TNF treatment response in rheumatoid arthritis. Pharmacogenomics J 17(5):403–411 139. Hensor EMA, McKeigue P, Ling SF et al (2019) Validity of a two-component imaging-derived disease activity score for improved assessment of synovitis in early rheumatoid arthritis. Rheumatology (Oxford) 58(8): 1400–1409 140. Gilani SS, Nair N, Plant D et al (2020) Pharmacogenetics of TNF inhibitor response in rheumatoid arthritis utilizing the two-component disease activity score. Pharmacogenomics 21(16):1151–1156 141. Guan Y, Zhang H, Quang D et al (2019) Machine learning to predict anti-tumor necrosis factor drug responses of rheumatoid arthritis patients by integrating clinical and genetic markers. Arthritis Rheumatol 71(12):1987–1996 142. Daien CI, Fabre S, Rittore C et al (2012) TGF beta1 polymorphisms are candidate predictors of the clinical response to rituximab in rheumatoid arthritis. Joint Bone Spine 79(5): 471–475 143. Coenen MJ, Gregersen PK (2009) Rheumatoid arthritis: a view of the current genetic landscape. Genes Immun 10(2):101–111 144. Boissier MC (2011) Cell and cytokine imbalances in rheumatoid synovitis. Joint Bone Spine 78(3):230–234 145. Guo Z, Binswanger U, Knoflach A (2002) Role of codon 10 and codon 25 polymorphisms on TGF-beta 1 gene expression and protein synthesis in stable renal allograft recipients. Transplant Proc 34(7):2904–2906

146. Ruyssen-Witrand A, Rouanet S, Combe B et al (2013) Association between -871C>T promoter polymorphism in the B-cell activating factor gene and the response to rituximab in rheumatoid arthritis patients. Rheumatology (Oxford) 52(4):636–641 147. Fabris M, Quartuccio L, Vital E et al (2013) The TTTT B lymphocyte stimulator promoter haplotype is associated with good response to rituximab therapy in seropositive rheumatoid arthritis resistant to tumor necrosis factor blockers. Arthritis Rheum 65(1): 88–97 148. Ruyssen-Witrand A, Rouanet S, Combe B et al (2012) Fcgamma receptor type IIIA polymorphism influences treatment outcomes in patients with rheumatoid arthritis treated with rituximab. Ann Rheum Dis 71(6):875–877 149. Quartuccio L, Fabris M, Pontarini E et al (2014) The 158VV Fcgamma receptor 3A genotype is associated with response to rituximab in rheumatoid arthritis: results of an Italian multicentre study. Ann Rheum Dis 73(4):716–721 150. Kastbom A, Coster L, Arlestig L et al (2012) Influence of FCGR3A genotype on the therapeutic response to rituximab in rheumatoid arthritis: an observational cohort study. BMJ Open 2(5):e001524 151. Sarsour K, Greenberg J, Johnston JA et al (2013) The role of the FcGRIIIa polymorphism in modifying the association between treatment and outcome in patients with rheumatoid arthritis treated with rituximab versus TNF-alpha antagonist therapies. Clin Exp Rheumatol 31(2):189–194 152. Pal I, Szamosi S, Hodosi K et al (2017) Effect of Fcgamma-receptor 3a (FCGR3A) gene polymorphisms on rituximab therapy in Hungarian patients with rheumatoid arthritis. RMD Open 3(2):e000485 153. Fabris M, Quartuccio L, Lombardi S et al (2012) The CC homozygosis of the -174G>C IL-6 polymorphism predicts a lower efficacy of rituximab therapy in rheumatoid arthritis. Autoimmun Rev 11(5): 315–320 154. Vosslamber S, Raterman HG, van der Pouw Kraan TC et al (2011) Pharmacological induction of interferon type I activity following treatment with rituximab determines clinical response in rheumatoid arthritis. Ann Rheum Dis 70(6):1153–1159 155. Juge PA, Gazal S, Constantin A et al (2017) Variants of genes implicated in type 1 interferon pathway and B-cell activation modulate the EULAR response to rituximab at

Pharmacogenetics of Drug Therapies in Rheumatoid Arthritis 24 weeks in rheumatoid arthritis. RMD Open 3(2):e000448 156. Juge PA, van Steenbergen HW, Constantin A et al (2014) SPP1 rs9138 variant contributes to the severity of radiological damage in anticitrullinated protein autoantibody-negative rheumatoid arthritis. Ann Rheum Dis 73(10):1840–1843 157. Wang J, Bansal AT, Martin M et al (2013) Genome-wide association analysis implicates the involvement of eight loci with response to tocilizumab for the treatment of rheumatoid arthritis. Pharmacogenomics J 13(3): 235–241 158. Enevold C, Baslund B, Linde L et al (2014) Interleukin-6-receptor polymorphisms rs12083537, rs2228145, and rs4329505 as predictors of response to tocilizumab in rheumatoid arthritis. Pharmacogenet Genomics 24(8):401–405

567

159. Liu X, Xu J, Hu CD et al (2014) The relationship between SNPs in the genes of TLR signal transduction pathway downstream elements and rheumatoid arthritis susceptibility. Tsitol Genet 48(3):24–29 160. Maldonado-Montoro M, Canadas-Garre M, Gonzalez-Utrilla A et al (2018) Influence of IL6R gene polymorphisms in the effectiveness to treatment with tocilizumab in rheumatoid arthritis. Pharmacogenomics J 18(1): 167–172 161. Ranganathan P, McLeod H (2006) Methotrexate pharmacogenetics: first step toward individualized therapy in rheumatoid arthritis. Arthritis and Rheumatism 54(5): 1366–1377 162. Ranganathan P (2005) Pharmacogenomics of tumor necrosis factor antagonists in rheumatoid arthritis. Pharmacogenomics 2(4): 279–282

Chapter 20 Pharmacogenomics in Children Michael J. Rieder and Abdelbaset A. Elzagallaai Abstract Historically genetics has not been considered when prescribing drugs for children. However, it is clear that genetics are not only an important determinant of disease in children but also of drug response for many important drugs that are core agents used in the therapy of common problems in children. Advances in therapy and in the ethical construct of children’s research have made pharmacogenomic assessment for children much easier to pursue. It is likely that pharmacogenomics will become part of the therapeutic decision-making process for children, notably in areas such as childhood cancer where weighing benefits and risks of therapy is crucial. Key words Children, Pharmacogenomics, Pharmacology, Drug safety, Genetics, Drug ontogeny, Childhood cancer

1

Children and Genetics While the expanding interest in pharmacogenomics and personalized medicine over the past decades suggests that this is a recent phenomenon, in fact there has been interest in how genetically determined variations may impact therapy for children for many years, dating back to comments made by Sir Archibald Garrod, the father of inborn errors of metabolism, who at the dawn of the last century observed that, in addition to controlling key metabolic pathways, genetics was also likely to control some of the variations observed in terms of response to drugs [1]. This should not be surprising. The role of genetics in human disease has been recognized by pediatricians for many years. Given the impact of disorders with a genetic basis such as cystic fibrosis and Down syndrome, children’s health-care clinicians and researchers have been sensitized to the importance of genetics to a greater extent than colleagues with a primary focus on adults. Indeed, children were among the patients studied by David Price Evans and colleagues in their landmark paper describing the genetic control of isoniazid metabolism [2]. However, despite the

Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_20, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

569

570

Michael J. Rieder and Abdelbaset A. Elzagallaai

clear and compelling evidence that genetically determined variations in drug action and clearance should be important as outlined above, pharmacogenomics has been a relative later comer to the forefront in pediatric research. There are several likely reasons for this. While it was appreciated that genetics may play an important role in understanding sources of variability in drug response in children, it was also appreciated that ontogeny – changes in drug disposition related to maturation of key elements in human drug disposition and clearance – was a major issue, certainly with respect to drug safety [3]. Indeed, much of the fundamental research conducted in pediatric clinical pharmacology over the past three decades has focused on understanding the impact of ontogeny on drug handling in children and the subsequent impact on efficacy and safety [4]. There have been additional pragmatic issues that have limited the extent to which pharmacogenomics could be studied in children. These included the amount of blood initially required to conduct studies and the practical problems in obtaining biological samples from children, especially very young children, as well as ethical issues relative to the fact that the person consenting for genetic studies was not the person on whom the study was being conducted [5, 6]. Additionally, there have been several myths with respect to drug utilization in children – that drugs were used relatively infrequently in the care of most children and then that the vast majority of drugs used were antibiotics – that have reduced the enthusiasm of investigators to pursue these studies in children [7]. Over the past decade, many of these challenges have been addressed, and there has been a substantial increase in the amount and quality of pharmacogenomic research being conducted in children. The volume of blood needed and the cost of doing analysis have dramatically decreased, while the use of alternate sample sources – such as saliva – has made the conduct of studies much more feasible. There has been research and discourse on the issues of the ethical conduct of genetic studies in children. It has also been appreciated that drug use in children is indeed both common and complex, with studies showing, for example, that on average a Canadian child has four prescriptions per year and that these are from a range of 2400 therapeutic entities [7]!

2

Sources of Variation in Drug Response in Children The paradox of drug development is that drugs are developed and evaluated for safety and efficacy in populations, but the clinician treats individual patients in whom drugs either work, do not work, or cause harm [8]. This paradox is especially germane for children, in that many drugs used for the routine care of children have been approved based on studies in adult patients. In this case, drugs are

Pharmacogenomics in Children

571

used off-label – that is to say, use of the drug is not guided by dosage, safety, or efficacy data from the product monograph [9– 11]. Off-label drug use is not uncommon – indeed off-label drug use in adults frequently happens for off-indication prescribing – but in pediatrics this is also frequently off-evidence [12–14]. While in many cases off-label drug use in adult medicine is in the context of robust data supporting the indication, dose, and safety profile of the drug in question, frequently this data is lacking for children. In children, off-label drug use has been associated with an increased risk of adverse drug reactions [14, 15]. A key issue is variability in drug response, which is to say that there are some children in whom the drug works and is safe, some in which it does not produce the desired therapeutic effect and some in which the drug produces adverse events. There are a number of sources of variation in drug response in children [16]. These include the usual sources of variability in adults such as host-dependent variability in drug disposition and clearance, the impact of the disease being treated or other concurrent disease, and the effect(s) of other drugs or environmental influences (Fig. 1). 2.1 Ontogeny as a Source of Drug Response Variation

The impact of ontogeny on drug disposition, effect, and clearance is extremely important, notably for children under a year of age [4]. During different stages of early development, children form distinct populations regarding response to drugs. This phenomenon has been observed early on signifying the fact that drug response in children cannot be merely extrapolated from adult Variability

Health of the child Concurrent disease Other drugs Ontogeny Genetics

Fig. 1 Sources of variability in drug response in children. These include factors inherent to the child, the effect(s) of the disease being treated or other diseases, the effects of other therapies, ontogeny, and genetically controlled variation in drug disposition, action, and clearance

572

Michael J. Rieder and Abdelbaset A. Elzagallaai

data [17]. Factors that differ by the child’s age and may affect drug response include body water and fat contents, gastric pH, gastric emptying rate and gastric motility, and lower plasma protein concentration [4, 18]. In addition, other parameters such as the development of the blood-brain barrier (BBB), the expression and activity of drug metabolizing enzymes, and the ability of the kidney to metabolize and excrete drugs and drug metabolites continue to develop during the early years of life resulting in variations in drug responses and adverse reactions [19–23]. A final factor that to date has been rarely considered for children is how genetically determined variations in drug disposition and clearance – or in drug response – may impact on variability [24–27]. As noted above, the fact that genetics may be important, the well-being of children is not a mystery to child health-care providers. However, many of the genetic disorders that were historically most relevant to child health were chromosomal polysomies such as Down syndrome or disorders with classical Mendelian or X-linked inheritance such as cystic fibrosis or Duchenne muscular dystrophy in which mechanism or effects can be traced to a single event (polysomy) or a clear mechanistic pathway (e.g., reduced function of the cystic fibrosis transmembrane conductance regulator or alterations in dystrophin), while many pharmacogenetically determined variations are due to more complex mechanisms or the interactions of several genes. The degree to which these variations are clinically relevant in children has been debated, and the ongoing controversy as to how useful pharmacogenetic testing is likely to be in patient care has been part of the reason that pharmacogenomic approaches have been relatively slow to come to the clinic in children compared to adults. This is perhaps best illustrated by comparing two drugs for whom pharmacogenomic determinants have been identified that determine toxicity – 6-mercaptorpurine and codeine. 2.2 6Mercaptopurine and Codeine: A Tale of Two Drugs

An enzyme for which pharmacogenetic variability was identified for some time is thiopurine methyltransferase (TPMT). This enzyme catalyzes the S-methylation of thiopurines such as the chemotherapeutic agent 6-mercaptopurine (Fig. 2) [28]. It was appreciated in the 1970s that there was considerable variability in toxicity when this drug was administered to patients with cancer. A landmark study conducted by Drs. Richard Weinshilboum and Susan Sladek at the Mayo Clinic established that there was a pharmacogenetic basis for this variability, an activity pattern consistent with autosomal codominant inheritance for alleles for low and high TPMT activity; 88.6% of subjects had high enzyme activity, 11.1% with intermediate activity, and 0.3% with no detectable activity [29]. Relevant to considerations of pharmacogenomics in children, 115 of the subjects in this study were children, on average age of 13 years. Translating this to clinical relevance, it has been demonstrated that

Pharmacogenomics in Children

573

Fig. 2 Metabolism of 6-mercaptopurine (6-MP) via xanthine oxidase (XO) to the inactive metabolite 6-thiouric acid (6-TU), thiopurine S-methyltransferase (TPMT) to the inactive metabolite 6-methymercaptopurine (6-MMP), and hypoxanthine guanine phosphoribosyl transferase (HPRT) to 6-thioinosine monophosphate (6-TIMP) which is then further metabolized to thioguanine nucleotides (6-TGN), 6-methylmercaptopurine ribonucleotides (6-MMPR), or 6-thio-inosine triphosphate (6-thio-ITP), these all being active metabolites

patients who are homogenous for low TPTM activity – the 0.3% with no detectable activity – were at significantly greater risk for toxicity – such as neutropenia – when being treated with 6-mercaptopurine [30]. TPMT is inherited as a monogenic autosomal codominant trait, and its activity can vary in three genetically distinct population, normal metabolizers, intermediate metabolizers, and poor metabolizers, which depends on the number of lost functional alleles [29]. TPMT*2, TPMT*3A, TPMT*3B, and TPMT*3C are the common variants representing around 90% of intermediate and poor activity in Caucasian populations [31]. It has been shown that patients who are poor and intermediate metabolizers are prone to higher rates of ADRs to treatment with 6-MP when treated with regular doses of the drug compared to normal metabolizers [32– 35]. Genome-wide association studies (GWAS) also identified other genetic variations that are associated with higher incidence of thiopurines-induced myelosuppression [36–45]. Given the potential mortality and known morbidity associated with febrile neutropenia, it may seem surprising that the routine use of TPMT genotyping in defining dose regimens for children with cancer did not gain early widespread acceptance in the broader community of pediatric oncologists [46]. There are several reasons for this. A pragmatic reason is the volume of blood required at the

574

Michael J. Rieder and Abdelbaset A. Elzagallaai

time to conduct genotyping studies. This was at the time significantly larger than the amount of blood needed to monitor for toxicity – this being by a complete blood count. The cost of genotyping has historically been a consideration, previously being several folds larger than the cost of most routine laboratory tests. The outcome of interest – neutropenia – had been known for some times, and protocols were in place to evaluate this using simple assessments – such as routine use of complete blood counts. The value of genotyping above and beyond conventional monitoring was initially well defined. Finally, there is the consideration of frequency. There was a debate as to just how valuable routine genotyping was for the broader population, given that the frequency of patients homozygous for low activity genes was in the range of 0.3–0.5%. A key element of the debate was the extent to which heterozygotes were likely to need dosage alteration [47]. While this was not clear initially, there has been emerging data that suggests that dose alteration may need to occur in other groups than the homozygous low activity patients – for example, patients with high activity may need increased dosage, while risk of secondary malignancies may be related in part to variability in TPMT activity [48, 49]. It has been increasingly appreciated that gene-gene interactions may play an important role in determining toxicity and that analysis to determine risk should factor in multiple variables including age and concurrent therapy as well as genotypic variation [50, 51]. In this case, a pharmacogenetic determinant of variability was described more than 30 years ago, and the precise role(s) of how this variant will be used to alter therapy are still being investigated and defined. However, it is safe to say that there is increasing acceptance that routine genotyping of TPMT is standard of care in the context of the therapy for childhood cancer. As well, there is increasing interest in the use of TPMT genotyping to guide dosing for other drugs and in other disorders, for example, in treating patients with inflammatory bowel disease with azathioprine [52]. Let us now consider codeine. Codeine is an opiate alkaloid that is the second most abundant alkaloid in opium. Although present naturally, codeine for therapeutic use is most commonly synthesized. Codeine has been recognized for more than a century as having analgesic properties and has enjoyed widespread use for this indication, including being listed as part of the WHO “pain ladder” [53–55]. Codeine is considered a weak opiate and has been recommended as part of a step-wise approach to treating pain. There are many advantages to using codeine; it is inexpensive, available as an oral formulation in both liquid and tablet form, and extremely stable. However, there are important – and until recently largely unrecognized – pharmacogenomic variables in drug disposition which can significantly alter the benefit-risk profile of codeine.

Pharmacogenomics in Children

575

H H

H H

Morphine 6-Glucuronide (M6G) (potent agonist)

UGT2B7 UGT1A1

Morphine 6-Sulfate (Active)

CYP 2D6

H Codeine

UGT2B7 UGT1A1

Codeine 6-Glucuronide

Morphine

N-demethylation

H Morphine 3-Sulfate (Inactive)

H H Morphine 3-Glucuronide (M3G) (partial agonist)

H Normorphine (Inactive)

Fig. 3 Metabolism of codeine to the active metabolites morphine and codeine-6-glucuronide, the demethylation to morphine being catalyzed by CYP2D6; morphine is further metabolized to morphine-3-glucuronide and morphine-6-glucuronide. Active and inactive metabolites are shown

To appreciate these differences, it is important to first recognize that codeine is not in and of itself an effective analgesic but rather is a pro-drug. Chemically codeine is 3-methylmorphine, which pharmacologically is a relatively inactive compound. After ingestion of codeine, the drug enters the liver via the portal circulation where it undergoes metabolism via both Phase I and Phase II pathways (Fig. 3) [56]. Codeine is demethylated by CYP2D6 to produce morphine, which historically has been viewed to be the major mechanism by which codeine exerts its analgesic effects [57]. Codeine also is conjugated by glucuronyltransferase (more specifically, UDP-glucuronosyltransferase-2B7) to codeine-3-glucuronide and codeine-6-glucuronide, with the 6-glucuronide being active as an analgesic [58]. A minority of a codeine dose is metabolized by CYP2D6, typically accounting for approximately 5% of the dose. Once metabolized to morphine, further metabolism occurs via glucuronidation to morphine-3-glucuronide and morphine 6-glucuronide, with the 6-glucuronide also being pharmacologically active, having roughly half the potency of morphine. The major route of conjugation is typically via 3-glucuronidation, typically eight- to tenfold greater than the production of the 6-glucuronide [56, 58]. Codeine toxicity – with the classical hallmarks of coma, miosis, and bradypnea – has long been recognized as an adverse event

576

Michael J. Rieder and Abdelbaset A. Elzagallaai

associated with codeine overdose [59]. However, in 2006 Koren and colleagues described the case of a breast-fed infant who died of opiate toxicity related to maternal use of codeine for analgesia post episiotomy [60]. In this case, a detailed analysis of blood and breast milk coupled with genetic studies suggested that, despite the mother taking prescribed and conventional doses of a codeineacetaminophen combination product, the infant did indeed have very high concentrations of morphine in the blood, the postmortem blood morphine concentration being 245 nM/L [60]. To put this in context, patients on chronic high-dose opiate therapy have been found to have blood concentrations of 190 nM/L [61]. The reason for this very high morphine concentration became clear when genetic studies were undertaken. The mother was an ultrarapid metabolizer for CYP2D6 and therefore converted much more codeine to morphine than expected, a fact validated in studies of the morphine concentration of her breast milk [60]. In this case the particular genetic variability had been known for some time. CYP2D6 is a polymorphic enzyme, in that it can be demonstrated in three distinct phenotypes – extensive metabolizers (EMs), poor metabolizer (PMs) and ultrarapid metabolizers (UMs) [62]. These phenotypes are the product of the extensive polymorphisms known with respect to the gene encoding CYP2D6. CYP2D6 is highly polymorphic with 1536 variants, of which 390 alleles and sub-alleles have been identified (https://www. pharmvar.org/gene/CYP2D6) [63, 64]. An additional factor complicating this polymorphism is the variable expression of these phenotypes in different populations (Table 1) [65, 66]. Codeine was originally isolated in France in 1832, and the original use of codeine was among northern European populations, among whom the UM polymorphism is uncommon. As the use of codeine has expanded – and codeine has been among the most popular opiates used worldwide – codeine therapy has become more common in Table 1 Ethnic distribution of CYP2D6 phenotypes Population

PM phenotype (%)

UM phenotype (%)

Northern European

7.7–8.9

1

Mediterranean Littoral

2

8

Horn of Africa

2

29

South African

19

Chinese

1

1

Saudi Arabian

1

21

Derived from Bernard et al. [155]

Pharmacogenomics in Children

577

populations with much higher rates of the UM genotype. It is not surprise that, in these populations, problems with codeine toxicity have emerged – not only among newborns but also among children following surgery [67–69]. The recognition of these toxicities led to a brisk response by regulators and hospitals, including a formal contraindication being issued by the US FDA for the use of codeine in children with obstructive sleep apnea after tonsillectomy or adenoidectomy [70]. Many hospitals and practice groups have removed codeine from their formularies – some with careful deliberation and some with changes that appear be less than fully thought out (e.g., replacing codeine with oxycodone – which is O-demethylated by CYP2D6). A common change has been to replace codeine with morphine, which is sensible given that much of the analgesic activity of codeine is probably related to metabolism to morphine [71]. It should be noted that many of these changes occurred within 5 years of the publication of the index case that triggered concerns as to genetically determined toxicity in children [60, 72, 73]. Here we have two drugs both of which have genetically determined toxicities, yet in one case, there has been a profound change in use and regulation within half a decade of the description of the issues of concern, yet in the case of the other drugs, there have been no widespread major changes in use based on genetic testing although more than three decades have passed since potential issues were identified. This probably reflects in part the fact that although codeine was in common use in ambulatory practice, severe outcomes were uncommon, albeit with a high incidence of fatal outcomes [74]. In contrast, while febrile neutropenia during chemotherapy is common fatal, while outcomes are relatively uncommon, they occur under closely monitored conditions, and there are well-defined, evidence-based protocols to guide management [75]. The success of pediatric oncology has largely been based on the use of carefully developed evidence-based protocols, and the requirements needed to change protocols are substantial. In addition, pediatric oncologists are prepared to take risks considerably larger than many other physicians given the severity of the disorders being treated and the fact that outcomes in pediatric cancer, despite the many adverse effects of the treatments prescribed, are in fact the best outcomes for almost any cancers, with approximately 75% of children with cancer reporting a cure of their disease [76]. The very fact that genetically determined variations in TPMT activity influence the concentration of chemotherapeutic agents has been recognized for more than three decades means that these great improvements in outcome have occurred without assessment of TPMT activity beyond the crude but doubtless relevant phenotype of febrile neutropenia. The fact that many of the technical and economic issues that made routine genotyping of children

578

Michael J. Rieder and Abdelbaset A. Elzagallaai

problematic three decades ago have been resolved can then to be seen to be less of a factor in deciding when to routinely use TPMT in management.

3

Pharmacogenomics and Childhood Cancer As noted above, drug use in children is in fact more common than is usually appreciated and tends to be concentrated in approximately 25% of children, who account for 70% of drug use [7, 77]. This would suggest that the most appropriate groups of children for whom pharmacogenomic testing may be of utility are among this 25%, and prominent among these children are children with cancer. Chemotherapy is the mainstay of the treatment of children with cancer and for this indication has been strikingly effective [76, 78, 79]. However, while survival rates for children with cancer have improved dramatically, this has not been without cost. Part of this has been economic, in that cancer treatment involves considerable time in hospital and clinic and substantial use of drugs and laboratory resources [80]. Part of this has been in terms of health-related quality of life [81]. Health-related quality of life among children with cancer is impacted most significantly during therapy and when children develop terminal illness [81]. As well, adverse drug reactions produce a significant burden, including being a very common cause of hospital admission for children with cancer [82]. Among the increasing number of survivors of childhood cancer, healthrelated quality of life is relatively high – except among children who have sustained comorbidities, which are frequently long-term adverse events of therapy [81]. This may be as common as two-third of all survivors, with a large follow-up study demonstrating that a quarter of the adult survivors of childhood cancer had a serious chronic health condition related to their therapy with one quarter having three or more chronic health problems [83]. Thus, adverse drug reactions are a problem not only during therapy but also for many years thereafter [81, 83–85]. Given these troubling numbers, clearly strategies to reduce the risk of adverse drug reactions – and long-term health risk – are important go-forward elements in research in childhood cancer. Genetics has already been incorporated into the care of patients with cancer; an example is the Philadelphia chromosome, a chromosomal anomaly produced as a result of reciprocal translocation between chromosomes 9 and 22 [t(9;22)(q34;q11)] that is most commonly seen as a marker of chronic myeloid leukemia [86]. The presence of this chromosomal anomaly is associated with altered prognosis – and the need for different therapy – when seen in the context of acute lymphoblastic leukemia in childhood [87]. An area of oncology care where genetics has not been widely used has been

Pharmacogenomics in Children

579

in the evaluation – and possibly prevention – of adverse drug reactions [88]. As noted above, adverse drug reactions are very common among children with cancer, and current strategies for prevention and monitoring are largely based on evaluation of clinical phenotypes [89]. This can be effective – an excellent example is the use of Mesna during ifosfamide therapy, which has been quite clearly shown to sharply reduce the risk of bladder toxicity [90]. However, these strategies can be difficult to develop and often require much more mechanistic insights that are currently available. Pharmacogenomic testing offers the potential of being able to identify children at altered risk for adverse events – notably serious adverse events – so that monitoring can be performed on a more regular or consistent manner or novel therapies can be used to reduce adverse drug reaction risk. Given how oncology treatment is undertaken in children, ideally this could be done at the time of diagnosis and initial work-up, at which time considerable effort is put into informing the family and in the implementation of evidence-based treatment protocols. Additionally, the protocols used in oncology are very clear both as to the drugs used and their dose and timing. Thus, pediatric oncology seems to be an area in which pharmacogenomics could make a relatively early entry into routine care. Three drugs commonly used in oncology illustrate the potential. Cisplatin is a platinum-based chemotherapeutic agent that forms complexes which in vivo bind to DNA producing crosslinkages which trigger apoptosis [91]. This drug is a highly effective chemotherapeutic agent used for the therapy of a variety of solid tumors. Although a very useful drug, cisplatin is associated with many adverse events, one of the most serious being ototoxicity [92, 93]. Cisplatin-induced ototoxicity is a feared adverse event associated with therapy, with known risk factors including concurrent therapy with other ototoxic drugs, male gender, and age; children under the age of 5 have 20-fold greater risk than do adults [94, 95]. Our group has described a genetic association with cisplatin-induced ototoxicity, in a cross-Canada study which included a network of 16 pediatric academic health science centers from which cases and matched controls were recruited [96]. We described the association of tag single-nucleotide polymorphisms (SNPs) in the thiopurine S-methyltransferase (TPMT) gene (rs12201199 and rs12201199) and in the catechol-O-methyltransferase (COMT) gene (rs9332377) with cisplatin-induced ototoxicity [97, 98]. This has been validated by other investigators [96]. Modeling the predictive value of these alleles, the presence of three or more risk alleles predicts a very high 5-year risk of not having normal hearing compared to children who have no risk alleles, who have a 60% chance of normal hearing (Fig. 4). While the place of genetic testing for cisplatin-induced ototoxicity is being

Michael J. Rieder and Abdelbaset A. Elzagallaai

70

Chance for normal hearing versus number of risk alleles (%)

580

60 50 40 30 20 10 0 0

1

2

>3

Fig. 4 5-year chance of having normal hearing following cisplatin therapy related to having 0, 1, 2, or 3 or more risk alleles for cisplatin-induced ototoxicity; percentages estimated based on data derived by our group and published in [98]

discussed, economic analysis suggests that there may be a significant health-care saving over the longer term associated with genetic evaluation of hearing risk at the onset of chemotherapy [99, 100]. It is worth commenting on how these SNPs were identified. One might consider that a logical approach would be to develop a study strategy investigating genes regulating pathways known to be key mechanistic elements in the pathogenesis of the disorder of interest. However, in the case of cisplatin-induced ototoxicity, the pathophysiological mechanism(s) remain controversial, and consequently our group elected to genotype patient and control samples for 1949 SNPs which captured genetic variation among 220 genes involved in drug metabolism and distribution (including Phase I and II enzymes, drug transporters, and drug receptors) and disease-specific genes related to physiological pathways impacted by cisplatin. This broad approach was an important aspect of the success of our study in identifying unique SNPs associated with the toxicity of interest as a more selective search may well have missed them. In addition to informing clinicians and patients as to risk, our findings have also provided insights into potential mechanism (s) which are now being evaluated to better define the pathophysiology of cisplatin-induced hearing loss. The mechanism by which variations in TPMT enzyme activity can contribute to cisplatin toxicity has been proposed as due to increased levels of S-adenosylmethionine (SAM), a methyl donor substrate in the methionine pathway [98]. Supporting this hypothesis is the fact that administration of SAM with cisplatin to mice resulted in substantial increase in toxicity (3–6.2-folds) [101].

Pharmacogenomics in Children

581

10 year risk of cardiotoxicity

70 60 50 40 30 20 10 0 Low

Intemediate

High

Fig. 5 10-year chance of developed cardiotoxicity related to being low, intermediate, or high risk; risk category based on number of protective versus risk alleles and developed estimated based on data derived by our group and published in [112]

Anthracyclines antitumor agents including doxorubicin are effective drugs whose mechanisms of action include intercalation between base pairs on the DNA/RNA strand, inhibition of topoisomerase II, generation of iron-mediated free oxygen radicals, and inducing histone eviction from chromatin [102]. The anthracyclines are the cornerstone of therapy for many of the more common cancers in children, including most of the hematoreticular malignancies [103]. Although very effective, anthracyclines are associated with serious adverse effects, the most feared of which is cardiotoxicity [104–107]. Risk factors for anthracycline-induced cardiotoxicity include cumulative dose as well as age, with children under the age of 4 years having a significantly higher risk for cardiotoxicity (Fig. 5) [105, 106]. Anthracyclines use is associated with asymptomatic cardiac dysfunction in 575 of patients and congestive heart failure in 16% of patients [108]. Additional risk factors include female gender, higher dose rates, and cranial irradiation; preventive strategies to date have not produced consistently robust results in terms of efficacy [109]. The mechanism of anthracyclineinduced cardiotoxicity is not well elucidated; however, the major pathological changes that have been observed include vascular degeneration of the sarcoplasmic reticulum, mitochondrial dysfunction, and degeneration of cardiac microfilaments apparently through generation of reactive oxygen species (ROS) [108, 110, 111]. Anthracycline-induced cardiotoxicity is associated with both significant morbidity and mortality, and consequently strategies to identify patients at risk would be of considerable utility [111].

582

Michael J. Rieder and Abdelbaset A. Elzagallaai

Using a similar approach to our studies of cisplatin-induced ototoxicity, we have identified a series of SNPs associated with anthracyline-induced cardiotoxicity in children with cancer in Canada, a finding that we have verified in a replication cohort from the Netherlands [112, 113]. We identified a series of risk and protective alleles that can be related at least in good part to the known pharmacology of the anthracyclines; these variants include protective variants characterized by loss-of-function for influx transporters for anthracyclines as well as risk variants characterized by loss-of-function for efflux transporters for anthracyclines (Table 2) [112]. A GWAS from our group identified variant alleles in the retinoic acid receptor γ (RARγ) gene, which is essential in cardiac development and remodeling, to be significantly associated with anthracycline-induced cardiotoxicity in children [114]. As well variations in genes encoding the solute carrier transporter (SLC) and uridine diphosphateglucuronosyltransferase (UGT) were found to be associated with anthracycline cardiotoxicity [112, 113, 115, 116]. Other recent discoveries regarding the association of the ATP-binding cassette (ABC) transporter family of proteins, which regulate the distribution of many drugs including anthracyclines, have been published [117–120]. A third drug to consider is the nitrogen mustard alkylating agent ifosfamide. Ifosfamide is a positional isomer of cyclophosphamide that is used as an alkylating agent in the treatment of solid tumors [113, 121]. Ifosfamide itself is a pro-drug that must be activated to ifosfamide mustard to produce tumoricidal effects, in the case of ifosfamide mustard by DNA alkylation at the N-7 position of guanine which leads to inter- and intra-strand cross-links causing cell death (Fig. 6) [121]. While very useful for Table 2 Risk and protective variants predicting anthracycline cardiotoxicity Gene

Predictive value

SNP rs-ID

UGT1A6

Risk

rs6759892

ABCB4

Risk

rs1149222

ABCC1

Risk

rs4148350

HNMT

Risk

rs17583889

SCL28A3

Protective

rs78583889

FMO2

Protective

rs2020870

SPG7

Protective

rs2019604

SLC10A2

Protective

rs9514091

SLC28A3

Protective

rs4877847

Derived from [112]

Pharmacogenomics in Children

583

H2 +

Ifosfamide CYP2B6 CYP3A4 CYP2C9 CYP2C19 CYP3A5

2-Dechloroethylifosfamide CYP3A4 CYP3A5 CYP2B6

H

Nephrotoxicity

Chloroacetyldehyde

+

Nephrotoxicity

H

3-Dechloroethylifosfamide Chloroacetyldehyde

OH 4-Hydroxyifosfamide

ALDH

H

H

Detoxification

HO O

O Aldoifosfamide , - elimination

Carboxyifosfamide

+

Bladder Toxicity

Ifosfamide Mustard Acrolein

Antineoplasm

Fig. 6 Metabolism of ifosfamide. Metabolism of the pro-drug ifosfamide produces either the desired chemotherapeutic agent, ifosfamide mustard, or the nephrotoxin chloroacetaldehyde. Potential polymorphisms of interest include CYP 3A5 polymorphisms (low and high expressors) such as CYP 3A5*3 and *6 ¼ absence of or low expression, CYP 3A5 *1 ¼ high expression and ALDH polymorphisms (low and high expressors) such as ALDH2*1 ¼ normal or ALDH2*2 ¼ low function

the management of solid tumors, ifosfamide has been associated with a high risk of nephrotoxicity, with known long-term complications in terms of morbidity and mortality [122–124]. While the risk factors have been debated, it does appear that age under 3 years is a significant risk factor for the development of ifosfamide-induced nephrotoxicity [123]. Approximately 75% of ifosfamide dose is metabolized in the liver by CYP 450 enzymes to 4-hydroxyifosfamide (6-HF) [125, 126]. 6-HF diffuses freely into cells and equilibrates with its tautomer, aldoifosfamide, which non-enzymatically decomposes to produce the active antineoplastic metabolite ifosfamide mustard and the toxic by-product acrolein

584

Michael J. Rieder and Abdelbaset A. Elzagallaai

(Fig. 6). Aldehyde dehydrogenase (ALDH) enzymes convert aldoifosfamide to the non-toxic metabolite carboxyifosfamide. Genetic variations in the ALDH gene family have been found to associate with increased ifosfamide toxicity [127]. The mechanism of ifosfamide-induced nephrotoxicity has been the topic of intense study which has yielded important clues as pathophysiology [90, 128–134]. As noted above, ifosfamide is a pro-drug that must undergo activation to ifosfamide mustard to exert its anticancer effects (Fig. 6). Ifosfamide metabolism can produce either ifosfamide mustard by ring hydroxylation or chloracetaldehyde by side chain oxidation [121, 128, 131]. It has been demonstrated that chloracetaldehyde produced by intra-renal metabolism can act as a potent renal toxin, both in vitro and in vivo [76, 77, 80]. It has also been clearly demonstrated that the use of concurrent antioxidant therapy – in the case of ifosfamide with N-acetylcysteine – can prevent ifosfamide-induced renal injury, again both in vitro and in vivo [128, 132, 133]. This approach has been shown in a small number of case reports to be effective in children with cancer [134]. In this case the metabolism and putative mechanistic basis of the toxicity of interest are known in some depth. How does pharmacogenomics factor in? The proposed course of action – concurrent therapy with N-acetylcysteine – is associated with some risk of adverse effects from N-acetylcysteine and also introduces additional complexity into therapy. Given that the risk for nephrotoxicity in the highest risk group is approximately 30%, this suggests that identifying factors which predict risk – such as genetically determined variation in drug activation – may be of considerable utility in better defining which patients should, and should not, receive concurrent anti-oxidant therapy. There are several steps along the metabolic pathway regulated by enzymes known to be polymorphic, suggesting that pharmacogenomic studies of these pathways could be “low hanging fruit” in defining the contribution of genetics to ifosfamide-induced nephrotoxicity and in helping to better define care that is optimally safe and effective (Fig. 6) [126, 133– 135]. These three examples serve to illustrate the considerable potential that pharmacogenomics offers in improving care in children, even in an area such as pediatric oncology where great strides in improving child health have already taken place.

4

Personalized Medicine for Children The examples cited above with respect to pediatric oncology should not be taken to mean that this is the only care area in which pharmacogenomics for children should be investigated, and indeed there are a number of other areas – notably for children with

Pharmacogenomics in Children

585

complex chronic disorders – where pharmacogenomics offers great promise. There are a number of drugs – ranging from morphine to warfarin to proton pump inhibitors – where there are promising findings suggesting that pharmacogenomics offers the link to develop truly personalized medicine for children [24, 136– 139]. As well, studies need to be conducted to correlate the results of findings in adults with children – for example, in the area of biomarkers for serious adverse events [24, 140]. One area in which pharmacogenomics and personalized medicine offers tremendous promise in bringing clear direction is the area of drugs for central nervous system disorders [141–146]. Over the past two decades, there has been a substantial increase in the use of drugs impacting on the central nervous system in children. This has resulted in a number of questions as to efficacy and safety. Given that in many schools in North America as many as 5% of students are taking some type of psychoactive drug, this strongly suggests that research into the variability of drug response – particularly genetically governed areas of variability – is of considerable importance in developing evidence-based optimal therapy that is both effective and safe. As well, understanding these sources of variability permits the development of guidelines and guides the creation of tools which can be used to direct safer therapy [147]. There are serious questions that need to be asked when considering such studies. One obvious one is which drugs to study? Given that more than 2400 drugs are routinely used in the care of children, which ones should be priority targets for study? Fortunately, this issue has been given some thought, and there are algorithms available which can be used to generate robust solutions when prioritizing as to which drugs would be the most suitable targets for further study [148]. There are practical issues in the conduct of studies, such as ensuring adequate numbers of patients; the creation of large national and even international networks greatly facilitates such studies [149]. While there have been technical and economic issues historically, samples now can be collected relatively noninvasively by using saliva, and the costs of genomic testing have fallen dramatically over the past decade [24, 150].

5

Ethical Issues No consideration of pharmacogenomic testing in children would be complete without a consideration of the ethical issues involved. The ethics of any type of genetic testing in children are complex. When obtaining informed consent, a complication is that the person giving the consent – typically a parent or guardian – is not the person from whom the genetic information is being obtained. This is particularly problematic when this information can be used for

586

Michael J. Rieder and Abdelbaset A. Elzagallaai

risk analysis for adverse events or disease that may occur at some time in the distant future [151, 152]. There have been a number of studies exploring the attitudes of patients and health-care practitioners toward pharmacogenomics and personalized medicine [151, 153, 154]. Patients appear to be reasonably optimistic with respect to the potential for improving outcomes; interestingly, a major reservation has been that patients have been concerned that their own health-care practitioners may not have the right knowledge set to adequately address the new information presented! The literature on attitudes of parents toward genetic testing of their children is much more sparse. We are in the process of studying this area and have found that overall parents – and, interestingly, non-parents – are reasonably comfortable with genetic testing to guide drug dosing children with the proviso that this should be directed and purposeful. That is to say, genetic testing done to direct the results of a therapy planned for now appears to be much more acceptable than genetic testing for less well-defined purposes, although interestingly even if the results of testing predicted outcome – without ability to influence the outcome – parents still wanted to know them. Germane to this is the relative lack of difficulty our group has had in recruiting large numbers of children – patients and controls – for our pharmacogenomic studies [98, 112, 113, 148]. Clearly much work needs to be done, but in general it would be reasonable to conclude that parental attitudes would not be a barrier to well-planned, well-communicated pharmacogenomic research in children. Interestingly, one concern raised by parents was how familiar their children’s health-care providers were with incorporating genetic knowledge into therapeutic decision-making.

6

Pharmacogenomics and Drug Development for Children It is clear that pharmacogenomics is an increasingly important part of the drug development and drug regulation process, not only for adults but for children as well. When developing new therapeutic agents, the potential for genetic variation impacting on efficacy and/or safety should be part of the research considerations. There is a potential risk in that drug development guided by pharmacogenomic testing could potentially restrict entry into clinical trials to those patient sub-groups most likely to benefit and least likely to experience harm. This situation should be avoided, notably as the therapeutic world is truly global and drugs need to be tested among the populations who are likely to use them. This is very germane in the case of children as drug research in children has often been conducted in children of European or African American ancestry, which excludes large groups of children worldwide. Clinical trials can and should take advantage of genetic studies to better define

Pharmacogenomics in Children

587

which patient populations are likely to benefit – or to be at special risk – from the agent under study. Historically children’s therapy has lagged behind therapy for adults – this should not be the case in the era of personalized medicine. References 1. Knox WE (1958) Sir Archibald Garrod’s inborn errors of metabolism. I. Cystinuria. Am J Hum Genet 10:3–32 2. Evans DA, Manley KA, Mc KV (1960) Genetic control of isoniazid metabolism in man. Br Med J 2:485–491 3. Choonara I, Rieder M (2002) Drug toxicity and adverse drug reactions in children – a brief historical review. Paediatr Perinat Drug Therapy 5:12–18 4. Kearns GL, Abdel-Rahman SM, Alander SW et al (2003) Developmental pharmacology-drug disposition, action, and therapy in infants and children. N Engl J Med 349: 1157–1167 5. Avard D, Silverstein T, Sillon G et al (2009) Researchers’ perceptions of the ethical implications of pharmacogenomics research with children. Public Health Genomics 12: 191–201 6. Norbert PW, Roses AD (2003) Pharmacogenetics and pharmacogenomics: recent developments, their clinical relevance and some ethical, social, and legal implications. J Mol Med (Berl) 81:135–140 7. Rieder M, Matsui M, Macleod S (2003) Myths and challenges – drug utilization for Canadian children. Paed Child Health 8:7 8. Castro-Pastrana LI, Carleton BC (2011) Improving pediatric drug safety: need for more efficient clinical translation of pharmacovigilance knowledge. J Popul Ther Clin Pharmacol 18:e76–e88 9. Mason J, Pirmohamed M, Nunn T (2012) Off-label and unlicensed medicine use and adverse drug reactions in children: a narrative review of the literature. Eur J Clin Pharmacol 68:21–28 10. Palmaro A, Bissuel R, Renaud N et al (2015) Off-label prescribing in pediatric outpatients. Pediatrics 135:49–58 11. Pratico AD, Longo L, Mansueto S et al (2018) Off-label use of drugs and adverse drug reactions in pediatric units: a prospective, multicenter study. Curr Drug Saf 13: 200–207

12. Pandolfini C, Bonati M (2005) A literature review on off-label drug use in children. Eur J Pediatr 164:552–558 13. Pandolfini C, Impicciatore P, Provasi D et al (2002) Off-label use of drugs in Italy: a prospective, observational and multicentre study. Acta Paediatr 91:339–347 14. Turner S, Nunn AJ, Fielding K et al (1999) Adverse drug reactions to unlicensed and off-label drugs on paediatric wards: a prospective study. Acta Paediatr 88:965–968 15. Kimland E, Odlind V (2012) Off-label drug use in pediatric patients. Clin Pharmacol Ther 91:796–801 16. Rieder M (2010) If children ruled the pharmaceutical industry: the need for pediatric formulations. Drug News Perspect 23: 458–464 17. Halpern S (1988) American pediatrics: the social dynamic of professionalism 1880–1980. University of california Press, Berkeley 18. Agunod M, Yamaguchi N, Lopez R et al (1969) Correlative study of hydrochloric acid, pepsin, and intrinsic factor secretion in newborns and infants. Am J Dig Dis 14: 400–414 19. de Wildt SN, Kearns GL, Sie SD et al (2003) Pharmacodynamics of intravenous and oral midazolam in preterm infants. Clin Drug Investig 23:27–38 20. Leeder JS, Kearns GL (1997) Pharmacogenetics in pediatrics. Implications for practice. Pediatr Clin N Am 44:55–77 21. Marshall J, Rodarte A, Blumer J et al (2000) Pediatric pharmacodynamics of midazolam oral syrup. Pediatric Pharmacology Research Unit Network. J Clin Pharmacol 40:578–589 22. Marshall JD, Kearns GL (1999) Developmental pharmacodynamics of cyclosporine. Clin Pharmacol Ther 66:66–75 23. Samardzic J, Allegaert K, Bajcetic M (2015) Developmental pharmacology: a moving target. Int J Pharm 492:335–337 24. Hawcutt DB, Thompson B, Smyth RL et al (2013) Paediatric pharmacogenomics: an overview. Arch Dis Child 98:232–237

588

Michael J. Rieder and Abdelbaset A. Elzagallaai

25. Brouwer KL, Aleksunes LM, Brandys B et al (2015) Human ontogeny of drug transporters: review and recommendations of the Pediatric Transporter Working Group. Clin Pharmacol Ther 98:266–287 26. Stevens A, Hanson D, Whatmore A et al (2013) Human growth is associated with distinct patterns of gene expression in evolutionarily conserved networks. BMC Genomics 14: 547 27. Takahashi H, Ishikawa S, Nomoto S et al (2000) Developmental changes in pharmacokinetics and pharmacodynamics of warfarin enantiomers in Japanese children. Clin Pharmacol Ther 68:541–555 28. Weinshilboum RM, Raymond FA, Pazmino PA (1978) Human erythrocyte thiopurine methyltransferase: radiochemical microassay and biochemical properties. Clin Chim Acta 85:323–333 29. Weinshilboum RM, Sladek SL (1980) Mercaptopurine pharmacogenetics: monogenic inheritance of erythrocyte thiopurine methyltransferase activity. Am J Hum Genet 32: 651–662 30. Lennard L, Lilleyman JS, Van Loon J et al (1990) Genetic variation in response to 6-mercaptopurine for childhood acute lymphoblastic leukaemia. Lancet 336:225–229 31. Collie-Duguid ES, Pritchard SC, Powrie RH et al (1999) The frequency and distribution of thiopurine methyltransferase alleles in Caucasian and Asian populations. Pharmacogenetics 9:37–42 32. Relling MV, Gardner EE, Sandborn WJ et al (2013) Clinical pharmacogenetics implementation consortium guidelines for thiopurine methyltransferase genotype and thiopurine dosing: 2013 update. Clin Pharmacol Ther 93:324–325 33. Relling MV, Gardner EE, Sandborn WJ et al (2011) Clinical Pharmacogenetics Implementation Consortium guidelines for thiopurine methyltransferase genotype and thiopurine dosing. Clin Pharmacol Ther 89:387–391 34. Relling MV, Hancock ML, Rivera GK et al (1999) Mercaptopurine therapy intolerance and heterozygosity at the thiopurine S-methyltransferase gene locus. J Natl Cancer Inst 91:2001–2008 35. Relling MV, Schwab M, Whirl-Carrillo M et al (2019) Clinical Pharmacogenetics Implementation Consortium guideline for thiopurine dosing based on TPMT and NUDT15 genotypes: 2018 update. Clin Pharmacol Ther 105:1095–1105

36. Cargnin S, Genazzani AA, Canonico PL et al (2018) Diagnostic accuracy of NUDT15 gene variants for thiopurine-induced leukopenia: a systematic review and meta-analysis. Pharmacol Res 135:102–111 37. Moradveisi B, Muwakkit S, Zamani F et al (2019) ITPA, TPMT, and NUDT15 genetic polymorphisms predict 6-mercaptopurine toxicity in middle eastern children with acute lymphoblastic leukemia. Front Pharmacol 10: 916 38. Moriyama T, Nishii R, Perez-Andreu V et al (2016) NUDT15 polymorphisms alter thiopurine metabolism and hematopoietic toxicity. Nat Genet 48:367–373 39. Moriyama T, Yang YL, Nishii R et al (2017) Novel variants in NUDT15 and thiopurine intolerance in children with acute lymphoblastic leukemia from diverse ancestry. Blood 130:1209–1212 40. Schaeffeler E, Jaeger SU, Klumpp V et al (2019) Impact of NUDT15 genetics on severe thiopurine-related hematotoxicity in patients with European ancestry. Genet Med 21:2145–2150 41. Singh M, Bhatia P, Khera S et al (2017) Emerging role of NUDT15 polymorphisms in 6-mercaptopurine metabolism and dose related toxicity in acute lymphoblastic leukaemia. Leuk Res 62:17–22 42. Yang SK, Hong M, Baek J et al (2014) A common missense variant in NUDT15 confers susceptibility to thiopurine-induced leukopenia. Nat Genet 46:1017–1020 43. Yi ES, Choi YB, Choi R et al (2018) NUDT15 variants cause hematopoietic toxicity with low 6-TGN levels in children with acute lymphoblastic leukemia. Cancer Res Treat 50:872–882 44. Zgheib NK, Akika R, Mahfouz R et al (2017) NUDT15 and TPMT genetic polymorphisms are related to 6-mercaptopurine intolerance in children treated for acute lymphoblastic leukemia at the Children’s Cancer Center of Lebanon. Pediatr Blood Cancer 64:146–150 45. Zhou H, Li L, Yang P et al (2018) Optimal predictor for 6-mercaptopurine intolerance in Chinese children with acute lymphoblastic leukemia: NUDT15, TPMT, or ITPA genetic variants? BMC Cancer 18:516 46. McLeod HL, Krynetski EY, Relling MV et al (2000) Genetic polymorphism of thiopurine methyltransferase and its clinical relevance for childhood acute lymphoblastic leukemia. Leukemia 14:567–572

Pharmacogenomics in Children 47. Bomgaars L, McLeod HL (2005) Pharmacogenetics and pediatric cancer. Cancer J 11: 314–323 48. Bo J, Schroder H, Kristinsson J et al (1999) Possible carcinogenic effect of 6-mercaptopurine on bone marrow stem cells: relation to thiopurine metabolism. Cancer 86:1080–1086 49. Relling MV, Rubnitz JE, Rivera GK et al (1999) High incidence of secondary brain tumours after radiotherapy and antimetabolites. Lancet 354:34–39 50. Adam de Beaumais T, Jacqz-Aigrain E (2012) Pharmacogenetic determinants of mercaptopurine disposition in children with acute lymphoblastic leukemia. Eur J Clin Pharmacol 68:1233–1242 51. Dorababu P, Naushad SM, Linga VG et al (2012) Genetic variants of thiopurine and folate metabolic pathways determine 6-MP-mediated hematological toxicity in childhood ALL. Pharmacogenomics 13: 1001–1008 52. Mazor Y, Koifman E, Elkin H et al (2013) Risk factors for serious adverse effects of thiopurines in patients with Crohn’s disease. Curr Drug Saf 8:181–185 53. Brunton TL (1888) On the use of codeine to relieve pain in abdominal disease. Br Med J 1: 1213–1214 54. Eddy NB, Friebel H, Hahn KJ et al (1969) Codeine and its alternates for pain and cough relief. 5. Discussion and summary. Bull World Health Organ 40:721–730 55. World Health Organization (1998) Cancer pain relief and palliative care in children. World Health Organization, Geneva 56. Desmeules J, Gascon MP, Dayer P et al (1991) Impact of environmental and genetic factors on codeine analgesia. Eur J Clin Pharmacol 41:23–26 57. Sindrup SH, Brosen K (1995) The pharmacogenetics of codeine hypoalgesia. Pharmacogenetics 5:335–346 58. Madadi P, Koren G (2008) Pharmacogenetic insights into codeine analgesia: implications to pediatric codeine use. Pharmacogenomics 9:1267–1284 59. von Muhlendahl KE, Scherf-Rahne B, Krienke EG et al (1976) Codeine intoxication in childhood. Lancet 2:303–305 60. Koren G, Cairns J, Chitayat D et al (2006) Pharmacogenetics of morphine poisoning in a breastfed neonate of a codeine-prescribed mother. Lancet 368:704 61. Goucke RC, Hackett PL, Ilett KF (1994) Concentrations of morphine, morphine-6-

589

glucuronide and morphine-3-glucuronide in serum and cerebrospinal fluid following morphine administration to patients with morphine-resistant pain. Pain 56:145–149 62. Neafsey P, Ginsberg G, Hattis D et al (2009) Genetic polymorphism in cytochrome P450 2D6 (CYP2D6): population distribution of CYP2D6 activity. J Toxicol Environ Health B Crit Rev 12:334–361 63. Crews KR, Gaedigk A, Dunnenberger HM et al (2012) Clinical Pharmacogenetics Implementation Consortium (CPIC) guidelines for codeine therapy in the context of cytochrome P450 2D6 (CYP2D6) genotype. Clin Pharmacol Ther 91:321–326 64. Matic M, Nijenhuis M, Soree B et al (2021) Dutch Pharmacogenetics Working Group (DPWG) guideline for the gene-drug interaction between CYP2D6 and opioids (codeine, tramadol and oxycodone). Eur J Hum Genet. https://doi.org/10.1038/s41431-02100969-9 65. Bernard S, Neville KA, Nguyen AT et al (2006) Interethnic differences in genetic polymorphisms of CYP2D6 in the U.S. population: clinical implications. Oncologist 11:126–135 66. Teh LK, Bertilsson L (2012) Pharmacogenomics of CYP2D6: molecular genetics, interethnic differences and clinical importance. Drug Metab Pharmacokinet 27:55–67 67. Ciszkowski C, Madadi P, Phillips MS et al (2009) Codeine, ultrarapid-metabolism genotype, and postoperative death. N Engl J Med 361:827–828 68. Kelly LE, Rieder M, van den Anker J et al (2012) More codeine fatalities after tonsillectomy in North American children. Pediatrics 129:e1343–e1347 69. Madadi P, Shirazi F, Walter FG et al (2008) Establishing causality of CNS depression in breastfed infants following maternal codeine use. Paediatr Drugs 10:399–404 70. The U.S. Food and Drug Administration (FDA) (2012) Safety review update of codeine use in children; new boxed warning and contraindication on use after tonsillectomy and/or adenoidectomy. Available from: https://www.fda.gov/ media/85072/download. Accessed Nov 2021 71. Poulsen L, Brosen K, Arendt-Nielsen L et al (1996) Codeine and morphine in extensive and poor metabolizers of sparteine: pharmacokinetics, analgesic effect and side effects. Eur J Clin Pharmacol 51:289–295

590

Michael J. Rieder and Abdelbaset A. Elzagallaai

72. Crews KR, Monte AA, Huddart R et al (2021) Clinical Pharmacogenetics implementation consortium guideline for CYP2D6, OPRM1, and COMT genotypes and select opioid therapy. Clin Pharmacol Ther 110: 888–896 73. Pratt VM, Cavallari LH, Del Tredici AL et al (2021) Recommendations for clinical CYP2D6 genotyping allele selection: a joint consensus recommendation of the Association for Molecular Pathology, College of American Pathologists, Dutch Pharmacogenetics Working Group of the Royal Dutch Pharmacists Association, and the European Society for Pharmacogenomics and Personalized Therapy. J Mol Diagn 23:1047–1064 74. Friedrichsdorf SJ, Nugent AP, Strobl AQ (2013) Codeine-associated pediatric deaths despite using recommended dosing guidelines: three case reports. J Opioid Manag 9: 151–155 75. Lehrnbecher T, Phillips R, Alexander S et al (2012) Guideline for the management of fever and neutropenia in children with cancer and/or undergoing hematopoietic stem-cell transplantation. J Clin Oncol 30:4427–4438 76. Rosoff PM (2006) The two-edged sword of curing childhood cancer. N Engl J Med 355: 1522–1523 77. Rieder MJ, Canadian Paediatric Society, Drug Therapy and Hazardous Substances Committee (2011) Drug research and treatment for children in Canada: a challenge. Paediatr Child Health 16:560–561 78. Hudson MM, Neglia JP, Woods WG et al (2012) Lessons from the past: opportunities to improve childhood cancer survivor care through outcomes investigations of historical therapeutic approaches for pediatric hematological malignancies. Pediatr Blood Cancer 58:334–343 79. Turner M, Gagnon D, Lagace M et al (2013) Effect of treatment for paediatric cancers on balance: what do we know? A review of the evidence. Eur J Cancer Care (Engl) 22:3–11 80. Russell HV, Panchal J, Vonville H et al (2013) Economic evaluation of pediatric cancer treatment: a systematic literature review. Pediatrics 131:e273–e287 81. Fakhry H, Goldenberg M, Sayer G et al (2013) Health-related quality of life in childhood cancer. J Dev Behav Pediatr 34: 419–440 82. Mitchell AA, Lacouture PG, Sheehan JE et al (1988) Adverse drug reactions in children leading to hospital admission. Pediatrics 82: 24–29

83. Oeffinger KC, Mertens AC, Sklar CA et al (2006) Chronic health conditions in adult survivors of childhood cancer. N Engl J Med 355:1572–1582 84. Elzagallaai AA, Greff M, Rieder MJ (2017) Adverse drug reactions in children: the double-edged sword of therapeutics. Clin Pharmacol Ther 101:725–735 85. Mlakar V, Huezo-Diaz Curtis P, Satyanarayana Uppugunduri CR et al (2016) Pharmacogenomics in pediatric oncology: review of gene-drug associations for clinical use. Int J Mol Sci 17:1502 86. Jabbour E, Kantarjian H (2012) Chronic myeloid leukemia: 2012 update on diagnosis, monitoring, and management. Am J Hematol 87:1037–1045 87. Bhojwani D, Howard SC, Pui CH (2009) High-risk childhood acute lymphoblastic leukemia. Clin Lymphoma Myeloma 9(Suppl 3): S222–S230 88. Rieder M (2012) New ways to detect adverse drug reactions in pediatrics. Pediatr Clin N Am 59:1071–1092 89. Ross CJ, Visscher H, Rassekh SR et al (2011) Pharmacogenomics of serious adverse drug reactions in pediatric oncology. J Popul Ther Clin Pharmacol 18:e134–e151 90. Chen N, Aleksa K, Woodland C et al (2006) Ontogeny of drug elimination by the human kidney. Pediatr Nephrol 21:160–168 91. Maccio A, Madeddu C (2013) Cisplatin: an old drug with a newfound efficacy – from mechanisms of action to cytotoxicity. Expert Opin Pharmacother 14:1839–1857 92. Brock PR, Knight KR, Freyer DR et al (2012) Platinum-induced ototoxicity in children: a consensus review on mechanisms, predisposition, and protection, including a new International Society of Pediatric Oncology Boston ototoxicity scale. J Clin Oncol 30:2408–2417 93. Rybak LP, Mukherjea D, Jajoo S et al (2009) Cisplatin ototoxicity and protection: clinical and experimental studies. Tohoku J Exp Med 219:177–186 94. Langer T, am Zehnhoff-Dinnesen A, Radtke S et al (2013) Understanding platinuminduced ototoxicity. Trends Pharmacol Sci 34:458–469 95. Ruggiero A, Trombatore G, Triarico S et al (2013) Platinum compounds in children with cancer: toxicity and clinical management. Anticancer Drugs 24:1007–1019 96. Yang JJ, Lim JY, Huang J et al (2013) The role of inherited TPMT and COMT genetic variation in cisplatin-induced ototoxicity in

Pharmacogenomics in Children children with cancer. Clin Pharmacol Ther 94: 252–259 97. Pussegoda K, Ross CJ, Visscher H et al (2013) Replication of TPMT and ABCC3 genetic variants highly associated with cisplatin-induced hearing loss in children. Clin Pharmacol Ther 94:243–251 98. Ross CJ, Katzov-Eckert H, Dube MP et al (2009) Genetic variants in TPMT and COMT are associated with hearing loss in children receiving cisplatin chemotherapy. Nat Genet 41:1345–1349 99. Dionne F, Mitton C, Rassekh R et al (2012) Economic impact of a genetic test for cisplatin-induced ototoxicity. Pharmacogenomics J 12:205–213 100. Zazuli Z, Vijverberg S, Slob E et al (2018) Genetic variations and cisplatin nephrotoxicity: a systematic review. Front Pharmacol 9: 1111 101. Ochoa B, Bobadilla N, Arrellin G et al (2009) S-Adenosyl-L-methionine increases serum BUN and creatinine in cisplatin-treated mice. Arch Med Res 40:54–58 102. Gewirtz DA (1999) A critical evaluation of the mechanisms of action proposed for the antitumor effects of the anthracycline antibiotics adriamycin and daunorubicin. Biochem Pharmacol 57:727–741 103. van Dalen EC, Raphael MF, Caron HN et al (2011) Treatment including anthracyclines versus treatment not including anthracyclines for childhood cancer. Cochrane Database Syst Rev:CD006647 104. de Ville de Goyet M, Moniotte S, Brichard B (2012) Cardiotoxicity of childhood cancer treatment: update and current knowledge on long-term follow-up. Pediatr Hematol Oncol 29:395–414 105. Harake D, Franco VI, Henkel JM et al (2012) Cardiotoxicity in childhood cancer survivors: strategies for prevention and management. Future Cardiol 8:647–670 106. Kucharska W, Negrusz-Kawecka M, Gromkowska M (2012) Cardiotoxicity of oncological treatment in children. Adv Clin Exp Med 21:281–288 107. Von Hoff DD, Layard MW, Basa P et al (1979) Risk factors for doxorubicin-induced congestive heart failure. Ann Intern Med 91: 710–717 108. Geisberg CA, Sawyer DB (2010) Mechanisms of anthracycline cardiotoxicity and strategies to decrease cardiac damage. Curr Hypertens Rep 12:404–410 109. Zerra P, Cochran TR, Franco VI et al (2013) An expert opinion on pharmacologic

591

approaches to reducing the cardiotoxicity of childhood acute lymphoblastic leukemia therapies. Expert Opin Pharmacother 14: 1497–1513 110. Peng X, Chen B, Lim CC et al (2005) The cardiotoxicology of anthracycline chemotherapeutics: translating molecular mechanism into preventative medicine. Mol Interv 5: 163–171 111. Elzagallaai AA, Carleton BC, Rieder MJ (2021) Pharmacogenomics in pediatric oncology: mitigating adverse drug reactions while preserving efficacy. Annu Rev Pharmacol Toxicol 61:679–699 112. Visscher H, Ross CJ, Rassekh SR et al (2012) Pharmacogenomic prediction of anthracycline-induced cardiotoxicity in children. J Clin Oncol 30:1422–1428 113. Visscher H, Ross CJ, Rassekh SR et al (2013) Validation of variants in SLC28A3 and UGT1A6 as genetic markers predictive of anthracycline-induced cardiotoxicity in children. Pediatr Blood Cancer 60:1375–1381 114. Aminkeng F, Bhavsar AP, Visscher H et al (2015) A coding variant in RARG confers susceptibility to anthracycline-induced cardiotoxicity in childhood cancer. Nat Genet 47: 1079–1084 115. Bock KW, Kohle C (2005) UDP-glucuronosyltransferase 1A6: structural, functional, and regulatory aspects. Methods Enzymol 400:57–75 116. Visscher H, Rassekh SR, Sandor GS et al (2015) Genetic variants in SLC22A17 and SLC22A7 are associated with anthracyclineinduced cardiotoxicity in children. Pharmacogenomics 16:1065–1076 117. Reichwagen A, Ziepert M, Kreuz M et al (2015) Association of NADPH oxidase polymorphisms with anthracycline-induced cardiotoxicity in the RICOVER-60 trial of patients with aggressive CD20(+) B-cell lymphoma. Pharmacogenomics 16:361–372 118. Vulsteke C, Pfeil AM, Maggen C et al (2015) Clinical and genetic risk factors for epirubicininduced cardiac toxicity in early breast cancer patients. Breast Cancer Res Treat 152:67–76 119. Wang X, Liu W, Sun CL et al (2014) Hyaluronan synthase 3 variant and anthracyclinerelated cardiomyopathy: a report from the children’s oncology group. J Clin Oncol 32: 647–653 120. Wojnowski L, Kulle B, Schirmer M et al (2005) NAD(P)H oxidase and multidrug resistance protein genetic polymorphisms are associated with doxorubicin-induced cardiotoxicity. Circulation 112:3754–3762

592

Michael J. Rieder and Abdelbaset A. Elzagallaai

121. Dechant KL, Brogden RN, Pilkington T et al (1991) Ifosfamide/mesna. A review of its antineoplastic activity, pharmacokinetic properties and therapeutic efficacy in cancer. Drugs 42:428–467 122. Johnstone EC, Lind MJ, Griffin MJ et al (2000) Ifosfamide metabolism and DNA damage in tumour and peripheral blood lymphocytes of breast cancer patients. Cancer Chemother Pharmacol 46:433–441 123. Loebstein R, Koren G (1998) Ifosfamideinduced nephrotoxicity in children: critical review of predictive risk factors. Pediatrics 101:E8 124. Skinner R, Pearson AD, English MW et al (1996) Risk factors for ifosfamide nephrotoxicity in children. Lancet 348:578–580 125. Chang TK, Yu L, Maurel P et al (1997) Enhanced cyclophosphamide and ifosfamide activation in primary human hepatocyte cultures: response to cytochrome P-450 inducers and autoinduction by oxazaphosphorines. Cancer Res 57:1946–1954 126. Huang Z, Roy P, Waxman DJ (2000) Role of human liver microsomal CYP3A4 and CYP2B6 in catalyzing N-dechloroethylation of cyclophosphamide and ifosfamide. Biochem Pharmacol 59:961–972 127. Fonseca F, Gratacos M, Escaramis G et al (2014) ALDH5A1 variability in opioid dependent patients could influence response to methadone treatment. Eur Neuropsychopharmacol 24:420–424 128. Aleksa K, Ito S, Koren G (2004) Renal-tubule metabolism of ifosfamide to the nephrotoxic chloroacetaldehyde: pharmacokinetic modeling for estimation of intracellular levels. J Lab Clin Med 143:159–162 129. Chen N, Aleksa K, Woodland C et al (2007) The effect of N-acetylcysteine on ifosfamideinduced nephrotoxicity: in vitro studies in renal tubular cells. Transl Res 150:51–57 130. Chen N, Aleksa K, Woodland C et al (2007) Prevention of ifosfamide nephrotoxicity by N-acetylcysteine: clinical pharmacokinetic considerations. Can J Clin Pharmacol 14: e246–e250 131. Chen N, Aleksa K, Woodland C et al (2008) N-Acetylcysteine prevents ifosfamide-induced nephrotoxicity in rats. Br J Pharmacol 153: 1364–1372 132. Dubourg L, Michoudet C, Cochat P et al (2001) Human kidney tubules detoxify chloroacetaldehyde, a presumed nephrotoxic metabolite of ifosfamide. J Am Soc Nephrol 12:1615–1623

133. Hanly L, Chen N, Rieder M et al (2009) Ifosfamide nephrotoxicity in children: a mechanistic base for pharmacological prevention. Expert Opin Drug Saf 8:155–168 134. Hanly LN, Chen N, Aleksa K et al (2012) N-acetylcysteine as a novel prophylactic treatment for ifosfamide-induced nephrotoxicity in children: translational pharmacokinetics. J Clin Pharmacol 52:55–64 135. Chang TK, Weber GF, Crespi CL et al (1993) Differential activation of cyclophosphamide and ifosphamide by cytochromes P-450 2B and 3A in human liver microsomes. Cancer Res 53:5629–5637 136. Downing HJ, Pirmohamed M, Beresford MW et al (2013) Paediatric use of mycophenolate mofetil. Br J Clin Pharmacol 75:45–59 137. Fukuda T, Chidambaran V, Mizuno T et al (2013) OCT1 genetic variants influence the pharmacokinetics of morphine in children. Pharmacogenomics 14:1141–1151 138. Lima JJ, Lang JE, Mougey EB et al (2013) Association of CYP2C19 polymorphisms and lansoprazole-associated respiratory adverse effects in children. J Pediatr 163:686–691 139. Vear SI, Stein CM, Ho RH (2013) Warfarin pharmacogenomics in children. Pediatr Blood Cancer 60:1402–1407 140. Amstutz U, Ross CJ, Castro-Pastrana LI et al (2013) HLA-A 31:01 and HLA-B 15:02 as genetic markers for carbamazepine hypersensitivity in children. Clin Pharmacol Ther 94: 142–149 141. Arango C (2011) Child and adolescent neuropsychopharmacology: now or never. Eur Neuropsychopharmacol 21:563–564 142. Faught E (2012) Antiepileptic drug trials: the view from the clinic. Epileptic Disord 14: 114–123 143. Noam Y, Raol YH, Holmes GL (2013) Searching for new targets for treatment of pediatric epilepsy. Epilepsy Behav 26: 253–260 144. Pidano AE, Honigfeld L (2012) Pediatric psychopharmacology: context, model programs, and considerations for care. Psychiatr Serv 63:929–934 145. Siegel M (2012) Psychopharmacology of autism spectrum disorder: evidence and practice. Child Adolesc Psychiatr Clin N Am 21: 957–973 146. Stroeh O, Trivedi HK (2012) Appropriate and judicious use of psychotropic medications in youth. Child Adolesc Psychiatr Clin N Am 21:703–711 147. Kelly LE, Chaudhry SA, Rieder MJ et al (2013) A clinical tool for reducing central

Pharmacogenomics in Children nervous system depression among neonates exposed to codeine through breast milk. PLoS One 8:e70073 148. Shaw K, Amstutz U, Castro-Pastrana L et al (2013) Pharmacogenomic investigation of adverse drug reactions(ADRs): the ADR prioritization tool, APT. J Popul Ther Clin Pharmacol 20:e110–e127 149. Loo TT, Ross CJ, Sistonen J et al (2010) Pharmacogenomics and active surveillance for serious adverse drug reactions in children. Pharmacogenomics 11:1269–1285 150. Wong WB, Carlson JJ, Thariani R et al (2010) Cost effectiveness of pharmacogenomics: a critical and systematic review. Pharmacoeconomics 28:1001–1013 151. De Marco M, Cykert S, Coad N et al (2010) Views on personalized medicine: do the attitudes of African American and white

593

prescription drug consumers differ? Public Health Genomics 13:276–283 152. Hedgecoe AM (2006) Context, ethics and pharmacogenetics. Stud Hist Philos Biol Biomed Sci 37:566–582 153. Fargher EA, Eddy C, Newman W et al (2007) Patients’ and healthcare professionals’ views on pharmacogenetic testing and its future delivery in the NHS. Pharmacogenomics 8: 1511–1519 154. Rogausch A, Prause D, Schallenberg A et al (2006) Patients’ and physicians’ perspectives on pharmacogenetic testing. Pharmacogenomics 7:49–59 155. Bernard S, Neville KA, Nguyen AT, Flockhart DA (2006) Interethnic differences in genetic polymorphisms of CYP2D6 in the U.S. population: clinical implications. Oncologist 11(2):126–135

Chapter 21 Genetic Ancestry Inference for Pharmacogenomics I. King Jordan, Shivam Sharma, Shashwat Deepali Nagar, Augusto Valderrama-Aguirre, and Leonardo Marin˜o-Ramı´rez Abstract Genetic ancestry inference can be used to stratify patient cohorts and to model pharmacogenomic variation within and between populations. We provide a detailed guide to genetic ancestry inference using genomewide genetic variant datasets, with an emphasis on two widely used techniques: principal components analysis (PCA) and ADMIXTURE analysis. PCA can be used for patient stratification and categorical ancestry inference, whereas ADMIXTURE is used to characterize genetic ancestry as a continuous variable. Visualization methods are critical for the interpretation of genetic ancestry inference methods, and we provide instructions for how the results of PCA and ADMIXTURE can be effectively visualized. Key words Admixture, Genetic ancestry inference, Pharmacogenomics, Health disparities, Genetic variants, Population-specific drug efficacy

1

Introduction Pharmacogenomic variants that mediate patients’ response to medications often show large allele frequency differences among population groups [1, 2]. These allele frequency differences have important implications for treatment decisions, with populationspecific effects observed for drug efficacy, dosage, and toxicity. Indeed, there are numerous examples of racial and ethnic differences in drug response, many of which can be attributed to allele frequency differences in pharmacogenomic variants [3–7]. It has been observed that up to 20% of newly approved drugs show distinct racial and ethnic response profiles, and differences of this kind can lead to group-specific treatment recommendations issued by the FDA [8].

The original version of this chapter was revised. The correction to this chapter is available at https://doi.org/ 10.1007/978-1-0716-2573-6_22 Qing Yan (ed.), Pharmacogenomics in Drug Discovery and Development, Methods in Molecular Biology, vol. 2547, https://doi.org/10.1007/978-1-0716-2573-6_21, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022, Corrected Publication 2023

595

596

I. King Jordan et al.

Nevertheless, it should be stressed that race and ethnicity are socially ascribed characteristics, based on shared origins, culture, heritage, and social experiences. Race and ethnicity are not biological categories and are therefore imprecise proxies for genetic diversity [9]. Genetic ancestry, on the other hand, is a characteristic of the genome. Genetic ancestry measures individuals’ biogeographical origins, based on correlated allele frequency differences among ancestral source populations [10]. Genetic ancestry can be defined independently of the social dimensions of race and ethnicity, and it can be characterized objectively and with precision, as either a categorical or a continuous variable. Accordingly, pharmacogenomic variation among populations is better modeled with genetic ancestry as opposed to race and ethnicity. The aim of this chapter is to provide a practical guide to genetic ancestry inference for pharmacogenetic researchers who may wish to stratify their study cohorts based on patterns of genetic diversity rather than, or in addition to, the more commonly used social categories of race and ethnicity. In light of the increasing availability of large-sale genomic datasets, we focus on genetic ancestry inference methods that make use of genome-wide genetic variant data, including whole-genome sequences, whole exome sequences, and wholegenome genotypes. We provide detailed protocols for two commonly used methods – principal components analysis (PCA) and ADMIXTURE analysis – and we emphasize visualization methods given their importance for large-scale data analysis and interpretation. PCA yields a high-level overview of the patterns of genetic diversity found in a genomic dataset and can be used to delineate genetic ancestry categories [11]. ADMIXTURE can be used to characterize genetic ancestry as a continuous variable, providing fractional estimates of ancestry components for each genomic sample [12].

2

Materials In order to perform genetic ancestry inference, users will need access to (1) a unix/linux operating system, (2) the Conda package manager and environment management system, (3) program installation files, (4) all necessary program dependencies, and (5) appropriately formatted genomic variant data. We provide an overview of the operating system, the package manger, and the genomic data formats that are needed for both of the genetic ancestry inference methods described here. We also provide details on the installation of the R studio package, which can be used to visualize the results of the genetic ancestry inference.

2.1 Operating Systems

Scientific computing, including genetic ancestry inference, is generally conducted in the command line interface provided by unix/ linx operating systems. There are numerous unix/linux operating systems available, many of which are provided free of charge. We

Genetic Ancestry Inference

597

recommend the freely available RedHat or Ubuntu Linux operating systems, and all protocols described here can be successfully executed in one of those operating systems. 1. RedHat https://ubuntu.com/download/desktop 2. Ubuntu https://access.redhat.com/downloads 2.2 Package Manager and Environment Management System

Installation and execution of scientific software packages often requires a specific environment along with a number of dependencies, i.e., other programs or libraries. Thus, environment specification and dependency installation is a rate-limiting step for the use of scientific software, including genetic ancestry inference packages. Conda is a freely available software package and environment management system that allows users to install and update software packages and their dependencies. Use of Conda can save a great deal of time and effort, allowing users to focus on software execution without the need for source code compilation. It should be noted that not all genetic ancestry inference software is made available through Conda, users may have to install and compile source files for some packages, but the tools described here can all be installed from Conda. 1. Conda version 4.9.2 https://repo.anaconda.com/miniconda/

2.3

PLINK

The program PLINK v1.90b6.21 64-bit, which can be used for PCA, is distributed through Conda. 1. https://anaconda.org/sjnewhouse/plink

2.4

ADMIXTURE

The program ADMIXTURE version 1.3.0 is distributed through Conda. 1. https://anaconda.org/bioconda/admixture

2.5 Genomic Variant Data

There are numerous sources of genomic variant data, and users can use their own appropriately formatted data to conduct the genetic ancestry inference analyses described here. The 1000 Genomes Project provides freely available human genomic variant data for samplesToPops.tsv

Genetic Ancestry Inference

599

2. Extract the genomic variant data corresponding to the sample identifiers for the four populations. $ cut -f1 samplesToPops.tsv > sampleIDs.tsv $ plink --vcf phase3.chr22.GRCh38.GT.crossmap.vcf.gz -keep-allele-order --keep-fam sampleIDs.tsv --make-bed -out 1000Genomes.4Pops. Chr22.GRCh38 The last command creates the three PLINK format files, which are needed to run PCA in PLINK: 1000Genomes.4Pops.Chr22.GRCh38.bed 1000Genomes.4Pops.Chr22.GRCh38.bim 1000Genomes.4Pops.Chr22.GRCh38.fam 3.5 Linkage Disequilibrium (LD) Pruning

Perform linkage disequilibrium (LD) pruning to yield a reduced set of unlinked genetic variants. See Note 7. 1. Filter variants with minor allele frequency of 1% and perform LD pruning. $ plink --bfile 1000Genomes.4Pops.Chr22.GRCh38 --keep-alleleorder --maf 0.01 --indep-pairwise 500 5 0.25 --out 1000Genomes.4Pops.Chr22.GRCh38 This command creates the file: 1000Genomes.4Pops.Chr22. GRCh38.prune.in with 20,936 variants. $ plink --bfile 1000Genomes.4Pops.Chr22.GRCh38 --keep-alleleorder --extract 1000Genomes.4Pops.Chr22.GRCh38.prune. in --make-bed --out 1000Genomes.4Pops.Chr22.GRCh38. Pruned This command creates the files: 1000Genomes.4Pops.Chr22.GRCh38.Pruned.bed 1000Genomes.4Pops.Chr22.GRCh38.Pruned.bim 1000Genomes.4Pops.Chr22.GRCh38.Pruned.fam

3.6

PCA Analysis

Run PCA analysis to characterize the genetic relationships among the samples (see Fig. 1). See Note 8. $ plink --bfile 1000Genomes.4Pops.Chr22.GRCh38.Pruned --pca --out 1000Genomes.4Pops.Chr22.GRCh38.Pruned.PCA This command creates the PCA results files, which will be subsequently visualized in R studio: 1000Genomes.4Pops.Chr22.GRCh38.Pruned.PCA.eigenval 1000Genomes.4Pops.Chr22.GRCh38.Pruned.PCA.eigenvec

600

I. King Jordan et al.

0.12

0.08

0.04

PC2

PopGroup YRI IBS PEL CLM

0.00

-0.04

-0.05

0.00

PC1

Fig. 1 Principal components analysis (PCA) of four human populations: Colombia (CLM – green), Peru (PEL – red), Spain (IBS – orange), Yoruba (YRI – blue). The first two principal components (PCs) are shown 3.7 Visualize PCA Results

Visualize PCA results in R studio. 1. Install R packages needed for visualization. install.packages("dplyr") install.packages("ggplot2") install.packages("reshape2") 2. Configure and import the libraries. options(scipen¼100, digits¼3) library(’dplyr’) library(’ggplot2’) 3. Read and process the eigenvector PCA output file and extract the top two PCs. read.table(’1000Genomes.4Pops.Chr22. eigenvec