Lung Biology in Health & Disease Volume 156 Genetic Models in Cardiorespiratory Biology

ISBN: 0-8247-0512-2 This book is printed on acid-free paper. Headquarters Marcel Dekker, Inc. 270 Madison Avenue, New York, NY 10016 tel: 212-696-9000; fax: 212-685-4540 Eastern Hemisphere Distribution Marcel Dekker AG Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland tel: 41-61-261-8482; fax: 41-61-261-8896 World Wide Web http://www.dekker.com The publisher offers discounts on this book when ordered in bulk quantities. For more information, write to Special Sales/Professional Marketing at the headquarters address above. Copyright  2001 by Marcel Dekker, Inc. All Rights Reserved. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the publisher. Current printing (last digit): 10 9 8 7 6 5 4 3 2 1 PRINTED IN THE UNITED STATES OF AMERICA

To my wife Karen, my children Christopher, Diana, and Justin, my late father George and mother Aida, who always gave me support and encouragement GGH To my wife Wan, my children Helen, Heather, and Hannah, my parents Xu Xuejin and Li Jun, and my teachers Spyros Artavanis-Tsakonas and Gerry Rubin TX and to Vincent T. Marchesi and Joseph B. Warshaw, for their unrelenting efforts to bridge science to medicine, foster interaction across disciplines, and bring us together to address scientific questions of clinical importance

INTRODUCTION

During the 20th century, cardiovascular and pulmonary medicine greatly benefited from invasive approaches to diagnosis and therapy. Perhaps the most significant advance was the advent of cardiac catheterization, a procedure that truly revolutionized the understanding of cardiovascular function in health and disease and led the way to spectacular new treatments. Today, all fields of medicine, including cardiovascular and pulmonary medicine, await new developments that will come from molecular and genetic approaches. The complete sequencing of the human genome and the genomes of other species at the beginning of this new century increases the hope for a greater—if not complete—understanding of pathogenesis, as well as the development of definitive therapies. It has been recognized for a long time that diseases are the result of our genetic makeup and its response to environmental factors. Tremendous progress in identifying genes and discovering their function is being heralded as one of the greatest advances in medicine. Undoubtedly, there is strong justification for this optimism. The study of DNA variations has made it possible for us to uncover specific susceptibility to disease, predict the progression or severity of a disease, and, most importantly, identify which drug may work best in specific individuals. But all of this represents the mere beginning of the full exploitation of the genomic era. The task before us is to uncover gene function and to couple it with phenotypic observation. This cannot be accomplished without using models that permit complete molecular dissection of biological and pathological processes, as well as manipulations of the genes that control these processes. That is what this book is about. Genetic Models in Cardiopulmonary Biology, edited by Dr. Gabriel Haddad and Dr. Tian Xu, presents the information that is necessary to bring about the genomic revolution. In this volume, the reader will find descriptions of many models and will be stimulated and challenged by the findings already at hand. The editors have called upon a group of highly respected and innovative scientists to produce a unique and exceptional monograph that brings the Lung Biology in Health and Disease series to a higher level. And most importantly, it v

vi

Introduction

is a contribution that advances medicine and benefits patients affected by cardiopulmonary and related diseases. As the executive editor of this series of monographs, I wish to thank the editors and authors for this major contribution. Claude Lenfant, M.D. Bethesda, Maryland

PREFACE

As is apparent to the scientific community at large, major scientific breakthroughs have flooded the literature in the past two decades, not only because of the phenomenal advances in molecular biology but also because of the use of animal models that have been genetically well characterized. These breakthroughs have been important for two reasons. First, they have defined, at the molecular level, normative cellular and developmental processes. Second, they have shed light on fundamental processes that can become aberrant and thereby lead to disease. Examples abound: embryogenesis, development of organs (e.g., heart, brain, trachea) and the basis for organ defects, cancer, aging, alcohol intoxication, growth and growth defects, and biological clocks—to name a few. Since we believe that the discoveries of the past decade have been very instrumental in moving molecular medicine forward in a quantum fashion and that these breakthroughs have started to play an important role in understanding mechanisms underlying diseases like never before, the basic idea behind this book is rather simple. We had two specific aims when we developed this volume with Dr. Claude Lenfant. The first was to provide the reader with a synopsis of the approaches used in model systems, whether these are in Drosophila melanogaster, Caenorhabditis elegans, the zebrafish, or the mouse. The second was to provide a panoramic view of some of the exciting areas in cardiorespiratory biology that are being studied at the genetic level and that, most likely, will have substantial impact on our understanding of disease and therapy. In order to implement this important exercise, we have assembled a group of highly qualified investigators to contribute to this volume. We have also chosen some very current and exciting areas of research and grouped them into three groups: respiratory development and tissue oxygenation, cardiovascular and red blood cell development, and modulators in cardiorespiratory biology. It is important to mention that we did not intend this volume to be comprehensive from the point of view of addressing all aspects of cardiorespiratory biology. The excitement of today’s possibilities in modern biology and their impact on disease management and therapy cannot be overemphasized. The future holds great promise to solve disease riddles and their pathophysiology. The future will vii

viii

Preface

be incredibly exciting as current medical and graduate students become able to provide the answers to questions that have never been solved before, whether at the bedside or at the bench. These genetic models are contributing to this modern biological revolution, and our aim will be satisfactorily fulfilled with the publication of this volume, in which the contributors paraded some of the major discoveries and breakthroughs using genetic model systems. Gabriel G. Haddad Tian Xu

CONTRIBUTORS

Ravi Allada, M.D. Assistant Professor, Department of Neurobiology, Physiology, and Pathology, Northwestern University, Evanston, Illinois Rolf Bodmer, Ph.D. Associate Professor, Department of Biology, University of Michigan, Ann Arbor, Michigan Clifford W. Bogue, M.D. Assistant Professor of Pediatrics, Section of Critical Care, Department of Pediatrics, Yale University School of Medicine, New Haven, Connecticut Martina Brueckner, M.D. Associate Professor of Pediatrics, Section of Cardiology, Department of Pediatrics, Yale University School of Medicine, New Haven, Connecticut M. Ernest Dodd, Ph.D. Institute of Molecular Medicine and Genetics, Medical College of Georgia, Augusta, Georgia Richard A. Flavell, Ph.D. Chairman, Section of Immunobiology, Yale University School of Medicine, and Investigator, Howard Hughes Medical Institute, New Haven, Connecticut Gabriel G. Haddad, M.D. Professor of Pediatrics and Cellular and Molecular Physiology, and Director, Section of Respiratory Medicine, Department of Pediatrics, Yale University School of Medicine, New Haven, Connecticut Ste´phane Hunot, Ph.D. Postdoctoral Fellow, Section of Immunobiology, Yale University School of Medicine, New Haven, Connecticut Michael R. Koelle, Ph.D. Associate Professor, Department of Molecular Biophysics and Biochemistry, Yale University School of Medicine, New Haven, Connecticut ix

x

Contributors

Alex K.-K. Kuan, Ph.D. Associate Research Scientist, Section of Neurobiology, Yale University School of Medicine, New Haven, Connecticut Shuo Lin, Ph.D. Associate Professor, Institute of Molecular Medicine and Genetics, Medical College of Georgia, Augusta, Georgia Margaret Liu, Ph.D. Research Investigator, Department of Biology, University of Michigan, Ann Arbor, Michigan Wendy K. Lockwood Department of Biology, University of Michigan, Ann Arbor, Michigan Joseph A. Madri, Ph.D., M.D. Professor and Director, Medical Studies, Department of Pathology, Yale University School of Medicine, New Haven, Connecticut Scott D. Marty Institute of Molecular Medicine and Genetics, Medical College of Georgia, Augusta, Georgia Christopher J. Potter Department of Genetics, Yale University School of Medicine and Howard Hughes Medical Institute, New Haven, Connecticut Peter J. Ratcliffe, M.D. Professor of Medicine, Institute of Molecular Medicine, John Radcliffe Hospital, Oxford, England Karen Reue, Ph.D. Associate Professor of Medicine, University of California, Los Angeles, and Veterans Administration Greater Los Angeles Healthcare System, Los Angeles, California Michael Rosbash, Ph.D. Professor, Department of Biology, Brandeis University, and Investigator, Howard Hughes Medical Institute, Waltham, Massachusetts Frank H. Ruddle, Ph.D. Sterling Professor of Biology, Department of Molecular Cell and Developmental Biology, and Professor of Human Genetics, Department of Genetics, Yale University School of Medicine, New Haven, Connecticut Gregg L. Semenza, M.D., Ph.D. Professor, Institute of Genetic Medicine, The Johns Hopkins University School of Medicine, Baltimore, Maryland

Contributors

xi

Cooduvalli S. Shashikant, Ph.D. Associate Professor of Molecular and Developmental Biology, The Pennsylvania State University, University Park, Pennsylvania Didier Y. Stainier, Ph.D. Professor, Department of Biochemistry and Biophysics, University of California, San Francisco, San Francisco, California Ming-Tsan Su Graduate Student, Department of Biology, University of Michigan, Ann Arbor, Michigan Gregory S. Turenchalk Department of Genetics, Yale University School of Medicine and Howard Hughes Medical Institute, New Haven, Connecticut Emily C. Walsh Department of Biochemistry and Biophysics, University of California, San Francisco, San Francisco, California Pablo Wappner, Ph.D. Instituto de Investigaciones Bioquı´micas, Fundacio´n Campomar, Buenos Aires, Argentina Tian Xu, Ph.D. Associate Professor, Department of Genetics, Yale University School of Medicine, and Assistant Investigator, Howard Hughes Medical Institute, New Haven, Connecticut Timothy S. Zheng, Ph.D. Department of Inflammation and Immunology, Biogen Incorporated, Cambridge, Massachusetts

CONTENTS

Introduction (Claude Lenfant) Preface Contributors Part One

v vii ix

OVERVIEW

1. Drosophila as a Genetic Model System for Understanding Human Biology and Disease

1

Gregory S. Turenchalk, Christopher J. Potter, Gabriel G. Haddad, and Tian Xu I. II. III. IV.

Introduction Useful and Unique Techniques Available in Drosophila Example: Use of Drosophila in Cancer Biology Research Conclusions References

1 3 15 16 17

2. Caenorhabditis elegans as a Model for Human Biology and Disease

21

Michael R. Koelle I. Introduction II. Why Use C. elegans to Investigate Biology? III. Is It Posssible to Use C. elegans as a Model for Human Biology and Disease? IV. How Can Gene Knockout Technology in Worms Be Used to Explore the Functions of Human Disease Gene Homologs in C. elegans? V. What Are the Prospects for C. elegans as a Model for Human Biology and Disease?

21 22 24

27 30 xiii

xiv

Contents References

3. Transgenic Mouse Models

31 35

Cooduvalli S. Shashikant and Frank H. Ruddle I. II. III. IV. V. VI. VII. VIII. IX. X. XI. XII.

Introduction Gene Transfer Methods Site of Integration Insert Size Recombinational Cloning and Modification Techniques Cis-Regulatory Analysis Overexpression and Misexpression Inducible Gene Expression Binary Systems Loss of Expression Insertional Mutagenesis Transgenic Mouse Models for the Study of Human Pathologies and Functional Genomics References

Part Two

35 36 39 40 42 42 44 46 47 48 51 51 52

RESPIRATORY DEVELOPMENT AND TISSUE OXYGENATION

4. Genetic Models of Respiratory Tract Development: From Invertebrates to Vertebrates

59

Clifford W. Bogue I. Introduction II. Pharynx and Endoderm Development in Caenorhabditis elegans III. Drosophila Tracheal and Gut Development IV. Mouse Foregut and Lung Development References

59 61 65 68 79

5. Development of Branched Structures and the Cellular Response to Hypoxia: An Evolutionary Perspective

91

Pablo Wappner and Peter J. Ratcliffe I. Genetic Network Controlling Development of Drosophila melanogaster Tracheal System II. Conserved Features of Branching Morphogenesis in Evolution

91 98

Contents

xv

III. Cellular Response to Low Oxygen Tension, Conserved Features IV. Conclusions References

104 122 123

6. Study of Anoxia Tolerance: Use of a Novel Genetic Approach and Animal Model

139

Gabriel G. Haddad I. Introduction II. Heterogeneity in Tolerance to Lack of O2 Between Species and Phyla III. Major Unresolved Questions from Studies on Anoxia-Tolerant Organisms IV. Other Options: Genetic Invertebrate Models V. Conclusions References

139

142 143 147 148

7. Control of Oxygen Homeostasis by Hypoxia-Inducible Factor 1: Essential Roles in Embryogenesis, Physiology, and Disease Pathogenesis

153

140

Gregg L. Semenza I. Introduction II. Hypoxia-Inducible Factor 1 Mediates Transcriptional Activation of Gene Expression in Response to Hypoxia III. Expression of HIF-1: Inducers, Inhibitors, and Signal Transduction Pathways IV. Generation and Analysis of HIF-1α–Deficient Embryonic Stem Cells V. Role of HIF-1 in Embryogenesis VI. Involvement of HIF-1 in the Pathophysiology of Hypoxic Pulmonary Hypertension VII. Involvement of HIF-1 in Ischemic Neovascularization VIII. Involvement of HIF-1 in Retinal Vascularization IX. Involvement of HIF-1 in Cerebral Ischemia X. Involvement of HIF-1 in Ischemic Preconditioning XI. Involvement of HIF-1 in Hypoxia-Mediated Apoptosis XII. Involvement of HIF-1 in Tumor Progression XIII. Conclusions References

153 154 156 160 160 161 164 166 166 167 167 168 169 170

xvi

Contents

Part Three

CARDIOVASCULAR AND RED BLOOD CELL DEVELOPMENT

8. A Genetic Model for Cardiac Pattern Formation and Cell Fate Determination

179

Wendy K. Lockwood, Margaret Liu, Ming-Tsan Su, and Rolf Bodmer I. II. III. IV. V. VI. VII. VIII.

Introduction Cardiac Morphogenesis of Insects and Vertebrates The Cardiogenic Role of tinman and Its Vertebrate Homologs Cardiac Induction: Signals and Context pannier and Vertebrate GATA Factors Cell Type Diversification: Lineages and Context Summary Conclusions References

9. Cardiac Development in Vertebrates

179 180 181 185 191 192 195 195 196 203

Emily C. Walsh and Didier Y. Stainier I. II. III. IV. V. VI. VII. VIII IX. X. XI.

Introduction Heart Morphogenesis Genetic Loci Involved in Cardiac Differentiation Loci Required for Chamber Differentiation Genes Involved in Secondary Convergence Loci Required in Tube Assembly Genes Implicated in Looping Loci Involved in Valve Formation Loci Involved in Cardiac Septation and Further Morphogenesis Loci Involved in Epicardial Development Conclusions References

203 204 213 215 218 223 224 224 227 232 234 235

10. Genetic Defects Resulting in Abnormalities of Vertebrate Left–Right Asymmetry Provide Insight Into the Underlying Developmental Mechanisms

239

Martina Brueckner I. Left–Right Asymmetry Is a Unique Feature of Vertebrate Development

239

Contents

xvii

II. There Is Extensive Anatomical Left–Right Asymmetry in Normal Heart and Lungs III. Pathology Resulting from Abnormal Development of Left– Right Asymmetry IV. Cardiovascular Malformations Are Associated with Heterotaxia V. Development of Left–Right Asymmetry VI. Genetic Defects Affecting Left–Right Development Provide Insight into the Molecular Pathway That Creates Organismal Asymmetry VII. Conclusions References

245 251 252

11. Zebrafish: A Developmental and Genetic Model for Hematopoiesis and Hematopoietic Disorders

255

240 241 241 243

Scott D. Marty, M. Ernest Dodd, and Shuo Lin I. II. III. IV. V.

Introduction to Hematopoiesis Background Information on Zebrafish Techniques Used to Study Zebrafish Development Hematopoietic Disorders in Zebrafish Conclusion References

12. Evolving Paradigms in Vasculogenesis and Angiogenesis

255 260 264 269 275 275 281

Joseph A. Madri I. Introduction II. Targeted Disruption and Overexpression Approaches Used to Study Vasculogenesis and Angiogenesis III. Conclusions References 13. Mouse Genetic Models in Atherosclerosis and Lipid Metabolism Research

281 282 302 302

313

Karen Reue I. II. III. IV.

Introduction Atherosclerosis in the Mouse Tools for Genetic Analysis in the Mouse Induced Mouse Mutants in Atherosclerosis Research

313 315 316 318

xviii

Contents

V. Identification of Novel Genes in Atherosclerosis and Lipid Metabolism References Part Four

330 334

MODULATORS IN CARDIORESPIRATORY BIOLOGY

14. Apoptosis in Development and Disease

341

Timothy S. Zheng, Alex K.-K. Kuan, Ste´phane Hunot, and Richard A. Flavell I. II. III. IV.

Introduction Apoptosis in Vertebrate Development Apoptosis and Human Diseases Conclusions References

15. Fly Clocks

341 348 353 358 359 365

Ravi Allada and Michael Rosbash I. II. III. IV. V. VI. VII. VIII. IX. X.

XI. XII.

Introduction Circadian Rhythms in Human Health and Disease A Mammalian Clock: The Suprachiasmatic Nucleus Circadian Rhythms Are Influenced by Genes: Behavioral Genetics and the period Mutants Circadian Clocks Consist of Circadian Transcriptional Feedback Loops: The period Gene Molecular Mechanisms of Timekeeeping Delays in the Feeedback Loop: The timeless Gene How Do Molecular Clocks Become Synchronized to External Light–Dark Cycles? Did Nature Create Different Clocks for Different Organisms?: The Search for Mammalian Clock Gene Homologs Konopka and Benzer in the Mouse: Mammalian Forward Genetics Nature Does Not Create New Ways of Making Clocks: The Genome Projects and the Cloning of New Evolutionarily Conserved period Genes Getting Cyc’d about Clocks: Flies and Mammals Share Circadian Transcription Factors The Postranslational Clock: The doubletime Kinase

365 367 368 369 370 371 373 374 374

375 376 378

Contents

xix

XIII. A Clockwork Blue: The Blue Light–Sensitive cryptochrome Genes and the Eyes of the Clock XIV. Clocks Are Everywhere! XV. Clocks and Outputs XVI. An Ode to Forward Genetics and Gene Discovery References

379 381 382 383 384

Author Index Subject Index

391 441

Part One OVERVIEW

1 Drosophila as a Genetic Model System for Understanding Human Biology and Disease GREGORY S. TURENCHALK, CHRISTOPHER J. POTTER and TIAN XU

GABRIEL G. HADDAD Yale University School of Medicine New Haven, Connecticut

Yale University School of Medicine and Howard Hughes Medical Institute New Haven, Connecticut

I.

Introduction

The Drosophila model system is an excellent choice for researchers trying to understand human biology and disease processes. The fly has a short life cycle and is easy to culture and observe. The genome of the fly is also relatively simple, and it has been essentially completely sequenced. For any given fly gene, there are typically four homologs in the mouse, and zebrafish usually have an additional duplication. When a gene is knocked out in one of these higher organisms with a more complex genome, it can often be difficult to discern a phenotype. The relative lack of redundancy in the Drosophila genome makes it much easier to associate phenotypes with single mutational events. The genes contained within the Drosophila genome are also highly conserved. It has been estimated that over 80% of human genes have fly homologs. This property makes the fly a useful starting point for discovering important mammalian genes. For example, potassium channels were first cloned in the fly, and this enabled researchers to discover their mammalian counterparts. Homeobox genes were also first identified in the fly, and this discovery led to the identification of similar homeobox genes in other organisms. It is not just individual genes that are conserved in the fly; entire functional pathways are conserved as 1

2

Turenchalk et al.

well. This high level of functional conservation has made the fly an excellent tool for working out the details of pathways that are important to mammalian researchers. The information gained from studies in the fly has been used as a blueprint for working out the intricacies of mammalian pathways. Much of our knowledge of the mammalian Hedgehog, Notch, EGFR/ras, Wnt, and TGFβ pathways has had its genesis in research done with the Drosophila model system [1–12]. Although flies are very divergent from humans in many important ways, it has become apparent that many biological processes and their underlying mechanisms are highly conserved between flies and humans. Many examples of this biological similarity have been well documented. For example, homeobox genes are involved in segmentation in both flies and mammals. The hedgehog gene has a function in limb development in both flies and mammals as well. Also, even though flies have an entirely different type of eye than humans, both organisms utilize homologous genes (eyeless and Pax6) in the early stages of eye development [13]. Even the immune systems of flies and humans have parallels; the involvement of the Toll pathway in innate immunity is conserved between both organisms [14,15]. Studies of fly imaginal discs have shown that basic processes such as cell proliferation and cell fate determination are also conserved between flies and humans. Imaginal discs are sacs of specialized epithelial cells that give rise to most of the structures in the adult fly. These discs are single cell–layer structures that undergo proliferation during larval stages to produce mature discs that have characteristic sizes and shapes [16]. The mature discs then differentiate to produce specific adult structures. The specialized epithelial cells that undergo proliferation and differentiation are diploid and have a cell cycle similar to that of mammalian cells, consisting of G1, S, G2, and M phases [17,18]. The similarity between the fly and mammalian cell cycle is not simply restricted to the general organizational level: the conservation also exists at the molecular level. The fundamental cell cycle machinery, the cyclins (A, B, D, and E types), and their cyclin-dependent kinase partners (Cdk1, Cdk2, Cdk4, or Cdk6) are highly conserved between flies and mammals [17]. The molecular conservation of the cell cycle also extends to cell cycle regulatory components as well. For example, many mammalian cell cycle regulators such as the retinoblastoma protein (pRb) and E2F also have Drosophila homologs (RBF [19] and dE2F [20]). The mechanism of cell fate determination has also been studied in fly imaginal discs. The differentiation of imaginal disc cells to produce the adult structures occurs as a result of communication with surrounding cells through a combination of direct cell–cell interactions and long-range signaling [21,22]. This mosaic type of cell fate specification is very similar to the way cell fate is determined in most mammalian tissues [23]. It is clear that most of the molecular pathways involved

Drosophila as a Genetic Model System

3

in cell fate determination are well conserved between flies and mammals. For example, the Notch receptor has been shown to govern cell fate choice in Drosophila and vertebrates through a similar cell–cell communication mechanism [24]. The similarity of underlying biological processes and the high level of gene conservation between flies and humans make it possible for the fly system to be used to model many types of human diseases. Drosophila research can often lead to the identification of new human disease genes. The discovery that patched (a member of the Hedgehog pathway) was a tumor suppressor, inspired researchers to use the Hedgehog pathway as a blueprint to look for other candidate cancer genes in humans. The identification of the novel tumor suppressor gene lats in the fly [25] led to the discovery of mammalian homologs, and mouse knockout experiments revealed that lats functions as a tumor suppressor in mammals as well as flies [26,27]. The fly system can also be used as a tool for better understanding the biology of an identified human disease gene. The Notch pathway has been implicated in several different diseases in humans, and experiments in the fly system have been instrumental in understanding the biology of the human disorders [28,29]. Many researchers are now exploring the use of Drosophila as a model system for a wide range of important biological questions which were not traditionally studied in the fly. Issues as diverse as response to hypoxia, tissue size control mechanisms, alcohol intoxication, aging, learning and memory, stem cell biology, and congenital heart disease are currently being examined using the fly system.

II. Useful and Unique Techniques Available in Drosophila Aside from the underlying biology which makes the fly such a powerful experimental genetic system, there are technical issues which combine to make Drosophila even more attractive as an organism for modeling human biology and diseases. There are many tools available to Drosophila researchers which are not available to those using other model organisms. The availability of tools and techniques such as P-element transformation, balancer chromosomes, and mosaic screens gives researchers the means to address a wide range of biological questions efficiently and creatively, and one of the strengths of the fly system is how these various methods can be combined synergistically to address research problems in powerful ways. A selection of some of the general techniques available will be discussed below along with examples of how these techniques have actually been combined and utilized by researchers in the laboratory to address cancer-related questions.

4

Turenchalk et al. A. Power of Drosophila as a Genetic Research System

One of the areas in which Drosophila truly shines is in its tractability as a genetic research tool. A wealth of mutant phenotypes have been discovered and characterized in the fly system. These mutations allow researchers to take advantage of multiply marked chromosomes, which is tremendously useful for the genetic mapping of novel mutants. When interesting new recessive mutants are obtained, they can be crossed to special strains with balancer chromosomes so they can be stably maintained as heterozygous stocks. The balancer chromosomes are marked with easily discernible dominant mutations, contain multiple inversions, and are lethal when homozygous. The balancer chromosome provides a wild-type copy of the mutant gene to keep heterozygous animals viable, but the multiple inversions on the balancer chromosomes prevent viable recombinant progeny from developing in the stock. Flies also have large collections of other specialized chromosomes which have features such as duplications or deficiencies, or other unique characteristics. The attached X chromosome, for example, has allowed for the efficient identification of mutations located on the X chromosome through the use of rapid F1 genetic screens. In Drosophila, sex is determined by the ratio of X chromosomes to autosomes, and the Y chromosome is required for male fertility. In an attached-X stock, the female has two X chromosomes which are fused together and are inherited as a unit as well as a Y chromosome, whereas males have the normal complement of a single X chromosome and a single Y chromosome. In every generation, the males in these stocks inherit the X chromosome from their fathers and the Y chromosome from their mothers. F1 screens are carried out by crossing mutagenized males to attached X female virgins. The male progeny of these crosses can be examined directly for mutant phenotypes. Many behavioral screens were carried out in the fly using this approach. B. P-Element–Mediated Gene Transfer

The ability to introduce DNA heritably into an organism is one of the most essential techniques for a genetic model organism. Transgenic flies can be reliably created by using the method of P-element–mediated germline transformation. DNA cloned in a marked P-element vector is injected with a fine needle into the posterior of embryos less than 1 h old. The P-elements insert into the DNA of the germline cells. Transformed flies can be identified in the resultant progeny by the presence of the genetic marker on the P-element. Researchers in the fly community have amassed and mapped large collections of P-element insertions. These P-elements are excellent tools for mutagenesis and cloning. The insertion of transposable elements within or in the vicinity of a gene of interest can cause a mutant phenotype. When a source of transposase is provided, P-elements can be induced to excise from the genome. The excision of the P-element can be either precise or imprecise. If a P-element is precisely

Drosophila as a Genetic Model System

5

excised, the flies can be examined to see if the mutant phenotype is reverted. If this is the case, it proves that the insertion is responsible for the original mutant phenotype. If instead the P-elements excisions are imprecise, they can be used to generate new alleles of a gene. This technique is often used to generate null alleles of a gene of interest. Often excised P-elements will reintegrate into the genome a short distance from the original insertion site in a phenomenon called local hopping. If no mutations are available in a gene of interest, it is often possible to find a line from the P-element collection which has an insertion close to gene of interest. Local hopping of the P-element can then be used to generate insertional mutations in genes neighboring the original insertion site. Genes disrupted by a P-element insertion are also much easier to clone than genes mutated by other methods, because the P-element can be excised imperfectly so that it carries flanking genomic DNA which can be used to jumpstart the cloning process. Since the collection of P-element insertions contains a large number of lethal genes, it is a convenient source of mutations for genetic screens. Such screens are popular because the mutations are already established as lines, and it is more difficult to reveal the molecular nature of genes identified by other mutational methods such as EMS or x-ray mutagenesis. Aside from its use as a mutagenesis and cloning tool, P-element transformation can also be used to introduce cDNA or genomic DNA back into the fly genome for use in rescuing mutant phenotypes as well as for ectopic expression studies and other specialized purposes. An important feature of P-element transformation in flies is that it can quickly and reliably be used to obtain animals with a single copy of the transgene. Current techniques in the mammalian and worm model systems have not yet achieved this benchmark, and the ability to study the effects of a single copy of the transgene can be vital if the gene being examined has dosage-sensitive effects. C. Studying Gene Function by Elevating Gene Expression

1. Ectopic Expression Using Endogenous Promoters

The power of the fly as a genetic model is not just derived from the relative ease with which DNA can be introduced into the genome but from the vast array of choices an investigator has available for controlling the expression of the introduced DNA. The fly system has a wide range of well-characterized promoters that can be used to drive gene expression. These promoters allow genes of interest to be expressed ubiquitously, or they can restrict the gene expression in a spatial or temporal manner. Spatial promoters may be tissue, compartment, or cell type specific. Temporal promoters may be naturally restricted to certain stages of development or they may be inducible. This amazingly versatile array of promoters can be used ectopically to overexpress genes of interest and allow researchers

6

Turenchalk et al.

to assay the biological effects of this overexpression. The diversity of promoters available also makes it likely that the overexpression of a given gene can be examined without causing lethality to the animal. For example, the eyeless promoter can be used to drive ectopic expression exclusively in the eye, allowing the rest of the fly to develop normally, and the heat-shock promoter can be used to limit the ectopic expression temporally allowing issues of early lethality to be avoided. In addition to studying the effects of the overexpression of wild-type genes, researchers can modify genes in vitro and test the effects of ectopically expressing mutated versions of genes. When studying processes such as cancer biology, examining ectopic gene expression can be useful, since oncogenes are often aberrantly activated (i.e., Ras) or overexpressed (i.e., cyclin D). The study of ectopic gene expression in the fly is not restricted to wild-type versions of genes. Researchers may modify genes of interest in vitro to reflect suspected oncogenic mutations such as deletions or point mutations. Such potential oncogenic mutations can then be assayed in vivo. For example, this technique was utilized to confirm that a mutant Ret gene found in MEN2B tumors was in fact hyperactivated [30]. Even if a gene is not an oncogene candidate, ectopic expression can still yield useful information. The overexpression of a gene can overwhelm the pathway that the gene participates in with increased signaling and cause a visible phenotype, thus yielding clues to the function of the gene. This provides a useful entry point for human cancer researchers who have identified a candidate human cancer gene. Despite the evolutionary distance between flies and humans, the expression of a human gene in the fly often has biologically relevant effects, and thus the functions of the human gene can be assayed in the fly without the necessity of first cloning the fly homolog. 2. Ectopic Gene Expression Using the GAL4/UAS System

Further technical innovations utilizing specialized promoters have expanded the ease and versatility of ectopic gene expression. The introduction of the yeast GAL4/UAS system into the fly has enabled researchers to examine the ectopic expression of a gene of interest by generating a single construct (Fig. 1a) [31]. A fly line containing a UAS construct of the gene of interest may subsequently be crossed to any number of existing lines that express GAL4 in specific patterns. Wherever GAL4 is present, it binds the UAS element and triggers the expression of the gene of interest. This allows the ectopic expression of the gene to be assayed rapidly and efficiently. 3. Ectopic Gene Expression Using the FLP-out System

The introduction of the yeast FLP/FRT system into Drosophila is another technical innovation that has greatly expanded the utility of the fly system for research

Figure 1 (a) The Gal4/UAS system can be used to activate the expression of any cloned gene. A tissue specific, constitutive, or heat-shock–inducible promoter drives the expression of GAL4, a yeast transcriptional activator. The GAL4 protein binds the UAS promoter to activate the transcription of the cloned gene. (b) The FLP-out technique can be used to activate the heritable expression of a cloned gene in a subset of tissues. A FLP-out construct contains a constitutive promoter (such as the actin promoter), followed by an FRT-element, a marker gene with a poly A site which serves as a transcriptional terminator, a second FRT-element, and the Gal4 gene. Induction of the FLPase gene leads to site-directed recombination between the FRT-elements and the excision of the DNA between the FRT sites. The constitutive promoter can then activate the transcription of the Gal4 gene.

8

Turenchalk et al.

[32]. FLP recombinase is a catalyst for the site-specific recombination between FLP recombination target (FRT) sites. When the FRT sites are located in cis on a chromosome, the expression of FLP recombinase can result in the excision of intervening DNA between the FRT sites. This ability has led to the development of the FLP-out technique to drive the heritable expression of a gene within a specific subset of tissues (Fig. 1b) [33]. A FLP-out construct consists of a constitutive promoter followed by an FRT site, a marker gene with a poly A site that serves as a transcriptional terminator, a second FRT site, and a cDNA for the gene of interest. A fly line containing this construct may be crossed to another line containing a source of FLPase under an inducible promoter such as the heatshock promoter. The FLPase expression causes the excision of the intervening DNA between the FRT sites within the FLP-out construct, removing the transcriptional terminator and marker gene, and causing the cDNA of the gene of interest to be heritably overexpressed within a random assortment of cells. The FLP-out technique can also be combined with the UAS–GAL4 system to allow even more powerful studies of gene overexpression. If the FLP-out construct is used to drive the expression of GAL4 instead of a single gene of interest, this line can be used to cause the expression of any number of UAS constructs. Therefore, the effects of ectopically expressing several genes at the same time can be carried out. Neufeld et al. have used this technique to determine the effects of cell cycle regulators such as Rbf and E2F on cell size, the cell cycle, and cell division rate [34]. D. Studying Gene Function by Inactivating Gene Expression

Although ectopic expression studies can be very useful, it is possible that overexpression of a gene may not produce a phenotype. It is important to study the effect of inactivating a gene of interest to get an accurate picture of the normal function of the gene. The fly is particularly useful for studying loss of function mutations in genes. The biology and developmental processes of the fly have been studied intensively for years, so there is an excellent context available for helping to decipher the functions of individual genes. Also, the fly genome is smaller and less complex than the human genome, allowing for less redundancy. This makes it more likely that a mutation in a gene of interest will reveal an obvious phenotype. Even when a mutation causes early lethality, the existing knowledge of fly development often provides clues that make it possible to discover the function of the mutated gene. 1. Using Epistasis to Build Genetic Pathways

If the effect of a mutation on development suggests involvement in a particular pathway, or if the mutation has a phenotype similar to that of another known mutant, epistasis experiments can be carried out by double-mutant analysis. In

Drosophila as a Genetic Model System

9

such an analysis, the phenotypes of two animals carrying distinct genetic alterations can be compared to the phenotypes of a third animal carrying both genetic alterations. This type of analysis makes it possible to organize the actions of genes into a genetic pathway. The ease of genetic manipulation in Drosophila, its short life cycle, and the availability of balancer chromosomes makes the fly an excellent organism to use for such experiments. 2. Studying Loss of Function Mutations in Genetic Mosaics

Quite often a mutation in a gene of interest will cause early lethality in homozygous mutant animals, and this may make it difficult study the functional requirement of a gene in a given biological process or tissue. In these cases, it is useful to be able to study these mutations in animals which are mosaics of wild-type, heterozygous, and homozygous mutant cells. The preponderance of tissue in the animal has at least one wild-type copy of the gene of interest, so the animal escapes the early lethality and allows the phenotype of the homozygous mutant clones to be more meaningfully studied. Such mosaic animals can be efficiently generated using chromosomes bearing FRT sites. When FRT sites are located in trans on homologous chromosome, the expression of FLP recombinase causes a high frequency of mitotic recombination [32,35]. When mitotic recombination occurs in an animal which is heterozygous for a mutation, it results in the formation of a homozygous mutant clone along with a homozygous wild-type twin-spot clone in an otherwise healthy heterozygous animal. The construction of a set of chromosomes carrying FRT sites near the centromeres allow this method to be used to produce genetic mosaics for more than 95% of the genes in the fly genome [35]. The use of a cell autonomous marker allows the clones to be distinguished form one another and also allows the direct examination of mutant phenotypes in developing and internal tissues. It is also possible to detect subtle phenotypes, such as change in growth rate, by comparing the sizes of mutant and twin-spot clones. Such a system of mosaic analysis is excellent for examining genes suspected of being tumor suppressors, which do not show a phenotype unless both copies of the gene are mutated. E. Identifying Genes of Interest Using Genetic Screens

One of the major advantages of a model system such as Drosophila is that the entire genome can be systematically screened for mutations of interest if the researchers can identify a suitable phenotype for a screen. The fly has a short life cycle and is easy to culture, allowing a large number of animals to be rapidly screened for mutations effecting a given biological process. A largescale mutagenesis screen is simply not practical for a mammalian model system. A typical mutagenesis screen involves mutagenizing flies and crossing them to a marked

10

Turenchalk et al.

balancer strain (Fig. 2a). The balancer strain prevents recombination on the mutant chromosome and the markers allow animals heterozygous for the new mutation to be identified. Heterozygotes resulting from this cross are segregated and crossed to the marked balancer strain again, allowing the individual mutant lines to be maintained as stocks. Heterozygous animals within a stock can then be crossed to each other to generate animals that are homozygous for the mutation and can be examined for obvious phenotypes. Although the traditional genetic screens have proven to be tremendously useful, there have been recent technical innovations that have led to the development of new types of genetic screens which are far superior to the traditional method of genetic screening. 1. Mosaic Genetic Screens

Carrying out a mosaic screen requires the use of fly strains that carry FRT sites near the centromere on every chromosome arm. Mutagenized males are crossed to females to produce a population of heterozygous embryos which each carry a distinct newly induced mutation (see Fig. 2b). Induction of FLPase expression during development causes a high frequency of mitotic recombination and generates homozygous mutant clones in mosaic animals that can be assayed for phenotypes. Mosaic screens have major advantages over traditional genetic screens. One important advantage is that mosaic screens are much more efficient than traditional screens. Traditional screens require at least three generations, and individual lines must be established for each mutagenized progeny (see Fig. 2a). These lines must be maintained until they can be assayed for phenotypes. Mosaic screens are performed in the first generation which allows screenFigure 2 A comparison of a conventional screen versus an F1 mosaic screen in the identification of tumor suppressor genes. (a) A conventional screen requires three generations (F3) before a mutant phenotype can be analyzed. Often the animal will die before a cancer-related phenotype can be detected. Approximately 30 days are required for the detection of a possible mutant phenotype. (b) An F1 mosaic screen requires only one generation (F1) for the detection of a mutant phenotype. The parental generation contains FRT-elements located proximal to a centromere. During development of the F1 generation, the FLPase gene is induced. The FLPase protein increases the frequency of mitotic recombination between homologous chromosomes, and can lead to the formation of a ⫺/⫺ mutant cell and a ⫹/⫹ wild-type cell. The mutant cell can be distinguished from the parental and twin-spot cell by the lack of marker gene expression. Such mitotic recombination can result in large clones of mutant cells(⫺/⫺) in a heterozygous (⫺/⫹) wild-type background. An F1 mosaic screen requires approximately 10 days for the identification of a mutant phenotype. Indeed, mosaic screens have been successful in identifying novel Drosophila tumor suppressor genes.

Drosophila as a Genetic Model System

11

12

Turenchalk et al.

ing for phenotypes in F1 individuals before establishing lines (see Fig. 2b). Only individuals carrying mutations of interest are kept, resulting in a dramatic increase in the efficiency of the screen. Another advantage of the mosaic screen is that it allows many mutations to be identified which would be missed by a traditional screen owing to early lethality. Many genes are required in multiple tissues and at multiple developmental stages, and mutations in these genes often cause early lethality in homozygous animals. Screens for mutations effecting late developmental stages or specific tissues are often unable to be recovered in a traditional screen. Mosaic screens allow the phenotypes of such lethal mutations to be observed in clonal patches of mutant cells in an otherwise healthy animal. The mosaic screen is a powerful screening technique which is unique to the fly model system. In practice, this type of screen has already been used to isolate important genes that have been missed in traditional genetic screens [25,36,37]. 2. Ectopic Activation Screens

If a loss-of-function phenotype for a gene is not informative or an alternative way of examining the function of a gene is desired, another type of screen, the ectopic activation screen, can be utilized (Fig. 3) [38]. This type of screen makes use of a modular misexpression system and allows systematic gain-of-function screening. In an ectopic activation screen, mobilizing a P-element that contains the UAS enhancer and a promoter (an EP element) generates target lines. This causes random insertions into the fly genome where the EP-elements are capable of influencing the transcription of sequences flanking the insertions. Since Pelements preferentially insert into the 5′ ends of genes, it makes it possible for the EP-elements to drive expression of essentially full-length transcripts from the gene downstream of the element insertion sites. The phenotypes caused by these insertions in the target lines are screened for by crossing the flies to other lines that express GAL4 in specific patterns. The GAL4 expression specifically activates the UAS enhancer causing the target line to misexpress the gene downstream of its EP insertion. The phenotypes observed in the target lines can be due to either expressing a gene at the wrong place or time, or they may be due to overexpressing the gene. If the system is used to screen for a tumor-formation phenotype, it should be useful in identifying oncogenes. 3. Modifier Screens

Once an interesting phenotype has been obtained by manipulations which either elevate or inactivate the expression of a gene of interest, another type of screen becomes available, the modifier screen. In a modifier screen, interacting genes are identified by searching for second site mutations that suppress or enhance

Drosophila as a Genetic Model System

13

Figure 3 An ectopic activation screen can identify genes whose overexpression might lead to tumor formation. A UAS promoter can be inserted randomly into the Drosophila genome by P-element–mediated transformation. Induction of the Gal4 protein leads to the ectopic expression of the trapped gene, which could result in an oncogenic phenotype.

the original phenotype (Fig. 4). These modifier screens are extremely powerful tools for helping to discover the components of pathways. Modifier screens have been successfully used for loss-of-function phenotypes as well as for gain of function phenotypes. An example of a loss-of-function modifier screen is the one that looked for Notch modifiers. A modifier screen for mutations that could suppress the lethality of Notch mutants yielded mutants in the Delta locus [39]. An example of a gain-of-function screen is the screen for modifiers of the Ras overexpression phenotype in the eye that was used to identify many of the components of the Ras pathway [40].

14

Turenchalk et al.

Figure 4 An example of a modifier screen that can be used to identify additional components of genetic pathways. Expression of the lats tumor suppressor under the GMR promoter (which drives gene expression in the differentiating cells of the developing Drosophila eye) leads to viable flies that exhibit a small eye phenotype. GMR-lats flies can then be crossed to mutagenized flies. In the resulting generation, the effect of the mutant gene on the GMR-lats small eye phenotype can be easily assayed by a change in eye size. This can rapidly lead to the identification of genes that function to activate or inhibit lats signaling.

Drosophila as a Genetic Model System

15

III. Example: Use of Drosophila in Cancer Biology Research Researchers seeking ways to discover the mechanisms of cancer development should not overlook the powerful and innovative techniques available in the fruit fly as a model organism. Since flies have a short life cycle, it is not intuitively obvious that they would have to contend with a phenomenon like tumor development, which is frequently caused by an accumulation of mutations that cause inappropriate inactivation or activation of genes. Although cancer may not be a major concern for flies in the wild, it turns out that much of the underlying molecular machinery involved in human cancer processes are highly conserved in the fly. In the late 1970s, the fly began as a cancer research organism when spontaneous mutations were identified that caused animals to die during larval stages with overproliferated internal tissues [41,42]. The discovery of this overproliferation phenotype allowed subsequent screens to be designed that were used successfully to recover dozens of genetic loci [41,43–45]. Many of these loci behaved as recessive loss-of-function mutations and were defined as tumor suppressors [46]. The identification of Drosophila tumor suppressors was promising, but characterization of these genes did not provide an obvious link to processes such as regulation of the cell cycle, which were part of the contemporary understanding of the mechanisms involved in tumor formation. These studies did, however, point to the importance of cell–cell communication in the regulation of cell proliferation [42,47,48]. The fly tumor suppressors also failed to show similarity to tumor suppressors that had been identified in humans up to that point in time [49,50]. These initial qualms about the similarity of tumor formation between flies and humans did not deter continued efforts in researching cancer in Drosophila. A. Fly and Human Homologs

Although the fly has approximately 10-fold fewer genes than humans, it turns out that a significant number of fly genes are homologs of human oncogenes and tumor suppressors. As a matter of fact there are more than 76 fly homologs of mammalian cancer genes which are currently under intensive investigation including homologs of human oncogenes such as c-src, E2F, RAS, and ret, and human tumor suppressor genes such as p53, PTEN, patched, ATM, NF1, NF2, and pRB [51]. Studies of fly homologs of human genes have contributed toward the understanding of the developmental functions of these genes as well as their actions at a molecular level, and they have allowed the genes to be placed in the context of genetic pathways. The study of cancer-related genes in a model organism like Drosophila, which is more readily tractable to experimentation than humans, provides an excellent opportunity to expand our understanding of the actions of the human counterparts of these genes.

16

Turenchalk et al. B. Relevance of the Fly Model to Human Cancer

Although fly homologs exist for numerous human cancer genes, mutations in the fly versions of the genes do not always lead to tumor formation. This raises the question of whether flies provide a useful model for mammalian cancer-related processes. This question has been answered by recent studies that indicate flies are capable of forming real tumors and exhibiting a full range of cancer-like behaviors including metastasis [25,52,53]. For example, research on the Drosophila tumor suppressor gene lats shows that somatic cells mutant for lats undergo extensive proliferation to form large tumor outgrowths with morphological characteristics very much like those of human tumors, demonstrating that flies can indeed grow tumors comparable to those found in human patients [25]. This finding is bolstered by studies of the mouse homolog of the gene Lats1. Mice deficient for Lats1 develop soft tissue sarcomas and ovarian stromal cell tumors and exhibit a heightened sensitivity to carcinogenic treatments [26]. This demonstration that a previously uncharacterized tumor suppressor gene discovered in the fly can also act as a tumor suppressor in the mouse indicates that studies of cancer genes in flies can be directly relevant to mammalian tumorigenesis. Further support for the relevance of using Drosophila as a model for human cancer processes comes from experiments with tumors derived from Drosophila lgl mutants [52,53]. These experiments clearly demonstrated that tumor metastasis can occur in the fly. Flies homozygous for lgl mutations die during the larval stage, but brain tumors transplanted from these mutants into normal adult flies exhibit metastasis by spreading into distant organs [52]. Although lgl homologs have not yet been shown to be involved in human tumor metastasis, the lgl metastatic cells exhibit an increase in type IV collagenase, which indicates that at least some of the biochemical mechanisms of metastasis are conserved between flies and humans [53,54]. This conservation is further supported by the discovery that the human gene nm23, which has been shown to be involved in metastasis, has a homolog in Drosophila called abnormal wing discs [55,56]. Mutations in awd can cause abnormal tissue morphology and widespread aberrant differentiation that is analogous to changes that occur in human malignant progression. IV. Conclusions In recent years, the development of powerful genetic techniques have provided Drosophila researchers with the opportunity to rapidly identify and characterize genes involved in human disease and development. The general high level of gene and pathway conservation, the similarity of the cellular processes, and evidence of the functional conservation of specific disease-related genes between

Drosophila as a Genetic Model System

17

flies and mammals argue that studies in flies can directly contribute to the understanding of human biology and disease. Now that the fly genome sequence has been completed, fly homologs for most of the known human disease genes have already been discovered, and many more will become evident as novel human disease genes are identified. Researchers accustomed to working with the mammalian system may find that the use of the fly model system can be a complement to their current studies and give them insight on aspects of human disease and biology that are not easily addressable in the mammalian system. The methods discussed are only a sampling of the techniques available to researchers using the Drosophila model system. These methods are constantly evolving; they are refined and combined in creative ways as investigators are forced to find new ways to answer the questions posed by their work. The Drosophila model has become one of the most powerful systems for not only addressing developmental and genetics questions but also for studying disease related processes.

References 1. Bray S. Notch signalling in Drosophila: three ways to use a pathway. Semin Cell Dev Biol 1998; 9:591–597. 2. Casci T, Freeman M. Control of EGF receptor signalling: lessons from fruitflies. Cancer Metastasis Rev 1999; 18:181–201. 3. Freeman M. Complexity of EGF receptor signalling revealed in Drosophila. Curr Opin Genet Dev 1998; 8:407–411. 4. Hammerschmidt M, Brook A, McMahon AP. The world according to hedgehog. Trends Genet 1997; 13:14–21. 5. Joutel A, Tournier-Lasserve E. Notch signalling pathway and human diseases. Semin Cell Dev Biol 1998; 9:619–625. 6. Murone M, Rosenthal A, de Sauvage FJ. Hedgehog signal transduction: from flies to vertebrates. Exp Cell Res 1999; 253:25–33. 7. Padgett RW, Das P, Krishna S. TGF-beta signaling, Smads, and tumor suppressors. Bioessays 1998; 20:382–390. 8. Perrimon N. Hedgehog and beyond. Cell 1995; 80:517–520. 9. Raftery LA, Sutherland DJ. TGF-beta family signal transduction in Drosophila development: from Mad to Smads. Dev Biol 1999; 210:251–268. 10. Siegfried E, Perrimon N. Drosophila wingless: a paradigm for the function and mechanism of Wnt signaling. Bioessays 1994; 16:395–404. 11. Wodarz A, Nusse R. Mechanisms of Wnt signaling in development. Annu Rev Cell Dev Biol 1998; 14:59–88. 12. Young MW, Wesley CS. Diverse roles for the Notch receptor in the development of D. melanogaster. Perspect Dev Neurobiol 1997; 4:345–355. 13. Wawersik S, Mass RL. Vertebrate eye development as modeled in Drosophila. Hum Mol Genet 2000; 9:917–925.

18

Turenchalk et al.

14. Dushay MS, Eldon ED. Drosophila immune responses as models for human immunity. Am J Hum Genet 1998; 62:10–14. 15. Kopp EB, Medzhitov R. The Toll-receptor family and control of innate immunity. Curr Opin Immunol 1999; 11:13–18. 16. Bryant PJ, Schmidt O. The genetic control of cell proliferation in Drosophila imaginal discs. J Cell Sci 1990; 13(suppl):169–189. 17. Edgar BA, Lehner CF. Developmental control of cell cycle regulators: a fly’s perspective. Science 1996; 274:1646–1652. 18. Orr-Weaver TL. Developmental modification of the Drosophila cell cycle. Trends Genet 1994; 10:321–327. 19. Du W, Vidal M, Xie J, Dyson N. RBF, a novel RB-related gene that regulates E2F activity and interacts with cyclin E in Drosophila. Genes Dev 1996; 10:1206– 1218. 20. Dynlacht BD, Brook A, Dembski M, Yenush L, Dyson N. DNA-binding and transactivation properties of Drosophila E2F and DP proteins. Proc Natl Acad Sci USA 1994; 91:6359–6363. 21. Basler K, Hafen E. Specification of cell fate in the developing eye of Drosophila. Bioessays 1991; 13:621–631. 22. Rooke JE, Xu T. Positive and negative signals between interacting cells for establishing neural fate. Bioessays 1998; 20:209–214. 23. Davidson EH. How embryos work: a comparative view of diverse modes of cell fate specification. Development 1990; 108:365–389. 24. Artavanis-Tsakonas S, Matsuno K, Fortini ME. Notch signaling. Science 1995; 268: 225–232. 25. Xu T, Wang, W, Zhang, S, Stewart, RA, Yu W. Identifying tumor suppressors in genetic mosaics: the Drosophila lats gene encodes a putative protein kinase. Development 1995; 121:1053–1063. 26. St. John MA, Tao W, Fei X, Fukumoto R, Carcangiu ML, Brownstein DG, Parlow AF, McGrath J, Xu T. Mice deficient of Lats1 develop soft-tissue sarcomas, ovarian tumours and pituitary dysfunction. Nat Genet 1999; 21:182–186. 27. Tao W, Zhang S, Turenchalk GS, Stewart RA, St John MA, Chen W, Xu T. Human homologue of the Drosophila melanogaster lats tumour suppressor modulates CDC2 activity. Nat Genet. 1999; 21:177–181. 28. Ellisen LW, Bird J, West DC, Soreng AL, Reynolds TC, Smith SD, Sklar J. TAN1, the human homolog of the Drosophila notch gene, is broken by chromosomal translocations in T lymphoblastic neoplasms. Cell 1991; 66:649–661. 29. Go MJ, Eastman DS, Artavanis-Tsakonas S. Cell proliferation control by Notch signaling in Drosophila development. Development 1998; 125:2031–2040. 30. Read RD, Cagan RL. Identifying the RetMEN2B Signaling Pathway: a Drosophila transgenic approach. Am Conf Drosaphila Res 1998; 39:163A. 31. Brand AH, Perrimon N. Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 1993; 118:401–415. 32. Golic KG, Lindquist S. The FLP recombinase of yeast catalyzes site-specific recombination in the Drosophila genome. Cell 1989; 59:499–509. 33. Struhl G, Basler K. Organizing activity of wingless protein in Drosophila. Cell 1993; 72:527–540.

Drosophila as a Genetic Model System

19

34. Neufeld TP, de la Cruz AF, Johnston LA, Edgar BA. Coordination of growth and cell division in the Drosophila wing. Cell 1998; 93:1183–1193. 35. Xu T, Rubin GM. Analysis of genetic mosaics in developing and adult Drosophila tissues. Development 1993; 117:1223–1237. 36. Rooke J, Pan D, Xu T, Rubin GM. KUZ, a conserved metalloprotease-disintegrin protein with two roles in Drosophila neurogenesis. Science 1996; 273:1227–1231. 37. Theodosiou NA, Zhang S, Wang WY, Xu T. slimb coordinates wg and dpp expression in the dorsal-ventral and anterior-posterior axes during limb development. Development 1998; 125:3411–3416. 38. Rorth P, Szabo K, Bailey A, Laverty T, Rehm J, Rubin GM, Weigmann K, Milan M, Benes V, Ansorge W, Cohen SM. Systematic gain-of-function genetics in Drosophila. Development 1998; 125:1049–1057. 39. Xu T, Rebay I, Fleming RJ, Scottgale TN, Artavanis-Tsakonas S. The Notch locus and the genetic circuitry involved in early Drosophila neurogenesis. Genes Dev 1990; 4:464–475. 40. Karim FD, Chang HC, Therrien M, Wassarman DA, Laverty T, Rubin GM. A screen for genes that function downstream of Ras1 during Drosophila eye development. Genetics 1996; 143:315–329. 41. Gateff E. Malignant neoplasms of genetic origin in Drosophila melanogaster. Science 1978; 200:1448–1459. 42. Mechler BM, McGinnis W, Gehring WJ. Molecular cloning of lethal(2)giant larvae, a recessive oncogene of Drosophila melanogaster. EMBO J 1985; 4:1551–1557. 43. Bryant PJ, Watson KL, Justice RW, Woods DF. Tumor suppressor genes encoding proteins required for cell interactions and signal transduction in Drosophila. Development 1993; (suppl):239–249. 44. Gateff E. Cancer, genes, and development: the Drosophila case. Adv Cancer Res 1982; 37:33–74. 45. Torok T, Tick G, Alvarado M, Kiss I. P-lacW insertional mutagenesis on the second chromosome of Drosophila melanogaster: isolation of lethals with different overgrowth phenotypes. Genetics 1993; 135:71–80. 46. Woods DF, Bryant PJ. Molecular cloning of the lethal(1)discs large-1 oncogene of Drosophila. Dev Biol 1989; 134:222–235. 47. Strand D, Raska I, Mechler BM. The Drosophila lethal(2)giant larvae tumor suppressor protein is a component of the cytoskeleton. J Cell Biol 1994; 127:1345–1360. 48. Woods DF, Hough C, Peel D, Callaini G, Bryant PJ. Dlg protein is required for junction structure, cell polarity, and proliferation control in Drosophila epithelia. J Cell Biol 1996; 134:1469–1482. 49. Knudson AG. Antioncogenes and human cancer. Proc Natl Acad Sci USA 1993; 90:10914–10921. 50. Pitot HC. The molecular biology of carcinogenesis. Cancer 1993; 72:962–970. 51. Potter CJ, Turenchalk GS, Xu T. Drosophila in cancer research. An expanding role. Trends Genet 2000; 16:33–39. 52. Woodhouse E, Hersperger E, Shearn A. Growth, metastasis, and invasiveness of Drosophila tumors caused by mutations in specific tumor suppressor genes. Dev Genes Evol 1998; 207:542–550. 53. Woodhouse E, Hersperger E, Stetler-Stevenson WG, Liotta LA, Shearn A. Increased

20

Turenchalk et al.

type IV collagenase in lgl-induced invasive tumors of Drosophila. Cell Growth Differ 1994; 5:151–159. 54. Ambiru S, Miyazaki M, Nakagawa K, Nakajima N. Increased serum type IV collagen 7-S levels in patients with hepatic metastasis. Am J Gastroenterol 1995; 90: 783–787. 55. Biggs J, Hersperger E, Steeg PS, Liotta LA, Shearn A. A Drosophila gene that is homologous to a mammalian gene associated with tumor metastasis codes for a nucleoside diphosphate kinase. Cell 1990; 63:933–940. 56. Freije JM, MacDonald NJ, Steeg PS. Nm23 and tumour metastasis: basic and translational advances. Biochem Soc Symp 1998; 63:261–271.

2 Caenorhabditis elegans as a Model for Human Biology and Disease

MICHAEL R. KOELLE Yale University School of Medicine New Haven, Connecticut

I.

Introduction

Advances in human genetics and genomics have made it increasingly easy and rapid to identify genes involved in human diseases. Upon their identification, however, it often remains obscure how such genes function in normal biology and how defects in such genes lead to disease. About three-quarters of human genes have homologs in Caenorhabditis elegans, and recent technical advances in C. elegans gene knockout technology make it straightforward to delete any gene from the C. elegans genome. The power of C. elegans genetics and the detailed understanding of the biology of this organism can thus be used to rapidly leverage identification of a human disease gene into a detailed understanding of its homolog’s function in C. elegans. Recent examples of this approach show that biological processes are highly enough conserved from worms to humans that the understanding thus achieved in the worm can often yield deep insights into the functions of the original disease gene in humans. In this chapter, I discuss the prospects for using this approach as a general strategy to investigate gene function in human biology and disease. 21

22

Koelle II. Why Use C. elegans to Investigate Biology?

For those unfamiliar with C. elegans, I begin with a brief review of the experimental advantages of the worm that have led to its status as a leading genetic model organism. In the 1960s, Brenner chose C. elegans as new a model organism for studies of development and the nervous system (Brenner, 1974). At that time, decades of experience with the advantages and limitations of genetics in Drosophila were already available. Brenner could thus understand, with an unprecedented level of sophistication, the technical issues that would allow him to choose an animal virtually ideally suited to both genetic and anatomical studies. A few of its advantages are outlined below. A. Powerful Genetics

C. elegans adults are ⬃1 mm long worms (Fig. 1) that are cheap and easy to culture (see Wood et al., 1988, and Riddle et al., 1997, for reviews of C. elegans biology and genetics). These worms eat Escherichia coli bacteria, and they can be grown on Petri dishes, in large liquid cultures, or for high-throughput studies, in microtiter wells. A generation takes only 3 days, and each animal has ⬃300 progeny. This speed and fecundity greatly facilitate genetic studies. C. elegans are typically hermaphrodites that reproduce when an individual worm fertilizes itself using both sperm and eggs it has produced. Such self-fertilization is a tremendous advantage in carrying out genetic screens: It automates the otherwise laborious process of inbreeding that is required to reveal the defects caused by recessive mutations. A typical genetic screen in C. elegans consists of treating worms with a mutagen, simply waiting 6 days for two generations of automatic inbreeding, and then examining the resulting animals for mutant defects. C. elegans can be frozen alive, and a typical worm genetics laboratory has hundreds or thousands of frozen mutant strains in storage ready for use within days of thawing. B. Simple and Well-defined Anatomy

C. elegans anatomy is understood at a level of detail unmatched for any other animal. There are precisely 959 somatic cells, each of which can be identified by its morphology and position in the animal. The pattern of cell divisions starting from the zygote that gives rise to all 959 somatic cells has been described, and is essentially invariant from animal to animal (Sulston and Horvitz, 1977; Sulston et al., 1983). The adult anatomy includes a muscular pharynx used to eat, an intestine, a gonad, body wall muscles for movement, an excretory system, and a nervous system. There are precisely 302 neurons, and the ⬃7000 synaptic connections made by these neurons have been described by examining serial section

C. elegans as a Model for Human Biology and Disease

23

Figure 1 C. elegans adult hermaphrodites are ⬃1 mm long worms. The head of the animal shown is to the left.

electron micrographs of the entire animal (White et al., 1986). C. elegans is the only animal for which such a wiring diagram of the complete nervous system has been determined. Because C. elegans is transparent, it is possible (with proper microscope optics) to see the nucleus of every cell in live animals. Transparency also makes it possible to visualize specific cells or subcellular structures in great detail in living animals by using transgenes to express the green fluorescent protein in the desired locations (Chalfie et al., 1994). Finally, transparency makes it possible to use a laser specifically to kill any cell in a living animal in order to determine the function of that cell (Bargmann and Avery, 1995). The powerful genetics and well-described anatomy of the worm synergize to allow a remarkably precise understanding of gene function. It is possible, for example, to characterize a mutant with abnormal development by simply observing its cell divisions and cell migrations during development and comparing the results to the known pattern for the wild type. Thus the precise point in development at which a mutant first deviates from the wild type can be determined, allowing primary defects caused by the mutation to be distinguished from secondary consequences. In a typical example, a mutant was identified in which animals appear anatomically normal but have abnormal behavior as adults. The primary cause leading to this ultimate defect was traced to one pair of neurons inappropriately undergoing programmed cell death late in development (Conradt, and Horvitz, 1998). Without this precise description of the mutant phenotype, it might have been concluded based on the behavioral defects, that the gene was involved in neural or muscular function rather than in the process of cell death. C. Molecular Genetics and Genomics

C. elegans was the first animal to have its genome sequence determined. More than 19,000 genes are predicted from the 97-megabase sequence, including homologs of 74% of known human genes (C. elegans Sequencing Consortium, 1998). Heritable transgenes can be produced by simply injecting the DNA of interest into animals (Mello and Fire, 1995). Genetic mapping and other approaches make it easy molecularly to identify genes identified by mutations. Techniques for inactivating C. elegans genes have recently been developed (see

24

Koelle

below) allowing investigations of the functions of genes identified in the genome sequence (Jansen, et al., 1997; Liu et al., 1999). In sum, essentially the full set of experimental techniques needed to investigate gene function have been developed in C. elegans to a level where each is rapid and convenient. The result is that once a biological process of interest is identified in worms, it is a remarkably rapid process to identify the genes and molecules responsible for that process. One simply obtains mutations that disrupt the process and clones the genes identified by the mutations. Further molecular and genetic analyses can then often yield deep mechanistic insights into the biological process. III. Is It Possible to Use C. elegans as a Model for Human Biology and Disease? Can the experimental advantages of the worm outlined above be used to shed light on problems in human biology and disease? This depends on whether the biological processes of interest are conserved across evolution from humans to worms, something that might at first appear unlikely based on the extreme differences in anatomical complexity and life style between these two organisms. Although it is true that humans and worms bear little resemblance at the macroscopic level, it turns out that conservation of molecular pathways between humans and worms is strikingly common, and that an understanding at the level of a molecular pathway is often what is required to advance understanding of a human disease. Only when one appreciates the subtlety that molecular pathways, rather than their macroscopic outcomes, are what is often conserved across evolution, does the potential power of exploiting worm genetics to understand human biology become apparent. Some examples are given below to illustrate the types of evolutionary conservation that are typically observed between humans and worms. These examples illustrate conservation of two types: the obvious type, in which both a molecular pathway and its macroscopic outcome are conserved, and the less obvious type in which a molecularly conserved pathway leads to apparently different outcomes in worms and humans. A. Obvious Evolutionary Conservation Between Humans and Worms

The most basic aspects of intermediary metabolism, such as DNA replication, the cytoskeleton, protein transport and secretion, protein synthesis, folding, and degradation, might be considered core functions in the biology of eukaryotes since they appear to be highly conserved across evolution. Indeed, about half of

C. elegans as a Model for Human Biology and Disease

25

the known human genes have homologs in both worms and yeast, and these appear to be largely involved in such core functions (Chervitz et al., 1998). There would appear to be relatively little benefit to using C. elegans genetics to investigate these core functions, since they might be more efficiently analyzed using the even more powerful genetics available in yeast. The value of C. elegans as a model system is clearer for those human biological functions that are not conserved in yeast. Whereas only half of human proteins have yeast homologs, about three-quarters have homologs in C. elegans (Chervitz et al., 1998). The conserved molecules found in worms, but not yeast, appear to exist largely to deal with issues of multicellularity. Different cell types (e.g., muscles, neurons) must arise in development and carry out specialized functions, and a very complex set of signal transduction machinery is required for the various cells of the animal to communicate with each other. Because defects in human disease are often associated with development or function of specific cell types, or with cell-signaling pathways, worms should often provide an appropriate model. Indeed, a recent survey of identified human disease genes shows that a large fraction have orthologs in C. elegans but not in simpler model organisms such as yeast or bacteria (Mushegian et al., 1997). In some cases, the value of the worm as a model is obvious because both a molecular pathway and its macroscopic outcome are conserved from worms to humans. The most famous example of such a case is that of programmed cell death (reviewed by Metzstein et al., 1998). One human gene involved in cell death, Bcl-2, was identified because it lies at a chromosomal translocation breakpoint associated with certain cancers in which programmed cell death has gone awry (Korsemeyer, 1999). However, as is typically the case upon identification of such a human disease gene, the normal role of Bcl-2 in the process of cell death and how Bcl-2 defects might perturb cell death remained obscure. The contribution of C. elegans to understanding programmed cell death in humans is a classic example of exploiting the worm as a model. From the complete description of the worm cell lineage, it was known that a large number of cell deaths occur in C. elegans (Sulston and Horvitz, 1977), and that the dying cells appeared to be similar to cells undergoing programmed death in humans. Using the power of worm genetics and the ease of observing extra cells in the transparent bodies of C. elegans, a series of mutations were identified that caused either too little or too much cell death (Metzstein et al., 1998). Using the molecular genetic tools available in the worm, the genes identified by these mutations were cloned. Among them was a homolog of Bcl-2, establishing that the molecular mechanisms of cell death in humans and worms were related (Hengartner and Horvitz, 1994). This spurred the search for human homologs of the remaining worm cell death genes, which were eventually identified and proved to function in a molecular pathway for cell death quite similar between worms and mammals.

26

Koelle B. Nonobvious Conservation Between Humans and Worms: Conserved Molecular Pathways that Lead to Different Macroscopic Outcomes

Cell death, in retrospect, appears to be an obvious case for the use of a C. elegans model, since similar types of cell deaths were known to occur in both humans and worms. However, the utility of worms as a model extends remarkably beyond what is immediately obvious. Often complex molecular processes defective in human disease are conserved from worms to humans, molecule for molecule, although the outcomes of the proper or improper functioning of these pathways may be unrecognizably different in the two organisms. One illustration of this point is the ras signaling pathway (reviewed by Sternberg and Han, 1998). The ras oncogene was identified as being one of the most frequently mutated genes in human tumors, but initially its role in normal biology and how defects in ras led to tumor formation remained obscure (Shih and Weinberg, 1982). A close homolog of ras was identified in C. elegans, because defects in the worm ras gene led to alterations in the pattern of cell divisions during the development of the C. elegans vulva (Beitel et al. 1990; Han and Sternberg, 1990). Subsequent genetic studies of vulva formation in worms led rapidly to the identification of many mutations that disrupt vulval development. Cloning of the genes identified by these mutations allowed identification of a set of at least nine proteins that function with ras in a growth factor–signaling pathway that is conserved, protein for protein, between humans and worms (Sternberg and Han, 1998). The identification and analysis of ras signaling components in worms has thus shed considerable light on human tumor formation. A second illustration of a nonobvious worm model for human disease is the case of Alzheimer’s disease. Rare inherited cases of Alzheimer’s disease are caused by mutations in genes encoding transmembrane proteins called ‘‘presenilins’’ (reviewed by Haass and De Strooper, 1999). At the time of their identification, the roles of presenilins in normal biology and Alzheimer’s disease were unknown. Although no mutations are known that cause defects in worms similar to the symptoms of Alzheimer’s disease, a C. elegans presenilin homolog, called sel-12 was identified by mutations that affect the pattern of cell divisions during worm development (Levitan and Greenwald, 1995). The normal human presenilin genes (but not the mutant versions found in Alzheimer’s patients) can substitute for the sel-12 gene in C. elegans, establishing the functional similarity of these genes across evolution (Levitan et al., 1996). By exploiting the power of C. elegans as an experimental system, a number of molecules that function in a molecular pathway with sel-12 have been identified. The presenilins were thus found to function in the Notch signal transduction pathway (whose components are conserved in humans, flies, and worms) and to control the proteolytic processing of the Notch signaling molecule (Struhl and Greenwald, 1999). Current studies of

C. elegans as a Model for Human Biology and Disease

27

the human presenilins focus on the potential role of Notch signaling in Alzheimer’s disease and the potential role of presenilins in the proteolytic process of the amyloid precursor protein, whose cleavage products accumulate in patients with Alzheimer’s disease. What were the key insights that allowed worms to be used to help understand the processes underlying tumor formation and Alzheimer’s disease? In one case, it was the realization that to study tumor formation in a worm model one needed to look not for mutant worms with tumors but for mutant worms with vulval defects. The ras signaling pathway that controls cell proliferation in mammals, although conserved in great detail at a molecular level, has a different macroscopic outcome (correct differentiation of vulval precursor cells) in worms. Similarly, in the case of the presenilins, it was the realization that to study Alzheimer’s disease in a worm model one needed to look not for mutant worms with amyloid plaque formation but for mutant worms with certain cell differentiation defects. As with the ras pathway, the molecular process involving presenilins appears to be remarkably highly conserved from humans to worms, but its activity ultimately manifests itself differently in each organism. The list of conserved molecular pathways that lead to outcomes in worms and humans that are not immediately obviously related is already extensive. Examples include the insulin-signaling pathway, involved in diabetes in humans, which in worms controls a developmental decision to enter the spore-like dauer form when food is scarce (Ogg and Ruvkun, 1998). Another example is of the retinoblastoma gene, which functions as a tumor suppressor in humans, but which functions in C. elegans in pathway controlling an aspect of vulval development (Lu and Horvitz, 1998). Since three-quarters of human proteins have worm homologs, it can be expected that the list of examples of nonobvious C. elegans models for human disease will be extended considerably in the future.

IV. How Can Gene Knockout Technology in Worms Be Used to Explore the Functions of Human Disease Gene Homologs in C. elegans? In the examples cited above, the connection of the human disease gene to its nonobvious worm model was made easily, since the worm homolog of the disease had already been analyzed, placing it in a molecular pathway controlling a process already under study by C. elegans molecular geneticists. Although these examples are useful to illustrate the potential for using worm models of human disease, they are not typical examples. Mutations currently are available in less than about 1700 of the 19,099 genes predicted by the worm genome sequence (Hodgkin and Martinelli, 1999), and only a small fraction of these available muta-

28

Koelle

tions have been analyzed in detail. A typical human disease gene will therefore have a C. elegans homolog in which no mutations are currently available, and whose function in C. elegans is unknown. Thus, in the majority of cases, the immediate challenge to exploiting C. elegans as a disease model is to identify a mutation in a C. elegans disease gene homolog. As argued above, the resulting phenotype is likely to bear little resemblance to the original human disease. Once the mutant phenotype is known, however, the full power of C. elegans genetics can potentially be harnessed to identify additional mutations that mimic, enhance, or suppress the original mutation, thus identifying a set of additional genes that function in a common process with the original disease gene. Because the molecular pathway thus delineated will typically be conserved in humans, this procedure should lead to further insights into the disease process in humans. Until recently, the sole hurdle to the procedure outlined above has been that the technology for knocking out C. elegans genes was quite primitive, so that the worm’s large number of disease gene homologs remained tantalizingly out of reach from analysis. Fully realizing the potential of C. elegans as a model for human biology required removing this last technical hurdle. Fortunately, a recent series of technical developments have allowed more than 100 C. elegans genes to be knocked out (Jansen et al., 1997; Liu et al., 1999), and improvements to this new gene knockout technology are likely to continue. The new technology produces true gene knockouts; i.e., strains of worms in which the gene of interest has been actually deleted from the genome. This is in contrast to the ‘‘RNA interference’’ technique (Tabara et al., 1998), also recently developed in C. elegans, in which injection or ingestion of a doublestranded RNA copy of a gene of interest produces only a transient and potentially partial inactivation of a target gene. True gene knockouts, because they yield an unending supply of animals with a fully and reproducibly inactivated gene, are better suited to the kinds of analyses, such as genetic enhancer or suppresser screens, that are required to delineate molecular pathways in which a gene of interest participates. Such true gene knockouts are currently routinely produced in a small number of C. elegans laboratories, but the technology is likely soon to be implemented in many, if not most, C. elegans laboratories. The new techniques used to isolate gene knockouts are summarized below and illustrated in Figure 2. Attempts to use technology based on homologous recombination, as is routine in yeast and mice, have not been successful in C. elegans. Instead, an entirely new approach has been developed (Jansen et al., 1997; Liu et al., 1999). The principle underlying the C. elegans gene knockout procedure is simple: A large number of worms are treated with a chemical mutagen that randomly damages their DNA and that at a low frequency will cause a chromosomal deletion of any gene of interest. DNA is prepared from some of the animals

C. elegans as a Model for Human Biology and Disease

29

Figure 2 A current implementation of the C. elegans gene knockout technology. Mutagenized animals are grown in microtiter wells, and a rare well may contain animals deleted for a gene of interest (shaded well in illustration). Some animals from each well are stored frozen, whereas some of their siblings are transferred to new microtiter plates and lysed to release DNA. Pools of these DNA samples are PCR amplified. Deletion mutations are detected as smaller PCR products and can be detected even when they represent a small (less than one part in 2000) proportion of the DNA pool. When a positive pool is identified, the 48 DNA samples from which it was derived are tested by PCR to identify the one containing the mutant DNA (shaded well in illustration). Animals from the corresponding frozen culture are thawed and tested by PCR to identify individual living animals carrying the gene deletion.

and live siblings of these animals (carrying the same mutations) are stored away. Polymerase chain reaction (PCR) primers flanking a gene of interest are used to amplify the gene from the DNA samples. Deletions in the gene of interest are detected, since they bring the primers closer together and thus produce smaller PCR products. After a mutation in the gene of interest is detected, a live animal carrying the mutation is recovered from among the stored siblings. The challenges in this strategy lie in the complex logistics of its execution. In current implementations of the technology, deletions in a gene of interest are detected only once every several hundred thousand mutagenized animals examined, so that mutant ‘‘libraries’’ representing 1 million mutagenized genomes or more must be generated, stored, and screened if gene knockouts are to be pro-

30

Koelle

duced with a high probability of success. Needless to say, all aspects of the procedure must be engineered for high throughput; meaning that all samples are in 96-well format and all pipetting is carried out with 12-channel pipettors. Several methods for implementing the knockout technology have been released either through traditional publications (Jansen et al., 1997; Liu et al., 1999) or as detailed protocols available via the internet (Moulder and Barstead, 1997; Shechner and Koelle, 1999). One such implementation is illustrated in Figure 2. In this implementation, it takes about 3 weeks to generate a million-genome mutant library, which can be stored indefinitely and screened about 200 times. It takes about 15–20 days of part-time work to screen the library and recover a knockout of a particular gene. Although current gene knockout procedures are quite successful and convenient, further improvements in the near future are likely to make the technique significantly faster, easier, and able to produce knockout mutations with greater reliability (success rates for recovering knockouts from a 1-million genome library are currently less than 100%). For example, better mutagenesis procedures that produce deletion mutations at higher frequencies could dramatically improve the technique. A project is being undertaken by a consortium of laboratories to generate deletions in many or all C. elegans genes (see http:/ / www.cigenomics.bc.ca/elegans/). If successful, this would eventually make C. elegans gene knockouts easily available to researchers outside of C. elegans laboratories who cannot readily generate such mutants themselves.

V.

What Are the Prospects for C. elegans as a Model for Human Biology and Disease?

A general procedure has been described above for developing C. elegans models for human biology and disease by knocking out C. elegans genes and using the power of worm molecular genetics to analyze the molecular pathways in which these genes participate. With the development of efficient C. elegans gene knockout technologies, all the technical barriers to applying this procedure have finally been overcome. What impact will this approach have in the coming years? Although the prospects for this approach are theoretically unlimited, several factors may curtail its use in practice. The first is the fact that a fair amount of specialized training and some specialized equipment is required fully to exploit C. elegans as an experimental system. The logic and methods involved in classic genetic techniques such as genetic screens and genetic mapping are not familiar enough to most molecular biologists and biochemists to allow them successfully to carry out these types of experiments without further training. Also, familiarity with the particulars of C. elegans biology is required to use this organism for such experiments. Thus researchers who are not able to collaborate with a local

C. elegans as a Model for Human Biology and Disease

31

C. elegans laboratory, or who cannot spend perhaps a year training in such a laboratory, are unlikely to be highly successful in using C. elegans as a model system to study the homolog of a gene of interest. For the foreseeable future, the use of worms as a model for human biology and disease may be limited by the pool of researchers trained as worm geneticists. This situation stands in contrast, for example, to the wide adoption of the yeast two-hybrid technique by large numbers of laboratories with no previous experience in yeast genetics (Colas and Brent, 1998). Two-hybrid screens rely on little specialized knowledge of either genetics or yeast biology, and once a screen is completed, researchers can return to familiar methods of molecular biology or biochemistry to pursue results of the screen. A second factor may limit the use of worm models in a way that is currently harder to assess. Some C. elegans gene knockouts cause only subtle or undetectable defects, presumably either due to redundancy with other genes or due to researchers’ lack of ability to find conditions or levels of analysis that reveal defects. Other gene knockouts result in early lethality, but observation may yield little information as to the specific cause of death. In either of these cases, it can be difficult to exploit such a knockout to define a genetic pathway in which the gene of interest participates. Over the next few years, as C. elegans gene knockouts begin to provide an increasing number of potential worm models for human disease, it should become apparent to what extent this experimental approach will have an impact on progress in understanding human biology. References Bargmann CI, Avery L. Laser killing of cells in Caenorhabditis elegans. Methods Cell Biol 1995; 48:115–250. Beitel G, Clark S, Horvitz HR. Caenorhabditis elegans ras gene let-60 acts as a switch in the pathway of vulval induction. Nature 1990; 348:503–509. Brenner S. The genetics of Caenorhabditis elegans. Genetics 1974; 77:71–94. C. elegans Sequencing Consortium. Genome sequence of the nematode C. elegans: A platform for investigating biology. Science 1998; 282:2012–2018. Chervitz SA, Aravind L, Sherlock G, Ball CA, Koonin EV, Dwight SS, Harris MA, Dolinkski, K, Mohr S, Smith T, Weng S, Cherry JM, Botstein D. Comparison of the complete protein sets of worm and yeast: orthology and divergence. Science 1998; 282:2022–2028. Chalfie M, Tu Y, Euskirchen G, Ward WW, Prasher DC. Green fluorescent protein as a marker for gene expression. Science 1994; 263:802–805. Colas P, Brent R. The impact of two-hybrid and related methods on biotechnology. Trends Biotechnol. 1998; 16:355–363. Conradt B, Horvitz HR. The C. elegans protein EGL-1 is required for programmed cell death and interacts with the Bcl-2-like protein CED-9. Cell 1998; 93:519–529.

32

Koelle

Haass, C, De Strooper, B. The presenilins in Alzheimer’s disease-proteolysis holds the key. Science 1999; 286:916–919. Han M, Sternberg PW. let-60, A gene that specifies cell fates during C. elegans vulval induction, encodes a ras protein. Cell 1990; 63:921–931. Hengartner MO, Horvitz HR. C. elegans cell survival gene ced-9 encodes a functional homolog of the mammalian proto-oncogene bcl-2. Cell 1994; 25:76(4):665–76. Hodgkin J, Martinelli S. 1999 genetic map of Caenorhabditis elegans. Jansen G, Hazendonk E, Thisjssen KL, Plasterk RHA. Reverse genetics by chemical mutagenesis in Caenorhabditis elegans. Nature Genet 1997; 17:119–121. Korsemeyer SJ. BCL-2 gene family and the regulation of programmed cell death. Cancer Res 1999; 59(suppl):1685s–1692s. Levitan D, Greenwald I. Facilitation of lin-12-mediated signalling by sel-12, a Caenorhabditis elegans S182 Alzheimer’s disease gene. Nature 1995; 377:351–354. Levitan D, Doyle TG, Brousseau D, Lee MK, Thinakaran G, Slunt HH, Sisodia SS, Greenwald I. Assessment of normal and mutant human presenilin function in Caenorhabditis elegans. Proc Natl Acad Sci USA 1996; 93:14940–14944. Liu XL, Spoerke JM, Mulligan EL, Chen J, Reardon B, Westlund B, Sun L, Abe K, Armstrong B, Hardiman G, King J, McCague L, Basson M, Clover R, Johnson CD. High-throughput isolation of Caenorhabditis elegans deletion mutants. Genome Res 1999; 9:859–867. Lu X, Horvitz HR. lin-35 and lin-53, two genes that antagonize a C. elegans Ras pathway, encode proteins similar to Rb and its binding protein RbAp48. Cell 1998; 95:981– 991. Metzstein MM, Stanfield GS, Horvitz HR. Genetics of programmed cell death in C. elegans: past, present, and future. Trends Genet 1998; 14:410–416. Mello C, Fire A. DNA transformation. Methods Cell Biol 1995; 48:452–482. Moulder G, Barstead R. 1997 http:/ /snmc01.omrf.uokhsc.edu/revgen/knockout.html. Mushegian AR, Bassett DE, Boguski MR, Bork P, Koonin EV. Positionally cloned human disease genes: patterns of evolutionary conservation and functional motifs. Proc Natl Acad Sci USA 1997; 94:5831–5836. Ogg S, Ruvkun G. The C. elegans PTEN homolog, DAF-18, acts in the insulin receptorlike metabolic signaling pathway. Mol 1998; Cell 2:887–93. Riddle DL, Blumenthal T, Meyer BJ, Priess JR, eds. C. elegans II Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1997. Shechner D, Koelle MR. 1999. http:/ /info.med.yale.edu/mbb/koelle/protocol list page.html. Shih C, Weinberg RA. Isolation of a transforming sequence from a human bladder carcinoma cell line. Cell 1982; 29:161–169. Sternberg PW, Han M. Genetics of RAS signaling in C. elegans. Trends Genet 1998; 14: 466–472. Struhl G, Greenwald I. Presenilin is required for activity and nuclear access of Notch in Drosophila. Nature 1999; 398:522–525. Sulston JE, Horvitz HR. Post-embryonic cell lineages of the nematode Caenorhabditis elegans. Dev Biol 1977; 56:110–156. Sulston JE, Shierenberg White JG, Thomson JN. The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev Biol 1983; 100:64–119.

C. elegans as a Model for Human Biology and Disease

33

Tabara H, Grishok A, Mello CC. RNAi in C. elegans: soaking in the genome sequence. Science 1998; 282:430–431. White JG, Southgate E, Thomson JN, Brenner S. The structure of the nervous system of the nematode Caenorhabditis elegans. Phil Trans R Soc Lond B 1986; 314:1–340. Wood WB, Community of C. elegans researchers, eds. The Nematode Caenorhabditis elegans. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, 1988;

3 Transgenic Mouse Models COODUVALLI S. SHASHIKANT

FRANK H. RUDDLE

The Pennsylvania State University University Park, Pennsylvania

Yale University School of Medicine New Haven, Connecticut

I.

Introduction

Interest in manipulating the genetic make-up of organisms has a long-standing history, beginning with selective breeding to the present cloning of mammals. The route to transgenesis can be traced to the convergence of gene transfer methods developed in mammalian cell culture systems and methods to manipulate early mouse embryos (see Refs. 1–3 for historical perspectives). The past two decades have witnessed three major technological breakthroughs that have significantly enhanced our abilities to alter the genetic make-up of mammals. First, in the early 1980s, it was shown that foreign DNA could be introduced into the mouse germline by directly injecting into pronuclei of one-cell mouse embryo [4]. This process of transforming the mouse genome by pronuclear injections was referred to as ‘‘transgenic’’ [5]. Transgenic technology, first demonstrated in the mouse, is now practiced widely in flies, worms, birds, fish, frogs, and mammals including mice, cows, sheep, and pigs. Second, in the late 1980s, it was shown that genes could be disrupted by a homologous recombination-based strategy in embryonic stem (ES) cells [6,7]. These ES cells are then used to generate mice that carry specific alterations in the germline [8,9]. This process of mutating genes was referred to as ‘‘gene knockout’’. Gene knockout technol35

36

Shashikant and Ruddle

ogy has resulted in the creation of a large number of mutant mouse strains, which has provided valuable insights into gene function at the organismic level (reviewed in Ref. 10). Third, in the mid 1990s, it was shown that animals could be generated by replacing the zygotic nucleus with the somatic nucleus [11,12]. This process, referred to as cloning or nuclear transfer, was first demonstrated in sheep and later in other organisms including mice [13]. The cloning of mammals has provoked an intense debate on the technological and ethical implications for the human cloning. In addition to these advances, a number of molecular biological techniques have emerged, raising the possibility of more sophisticated genome modifications. Many of these manipulations are carried out extensively in mice. In this chapter, we review emerging technologies that will enable us to generate better mouse models for the study of gene function. II. Gene Transfer Methods There are four different methods for generating transgenic mice: pronuclear injection, viral infection, and cell- and sperm-mediated gene transfers [14]. A. Pronuclear Injections

In the mouse egg after fertilization, the male and female nuclei remain independent for a few hours before they fuse to make the zygotic nucleus. These independent nuclei can be visualized under a phase contrast microscope for several hours and, with the aid of micromanipulators, DNA can be injected into one of them. Several molecules of tandemly joined DNA then integrate into random sites within the egg cell’s genome. Viable eggs that survive injections are transferred into the oviduct of pseudopregnant mice [4] (reviewed in Ref. 2). Founder generation mice that carry the transgene can be identified by Southern blot or polymerase chain reaction (PCR) analyses (Fig. 1). Transgene expression and its phenotypic consequences are analyzed in subsequent generations. The original methods described for pronuclear injections by Gordon et al. [4] are followed with little modifications in generating transgenic mice. One major disadvantage of the pronuclear injection is the low number of transgenic mice that are generated which appropriately express the transgene. However, pronuclear injection remains a method of choice for generating transgenic mice. B. Viral Infection and Retroviral Vectors

Exogenous DNA can be introduced into the germline by infecting mouse embryos with a retrovirus [15]. Retroviral vectors carrying foreign DNA can be used to infect early embryos (eight-cell stage), and infected embryos can be implanted into a pseudopregnant mouse [16–18]. As the retroviral genome is copied and

Transgenic Mouse Models

37

Figure 1 Pronuclear injection method to generate transgenic mouse.

integrated at specific sites in the mouse genome as a proviral DNA, there is a high probability of integration of the transgene into the genome. However, retrovirus-based vectors have limits on the size of foreign DNA that they can carry (8 kb). Thus, at least in mice, retroviral vectors are not that popular. In the chick, however, replication-competent and replication-incompetent retroviral vectors provide a means of introducing transgenes at limited stages in specific tissue [19]. Besides retroviral vectors, transposons can be used to integrate foreign DNA into the genome. Recently, the Drosophila element, mariner’s transposon element, was shown to integrate into the chick genome and transpose at high efficiency [20]. Vectors based on transposons may well be tested for their usefulness in generating transgenic animals. Electroporation of nonviral DNA has been used to introduce nonviral DNA into early chick embryos [21,22]. However, these methods are not useful in introducing DNA into the germline. In general, a method based on infection, transfection, or electroporation is likely to be less labor intensive than pronuclear injections. However, no such reliable and efficient method of making transgenic mice currently exists. C. Cell-Mediated Gene Transfers

Genes can be transferred into mammalian cells by a variety of methods including fusion, transfection, electroporation, and direct injection. There are two methods of transferring DNA from cells to the mouse germline. The first method is based

38

Shashikant and Ruddle

on ES cell technology. Embryonic stem cells are derived from mouse blastocysts. These cells can be maintained in culture without differentiation, transfected with genes of interest, prescreened for correct targeting and proper expression, and introduced into mouse blastocysts by direct injection. ES cells contribute extensively to different tissues of the chimeric mouse including the germline. Thus, a transgenic line can be established via ES cell technology. This method at present, however, is not attractive, because it requires breeding for two generations to establish transgenic founders. However, this route to transgenesis may become important, particularly if the probability of generating expression lines is enhanced with preselected sites of transgene insertion [23]. At present, ES cells are extensively used to disrupt genes by homologous recombination, creating mutant mice [10]. ES cell technology is frequently used in creating gene traps [24,25]. A second method of cell-mediated gene transfer is through nuclear transplantation (Fig. 2). Recent findings which suggest that nuclei isolated from adult

Figure 2 mouse.

Schematic of the nuclear transplantation strategy employed in cloning of

Transgenic Mouse Models

39

mammalian cells can replace the zygotic nucleus and undergo development has raised the possibility that genes transferred into cells other than ES cells can be used to generate transgenic animals [13]. Cloning by nuclear transplantation has been demonstrated in several mammals including mice. This dogma-shattering advance is yet to find its niche in transgenic mouse technology. In large mammals, however, where generating transgenic animals by pronuclear injection has been an expensive proposition, nuclear transplantation has become the method of choice. To date, cloning has been possible in the mouse from nuclei derived from ES cells, cumulus, and skin cells [26–30]. Again, if proven feasible, these methods will provide a means specifically to select cells expressing the transgene in vitro for generating transgenic mice. Unlike ES cell–mediated gene transfers, nuclear transplantation provides a method for generating founder lines within one generation. D. Sperm-Mediated Gene Transfer

There has been a considerable effort in developing sperm as a vehicle for DNA delivery. DNA can be absorbed onto sperm and delivered into oocytes either by in vitro fertilization or intracytoplasmic injections [31,32]. Initial efforts for transgenesis based on in vitro fertilization techniques fell into disrepute owing to lack of reproducibility and the unclear understanding of mechanisms of DNA delivery [32,33]. Coinjection of permeabilized sperm heads and exogenous DNA into unfertilized mouse eggs has recently been found to produce a remarkably high number of eggs expressing the transgene [31]. These experiments may call for a reexamination on the utility of sperm-mediated gene transfers. At present, the reported efficiency, which is comparable to that of pronuclear injections in mice, does not necessarily make this method competitive for mouse models. However, for nonmouse species, where pronuclear injections are difficult, spermmediated gene transfer may become an attractive method of generating transgenic animals. In summary, although pronuclear injection has remained the most successful method, for the generation of transgenic animals, other recent developments promise to modify and generate additional methods.

III. Site of Integration The mechanism by which a transgene inserts into the genome following pronuclear injections remains poorly understood [2]. Transgenes form head-to-tail concatamers and integrate into a single site in the genome. Usually these concatemers contain at least one complete copy of the transgene. Occasionally, integration may occur at more than one site, and it is possible to establish two separate

40

Shashikant and Ruddle

transgenic lines from a single founder mouse. The transgene integration occurs by illegitimate recombination at random sites in the genome. Transgene integration usually occurs in one-cell embryos, but occasionally mosaic founders can arise by integration in two- or four-cell embryos. In mosaic founders, not every cell carries the transgene, resulting in lower than 50% frequency of transmission. Mosaic founders are advantageous for the study of transgenes which are embryonically lethal when fully expressed [34]. The mosaic founder remains viable. At a low frequency, the lethal transgene is transmitted to the next generation, thus providing a constant source of transgenic embryos, where lethal phenotype can be systematically studied. The site of integration in the genome influences the expression of the transgene. In fact, a major concern with transgenic technology is that not all founders that carry the transgene express it. The chromatin environment in which the transgene randomly inserts is critical in determining the level of expression. Location of the transgene in heterochromatin or in nuclear compartments where transcription is silenced results in a lack of transgene expression. In contrast, location of the transgene in euchromatin or in the nuclear compartment where active transcription is sustained results in expression of the transgene [2,35]. Furthermore, transgene expression is often copy number independent, further strengthening the significance of the local chromatin environment. Transgene expression is also influenced by cis-acting sequences located in the vicinity of the transgene integration. This often results in transgene expression at additional spatial and temporal sequences. Local cis-acting elements influence the level of transgene expression. Further understanding of factors influencing transcriptional activity is critical for improving methods by which consistent transgene expression can be obtained. Characterization of boundary elements, insulators, and silencers will contribute toward the better transgene design [36,37]. The presence of locus control regions (LCRs), strong enhancers or nuclear matrix attachment sequences can direct the formation of transcriptionally active domain [38,39]. Sequences which can direct position-independent, copy number dependent expression in multiple tissues have recently been identified [39–44]. However, these sequences have not yet been incorporated into a general strategy for designing transgenes. Alternatively, one can preselect sites at which the transgene be integrated [23]. This can be achieved by the cell-mediated transgenesis approach. Strategies can be designed specifically to integrate a single copy of the transgene at a given site by homologous recombination in ES cells. These targeted cells can then be used to generate transgenic mice [23,45].

IV. Insert Size Conventional transgenic constructs, generated in pBR322- or pUC-based vectors are limited in their size. Hence the constructs designed excludes many sequences

Transgenic Mouse Models

41

that are important for proper levels and patterns of gene expression. Small transgenes are more vulnerable to the influence of local cis-acting sequences and chromatin environment. Although small-insert transgenic mice are widely employed in various studies, in general, they are poor models of gene expression. Many genes and gene clusters are often large in size with regulatory elements scattered in the immediate 5′, 3′, and intronic regions or at considerable distances from coding regions. For the proper expression of even modest-sized genes in transgenic mice, it is important to provide all the regulatory elements and a suitable chromatin environment. Sequences located at greater distances from the coding region of the gene can be included in the design of the transgene. Many of these shortcomings of the small-insert transgenics have been resolved with the development of large-insert cloning vehicles, especially yeast and bacterial artificial chromosomes (YACs and BACs, respectively), and more recently a yeast–bacteria shuttle vector, pClasper, which can hold several hundred kilobases of DNA fragments [46]. Yeast artificial chromosomes were developed by combining yeast sequences that define centromeres, telomeres, and a replication origin (ori) along with selectable markers [47]. These sequences were combined to create a linear cloning vehicle that can hold up to 2000 kb. Bacterial artificial chromosomes, containing the F-factor ori and P1 artificial chromosomes (PACs) containing P1 bacteriophage replicon that can hold inserts in the range of 300 kb were developed later as additional large-insert vehicles [48–50]. Although YACs provide a larger cloning capacity and easier insert manipulation, there are significant drawbacks in the use of YACs for transgenic studies. These include poor cloning efficiency and chronic chimerism in most YAC libraries. In contrast, bacteria-based vectors, BACs and PACs, are relatively easy to handle, and the inserts are more stable and less prone to chimerism. However, methods to modify large inserts in bacteria are poorly developed compared to yeast. The yeast– bacteria shuttle vector, pClasper, combines the advantages of BACs and YACs without their apparent disadvantage [51]. The inserts cloned in pClasper can be stably maintained in yeast and manipulated using the yeast genetic system and then shuttled into bacteria and prepared for isolating DNA fragments to generate transgenic mice [51]. Yeast artificial chromosomes can be introduced into mice either by direct pronuclear injection or by ES cell technology (reviewed in Refs. 52 and 53). In the latter method, yeast spheroplasts containing YACs are fused with ES cells. Alternatively, YACs are introduced into ES cells by lipofection. Embryonic stem cells carrying YAC DNA are then injected into mouse blastocysts to create chimeric mice. The inserts from BACs and pClasper are introduced into mice by direct pronuclear injections. Large-insert cloning vectors make it possible to examine large genes and gene complexes; to identify long-acting cisregulatory elements; to create transgenic models which more faithfully reproduce the functional aspects of endogenous loci; to complement known genetic mutations; and to study higher order chromosomal structures such as X chromosome inactivation.

42

Shashikant and Ruddle V.

Recombinational Cloning and Modification Techniques

One problem associated with large-insert transgenics is that the molecular subcloning techniques based on availability of unique or rare restriction sites are inadequate in cloning and modification of the inserts. Further refinement of functional analysis of genes will require modifications of large inserts. Precise alterations in both the coding and regulatory sequences are essential for understanding the role of specific sequences in gene function. Methods based on homologous recombination can be used to generate DNA constructs for transgenic analysis. In yeast, compared to other organisms, the frequency of homologous recombination is relatively high. Transformed DNA (lacking autonomous replicating sequences) invariably integrates into the yeast genome by homologous recombination. This principle can be exploited to modify inserts in yeast by simple transformation and selection protocols [46]. Two popular methods to manipulate the yeast genome are pop-in/pop-out and one-step gene replacement methods. These methods can be used to make precise deletion or mutation of sequences or for insertion of reporter genes or modified sequences in the transgene. Similar methods for modification of large-insert transgene are currently being developed in bacterial system as well [54–56]. Although transgenic applications of recombinational cloning have been relatively few so far, an increased application of these methods is likely to occur in the near future.

VI. Cis-Regulatory Analysis Transgenic mice provide a reliable system for mapping cis-regulatory elements that control spatial and temporal patterns of gene expression. Although many studies employ cell cultures or DNA protein binding and transcription assays for regulatory analyses, the utility of transgenic mice in corroborating in vitro experiments is well appreciated. In many instances, regulatory sequences identified in cell culture experiments have proven to be inconsequential for regulating gene expression in the whole organism. Sequences which control spatial patterns of gene expression in multiple tissues can be identified reliably by transgenic analysis. However, transgenic studies are labor intensive, expensive, and often time consuming. Further, mapping cis-acting elements is not an easy undertaking, because these elements may be located at multiple sites either in the 5′, 3′, or intronic regions, in the immediate vicinity of the gene, or at a considerable distance. These difficulties make cis-regulatory analysis an unattractive proposition. However, mapping of regulatory elements which direct tissue-specific expression in a developmentally regulated pattern is critical for understanding the function of a gene. These studies establish regulatory pathways that underlie tissue differentiation and provide powerful tools to manipulate developmental processes. A

Transgenic Mouse Models

43

collection of regulatory sequences with distinct spatial and temporal specificity will be a valuable biological resource. By careful deletion and mutation analysis it is possible to engineer regulatory sequences that direct gene expression with exquisite specificity. In the enhancer world point of view, complex patterns of gene expression are achieved by the combination of distinct, modular regulatory units. Each module directs distinct tissue-specific expression patterns. Identification of such modules is a matter of fine resolution mapping of regulatory elements. Cis-regulatory elements are generally identified by reporter gene analysis (reviewed in Refs. 57 and 58). The bacterial β-galactosidase gene is the most commonly used reporter, as its expression can be visualized by staining tissues with X-gal. Recently, the gene encoding green fluorescent protein (GFP) has been engineered for use in transgenic mice (reviewed in Ref. 59). Fluorescence allows detection of the reporter gene expression by noninvasive methods. The other reporter gene being used is human alkaline phosphatase. Bicistronic reporter gene constructs containing two reporter genes are also available [60]. Typically, a reporter gene is inserted in the frame of the coding sequence and the transgene, containing the surrounding genomic region, is examined for expression in transgenic mice. Alternatively, fragments of DNA surrounding the gene can be cloned upstream of a heterologous promoter, for instance, the mouse heat-shock protein 68 (hsp68) promoter, driving a reporter gene [61]. Although, cis-regulatory elements may be scattered over large distances, the first place to look for them is in the immediate vicinity of the gene. Typically, several kilobases of DNA immediately upstream of the translation start sites are examined for regulatory sequences. If the functional elements are not identified in these regions, the next step is to include larger 5′ and 3′ DNA fragments and all introns for analysis. Here, DNA fragments cloned in the large-insert vectors described above become useful. To date, only a few studies have taken advantage of homologous recombination to generate reporter gene constructs. In a study of Hoxc8 gene regulation, pClasper was used to rescue a 27-kb DNA fragment surrounding Hoxc8 from a YAC clone and a lacZ-URA3 cassette was then inserted in the frame of the coding sequence by the one-step replacement method [62]. A 29-kb DNA fragment surrounding a lamprey HoxQ8 gene was cloned into pClasper by in vitro subcloning and the LacZ-URA3 cassette was similarly inserted by the one-step replacement method [63]. In another study, A 100-kb DNA insert that contains regions of the zebrafish HoxB cluster was captured by homologous recombination, and in a second round of targeting, a LacZ-URA3 was inserted into HoxA11B and expression was observed in transgenic mice [64]. Cis-acting elements capable of directing region-specific expression in transgenic embryos were detected with constructs generated by recombinational cloning. Using homologous recombination in bacteria, a IRES-lacZ gene was introduced into a 130-kb BAC containing the mouse zinc finger gene, RU49 [56]. In these

44

Shashikant and Ruddle

cases, long-range cis-regulatory elements not identifiable with conventional transgenic constructs were detected. Recombination-targeted cloning has thus made long-range cis-regulatory analysis more feasible. These simple methods of introducing reporter genes or altering specific sequences should not only identify cis-acting elements but also enable more difficult and challenging experiments. This includes elucidating interactions at multiple cis-acting elements, interactions between global regulatory sequences and gene-specific enhancers, and sharing of enhancers between two or more closely linked genes (reviewed in Ref. 46). A combination of large-insert cloning methods and recombination-targeted modification thus provides a basis for ensuring that the complete set of cis-acting sequences required for gene regulation will be identified. What is missing is a rapid method for generating deletion constructs, so that cis-acting sequences can be limited to smaller stretches of sequences. Since transcription factors typically interact with short stretches of DNA (6–18 bp length), it is critical to identify such sequences by a combination of deletion and mutation analyses. One possibility is to adopt a transposon-based deletion strategy [65, 66]. In this method, a transposon can be designed to integrate randomly in the insert but can be resolved with specific selection such that the resolved product contains specific regions of the insert. By examining several clones, it is possible to arrange constructs, which contain progressive, directional deletions in the insert. Application of this principle to transgenic studies will fill a major gap in regulatory analysis. In summary, the newer technological innovations outlined above will make identification of regulatory elements an attractive proposition. A bank of regulatory elements thus identified will find many applications, including creation of tissue-specific gene knockouts described below.

VII. Overexpression and Misexpression Two types of gene expression are achieved in transgenic mice; one in which the transgene is expressed with its own regulatory region, thereby creating an altered level of expression within its normal spatiotemporal domain and one, in which the expression of the transgene is directed by heterologous promoter and enhancer elements. The expressed gene may be similar to that of the normal gene or carry mutations/alterations either in the coding or regulatory regions. There is a tremendous literature on such studies; impacting on every field of modern biology including developmental biology, neurobiology, and immunology. Models created by these studies have provided critical information on many diseases and abnormalities including cancer and neurological diseases. Many studies have examined the ability of genes to be expressed normally in the mouse. As discussed in the preceding sections, with the advent of largeinsert cloning vehicles, it has been possible to achieve accurate expression of the

Transgenic Mouse Models

45

transgene by including larger fragments of DNA. These experiments essentially provide information on the gene dosage effect in mouse development. For instance, trisomy in human chromosome 21 leads to a number of phenotypes associated with Down’s syndrome. Overexpression of several large fragments derived from chromosome 21 have resulted in dosage imbalance and many characteristic phenotypes associated with the Down’s syndrome [67]. Using BAC clones carrying the Zipro1 gene, different dosages of the transgene have been correlated with different aspects of the Zipro1 function in cerebellar and skin development [68]. A second application of this strategy is to complement known genetic disorders or null mutations caused by gene knockout strategies. With a growing trend for large-scale mutagenesis strategies in mice ranging from chemical treatment to gene trap and chromosome engineering, the repertoire of mouse mutants are likely to increase dramatically in the next decade. Strategies aimed at complementing these mutations with large-insert transgenics using YAC and BAC clones have emerged [69]. A single BAC clone containing the Clock gene can restore circadian rhythm to mice with mutated Clock [70]. A third example of this strategy is to examine the significance of naturally occurring variations either in the regulatory or coding regions of the gene and find correlation with the disease state. Transgenic mice carrying mutations in the coding region of the βadrenergic receptor gene exhibit many features of cardiac dysfunction [71]. With the advent of pharmacogenomics and the concept of documenting variations in the human population, transgenic mice models provide an excellent means of assaying functional significance of such variations. A fourth example of this strategy is to use mice as bioreactors to produce proteins of pharmacological importance. For instance, with YAC transgenics, it is possible to generate human antibodies in mice [72,73]. The expression of a gene at ectopic sites is achieved by placing a gene of interest under the regulatory control of a heterologous promoter or enhancer. In many instances, expression of the transgene in tissues where the endogenous gene is normally not expressed results in interference with normal development or cell differentiation. These experiments provide insight on the role of genes in normal and abnormal processes. Altered expression of oncogenes, growth hormone genes, and homeobox genes, for example, have provided valuable information on growth, development, and differentiation processes. There is a considerable interest in using transgenic animals as bioreactors to produce proteins which retain their biological activities. Proteins produced by recombinant DNA technology with bacterial or baculoviral expression systems are unlikely to possess mammalian-specific protein modifications including glycosylation and phosphorylation. Transgenic mice in which gene expression can be directed by liver- or mammary gland–specific promoters can be used to generate microgram quantities of proteins with superior biological activities. Current strategies for functional genomic studies have resulted in designing methods to analyze gene expression

46

Shashikant and Ruddle

patterns using microarrays and microchips. However, these methods do not necessarily provide information on biological function of the gene. One could envisage the use of transgenic mice in conjunction with large-scale expression and phenotypic analyses to identify gene function.

VIII. Inducible Gene Expression Transgenic mice generated by the above strategies that use defined regulatory regions limit the ability to precisely control gene expression. Inducible systems, where gene expression can be turned on and off at different periods of mouse development, provide invaluable information on gene function. For instance, in Drosophila, the heat-inducible gene expression system has been used successfully to determine the role of developmental genes in patterning of the embryonic axis. Transgene expression is placed under the control of cis-acting elements that respond to inducers. Transgenic mice thus generated are treated with an inducer, resulting in controlled spatial and temporal gene expression in the animals. A major problem with this strategy is that there are no perfectly controllable inducible systems in mice. First, the response elements generally lack specificity as they respond to other transcriptional activators besides their main inducers, thus resulting in leaky, uncontrolled expression. Second, the inducer may affect expression of endogenous genes other than the transgene, thus making it difficult to absolutely correlate the resulting phenotype to the induction of transgene expression. Very often the concentration at which the inducer is used may have deleterious effects on mice. Thus, it is challenging to design an inducer, which at low concentrations can result in dose-dependent induction of high levels of gene expression by interacting with specific cis-acting elements, which in turn are refractory to other trans-acting factors affecting mouse developmental and differentiation processes. Despite these difficulties, a variety of inducible gene expression systems are currently being used in transgenic mice (reviewed in Refs. 74 and 75). An early example of inducible transgenic mice is the metal-induced growth hormone mice [76]. In this experiment, the human growth hormone gene was placed under the control of a metallothionein promoter. Transgenic mice fed with zinc showed dramatic growth, thereby inspiring many with the power of transgenic technology. However, the metallothionein promoter also responds to other trans-acting factors, some of which are constitutively expressed, resulting in a basal level of expression, thus limiting its use. The interferon-inducible promoters show rapid and high-level induction of transgene expression [77]. However, besides leakiness of the promoter, interferons evoke diverse physiological responses in mice. Steroid response elements are among the best characterized cis-acting sequences. Accordingly, inducible systems based on glucocorticoid, estrogen, pro-

Transgenic Mouse Models

47

gesterone, and other ligands capable of interactions with steroid response elements have been used in transgenic mice studies (reviewed in Refs. 74 and 75). Recently, it has been shown that the mutant estrogen receptor (ER) responds to an antiestrogen drug, tamoxifen, instead of its natural ligand, 17β-estradiol [78, 79]. Fusion proteins containing the tamoxifen-binding domain of the ER remain inactive in the absence of ligand, presumably due to steric hindrance caused by its binding to heat-protein 90 (HSP90) protein. When treated with tamoxifen, heat shock protein dissociates, resulting in activation of the fusion protein. This strategy is currently finding more applications, especially in inducing Cre recombinase activity in transgenic mice (reviewed in Ref. 10). The inducible system based on the Drosophila molting hormone ecdysone appears to be capable of high-level induction and low basal activity in transgenic mice. For example, a fusion protein, VP16-ecdysone receptor, is activated by binding to the ligand muristerone, which combines with retinoid receptor (RXR) to form a heterodimer, which induces expression of the transgene placed under the control of ecdysone response elements [80,81]. Two bacterial inducible systems employed in transgenic studies are based on the lac and tet operons (reviewed in Refs. 74 and 75). The lac repressor (lacR) binds to the lac operator sequences and inactivates transcriptional activity of the promoter. The synthetic inducer, IPTG, relieves this repression, which can activiate transcription by several hundredfold. However, the complexity of lac operator regulation and the high level of IPTG required to induce the promoter makes this system rather unattractive. An inducible system based on tetracycline transactivation appears to be promising because of the low tetracycline concentration required, relieving repression mediated by tetR and the exquisite specificity with which tetR binds to tetO sequences. A VP16-tetR fusion protein (tTA) which functions as an activator is effective in the mammalian system. In the absence of tetracycline (or doxycyclin), tTA induces expression of the transgene placed under the control of tetO sequences, but in the presence of tetracycline inhibits transcription. A variant of this fusion protein, named reverse tTA, acts in the opposite manner. Both tTA- and rtTA-inducible systems are currently being used widely in transgenic systems (reviewed in Refs. 10, 74, and 75).

IX. Binary Systems To overcome the problem of establishing transgenic lines carrying genes that are potentially embryonically lethal, binary mouse systems were designed [82]. The binary system consists of two transgenic mouse lines. The first one contains a potentially lethal gene regulated by a normally inactive promoter that can be induced by a specific activator. The second mouse line contains the specific activator regulated by a tissue-specific or inducible promoter. The expression of the

48

Shashikant and Ruddle

specific activator in the second line should not affect the normal development of the mouse, presumably for lack of its target promoter in the mouse genome. When these two lines are crossed, the expression of the lethal gene is induced in those progenies, which inherit both transgenes. The feasibility of the binary system was first demonstrated using the viral activator VP16 and its responsive, immediate early promoter [82]. The responder line that carries Hoxc8 under the control of the immediate early promoter is crossed with an activator line that carries VP16 under the control of the Hoxc8 enhancer. When the responder and activator lines are crossed, Hoxc8 is activated in appropriate tissues, resulting in an embryonic phenotype of the axial skeleton [83]. A second binary system uses the GAL4/UAS system widely used in Drosophila [84,85]. In general, binary systems have been less popular owing to problems of toxicity, leakiness, and reproducible activation of responder transgene in every cell. One significant binary system that has found wider application utilizes Cre recombinase and loxP sites [10,45]. Cre recombinase can excise sequences flanked by loxP sites. At present, this appears to be a very useful system to generate tissue-specific alteration of gene expression. A transgenic line is established in which Cre recombinase activity is placed under the control of a tissue-specific promoter. A second line is established via ES cell technology in which sequences to be removed are flanked by two loxP (floxed) sites. The two mouse lines are crossed, and in the progenies inheriting Cre recombinase and the floxed gene, tissue-specific deletion occurs. To achieve tissue-specific activation of gene expression, the floxed sequences are designed to inactivate the gene. Recombinasemediated removal of these sequences brings the necessary sequences in contiguity for transcription and translation to occur. In contrast, for inactivation, loxP is placed in regions flanking coding regions of the gene. Removal of loxP and the intervening coding region by Cre recombinase results in gene inactivation. Transgenic mouse lines generated with various inducible and tissue-specific promoters form valuable resources, since these lines can be used to alter expression of many genes. Many interesting Cre recombinase mice are available, which include Cre recombinase expressed only in spermatocytes during the meiotic phase of cell division, in one-cell embryos, and in a few defined brain cells (see http://www.mshri.on.ca/nagy/Cre-pub.html on the web). The advantage of the binary system is that multiple crosses can be carried out to achieve different types of tissue-specific alterations in gene expression. This will provide an extensive database on gene function in multiple tissues at different developmental stages. However, the underlying limitation of all these transgenic strategies is the availability of well-defined promoter elements with distinct and unique specificity. X.

Loss of Expression

Loss of function experiments are carried out in transgenic mice by directing the expression of dominant negative variants, antisense, or ribozymes to specific tis-

Transgenic Mouse Models

49

sues at desired stages of development. Among these, dominant negative mutations have become a powerful alternative to gene knockout studies. The importance of dominant negative mutations was elegantly outlined by Herskowitz [86]. The basic strategy involves the inhibition of a wild-type gene product by an overproduced variant of the same product. The variant is designed to contain some functional subsets, but either missing or defective in other subsets of the domains of the parent protein. An overproduced variant forms nonfunctional complexes with the wild-type protein and some of its targets resulting in inhibition of a specific function. For example, dominant negative variants can be designed to inactivate recognition, conduction, or amplification in a signal transduction pathway. These mutants can probe various functions within a protein complex, thus exploring the relationships between each function [87,88]. Similarly, dominant negative variants can be designed to disrupt protein–protein interactions, DNA-binding, or transcriptional activation. Dominant negative, as the name suggests, is a dominant phenotype which can be studied in heterozygous conditions. Although dominant negative variants can be designed based on biochemical or structural information, how these variants function in a complex cellular environment is often unpredictable. Additional problems lie in identifying promoters that are functional in the same cells as the wild-type gene but achieve higher levels of variant expression. In spite of these limitations, transgenic mice derived from dominant negative variants have provided useful information on gene function. The use of antisense strategy to generate loss of function in transgenic mice has recently been reviewed [89]. Antisense strategy involves the design of complementary oligonucleotides that hybridize with the 5′ region of the mRNA, thereby inhibiting its translation. Although antisense methods have been successful in plants, fewer published reports are available in transgenic mice. It is believed that antisense expression must be at least 10- to 20-fold greater than the specific mRNA to inhibit translation. The use of heterologus promoters and design of transgenes which encode longer antisense RNA are believed to improve the efficiency of inactivating genes in transgenic mice. Ribozymes, sequence-specific endoribonucleases, have great potential to create loss of function by cleaving specific RNA molecules. Ribozymes have been successfully used to target and destroy specific viral and cellular RNAs. They are regarded as an important therapeutic tool as antiviral agents. Several studies have used ribozymes in transgenic mice to knock down the expression of specific target genes (reviewed in Ref. 90). Two types of ribozymes, hairpin and hammerhead ribozymes, which have specific nucleotide cleavage sites, have been employed in these studies. Although many in vitro and cell culture studies have shown that ribozymes are effective tools in reducing or eliminating gene expression, their application in transgenic mice are somewhat limited. Examples of ribozyme transgenics include β2-microglobulin, glucokinase, α-lactalbumin, growth hormone, amelogenin, and HIV-1, among others [91–94]. Lack of toxic-

50

Shashikant and Ruddle

ity and relatively unique specificity make ribozymes a very attractive tool for transgenesis. Transgenic models are perceived as rapid and inexpensive mouse models to study and evaluate ribozymes, which have a potential as gene therapeutic agents. Although transgenic methods described above are useful, a major approach for studying the loss of gene expression in mice is gene targeting in ES cells and transmission of the mutation through the mouse germline (reviewed in Ref. 10). Two important steps involved in gene knockout technology are (1) the development of a gene targeting method based on the principle of homologous recombination and (2) the demonstration that when ES cells are injected into a mouse blastocyst they contribute to various tissues including the germline. Strategies involved in designing the targeting construct have been extensively reviewed [10]. A commonly used method is to flank the neoR gene with homologous fragments derived from the gene to be targeted (Fig. 3). The targeting construct also contains a thymidine kinase (tk) gene linked to one of the homologous fragments. In the event of correct gene targeting by homologous recombination, the endogenous gene is disrupted by neoR, creating a null mutation and the tk gene is lost. The targeting construct is electroporated into ES cells and selected for the presence of neoR (positive selection) and the absence of tk (negative selection). ES cells in which one of the endogenous allele is disrupted is selected and injected into the blastocyst. The chimeric mice obtained are screened for germline transmission of the mutation. The phenotypic effect of mutation is studied in both

Figure 3 Schematic of a commonly used gene-targeting strategy in ES cells. Fragments A and B represent regions of homology involved in recombination. As a consequence of a correct targeting event, neo R gene is introduced into an endogenous locus and tk gene is lost.

Transgenic Mouse Models

51

heterozygous and homozygous mice. This seemingly intricate technology is now perfected in many laboratories and is subject to many innovations that allow simple inactivation of genes, changing the regulatory context of a gene, and creating deletions, inversions, and other chromosomal rearrangements [10]. A major problem with the gene knockout strategy is that often gene function in an adult cannot be studied because of early embryonic lethality. To a large extent, the binary system described earlier utilizing Cre/loxP provides a means of obtaining tissue- and stage-specific gene knockouts. The ever-growing list of mouse mutants generated by gene-targeting technology can be accessed on the web (http://www.jax.org/resources/documents/imr). XI. Insertional Mutagenesis Another method to identify developmental genes involves insertional mutagenesis. Random integration of transgenes occasionally results in disruption of an endogenous gene, resulting in the serendipitous identification of a novel gene. A consequence of random transgene integration is that it may disrupt an endogenous locus. Mutant phenotypes may become evident when transgenic mice are bred to homozygosity. The transgene provides a tag to identify fragments flanking the point of insertion. Using these end fragments, one can clone the genetic locus, and the normal gene that was disrupted in the transgenic mice can be identified. In many instances, however, transgene insertion caused gross rearrangements and large deletions at the site of integration, resulting in difficulties in the identification of the endogenous locus. The development of gene-trap strategies in ES cells has largely replaced transgene insertion as a general method to screen for novel genes in early mouse development [24,25]. A large number of gene-trapped ES cell clones, each containing a single mutation, are commercially available (see http://www.lexgen.com on the web). XII. Transgenic Mouse Models for the Study of Human Pathologies and Functional Genomics Transgenic technology, with the array of innovations outlined above, makes mice a powerful experimental organism to study problems relevant to human pathologies. No other experimental organism can be manipulated in as many different ways as a mouse can be. Together with vast resources of inbred lines, congenic strains, and knockout and transgenic lines, the mouse will continue to be a lead experimental organism for functional genomics. Since mouse physiology closely resembles that of the human, the mouse is the experimental organism of choice to generate models for human disorders and abnormalities. The importance of mice for the study of human pathologies has recently been reviewed [95,96].

52

Shashikant and Ruddle

There are a few problems associated with mouse models. One is the longer time required for obtaining meaningful results, as compared to in vitro studies; for instance, those based on profiling change of expression using microarrays. Second, working with mice requires an elaborate infrastructure and is very expensive. Third, some of the techniques outlined in this chapter are, by industrial standards, considered to be very inefficient. Several technological innovations are required to make mouse models more efficient. These include large-insert transgenics; better methods to manipulate transgene constructs; extensive characterization of tissue-specific promoters; an improved efficiency in achieving a good level of transgene expression in every founder line established; molecular characterization of novel methods to manipulate gene function; and extrapolation of these results at the cellular and organismic levels. As the human genome-sequencing project is nearing completion, the focus is shifting toward functional genomics. At present, functional genomics are associated with the microarray-based expression systems. This technology allows simultaneous documentation of the variation in the expression pattern of many genes in response to a particular physiological stimulus. Additionally, sophisticated sequence analysis programs can provide educated guesses on potential functional domains in proteins, thereby classifying genes involved in signaling pathways, transcription, transport, secretion, among others. However, further understanding of gene function should rely on misexpression, which can be studied at the cellular and organismic levels. Transgenic mouse models will continue to be useful tools for functional genomics. Acknowledgments The authors wish to acknowledge Drs. Daniel Seufert, Gina Pighetti, and Cindy McKinney for critical reading of the manuscript and Tannaz Amalsadvala and Brooke Scheremeta for help with figures. This work was supported by USDA Hatch Project 3745 to C.S.S. and NIH GM09966 to F.H.R. References 1. Hogan BB, R, Constantini F. and Lacey E.: Manipulating the Mouse Embryo: A Laboratory Manual. CSHL Press, Cold Spring Harbor, NY, 1994. 2. Gordon JW. Production of transgenic mice. Methods Enzymol 1993; 225:747–771. 3. Scangos G, Ruddle FH. Mechanisms and applications of DNA-mediated gene transfer in mammalian cells—a review. Gene 1981; 14:1–10. 4. Gordon JW, Scangos GA, Plotkin DJ, Barbosa JA, Ruddle FH: Genetic transformation of mouse embryos by microinjection of purified DNA. Proc Natl Acad Sci USA 1980; 77:7380–7384. 5. Gordon JW, Ruddle FH. Integration and stable germ line transmission of genes injected into mouse pronuclei. Science 1981; 214:1244–1246.

Transgenic Mouse Models

53

6. Doetschman T, Gregg RG, Maeda N, Hooper ML, Melton DW, Thompson S, Smithies O. Targetted correction of a mutant HPRT gene in mouse embryonic stem cells. Nature 1987; 330:576–578. 7. Thomas KR, Capecchi MR. Site-directed mutagenesis by gene targeting in mouse embryo-derived stem cells. Cell 1987; 51:503–512. 8. Schwartzberg PL, Goff SP, Robertson EJ. Germ-line transmission of a c-abl mutation produced by targeted gene disruption in ES cells. Science 1989; 246:799– 803. 9. Zijlstra M, Li E, Sajjadi F, Subramani S, Jaenisch R. Germ-line transmission of a disrupted beta 2-microglobulin gene produced by homologous recombination in embryonic stem cells. Nature 1989; 342:435–438. 10. Muller U. Ten years of gene targeting: targeted mouse mutants, from vector design to phenotype analysis. Mech Dev 1999; 82:3–21. 11. Wilmut I, Schnieke AE, McWhir J, Kind AJ, Campbell KH. Viable offspring derived from fetal and adult mammalian cells [see comments]. [Published erratum appears in Nature 1997 Mar 13;386(6621):200.] Nature 1997; 385:810–813. 12. Campbell KH, McWhir J, Ritchie WA, Wilmut I. Sheep cloned by nuclear transfer from a cultured cell line [see comments]. Nature 1996; 380:64–66. 13. Gurdon JB, Colman A. The future of cloning [news]. Nature 1999; 402:743–746. 14. Glick BaP, JJ: Molecular Biotechnology: Principles and Applications of Recombinant DNA. ASM Press, Washington, DC, 1998. 15. Jaenisch R. Germ line integration and Mendelian transmission of the exogenous Moloney leukemia virus. Proc Natl Acad Sci USA 1976; 73:1260–1264. 16. Jahner D, Haase K, Mulligan R, Jaenisch R. Insertion of the bacterial gpt gene into the germ line of mice by retroviral infection. Proc Natl Acad Sci USA 1985; 82: 6927–6931. 17. Mann R, Mulligan RC, Baltimore D. Construction of a retrovirus packaging mutant and its use to produce helper-free defective retrovirus. Cell 1983; 33:153–159. 18. van der Putten H, Botteri FM, Miller AD, Rosenfeld MG, Fan H, Evans RM, Verma IM. Efficient insertion of genes into the mouse germ line via retroviral vectors. Proc Natl Acad Sci USA 1985; 82:6148–6152. 19. Federspiel MJ, Hughes SH. Retroviral gene delivery. Methods Cell Biol 1997; 52: 179–214. 20. Sherman A, Dawson A, Mather C, Gilhooley H, Li Y, Mitchell R, Finnegan D, Sang H. Transposition of the Drosophila element mariner into the chicken germ line [see comments]. [Published erratum appears in Nat Biotechnol 1999 Jan;17(1):81.] Nat Biotechnol 1998; 16:1050–1053. 21. Muramatsu T, Mizutani Y, Ohmori Y, Okumura J. Comparison of three nonviral transfection methods for foreign gene expression in early chicken embryos in ovo. Biochem Biophys Res Commun 1997; 230:376–380. 22. Itasaki N, Bel-Vialar S, Krumlauf R. ‘Shocking’ developments in chick embryology: electroporation and in ovo gene expression. Nat Cell Biol 1999; 1:E203–E207. 23. Bronson SK, Plaehn EG, Kluckman KD, Hagaman JR, Maeda N, Smithies O. Single-copy transgenic mice with chosen-site integration [See comments]. Proc Natl Acad Sci USA 1996; 93:9067–9072. 24. Kitajima K, Takeuchi T. Mouse gene trap approach: identification of novel genes

54

25. 26. 27. 28. 29.

30.

31.

32.

33. 34. 35. 36. 37.

38. 39.

40.

41.

42.

Shashikant and Ruddle and characterization of their biological functions. Biochem Cell Biol 1998; 76:1029– 1037. Zambrowicz BP, Friedrich GA. Comprehensive mammalian genetics: history and future prospects of gene trapping in the mouse. Int J Dev Biol 1998; 42:1025–1036. Wakayama T, Tateno H, Mombaerts P, Yanagimachi R. Nuclear transfer into mouse zygotes. Nat Genet 2000; 24:108–109. Wakayama T, Yanagimachi R. Cloning of male mice from adult tail-tip cells [news]. Nat Genet 1999; 22:127–128. Wakayama T, Rodriguez I, Perry AC, Yanagimachi R, Mombaerts P. Mice cloned from embryonic stem cells. Proc Natl Acad Sci USA 1999; 96:14984–14989. Rideout WM, 3rd, Wakayama T, Wutz A, Eggan K, Jackson-Grusby L, Dausman J, Yanagimachi R, Jaenisch R. Generation of mice from wild-type and targeted ES cells by nuclear cloning. Nat Genet 2000; 24:109–110. Wakayama T, Perry AC, Zuccotti M, Johnson KR, Yanagimachi R. Full-term development of mice from enucleated oocytes injected with cumulus cell nuclei [see comments]. Nature 1998; 394:369–374. Perry AC, Wakayama T, Kishikawa H, Kasai T, Okabe M, Toyoda Y, Yanagimachi R. Mammalian transgenesis by intracytoplasmic sperm injection [see comments]. Science 1999; 284:1180–1183. Lavitrano M, Camaioni A, Fazio VM, Dolci S, Farace MG, Spadafora C. Sperm cells as vectors for introducing foreign DNA into eggs: genetic transformation of mice [see comments]. Cell 1989; 57:717–723. Brinster RL, Sandgren EP, Behringer RR, Palmiter RD. No simple solution for making transgenic mice [letter; Comment]. Cell 1989; 59:239–241. Pollock RA, Jay G, Bieberich CJ. Altering the boundaries of Hox3.1 expression: evidence for antipodal gene regulation. Cell 1992; 71:911–923. Jasin M, Moynahan ME, Richardson C. Targeted transgenesis [comment]. Proc Natl Acad Sci USA 1996; 93:8804–8808. Sun FL, Elgin SC. Putting boundaries on silence. Cell 1999; 99:459–462. Francastel C, Walters MC, Groudine M, Martin DI. A functional enhancer suppresses silencing of a transgene and prevents its localization close to centrometric heterochromatin. Cell 1999; 99:259–269. Cockell M, Gasser SM. Nuclear compartments and gene regulation. Curr Opin Genet Dev 1999; 9:199–205. Grosveld F, van Assendelft GB, Greaves DR, Kollias G. Position-independent, highlevel expression of the human beta-globin gene in transgenic mice. Cell 1987; 51: 975–985. Krempen K, Grotkopp D, Hall K, Bache A, Gillan A, Rippe RA, Brenner DA, Breindl M. Far upstream regulatory elements enhance position-independent and uterus-specific expression of the murine alpha1(I) collagen promoter in transgenic mice. Gene Expr 1999; 8:151–163. Blom van Assendelft G, Hanscombe O, Grosveld F, Greaves DR. The beta-globin dominant control region activates homologous and heterologous promoters in a tissue-specific manner. Cell 1989; 56:969–977. Shewchuk BM, Asa SL, Cooke NE, Liebhaber SA. Pit-1 binding sites at the somatotrope-specific DNase I hypersensitive sites I, II of the human growth hormone locus

Transgenic Mouse Models

43.

44.

45. 46. 47. 48.

49. 50.

51.

52. 53. 54. 55.

56.

57. 58. 59.

55

control region are essential for in vivo hGH-N gene activation. J Biol Chem 1999; 274:35725–35733. Talbot D, Descombes P, Schibler U. The 5′ flanking region of the rat LAP (C/EBP beta) gene can direct high-level, position-independent, copy number-dependent expression in multiple tissues in transgenic mice. Nucleic Acids Res 1994; 22:756– 766. Willoughby DA, Vilalta A, Oshima RG. An Alu element from the K18 gene confers position-independent expression in transgenic mice. J Biol Chem 2000; 275:759– 768. Sauer B. Inducible gene targeting in mice using the Cre/lox system. Methods 1998; 14:381–392. Shashikant CS, Carr JL, Bhargava J, Bentley KL, Ruddle FH. Recombinogenic targeting: a new approach to genomic analysis—a review. Gene 1998; 223:9–20. Burke DT, Carle GF, Olson MV. Cloning of large segments of exogenous DNA into yeast by means of artificial chromosome vectors. Science 1987; 236:806–812. Ioannou PA, Amemiya CT, Garnes J, Kroisel PM, Shizuya H, Chen C, Batzer MA, de Jong PJ. A new bacteriophage P1-derived vector for the propagation of large human DNA fragments. Nat Genet 1994; 6:84–89. Sternberg NL. Cloning high molecular weight DNA fragments by the bacteriophage P1 system. Trends Genet 1992; 8:11–16. Shizuya H, Birren B, Kim UJ, Mancino V, Slepak T, Tachiiri Y, Simon M. Cloning and stable maintenance of 300-kilobase-pair fragments of human DNA in Escherichia coli using an F-factor-based vector. Proc Natl Acad Sci USA 1992; 89:8794– 8797. Bradshaw MS, Bollekens JA, Ruddle FH. A new vector for recombination-based cloning of large DNA fragments from yeast artificial chromosomes. Nucleic Acids Res 1995; 23:4850–4856. Lamb BT, Gearhart JD. YAC transgenics and the study of genetics and human disease. Curr Opin Genet Dev 1995; 5:342–348. Peterson KR, Clegg CH, Li Q, Stamatoyannopoulos G. Production of transgenic mice with yeast artificial chromosomes. Trends Genet 1997; 13:61–66. Dabert P, Smith GR. Gene replacement with linear DNA fragments in wild-type Escherichia coli: enhancement by Chi sites. Genetics 1997; 145:877–889. Jessen JR, Meng A, McFarlane RJ, Paw BH, Zon LI, Smith GR, Lin S. Modification of bacterial artificial chromosomes through chi-stimulated homologous recombination and its application in zebrafish transgenesis. Proc Natl Acad Sci USA 1998; 95:5121–5126. Yang XW, Model P, Heintz N. Homologous recombination based modification in Escherichia coli and germline transmission in transgenic mice of a bacterial artificial chromosome [see comments]. Nat Biotechnol 1997; 15:859–865. Cui C, Wani MA, Wight D, Kopchick J, Stambrook PJ. Reporter genes in transgenic mice. Transgenic Res 1994; 3:182–194. Gilthorpe JD, Rigby PW. Reporter genes for the study of transcriptional regulation in transgenic mouse embryos. Methods Mol Biol 1999; 97:159–182. Ikawa M, Yamada S, Nakanishi T, Okabe M. ‘Green mice’ and their potential usage in biological research. FEBS Lett 1998; 430:83–87.

56

Shashikant and Ruddle

60. Li X, Wang W, Lufkin T. Dicistronic LacZ and alkaline phosphatase reporter constructs permit simultaneous histological analysis of expression from multiple transgenes. Biotechniques 1997; 23:874–878, 880, 882. 61. Kothary R, Clapoff S, Darling S, Perry MD, Moran LA, Rossant J. Inducible expression of an hsp68-lacZ hybrid gene in transgenic mice. Development 1989; 105:707– 714. 62. Bradshaw MS, Shashikant CS, Belting HG, Bollekens JA, Ruddle FH. A long-range regulatory element of Hoxc8 identified by using the pClasper vector. Proc Natl Acad Sci USA 1996; 93:2426–2430. 63. Carr JL, Shashikant CS, Bailey WJ, Ruddle FH. Molecular evolution of Hox gene regulation: cloning and transgenic analysis of the lamprey HoxQ8 gene. J Exp Zool 1998; 280:73–85. 64. Chiu CH, Amemiya CT, Carr JL, Bhargava J, Hwang JK, Shashikant CS, Ruddle FH, Wagner GP. A recombinogenic targeting method to modify large-inserts for cis-regulatory analysis in transgenic mice: construction and expression of a 100-kb, zebrafish Hoxa-11b-lacZ reporter gene. Dev Genes Evol 2000; 210:105–109. 65. Krishnan BR, Jamry I, Berg DE, Berg CM, Chaplin DD. Construction of a genomic DNA ’feature map’ by sequencing from nested deletions: application to the HLA class I region. Nucleic Acids Res 1995; 23:117–122. 66. LaBanca F, Krishnan BR, Chaplin DD, Berg DE, Berg CM. Restriction map of a 35-kb HLA fragment constructed by nested deletion ’drop-out’ mapping. Gene 1995; 164:335–339. 67. Smith DJ, Stevens ME, Sudanagunta SP, Bronson RT, Makhinson M, Watabe AM, O’Dell TJ, Fung J, Weier HU, Cheng JF, Rubin EM. Functional screening of 2 Mb of human chromosome 21q22.2 in transgenic mice implicates minibrain in learning defects associated with Down syndrome [see comments]. Nat Genet 1997; 16:28– 36. 68. Yang XW, Wynder C, Doughty ML, Heintz N. BAC-mediated gene-dosage analysis reveals a role for Zipro1 (Ru49/Zfp38) in progenitor cell proliferation in cerebellum and skin [see comments]. Nat Genet 1999; 22:327–335. 69. Magdaleno SM, Curran T. Gene dosage in mice—BAC to the future [news; comment]. Nat Genet 1999; 22:319–320. 70. Antoch MP, Song EJ, Chang AM, Vitaterna MH, Zhao Y, Wilsbacher LD, Sangoram AM, King DP, Pinto LH, Takahashi JS. Functional identification of the mouse circadian Clock gene by transgenic BAC rescue. Cell 1997; 89:655–667. 71. Turki J, Lorenz JN, Green SA, Donnelly ET, Jacinto M, Liggett SB. Myocardial signaling defects and impaired cardiac function of a human beta 2-adrenergic receptor polymorphism expressed in transgenic mice. Proc Natl Acad Sci USA 1996; 93: 10483–10488. 72. Bruggemann M, Taussig MJ. Production of human antibody repertoires in transgenic mice. Curr Opin Biotechnol 1997; 8:455–458. 73. Nicholson IC, Zou X, Popov AV, Cook GP, Corps EM, Humphries S, Ayling C, Goyenechea B, Xian J, Taussig MJ, Neuberger MS, Bruggemann M. Antibody repertoires of four- and five-feature translocus mice carrying human immunoglobulin heavy chain and kappa and lambda light chain yeast artificial chromosomes. J Immunol 1999; 163:6898–6906.

Transgenic Mouse Models

57

74. Gingrich JR, Roder J. Inducible gene expression in the nervous system of transgenic mice. Annu Rev Neurosci 1998; 21:377–405. 75. Saez E, No D, West A, Evans RM. Inducible gene expression in mammalian cells and transgenic mice. Curr Opin Biotechnol 1997; 8:608–616. 76. Palmiter RD, Brinster RL, Hammer RE, Trumbauer ME, Rosenfeld MG, Birnberg NC, Evans RM. Dramatic growth of mice that develop from eggs microinjected with metallothionein-growth hormone fusion genes. Nature 1982; 300:611–615. 77. Arnheiter H, Skuntz S, Noteborn M, Chang S, Meier E. Transgenic mice with intracellular immunity to influenza virus. Cell 1990; 62:51–61. 78. Feil R, Brocard J, Mascrez B, LeMeur M, Metzger D, Chambon P. Ligand-activated site-specific recombination in mice. Proc Natl Acad Sci USA 1996; 93:10887– 10890. 79. Zhang Y, Riesterer C, Ayrall AM, Sablitzky F, Littlewood TD, Reth M. Inducible site-directed recombination in mouse embryonic stem cells. Nucleic Acids Res 1996; 24:543–548. 80. No D, Yao TP, Evans RM. Ecdysone-inducible gene expression in mammalian cells and transgenic mice. Proc Natl Acad Sci USA 1996; 93:3346–3351. 81. Yao TP, Forman BM, Jiang Z, Cherbas L, Chen JD, McKeown M, Cherbas P, Evans RM. Functional ecdysone receptor is the product of EcR and Ultraspiracle genes. Nature 1993; 366:476–479. 82. Byrne GW, Ruddle FH. Multiplex gene regulation: a two-tiered approach to transgene regulation in transgenic mice. Proc Natl Acad Sci USA 1989; 86:5473– 5477. 83. Yueh YG, Gardner DP, Kappen C. Evidence for regulation of cartilage differentiation by the homeobox gene Hoxc-8. Proc Natl Acad Sci USA 1998; 95:9956–9961. 84. Rowitch DH, B S-J, Lee SM, Flax JD, Snyder EY, McMahon AP. Sonic hedgehog regulates proliferation and inhibits differentiation of CNS precursor cells. J Neurosci 1999; 19:8954–8965. 85. Ornitz DM, Moreadith RW, Leder P. Binary system for regulating transgene expression in mice: targeting int-2 gene expression with yeast GAL4/UAS control elements. Proc Natl Acad Sci USA 1991; 88:698–702. 86. Herskowitz I. Functional inactivation of genes by dominant negative mutations. Nature 1987; 329:219–222. 87. Perlmutter RM, Alberola-Ila J. The use of dominant-negative mutations to elucidate signal transduction pathways in lymphocytes. Curr Opin Immunol 1996; 8:285– 290. 88. Yuen PS. Dominant negative mutants of guanylyl cyclase: probes for global functions and intramolecular mechanisms. Methods 1999; 19:532–544. 89. Erickson RP. Antisense transgenics in animals. Methods 1999; 18:304–310. 90. Welch PJ, Barber JR, Wong-Staal F. Expression of ribozymes in gene transfer systems to modulate target RNA levels. Curr Opin Biotechnol 1998; 9:486–496. 91. L’Huillier PJ, Soulier S, Stinnakre MG, Lepourry L, Davis SR, Mercier JC, Vilotte JL. Efficient and specific ribozyme-mediated reduction of bovine alpha-lactalbumin expression in double transgenic mice. Proc Natl Acad Sci USA 1996; 93:6698– 6703. 92. Luyckx VA, Leclercq B, Dowland LK, Yu AS. Diet-dependent hypercalciuria in

58

93.

94.

95. 96.

Shashikant and Ruddle transgenic mice with reduced CLC5 chloride channel expression. Proc Natl Acad Sci USA 1999; 96:12174–12179. Alami R, Gilman JG, Feng YQ, Marmorato A, Rochlin I, Suzuka SM, Fabry ME, Nagel RL, Bouhassira EE. Anti-beta s-ribozyme reduces beta s mRNA levels in transgenic mice: potential application to the gene therapy of sickle cell anemia. Blood Cells Mol Dis 1999; 25:110–119. Andang M, Hinkula J, Hotchkiss G, Larsson S, Britton S, Wong-Staal F, Wahren B, Ahrlund-Richter L. Dose-response resistance to HIV-1/MuLV pseudotype virus ex vivo in a hairpin ribozyme transgenic mouse model. Proc Natl Acad Sci USA 1999; 96:12749–12753. Bedell MA, Largaespada DA, Jenkins NA, Copeland NG. Mouse models of human disease. Part II: recent progress and future directions. Genes Dev 1997; 11:11–43. Bedell MA, Jenkins NA, Copeland NG. Mouse models of human disease. Part I: techniques and resources for genetic analysis in mice. Genes Dev 1997; 11:1–10.

Part Two RESPIRATORY DEVELOPMENT AND TISSUE OXYGENATION

4 Genetic Models of Respiratory Tract Development From Invertebrates to Vertebrates

CLIFFORD W. BOGUE Yale University School of Medicine New Haven, Connecticut

I.

Introduction

The mammalian respiratory system is a complex organ comprising two parts— the trachea and the lung—which both develop from the foregut. The respiratory system has arisen relatively late in vertebrate evolution, and its primary function is gas exchange, a function critical to survival of the organism. The biological process and genetic pathways that influence lung development begin long before the lung buds form as ventral outpouching of the foregut. At embryonic day 6 (E6), just prior to onset of gastrulation, the embryonic portion of the developing embryo is composed of two epithelial layers—the epiblast, which forms the entire embryo, and the visceral endoderm (or hypoblast), which is replaced by definitive gut endoderm during gastrulation and eventually contributes only to the extraembryonic yolk sac [1,2]. Gastrulation begins at E6.5 with the formation of the primitive streak, with epiblast cells on one side undergoing an epithelial–mesenchymal transformation to generate mesoderm. With the formation of the primitive streak on the posterior side of the embryo, the anteroposterior (AP) axis of the embryo becomes morphologically obvious. Over the ensuing 24 h, the streak elongates and moves along the rim of the egg cylinder toward its distal tip. It is here, at the anterior end of the primitive streak, that a specialized structure known 59

60

Bogue

as the node forms. The node generates axial mesendoderm, which includes the mesoderm that contributes to the prechordal plate and notochord as well as the definitive gut endoderm. The posterior end of the primitive streak gives rise to extraembryonic mesoderm, and the lateral plate and paraxial mesoderm arise from the intervening regions of the streak. The definitive endoderm moves anteriorly, replacing the anterior visceral endoderm (AVE) and eventually gives rise to the endoderm of the foregut and midgut. The foregut begins as a ventral folding of endodermal tissue at the anterior end of the embryo. As the foregut begins to close off, several tissue buds begin to form and later become crucial parenchymal organs such as the thyroid, thymus, lung, liver, and pancreas. Therefore, proper formation of the foregut is a crucial step in lung development. Interestingly, relatively little is known about endodermal development compared to the knowledge that has accumulated regarding the mechanisms involved in mesodermal and ectodermal formation. In the mouse, the trachea and lungs develop from the foregut at E9.5, a stage when the foregut is a single tube of epithelial endoderm surrounded by splanchnic mesoderm. Lung and tracheal development are first recognized at this stage as a ventral epithelial outpouching of the foregut comprising two parts— the future trachea and two primordial lung buds. Interestingly, these two parts of the mammalian respiratory system are formed and develop in completely different ways. Namely, the trachea develops by the division of the foregut into two tubes—the esophagus (dorsal) and the trachea (ventral)—by a longitudinal septum. In contrast, the lung, which includes the two main bronchi, subsequent distal bronchial branches, and alveoli, is formed primarily by branching morphogenesis of the two primary lung buds that are seen forming from the ventral foregut (excellently reviewed in Ref. 3). It is interesting to note that the early pattern of budding and branching during lung morphogenesis is asymmetrical and remarkably reproducible. This pattern of development, which is reproduced with high fidelity from animal to animal and even in lungs in culture, suggests that the budding lung has dorsal-ventral (D-V), medial-lateral (M-L) as well as left-right (L-R) identity. Genetic factors that control the orientation of these axes have only recently begun to be elucidated. For instance, the determination of LR identity involves a complex cascade of factors (see Chap. 10) [4] with Pitx2, which is expressed in the left lung bud but not the right [5], likely to be one key factor. One remarkable aspect of embryonic development that has become apparent in recent years is the conservation of patterning pathways and genetic interactions between invertebrates and vertebrates. The identification and characterization of large families of related genes that regulate gene expression in invertebrates, such as Caenorhabditis elegans and Drosophila, has contributed greatly to the elucidation of the mechanisms involved in vertebrate embryogenesis. The goal of this chapter is to review selected aspects of pharyngeal

Genetic Models of Respiratory Tract Development

61

development in C. elegans, foregut and tracheal development in Drosophila, and foregut and lung development in the mouse, focusing on the remarkable similarities of these processes at the genetic level. Hopefully, this overview will highlight the importance of taking a broad view of the many advances that have been made in developmental biology and searching for the common themes that may link one discovery to another. It is through utilizing this kind of approach that biologists will be in a position most fully to exploit the powerful genetic systems of worms and flies (and many other invertebrate and vertebrate species) to gain critical insight into the normal mechanisms of development involved in the genesis and formation of the mammalian respiratory system.

II. Pharynx and Endoderm Development in Caenorhabditis elegans The nematode pharynx is a large neuromuscular organ used for feeding, functioning to grind up food and deliver it to the intestine. Similar to the foregut compartments of much more complex animals, the C. elegans pharynx is composed of 80 cells that can be grouped into 5 cell types: muscles, glands, neurons, epithelia, and structural cells called marginal cells [6]. Two early blastomeres, ABa and MS, give rise to the pharynx (as well as other organs) [7]. Pharynx development from these two blastomeres involves two distinct developmental pathways involving maternal-effect genes and precise cell–cell interactions: an inductive pathway, in which the cells of the ABa lineage are induced to produce the anterior pharynx, and an autonomous pathway where the cells of the MS lineage give rise to the posterior pharynx (reviewed in Ref. 8). These two pathways, in which entire cell lineages and broad domains of axial patterning are controlled, involve maternally expressed genes such as glp-1 [9,10] and skn-1 [11,12]. However, the mechanisms that act downstream of these early specification events to mediate the development of this apparently undifferentiated pharynx primordium into a fully functioning and morphologically complex pharyngeal organ are only recently becoming elucidated. In C. elegans, the intestine is the only part of the gut that is considered solely endodermal, whereas the pharyngeal (anterior) and rectal (posterior) regions are considered both ectodermal and mesodermal. Thus, as is also seen in the fly, the gut is derived from all three germ layers [13], with the digestive component being the only part that is solely endoderm. This contrasts to the situation in vertebrates, where a larger portion of the gut is derived from the endoderm. Although there are differences in the morphogenesis and germ layer origin of the gut tube between invertebrates and vertebrates, there appear to be common biological mechanisms involved in the development of gut/endoderm and endoderm-derived organs.

62

Bogue A. pha-4

In 1994, the pha-4 gene was first shown to be necessary for development of the pharyngeal primordium [14]. In the pha-4 mutants, no pharyngeal primordium, and hence no pharynx, was formed during embryogenesis. Interestingly, the lack of a pharyngeal primordium in these mutants was not associated with alterations either in the identity of the EMS blastomere or in other aspects of anterior identity. Rather the specific cells affected by the pha-4 mutation were related to each other only by virtue of their fate to become pharynx and not by their position or cell lineage. These findings suggested that pha-4 functioned as an ‘‘organ-identity factor,’’ specifying which cells will form the pharynx, not specifying their particular cell type (i.e., muscle or nerve) [14]. Recently, two groups of investigators have demonstrated that pha-4 encodes a fork head/HNF-3 homolog that is expressed in all pharyngeal precursors and establishes their fate [15,16]. In keeping with its putative role as a organ-identity factor, ectopic expression of pha-4 led to ectopic pharynx formation, although this effect was partial. In addition, Kalb et al. [15] showed that PHA-4 protein expression in the pharynx persists in the adult and identified two potential targets of pha-4, the NK-2 homeobox genes ceh-22 and myo-2, genes involved in the development of the pharyngeal musculature. PHA-4 is also expressed at high levels in the rectum and at low levels in the midgut. There are at least three other cloned C. elegans genes that encode proteins containing the fork head/HNF-3 DNA binding domain—lin-31, pes-1, and daf-16 [17–20]. However, PHA-4 remains the protein most similar to Drosophila fork head ( fkh) [21]. Bolstering the proposition that pha-4, fkh, and HNF-3 genes perform homologous functions in worms, flies, and mice are the similarities between the domains of expression, phenotypes of the mutants, and the factors controlling their expression. Both C. elegans pha-4 and Drosophila fkh are highly expressed in comparable regions of their respective digestive tracts, with high levels of expression in the most anterior (pharynx/foregut) and posterior (rectum/hindgut) regions of the gut tube and lower levels of expression in the intervening gut proper [15,22]. Prior to organogenesis in the mouse, HNF-3β is expressed in the visceral and definitive endoderm, node, notochord, and floor plate [23–26]. In the gut endoderm, expression is seen in both the foregut and hindgut, like its homologs in C. elegans and Drosophila. Mutations in these genes in the three species show similar phenotypes. As mentioned above, mutations in C. elegans pha-4 show abnormal/absent specification of both pharyngeal and rectal precursors. In Drosophila fkh mutants, foregut and hindgut cells are transformed into head sclerite structures [22]. Likewise, in mice in which the HNF-3β gene has been deleted, severe abnormalities are seen in the development of the foregut [26,27], and this abnormality is due to specific loss of function in the definitive gut endoderm and not due to a defect in the visceral endoderm [28].

Genetic Models of Respiratory Tract Development

63

B. ceh-22

The NK-2 gene family is a phylogenetically conserved homeobox gene family with diverse function and expression patterns in during development. The family was initially identified by Kim and Nirenberg in Drosophila [29], and since their initial discovery, many additional family members have been cloned from a number of vertebrates and invertebrates (reviewed in Refs. 30 and 31). Recently, the C. elegans NK-2 homeobox gene ceh-22 was described and found to be exclusively expressed in the pharyngeal muscles [32]. The homeodomain of CEH-22 is 67% identical to the homeodomain of the murine NK-2 homeobox gene Nkx2.5, suggesting the possibility of conservation of function [32]. In the mouse, Nkx2.5 is expressed in early cardiac precursors [33], and it has been shown to be an important regulator of heart development [34]. Along with being expressed exclusively in the pharyngeal musculature, ceh-22 also binds the enhancer of the pharyngeal muscle–specific myo-2 gene [32], and ceh-22 null mutants display abnormalities in pharyngeal development and function [35]. It has been suggested that there is a relationship between pharyngeal development in nematodes and heart development in insects and vertebrates [30,35,36]. Supportive of this concept are data that show that mouse Nkx2.5 can partially rescue the phenotype seen in D. tinman mutants [31]. Specifically, mouse Nkx2.5 was able to rescue the visceral mesoderm but not the heart defect, indicating that the genetic functions of NK2 homeoproteins tinman and NKX-2.5 in cardiogenesis are not freely interchangeable between Drosophila and the mouse. However, when zebrafish Nkx2.5 was tested for its ability to substitute for ceh-22 in transgenic C. elegans, it was found that Nkx2.5 could indeed activate the expression of the ceh-22 target gene myo-2 and could rescue a ceh-22 mutant when expressed in the pharyngeal muscle [36]. These results bolster the contention that development of the vertebrate heart and nemotode pharynx share conserved genetic regulatory pathways. It is interesting to note that the conservation of this developmental pathway is probably involved in more than just cardiac muscle development. For instance, Nkx2.5, and the related genes Nkx2.3 and Nkx2.7 are expressed in the anterior endoderm, which gives rise to the anterior digestive tract as well as the vertebrate respiratory system, in addition to the lateral plate mesoderm, from which the heart develops [33,37–40]. Other members of the NK-2 family are known to be involved in the development of endodermally derived organs in mammals (see below). Therefore, some of the NK-2 genes may function across phyla in a conserved manner to direct the development of a number of anterior structures, including the anterior gut, respiratory system, and heart. C. end-1 and elt-2

The endoderm in nematodes is exclusively derived from the E blastomere, which is one of the six founder cells [7,41]. Two GATA factors, end-1 and elt-2, that

64

Bogue

are expressed in C. elegans and play important roles in endoderm formation. In 1997, Zhu et al. identified a single genomic region that was required for the production of the entire C. elegans endoderm [42]. In embryos lacking this region, the E cell fails to respond to its inductive signal and forms mesoderm and ectoderm rather than endoderm. end-1 was identified as a gene within this region that was able to restore the E cell’s ability to form endoderm. It is expressed when the E cell is first formed, and its expression persists until the division of the E daughter cells [42]. In addition, ectopic expression of END-1 in nonendodermal embryonic blastomeres results in their respecification to an endodermal fate [43]. A second GATA factor expressed in the developing endoderm, elt-2, has also been shown to be critical for gut development. elt-2 is expressed in a completely gut-specific manner at the 2E cell stage of gut development—a stage that is one cell cycle after the gut has become clonally established [44,45] and subsequent to the expression of end-1. elt-2 expression persists throughout endoderm development, and the lack of elt-2 results in major abnormalities in gut morphogenesis [45]. However, elt-2 mutant embryos express the gut-specific gene ges-1, indicating that elt-2 is crucial for terminal gut differentiation rather than initial gut specification. Misexpression of elt-2 does not result in ectopic gut formation but does

Figure 1 Genetic interactions in gut/endodermal development in C. elegans, Drosophila, and mouse. Presented is a simplified schematic of the genetic interactions, highlighting the similarities of the gut developmental programs that have been elucidated in the three organisms, especially with respect to the roles of the HNF 3/fork head, GATA, and Nkx families. See text for details.

Genetic Models of Respiratory Tract Development

65

induce ectopic gut gene expression. Thus, it appears that normal gut development requires the sequential expression of end-1 and elt-2. These genes are highly similar to the Drosophila GATA factor serpent and to the mouse GATA-4, -5, -6 factors in sequence, expression, and gene interactions, and they apparently function in endoderm and gut development (see below). Taken together with the data on the role of NK-2 and fkh genes in C. elegans pharynx and endoderm development, it is likely that these genes form an evolutionarily conserved molecular mechanism regulating endoderm and gut development that is at work in both invertebrates and vertebrates (Fig. 1). III. Drosophila Tracheal and Gut Development Development of the Drosophila tracheal, or respiratory, system involves the formation of a complex, highly ramifying network of thousands of branching epithelial tubes that conduct oxygen from the external openings called spiracules to the internal tissues. The pattern of branching is highly stereotypical and involves reciprocal interactions between the epithelium lining the branching tubes and the surrounding tissues (for excellent reviews, see Refs. 46 and 47). Recent genetic evidence suggests that the factors involved in the branching morphogenesis of the Drosophila trachea are quite similar to the factors involved in mouse lung development. The functional homologies that exist between the genes regulating Drosophila tracheal development and mouse lung development have provided important new insight into the mechanisms involved in the formation of the respiratory system. Likewise, Drosophila gut development has significant similarity to gut development in both worms and mice. This section will highlight some of the genes involved in Drosophila tracheal and gut development and will reveal similarities with genetic factors and mechanisms that are necessary for the development of the mouse foregut and lung. A. Tracheal Development

The Drosophila trachea arises from tracheal placodes that are laterally distributed in a segmental fashion on the larval surface. These placodes are established from the embryonic ectoderm before any branching occurs. The establishment of the tracheal placodes requires induction of trachealess, a basic helix–loop–helix (bHLH)–PAS domain transcription factor whose transcription is activated in the placodes 1–2 hr before tracheal sac formation occurs [48–50]. In addition, the formation of the tracheal placodes also involves positional cues from the activity of the genes decapentaplegic (dpp), wingless (wg), and hedgehog (hh) [51]. Trachealess forms a complex with the gene Tango, a bHLH-PAS protein with a broad expression pattern that is homologous to the mammalian gene ARNT [52,53], and presumably the interaction of this heterodimeric complex in the tracheal cells,

66

Bogue

Figure 2 Schematic of Drosophila tracheal development. A portion of an epithelial sac is shown sprouting on primary, two secondary, and many terminal branches. Mutations in different genes block or cause misregulation of the process at the indicated steps. TCF ⫽ ternary complex factor. Reprinted with permission from Metzger RJ, Krasnow RJ, Science 1999; 284(5420):1635–1639. Copyright 1999 American Association for the Advancement of Science (Ref. 46).

along with the expression of the POU-domain transcription factor ventral veinless (vv) [51] is necessary for sac formation and for preparing the tracheal sac cells for primary branching [48,50,52]. The subsequent branching of the tracheal system can be divided into three stages—primary, secondary, and terminal—and occurs, surprisingly, by changes in cell shape and by directed cell migration in the absence of cell proliferation [49]. Each stage is characterized by the expression of specific genes, as shown in Figure 2, which allow the entire process to be followed both at the anatomic as well as the molecular level. Primary tracheal branches are multicellular tubular structures that arise by cell migration. Secondary branches are unicellular tubes that are extensions of individual tracheal cells at the tips of primary branches. Terminal branches are cytoplasmic extensions that arise from secondary branch cells. The formation of the primary and secondary branches is fairly rigidly fixed, whereas the structure of the terminal branches is highly variable and is regulated by the oxygen need of the tissues [46]. The critical determinant of the branching pattern of the Drosophila trachea is the branchless (bnl) gene, a member of the FGF gene family [54]. The expression bnl is seen in groups of cells surrounding the tracheal sacs just prior to the time when primary branches form. The receptor for bnl is breathless (btl), a member of the FGF receptor family. Secreted bnl binds to the btl receptor on adjacent tracheal cells, resulting in a signaling cascade which guides the migration of the tracheal cells during the primary budding process [55–58]. Expression of bnl in tracheal development is dynamic, with expression levels dropping significantly as the primary branch moves toward a cluster of cells expressing bnl and the primary branch stops growing. Likewise, in other areas, when a new patch of cells begins expressing bnl, tracheal bud growth begins to migrate toward

Genetic Models of Respiratory Tract Development

67

the new patch of bnl-expressing cells. If bnl is ectopically expressed in novel sites, new tracheal branches form at those sites. Thus, bnl is responsible for determining the pattern of primary branches. It is apparent that cells in the tracheal placode that migrate in response to a nearby bnl signal receive genetic signals prior to migration that predetermine how they will behave. Two potential candidates for these signals are Dpp (tumor growth factor-β [TGF-β] superfamily) and spitz (epidermal growth factor [EGF] superfamily). The dpp gene is expressed in ectodermal cells adjacent to the tracheal placodes dorsally and ventrally and spitz is expressed within the placode itself. Evidence that the dpp and spitz signaling pathways control the migratory activity of the tracheal cells is evidenced by the fact that simultaneous inactivation of both genes results in a block in cell migration out of the tracheal placode [59]. In particular, dpp is required to control the dorsal expression of bnl and for the recruitment and migration of dorsal and ventral tracheal cells in dorsal and ventral directions [60]. Thus, once tracheal cells are initially defined, both dpp and spitz pathways are involved in specifying branch fates in the dorsal–ventral and anterior–posterior axes by determining the particular sites of bnl expression [56,61]. The formation of secondary and terminal branches, like the primary branch, is also controlled by bnl and btl, but there are additional factors and mechanisms involved. As the primary branches extend toward cells expressing bnl, the cells at the tips of the branch are exposed to locally higher levels of bnl, which induces the expression of secondary branch genes such as pointed, an ETS-domain transcription factor involved in secondary branch formation [62]. High levels of bnl also induce the gene sprouty, which inhibits the FGF pathway, thereby limiting the range of bnl action so that secondary branching occurs only in the cells closest to the bnl signaling center [63,64]. A null mutation of the sprouty gene results in a gain of function mutation whose phenotype results in ectopic branches that are induced on the stalks of primary branches [64]. Thus, the pattern of secondary branching is refined through the induction of factors that both promote and inhibit tracheal cell migration. The formation of the terminal branches is also primarily governed by the FGF pathway, with additional genes such as pruned, trimmed, and cropped being induced that serve to finely regulate the branching pattern [65]. Interestingly, recent evidence shows that hypoxia induces terminal branching by stimulating the expression of bnl, which functions as a chemoattractant guiding new branches to the bnl expressing cells [66]. B. Gut Development

In Drosophila, gut formation begins after the segmentation genes have determined the future body plan (reviewed in Ref. 67). Shortly after egg deposition, cells at the anterior (ventral) and posterior (dorsal) regions of the embryo invagi-

68

Bogue

nate to form the stomodeum and proctodeum, respectively. The stomodeum gives rise to the foregut, the proctodeum gives rise to the hindgut, and the midgut is derived from two primordia that abut the stomodeal and proctodeal invaginations and fuse to form a continuous gut tube. The fkh gene, a transcription factor with both worm (pha-4) and mouse (HNF-3/fork head) homologs, is expressed in the anlagen of the developing gut. A loss of function mutation in fkh results in lack of gut formation [22], indicating the critical role that f kh plays in gut development. Interestingly, all three regions of the gut are affected, indicating that the effect transcends germ layer origin. Target genes for f kh involved in mediating gut development are largely unknown. However, recently it has been shown that f kh activates the genes hh, wg, and dpp in signaling centers in the foregut and hindgut primordia, which results in regionalization of the gut canal and promotes formation of additional gut organs, such as the malpighian tubules and proventriculus [67]. In addition, the expression of hh in the developing gut epithelium is required for the subsequent expression of the homeobox gene bagpipe (bap), a gene required for visceral mesodermal development [67]. The development of the midgut endoderm requires the expression of the GATA factor homolog serpent (srp). Mutants that do not express srp, which is expressed in the midgut primordia present in the developing foregut and hindgut, lack the entire midgut and do not show evidence of endodermal differentiation [13]. Thus, it appears that normal midgut development requires the expression of both a fkh gene and GATA factor. These data suggest that there is a evolutionarily conserved genetic program involved in gut and endoderm development in the worm (see above), fly, and mouse (see below) which involves members of several gene families, including HNF/ fork head, GATA, dpp (TGF-β), hedgehog, and wingless (see Fig. 1).

IV. Mouse Foregut and Lung Development A. HNF-3/Fork Head Genes

The HNF-3 (hepatocyte nuclear factor 3) proteins were first discovered as factors abundantly expressed in liver [68,69] and functioned as transcription factors that regulated the expression of a host of gut-derived endodermal cells [70]. There are three HNF isoforms encoded by different genes (designated α, β, γ) and they have a high degree of sequence similarity in their DNA binding domain, known as the HNF-3/fork head, or winged helix domain [71]. Over the past several years, many members of this large family of genes have been identified in several species [72]. At the time of gastrulation (E6.5), HNF-3β is first expressed in the node at the anterior end of the primitive streak, and it is seen shortly thereafter in the definitive endoderm as well as the notochord and the neural floor plate [23– 25,73]. HNF-3β is capable of inducing the expression of shh during development

Genetic Models of Respiratory Tract Development

69

[74,75], and shh expression appears to maintain the expression of HNF-3β [76,77]. A crucial role for HNF-3β in foregut development has been established through gene-targeting experiments. Homozygous inactivation of HNF-3β leads to severe defects in foregut morphogenesis as well as abnormalities in notochord formation [26,27]. Recently, Dufort et al. demonstrated an intrinsic requirement for HNF-3β in foregut/midgut endodermal development in chimeric mice [28]. HNF-3β expression in the lung is confined to the pulmonary epithelium during development and is also seen in fully differentiated bronchiolar–alveolar epithelium in adult lung [23,78]. In vitro studies have revealed that several epitheliumspecific genes are regulated by HNF-3β, including Nkx2.1 [79], human surfactant protein B (SP-B) [80,81], and Clara cell secretory protein (CCSP, or CC10) [82– 84]. HNF-3α is also expressed in the foregut endoderm and later in the pulmonary epithelium, and inactivation of this gene results in defects in pancreatic function [82,85]. HNF-3γ is expressed in the gut endoderm after the formation of the gut, and inactivation of the gene does not disrupt lung function and is not lethal [86]. Recently, the related family member HFH-4 gene has been shown to have a role in lung development as well. HFH-4 is expressed in ciliated epithelial cells in the developing lung [87–89]. Overexpression of HFH-4 in the distal respiratory epithelium leads to formation of columnar epithelium where only cuboidal epithelium normally is present [90], and homozygous inactivation of HFH-4 results in mice that lack cilia and have situs inversus [91]. B. GATA Family

The GATA family of transcription factors are a developmentally important group of genes that are nuclear proteins that contain one or two distinct zinc finger domains that bind DNA [92]. The mouse family of GATA factors currently consists of six members. GATA-1, -2, and -3 have important roles in vertebrate hematopoiesis as demonstrated by gene inactivation studies [93–96]. GATA-4, -5, and -6 were identified more recently and are expressed in several nonhematopoietic tissues, including the heart and foregut [97–100]. GATA-4 is expressed in the precardiac splanchnic mesoderm and adjacent foregut endoderm during the time of foregut invagination and heart tube formation, and homozygous inactivation of the gene results in severe defects of foregut and heart morphogenesis and absence of ventral endoderm [101,102]. Data obtained from mouse chimera studies [103], ectopic expression studies in frogs [104], and antisense oligonucleotide experiments in chicks [105] suggest that GATA-4 is not required for heart development and that, most likely, the roles of GATA-4, -5, and -6 are redundant and/ or complementary during cardiac morphogenesis. However, there appears to be an absolute requirement for GATA-4 in endoderm and foregut development [106,107], placing GATA-4 near the top of the hierarchy of genes involved in gut and endoderm formation. GATA-5 and -6 are expressed in the developing

70

Bogue

lung, whereas GATA-4 is not [98]. GATA-5 and -6 appear to be expressed in complementary patterns in the developing lung—GATA-5 is expressed in the pulmonary mesenchyme at E12.5 and later is found throughout the lung parenchyma, with highest levels being in the bronchial smooth muscle, and GATA-6 is found at high levels in the lung epithelium at E12.5 and at E18.5 is expressed in the lung parenchyma and in vascular smooth muscle cells [98]. GATA-6 has been shown, in vitro, to activate the promoter of the divergent homeobox gene Nkx2.1, suggesting a genetic interaction important in lung development [108]. Recent studies in cardiac gene expression and heart development reinforce the concept that GATA factors and Nkx genes interact in development. This interaction was shown to occur by Nkx2.5 and GATA-4 cooperatively regulating, through direct physical interaction, the transcription of target genes in the heart [109] and by direct activation of the GATA-6 cardiac enhancer by Nkx2.5 in transgenic mice [110]. C. Homeobox Genes

Homeobox genes are a family of transcription factors that are involved in axial patterning during embryogenesis [111]. The name of this gene family derives from a highly conserved 60–amino acid (180-bp) region referred to as the homeodomain. Since the description of the original members of this gene family almost 15 years ago, evidence has accumulated that they function as transcription factors. The Hox genes are a murine subfamily of homeobox genes containing over 30 members that are located in four clusters, each on a separate chromosome (chromosomes 2, 6, 11, and 15). Characteristic of this subfamily is a homeodomain with a high degree of homology to that of Antennapedia (Antp), a Drosophila homeobox gene that controls leg versus antenna development [112]. In the developing mouse, the temporal and spatial pattern of expression of each Hox gene is determined by its position on the chromosome relative to the other genes in the Hox cluster—the 5′ genes are expressed later and more posteriorly, whereas the 3′ genes are expressed earlier and more anteriorly [111]. Pattern formation appears to be governed, in large part, by a network of these transcription factors, as evidenced by the fact that they have both self- and cross-regulatory properties [113]. A number of Hox genes are known to be expressed in the developing lung, with the Hoxa and Hoxb clusters being particularly well represented in surveys of pulmonary Hox gene expression that have been performed [114–120]. Data examining the expression of Hoxb-2, b-3, b-4, and b-5 in the foregut at the time of lung budding (E9.5) showed that Hoxb-2 was expressed most proximally and b-3, b-4, and b-5 were expressed progressively more distally in the foregut mesenchyme [118]. This pattern of expression is colinear with a gene’s position on the chromosome and is consistent with the general pattern of Hox gene expression in the developing central nervous system where genes on the 3′ end of the cluster have spatial expression domains that are more anterior than genes that are more

Genetic Models of Respiratory Tract Development

71

5′ in the cluster. Interestingly, the lung buds form in the region of the foregut that corresponds to the anterior boundary of Hoxb-5 expression, which may suggest that Hoxb-5 helps determine the site of lung budding, similar to the role that it plays in determining the position of the limb buds along the A–P axis [121]. In later stages of development, Hoxa and Hoxb genes have dynamic patterns of expression in the lung, and several have been shown to be responsive to treatment with retinoic acid [119,122] as well as dexamethasone and growth factors [123]. Yet the disruption of only two Hox genes has been associated with a pulmonary phenotype—Hoxa-3 and Hoxa-5. Mice in which the Hoxa-3 gene was disrupted show, along with skeletal defects, abnormalities of the thyroid, parathyroids, thymus, heart, esophagus, and respiratory tract [124,125]. These animals appear to die at birth of respiratory failure and have abnormal tracheal epithelium as well as small tracheae and bronchi. Mice with a homozygous deletion of Hoxa-5 die of respiratory failure due to altered lung and tracheal morphogenesis, tracheal occlusion, and decreased surfactant protein production [126]. Interestingly, although Hoxa-5 expression is confined entirely to the mesenchyme of the lung, Hoxa-5 mutant animals have altered levels of expression of Nkx2.1, HNF-3β and N-myc, genes that are all expressed in the epithelium of the lung. This underscores the role of epithelial–mesenchymal interactions during lung development. Much more needs to be determined about the role of Hox genes in lung development. In addition to the Hox genes, there are a number of other homeobox-containing genes expressed in the developing mouse whose homeobox sequence is variably divergent from the Antp and Hox homeoboxes. These genes have been referred to as nonclustered, orphan, or divergent homeoboxes and usually have a more restricted spatial pattern of expression during embryogenesis than the Hox genes. Additionally, these genes are often located on different chromosomes than the Hox clusters. The gene Nkx2.1, also known as TTF-1 and T/ebp-1, is one such divergent homeobox gene expressed in the lung. Nkx2.1 is expressed in the endoderm of the ventral foregut at E9.5, as the foregut is undergoing septation into the ventral trachea and dorsal esophagus [127], as well as in the thyroid and discrete regions of the brain [128]. Mice with a null mutation in Nkx2.1 have no septation of the foregut into trachea and esophagus, similar to the anatomical findings of humans with tracheoesophageal fistula. In addition, the lungs form the two main bronchi, but fail to undergo subsequent branching morphogenesis. The result is highly dilated sacs lined with columnar epithelium and no distal epithelium instead of well-formed lungs. The Nkx2.1 ⫺/⫺ phenotype is reminiscent of the shh ⫺/⫺ phenotype, as in both cases the foregut fails to septate normally and the lungs are present but dysplastic [129,130]. However, Nkx2.1 mutant lungs do not develop distal cuboidal epithelium and have markedly reduced levels of surfactant proteins and Bmp4, unlike shh mutants which do develop distal lung cells. Additionally, Nkx2.1 has been shown to interact with the promoters of a number of distal respiratory epithelial genes, such as SP-A, -B, and -C, and CC10, suggesting a role for

72

Bogue

Nkx2.1 in the proximal–distal patterning of the lung. The promoter of Nkx2.1 has been shown to be activated by both GATA-6 [108] and HNF-3β [79]. Other divergent homeobox genes expressed in the lung include Pitx2 and Hex. Pitx2 is a bicoid type homeobox gene that was first identified as the gene responsible for Reiger’s syndrome in humans [131]. Pitx2 is involved in determining the L–R asymmetry of the vertebrate body axis (see Chap. 10) and animals homozygous for a deletion of Pitx2 exhibit right pulmonary isomerism, a condition where two right lungs are present [132–135]. A more detailed analysis of lung development has not been performed in these animals as yet. Hex is a divergent homeobox gene expressed as early as E4.5 in the primitive endoderm which is later expressed asymmetrically in the anterior visceral endoderm, a tissue thought to be involved in patterning the anterior of the embyro [136]. At the time of gastrulation, Hex is expressed in the definitive endoderm emerging from the node and its expression persists in foregut endoderm and organs that develop from the foregut endoderm, including the thyroid, thymus, lung, liver, pancreas, and gallbladder [136–140]. In the lung, Hex is not expressed in the foregut in the region of lung budding prior to the onset of lung development. By E11.5, however, Hex is highly expressed in the endoderm of lung buds with lower levels in the mesenchyme (Figure 3). One day later (E12.5), distal endodermal expression remains high and mesenchymal expression increases. In the lung, Hex transcripts are seen at all embyronic ages and in the adult lung as well (Bogue, unpublished data). The specific function of Hex in lung development, as well as its function in other aspects of mouse development, remains unknown. However,

Figure 3 Hex expression in early lung development. Whole mount in situ hybridization of E11.5 and E12.5 mouse lungs with digoxigenin labelled Hex antisense RNA. At E11.5, Hex is most highly expressed in the lung endoderm, especially at the tips of developing lung buds (arrows). Hex is also expressed at a low level in the lung mesenchyme. At E12.5, Hex expression persists at high levels in the endoderm of distal lung buds (arrows) and expression is more prominent in the mesoderm than at E11.5.

Genetic Models of Respiratory Tract Development

73

recent data indicate that the promoter of the bile acid transporter ntcp is activated by Hex in vitro and Hex, in turn, is regulated both by HNF-3β and GATA-4 as well as members of the Sp family of transactivators [141,142]. This, as well as the data from the study of Nkx2.1 in the lung, is consistent with the common theme of HNF/forkhead, GATA factors, and homeobox genes interacting in a common genetic pathway in gut endoderm and lung development. D. shh and Gli Genes

Sonic hedgehog (shh) is a vertebrate homolog of Drosophila hedgehog (hh), a segment polarity gene that encodes a secreted signaling protein regulating dpp and wingless (wg) in target cells [143–146]. In Drosophila, patched (ptc), a tumor suppressor gene, encodes a large protein thought to be a HH receptor and the zinc finger gene cubitus interruptus (Ci) encodes a transcriptional mediator of HH signaling. In vertebrates, shh exists as a proprotein that is cleaved into two functional moieties—a 19-kD NH2 terminal petide that is modified by the addition of cholesterol and serves as a cell membrane anchor and a 26- to 28-kD COOH terminal peptide that freely diffuses between cells and is able to exert distant effects in a concentration-dependent manner (reviewed in Ref. 147). shh mediates its effects on adjacent cells by binding to the transmembrane receptor ptc, which in turn results in increased levels of the Gli genes, mammalian homologs of Ci. In the lung, shh is expressed in the epithelium throughout development and into adulthood. At E9.5, shh is prominently expressed in the tracheal diverticulum, an endodermal outpouching from the ventral foregut [129], and as lung development proceeds, expression is particularly high in the epithelium of the distal tips of the lung buds [148–150]. Not unexpectedly, perturbations in the levels of shh expression in the developing lung result in abnormalities of foregut and lung formation. Targeted overexpression of shh in the lung epithelium results in excessive mesoderm accumulation and a higher than normal rate of mesodermal cell proliferation, as well as increased levels of ptc expression [148]. The expression of ptc in developing lung is predominately mesenchymal and is seen most highly in the mesenchyme around distal lung buds adjacent to regions of highest shh expression [129,130,148]. This complementary expression pattern coupled with the fact that overexpression of shh in the epithelium (using the SPC promoter) results in effects that are predominately mesenchymal suggests that Shh protein secreted by the lung buds is signaling to the adjacent mesenchyme through a PTC-dependent pathway. Null mutations in shh have confirmed its role in foregut and lung development. Mice homozygous for a deletion of shh have abnormalities of septation and development of the esophagus and trachea as well as abnormalities in branching morphogenesis of the lung [129,130]. Several interesting observations have been made in the analysis of the shh null mutants. First, the abnormal division of the foregut into the anteriorly placed trachea and posteri-

74

Bogue

orly placed esophagus indicates that this developmental process is likely to be dependent on epithelial–mesenchymal interactions that require shh signaling. Second, although the lungs are quite abnormal, consisting of simple sacs lined with respiratory epithelia covered by a sparse, loose mesenchyme, they are present. Therefore, although shh is clearly necessary for normal branching of the lung to occur, it is not necessary for the initiation of lung budding from the foregut. Interestingly, the lungs of shh null mutants showed evidence of development of both proximal and distal respiratory epithelium, indicating that shh is not necessary for normal proximal–distal epithelial development to occur. As was observed in the SP-C–shh transgenics, the absence of shh expression in the lung preferentially affected the mesenchyme. In particular, mutant lungs exhibited enhanced cell death and decreased cell proliferation in the mesenchyme. Finally, analysis of the expression of two major shh effector genes, ptc and Gli, in shh mutant lungs supports the premise that shh is a signaling molecule that is secreted from the endoderm and induces gene expression and cell proliferation in the adjacent pulmonary mesenchyme. Predictably, the absence of shh resulted in decreased expression of ptc1, Gli1, and Gli3 in the mesoderm, whereas the expression of the endodermally expressed genes HNF-3β, FGFR-2, and Nkx 2.1 was not altered [129,130]. The vertebrate Gli gene family currently consists of three members, Gli, Gli2, and Gli3 that encode DNA binding proteins with five zinc fingers and are homologous to Drosophila Ci [151]. During development of the foregut and lung, all three Gli genes are expressed in the splanchnic and lung mesenchyme, with higher levels generally being seen in the distal rather than the proximal mesoderm [151,152]. The role of all three Gli genes in foregut and lung development has been examined by targeted deletion. Mice with a deletion of the Gli3 gene have the mildest lung phenotype consisting of localized size reductions and changes in the shape of the lobes [152]. A much more dramatic phenotype is seen in embryos with mutations in Gli2 and in embryos with compound mutations in both Gli2 and Gli3 [153]. At E11.5, Gli2 ⫺/⫺ animals develop only one lobe on the right instead of the usual four, whereas the left lung appears to branch normally. As development proceeds, both right and left lungs are hypoplastic, characterized by thick mesenchymal layers and small air sacs. BrdU incorporation studies suggest that the lung hypoplasia is due to a decrease in cell proliferation at later stages of lung development. Additionally, at later stages of development, both the trachea and esophagus are hypoplastic, indicating a more generalized effect on the foregut. The phenotype of Gli2 ⫺/⫺; Gli3 ⫹/⫺ lungs is even more severe, consisting of animals with esophageal atresia, tracheoesophageal fistulae, and a single-lobed lung in the midline. Two other foregut-derived organs, the pancreas and liver, appeared to bud normally from the foregut endoderm at E9.5. The most severe phenotype was seen in Gli2 ⫺/⫺; Gli3 ⫺/⫺ animals, which completely lack esophagus, trachea, or lungs. Other foregut derivatives such as the stomach, liver, pancreas, and thymus are reduced in size but

Genetic Models of Respiratory Tract Development

75

are present, and the expression of HNF-3β appears to be decreased. These results suggest that patterning of the foregut, in both the dorsal—ventral and proximal— distal axes, and the control of lung bud initiation involves both epithelial–mesenchymal interactions and the shh–ptc–Gli gene regulatory network. E. Bmp4

Bmp4 is a member of a large family of evolutionarily conserved signaling molecules that belongs to the TGF-β superfamily. The BMP family members are known to be involved in a number of embryonic patterning events (reviewed in Ref. 154) and mediate their effects through transmembrane serine–threonine kinase receptors that form heterodimers/tetramers of two different subunits [155]. Bmp4 has been placed into a subfamily with Bmp2 and Drosophila dpp based on a comparison of the sequence of the carboxy terminal mature region of the protein [154]. Transcripts for Bmp5, Bmp7, and Bmp4 have been detected in the developing lung. Bmp5 is expressed ubiquitously in the mesenchyme [156], whereas Bmp7 is expressed ubiquitously in the endoderm [157]. Bmp4 expression is dynamic, and there appear to be two distinct sites of expression during lung development. Prior to lung budding, Bmp4 is expressed in the ventral mesenchyme surrounding the foregut and expression in the ventral mesenchyme persists during the early stages of lung development (until approximately E11.5). Starting at E10.0, expression can be detected in the lung endoderm, and by E11.5, the endodermal expression is limited to the distal lung bud. Expression in the distal endoderm is maintained at least through E17.5 [158]. Thus, Bmp4 is a likely candidate for regulating lung branching. Unfortunately, Bmp4 ⫺/⫺ mice die by approximately E10, precluding the use of the null mutants in studying the role of Bmp4 in lung development [159,160]. In order to study the function of Bmp4 in lung development, Bellusci et al. overexpressed Bmp4 in transgenic mice under the control of the SP-C promoter, which targets gene expression to the distal respiratory epithelium [157]. They found that the lungs of SP-C–Bmp4 transgenic mice were abnormal, showing the formation of cystic terminal sacs and decreased epithelial proliferation. Weaver et al. recently showed that inhibition of Bmp4 signaling in transgenic mice expressing either the Bmp antagonist Xnoggin or a dominant negative Bmp receptor in the distal lung epithelium resulted in a marked reduction in distal epithelial cell types and an increase in proximal cell types in the distal part of the lung [158]. These data suggest that Bmp4 may play a role in lung bud formation from the foregut as well as in patterning the lung epithelium along the proximal–distal axis. F. Fibroblast Growth Factor Family

The fibroblast (FGF) family of genes includes at least 18 members that encode growth factors that regulate a number of different biological processes during development, including cellular proliferation, migration, and differentiation. Fi-

76

Bogue

broblast growth factors bind to their receptors (FGFR), which are known to be tyrosine kinase receptors. During embryogenesis, interactions between FGFs and FGFRs have important roles in mediating epithelial–mesenchymal interactions [161,162]. Several members of the FGF family and their receptors are known to be expressed in the developing lung and have been shown to be necessary for normal lung development to occur. At least 3 FGFs appear to have important roles in lung development and branching morphogenesis. A number of recent studies indicate that FGF-10 is the most critical FGF for branching morphogenesis of the lung. Early in lung development, Fgf10 expression is present in the mesenchyme of the primary lung buds at E9.75 [163]. As development proceeds, the expression of Fgf10 is dynamic, with highest levels being seen in the mesoderm where the endoderm is forming a bud or where bud outgrowth is continuing [163,164]. In fact, the expression domain of Fgf10 in the developing mesoderm appears to predetermine the site of endodermal budding, which is consistent with the observations that FGF-10 can induce budding from endodermal lung buds cultured in the absence of mesenchyme [163] and is a potent chemotactic factor that is able to direct the outgrowth of lung buds in whole lung organ culture [165]. Interestingly, during lung branching morphogenesis, there are regions of mesoderm lateral to the developing lung bud that continue to express low levels of Fgf10, suggesting there may be some factor or factors produced by the mesoderm that locally inhibit or downregulate Fgf10 expression, thereby controlling the sites of lateral branching. Gene ablation studies have unequivocally proved a role for FGF-10 in lung development. In Fgf10 ⫺/⫺ mice, septation of the foregut into the trachea and esophagus occurred but formation of lung buds did not occur [166,167]. The Fgf receptor Fgfr2IIIb, which binds both FGF-7 and FGF-10 with high affinity, is highly expressed in the lung endoderm at early stages of lung development [168,169]. Transgenic mice in which a dominant-negative form of the Fgfr2IIIb receptor was expressed in the lung epithelium under the control of the SP-C promoter or a soluble dominant-negative form of the receptor was ubiquitously expressed developed normal tracheae and main stem bronchi but did not develop secondary or terminal lung buds [168,170], which is consistent with FGF-10 playing a crucial role in lung budding. The role of FGF-7 in lung branching is less clear. FGF-7 promotes proliferation and branching in lung epithelium in culture [169,171], and overexpression in the lung epithelium in transgenic mice leads to marked pulmonary malformations with cystic lung formation [172]. However, Fgf7 ⫺/⫺ mice had no discernible lung malformations [173]. Recent data by Lebeche et al. suggest that FGF-7 levels are highly dynamic during lung morphogenesis and that FGF-7 is most likely part of a complex regulatory pathway involving the genes shh, ptc, and Fgf1 [174]. Recently, mammalian homologs of the Drosophila bnl inhibitor sprouty have been identified and shown to be present in the developing lung [175,176]. Inhibition of mSpry-2, using antisense

Genetic Models of Respiratory Tract Development

77

Figure 4 Two models for patterning of bud site in the early mouse lung. Schematic representation of the generation of three secondary buds (1, 2, and 3) on the dorsal (D), ventral (V), and lateral sides, respectively, of a hypothetical right primary bud as is grows along the proximodistal axis (dotted line in A1). Model A, on the left, predicts that the domains of Fgf10 expression (dotted circles in A1) are ‘‘hardwired’’ or predetermined at the time of formation of the first lung bud (A1). As the lung grows, buds begin to grow at sites of increased Fgf10 expression, with the sites of lung budding retaining their relative positions along the three axes of the lung buds. The tips of lung buds express high levels of Bmp4 and shh, which control and refine the budding process. In the second, or ‘‘generative,’’ model (B), lateral inhibitory mechanisms emanating from the distal tip endoderm of both primary and secondary buds play a major role in determining the sites of budding. As the primary bud grows (B1 to B2), lateral inhibition is first relieved in the region where bud 1 will form, then bud 2, bud 3 and so on. As the buds form in response to low levels of Fgf10 in the mesoderm, they produce factors that upregulate Fgf10 in the surrounding mesoderm which inhibits the formation of lung buds in the vicinity. Reprinted with permission from Hogan BL. Morphogenesis. Cell 1999; 96(2):225–233 (Ref. 3).

78

Bogue

oligonucleotides in lungs in culture, results in a marked increase in lung branching as well as increased expression of the mRNAs for SP-A, -B, and -C, suggesting a conservation of function of the sprouty genes in inhibiting the Fgf family genes in both flies and mice. This conservation of function was demonstrated by Minowada et al. in limb development, where they showed through both gain and loss of function experiments that Fgf signaling induced the expression of spry gene expression in both the mouse and chick embryo and that spry expression subsequently antagonized Fgf function in the developing limb [177]. The dynamic spatial and temporal pattern of Fgf10 expression coupled with its role in lung budding is very similar to the dynamic expression pattern of bnl and its role in the development of the Drosophila tracheal system. Like bnl, Ffg10 is expressed in the endoderm, is necessary for branching morphogenesis, and misexpression results in ectopic branch formation. In addition, both bnl and Fgf10 appear to induce programs involved in airway branching after the formation of the initial bud [46]. For instance, the expression of Bmp4 in the mouse lung appears to be induced by FGF-10 [167,174]. It has been suggested recently that the lung endoderm acts as an active signaling center similar to the signaling seen in the growing tips of tracheal branches in Drosophila and in the developing limb bud [3,46,158,174]. In this model, Fgf10 and Bmp4 act in combination to coordinate branching morphogenesis and proximal–distal development of the lung epithelium. Factors such as shh and mouse sprouty function as inhibitors of branching to limit lateral branching to specific anatomical sites or specific distances from the primary branch. Hogan has proposed two separate models for lung morphogenesis that incorporate the roles of Fgf10, shh, and Bmp4 as major regulators of the process (Fig. 4). In one model, the sites of lung budding are genetically predetermined, or ‘‘hardwired’’ as in Drosophila tracheal development, and are related to the dorsal–ventral, proximal–distal, and medial–lateral axes of the foregut and primary lung buds. The second model is termed ‘‘generative’’ and is a dynamic process in which multiple factors expressed in both the endoderm and mesoderm interact to determine where lung budding and branching will occur and involves both stimulatory and inhibitory genetic interactions. Whether lung development follows one of these developmental models, a combination of both, or a totally different developmental program remains to be seen. However, much insight into human development has and will be gleaned from studying the development of a variety of organisms, such as worms, flies, and mice.

Acknowledgments The author would like to thank Tina Brueckner for helpful discussions. The work described in this manuscript was funded, in part, by a research grant from the American Lung Association and by a KO8 award from the NHLBI.

Genetic Models of Respiratory Tract Development

79

References 1. Lawson KA, Meneses JJ, Pedersen RA. Clonal analysis of epiblast fate during germ layer formation in the mouse embryo. Development 1991; 113(3):891–911. 2. Gardner RL, Rossant J. Investigation of the fate of 4–5 day post-coitum mouse inner cell mass cells by blastocyst injection. J Embryol Exp Morphol 1979; 52: 141–52. 3. Hogan BL. Morphogenesis. Cell 1999; 96(2):225–233. 4. Beddington RS, Robertson EJ. Axis development and early asymmetry in mammals. Cell 1999; 96(2):195–209. 5. Meno C, Shimono A, Saijoh Y, Yashiro K, Mochida K, Ohishi S, Noji S, Kondoh H, Hamada H. lefty-1 is required for left-right determination as a regulator of lefty2 and nodal. Cell 1998; 94(3):287–97. 6. Albertson DG, Thomson JN. The pharynx of Caenorhabditis elegans. Philos Trans R Soc Lond B Biol Sci 1976; 275(938):299–325. 7. Sulston JE, Schierenberg E, White JG, Thomson JN. The embryonic cell lineage of the nematode Caenorhabditis elegans. Dev Biol 1983; 100(1):64–119. 8. Schnabel R, Priess JR. Specification of cell fates in the early embryo. In: Riddle DL, Blumenthal T, Meyer BJ, Priess JR, eds. C. elegans II. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press; 1997, pp 361–382. 9. Priess JR, Thomson JN. Cellular interactions in early C. elegans embryos. Cell 1987; 48(2):241–50. 10. Priess JR, Schnabel H, Schnabel R. The glp-1 locus and cellular interactions in early C. elegans embryos. Cell 1987; 51(4):601–611. 11. Bowerman B, Draper BW, Mello CC, Priess JR. The maternal gene skn-1 encodes a protein that is distributed unequally in early C. elegans embryos. Cell 1993; 74(3): 443–452. 12. Bowerman B, Eaton BA, Priess JR. skn-1, a maternally expressed gene required to specify the fate of ventral blastomeres in the early C. elegans embryo. Cell 1992; 68(6):1061–1075. 13. Reuter R. The gene serpent has homeotic properties and specifies endoderm versus ectoderm within the Drosophila gut. Development 1994; 120(5):1123–1135. 14. Mango SE, Lambie EJ, Kimble J. The pha-4 gene is required to generate the pharyngeal primordium of Caenorhabditis elegans. Development 1994; 120(10):3019– 3031. 15. Kalb JM, Lau KK, Goszczynski B, Fukushige T, Moons D, Okkema PG, McGhee JD. pha-4 is Ce-fkh-1, a fork head/HNF-3alpha,beta,gamma homolog that functions in organogenesis of the C. elegans pharynx. Development 1998; 125(12): 2171–2180. 16. Horner MA, Quintin S, Domeier ME, Kimble J, Labouesse M, Mango SE. pha-4, an HNF-3 homolog, specifies pharyngeal organ identity in Caenorhabditis elegans. Genes Dev 1998; 12(13):1947-1952. 17. Lin K, Dorman JB, Rodan A, Kenyon C. daf-16: An HNF-3/forkhead family member that can function to double the life-span of Caenorhabditis elegans. Science 1997; 278(5341):1319–1322.

80

Bogue

18. Miller LM, Gallegos ME, Morisseau BA, Kim SK. lin-31, a Caenorhabditis elegans HNF-3/fork head transcription factor homolog, specifies three alternative cell fates in vulval development. Genes Dev 1993; 7(6):933–947. 19. Hope IA. PES-1 is expressed during early embryogenesis in Caenorhabditis elegans and has homology to the fork head family of transcription factors. Development 1994; 120(3):505–514. 20. Ogg S, Paradis S, Gottlieb S, Patterson GI, Lee L, Tissenbaum HA, Ruvkun G. The Fork head transcription factor DAF-16 transduces insulin-like metabolic and longevity signals in C. elegans. Nature 1997; 389(6654):994–999. 21. Azzaria M, Goszczynski B, Chung MA, Kalb JM, McGhee JD. A fork head/HNF3 homolog expressed in the pharynx and intestine of the Caenorhabditis elegans embryo. Dev Biol 1996; 178(2):289–303. 22. Weigel D, Jurgens G, Kuttner F, Seifert E, Jackle H. The homeotic gene fork head encodes a nuclear protein and is expressed in the terminal regions of the Drosophila embryo. Cell 1989; 57(4):645–658. 23. Monaghan AP, Kaestner KH, Grau E, Schutz G. Postimplantation expression patterns indicate a role for the mouse forkhead/HNF-3 alpha, beta and gamma genes in determination of the definitive endoderm, chordamesoderm and neuroectoderm. Development 1993; 119(3):567–578. 24. Sasaki H, Hogan BL. Differential expression of multiple fork head related genes during gastrulation and axial pattern formation in the mouse embryo. Development 1993; 118(1):47–59. 25. Ang SL, Wierda A, Wong D, Stevens KA, Cascio S, Rossant J, Zaret KS. The formation and maintenance of the definitive endoderm lineage in the mouse: involvement of HNF3/forkhead proteins. Development 1993; 119(4):1301–1315. 26. Ang SL, Rossant J. HNF-3 beta is essential for node and notochord formation in mouse development. Cell 1994; 78(4):561–574. 27. Weinstein DC, Ruiz i Altaba A, Chen WS, Hoodless P, Prezioso VR, Jessell TM, Darnell JE, Jr. The winged-helix transcription factor HNF-3 beta is required for notochord development in the mouse embryo. Cell 1994; 78(4):575–588. 28. Dufort D, Schwartz L, Harpal K, Rossant J. The transcription factor HNF3beta is required in visceral endoderm for normal primitive streak morphogenesis. Development 1998; 125(16):3015–3025. 29. Kim Y, Nirenberg M. Drosophila NK-homeobox genes. Proc Natl Acad Sci USA 1989; 86(20):7716–7720. 30. Harvey RP. NK-2 homeobox genes and heart development. Dev Biol 1996; 178(2): 203–216. 31. Ranganayakulu G, Elliott DA, Harvey RP, Olson EN. Divergent roles for NK-2 class homeobox genes in cardiogenesis in flies and mice. Development 1998; 125(16):3037–3048. 32. Okkema PG, Fire A. The Caenorhabditis elegans NK-2 class homeoprotein CEH22 is involved in combinatorial activation of gene expression in pharyngeal muscle. Development 1994; 120(8):2175–2186. 33. Lints TJ, Parsons LM, Hartley L, Lyons I, Harvey RP. Nkx-2.5: a novel murine homeobox gene expressed in early heart progenitor cells and their myogenic dscendants. 1993; 119:419–431.

Genetic Models of Respiratory Tract Development

81

34. Lyons I, Parsons LM, Hartley L, Li R, Andrews JE, Robb L, Harvey RP. Myogenic and morphogenetic defects in the heart tubes of murine embryos lacking the homeo box gene Nkx2-5. Genes Dev 1995; 9(13):1654–1666. 35. Okkema PG, Ha E, Haun C, Chen W, Fire A. The Caenorhabditis elegans NK-2 homeobox gene ceh-22 activates pharyngeal muscle gene expression in combination with pha-1 and is required for normal pharyngeal development. Development 1997; 124(20):3965–3973. 36. Haun C, Alexander J, Stainier DY, Okkema PG. Rescue of Caenorhabditis elegans pharyngeal development by a vertebrate heart specification gene. Proc Natl Acad Sci USA 1998; 95(9):5072–5075. 37. Chen JN, Fishman MC. Zebrafish tinman homolog demarcates the heart field and initiates myocardial differentiation. Development 1996; 122(12):3809–3816. 38. Lee KH, Xu Q, Breitbart RE. A new tinman-related gene, nkx2.7, anticipates the expression of nkx2.5 and nkx2.3 in zebrafish heart and pharyngeal endoderm. Dev Biol 1996; 180(2):722–731. 39. Schultheiss TM, Xydas S, Lassar AB. Induction of avian cardiac myogenesis by anterior endoderm. Development 1995; 121(12):4203–4214. 40. Tonissen KF, Drysdale TA, Lints TJ, Harvey RP, Krieg PA. XNkx-2.5, a Xenopus gene related to Nkx-2.5 and tinman: evidence for a conserved role in cardiac development. Dev Biol 1994; 162(1):325–328. 41. Deppe U, Schierenberg E, Cole T, Krieg C, Schmitt D, Yoder B, von Ehrenstein G. Cell lineages of the embryo of the nematode Caenorhabditis elegans. Proc Natl Acad Sci USA 1978; 75(1):376–380. 42. Zhu J, Hill RJ, Heid PJ, Fukuyama M, Sugimoto A, Priess JR, Rothman JH. end-1 encodes an apparent GATA factor that specifies the endoderm precursor in Caenorhabditis elegans embryos. Genes Dev 1997; 11(21):2883–2896. 43. Zhu J, Fukushige T, McGhee JD, Rothman JH. Reprogramming of early embryonic blastomeres into endodermal progenitors by a Caenorhabditis elegans GATA factor. Genes Dev 1998; 12(24):3809–3814. 44. Hawkins MG, McGhee JD. elt-2, a second GATA factor from the nematode Caenorhabditis elegans. J Biol Chem 1995; 270(24):14666–14671. 45. Fukushige T, Hawkins MG, McGhee JD. The GATA-factor elt-2 is essential for formation of the Caenorhabditis elegans intestine. Dev Biol 1998; 198(2):286–302. 46. Metzger RJ, Krasnow MA. Genetic control of branching morphogenesis. Science 1999; 284(5420):1635–1639. 47. Hogan BL, Yingling JM. Epithelial/mesenchymal interactions and branching morphogenesis of the lung. Curr Opin Genet Dev 1998; 8(4):481–486. 48. Isaac DD, Andrew DJ. Tubulogenesis in Drosophila: a requirement for the trachealess gene product. Genes Dev 1996; 10(1):103–117. 49. Samakovlis C, Hacohen N, Manning G, Sutherland DC, Guillemin K, Krasnow MA. Development of the Drosophila tracheal system occurs by a series of morphologically distinct but genetically coupled branching events. Development 1996; 122(5):1395–407. 50. Wilk R, Weizman I, Shilo BZ. trachealess encodes a bHLH-PAS protein that is an inducer of tracheal cell fates in Drosophila. Genes Dev 1996; 10(1):93–102. 51. Llimargas M, Casanova J. ventral veinless, a POU domain transcription factor,

82

52.

53.

54.

55.

56.

57.

58.

59.

60.

61. 62.

63.

64.

65. 66.

Bogue regulates different transduction pathways required for tracheal branching in Drosophila. Development 1997; 124(17):3273–3281. Ohshiro T, Saigo K. Transcriptional regulation of breathless FGF receptor gene by binding of TRACHEALESS/dARNT heterodimers to three central midline elements in Drosophila developing trachea. Development 1997; 124(20):3975– 3986. Sonnenfeld M, Ward M, Nystrom G, Mosher J, Stahl S, Crews S. The Drosophila tango gene encodes a bHLH-PAS protein that is orthologous to mammalian Arnt and controls CNS midline and tracheal development. Development 1997; 124(22): 4571–4582. Sutherland D, Samakovlis C, Krasnow MA. branchless encodes a Drosophila FGF homolog that controls tracheal cell migration and the pattern of branching. Cell 1996; 87(6):1091–1101. Glazer L, Shilo BZ. The Drosophila FGF-R homolog is expressed in the embryonic tracheal system and appears to be required for directed tracheal cell extension. Genes Dev 1991; 5(4):697–705. Shilo BZ, Gabay L, Glazer L, Reichman-Fried M, Wappner P, Wilk R, Zelzer E. Branching morphogenesis in the Drosophila tracheal system. Cold Spring Harb Symp Quant Biol 1997; 62:241–247. Klambt C, Glazer L, Shilo BZ. breathless, a Drosophila FGF receptor homolog, is essential for migration of tracheal and specific midline glial cells. Genes Dev 1992; 6(9):1668–78. Lee T, Hacohen N, Krasnow M, Montell DJ. Regulated Breathless receptor tyrosine kinase activity required to pattern cell migration and branching in the Drosophila tracheal system. Genes Dev 1996; 10(22):2912–2921. Wappner P, Gabay L, Shilo BZ. Interactions between the EGF receptor and DPP pathways establish distinct cell fates in the tracheal placodes. Development 1997; 124(22):4707–4716. Vincent S, Ruberte E, Grieder NC, Chen CK, Haerry T, Schuh R, Affolter M. DPP controls tracheal cell migration along the dorsoventral body axis of the Drosophila embryo. Development 1997; 124(14):2741–2750. Krasnow MA. Genes that control organ form: lessons from bone and branching morphogenesis. Cold Spring Harb Symp Quant Biol 1997; 62:235–240. Klambt C. The Drosophila gene pointed encodes two ETS-like proteins which are involved in the development of the midline glial cells. Development 1993; 117(1): 163–176. Kramer S, Okabe M, Hacohen N, Krasnow MA, Hiromi Y. Sprouty: a common antagonist of FGF and EGF signaling pathways in Drosophila. Development 1999; 126(11):2515–2525. Hacohen N, Kramer S, Sutherland D, Hiromi Y, Krasnow MA. Sprouty encodes a novel antagonist of FGF signaling that patterns apical branching of the Drosophila airways. Cell 1998; 92(2):253–263. Warburton D, Zhao J, Berberich MA, Bernfield M. Molecular embryology of the lung: then, now, and in the future. Am J Physiol 1999; 276(5 Pt 1):L697–704. Jarecki J, Johnson E, Krasnow MA. Oxygen regulation of airway branching in Drosophila is mediated by branchless FGF. Cell 1999; 99(2):211–220.

Genetic Models of Respiratory Tract Development

83

67. Hoch M, Pankratz MJ. Control of gut development by fork head and cell signaling molecules in Drosophila. Mech Dev 1996; 58(1–2):3–14. 68. Lai E, Prezioso VR, Smith E, Litvin O, Costa RH, Darnell JE, Jr. HNF-3A, a hepatocyte-enriched transcription factor of novel structure is regulated transcriptionally. Genes Dev 1990; 4(8):1427–1436. 69. Lai E, Prezioso VR, Tao WF, Chen WS, Darnell JE Jr. Hepatocyte nuclear factor 3 alpha belongs to a gene family in mammals that is homologous to the Drosophila homeotic gene fork head. Genes Dev 1991; 5(3):416–427. 70. Zaret K. Developmental competence of the gut endoderm: genetic potentiation by GATA and HNF3/fork head proteins. Dev Biol 1999; 209(1):1–10. 71. Weigel D, Jackle H. The fork head domain: a novel DNA binding motif of eukaryotic transcription factors? [letter]. Cell 1990; 63(3):455–456. 72. Kaufmann E, Knochel W. Five years on the wings of fork head. Mech Dev 1996; 57(1):3–20. 73. Ruiz i Altaba A, Prezioso VR, Darnell JE, Jessell TM. Sequential expression of HNF-3 beta and HNF-3 alpha by embryonic organizing centers: the dorsal lip/ node, notochord and floor plate. Mech Dev 1993; 44(2–3):91–108. 74. Epstein DJ, McMahon AP, Joyner AL. Regionalization of Sonic hedgehog transcription along the anteroposterior axis of the mouse central nervous system is regulated by Hnf3-dependent and -independent mechanisms. Development 1999; 126(2):281–292. 75. McMahon AP. Mouse development. Winged-helix in axial patterning. Curr Biol 1994; 4(10):903–906. 76. Echelard Y, Epstein DJ, St-Jacques B, Shen L, Mohler J, McMahon JA, McMahon AP. Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity. Cell 1993; 75(7):1417–1430. 77. Sasaki H, Hogan BL. HNF-3 beta as a regulator of floor plate development. Cell 1994; 76(1):103–115. 78. Zhou L, Lim L, Costa RH, Whitsett JA. Thyroid transcription factor-1, hepatocyte nuclear factor-3beta, surfactant protein B, C, and Clara cell secretory protein in developing mouse lung. J Histochem Cytochem 1996; 44(10):1183–1193. 79. Ikeda K, Shaw-White JR, Wert SE, Whitsett JA. Hepatocyte nuclear factor 3 activates transcription of thyroid transcription factor 1 in respiratory epithelial cells. Mol Cell Biol 1996; 16(7):3626–3636. 80. Clevidence DE, Overdier DG, Peterson RS, Porcella A, Ye H, Paulson KE, Costa RH. Members of the HNF-3/forkhead family of transcription factors exhibit distinct cellular expression patterns in lung and regulate the surfactant protein B promoter. Dev Biol 1994; 166(1):195–209. 81. Bohinski RJ, Di Lauro R, Whitsett JA. The lung-specific surfactant protein B gene promoter is a target for thyroid transcription factor 1 and hepatocyte nuclear factor 3, indicating common factors for organ-specific gene expression along the foregut axis. Mol Cell Biol 1994; 14(9):5671–5681. 82. Bingle CD, Hackett BP, Moxley M, Longmore W, Gitlin JD. Role of hepatocyte nuclear factor-3 alpha and hepatocyte nuclear factor-3 beta in Clara cell secretory protein gene expression in the bronchiolar epithelium. Biochem J 1995; 308(Pt 1):197–202. 83. Sawaya PL, Stripp BR, Whitsett JA, Luse DS. The lung-specific CC10 gene is

84

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

Bogue regulated by transcription factors from the AP-1, octamer, and hepatocyte nuclear factor 3 families. Mol Cell Biol 1993; 13(7):3860–3871. Sawaya PL, Luse DS. Two members of the HNF-3 family have opposite effects on a lung transcriptional element; HNF-3 alpha stimulates and HNF-3 beta inhibits activity of region I from the Clara cell secretory protein (CCSP) promoter. J Biol Chem 1994; 269(35):22211–22216. Kaestner KH, Katz J, Liu Y, Drucker DJ, Schutz G. Inactivation of the winged helix transcription factor HNF-3 alpha affects glucose homeostasis and islet glucagon gene expression in vivo. Genes Dev 1999; 13(4):495–504. Kaestner KH, Hiemisch H, Schutz G. Targeted disruption of the gene encoding hepatocyte nuclear factor 3 gamma results in reduced transcription of hepatocytespecific genes. Mol Cell Biol 1998; 18(7):4245–4251. Pelletier GJ, Brody SL, Liapis H, White RA, Hackett BP. A human forkhead/ winged-helix transcription factor expressed in developing pulmonary and renal epithelium. Am J Physiol 1998; 274(3 Pt 1):L351–359. Tichelaar JW, Wert SE, Costa RH, Kimura S, Whitsett JA. HNF-3/forkhead homologue-4 (HFH-4) is expressed in ciliated epithelial cells in the developing mouse lung. J Histochem Cytochem 1999; 47(6):823–832. Blatt EN, Yan XH, Wuerffel MK, Hamilos DL, Brody SL. Forkhead transcription factor HFH-4 expression is temporally related to ciliogenesis. Am J Respir Cell Mol Biol 1999; 21(2):168–76. Tichelaar JW, Lim L, Costa RH, Whitsett JA. HNF-3/forkhead homologue-4 influences lung morphogenesis and respiratory epithelial cell differentiation in vivo. Dev Biol 1999; 213(2):405–417. Chen J, Knowles HJ, Hebert JL, Hackett BP. Mutation of the mouse hepatocyte nuclear factor/forkhead homologue 4 gene results in an absence of cilia and random left-right asymmetry. J Clin Invest 1998; 102(6):1077–1082. Omichinski JG, Clore GM, Schaad O, Felsenfeld G, Trainor C, Appella E, Stahl SJ, Gronenborn AM. NMR structure of a specific DNA complex of Zn-containing DNA binding domain of GATA-1. Science 1993; 261(5120):438–446. Pevny L, Lin CS, D’Agati V, Simon MC, Orkin SH, Costantini F. Development of hematopoietic cells lacking transcription factor GATA-1. Development 1995; 121(1):163–172. Pandolfi PP, Roth ME, Karis A, Leonard MW, Dzierzak E, Grosveld FG, Engel JD, Lindenbaum MH. Targeted disruption of the GATA-3 gene causes severe abnormalities in the nervous system and in fetal liver haematopoiesis. Nat Genet 1995; 11(1):40–44. Tsai FY, Keller G, Kuo FC, Weiss M, Chen J, Rosenblatt M, Alt FW, Orkin SH. An early haematopoietic defect in mice lacking the transcription factor GATA-2. Nature 1994; 371(6494):221–226. Fujiwara Y, Browne CP, Cunniff K, Goff SC, Orkin SH. Arrested development of embryonic red cell precursors in mouse embryos lacking transcription factor GATA-1. Proc Natl Acad Sci USA 1996; 93(22):12355–12358. Morrisey EE, Ip HS, Lu MM, Parmacek MS. GATA-6: a zinc finger transcription factor that is expressed in multiple cell lineages derived from lateral mesoderm. Dev Biol 1996; 177(1):309–322.

Genetic Models of Respiratory Tract Development

85

98. Morrisey EE, Ip HS, Tang Z, Lu MM, Parmacek MS. GATA-5: a transcriptional activator expressed in a novel temporally and spatially-restricted pattern during embryonic development. Dev Biol 1997; 183(1):21–36. 99. Arceci RJ, King AA, Simon MC, Orkin SH, Wilson DB. Mouse GATA-4: a retinoic acid-inducible GATA-binding transcription factor expressed in endodermally derived tissues and heart. Mol Cell Biol 1993; 13(4):2235–2246. 100. Laverriere AC, MacNeill C, Mueller C, Poelmann RE, Burch JB, Evans T. GATA4/5/6, a subfamily of three transcription factors transcribed in developing heart and gut. J Biol Chem 1994; 269(37):23177–23184. 101. Kuo CT, Morrisey EE, Anandappa R, Sigrist K, Lu MM, Parmacek MS, Soudais C, Leiden JM. GATA-4 transcription factor is required for ventral morphogenesis and heart tube formation. Genes Dev 1997; 11(8):1048–1060. 102. Molkentin JD, Lin Q, Duncan SA, Olson EN. Requirement of the transcription factor GATA-4 for heart tube formation and ventral morphogenesis. Genes Dev 1997; 11(8):1061–1072. 103. Narita N, Bielinska M, Wilson DB. Cardiomyocyte differentiation by GATA-4– deficient embryonic stem cells. Development 1997; 124(19):3755–3764. 104. Jiang Y, Evans T. The Xenopus GATA-4/5/6 genes are associated with cardiac specification and can regulate cardiac-specific transcription during embryogenesis. Dev Biol 1996; 174(2):258–270. 105. Jiang Y, Tarzami S, Burch JB, Evans T. Common role for each of the cGATA-4/ 5/6 genes in the regulation of cardiac morphogenesis. Dev Genet 1998; 22(3):263– 277. 106. Narita N, Bielinska M, Wilson DB. Wild-type endoderm abrogates the ventral developmental defects associated with GATA-4 deficiency in the mouse. Dev Biol 1997; 189(2):270–274. 107. Soudais C, Bielinska M, Heikinheimo M, MacArthur CA, Narita N, Saffitz JE, Simon MC, Leiden JM, Wilson DB. Targeted mutagenesis of the transcription factor GATA-4 gene in mouse embryonic stem cells disrupts visceral endoderm differentiation in vitro. Development 1995; 121(11):3877–3888. 108. Shaw-White JR, Bruno MD, Whitsett JA. GATA-6 activates transcription of thyroid transcription factor-1. J Biol Chem 1999; 274(5):2658–2664. 109. Durocher D, Charron F, Warren R, Schwartz RJ, Nemer M. The cardiac transcription factors Nkx2–5 and GATA-4 are mutual cofactors. EMBO J 1997; 16(18): 5687–5696. 110. Molkentin JD, Antos C, Mercer B, Taigen T, Miano JM, Olson EN. Direct Activation of a GATA-6 cardiac enhancer by Nkx2.5: evidence for a reinforcing regulatory network of Nkx2.5 and GATA transcription factors in the developing heart. Dev Biol 2000; 217(2):301–309. 111. McGinnis W, Krumlauf R. Homeobox genes and axial patterning. Cell 1992; 68: 283–302. 112. Graham A, Papalopulu N, Krumlauf R. The murine and drosophila homeobox gene complexes have common features of organization and expression. Cell 1989; 57: 367–378. 113. Affolter M, Schier A, Gehring WJ. Homeodomain proteins and the regulation of gene expression. Curr Opin Cell Biol 1990; 2:485–495.

86

Bogue

114. Graham A, Papalopulu N, Lorimer J, McVey JH, Tuddenham EGD, Krumlauf R. Characterization of a murine homeo box gene, Hox-2.6, related to the Drosophila Deformed gene. Genes Dev 1988; 2:1424–1438. 115. Cardoso WV. Transcription factors and pattern formation in the developing lung. Am J Physiol 1995; 269(4 Pt 1):L429–442. 116. Gaunt SJ, Krumlauf R, Duboule D. Mouse homeo-genes within a subfamily, Hox1.4, -2.6 and -5.1, display similar anteroposterior domains of expression in the embryo, but show stage- and tissue-dependent differences in their regulation. Development 1989; 107:131–141. 117. Krumlauf R, Holland PWH, McVey JH, Hogan BLM. Developmental and spatial patterns of expression of the mouse homeobox gene, Hox 2.1. Development 1987; 99:603–617. 118. Bogue CW, Lou LJ, Vasavada H, Wilson CM, Jacobs HC. Expression of Hoxb genes in the developing mouse foregut and lung. Am J Respir Cell Mol Biol 1996; 15:163–171. 119. Bogue CW, Gross I, Vasavada H, Dynia DW, Wilson CM, Jacobs HC. Identification of Hox genes in newborn lung and effects of gestational age and retinoic acid on their expression. Am J Physiol 1994; 266(4 Pt 1):L448–454. 120. Mollard R, Dziadek M. Homeobox genes from clusters A and B demonstrate characteristics of temporal colinearity and differential restrictions in spatial expression domains in the branching mouse lung. Int J Dev Biol 1997; 41(5):655– 666. 121. Rancourt DE, Tsuzuki T, Capecchi MR. Genetic interaction between hoxb-5 and hoxb-6 is revealed by nonallelic noncomplementation. Genes Dev 1995; 9(1):108– 122. 122. Bernacki SH, Nervi C, Vollberg TM, Jetten AM. Homeobox 1.3 expression: induction by retinoic acid in human bronchial fibroblasts. 1992; 7:3–9. 123. Chinoy MR, Volpe MV, Cilley RE, Zgleszewski SE, Vosatka RJ, Martin A, Nielsen HC, Krummel TM. Growth factors and dexamethasone regulate Hoxb5 protein in cultured murine fetal lungs. Am J Physiol 1998; 274(4 Pt 1):L610–20. 124. Chisaka O, Capecchi MR. Regionally restricted developmental defects resulting from targeted disruption of the mouse homeobox gene hox-1.5. Nature 1991; 350: 473–479. 125. Manley NR, Capecchi MR. The role of Hoxa-3 in mouse thymus and thyroid development. Development 1995; 121(7):1989–2003. 126. Aubin J, Lemieux M, Tremblay M, Berard J, Jeannotte L. Early postnatal lethality in Hoxa-5 mutant mice is attributable to respiratory tract defects. Dev Biol 1997; 192(2):432–445. 127. Minoo P, Su G, Drum H, Bringas P, Kimura S. Defects in tracheoesophageal and lung morphogenesis in Nkx2.1 (⫺/⫺) mouse embryos. Dev Biol 1999; 209(1):60– 71. 128. Lazzaro D, Price M, De Felice M, Di Lauro R. The transcription factor TTF-1 is expressed at the onset of thyroid and lung morphogenesis and in restricted regions of the foetal brain. Development 1991; 113:1093–1104. 129. Litingtung Y, Lei L, Westphal H, Chiang C. Sonic hedgehog is essential to foregut development. Nat Genet 1998; 20(1):58–61.

Genetic Models of Respiratory Tract Development

87

130. Pepicelli CV, Lewis PM, McMahon AP. Sonic hedgehog regulates branching morphogenesis in the mammalian lung. Curr Biol 1998; 8(19):1083–1086. 131. Semina EV, Reiter R, Leysens NJ, Alward WL, Small KW, Datson NA, SiegelBartelt J, Bierke-Nelson D, Bitoun P, Zabel BU and others. Cloning and characterization of a novel bicoid-related homeobox transcription factor gene, RIEG, involved in Rieger syndrome. Nat Genet 1996; 14(4):392–399. 132. Gage PJ, Suh H, Camper SA. Dosage requirement of Pitx2 for development of multiple organs. Development 1999; 126(20):4643–4651. 133. Kitamura K, Miura H, Miyagawa-Tomita S, Yanazawa M, Katoh-Fukui Y, Suzuki R, Ohuchi H, Suehiro A, Motegi Y, Nakahara Y, et al. Mouse Pitx2 deficiency leads to anomalies of the ventral body wall, heart, extra- and periocular mesoderm and right pulmonary isomerism. Development 1999; 126(24):5749–5758. 134. Lin CR, Kioussi C, O’Connell S, Briata P, Szeto D, Liu F, Izpisua-Belmonte JC, Rosenfeld MG. Pitx2 regulates lung asymmetry, cardiac positioning and pituitary and tooth morphogenesis. Nature 1999; 401(6750):279–282. 135. Lu MF, Pressman C, Dyer R, Johnson RL, Martin JF. Function of Rieger syndrome gene in left-right asymmetry and craniofacial development. Nature 1999; 401(6750):276–278. 136. Thomas PQ, Brown A, Beddington RS. Hex: a homeobox gene revealing periimplantation asymmetry in the mouse embryo and an early transient marker of endothelial cell precursors. Development 1998; 125(1):85–94. 137. Yatskievych TA, Pascoe S, Antin PB. Expression of the homeobox gene Hex during early stages of chick embryo development. Mech Dev 1999; 80:107–109. 138. Keng VW, Fujimori KE, Myint Z, Tamamaki N, Nojyo Y, Noguchi T. Expression of Hex mRNA in early murine postimplantation embryo development. FEBS Lett 1998; 426(2):183–186. 139. Newman CS, Chia F, Krieg PA. The XHex homeobox gene is expressed during development of the vascular endothelium: overexpression leads to an increase in vascular endothelial cell number. Mech Dev 1997; 66(1–2):83–93. 140. Bedford FK, Ashworth A, Enver T, Wiedemann LM. HEX: a novel homeobox gene expressed during haematopoiesis and conserved between mouse and human. Nucleic Acids Res 1993; 21(5):1245–1249. 141. Denson LA, McClure MH, Bogue CW, Karpen SJ, Jacobs HC. HNF-3β and GATA-4 transactivate the liver-enriched homeobox gene, Hex. Gene 2000; 246: 311–320. 142. Myint Z, Inazu T, Tanaka T, Yamada K, Keng VW, Inoue Y, Kuriyama M, Noguchi T. Genomic organization and promoter analysis of a mouse homeobox gene, hex. J Biochem (Tokyo) 1999; 125(4):795–802. 143. Capdevila J, Pariente F, Sampedro J, Alonso JL, Guerrero I. Subcellular localization of the segment polarity protein patched suggests an interaction with the wingless reception complex in Drosophila embryos. Development 1994; 120(4):987–998. 144. Heberlein U, Wolff T, Rubin GM. The TGF beta homolog dpp and the segment polarity gene hedgehog are required for propagation of a morphogenetic wave in the Drosophila retina. Cell 1993; 75(5):913–926. 145. Ingham PW. Localized hedgehog activity controls spatial limits of wingless transcription in the Drosophila embryo. Nature 1993; 366(6455):560–562.

88

Bogue

146. Basler K, Struhl G. Compartment boundaries and the control of Drosophila limb pattern by hedgehog protein. Nature 1994; 368(6468):208–214. 147. Tabin CJ, McMahon AP. Recent advances in Hedgehog signaling. Trends Cell Biol 1997; 7:442–446. 148. Bellusci S, Furuta Y, Rush MG, Henderson R, Winnier G, Hogan BL. Involvement of Sonic hedgehog (Shh) in mouse embryonic lung growth and morphogenesis. Development 1997; 124(1):53–63. 149. Bitgood MJ, McMahon AP. Hedgehog and Bmp genes are coexpressed at many diverse sites of cell-cell interaction in the mouse embryo. Dev Biol 1995; 172(1): 126–138. 150. Urase K, Mukasa T, Igarashi H, Ishii Y, Yasugi S, Momoi MY, Momoi T. Spatial expression of Sonic hedgehog in the lung epithelium during branching morphogenesis. Biochem Biophys Res Commun 1996; 225(1):161–166. 151. Hui CC, Slusarski D, Platt KA, Holmgren R, Joyner AL. Expression of three mouse homologs of the Drosophila segment polarity gene cubitus interruptus, Gli, Gli-2, and Gli-3, in ectoderm- and mesoderm-derived tissues suggests multiple roles during postimplantation development. Dev Biol 1994; 162(2):402–413. 152. Grindley JC, Bellusci S, Perkins D, Hogan BL. Evidence for the involvement of the Gli gene family in embryonic mouse lung development. Dev Biol 1997; 188(2): 337–348. 153. Motoyama J, Liu J, Mo R, Ding Q, Post M, Hui CC. Essential function of Gli2 and Gli3 in the formation of lung, trachea and oesophagus. Nat Genet 1998; 20(1): 54–57. 154. Hogan BL. Bone morphogenetic proteins: multifunctional regulators of vertebrate development. Genes Dev 1996; 10(13):1580–1594. 155. Massague J, Weis-Garcia F. Serine/threonine kinase receptors: mediators of transforming growth factor beta family signals. Cancer Surv 1996; 27:41–64. 156. King JA, Marker PC, Seung KJ, Kingsley DM. BMP5 and the molecular, skeletal, and soft-tissue alterations in short ear mice. Dev Biol 1994; 166(1):112–122. 157. Bellusci S, Henderson R, Winnier G, Oikawa T, Hogan BL. Evidence from normal expression and targeted misexpression that bone morphogenetic protein (Bmp-4) plays a role in mouse embryonic lung morphogenesis. Development 1996; 122(6): 1693–1702. 158. Weaver M, Yingling JM, Dunn NR, Bellusci S, Hogan BL. Bmp signaling regulates proximal-distal differentiation of endoderm in mouse lung development. Development 1999; 126(18):4005–4015. 159. Lawson KA, Dunn NR, Roelen BA, Zeinstra LM, Davis AM, Wright CV, Korving JP, Hogan BL. Bmp4 is required for the generation of primordial germ cells in the mouse embryo. Genes Dev 1999; 13(4):424–436. 160. Winnier G, Blessing M, Labosky PA, Hogan BL. Bone morphogenetic protein-4 is required for mesoderm formation and patterning in the mouse. Genes Dev 1995; 9(17):2105–2116. 161. Martin GR. The roles of FGFs in the early development of vertebrate limbs. Genes Dev 1998; 12(11):1571–1586. 162. Goldfarb M. Functions of fibroblast growth factors in vertebrate development. Cytokine Growth Factor Rev 1996; 7(4):311–325.

Genetic Models of Respiratory Tract Development

89

163. Bellusci S, Grindley J, Emoto H, Itoh N, Hogan BL. Fibroblast growth factor 10 (FGF-10) and branching morphogenesis in the embryonic mouse lung. Development 1997; 124(23):4867–4878. 164. Hogan BL, Grindley J, Bellusci S, Dunn NR, Emoto H, Itoh N. Branching morphogenesis of the lung: new models for a classical problem. Cold Spring Harb Symp Quant Biol 1997; 62:249–256. 165. Park WY, Miranda B, Lebeche D, Hashimoto G, Cardoso WV. FGF-10 is a chemotactic factor for distal epithelial buds during lung development. Dev Biol 1998; 201(2):125–34. 166. Min H, Danilenko DM, Scully SA, Bolon B, Ring BD, Tarpley JE, DeRose M, Simonet WS. Fgf-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless. Genes Dev 1998; 12(20): 3156–3161. 167. Sekine K, Ohuchi H, Fujiwara M, Yamasaki M, Yoshizawa T, Sato T, Yagishita N, Matsui D, Koga Y, Itoh N and others. Fgf-10 is essential for limb and lung formation. Nat Genet 1999; 21(1):138–141. 168. Peters K, Werner S, Liao X, Wert S, Whitsett J, Williams L. Targeted expression of a dominant negative FGF receptor blocks branching morphogenesis and epithelial differentiation of the mouse lung. Embo J 1994; 13(14):3296–3301. 169. Cardoso WV, Itoh A, Nogawa H, Mason I, Brody JS. FGF-1 and FGF-7 induce distinct patterns of growth and differentiation in embryonic lung epithelium. Dev Dyn 1997; 208(3):398–405. 170. Celli G, LaRochelle WJ, Mackem S, Sharp R, Merlino G. Soluble dominant-negative receptor uncovers essential roles for fibroblast growth factors in multi-organ induction and patterning. EMBO J 1998; 17(6):1642–1655. 171. Post M, Souza P, Liu J, Tseu I, Wang J, Kuliszewski M, Tanswell AK. Keratinocyte growth factor and its receptor are involved in regulating early lung branching. Development 1996; 122(10):3107–3115. 172. Simonet WS, DeRose ML, Bucay N, Nguyen HQ, Wert SE, Zhou L, Ulich TR, Thomason A, Danilenko DM, Whitsett JA. Pulmonary malformation in transgenic mice expressing human keratinocyte growth factor in the lung. Proc Natl Acad Sci USA 1995; 92(26):12461–21265. 173. Guo L, Degenstein L, Fuchs E. Keratinocyte growth factor is required for hair development but not for wound healing. Genes Dev 1996; 10(2):165–175. 174. Lebeche D, Malpel S, Cardoso WV. Fibroblast growth factor interactions in the developing lung. Mech Dev 1999; 86(1–2):125–136. 175. Tefft JD, Lee M, Smith S, Leinwand M, Zhao J, Bringas P, Jr., Crowe DL, Warburton D. Conserved function of mSpry-2, a murine homolog of Drosophila sprouty, which negatively modulates respiratory organogenesis. Curr Biol 1999; 9(4):219– 222. 176. de Maximy AA, Nakatake Y, Moncada S, Itoh N, Thiery JP, Bellusci S. Cloning and expression pattern of a mouse homologue of drosophila sprouty in the mouse embryo. Mech Dev 1999; 81(1–2):213–216. 177. Minowada G, Jarvis LA, Chi CL, Neubuser A, Sun X, Hacohen N, Krasnow MA, Martin GR. Vertebrate Sprouty genes are induced by FGF signaling and can cause chondrodysplasia when overexpressed. Development 1999; 126(20):4465–4475.

5 Development of Branched Structures and the Cellular Response to Hypoxia An Evolutionary Perspective

PABLO WAPPNER

PETER J. RATCLIFFE

Instituto de Investigaciones Bioquı´micas Fundacio´n Campomar Buenos Aires, Argentina

Institute of Molecular Medicine John Radclife Hospital Oxford, England

I.

Genetic Network Controlling Development of Drosophila melanogaster Tracheal System

Oxygenation of tissues in terrestrial vertebrates is achieved through the coordinated function of the respiratory and circulatory systems. The respiratory system fulfills an active ventilatory role in which the lungs force renewal of the air that contacts the alveolar epithelium and is juxtaposed to a dense blood capillary plexus. At this interphase, hemoglobin present in red blood cells releases carbon dioxide to the lungs and is loaded with oxygen. Thus, hemoglobin carries oxygen through very long distances with high efficiency, delivering it to the tissues and organs of the body. Insects have developed a much simpler, although very efficient, mechanism for oxygen delivery. The circulatory system is simple and does not fulfill an O2carrying function. Air reaches the tissues by passive diffusion through a highly ramified tubular network called the tracheal system (Fig. 1) [1,2]. The main longitudinal tubes in the tracheal system, the dorsal trunks, are linked to the outside by four small orifices, the anterior and posterior tracheal spiracles. Moving inward from the spiracles, the tubes ramify and narrow, constituting secondary and ter91

92

Wappner and Ratcliffe

Figure 1 The Drosophila tracheal network. Third instar larva tracheal system of a transgenic line expressing the green fluorescent protein (GFP) specifically in all tracheal nuclei (arrowhead ). Tracheal tubes can be seen by transmitted light (white arrow). A branching point is marked by black arrow.

tiary branches (also called terminal branches). Terminal branches are formed by single cells from which numerous cellular processes extend, reaching every single cell in the body. What are the molecular mechanisms underlying morphogenesis of such a complex spatial network? Primary and secondary branches develop during the embryonic stage and have a largely stereotyped structure and position [2] as is the case in vertebrate lung development and vasculogenesis [3,4]. In contrast, tracheal terminal branches are extremely plastic and respond to local oxygen requirements while they develop [5]. The formation of terminal branches occurs mainly at the end of embryogenesis and during the three larval stages [6]. Because

Branching Morphogenesis and Hypoxia

93

of its oxygen-dependent plasticity, the process can be compared to angiogenesis in vertebrates. A. Tracheal Cell Fate Differentiation

The embryonic tracheal system develops from 10 clusters of ectodermal cells on each side of the body. These 20 cell clusters, called tracheal placodes, differentiate at the embryonic stage 8 when they begin to express specific transcription factors that behave as master regulator genes for tracheal development. These genes encode Trachealess [7,8], a basic helix loop helix–PAS (bHLH-PAS) protein that forms a heterodimeric transcription factor with a second bHLH-PAS protein, Tango [9–11], and a POU-domain transcription factor called Drifter [12– 14]. The three proteins are most likely part of the same transcription complex, one that provides ‘‘tracheal identity’’ to the cells. Whereas tango is ubiquitously expressed [9], trachealess and drifter expression are spatially restricted and define the size and shape of the tracheal placodes [7,8]. The molecules that set up the spatial cues for such a restricted expression pattern on the ectoderm are unknown. Presumably, the precise position of the placodes is fixed by the intersection of anteroposterior and dorsoventral positional determinants [15]. After the initial expression pattern of both trachealess and drifter is set up, further expression is maintained by autoregulation throughout tracheal development [7]. In the absence of either trachealess or drifter, tracheal development does not take place [7,8,14–16]. In addition to controlling their own expression, trachealess and drifter induce a number of downstream genes that are necessary for the initial stages of tracheogenesis [10–12]. B. Tracheal System Is Shaped Exclusively by Cell Migration. Role of the FGF Receptor Pathway

Initially, tracheal placode cells undergo two rounds of divisions, reaching the 80-cell placode stage (Fig. 2A). Remarkably, from this stage onward, tracheal development proceeds solely by cell migration without any further proliferation [1,2]. At this stage, tracheal cells invaginate, bringing about the tracheal pits (Fig. 2B) in a process dependent on the Drosophila epidermal growth factor receptor (EGFR) pathway (see below) [17]. Thereafter, cells start to migrate in six precisely defined directions forming the primordia of the six main primary tracheal branches, which are the dorsal trunk anterior (DTA) and posterior (DTP), the dorsal branch (DB), the lateral trunk anterior (LTA) and posterior (LTP), and the visceral branch (VB) (Fig. 2C–E) [18,19]. The migration of tracheal cells depends on the fibroblast growth factor (FGF) homolog encoded by the branchless gene [20] and its receptor encoded by breathless [21–24]. The receptor Breathless is expressed in all tracheal

94

Wappner and Ratcliffe

Figure 2 Treacheal development during embryogenesis in Drosophila. Development of the tracheal system throughout embryogenesis. Tracheal cells are visualized with an antibody specific for the trachealess protein. (A) Stage 10 embryos (4.20–5.20 hr), dorsal view. Tracheal placodes are already differentiated but invagination has not started. (B) Stage 11 embryos (5.20–7.20 hr), doral view. Tracheal placodes have undergone invagination, giving rise to the tracheal pits. (C) Stage 12 embryo (7.20–9.20 hr), lateral view. Tracheal cells have started to migrate in six different directions to bring about the tracheal primary branches. (D) Stage 14 embryo (10.20–11.20 hr), lateral view. Primary branches are already formed. Fusion of tracheal tubes of adjacent segments is not complete. (E) Tracheal system of a 14-stage embryo, closer lateral view; transgenic fly line expressing beta-galactosidase under control of trachealess promoter. Tracheal cells are visualized with an anti–β-galactosidase antibody. DB, dorsal branch; Dta, dorsal trunk anterior; DTp, dorsal trunk posterior; VB, visceral branch; LTa, lateral trunk anterior; LTp, lateral trunk posterior.

cells. In contrast, Branchless/FGF is expressed outside the tracheal placode in small clusters of ectodermal cells (see Fig. 4). The Branchless expression pattern predicts the direction of migration of the cells that bring about the primary tracheal branches. Ectopic expression of branchless in an unusual ectodermal location causes tracheal cell migration and branch outgrowth toward the new location of expression [20,25]. Thus, the shape of the tracheal tree is set up by Branchless/ FGF expression pattern in the target tissues and consequent activation of Breathless in the tracheal cells.

Branching Morphogenesis and Hypoxia

95

When tracheal branches begin to form, a leading cell in each branch differentiates from the rest of the cells [18]. The binding of Branchless/FGF to the Breathless receptor activates the canonical Ras cascade, and this activation occurs exclusively in the leading cell. Such a restricted activation pattern can be monitored in situ with an antibody that recognizes the diphosphorylated (activated) form of mitogen activated protein (MAP) kinase [26,27]. How is that Breathless activation is restricted to the leading cell? A mechanism preventing activation of the neighboring cells does indeed occur. The gene sprouty is expressed exclusively in the leading cell and exerts inhibition on the neighboring cells, thus preventing Breathless activation. sprouty Mutations bring about a multi-leading cells phenotype in which many ectopic terminal branches are formed [28]. Recently, sprouty was shown to be a general negative effector for the canonical Ras pathway [29,31]. How does branch formation proceed after the onset of cell migration? The branchless expression pattern in the target tissues is very dynamic. Once the leading cell has reached the Branchless positive cluster, branchless expression ceases and the gene is turned on in a new cluster of cells a bit further on the track of the forming branch. Thus, branchless turns on and off many times along the path of each of the growing branches throughout tracheal development [20]. C. Subdivision of the Tracheal Placode into Different Subfates. An Interplay Between the Transforming Growth Factor-␤ and EGF Receptor Pathways

How are cells of the tracheal placode allocated to each branch primordia in the correct number and position? We discussed above the mechanism by which Branchless determines the direction of migration of the leading cell in each of the growing branches. Clearly, the information provided by the Branchless/ Breathless system is insufficient to instruct the cells of the tracheal placode as to how they should split and which branch they should join. Interactions between the EGF and transforming growth factor-β (TGF-β) receptor pathways establish distinct cell fates in the tracheal placodes [19]. Phenotypic analysis of mutants and of the consequences of ectopic gene expression have shed light on the mechanism of action of these two pathways. Tracheal mutant phenotypes for genes of the EGFR pathway indicate that it is responsible for patterning the branches that extend in the anteroposterior direction; namely, the dorsal trunk and visceral branches. Homozygous mutations in any of the genes in this pathway do not interfere with dorsoventral cell migration and, therefore, do not alter development of the dorsal branch and lateral trunks [19]. The EGFR pathway in the tracheae is activated by the ligand Spitz, which is synthesized as an inactive membrane precursor. In order to become active,

96

Wappner and Ratcliffe

Spitz is cleaved by an unknown protease that depends on the function of two additional membrane proteins called Rhomboid and Star [32,33]. In every tissue where Spitz functions as the EGFR ligand, Rhomboid is the only component of the pathway with a spatial and temporal restricted expression pattern. All the other components, including the receptor and Spitz precursor, are ubiquitously expressed. Thus, the Rhomboid expression pattern marks the activation of the pathway [34]. Rhomboid is transiently expressed in a subset of cells lying at the center of the tracheal placode before the onset of invagination [19]. Thus, it can be inferred that determination of branch fates by the EGFR pathway precedes the onset of migration of tracheal cells. The MAP kinase activation pattern visualized with the antidiphospho-Erk antibody supports this notion [27]. The TGF-β homolog Decapentaplegic (Dpp) has a complementary role: it patterns the dorsoventral branches; namely, the dorsal branch and lateral trunks. Mutants for either type I or type II subunits of the TGF-β receptor lack dorsoventral branches without showing defects in the development of anteroposterior branches [35]. At the placode stage, the ligand Dpp is expressed on the ectoderm in two stripes abutting the tracheal placode. In accordance with this, the type II subunit of the receptor is expressed exclusively inside the tracheal placode [36,37]. Thus, Dpp diffuses into the tracheal placode activating the signaling pathway at the dorsal and ventral parts of the placode. Presumably, cells that migrate dorsally and ventrally arise from such regions of the placode. In addition, antagonistic interactions between EGFR and Dpp pathways have been demonstrated. Overexpression of Dpp or a constitutively active form of the receptor bring about a phenotype that resembles that of EGFR pathway mutants [35]. Conversely, overexpression of a secreted form of Spitz on a Dppsensitized background results in a phenotype similar to TGF-β pathway mutants. Thus, mutant phenotypes of each of the two pathways resemble overexpression phenotypes of the other pathway. Hypomorphic mutations in either the Dpp or EGFR pathway lead to phenotypes indistinguishable from those of their null mutations. Nevertheless, when such hypomorphic mutations in both pathways are combined in the same embryo, normal tracheal cell migration is largely restored. This result confirms that the direction of tracheal cell migration emerges from a delicate balance between the two signaling pathways. Predominance of Dpp, either by overexpression or due to a mutation in the EGFR pathway, instructs the cells to migrate dorsoventrally. In contrast, predominance of the EGFR pathway leads to anteroposterior migration of the tracheal cells (Fig. 3) [19]. D. Genetic Hierarchy Controlling Tracheal Branching

Primary branches are formed at the beginning of tracheogenesis when multiple cells adopt a tubular array as they migrate. Primary branch formation begins at

Branching Morphogenesis and Hypoxia

97

Figure 3 Model for patterning the tracheal cells by the EGFR and DPP pathways. At stages 10–11, the EGFR pathway is activated in the tracheal placode. Owing to Rhomboid expression in the central part of the placode and restricted diffusion of secreted Spitz, stronger activation levels are observed in the central part. In contrast, DPP is expressed on the ectoderm in two stripes abutting the tracheal placode. Thus, higher levels of DPP pathway activation may be induced in the dorsal and ventral part of the placode. As a result, two different cell populations are determined in the trachea; the EGFR-induced cells in the central part of the DPP-induced cells in the dorsal and vental regions. When tracheal cell migration begins at stage 12, different tracheal cells are recruited to different tracheal branches according to the fate they assumed. It is possible that activation of Breathless by Branchless induces only the migration of the tip cells (marked by *) and the fate of the other cells would determine which leading tip cell they will join. (From Ref. 19, used with permission).

the placode stage and continues until embryonic stage 15. By stage 15, secondary branches arise either at the end or at internal positions of primary branches [1,2]. From embryonic stage 16 onward, secondary branches sprout out many subcellular processes that ramify more and more finely, penetrate into the target tissues, and constitute the terminal branches [6,21,38]. Terminal branch formation continues throughout the three larval stages. Ultrastructures of primary, secondary, and terminal branches are quite different [18]. A cross section of a primary branch shows an array of multiple cells linked by intercellular junctions. In contrast, a cross section of secondary branches exhibits unicellular tubes in which the cells

98

Wappner and Ratcliffe

are closed on themselves by intracellular junctions. Since terminal branches are formed by subcellular projections, a cross section shows a continuous tube wall with no junctions [18]. Different genes are activated at each stage of branching morphogenesis. Such genes are organized in a regulatory hierarchy and, therefore, have a dual function. On one hand, they regulate a particular level of branching. On the other, they induce genes that control the next level of branching. ‘‘Pantip’’ markers, like the ETS transcription factor Pointed, are expressed in secondary branches and induce the expression of ‘‘Terminal’’ and ‘‘Fusion’’ markers [18]. Terminal markers, like the serum response factor (SRF), are expressed at the tip of secondary branches and turn on the sprouting program that brings about terminal branches [6,38]. Fusion markers, like the zinc finger protein Escargot, are expressed in the cells that undergo fusion between tracheal tubes while they repress the terminal branching program [39–41]. How are fusion and terminal cells distinguished from the rest of the cells in a given branch? The Notch pathway has a key role in this process, and its mechanism of action involves two phases [42,43]. First, the Notch pathway helps to single out cells that participate in tracheal fusion. Activation of the pathway leads to inhibition of differentiation toward a ‘‘fusion fate.’’ Localized expression of the membrane-bound ligand Delta mediates this inhibition: Delta accumulates at the tips of the primary branches, thus activating the receptor Notch at the second cell of the branch. In consequence, this second cell of the branch is prevented from acquiring a fusion fate. It is important that only one cell (the tip cell) acquires the fusion fate, and by this mechanism, the system avoids the simultaneous occurrence of two fusion cells in the same branch. The occurence of multiple fusion cells in a single branch would interfere with the normal fusion process. In a second phase, perhaps later during tracheal branch development, the Notch pathway is activated again to participate on the selection of cells that extend terminal branches. The mechanism seems to be analogous: Localized expression of Delta activates the Notch pathway in the cell that is adjacent to the one that acquires the ‘‘terminal fate.’’ This mechanism prevents occurrence of multifusion fates within a single branch [42].

II. Conserved Features of Branching Morphogenesis in Evolution Mammalian branching morphogenesis shares many cellular and molecular features with tracheal development in Drosophila [3]. Development of both the vasculature and the respiratory systems are paradigmatic examples of branching morphogenesis in mammals. We will discuss briefly both processes and draw a comparison between them and Drosophila tracheal development.

Branching Morphogenesis and Hypoxia

99

A. Lung Development

Lung primordia develop from epithelial cells on the ventral gut that sprout into the mesenchyme. The two primary buds undergo several rounds of branching, giving rise to smaller and smaller bronchi [4]. Genetic mechanisms controlling lung development are less well characterized than the ones responsible for tracheal system development in Drosophila. For instance, it is still unclear how location of lung bud initiation is determined. Hox genes are probably involved in positioning the lung buds along the anteroposterior axis, since in Hoxb5 mutants lung buds are shifted to a more anterior position [44]. The homology between molecular mechanisms underlying branching morphogenesis in mammalian lung and Drosophila tracheae is nevertheless striking (Fig. 4) [3,4]. Early lung development is highly stereotyped in both the timing and position of the branched structure. The branching pattern emerges from cross talk between the epithelium and the mesenchyme [45]. Transplantation studies indicate that the mesenchyme that surrounds the distal part of the growing buds has an instructive activity on the epithelium, providing information for the correct branching pattern [46]. FGF-10 appears to have a major role in lung morphogenesis, since null mutant mice have no lungs or lung buds at all [47]. Moreover, FGF-10 is expressed in the mesenchyme and seems to fulfill an instructive role similar to that of Branchless/FGF in Drosophila [20,48,49]. Expression of a dominant negative form of the FGF-10 receptor, Fgfr2IIIb, in the epithelium prevents branching just as Breathless/FGF dominant negative receptor does in Drosophila [21,50]. FGF-10 expression in the mesenchyme has a highly dynamic pattern very reminiscent to that of Branchless during Drosophila embryonic development. When a growing branch reaches an FGF-10 expression domain, the gene is switched off and its expression restarts slightly further on in a new location. As in Drosophila tracheal development, iteration of this process provides a guidance mechanism for the migrating cells (see Fig. 4) [4]. What other molecules contribute to the cross talk between the bronchi and the surrounding mesenchyme? Sonic Hedgehog (Shh), which is expressed in the epithelium, seems to be important in the process. Shh null mutant mice show over expression of FGF-10 in the adjacent mesenchyme, indicating that Shh downregulates FGF-10 in the neighboring cells [44,48]. Further research is required to address whether other molecular similarities underlie both branching morphogeneic processes. However, molecular homology between tracheal and lung development extends even further in other ways. The Drosophila sprouty (sty) gene is a non– cell autonomous branching inhibitor expressed in terminal cells, preventing adjacent cells from acquiring a terminal fate [28]. Four mammalian sty orthologs have been identified so far [51,52]. Interestingly, one of these orthologs, the murine sprouty-2 gene, is expressed in the lungs. An antisense oligonucleotide strat-

100

Wappner and Ratcliffe

Fgfr2

Figure 4 Tracheal and bronchial cell migration is guided by growth factors from the FGF family. Branchless/FGF expression in clusters of ectodermal or mesodermal cells guides the migration of forming tracheal branches which express the FGF receptor homologue, Breathless. The scheme shows a stage 12 developing tracheal system. Branchless expression is very dynamic. Once the growing tracheal branch has reached the cluster that expresses branchless, expression ceases and the gene is turned on in a new cluster of cells a bit further on the track of the forming branch. A similar situation is reproduced during the formation of bronchial branches in lung development. FGF-10 is expressed in clusters of cells in the mesenchyme that surround the forming bronchi. Bronchial cells which express FGFR2 are guided toward FGF-10–expressing cell clusters.

Branching Morphogenesis and Hypoxia

101

egy has been used in order to investigate the function of this gene in embryonic murine lungs in culture. Interestingly, suppression of Sty expression led to a dramatic increase in branching, showing that sty function has also been conserved [53]. Overall, comparisons between the molecular mechanisms underlying development of Drosophila tracheae and development of mammalian lung shows that the basic developmental machinery has been largely conserved over 600 million years of metazoan evolution. B. Vasculogenesis and Angiogenesis

Two distinct processes, vasculogenesis and angiogenesis, are involved in the development of the vascular system. Vasculogenesis is the formation of vascular structures from precursor cells called angioblasts, whereas angiogenesis involves formation of new blood vessels from preexisting ones [54]. Vasculogenesis is analogous to the stereotyped stages of the treacheal development, whereas angiogenesis parallels the formation of terminal branches. Cells involved in vasculogenesis derive from bipotential precursors called hemangioblasts. Such a cell lineage has the ability to differentiate toward hematopoietic precursors or alternatively become angioblasts involved in development of the vascular system [55,56]. Vasculogenesis begins with the formation of the primary capillary plexus, a network of homogeneously sized blood vessels without a defined directionality or hierarchical branching pattern. Subsequently, the capillary plexus undergoes maturation in a process that is part of the angiogenic stage of vascular development [57]. At this stage, angiogenesis occurs through two alternative mechanisms: sprouting and nonsprouting angiogenesis. Both mechanisms operate concurrently in the vascularization of several organs like the lung and the heart. Nonsprouting angiogenesis involves the splitting of preexisting blood vessels. In this process, endothelial cells proliferate inside a vessel, subdividing it down its long axis. In contrast, sprouting angiogenesis involves chemotactic migration and proliferation of endothelial cells from preexisting blood vessels. During migration these cells form a lumen, and finally the process is completed with the maturation of the endothelium [57]. The manner in which cells migrate and form tubules in vasculogenesis closely resembles tracheal development. However, it is of note that Drosophila tracheal tubes are formed exclusively by cell migration with complete absence of cell division [1,2]. This might be seen as a major difference between the two processes. Nevertheless, tracheal development in more primitive insects such as the blood-sucking bug Rhodnius prolixus involves not only cell migration but also cell proliferation [5,58]. Therefore, suppression of cell division during tubulogenesis is probably a secondary and relatively late event in insect phylogeny. The basic morphogenetic mechanism has most likely arisen only once in evolution and has been coopted for various developmental processes in different phyla.

102

Wappner and Ratcliffe

Experiments with knockout mice, as well as concurrent data from many ex vivo and in vitro models, has delineated the role of several growth factors at different stages of vasculogenesis and angiogenesis [57]. Growth factors of the vascular endothelial growth factor family (VEGF-A–D and P1GF) and their tyrosine kinase receptors (VEGF-R1–R3) play a major role in all stages of vascular development [59]. Knockout mice for the VEGF-R2 or VEGF-A have defects in both hematopoiesis and vasculogenesis. This observation suggests that a member of the family, VEGF-A, and its receptor VEGF-R2 are involved in differentiation of hemangioblasts from the mesodermal precursor cells [56,60]. Experimental evidence suggests that the next step in the differentiation cascade, namely, the binary choice between hematopoiesis or vasculogenesis, seems also to involve VEGF-A [60]. VEGF-R2 expression persists in angioblasts, whereas expression ceases in the hematopoietic lineage [54]. VEGF-R1 has been implicated in the formation of the primary plexus. Knockout mice for this receptor produce normal angioblasts but fail to assemble them in order to generate a normal vascular system [60–62]. Interestingly, VEGF-R3–defective mice exhibit an abnormal vascular organization characterized by irregular large vessels with defective lumens [63,64]. VEGFs also play a central role in angiogenesis as well, since they exhibit potent mitogen and chemoattractant activity on endothelial cells in all experimental settings. In most developing organs, the nascent blood vessels stain positively for all VEGF receptors, whereas the stromal cells that surround the forming vasculature express VEGFs [65,66]. The role of VEGF in vascular development resembles closely the participation of Branchless in tracheogenesis. Both proteins, acting through their tyrosine kinase receptors, exert an instructive role, guiding the leading cells toward their target tissues. Such homology extends not only across the stereotyped stage of vasculature and tracheal development but also to hypoxia-dependent angiogenesis and plastic development of tracheal terminal branches (discussed in the next section). It is interesting to stress that, by sequence comparison, Branchless is most closely related to a different mammalian family of molecules, the fibroblast growth factors (FGFs) [20]. Though basic fibroblast growth factor (bFGF) is a potent angiogenic factor in vitro with strong chemoattractant and mitogenic activity in mammalian systems [67], its role in the development of vascular system is elusive. Knockout mice for bFGF show no developmental defect. Since functional redundancy with other members of the FGF family probably occurs, a role of bFGF in the formation of blood vessels during development cannot be ruled out [57]. Nevertheless, it would seem that although organization of functional pathways is highly conserved between insects and mammals, related genes may serve different roles in the two phyla. In mammals, angiopoietins (Ang-1–4) are also known to play a major role in vasculogenesis and angiogenesis [68–71]. Two angiopoietin receptor tyrosine

Branching Morphogenesis and Hypoxia

103

kinases have been identified, termed Tie-1 and Tie-2. Whereas the function of Tie-1 is still undetermined, Tie-2 has a clear role in vasculature development [59]. Ang-1 and Ang-3 ligands have been shown to activate Tie-2 receptor [72]. Ang-2 and Ang-4, on the other hand, are competitive inhibitors at the Tie-2 receptor, and counterbalance the action of the stimulatory ligands [73,74]. The function of angiopoietins is very different, rather complementary when compared with VEGFs [75]. They are unable to exert a mitogenic or chemoattractant effect on endothelial cells. Rather the Ang/Tie system participates in the recruitment of support cells that surround endothelial cells in the blood vessels; namely, pericytes and smooth muscle cells. Support cells secrete angiopoietins that stimulate the Tie-2 receptor expressed on the endothelial cells. The vascular system in knockout mice for Ang-1 or Tie-2 is grossly abnormal, although a primitive vasculature is formed. Further stages of development, namely, consolidation and remodeling of such a primitive vasculature, are deficient. Thus, the angiopoietinmediated cross talk between endothelial and support cells is critical for vasculature maturation and remodeling. Remarkably, transgenic mice overexpressing Ang-1 exhibit a hypervascularized phenotype that resembles the Ang-2 loss of function phenotype. This observation provides functional evidence that Ang-1 and Ang-2 have antagonistic effects in vivo. It has been suggested that Ang-1 stabilizes the vasculature, whereas Ang-2 causes destabilization leading to remodeling of the vascular structure [74]. Thus, Ang-2 has a cardinal role in actively remodeling tissues such as the female reproductive tract, and Ang-1 prevails in nonremodeling tissues [59]. Angiopoietin homologs have not been identified in Drosophila as yet. Although Drosophila tracheal tubules and vertebrate blood vessels share many cellular and molecular features, particular ultrastructural differences should be highlighted. Tracheal cells secrete a chitinous cuticle through their luminal membrane, which contributes to maintenance of the shape and structure of tracheal tubes, thus avoiding the requirement for additional cells [5]. In contrast, endothelial cells in blood vessels require accessory support cells in order to differentiate properly and achieve the correct structure. Angiopoietins seem to be involved exclusively in the cross talk between endothelial and support cells. Such a difference might explain why angiopoietins most likely do not exist in the Drosophila tracheal system. Recent reports have shed light on the unexpected importance of another new family of molecules in vascular development: the Ephrins, molecules which were originally found to contribute to neural development, were recently implicated in setting up the final vascular pattern [76]. To date, 14 Ephrins receptor tyrosine kinases and 8 ligands have been identified [59]. It is of interest that Ephrins are membrane-bound ligands that need to form clusters in order to activate their receptors on the adjacent cells. Monomeric forms appear to block receptor activation [77]. Detailed studies have focused on the interaction between the

104

Wappner and Ratcliffe

ligand Ephrin-B2 and the receptor Eph-4. Interestingly, their expression pattern in forming blood vessels revealed an unexpected role: They set up the borders between arteries and veins during maturation of the capillary plexus [78]. EphrinB2 (ligand) is expressed in the future arterial but not venous endothelial cells, whereas Eph-4 (receptor) marks exclusively the cells that will become part of the veins. Such an interaction appears to be essential for proper development of the vasculature, since null homozygous mice for ephrin-B2 exhibit major defects in maturation of the vascular system [76]. Ephrin ligand and receptor homologues have been identified in lower metazoans, such as Caenorabditis elegans and Drosophila melanogaster. The C. elegans Eph receptor homolog, VAB-1, participates in neural and epithelial morphogenesis [79]. Neuronal expression of VAB-1 is required for both neuroblast movement during closure of the ventral gastrulation cleft and also for epidermal cell movement during ventral enclosure of the embryo by the epidermis. The fact that neuronal expression is also needed for epidermal cell movement implies that cross talk between the two cell types is critical in the migration mechanism [79]. The Drosophila Eph receptor homolog, Dek, is expressed exclusively in a large subset of embryonic interneurons and targeted to the axons when pathfinding takes place [80]. Functional redundancy probably occurs, since overexpression of the protein in neurons does not cause any detectable phenotype. Remarkably, another Drosophila protein called DOCK that is needed for proper retinal neuron assembly [81–83] interacts in a yeast two-hybrid system with a mammalian EphB1 receptor [84], indicating that Ephrins are likely to be involved in neuronal pathfinding in Drosophila. So far, Ephrins have not been implicated in tracheal development, although such a possibility seems likely a priori. As in mammals, a large family of Ephrins and Ephrin receptors probably exist in Drosophila. Isolation and study of novel members of this family may uncover a role in tracheal pathfinding, tubular differentiation, or branch fusion.

III. Cellular Response to Low Oxygen Tension, Conserved Features In the previous section, we discussed the mechanisms that control angiogenesis during development. In adult organisms, local hypoxia is a major stimulus for angiogenesis in various physiological and pathological situations. The control of angiogenesis has become a major therapeutic goal [85]. In ischemic pathologies such as coronary heart disease and chronic limb ischemia, enhanced angiogenesis may improve outcome [86,87]. In contrast, tumorigenesis and metastasis require the outgrowth of new blood vessels [88]. When tumors grow over a certain size, the inner cells become hypoxic and, if additional blood supply is not received, tumor growth cannot continue [89–91]. Although the role of hypoxia in angio-

Branching Morphogenesis and Hypoxia

105

genesis has long been appreciated [92], the mechanisms have become better understood only recently following the definition of angiogenic growth factors. For instance, VEGF levels are increased up to 10- to 15-fold in hypoxic cells [93,94]. Part of this induction is due to an increase in mRNA stability and part is due to transcriptional induction. A. Hypoxia Inducible Factor-1. General Features

Transcriptional induction of VEGF is known to be mediated by a DNA binding complex termed hypoxia inducible factor-1 (HIF-1) that behaves as a general inducer of genes that respond to hypoxia [95,96]. HIF-1 is a heterodimeric transcription factor composed by two subunits that belong to the basic helix loop helix–PAS (bHLH-PAS) family [97–99]. The alpha subunit is stabilized in hypoxia, whereas the beta subunit, a common partner for several other bHLH-PAS proteins, is constitutively expressed regardless of oxygen tension [100–102]. Mutant cell lines lacking the beta subunit are unable to respond to hypoxia, indicating that it is essential for the alpha subunit to fulfill its function [103]. The PAS family was first defined as a group composed by two mammalian transcription factors, the aryl hydrocarbon receptor (AHR) and aryl hydrocarbon nuclear translocator (ARNT) and two Drosophila proteins, Period (Per) and Single-minded (Sim) [104]. All bHLH-PAS proteins form heterodimers and have a stereotypic structure consisting of several domains with diverse levels of conservation [105]. The basic helix loop helix domain is involved in DNA binding and protein dimerization, and the PAS domain participates in target gene specificity and protein– protein interactions. The role of the PAS domain in Drosophila bHLH-PAS proteins was clearly demonstrated in vivo using Trachealess (Trh) and single minded (Sim) as model proteins [10]. TRH is a master regulator protein for tracheal development [7,8] and Sim is involved in patterning the ventral embryonic midline [106–108]. Hence, each bHLH-PAS transcription factor activates a different set of target genes that determine different developmental outcomes. TRH and SIM share a common beta subunit and recognize the same regulatory sequence in the DNA. The question of how the target specificity of Trh versus Sim is determined was answered by constructing chimeras between the two proteins and expressing them in transgenic files [10]. A Sim protein whose PAS domain was replaced by the PAS domain of Trh behaved as a wild-type Trh in terms of target gene specificity. Interestingly, recent sequence alignments suggest that these PAS proteins form part of a superfamily which extends across all living organisms and contains members with sensing functions for oxygen, redox potential, and light [109,110]. One of the best-characterized examples is the FixL protein which functions as a heme protein oxygen sensor in Rhizobium meliloti [111,112] (see below). Structural studies indicate that the heme center binds to the PAS domain of FixL. Surprisingly, studies of HIF-1α regulation by hypoxia have suggested that in this

106

Wappner and Ratcliffe

molecule regulatory sequences lie C terminal to the PAS domain (see below), so that it is not yet clear whether the molecules function in an analogous way despite both being regulated by oxygen. HIF-1 not only induces VEGF, since it drives the expression of a plethora of hypoxia-responsive genes [100,102,113]. These include genes involved in systemic responses and genes concerned with cellular adaptive responses. Systemic response genes include erythropoietin and transferrin (involved in erythropoiesis) [97,116], the inducible nitric oxide synthase (vasomotor responses) [117], and tyrosine hydroxylase (catecholamine biosynthesis and the ventilatory response) [118,119]. Cellular adaptive responses involve induction of the glucose transporter 1 (Glut1) and many glycolytic enzymes such as aldolase A (ALDA), enolase 1 (ENO1), phosphofructokinase (PFK), phosphogliceratokinase 1 (PGK1), and lactate dehydrogenase A (LDH-A) [102,120–124]. Upregulation of glycolytic enzymes has the potential to enhance anaerobic metabolism and partially compensate for reduction in oxidative phosphorylation [114]. Another important HIF-1–mediated cellular response appears to involve stabilization of P-53 protein and activation of apoptosis [125,126]. Interestingly, in one report, HIF-1α null embryonic stem cells were found to give rise to large tumors that grew in nude mice despite impaired vascular function [125]. This result contrasts with the majority of studies where inhibition of angiogenesis retards tumor growth and in some cases leads to tumor retraction. One interpretation is that the effect arises from two conflicting roles of HIF-1, a proangiogenic effect supporting growth, and a proapoptotic or antiproliferative effect which normally suppresses the growth of hypoxic cells by an active process [125]. In all genes in which HIF-1 activates hypoxia-dependent transcription, enhancer regions were found to contain one or more HIF-1 responsive elements (HREs) with the core sequence 5′-RCGTG-3′ [97,119–124]. Electromobility shift assay (EMSA) experiments have shown that short sequences surrounding the core binding site are sufficient for HIF-1 binding [96,97]. Nevertheless, deletional and mutational analysis of hypoxia-responsive promoters have shown that, in order to induce transcription, HIF-1 requires at least one secondary consensus sequence presumably binding a cooperative factor [100,113,114]. In some cases, the precise elements and their binding factors are still unknown, although some have been identified. These include a cAMP-responsive element in the LDH-A promoter [121] which binds CREB-1 and ATF-1 [127] and a nuclear receptor response element in the erythropoietin enhancer [114] which binds HNF-4 [128]. Relatively few data on hypoxia-responsive genes are available in Drosophila. Phosphoglycerate kinase mRNA is upregulated in Drosophila Schneider (SL2) cells [129] subjected to 1% oxygen, indicating that genes encoding glycolytic enzymes appear to respond to oxygen in Drosophila as well as in mammalian cells. In keeping with this, Drosophila PFK and ALDA genes bear typical hypoxia-responsive elements in their enhancer regions (P. Wappner, unpublished

Branching Morphogenesis and Hypoxia

107

data). However, functional studies are required in order to assess if these genes respond to hypoxia. The existence of a Drosophila HIF-1 homolog was first detected by EMSA [129]. In these studies, nuclear extracts from normoxic or hypoxic SL2 cells were incubated with a labeled oligonucleotide derived from mouse erythropoietin and phosphoglycerate kinase-1 enhancers. The experiments revealed the formation of a hypoxia inducible complex, strongly suggesting the occurrence of a homologous system of gene regulation in flies which is able to interact with mammalian a DNA recognition site for HIF-1 (Fig. 5) Conservation between mammalian and insect responses to hypoxia are also seen at the functional level. The oxygen-dependent development of tracheal terminal branches shares many features with mammalian oxygen-dependent angiogenesis [3]. Classic experiments performed by insect physiologists first revealed the extreme plasticity and compensatory capacity of the insect tracheal system [58]. In the blood-sucking bug Rhodnius prolixus, a major abdominal tracheal branch was surgically severed, thus generating an hypoxic area. A few days later, a dramatic compensatory response was observed, with branches from neighboring segments extending terminal branches into the hypoxic zone. In another classic experiment, a metabolically active heterologous organ was implanted in a Rhodnius abdomen. The organ was rapidly invaded by projections emanating from neighboring tracheal branches that provide the additional oxygen demand [58]. Much more recently, taking advantage of the genetic tools available in Drosophila, it was shown that homology between angiogenesis and the terminal stages of tracheal development extends to the molecular response to hypoxia [130]. In a manner similar to the function of VEGF in mammalian angiogenesis, Drosophila Branchless/FGF guides tracheal migration during development and also drives extension of plastic terminal branches as well. As with VEGF in mammalian systems, Western blot experiments have shown that Branchless itself is induced by hypoxia [130]. An interesting observation exemplifies the role of Branchless in oxygen-dependent tracheal plasticity. Larval tracheal branches occasionally break spontaneously, thus creating a hypoxic area. In these areas, dramatic upregulation of branchless expression occurs and tracheal extensions are attracted, correcting the deficit in oxygen supply. Another example is seen in experimental tumors which can be generated in Drosophila larvae by genetic manipulations. As with mammalian tumorigenesis, the tumor core becomes hypoxic, expresses Branchless (equivalent to VEGF in mammals), and is subsequently invaded by tracheal projections (equivalent to blood capillaries in mammals) that bring extra oxygen to the tumor core [131]. The mechanisms controlling branchless expression in Drosophila remain mysterious. However, circumstantial evidence suggest that branchless is probably induced by a transcription factor from the bHLH-PAS family [132]. A transgenic line bearing a LacZ transcriptional reporter under control of a HIF-1 binding site concatemer has indicated the existence of an unknown bHLH-PAS

108

Wappner and Ratcliffe

Figure 5 Electromobility shift assay (EMSA) showing the formation of a hypoxia inducible protein in Drosophila that binds HIF-1 responsive elements on the DNA. Specific binding of a Drosophila HIF-1 homolog to mouse erythropoietin (Epo) or phosphoglycerate kinase (PGK) probes. Five micrograms of nuclear extract from normoxic (N) or hypoxic (H) HeLa (lanes 1–2) or SL2 cells (lanes 3–14) were incubated with labeled Epowt (lanes 1–8) or PGK-wt (lanes 9–14) oligonucleotides. Competition with a 200-fold molar excess of unlabeled oligonucleotide was as follows: Epo-wt in lanes 5–6, Epo-mut in lanes 7–8, PGK-wt in lanes 11–12, PGK-mut in lanes 13–14. Extracts were incubated in binding buffer for 5 min. Competitors followed immediately by probe were added, and after a further 10 min incubation, samples were separated on a 5% polyacrylamide EMSA gel. Positions of inducible (I) and constitutively (C) retarded species are marked (From Ref. 129, used with permission).

Branching Morphogenesis and Hypoxia

109

protein that is expressed exactly in the same pattern as branchless (P. Wappner and B.-Z. Shilo, unpublished data). Genetic experiments indicated that the expression of the unknown bHLH-PAS protein is not under branchless control, strongly suggesting that, conversely, the bHLH-PAS transcription factor is the factor that regulates branchless gene expression [132]. B. Regulation of HIF-1 Stability by Oxygen Levels

HIF-1 protein is virtually undetectable in cells kept in normoxic cells, and its levels raise dramatically over a few hours when cells are transferred to a hypoxic atmosphere [98]. Although it was initially reported that HIF-1 mRNA levels increase with hypoxia, most laboratories have not demonstrated hypoxic induction at the mRNA level, indicating that regulation is posttranscriptional. Cycloheximide chase experiments indicated that the protein half-life was long in hypoxic cells in contrast with rapid disappearance when cells are reoxygenated [101]. This suggested that the primary mode of regulation is posttranslational, occurring through changes in protein stability. In fact, the protein has a very short half-life in normoxia (half-life of few minutes) and becomes much more stable in hypoxia (half-life greater than 12 hr). Many independent experiments have indicated that the beta subunit is not affected by oxygen levels and the alpha subunit is stabilized in hypoxia, thus increasing the amount of functional HIF-1 [100,102]. C. Transferrable Degradation Domains

Which domains of HIF-1α are involved in oxygen-dependent instability? This question has been addressed by fusing HIF-1α sequences to the DNA binding domain of a heterologous transcription factor; for example, yeast Gal4 [133– 135]. Such fusion proteins can be assayed for oxygen-dependent properties either functionally (using a Gal4 UAS-linked luciferase reporter) or for effects on stability (by immunoblotting). By this strategy, two regulatory domains have been defined in HIF-1α, one lying adjacent to the C terminal transactivation domain (CAD), and another covering a substantial internal portion of the molecule from amino acids 400–600 and overlapping the N terminal transactivation domain (NAD) [133,134]. Analysis of the minimal sequences required for regulatory behavior defined regions adjacent or overlapping the NAD (amino acids 549– 582) and CAD (amino acids 775–826), domains which were sufficient to confer some level of oxygen-regulated activity [133–136]. Both NAD and CAD–Gal4 fusions were able to induce the UAS-luciferase reporter in an oxygen-dependent manner [133–135]. Nevertheless, Western blot analysis of the chimeras indicated that NAD- and CAD-related regulatory domains operate through very different mechanisms, since only the former is stabilized in hypoxia [134,136]. In contrast, CAD-mediated induction involves posttranslational modifications that enhance transcriptional activity irrespective of the protein level [136] (discussed on pages 115–117).

110

Wappner and Ratcliffe

EPAS-1 (HIF-2α), a second regulatory protein in the hypoxic response, was also subjected to structure–function studies and found to have a similar organization [137]. As for HIF-1α, separable domains related to both the N and C terminals were found to be oxygen responsive. Again similar to HIF-1, oxygen concentrations modulate protein levels and transcriptional activity through the two domains independently [137]. One other interesting characteristic was shared by the two proteins. In each case, the internal domain overlapping the NAD (and regulated through protein stability) could be divided into portions which could operate independently to confer spatial instability [137]. A Drosophila protein called Similar (Sima) appears to be a HIF-1 homolog [138]. Sima has an overall 45% identity with mammalian HIF-1α (the highest among Drosophila bHLH-PAS proteins) and a 63% identity in the basic and helix loop helix domains. Sequence homology between Sima and HIF-1α is reinforced by a clear functional homology [139]. Sima is expressed at low levels in Drosophila SL2 cells and could not be detected by Western blotting of crude extracts. However, when immunoprecipitates of SL2 extracts were prepared using Sima antisera and subjected to Western blot, an increase of Sima levels in hypoxia could be detected [139]. In order to answer whether Sima regulation is posttranscriptional and try to define which domains of Sima are responsible for hypoxic induction, a Gal4 fusion approach similar to that used for HIF-1 was followed [134,135]. As indicated above, amino acid sequence reveals the highest level of conservation in the N terminal region [138]. Although much lower than in the bHLH and PAS regions, significant homology is also observed between the C terminal of HIF-1α and an internal domain of Sima. In contrast, most of the C terminal region of Sima corresponding approximately to the last 500 amino acids (residues 1000–1505) has no homology with HIF-1α [139]. In a similar analysis to that performed for HIF-1α, a deleted form of Sima lacking the basic and HLH domains was fused to the Gal4 DNA binding domain and cotransfected in Hep3B (mammalian) cells together with an UAS–luciferase (Gal4 binding) reporter. The chimera was able to drive hypoxia-dependent induction of the reporter, showing that Sima regulation occurs at the protein level. Further fusions to several deleted forms of Sima extend the paradigm of structure–function conservation with HIF. Deletions to amino acids 527, 665, and 816, all of which removed the bHLH and PAS domains of Sima but retained differing extents of the central region, showed inducible activity. In contrast, two deletions reaching amino acids 1043 and 1135, which lacked the whole internal domain, were not hypoxia responsive [139]. Unexpectedly, fly lines bearing a chromosomal deficiency that covers the sima gene showed an unchanged pattern of induction of the bHLH-PAS LacZ reporter that parallels branchless expression (P. Wappner, unpublished data) (discussed above). This result indicates that the unknown bHLH-PAS protein that regulates branchless is not Sima. Further research is required in order to define

Branching Morphogenesis and Hypoxia

111

the physiological role of Sima and the precise function of the unknown bHLHPAS protein that regulates the expression of branchless in Drosophila. D. Involvement of Ubiquitin-Mediated Proteasomal Pathway in HIF-1␣ Degradation

Experiments with specific inhibitors of the proteasome have indicated that the ubiquitin proteasome pathway is involved in HIF-1α oxygen-dependent degradation [133,140,141]. Treatment of mammalian cells with either lactacystin or MG132 induced high levels of HIF-1 regardless of the oxygen concentration. Involvement of the ubiquitin proteasome pathway in HIF-1 degradation was also recently confirmed by immunoprecipitation and Western blot experiments in which ubiquitin/HIF-1 adducts were detected [133,141]. The ubiquitin/proteasome pathway plays a key role in many basic cellular processes. In addition to degradation of defective or misfolded proteins, a critical regulatory role for this system has been defined in studies of the cell cycle. Strict control of the destruction of factors that regulate passage through control points in the cycle is crucial for normal cell division and proliferation. Many of these controls involve protein phosphorylation and regulatory mechanisms include both synthesis and degradation of cyclins and cyclin-dependent kinases inhibitors (CdKs). Such degradation occurs through the ubiquitin/proteasome pathway [142–154]. Similarly, some major signal transduction pathways which are of great importance in developmental and cell differentiation are known to be controlled in the same way. The nuclear factor-κB (NF-κB) pathway is controlled at the level of I-κ B proteolysis [155]. The Hedgehog pathway that controls segmentation and appendage development in Drosophila as well as many developmental processes in mammals is regulated by the proteolysis of the Cubitus interruptus protein that transduces the Hedgehog signal intracellularly [156–161]. Wnt/Wingless, another major signaling pathway involved in development and oncogenesis, is controlled by the ubiquitin/proteasome pathway. In this case, βcatenin is the target [162–165] and the adenomatous polyposis coli (APC) gene product is required for its degradation [166–168]. Virtually all types of colorectal cancer derive from abnormally high levels of β-catenin owing to defects in degradation [169–170). As with signaling molecules, the list of transcription factors whose half-life is controlled by the ubiquitin-proteasome pathway has enlarged dramatically over the past few years. It now includes NF-κB, Fos, Jun, Myc, IPR2, and HIF-1α among others [171–176]. How do eukaryotic cells decide which proteins are due to enter the ubiquitin/proteasome pathway and when such proteins should do so? This major question is currently the focus of a great deal of attention. In most cases, the target protein must be conjugated to ubiquitin in order to be recognized and degraded by the 26S proteasome [175]. Ubiquitin is conjugated to the substrate through an isopeptidic bond between the carboxy terminal amino acid of ubiquitin and spe-

112

Wappner and Ratcliffe

cific lysines of the target protein. Typically, the conjugation reaction occurs in three successive steps [177]. First, ubiquitin is conjugated through the C terminal residue to an ubiquitin-activating enzyme (E1). The reaction requires ATP and results in the formation of a thioester bond. Second, ubiquitin is transferred from the E1 enzyme to a second enzyme termed ubiquitin-conjugating enzyme (E2) in a reaction in which transference of the thioester bond occurs. Finally, ubiquitin is transferred to the substrate by a (E3) protein–ubiquitin ligase. E3 enzymes catalyze the formation of a covalent bond between epsilon amino groups of lysines in the substrate and ubiquitin C terminal. The E2 enzyme remains as part of the complex once conjugation was completed. In most cases, the sequential action of E1, E2, and E3 is repeated many times on the same substrate molecule, determining the formation of a polyubiquitin chain that is recognized by the 26S proteasome, functioning as a strong degradation signal [175–177]. Many laboratories have demonstrated that ubiquitin conjugation is the step which defines target specificity for degradation [178]. More recently, several reports showed that E3 ubiquitin-protein ligases are the enzymes responsible for such specificity [179]. It has not been straightforward to identify E3 enzymes by sequence homology, since they constitute a highly divergent family. Although some E3 enzymes are single polypeptides [175,180,181], work from many groups in the field suggests that multiprotein complexes are most common. The halflife of cell cycle regulators depends on two kinds of multiprotein complexes with E3 enzymatic activity. These are the cyclosome or anaphase-promoting (APC) complex (182) and the SCF complex [147]. In the cell cycle, the cyclosome controls the half-life of proteins that regulate the transition from metaphase to anaphase, whereas the SCF complexes are involved in the transition from G1 to S phase [147,149,151]. Recent studies on SCF complexes have shed light on the molecular basis of specificity in ubiquitin conjugation. SCF complexes (initially described in yeast) are composed of three proteins, each with homologues in humans and in Caenorabditis elegans [148,154,183,184]. These three proteins are Skp1, Cdc53/ Cullin1 and a third variable protein that belongs to a family defined by a consensus sequence called the F-box (S C F ). Cdc53/Cul1 interacts simultaneously with Skp1, with a specific E2 enzyme, and with an F-box protein, fulfilling the function of scaffold for the whole complex [145,154]. Recent publications reported the occurrence of a fourth protein in the complex. This new component has been termed Rbx/ROC. It is evolutionary conserved and was shown to be essential for ubiquitin-protein ligase activity [185–189]. Thus, the SCF complex contains three polypeptides (Skp1, Cul1, and Rbx/ ROC), which appear to be common to many degradation complexes and one variable component (F-box protein), which is specific to a particular target or small group of targets. Evidence from several laboratories indicates that F-box proteins behave as adapters between substrate proteins and the rest of the complex

Branching Morphogenesis and Hypoxia

113

[183]. Thus, the so-called ‘‘combinatorial F-box hypothesis’’ proposes that many different substrates can be recognized by various SCF complexes that differ in their F-box adapter protein [162,183,190,191]. Despite this advance in understanding the specificity of proteolysis, understanding how precise regulation of the process is achieved still presents a challenge. In some cases, proteins must undergo posttranscriptional modifications, like phosphorylation or redox changes, in order to be recognized by the E3 complex and ubiquitinated [175,192]. Such mechanisms are clearly important in determining the timing of degradation in the cell cycle and may also mediate responses to a specific physiological stimulus. A recent publication gave an unexpected twist to this interesting and rapidly evolving field. In this work, the Von Hippel–Lindau (VHL) tumor suppressor protein was shown to be critical for HIF-1α proteolysis [193]. It therefore appears that VHL controls HIF-1α degradation by the 26S proteasome and is a key factor in the pleiotropic response of cells to hypoxia. Von Hippel–Lindau disease is a dominantly inherited cancer syndrome characterized by a predisposition to develop several kinds of highly vascularized tumors [194]. The most frequently reported tumors are the sporadic clear cell renal carcinoma (RCC), retinal angiomas, central nervous system hemangioblastomas, and pheochromocytomas [195,196]. The VHL-associated tumors most probably owe their highly angiogenic phenotype to constitutive upregulation of the normally hypoxia inducible gene encoding VEGF [197]. In the recent work, it was shown that a whole battery of hypoxia-responsive genes is upregulated in a VHL mutant cell line regardless of oxygen levels. Such a battery of genes include VEGF, Glut-1, TGF-β1, ALDA, PGK-1, PFK-C, and LDH-A [193]. HIF-1 itself was examined by EMSA in the VHL mutant line and showed high levels of HIF-1/DNA complex formation regardless of oxygen levels. When the cell line was stably transfected with a plasmid encoding wild-type VHL, HIF-1/DNA complexes disappeared from normoxic cells and were induced in hypoxic conditions only [193]. In keeping with this, Western blot analysis of a series of VHL mutant cell lines revealed high levels of both HIF-1 and EPAS-1, which are known to be regulated by a common mechanism. Moreover, immunoprecipitation experiments revealed physical interactions between VHL and HIF-1α subunits in cells treated with proteasomeblocking agents [193]. Most interestingly, the interaction between VHL and HIF-1α was shown to be iron dependent and was not observed in cells treated with desferrioxamine or cobalt. This provides a possible explanation as to how these agents stabilize HIF-1 and upregulate the system. It also has implications for models of the oxygen sensing mechanism (see below). While writing this review, another interesting paper was published confirming the participation of VHL in an E3 ubiquitin–protein ligase complex [198]. It is reported that VHL promotes ubiquitination in vivo while forming part of a complex that includes components related to the SCF complex. The new E3 complex includes elongins B and C, with remarkable homology with ubiquitin and

114

Wappner and Ratcliffe

Skp1, respectively; Cullin2, which is very similar to Cullin-1, and finally Rbx1, which is a common component with the SCF complex. As expected, ubiquitin and E2s are also found in such new complexes. On this basis, a new kind of E3 complex has been defined and termed VCB (Von Hippel–Lindau, Cullin2, elongin BC) [198–200] (Fig. 6). HIF-1 is therefore a substrate for ubiquitination by the VCB/E3 ligase complex (N. Masson and P.J. Ratcliffe unpublished data). Ubiquitination of HIF-1 occurs in an oxygen-dependent manner, but unlike treat-

Figure 6 Comparison of SCF and VCB E3 ligase complexes. A common general structure seems to underlie the two different E3 ligase complexes that catalyze specific ubiquitination of many cellular proteins. Ubiquitination involves a cascade of reactions in which ubiquitin is passed from an E1 ubiquitin-activating enzyme to an E2 ubiquitin-conjugating enzyme and finally to the protein substrate in a reaction catalyzed by an E3 ubiquitin ligase complex. SCF and VCB are two of such complexes. Both complexes share a common general architecture. Cul1/Cdc53 is highly homologous to Cul2; Skp1 that functions as a scaffold in SCF complex is structurally homologous to elongin B-C and Rbx1 is common to SCF and VCB complexes. Each complex interacts with a set of different adapter proteins that mediates recognition of protein substrates through specific protein– protein interactions. In SCF complexes, adapter proteins exhibit a typical protein domain called F-box that allows interaction with Skp1. In VCB complexes the adapters are proteins bearing SOCS box domains.

Branching Morphogenesis and Hypoxia

115

ment with iron chelators or cobalt, oxygen availability does not appear to affect the association between HIF-1α and pVHL. Possible mechanisms by which oxygen controls the ability of VCB to ubiquitinate HIF-1 will be discussed below. As in all other eukaryotic systems, the ubiquitin/proteasome pathway functions in Drosophila melanogaster. Several genes encoding ubiquitin as well as various E1 and E2 enzymes have been identified in Drosophila [201]. Although SCF complexes have not been reported so far, two proteins bearing an F-box motif were described [160,202,203]. One of such proteins, Slimb, was characterized in detail and careful genetic and functional studies were performed. As with mammals, Slimb mediates proteolytic control over several important signal transduction pathways like Wingless, Hedgehog, and NF-κ-B. Genetic and molecular experiments have revealed that the mechanism of action of Slimb involves activation of proteolysis of different components that participate in such signaling pathways [160,204]. The conservation of Slimb function in Drosophila provides a remarkable example of how molecular basic mechanisms have been used repeatedly in evolution in order to control homologous developmental programs even when morphological outcomes are so diverse. E. Oxygen Regulation of HIF-1 by Nonproteolytic Mechanisms

As outlined above, oxygen-dependent control of proteolytic degradation by the ubiquitin-proteasome system is not the only regulatory process affecting HIF-1 activation (Fig. 7). Although its function was not at first recognized, HIF-1α was cloned independently as a protein interacting with the transcriptional coactivator P300 [205]. Several papers have now demonstrated a role for P300 transcription factor and its homolog CREB-binding protein (CBP) in HIF-1 hypoxic induction [127,136,205,206]. Physical interaction between HIF-1 and P300/CBP has been demonstrated both in vivo and in vitro. EMSA experiments showed that P300/ CBP existed within the HIF-1/DNA complex. Functional assays indicated that cells transfected with P300/CBP show an enhanced inducibility of the erythropoietin (Epo) enhancer in response to hypoxia [127,136,205,206]. Furthermore, the E1A oncoprotein is known to target and inhibit P300/CBP. Consistent with this, E1A transfection inhibits hypoxia-induced VEGF and Epo transcription [205]. P300 interacts with the C terminal activation domain (CAD) and probably also with the NAD [136]. This interaction appears to be regulated by oxygen independently from HIF-1α proteolysis. There is some uncertainty as to whether coactivator recruitment is essential for transcriptional activation. Although inhibitors of proteasomal function stabilize HIF-1 in normoxia, they do not cause transcriptional activation. It has therefore been suggested that this indicates that other processes (such as coactivator recruitment) are essential for activation. However, the result is difficult to interpret because of nonspecific effects of these compounds on transcription. Other experiments indicate that stabilization of HIF-1α

116

Wappner and Ratcliffe

Figure 7 Integrated model of HIF-1 regulation by oxygen levels. In normoxia, HIF1α subunits are ubiquitinated and degraded by the 26S proteasome. The ubiquitination reaction is probably catalyzed by VCB/E3 ligase complex which is composed of the Von Hippel–Lindau protein, Cullin 2, Rbx-1 and elongin B-C. In hypoxia, it is proposed that ubiquitination is prevented, HIF-1α accumulates in the nucleus, dimerizes with the beta subunit, and undergoes conjugation with P300 transcription factor. Conjugation with P300 is redox dependent and involves reduction of specific disulfide bonds catalyzed by thioredoxin. Thioredoxin nuclear entry is also dependent on hypoxia. Activated HIF-1 dimer is a pleiotropic inducer of the hypoxic response. Oxygen-dependent genes include those encoding most glicolytic enzymes, glucose transporters, erythropoietin, transferrin, vascular endothelial growth factor (VEGF), tyrosine hydroxilase, and endothelial nitric oxide synthase. HIF-1 also mediates stabilization of P53 protein.

Branching Morphogenesis and Hypoxia

117

is sufficient for at least some degree of transactivation. For instance, removal of the degradation domain of HIF-1α results in a constitutively active protein [133], whereas stabilization of HIF-1α subunits in the context of pVHL defective cells is sufficient to activate HIF-1 target genes [193]. Most probably, coactivator recruitment serves to modify and amplify the hypoxia inducible response (see Fig. 7). The mechanism of oxygen-dependent recruitment however remains unclear. The results of recent experiments have indicated that thiol-based redox changes are likely to be involved in the response to hypoxia. Exposure of cells to the thiol-reductive agent N-(2-mercaptopropionyl)-glycine (NMPG) activates the system even in normoxia [140,207]. This is consistent with previously reported data showing that the reducing enzymes thioredoxin and Ref-1 potentiate hypoxia-induced gene transcription [101,207,208]. These agents promote the interaction of the HIF-1α CAD with P300/CBP as assessed from the activity of a Gal4 fusion to the HIF-1α CAD [136]. Thiol active agents also affect the interaction of the HIF-1α CAD with P300 in vitro suggesting that they might act directly on the CAD. Mutation of the single conserved cysteine in the CAD to serine abolished both the interaction with CBP/P300 and transactivation supporting this view [136]. However, in another study, this cysteine was mutated to alanine; hypoxia inducible activity of the isolated CAD persisted [137], indicating that effects of sulfydryl agents are not likely to be dependent on that residue and seem to be indirect. An unexpected additional piece of data also seems likely to be relevant to this process. Thioredoxin was found to translocate to the nucleus in response to hypoxia in several different cell lines, although the precise connection with HIF-1 activation is unclear [136]. Overall, the mechanism of HIF-1 activation is therefore complex with at least three processes contributing to transcriptional activation by the hypoxic stimulus. 1. VCB-dependent degradation by the proteasome is inhibited, rendering the alpha subunits stable. The regulatory trigger is unknown, although some oxygen-dependent modification of the degradation domain is currently the favored candidate. 2. HIF-1 and thioredoxin are translocated to the nucleus; again the mechanism of such oxygen-dependent nuclear localization is unclear, although critical sequences in HIF-1α have been defined [206]. 3. HIF-1 interacts with P300/CBP coactivators through its activation domains. These molecules themselves possess transcriptional activation domains. They also possess histone deacetylase activity, and so may have additional actions on endogenous gene expression from chromatinized DNA.

118

Wappner and Ratcliffe

Functional studies on Drosophila Sima and other oxygen-responsive bHLH-PAS proteins will be important in addressing whether these molecular aspects of the hypoxic response are conserved in Drosophila. F. Oxygen Sensing and Signal Transduction: Some Clues and Many Open Questions

As discussed above, the mechanism responsible for HIF-1 activation are quite well understood. Much less is known about upstream events that trigger the hypoxic response. The specificity of the response for hypoxia (as opposed to more broadly based responses to cellular stressors) has led to the concept of a distinct oxygen sensor. Such a molecule is proposed to undergo biochemical modification in an oxygen-dependent manner, leading to activation of a signaling process that culminates in HIF-1–dependent gene expression [100,209]. Several such models have been proposed, and quite possibly cellular oxygen-sensing processes involve a number of different mechanisms perhaps interacting with HIF-1 through different signaling pathways. We will discuss briefly some of the current hypotheses. Studies in bacteria and yeast lend strong support to the concept of specific sensing systems operating on hierarchical cascades of gene regulation. Molecules have been defined which respond to aspects of oxygen metabolism such as a particular redox couple or the level of a particular oxygen radical species. However, at least one class of sensing system interacts directly with dioxygen and therefore corresponds to the concept of an oxygen sensor in its purest form. In nitrogen-fixing symbiotic bacteria such as Rhizobium meliloti, genes expressed late in nodule formation are induced by hypoxia. In this case, the sensing and transduction mechanisms have been well characterized both from a genetic and from a biochemical point of view [210]. The oxygen-regulated response depends on a two-component signal transduction system. A protein termed FixL functions as the oxygen sensor [211]. FixL presents two distinct domains: one domain (a PAS sensor domain; see above) close to the N terminal binds a heme prostetic group, whereas the other domain, adjacent to the C terminal, has histidine kinase activity (transmitter domain). Oxygen behaves as an allosteric modifier of FixL. Hence, at low oxygen levels, histidine kinase activity is maximal and FixL undergoes autophosphorylation on a specific histidine residue of the catalytic domain. Thereafter phosphate is transferred to an aspartate residue of the second component of the system, a protein called FixJ. Phosphorylation converts FixJ into an active transcription factor that is now able to induce downstream hypoxia-responsive genes [212–214]. Interestingly, sometime before this system was characterized, certain characteristics of the response of the erythropoietin gene to hypoxia had led to the proposal of heme-containing protein as the mammalian oxygen sensor [215]. Cells treated with cobalt, nickel, manganese, or the iron chelator deferoxamine

Branching Morphogenesis and Hypoxia

119

(DFO) mimic the hypoxic response [215]. Cobalt and DFO function reliably as inducers which mimic low oxygen in many different contexts. Both agents are able to elicit the formation of HIF-1/DNA complexes in EMSA experiments and both are able to induce hypoxia-dependent gene expression [100,102,113]. These data represent very strong evidence that at least some form of iron center is involved in sensing or transducing hypoxia. In one particular model, it has been proposed that cobalt or nickel could replace iron in the heme group, thus fixing the putative heme-containing protein in a deoxy state and triggering hypoxic gene expression. In such a model, DFO treatment of cells might lead to production of an apolipoprotein form of the protein that is also able to generate the deoxy signal constitutively [100,215]. In keeping with this hypothesis, it was demonstrated that cobalt is a substrate for ferrochelatase, the enzyme that incorporates iron during the synthesis of heme [216]. However, DFO is unable to chelate heme iron, and if such a model is correct, then it seems likely that the action is on the synthesis of a rapidly turning over protein rather than interference with heme iron after synthesis. Evidence in favor of the heme sensor hypothesis has been provided by the ability of carbon monoxide (CO) to inhibit the hypoxic response [217]. Cobalt or nickel-substituted heme groups are unable to bind CO, and CO does not affect the response to these agents [215]. However, not all groups have reported a major effect of CO on the response to hypoxia. In one recent study, cells treated with inhibitors of heme biosynthesis did not show impairment of the hypoxic response and the effect of CO was reported to be only minor [218]. Another line of evidence points toward hydrogen peroxide as a chemical mediator of the hypoxic response [219]. Hydrogen peroxide levels diminish in hypoxia [219], and overexpression of catalase can increase expression of hypoxia-dependent transcription [220]. In this model, it has been proposed that the hydrogen peroxide is derived from superoxide by the action of superoxide dismutase and that an iron center catalyzes the conversion of hydrogen peroxide to more reactive oxygen intermediates (ROIs) such as hydroxyl radical and singlet oxygen [221]. Such a reaction is known as the Fenton reaction [222]. ROIs might operate on HIF-1 distantly by the activation of a signal transduction cascade. Alternatively, a local Fenton reaction arising at an iron center within the HIF-1 complex might oxidize particular amino acids, thus triggering the degradation mechanism. Indirect support for such a possibility has recently been provided by studies of the HIFα/pVHL interaction [193]. Although this interaction appears to be independent of oxygen, the proteins are dissociated in cells treated with iron chelators or cobalt ions. This indicates that the mechanism of stabilization of HIF in response to hypoxia and iron chelation are different. Conceivably this might indicate that existence of two separate pathways of HIF induction both acting remotely on the VCB complex. Another explanation is that (as proposed above) the point of interaction is at the site of sensing and therefore that this

120

Wappner and Ratcliffe

center is local to the complex. Such an iron center might interact directly with dioxygen. Alternatively (as outlined above), it might react with partially reduced species such as superoxide or hydrogen peroxide. In the latter case, what is the source of these species? A widely expressed low-output isoform of the neutrophil NADPH oxidase has been proposed [221]. In mutant cells defective for the neutrophil form, the hypoxia inducible response is unaffected, indicating that this enzyme itself is not critical [223]. Interestingly closely related genes have recently been described, and they remain candidates [224]. Others have proposed that mitochondrial production of oxygen radicals is responsible for the signal [225]. It has long been clear that mitochondrial inhibitors such as cyanide do not mimic hypoxia in respect of HIF activation, indicating that high-energy phosphate levels themselves are unlikely to be relevant to the sensing process [102,209]. However, recent work has suggested that mitochondrial function might be involved in sensing in quite a different way. In yeast [226] and mammalian cells [225], mitochondrial inhibitors and mutants which are defective at several different points in the electron chain appear to be unable to induce certain hypoxia-responsive genes. Some signal from the mitochondria therefore appears to be required, and the authors proposed that the molecular sensor could be the electron transport chain complex IV (cytochrome c oxidase) [226]. The nature of the signal is unclear. The authors found the effect to persist even in complete anoxia, and they therefore proposed that oxygen radical species were unlikely to be involved, although presumably other radicals associated with ‘‘reductive stress’’ would not be excluded by this finding [226]. In mammalian cells, somewhat different data have been reported. Here inhibitors of complex I of the respiratory chain were noted to lead to suppression of the hypoxic response [225]. Similar effects were observed in rho zero cells which have been depleted of mitochondrial DNA but not with pharmacological inhibitors acting at other sites in the electron transport chain. It was proposed that the signal arose from increased production of reactive oxygen species from site 1 in hypoxia [225]. This model is at odds with ideas arising from the effects of hydrogen peroxide where the reverse action of radical species has been proposed, and further work is clearly needed to resolve the discrepancies. There is at present little evidence as to whether any of these mechanisms participate in the Drosophila hypoxic response. The Drosophila bHLH-PAS protein Sima was tested for its ability to convey a hypoxic response when expressed as a Gal4 fusion in mammalian cells [139]. In this context, Sima behaved very much like mammalian HIF-1. Cobalt and DFO mimicked the hypoxic response, suggesting that Drosophila sequences were able to interact in some way with the mammalian system [139]. A recent remarkable paper has indicated a likely role for the nitric oxide synthase (NOS)/protein kinase G (PKG) pathway in oxygen sensing in Drosophila [227]. Nitric oxide (NO) is a highly reactive compound that targets soluble

Branching Morphogenesis and Hypoxia

121

guanylate cyclase (sGC), which is a heme-containing protein [228–232]. Soluble GC hydrolyzes GTP to produce cGMP. cGMP targets and activates PKG, which in turn phosphorylates several unknown downstream proteins. In this report [227], a wide spectrum of biological responses to hypoxia was analyzed and shown to be affected in flies in which the NOS/PKG pathway was either increased or reduced through genetic or pharmacological manipulations. Such responses included behavioral, developmental, cellular, and biochemical effects. Drosophila larvae display a striking behavior in response to low oxygen levels that consists of interruption of feeding and moving away from the culture medium (roving behavior). The authors observed that PKG mutants exhibit impairment in the hypoxia-triggered roving behavior [227]. Instead of leaving the culture medium, the mutant larvae remain inside and continue feeding. A similar abrogation of the roving behavior in hypoxia was observed when larvae were fed with L-NAME, an inhibitor of NOS. Conversely, when the pathway was hyperactivated by ectopic expression of NOS in a transgenic line, the roving behavior was enhanced significantly. In another set of experiments, the NOS/PKG pathway was shown to affect embryonic survival in hypoxia. Eight-hour-old wild-type embryos survive for several days in severe hypoxia. In contrast, PKG mutant embryos are much more sensitive and only a small fraction survive [227]. These interesting results were complemented with studies on the cell cycle. Wild-type embryos subjected to hypoxia arrest development and resume it only when they are returned to normoxia. The underyling mechanism was found to be due to a block in S phase of the cell cycle. The NOS/PKG pathway was also found to be involved in this hypoxia-mediated block. NOS overexpression in a transgenic line or treatment of embryos with a hydrolysis-resistant analog of cGMP blocked the cycle at S phase. Accordingly, inhibition of NOS with L-NAME alleviates the block in S phase imposed by hypoxia. Similarly, PKG mutants were not blocked in S phase at low oxygen levels [227]. The work was completed with observations on the final stages of tracheal development. As discussed above, development of terminal tracheal branches is plastic and depends largely on local oxygen concentrations [1–3]. The NOS/ PKG pathway was shown to be involved in this hypoxia-dependent process as well. The PKG mutant and L-NAME–treated larvae have a diminished number of terminal branches when compared with the controls. In contrast, the number of terminal branches in the transgenic line overexpressing NOS was increased [227]. Further research is required in order to address whether the pathway modifies directly HIF-1/Sima–dependent transcription of hypoxia inducible genes. The fact that NOS and sGC are both heme-containing proteins raises the interesting possibility that one or both eyzymes may function as a molecular sensor. Interestingly, sGC is activated by NO and reversibly inhibited by oxygen [233–

122

Wappner and Ratcliffe

234]. This is consistent with a mechanism in which oxygen regulates the enzymatic activity and NO functions as a modulator that fine tunes the threshold for oxygen sensitivity in response to particular physiological requirements. It is still unclear how oxygen binds to sGC, but binding to the heme group has been reported to be unlikely. Instead, oxygen seems to affect some critical sulfhydril groups at the active site of the enzyme. Whereas reduction of such sulfhydril groups would render the enzyme active (hypoxia), oxidation would make it inactive (normoxia/hyperoxia). Although several groups have described interactions of NO with HIF activation in mammalian cells, further research is required to determine the extent of similarities with the Drosophila pathway.

IV. Conclusions Drosophila tracheal development shares many features with branching morphogenesis in mammals. First, there is striking conservation between molecular mechanisms underlying tracheal development and lung branching morphogenesis. The genetic control of branching during lung development is determined by localized expression of FGF-10 in the mesenchyme that stimulates its receptor on the growing bronchial branch. Sprouty expressed at the tip of the branch antagonizes FGF activation on the neighboring cells. Similarly, during Drosophila tracheal development, Breathless/FGF receptor is expressed in the growing tracheal branches and is stimulated by the ligand Branchless which is expressed in the surrounding tissues and guides branch migration. Sprouty is again the inhibitor of branching on the cells that are not destined to emit ramifications. Mammalian vascular development also has certain elements in common with developmental processes in insect tracheae. One of the most conserved features is seen in the ability of insect terminal tracheal branches to respond to hypoxia, which is very much like mammalian blood capillaries during tumor angiogenesis and in the collateral response to ischemia. Molecular mechanisms underlying such a hypoxic response and oxygendependent gene expression are highly conserved as well. Mammalian VEGF and fly Branchless/FGF are both induced by local hypoxia, and guide respectively the outgrowth of capillaries or tracheal terminal branches that supply oxygen to the hypoxic area. The flanking sequences of Drosophila genes whose mammalian homologs show hypoxic induction bear conserved motifs typical of oxygenresponsive genes, although hypoxic induction of such Drosophila genes still remains to be confirmed. As in mammals, hypoxia in Drosophila induces the accumulation of at least one nuclear protein of the bHLH-PAS family that is able to bind an oligonucleotide derived from a mammalian oxygen-responsive enhancer. Such a Drosophila

Branching Morphogenesis and Hypoxia

123

bHLH-PAS protein could be the one encoded by the similar (sima) gene, although it seems likely that at least one other protein is involved. Among all the Drosophila bHLH-PAS proteins so far defined, Sima has the highest homology with mammalian HIF-1. In keeping with this, Sima sequences can convey hypoxia inducible responses in a mammalian cell line, implying the potential for interaction with the mammalian pathway. Drosophila has proved to be a very suitable genetic model for studying developmental processes such as the formation of tubular networks or for dissecting the hypoxic pathway at the molecular level. In the near future, genetic models like Drosophila melanogaster are expected to open new exciting perspectives in the study of developmental and physiological aspects in the adaptation of organisms to variations in oxygen availability.

References 1. Krasnow MA. Genes that control organ form: lessons from bone and branching morphogenesis. Cold Spring Harb Symp Quant Biol 1997; 62:235–240. 2. Shilo BZ, Gabay L, Glazer L, Reichman-Fried M, Wappner P, Wilk R, Zelzer E. Branching morphogenesis in the Drosophila tracheal system. Cold Spring Harb Symp Quant Biol 1997; 62:241–247. 3. Metzger RJ, Krasnow MA. Genetic control of branching morphogenesis. Science 1999; 284:1635–1639. 4. Hogan BL, Yingling JM. Epithelial/mesenchymal interactions and branching morphogenesis of the lung. Curr Opin Genet Dev 1998; 8:481–486. 5. Manning GE, Krasnow MA. Development of the Drosophila tracheal system. In: Bate M, Martinez Arias A, eds. The Development of Drosophila melanogaster. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory Press, 1993, pp 609– 685. 6. Guillemin K, Groppe J, Ducker K. Treisman R, Hafen E, Affolter M, Krasnow, MA. The pruned gene encodes the Drosophila serum response factor and regulates cytoplasmic outgrowth during terminal branching of the tracheal system. Development 1996; 122:1353–1562. 7. Wilk R, Weizman I, Shilo BZ. trachealess encodes a bHLH-Pas protein that is an inducer of tracheal cell fates in Drosophila. Genes Dev 1996; 10:93–102. 8. Isaac DD, Andrew DJ. Tubulogenesis in Drosophila: a requirement for the trachealess gene product. Genes Dev 1996; 10:103–117. 9. Sonnenfeld M, Ward M, Nystrom G, Mosher J, Stahl S, Crews S. The Drosophila tango gene encodes a bHLH-PAS protein that is orthologous to mammalian ARNT and controls CNS midline and tracheal development. Development 1997; 124: 4571–4582. 10. Zelzer E, Wappner P, Shilo BZ. The PAS domain confers target gene specificity of Drosophila bHLH-PAS proteins. Genes Dev 1997;2079–2089. 11. Ohshiro T, Saigo K. Transcriptional regulations of breathless FGF receptor gene

124

12.

13.

14.

15. 16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

Wappner and Ratcliffe by binding of TRACHEALESS/dARNT heterodimers to three central midline elements in Drosophila developing trachea. Development 1997; 124:3975–3986. Anderson MG, Certel SJ, Certel K, Lee T, Montell DJ, Johnson WA. Function of the Drosophila POU domain transcription factor drifter as an upstream regulator of Breathless receptor tyrosine kinase expression in developing trachea. Development 1996; 122:4169–4178. Certel K, Anderson MG, Shrigley RJ, Johnson WA. Distinct variant DNA-binding sites determine cell-specific autoregulated expression of the Drosophila POU domain transcription factor drifter in midline glia or trachea. Mol Cell Biol 1996; 16: 1813–1823. Anderson MG, Perkins GL, Chittick P, Shrigley RJ, Johnson WA. Drifter, a Drosophila POU-domain transcription factor, is required for correct differentiation and migration of tracheal cells and midline glia. Genes Dev 1995; 9:123–137. Goto S, Hayashi S. Specification of the embryonic limb primordium by graded activity of Decapentaplegic. Development 1997; 124:125–132. Llimargas M, Casanova J. ventral veinless, A POU domain transcription factor, regulates different transduction pathways required for tracheal branching in Drosophila. Development 1997; 124:3273–3281. Llimargas M, Casanova J. EGF signalling regulates cell invagination as well as cell migration during formation of tracheal system in Drosophila. Dev Genes Evol 1999; 209:174–179. Samakovlis C, Hacohen N, Manning G, Sutherland DC, Guillemin K, Krasnow MA. Development of the Drosophila tracheal system occurs by a series of morphologically distinct but genetically coupled branching events. Development 1996; 122:1395–1407. Wappner P, Gabay L, Shilo BZ. Interactions between the EGF receptor and DPP pathways establish distinct cell fates in the tracheal placodes. Development 1997; 124:4707–4716. Sutherland D, Samakovlis C, Krasnow MA. Branchless encodes a Drosophila FGF homolog that controls tracheal cell migration and the pattern of branching. Cell 1996; 87:1091–1101. Reichman-Fried M, Shilo BZ. Breathless, a Drosophila FGF receptor homolog, is required for the onset of tracheal cell migration and tracheole formation. Mech Dev 1995; 52:265–273. Reichman-Fried M, Dickson B, Hafen E, Shilo BZ, Elucidation of the role of breathless, a Drosophila FGF receptor homolog, in tracheal cell migration. Genes Dev 1994; 8:428–439. Klambt C, Glazer L, Shilo BZ. Breathless, Drosophila FGF receptor homolog, is essential for migration of tracheal and specific midline glial cells. Genes Dev 1992; 6:1668–1678. Glazer L, Shilo BZ. The Drosophila FGF-R homolog is expressed in the embryonic tracheal system and appears to be required for directed tracheal cell extension. Genes Dev 1991; 5:697–705. Lee T, Hacohen N, Krasnow M. Montell DJ. Regulated Breathless receptor tyrosine kinase activity required to pattern cell migration and branching in the Drosophila tracheal system. Genes Dev 1996; 10:2912–2921.

Branching Morphogenesis and Hypoxia

125

26. Gabay L, Seger, R, Shilo BZ. MAP kinase in situ activation atlas during Drosophila embryogenesis. Development 1997; 124:3535–3541. 27. Gabay L, Seger R, Shilo BZ. In situ activation pattern of Drosophila EGF receptor pathway during development. Science 1997; 277:1103–1106. 28. Hacohen N, Kramer S, Sutherland D, Hiromi Y, Krasnow MA. Sprouty encodes a novel antagonist of Fgf signaling that patterns apical branching of the Drosophila airways. Cell 1998; 92:253–263. 29. Kramer S, Okabe M, Hacohen N. Krasnow MA, Hiromi Y. Sprouty: a common antagonist of FGF and EGF signaling pathways in Drosophila. Development 1999; 126: 2515–2525. 30. Reich A, Sapir A, Shilo B. Sprouty is a general inhibitor of receptor tyrosine kinase signaling. Development 1999; 126:4139–4147. 31. Casci T, Vinos J, Freeman M. Sprouty, an intracellular inhibitor of Ras signaling. Cell 1999; 96:655–665. 32. Schweitzer R, Shaharabany M, Seger R, Shilo BZ. Secreted Spitz triggers the DER signaling pathway and is a limiting component in embryonic ventral ectoderm determination. Genes Dev 1995; 9:1518–1529. 33. Guichard A, Biehs B, Sturtevant MA, Wickline L, Chacko J, Howard K, Bier E. rhomboid and Star interact synergistically to promote Egfr/Mapk signaling during Drosophila wing vein development. Development 1999; 126:2663–2676. 34. Schweitzer R, Shilo BZ. A thousand and one roles for the Drosophila EGF receptor. Trends Genet 1997; 13:191–196. 35. Vincent S, Ruberte E, Grieder NC, Chen CK, Haerry T, Schuh R, Affolter M. DPP controls tracheal cell migration along the dorsoventral body axis of the Drosophila embryo. Development 1997; 124:2741–2750. 36. Ruberte E, Marty T, Nellen D, Affolter M, Basler K. An absolute requirement for both the type II and type I receptors, punt and thick veins, for dpp signaling in vivo. Cell 1995; 80:889–897. 37. Affolter M, Nellen D, Nussbaumer U, Basler K. Multiple requirements for the receptor serine/threonin kinase thick veins reveal novel functions of TGF beta homologs during Drosophila embryogenesis. Development 1994; 120: 3105–3117. 38. Affolter M, Montagne J, Walldorf U, Groppe J, Kloter U, LaRosa M, Gehring WJ. The Drosophila SRF homolog is expressed in a subset of tracheal cells and maps within a genomic region required for tracheal development. Development 1994; 120:743–753. 39. Tanaka-Matakatsu M, Uemura T, Oda H, Takeichi M, Hayashi S. Cadherin-mediated cell adhesion and cell motility in Drosophila trachea regulated by the transcription factor Escargot. Development 1996; 122:3697–3705. 40. Hayashi S, Hirose S, Metcalfe T, Shirras AD. Control of imaginal cell development by the escargot gene of Drosophila. Development 1993; 118:105–115. 41. Whiteley M, Noguchi PD, Sensabaugh SM, Odenwald WF, Kassis JA. The Drosophila gene escargot encodes a zinc finger motif found in snail-related genes. Mech Dev 1992; 36:117–27. 42. Llimargas M. The Notch pathway helps to pattern the tips of the Drosophila tracheal branches by selecting cell fates. Development 1999; 126:2355–2364. 43. Ikeya T, Hyashi S. Interplay of Notch and FGF signaling restricts cell fate and

126

44.

45. 46.

47.

48.

49.

50.

51.

52. 53.

54. 55. 56.

57. 58. 59.

60.

61.

Wappner and Ratcliffe MAPK activation in the Drosophila trachea. Development 1999; 126:4455– 4463. Bellusci S, Furuta Y, Rush MG, Henderson R, Winnier G, Hogan BL. Involvement of Sonic hedgehog (Shh) in mouse embryonic lung growth and morphogenesis. Development 1997; 124:53–63. Hogan BL. Morphogenesis. Cell 1999; 96:225–233. Motoyama J, Liu J, Mo R, Ding Q, Post M, Hui CC. Essential function of Gli2 and Gli3 in the formation of the lung, trachea and oesophagus. Nat Genet 1998; 20:54–57. Min H, Danilenko DM, Scully SA, et al. FGF-10 is required for both limb and lung development and exhibits striking functional similarity to Drosophila branchless. Genes Dev 1998; 12:3156–3161. Bellusci S, Grindley J, Emoto H, Itoh N, Hogan BL. Fibroblast growth factor 10 (FGF10) and branching morphogenesis in the embryonic mouse lung. Development 1997; 124:4867–4878. Park WY, Miranda B, Lebeche D, Hashimoto G, Cardoso WV. FGF-10 is a chemotactic factor for distal epithelial buds during lung development. Dev Biol 1998; 201:125–134. Peters K, Werner S, Liao X, Wert S, Whitsett J, Williams L. Targeted expression of a dominant negative FGF receptor blocks branching morphogensis and epithelial differentiation of the mouse lung. EMBO J 1994; 13:3296–3301. de Maximy AA, Nakatake Y, Moncada S, Itoh N, Thiery JP, Bellusci S. Cloning and expression pattern of a mouse homologue of Drosophila sprouty in the mouse embryo. Mech Dev 1999; 81:213–216. Placzek M, Skaer H. A paradigm for restricted signalling. Curr Biol 1999; 9:R506– 510. Tefft JD, Lee M, Smith S, Leinwand M, Zhao J, Bringas P Jr, Crowe DL, Warburton D. Conserved function of mSpry-2, a murine homolog of Drosophila sprouty, which negatively modulates respiratory organogenesis. Curr Biol 1999; 9:219–222. Risau W. Mechanisms of angiogenesis. Nature 1997; 386:671–674. Cleaver O, Krieg PA. VEGF mediates angioblast migration during development of the dorsal aorta in Xenopus. Development 1998; 125:3905–3914. Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu XF, Breitman ML, Schuh AC. Failure of blood-island formation and vasculogenesis in Flk-1–deficient mice. Nature 1995; 376:62–66. Beck L Jr, D’amore PA. Vascular development: cellular and molecular regulation. FASEB J 1997; 11:365–373. Wigglesworth VB. The physiology of insect tracheoles. Adv Insect Physiol 1983; 17:86–148. Gale NW, Yancopoulos GD. Growth factors acting via endothelial cell-specific receptor tyrosine kinases: VEGFs, angiopoietins, and ephrins in vascular development. Genes Dev 1999; 13:1055–1066. Ferrara N, Carver-Moore K, Chen H, Dowd M, Lu LO, Shea KS, Powell-Braxton L, Hillan KJ, Moore MW. Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene. Nature 1996; 380:439–442. Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M, Fahrig

Branching Morphogenesis and Hypoxia

62.

63.

64.

65.

66.

67. 68.

69.

70. 71.

72.

73.

74.

75.

127

M, Vandenhoeck A, Harpal K, Eberhardt C, Declercq C, Pawling J, Moons L, Collen D, Risau W, Nagy A. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 1996; 380:435–439. Fong GH, Rossant J, Gertsenstein M, Breitman ML. Role of the Flt-1 receptor tyrosine kinase in regulating the assembly of vasular endothelium. Nature 1995; 376:66–70. Cao Y, Linden P, Farnebo J, Cao R, Ericksson A, Kumar V, Qi JH, Claesson Welsh L, Alitalo K. Vascular endothelial growth factor C induces angiogenesis in vivo. Proc Natl Acad Sci USA 1998; 95:14389–14394. Dumont DJ, Jussila L, Taipale J, Lymboussaki A, Mustonen T, Pajusola K, Breitman M, Alitalo K. Cardiovascular failure in mouse embryos deficient in VEGF receptor-3. Science 1998; 282:946–949. Millauer B, Wizigmann-Voos S, Schnurch H, Martinez R, Moller NP, Risau W, Ullrich A. High affinity VEGF binding and developmental expression suggest Flk-1 as a major regulator of vasculogenesis and angiogenesis. Cell 1993; 72:835– 846. Dumont DJ, Fong GH, Puri MC, Gradwohl G, Alitalo K, Breitman ML. Vascularization of the mouse embryo: a study of flk-1, tek, tie, and vascular endothelial growth factor expression during development. Dev Dyn 1995; 203:80–92. Flamme I, Risau W. Induction of vasculogenesis and hematopoiesis in vitro. Development 1992; 116:435–439. Suri C, McClain J, Thurston G, McDonald DM, Zhou H, Oldmixon EH, Sato TN, Yancopoulos GD. Increased vascularization in mice overexpressing angiopoietin1. Science 1998; 28:468- 471. Holash J, Maisonpierre PC, Compton D, Boland P, Alexander CR, ZagZag D, Yancopoulos GD, Wiegand SJ. Vessel cooption, regression, and growth in tumors mediated by angiopoietins and VEGF. Science 1999; 284:1994–1998. Davis S, Yancopoulos GD. The angiopoietins: Yin and Yang in angiogenesis. Curr Top Microbiol Immunol 1999; 237:173–185. Valenzuela DM, Griffiths JA, Rojas J, Aldrich TH, Jones PF, Zhou H, McClain J, Copeland NG, Gilbert DJ, Jenkins NA, Huang T, Papadopoulos N, Maisonpierre PC, Davis S, Yancopoulos GD. Angiopoietins 3 and 4: diverging gene counterparts in mice and humans. Proc Natl Acad Sci USA 1999; 96:1904–1909. Davis S, Aldrich TH, Jones PF, Acheson A, Compton DL, Jain V, Ryan TE, Bruno J, Radziejewski C, Maisonierre PC, Yancopoulos GD. Isolation of angiopoietin-1, a ligand for the TIE2 receptor, by secretion-trap expression cloning. Cell 1996; 87: 1161–1169. Witzenbichler B, Miasonpierre PC, Jones P, Yancopoulos GD, Isner JM. Chemotactic properties of angiopoietin-1 and -2, ligands for the endothelial-specific receptor tyrosine kinase Tie2. J Biol Chem 1998; 273:18514–18521. Maisonpierre PC, Suri C, Jones PF, Bartunkova S, Wiegand SJ, Radziejewski C, Compton D, McClain J, Aldrich TH, Papadopoulos N, Daly TJ, Davis S, Sato TN, Yancopoulos GD. Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis. Science 1997; 277:55–60. Folkman J, D’Amore PA. Blood vessel formation: what is its molecular basis? Cell 1996; 87:1153–1155.

128

Wappner and Ratcliffe

76. Yancopoulos GD, Klagsbrun M, Folkman J. Vasculogenesis, angiogenesis, and growth factors: ephrins enter the fray at the border. Cell 1998; 93:661–664. 77. Davis S, Gale NW, Aldrich TH, Maisonpierre PC, Lhotak V, Pawson T, Goldfarb M, Yancopoulos GD. Ligands for EPH-related receptor tyrosine kinases that require membrane attachment or clustering for activity. Science 1994; 266:816–819. 78. Wang HU, Chen ZF, Anderson DJ. Molecular distinction and angiogenic interaction between embryonic arteries and veins revealed by ephrin-B2 and its receptor Eph-B4. Cell 1998; 93:741–753. 79. George SE, Simokat K, Hardin J, Chisholm AD. The VAB-1 Eph receptor tyrosine kinase functions in neural and epithelial morphogenesis in C. elegans. Cell 1998; 92:633–643. 80. Scully AL, McKeown M, Thomas JB. Isolation and characterization of Dek, a Drosophila eph receptor protein tyrosine kinase. Mol Cell Neurosci 1999; 13:337–347. 81. Hing H, Xiao J, Harden N, Lim L, Zipursky SL. Pak functions downstream of Dock to regulate photoreceptor axon guidance in Drosophila. Cell 1999; 97:853– 863. 82. Desai CJ, Garrity PA, Keshishian H, Zipursky SL, Zinn K. The Drosophila SH2– SH3 adapter protein Dock is expressed in embryonic axons and facilitates synapse formation by the RP3 motoneuron. Development 1999; 126:1527–1535. 83. Rao Y, Zipursky SL. Domain requirements for the Dock adapter protein in growthcone signaling. Proc Natl Acad Sci USA 1998; 95:2077–2082. 84. Stein E, Huynh-Do U, Lane AA, Cerretti DP, Daniel TO. Nck recruitment to Eph receptor, EphB1/ELK, couples ligand activation to c-Jun kinase. J Biol Chem 1998; 273:1303–1308. 85. Isner JM, Losordo DW. Therapeutic angiogenesis for heart failure. Nat Med 1999; 5:491–492. 86. Pearlman JD, Hibberd MG, Chuang ML, et al. Magnetic resonance mapping demonstrates benefits of VEGF-induced myocardial angiogenesis. Nat Med 1995; 1: 1085–1089. 87. Isner JM, Pieczek A, Schainfeld R, Blair R, Haley L, Asahara T, Rosenfield K, Razvi S, Walsh K, Symes JF. Clinical evidence of angiogenesis after arterial gene transfer of phVEGF165 in patient with ischaemic limb. Lancet 1996; 348:370– 374. 88. Folkman J. Fighting cancer by attacking its blood supply. Sci Am 1996; 275:150– 154. 89. Hanahan D, Folkman J. Patterns and emerging mechanisms of the angiogenic switch during tumorigenesis. Cell 1996; 86:353–364. 90. Parangi S, M OR, Christofori G, Holmgren L, Grosfeld J, Folkman J, Hanahan D. Antiangiogenic therapy of transgenic mice impairs de novo tumor growth. Proc Natl Acad Sci USA 1996; 93:2002–2007. 91. Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ, Lowe SW, Giaccia AJ. Hypoxia-mediated selection of cells with diminished apoptotic potential in solid tumours. Nature 1996; 379:88–91. 92. Klagsbrun M, Knighton D, Folkman J. Tumor angiogenesis activity in cells grown in tissue culture. Cancer Res 1976; 36:110–114. 93. Shweiki D, Itin A, Soffer D, Keshet E. Vascular endothelial growth factor induced

Branching Morphogenesis and Hypoxia

94.

95.

96.

97. 98.

99. 100. 101.

102. 103.

104. 105. 106.

107.

108.

109. 110. 111.

129

by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 1992; 359:843– 845. Shweiki D, Neeman M, Itin A, Keshet E. Induction of vascular endothelial growth factor expression by hypoxia and by glucose deficiency in multicell spheroids: implications for tumor angiogenesis. Proc Natl Acad Sci USA 1995; 92:768–772. Maxwell PH, Pugh CW, Ratcliffe PJ. Inducible operation of the erythropoietin 3′ enhancer in multiple cell lines: evidence for a widespread oxygen-sensing mechanism. Proc Natl Acad Sci USA 1993; 90:2423–2427. Wang GL, Semenza GL. General involvement of hypoxia-inducible factor 1 in transcriptional response to hypoxia. Proc Natl Acad Sci USA 1993; 90:4034– 4308. Wang GL, Semenza GL. Characterization of hypoxia-inducible factor 1 and regulation of DNA binding activity by hypoxia. J. Biol Chem 1993; 268:21513–21518. Wang GL, Jiang BH, Rue EA, Semenza GL. Hypoxia-inducible factor 1 is a basichelix-loop-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci USA 1995; 92:5510–5514. Wang GL. Semenza GL. Purification and characterization of hypoxia-inducible factor 1. J Biol Chem 1995; 270:1230–1237. Bunn HF, Poyton RO. Oxygen sensing and molecular adaptation to hypoxia. Physiol Rev 1996; 76:839–885. Huang LE, Arany Z, Livingston DM, Bunn HF. Activation of hypoxia-inducible transcription factor depends upon redox-sensitive stabilization of its alpha subunit. J Biol Chem 1996; 271:32253–32259. Gleadle JM, Ratcliffe PJ. Hypoxia and the regulation of gene expression. Mol Med Today 1998; 4:122–129. Wood SM, Gleadle JM, Pugh CW, Hankinson O, Ratcliffe PJ. The role of the aryl hydrocarbon receptor nuclear translocator (ARNT) in hypoxic induction of gene expression. Studies in ARNT-deficient cells. J Biol Chem 1996; 271:15117–15123. Huang ZJ, Edery I, Rosbash M. PAS is a dimerization domain common to Drosophila period and several transcription factors. Nature 1993; 364:259–262. Crews ST. Control of cell lineage–specific development and transcription by bHLH-PAS proteins. Genes Dev 1998; 12:607–620. Wharton KA Jr, Franks RG, Kasai Y, Crews ST. Control of CNS midline transcription by asymmetric E-box-like elements: similarity to xenobiotic responsive regulation. Development 1994; 120:3563–3569. Franks RG, Crews ST. Transcriptional activation domains of the Single-minded bHLH protein are required for CNS midline cell development. Mech Dev 1994; 45:269–277. Crews S, Franks R, Hu S, Matthews B, Nambu J. Drosophila single-minded gene and the molecular genetics of CNS midline development. J Exp Zool 1992; 261: 234–244. Ponting CP, Aravind L. PAS: a multifunctional domain family come to light. Curr Biol 1997; 7:R674–677. Taylor BL, Zhulin IB. PAS domains: internal sensors of oxygen, redox potential, and light. Microbiol Mol Biol Rev 1999; 63:479–506. Monson EK, Ditta GS, Helsinki DR. The oxygen sensor protein FixL, of Rhizobium

130

112.

113. 114.

115.

116.

117.

118.

119. 120.

121.

122.

123.

124.

125.

Wappner and Ratcliffe meliloti. Role of histidine residues in heme binding, phosphorylation, and signal transduction. J. Biol Chem 1995; 270:5243–5250. Monson EK, Weinstein M, Ditta GS, Helsinski DR. The FixL protein of Rhizobium meliloti can be separated into a heme-binding oxygen sensing domain and a functional C-terminal kinase domain. Proc Natl Acad Sci USA 1992; 89:4280–4284. Semenza GL. Hypoxia-inducible factor 1: master regulator of O2 homeotasis. Curr Opin Genet Dev 1998; 8:588–594. Blanchard KL, Acquaviva AM, Galson DL, Bunn HF. Hypoxic induction of the human erythropoietin gene: cooperation between the promoter and enhancer, each of which contains steroid receptor response elements. Mol Cell Biol 1992; 12: 5373–5385. Ho VT, Bunn HF. Effects of transition metals on the expression of the erythropoietin gene: further evidence that the oxygen sensor is a heme protein. Biochem Biophys Res Commun 1996; 223:175–180. Rolfs A, Kvietikova I, Gassmann M, Wenger RH. Oxygen-regulated transferrin expression is mediated by hypoxia-inducible factor-1. J Biol Chem 1997; 272: 20055–20062. Melillo G, Taylor LS, Brooks A, Musso T, Cox GW, Varesio L. Functional requirement of the hypoxia-reponsive element in the activation of the inducible nitric oxide synthase promoter by the iron chelator desferrioxamine. J Biol Chem 1997; 272: 12236–12243. Czyzyk-Krzeska MF, Furnari BA, Lawson EE, Millhorn DE. Hypoxia increases rate of transcription and stability of tyrosine hydroxylase mRNA in pheochromocytoma (PC12) cells. J Biol Chem 1994; 269:760–764. Norris ML, Millhorn DE. Hypoxia-induced protein binding to O2-responsive sequences on the tyrosine hydroxylase gene. J Biol Chem 1995; 270:23774–23779. Firth JD, Ebert BL, Pugh CW, Ratcliffe PJ. Oxygen-regulated control elements in the phosphoglycerate kinase 1 and lactate dehydrogenase A genes: similarities with the erythropoietin 3′ enhancer. Proc Natl Acad Sci USA 1994; 91:6496–6500. Firth JD, Ebert BL. Ratcliffe PJ. Hypoxic regulation of lactate dehydrogenase A. Interaction between hypoxia-inducible factor 1 and cAMP response elements. J Biol Chem 1995; 270:21021–21027. Ebert BL, Gleadle JM, JF OR, Bartlett SM, Poulton J, Ratcliffe PJ. Isoenzymespecific regulation of genes involved in energy metabolism by hypoxia: similarities with the regulation of erythropoietin. Biochem J 1996; 313:809–814. Semenza GL, Roth PH, Fang HM, Wang GL. Transcriptional regulation of genes encoding glycolytic enzymes by hypoxia-inducible factor 1. J Biol Chem 1994; 269:23757–23763. Semenza GL, Jiang BH, Leung SW, Passantino R, Concordet JP, Marie P, Giallongo A. Hypoxia response elements in the aldolase A, enolase 1, and lactate dehydrogenase A gene promoters contain essential binding sites for hypoxia-inducible factor 1. J Biol Chem 1996; 271:32529–32537. Carmeliet P, Dor Y, Herbert JM, Fukumura D, Brusselmans K, Dewerchin M, Neeman M, Bono F, Abramovitch R, Maxwell P, Koch CJ, Ratcliffe P, Moons L, Jain RK, Collen D, Keshet E. Role of HIF-1 alpha in hypoxia-mediated apoptosis, cell proliferation and tumor angiogenesis. Nature 1998; 394:485–490.

Branching Morphogenesis and Hypoxia

131

126. An WG, Kanekal M, Simon MC, Maltepe E, Blagosklonny MV, Neckers LM. Stabilization of wild-type p53 hypoxia-inducible factor 1 alpha. Nature 1998; 392: 405–408. 127. Ebert BL, Bunn HF. Regulation of transcription by hypoxia requires a multiprotein complex that includes hypoxia-inducible factor 1, an adjacent transcription factor, and p300/CREB binding protein. Mol Cell Biol 1998; 18:4089–4096. 128. Galson DL, Tsuchiya T, Tendler DS, Huang LE, Ren Y, Ogura T, Bunn HF. The orphan receptor hepatic nuclear factor 4 functions as a transcriptional activator for tissue-specific and hypoxia-specific erythropoietin gene expression and is antagonized by EAR3/COUP-TF1. Mol Cell Biol 1995; 15:2135–2144. 129. Nagao M, Ebert BL, Ratcliffe PJ, Pugh CW. Drosophila melanogaster SL2 cells contain a hypoxically inducible DNA binding complex which recognises mammalian HIF-binding sites. FEBS Lett 1996; 387:161–166. 130. Jarecki J, Johnson E, Krasnow MA. Oxygen regulation of airway branching in Drosophila is mediated by Branchless FGF. Cell 1999; 99:211–220. 131. Johnson E, Jarecki J, Krasnow M. Oxygen-dependent patterning of tracheal branching is mediated by the Branchless FGF. 40th Annual Drosophila Research Conference 1999, Bellevue, Wasington, p 139. 132. Wappner P, Zelzer E, Bacon N, Ratcliffe PJ, Shilo B-Z. Sima may be the hypoxic regulator in Drosophila. 39th Annual Drosophila Research Conference 1998, Washington, DC, p 381A. 133. Huang LE, Gu J, Schau M, Bunn HF. Regulation of hypoxia-inducible factor 1 alpha is mediated by an O2-dependent degradation domain via the ubiquitin-proteasome pathway. Proc Natl Acad Sci USA 1998; 95:7987–7992. 134. Pugh CW, JF OR, Nagao M, Gleadle JM, Ratcliffe PJ. Activation of hypoxiainducible factor-1; definition of regulatory domains within the alpha subunit. J Biol Chem 1997; 272:11205–11214. 135. Jiang BH, Zheng JZ, Leung SW, Roe R, Semenza GL. Transactivation and inhibitory domains of hypoxia-inducible factor 1 alpha. Modulation of transcriptional activity by oxygen tension. J Biol Chem 1997; 272:19253–19260. 136. Ema M, Hirota K, Mimura J, Abe H, Yodoi J, Sogawa K, Poellinger L, FujiiKuriyama Y. Molecular mechanisms of transcription activation by HLF and HIF1 alpha in response to hypoxia: their stabilization and redox signal-induced interaction with CBP/p300. EMBO J 1999; 18:1905–1914. 137. O’Rourke JF, Tian YM, Ratcliffe PJ, Pugh CW. Oxygen-regulated and transactivating domains in endothelial PAS protein 1: comparison with hypoxia-inducible factor-1 alpha. J Biol Chem 1999; 274:2060–2071. 138. Nambu JR, Chen W, Hu S, Crews ST. The Drosophila melanogaster similar bHLHPAS gene encodes a protein related to human hypoxia-inducible factor 1 alpha and Drosophila single-minded. Gene 1996; 172:249–254. 139. Bacon NC, Wappner P, JF OR, Bartlett SM, Shilo B, Pugh CW, Ratcliffe PJ. Regulation of the Drosophila bHLH-PAS protein Sima by hypoxia: functional evidence for homology with mammalian HIF-1 alpha. Biochem Biophys Res Commun 1998; 249:811–816. 140. Salceda S, Caro J. Hypoxia-inducible factor 1 alpha (HIF-1 alpha) protein is rapidly degraded by the ubiquitin-proteasome system under normoxic conditions. Its stabil-

132

141.

142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154.

155.

156.

157. 158.

159. 160.

Wappner and Ratcliffe ization by hypoxia depends on redox-induced changes. J Biol Chem 1997; 272: 22642–22647. Kallio PJ, Wilson WJ, S OB, Makino Y, Poellinger L. Regulation of the hypoxiainducible transcription factor 1 alpha by the ubiquitin-proteasome pathway. J Biol Chem 1999; 274:6519–6525. Hershko A, Ciechanover A. The ubiquitin system. Annu Rev Biochem 1998; 67: 425–679. Ciechanover A. The ubiquitin-proteasome pathway: on protein death and cell life. EMBO J 1998; 17:7151–7160. Pahl HL, Baeuerle PA. Control of gene expression by proteolysis. Curr Opin Cell Biol 1996; 8:340–347. Jackson PK. Cell cycle: cull and destroy. Curr Biol 1996; 6:1209–1212. Hodges M, Tissot C, Freemont PS. Protein regulation: tag wrestling with relatives of ubiquitin. Curr Biol 1998; 8:R749–752. Peters JM. SCF and APC: the Yin and Yang of cell cycle regulated proteolysis. Curr Opin Cell Biol 1998; 10:759–768. Elledge SJ, Harper JW. The role of protein stability in the cell cycle and cancer. Biochim Biophys Acta 1998; 1377:M61–70. Townsley FM, Ruderman JV. Proteolytic ratchets that control progression through mitosis. Trends Cell Biol 1998; 8:238–244. Hochstrasser M. There’s the rub: a novel ubiquitin-like modification linked to cell cycle regulation. Genes Dev 1998; 12:901–907. Pagano M. Cell cycle regulation by the ubiquitin pathway. FASEB J 1997; 11: 1067–1075. Hershko A. Roles of ubiquitin-mediated proteolysis in cell cycle control. Curr Opin Cell Biol 1997; 9:788–799. Zhang H, Kobayashi R, Galaktionov K, Beach D. p19skp1 and p45skp2 are essential elements of the cyclin a-Cdk2 S phase kinase. Cell 1995; 82:915–925. Feldman RM, Correll CC, Kaplan KB, Deshaies RJ, A complex of Cdc4p, Skp1p, and Cdc53p/cullin catalyzes ubiquitination of the phosphorylated CDK inhibitor Sic1p. Cell 1997; 91:221–230. Li FN, Johnston M. Grr1 of Saccharomyces cerevisiae is connected to the ubiquitin proteolysis machinery through Skp1: coupling glucose sensing to gene expression and the cell cycle. EMBO J 1997; 16:5629–5638. Methot N, Basler K. Hedgehog controls limb development by regulating the activities of distinct transcriptional activator and repressor forms of Cubitus interruptus. Cell 1999; 96:819–831. Ohlmeyer JT, Kalderon D. Hedgehog stimulates maturation of Cubitus interruptus into a labile transcriptional activator. Nature 1998; 396:749–753 Chen Y, Gallaher N, Goodman RH, Smolik SM. Protein kinase A directly regulates the activity and proteolysis of Cubitus interruptus. Proc Natl Acad Sci USA 1998; 95:2349–2354. Kalderon D. Hedgehog signalling: Ci complex cuts and clasps. Curr Biol 1997; 7: R759–762. Jiang J, Struhl G. Regulation of the Hedgehog and Wingless signalling pathways by the F-box/WD40–repeat protein Slimb. Nature 1998; 391:493–496.

Branching Morphogenesis and Hypoxia

133

161. Aza-Blanc P, Ramirez-Weber FA, Laget MP, Schwartz C, Kornberg TB. Proteolysis that is inhibited by Hedgehog targets Cubitus interruptus protein to the nucleus and converts it to a repressor. Cell 1997; 89:1043–1053. 162. Kitagawa M, Hatakeyama S, Shirane M, Matsumoto M, Ishida N, Hattori K, Nakamichi I, Kikuchi A, Nakayama K. An F-box protein, FWD1, mediates ubiquitindependent proteolysis of beta-catenin. EMBO J 1999; 18:2401–2410. 163. Latres E, Chiaur DS, Pagano M. The human F box protein beta-Trcp associates with the Cul1/Skp1 complex and regulates the stability of beta-catenin. Oncogene 1999; 18:849–854. 164. Liu C, Kato Y, Zhang Z, Do VM, Yankner BA, He X. beta-Trcp couples betacatenin phosphorylation-degradation and regulates Xenopus axis formation. Proc Natl Acad Sci USA 1999; 96:6273–6278. 165. Winston JT, Strack P, Beer-Romero P, Chu CY, Elledge SJ, Harper JW. The SCFbeta–TRCP–ubiquitin ligase complex associates specifically with phosphorylated destruction motifs in IkappaBalpha and beta-catenin and stimulates Ik-Bα ubiquitination in vitro. Genes Dev 1999; 13:270–283. 166. Kinzler KW, Vogelstein B. Lessons from hereditary colorectal cancer. Cell 1996; 87:159–170. 167. Easwaran V, Song V, Polakis P, Byers S. The ubiquitin-proteasome pathway and serine kinase activity modulate adenomatous polyposis coli protein-mediated regulation of beta-catenin–lymphocyte enhancer–binding factor signaling. J Biol Chem 1999; 274:16641–16645. 168. Polakis P. The adenomatous polyposis coli (APC) tumor suppressor. Biochim Biophys Acta 1997; 1332:F127–147. 169. Polakis P. The oncogenic activation of beta-catenin. Curr Opin Genet Dev 1999; 9:15–21. 170. Tetsu O, McCormick F. Beta-catenin regulates expression of cyclin D1 in colon carcinoma cells. Nature 1999; 398:422–426. 171. Salghetti SE, Kim SY, Tansey WP. Destruction of Myc by ubiquitin-mediated proteolysis: cancer-associated and transforming mutations stabilize Myc. EMBO J 1999; 18: 717–726. 172. Hofmann RM, Pickart CM. Noncanonical MMS2-encoded ubiquitin-conjugating enzyme functions in assembly of novel polyubiquitin chains for DNA repair. Cell 1999; 96:645–653. 173. Sears R, Leone G. DeGregori J, Nevins JR. Ras enhances Myc protein stability. Mol Cell 1999; 3:169–179. 174. Gross-Mesilaty S, Reinstein E, Bercovich B, Tobias KE, Schwartz AL, Kahana C, Ciechanover A. Basal and human papillomavirus E6 oncoprotein-induced degradation of Myc proteins by the ubiquitin pathway. Proc Natl Acad Sci USA 1998; 95: 8058–8063. 175. Hershko A, Ciechanover A. The ubiquitin system for protein degradation. Annu Rev Biochem 1992; 61:761–807. 176. Laney JD, Hochstrasser M. Substrate targeting in the ubiquitin system. Cell 1999; 97:427–430. 177. Scheffner M, Nuber U, Huibregtse JM. Protein ubiquitination involving an E1-E2E3 enzyme ubiquitin thioester cascade. Nature 1995; 373:81–83.

134

Wappner and Ratcliffe

178. Krek W. Proteolysis and the G1-S transition: the SCF connection. Curr Opin Genet Dev 1998; 8:36–42. 179. Patton EE, Willems AR, Tyers M. Combinatorial control in ubiquitin-dependent proteolysis: don’t Skp the F-box hypothesis. Trends Genet 1998; 14:236– 243. 180. Huibregtse JM, Scheffner M, Beaudenon S, Howley PM. A family of proteins structurally and functionally related to the E6-AP ubiquitin-protein ligase. Proc Natl Acad Sci USA 1995; 92:2563–2567. 181. Scheffner M, Huibregtse JM, Vierstra RD, Howley PM. The HPV-16 E6 and E6AP complex functions as a ubiquitin-protein ligase in the ubiquitination of p53. Cell 1993; 75:495–505. 182. Zachariae Z, Nasmyth F. Whose end is destruction: cell division and the anaphasepromoting complex. Genes Dev 1999; 13:2039–2058. 183. Lyapina SA, Correll CC, Kipreos ET, Deshaies RJ. Human CUL1 forms an evolutionary conserved ubiquitin ligase complex (SCF) with SKP1 and an F-box protein. Proc Natl Acad Sci USA 1998; 95:7451–7456. 184. Kipreos ET, Lander LE, Wing JP, He WW, Hedgecock EM. Cul-1 is required for cell cycle exit in C. elegans and identifies a novel gene family. Cell 1996; 85:829– 839. 185. Seol JH, Feldman RM, Zachariae W, Shevchenko A, Correll CC, Lyapina S, Chi Y, Galova M, Claypool J, Sandmeyer S, Nasmyth K, Deshaies RJ, Shevchenko A, Deshaies RJ. Cdc53/cullin and the essential Hrt1 RING-H2 subunit of SCF define a ubiquitin ligase module that activates the E2 enzyme Cdc34. Genes Dev 1999; 13:1614–1626. 186. Tan P, Fuchs SY, Chen A, Wu K, Gomez C, Ronai Z, Pan ZQ. Recuitment of a ROC1-CUL1 ubiquitin ligase by Skp1 and HOS to catalyze the ubiquitination of Ik-Bα. Mol Cell 1999; 3:527–533. 187. Skowyra D, Koepp DM, Kamura T, Conrad MN, Conaway RC, Conaway JW, Elledge SJ, Harper JW. Reconstitution of G1 cyclin ubiquitination with complexes containing SCFGrr1 and Rbx1. Science 1999; 284:662–665. 188. Ohta T, Michel JJ, Schottelius AJ, Xiong Y. ROC1, a homolog of APC11, represents a family of cullin partners with an associated ubiquitin ligase activity. Mol Cell 1999; 3:535–541. 189. Kamura T, Koepp DM, Conrad MN, Skowyra D, Moreland RJ, Iliopoulos O, Lane WS, Kaelin WG Jr, Elledge SJ, Conaway RC, Harper JW, Conaway JW. Rbx1, a component of the VHL tumor suppressor complex and SCF ubiquitin ligase. Science 1999; 284:657–661. 190. Wolf DA, McKeon F, Jackson PK. F-box/WD–repeat proteins pop1p and Sud1p/ Pop2p form complexes that bind and direct the proteolysis of cdc18p. Curr Biol 1999; 9:373–376. 191. Margottin F, Bour SP, Durand H. Selig L, Benichou S, Richard V, Thomas D, Strebel K, Benarous RA. Novel human WD protein, h-beta TrCp, that interacts with HIV-1 Vpu connects CD4 to the ER degradation pathway through an F-box motif. Mol Cell 1998; 1:565–574. 192. Skowyra D, Craig KL, Tyers M, Elledge SJ, Harper JW. F-box proteins are recep-

Branching Morphogenesis and Hypoxia

193.

194. 195. 196.

197.

198.

199.

200.

201. 202.

203.

204. 205. 206.

207.

208.

135

tors that recruit phosphorylated substrates to the SCF ubiquitin-ligase complex. Cell 1997; 91:209–219. Maxwell PH, Wiesener MS, Chang GW, Clifford SC, Vaux EC, Cockman ME, Wykoff CC, Pugh CW, Maher ER, Ratcliffe PJ. The tumour suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 1999; 399:271–275. Kaelin WG Jr, Maher ER. The VHL tumour-suppressor gene paradigm. Trends Genet 1998; 14:423–426. Iliopoulos O, Kaelin WG, Jr. The molecular basis of von Hippel-Lindau disease. Mol Med 1997; 3:289–293. Pause A, Lee S, Lonergan KM, Klausner RD. The von Hippel-Lindau tumor suppressor gene is required for cell cycle exit upon serum withdrawal. Proc Natl Acad Sci USA 1998; 95:993–998. Iliopoulos O, Levy AP, Jiang C, Kaelin WG Jr, Goldberg MA. Negative regulation of hpoxia-inducible genes by the von Hippel-Lindau protein. Proc Natl Acad Sci USA 1996; 93:10595–10599. Lisztwan J, Imbert G, Wirbelauer C, Gstaiger M, Krek W. The von Hippel-Lindau tumor suppressor protein is a component of an E3 ubiquitin-protein ligase activity. Genes Dev 1999; 13:1822–1833. Liakopoulos D, Busgen T, Brychzy A, Jentsch S, Pause A. Conjugation of the ubiquitin-like protein NEDD8 to cullin-2 is linked to von Hippel-Lindau tumor suppressor function. Proc Natl Acad Sci USA 1999; 96:5510–5515. Stebbins CE, Kaelin WG Jr, Pavletich NP. Structure of the VHL-ElonginC-ElonginB complex: implications for VHL tumor suppressor function. Science 1999; 284:455–461. Haas AL, Siepmann TJ. Pathways of ubiquitin conjugation. FASEB J 1997; 11: 1257–1268. Theodosiou NA, Zhang S, Wang WY, Xu T. Slimb Coordinates Wg and Dpp expression in the dorsal-ventral and anterior-posterior axes during limb development. Development 1998; 125:3411–3416. Dong X, Tsuda L, Zavitz KH, Lin M, Li S, Carthew RW, Zipursky S. L ebi Regulates epidermal growth factor receptor signaling pathways in Drosophila. Genes Dev 1999; 13:954–965. Spencer E, Jiang J, Chen ZJ. Signal-induced ubiquitination of IkappaBalpha by the F-box protein Slimb/beta-TrCP. Genes Dev 1999; 13:284–294. Arany Z, Huang LE, Eckner R, et al. An essential role for p300/CBP in the cellular response to hypoxia. Proc Natl Acad Sci USA 1996; 93:12969–12973. Kallio PJ, Okamoto K, O’Brien S, Carrero P, Makino Y, Tanaka H, Poellinger L. Signal transduction in hypotic cells: inducible nuclear translocation and recruitment of the CBP/p300 coactivator by the hypoxia-inducible factor-1 alpha. EMBO J 1998; 17:6573–6586. Srinivas V, Zhang LP, Zhu XH, Caro J. Characterization of an oxygen/redox– dependent degradation domain of hypoxia-inducible factor alpha (HIF-alpha) proteins. Biochem Biophys Res Commun 1999; 260:557–561. Wang GL, Jiang BH, Semenza GL. Effect of altered redox states on expression

136

209. 210. 211. 212.

213.

214.

215. 216.

217.

218.

219. 220. 221. 222. 223.

224.

225.

Wappner and Ratcliffe and DNA-binding activity of hypoxia-inducible factor 1. Biochem Biophys Res Commun 1995; 212:550–556. Semenza GL. Perspectives on oxygen sensing Cell 1999; 98:281–284. Fischer HM. Genetic regulation of nitrogen fixation in rhizobia. Microbiol Rev 1994; 58:352–386. Gilles-Gonzalez MA, Ditta GS, Helinski DR. A haemoprotein with kinase activity encoded by the oxygen sensor of Rhizobium meliloti. Nature 1991; 350:170–172. Lois AF, Weinstein M, Ditta GS, Helinski DR. Autophosphorylation and phosphatase activities of the oxygen-sensing protein FixL of Rhizobium meliloti are coordinately regulated by oxygen. J Biol Chem 1993; 268:4370–4375. Lois AF, Ditta GS, Helinski DR. The oxygen sensor FixL of Rhizobium meliloti is a membrane protein containing four possible transmembrane segments. J Bacteriol 1993; 175:1103–1109. Gilles-Gonzalez MA, Gonzalez G, Perutz, MF, Kiger L, Marden MC, Poyart CC. Heme-based sensors, exemplified by the kinase FixL, are a new class of heme protein with distinctive ligand binding and autoxidation. Biochemistry 1994; 33:8067– 8073. Goldberg MA, Dunning SP, Bunn HF. Regulation of the erythropoietin gene: evidence that the oxygen sensor is a heme protein. Science 1988: 242:1412–1415. Sinclair P, Gibbs AH, Sinclair JF, de Matteis F. Formation of cobalt protoporphyrin in the liver of rats. A mechanism for the liver haem biosynthesis by inorganic cobalt. Biochem J 1979; 178:529–538. Kietzmann T, Schmidt H, Unthan-Fechner K, Probst I, Jungermann K. A ferroheme protein senses oxygen levels, which modulate the glucagon-dependent activation of the phosphoenolpyruvate carboxykinase gene in rat hepatocyte cultures. Biochem Biophys Res Commun 1993; 195:792–798. Srinivas V, Zhu X, Salceda S, Nakamura R, Caro J. Hypoxia-inducible factor 1 alpha (HIF-1 alpha) is a non-heme iron protein. Implications for oxygen sensing J Biol Chem 1999; 273:18019–18022. Fandrey J, Frede S, Jelkmann W. Role of hydrogen peroxide in hypoxia-induced erythropoietin production. Biochem J 1994; 303:507–510. Kroll SL C-KM. Role of H2O2 and heme-containing O2 sensors in hypoxic regulation of tyrosine hydroxylase gene expression. Am J Physiol 1998; 274:C167–174. Acker H. Cellular oxygen sensors. Ann NY Acad Sci 1994; 718:3–10. Fenton HJH. On a new reaction of tartaric acid. Chem News 1876; 33:190. Wenger RH, Marti HH, Schuerer-Maly CC, Kvietikova I, Bauer C, Gassmann M, Maly FE. Hypoxic induction of gene expression in chronic granulomatous diseasederived B-cell lines: oxygen sensing is independent of the cytochrome b558-containing nicotinamide adenine dinucleotide phosphate oxidase. Blood 1996; 87:756– 761. Suh YA, Arnold RS, Lassegue B, Shi J, Xu X, Sorescu D, Chung AB, Griendling KK, Lambeth JD. Cell transformation by the superoxide-generating oxidase Mox1. Nature 1999; 401:79–82. Chandel NS, Maltepe E, Goldwasser E, Mathieu CE, Simon MC, Schumacker PT. Mitochondrial reactive oxygen species trigger hypoxia-induced transcription. Proc Natl Acad Sci USA 1998; 95:11715–11720.

Branching Morphogenesis and Hypoxia

137

226. Kwast KE, Burke PV, Staahl BT, Poyton RO. Oxygen sensing in yeast: evidence for the involvement of the respiratory chain in regulating the transcription of a subset of hypoxic genes. Proc Natl Acad Sci USA 1999; 96:5446–5451. 227. Wingrove JA, O’Farrell PH. Nitric oxide contributes to behavioral, cellular, and developmental responses to low oxygen in Drosophila. Cell 1999; 98:105–114. 228. Stone JR, Marletta MA. Soluble guanylate cyclase from bovine lung: activation with nitric oxide and carbon monoxide and spectral characterization of the ferrous and ferric states. Biochemistry 1994; 33:5636–5640. 229. Vaandrager AB, de Jonge HR. Signalling by cGMP-dependent protein kinases. Mol Cell Biochem 1996; 157:23–30. 230. Hobbs AJ, Ignarro LJ. Nitric oxide-cyclic GMP signal transduction system. Methods Enzymol 1996; 269:134–148. 231. Stone JR, Marletta MA. Heme stoichiometry of heterodimeric soluble guanylate cyclase. Biochemistry 1995; 34:14668–14674. 232. Foerster J, Harteneck C, Malkewitz J, Schultz G, Koesling D. A functional hemebinding site of soluble guanylyl cyclase requires intact N-termini of alpha 1 and beta 1 subunits. Eur J Biochem 1996; 240:380–386. 233. Dierks EA, Burstyn JN. The deactivation of soluble guanylyl cyclase by redoxactive agents. Arch Biochem Biophys 1998; 351:1–7. 234. Schrammel A, Behrends S, Schmidt K, Koesling D, Mayer B. Characterization of 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one as a heme-site inhibitor of nitric oxide-sensitive guanylyl cyclase. Mol Pharmacol 1996; 50:1–5.

6 Study of Anoxia Tolerance Use of a Novel Genetic Approach and Animal Model

GABRIEL G. HADDAD Yale University School of Medicine New Haven, Connecticut

I.

Introduction

Tissue O2 deprivation is a clinically frequent event, but the cellular and molecular cascades that underlie the responses of cells and tissues to lack of oxygenation are not well defined or characterized. The lack of a well-developed definition is not due to the lack of studies, since there has been literally thousands of studies in the past few decades trying to understand the responses that occur and their molecular basis. Although we have, as a scientific community interested in these hypoxic responses, advanced knowledge over the past 10–20 years in finding answers related to the molecular definition of hypoxia in various tissues, animal species, and phyla, and at various ages [1], there are still major unanswered questions. These questions are not only about specific scientific issues but also about approaches used to develop a comprehensive understanding of what hypoxia is and how tissues respond to it. This chapter focuses on the heterogeneity of tissue responses to lack of O2. Examples some of the advances that we have made in the past decade regarding this heterogeneity in tissue and cell responses will be illustrated. Some of the genetic techniques and the use of invertebrate models that we have recently em139

140

Haddad

barked on will be discussed in order to address such questions. Heterogeneity of tissue responses and resistance and susceptibility to lack of O2 are major issues that have still not been solved. If appreciated and well understood, such issues can potentially lead to clinically important therapeutic interventions that could lessen human morbidity and mortality from hypoxic injury across the age spectrum whether in the form of stroke and myocardial infarction in the adult or intraventricular hemorrhage and perinatal asphyxic insult in early life. II. Heterogeneity in Tolerance to Lack of O2 Between Species and Phyla One of the very interesting findings in the past few decades is the wide heterogeneity in the response (or lack thereof) of cells and tissues to O2 deprivation [1]. The organism par excellence that has been studied enormously and proven to be exceedingly tolerant to lack of O2 is the turtle [1–6]. The turtle can survive weeks and months of severe hypoxic environments, whereas, in comparison, a rat brain would not survive beyond 10 min [1]. There is clearly a spectrum of responses, and a number of organisms would probably fit in between the rat and the turtle. For example, the newly born mammal is also resistant to hypoxia but not nearly as much as the turtle [5–8]. In order to understand how turtles, such as the freshwater turtle Pseudemys scripta, survive severe hypoxic stress, many groups of investigators have studied various aspects of turtle physiology. They have attempted to learn about, for example, turtle metabolism, neurotransmitter release, neural activity patterns, ionic homeostasis, and a huge body of literature has been written on the turtle. Although this chapter is not intended to review comprehensively the turtle literature, nevertheless some of the relevant research that has been done in the turtle will be summarized first. The hope is to be able to derive some crucial and important questions that we, as a scientific community, have not been able to address using these types of approaches. A. Electrophysiological Aspects

The electroencephalographic activity in the turtle brain is preserved for quite some time before it decreases during anoxia [2,9]. In the rat brain, in contrast, brain activity reaches an isoelectric line in a few minutes depending on the insult imposed [1,2,9]. In anoxia, brain activity reaches zero in the rat and in mammals in general in about 45–90 s. Irreversible injury also occurs within a few to several minutes. In the turtle, the root mean square voltage reaches 20% of control after 100 min of anoxia [2,9]. Of course, this decrease in activity is totally reversible in the turtle even when this animal sustains hypoxia/anoxia for weeks. Of interest is that, in spite of the fact that K⫹ ions leak to the outside of cells in turtle,

Anoxia Tolerance

141

albeit at a much smaller rate compared to the efflux in the rat, overall cellular conductance is reduced in turtle neurons [9,10]. This contrasts markedly with rat neurons in which conductance increases enormously, and this is mostly based on the opening of K⫹ channels [10–15]. This decrease in conductance in turtle neurons has earned the label of ‘‘channel arrest’’ [3,9,10]. Some investigators have used this idea of channel arrest, often without much data, to suggest that it is the lack of imbalance in the ionic homeostasis that allows the turtle to survive long periods in anoxia. And that O2-sensing and the signal-transduction system that is then activated (or inhibited) is responsible for the coordination of the various components of metabolism and ionic homeostatic mechanisms in the turtle. Clearly, it is not known how ionic homeostasis is preserved in the turtle. B. Neurotransmitters and Ionic Mechanisms

This research aspect on the turtle brain has also been an important contribution over the past several years. It seems, here again, that there is a potential number of transmitters that are related in an important way to hypoxic tolerance [16– 23] and that they also interact with each other during anoxia [18–20]. For example, whereas adenosine receptors lose their affinity and decrease in expression after prolonged hypoxia in the rat, these same receptors increase their affinity and keep their expression stable in the turtle central nervous system [17,18,23]. Furthermore, there is evidence now that the lack of decrease in intracellular calcium during anoxia in the turtle may be in part explained by the effect of adenosine on the N-methyl-D-aspartate (NMDA glutamate) receptors [17,18]. These receptors, which form a moeity with an ion channel that mediate calcium influx into cells, have an open probability that gets downregulated by adenosine during anoxia [17,18]. Of interest also is the observation that Na⫹ channels decrease in their expression on the plasma membrane after 4 hr of anoxia in the turtle [24]. Moreover, the Na/K ATPase activity drops by about one third in turtle neurons; also indicating energy saving as a potential means for survival [25]. K⫹ channels, such as ATP-sensitive K⫹ channels, do not seem to play a major role in the turtle [16,26]. C. Metabolic Considerations

It is clear that the turtle downregulates its O2 demands during anoxia [1,3,10]. Anoxia still results in a drop in phosphocreatine (about 40% in 3 hr) and 20% in ATP levels after about 2 hr [27]. Further, because of increased anaerobic glycolytic activity to maintain relatively well its ATP content, pH drops substantially from 7.2 to about 6.6 [27]. These changes are also completely reversible with reoxygenation [27]. Clearly, neither the stimulus nor the signals that mediate the downregulation of metabolic demands and for tolerating low pHs remains elusive.

142

Haddad D. Oxidative Stress During Reoxygenation

That the turtle does not seem to suffer tissue injury at the time of reoxygenation, after prolonged anoxia, has been fascinating and has enticed investigators to determine the levels of their antoxidant systems. From these studies, it seems that they have high levels of antoxidants such as catalase and superoxide dismutase [28–30] and that lipid peroxidation does not happen in the transition between anoxia and reoxygenation. Furthermore, it seems that turtles can regulate their glutathione-related enzymes in order to maintain high reduced/oxidized glutathione ratios [28–30]. How this is achieved is not clear at present. III. Major Unresolved Questions from Studies on AnoxiaTolerant Organisms Although a number of investigators have made substantial contributions in our understanding of anoxic tolerance using the turtle model over the past 1–2 decades, there are still a number of questions that have begged, the physiologist and biologist alike, to probe further. These have really stemmed from the advances already made using mostly physiological and biochemical techniques. Some of these questions are: 1. Is there a unifying ‘‘principle’’ which acts as a master switch in cells of anoxia-tolerant animals to regulate and coordinate the various activities in a hierarchical manner? If this is the case, what would an experimental approach be in order to uncover such a principle? 2. What is the signal or signal(s) for metabolism to be downregulation? Although this downregulation is important in survival, we have no clear idea as to what induces this downregulation. 3. What are the sensing pathways of a lack of O2 in these tolerant organisms? What is the molecular and genetic network that is involved in the pathway leading to survival? 4. What is the basis for ‘‘channel arrest’’ or for the decrease in cellular conductance? What are the factors that determine this arrest and what controls these factors? 5. What is the short-term versus the longer adaptive strategy in the tolerant organisms? What are the respective sensing mechanisms? What are their molecular and genetic bases? 6. What is the molecular and genetic basis for the species-dependent survival strategies? What is the basis for the difference in strategy between that of the turtle versus that of the crucian carp survival plan? Are there different survival plans in various species?

Anoxia Tolerance

143

7. Most importantly, what approach can be used to study such important questions without biasing the research for the study of a particular molecule rather than for another and for a pathway versus another? IV. Other Options: Genetic Invertebrate Models With some of these questions in mind, and while searching for an animal model that would be suitable to address some of these questions, we have discovered recently that the Drosophila melanogaster is an interesting organism, since these flies are very tolerant to a lack of O2 [31–34]. Drosophila can survive hours of total anoxia and then fly away. What was interesting about the Drosophila phenotype is that these flies sensed anoxia well as they lose coordination and muscle tone within seconds of being made anoxic [31–34]. What is striking is not their resistance in the sensing; that is, it is not that they lack the sensitivity to anoxia. Indeed, what is striking is their recovery, seemingly uninjured, after hours of total anoxia [31–32]. This discovery was important, because these flies have been so well studies in the past century and a great many tools are available for manipulating their genome, their physiology, and their molecular biology (see Chap. 1). Furthermore, with the evolutionary conservation of genetic material, genes, and their pathways, the study of Drosophila is no more simply a study of Drosophila biology but it is much broader in its implications, reaching mammals and humans. In this section, some of the studies that we have recently performed in flies to try to understand the basis for their anoxia tolerance will be reviewed. Although we have adopted several strategies [32–38], the most powerful and useful one is probably the forward genetic approach. Hence, concentration will be on the results obtained using this approach. A. Specific Questions We Asked in Drosophila

There were two major questions that we were interested in. (1) What is the genetic basis for the wide heterogeneity in the response to lack of O2? (2) Can we use anoxia-tolerant organisms to understand the basis of O2 responsiveness? It was also clear to us at that time that the freshwater turtle, which was considered the par excellence model for an anoxia-tolerant organism, was not an optimal organism if cutting-edge molecular techniques were to be employed and if the questions that we listed above were to be addressed. Furthermore, although some mammalian tissues are much more tolerant to lack of O2 than others, and these tissues could be used to analyze the basis for anoxia tolerance, genetic experiments in laboratory rodents may not be easy. When we subjected Drosophila flies to extremely low oxygen concentrations (ⱕ0.01%) and their physiological and behavioral responses studied before,

144

Haddad

during, and after anoxia, they show a very interesting phenomenon [31–32]. When they are first exposed to anoxia (complete lack of O2), Drosophila lose muscle tone and coordination, fall, and become motionless after about 30–60 s in anoxia. They stay motionless for the duration of anoxia. They tolerate a complete N2 atmosphere for several hours following which they seem to recover totally without apparent injury and with the ability to mate, fly, and see normally. Of interest, mean O2 consumption per gram of Drosophila tissue is substantially reduced in low O2 concentrations such as in a 1–2% fractional inspired oxygen level (20% of control) [31–32]. It is very interesting that they survive several hours at this concentration while preserving movements and muscle tone even at 1–2% O2. Also note that the resistance to anoxia is clearly manifested differently in flies as compared to turtles. For example, Drosophila definitely senses the lack of O2, as it responds quickly in a similar way to mammals; that is, flies very quickly develop anoxic stupor when the O2 level is very low [31–32]. In addition, they show a physiological response that is commensurate with this behavioral response. Indeed, extracellular recordings from flight muscles in response to giant fiber stimulation (a well-characterized nerve–muscle system and connections) reveal that muscle evoked potentials (EP) are totally silenced with anoxia within a very short period [31–32]. Furthermore, there is a complete recovery of muscle EP with reoxygenation after a latency that is proportional to the anoxic period, a physiological response that is again similar to the behavioral one. On the other hand, turtles do not seem to lose neuronal activity quickly and can sustain neuronal activity even after many minutes of anoxia [2,5,9,39]. Some neurons in the turtle fire for hours in the total absence of O2 [5]. Hence, unlike turtles, flies seem to have different strategies for survival under very low O2 conditions. From the point of view of sensing, flies seem to behave phenotypically in a manner that is similar to mammals, since they sense the low O2 conditions and respond to it like mammals [31–32]. However, flies seemingly do recover from prolonged anoxia, whereas this is not the case in mammalian organisms and tissues. Hence, the question is how can flies (and other organisms capable of tolerating anoxia) recover from anoxia and survive the severe stress? B. Previous Successes of Genetic Models

Although some of the genetic models, including Drosophila, have been in use for many decades, the conservation of complements of genes with evolution, from prokaryotes to eukaryotes and from yeast, Drosophila, Caenorhabditis elegans, and zebrafish to humans, has become more appreciated in the past decade. With this important discovery of gene conservation, these models have become even more timely and useful in trying to solve problems relevant to human physiology, biology, and disease. Genes responsible for functions as varied as circadian rhythms, aging, alcohol intoxication, development of tracheal buds, heart cham-

Anoxia Tolerance

145

bers, and the eye and central nervous system (CNS), have all been identified first in model systems [40–51] and then studied in mammals as well as humans. Genes responsible for programmed cell death were first cloned in C. elegans (see Chap. 1), and their homologs were found in mammals and humans afterward. The whole field of molecular physiology and, in particular, ion channels received a major boost after cloning the first K⫹ channel from the fruit fly. There are a number of reasons for the success of these genetic models in understanding normal biology. One major reason is that they enjoy important advantages, some conceptual and others technical. Although the space allocated for this minireview does not allow detailing these advantages, a few that have been particularly helpful in work done in Drosophila will be briefly mentioned. (1) The genus Drosophila has a number of useful characteristics of a genetic model: a small number of chromosomes, a generation time of 10 days at 25°C, and a generation size of more than 200–300 per female. (2) There is an enormous number of mutant lines (such as deficiencies, inversions, P-elements) and chromosomal markers available for use. (3) Molecular tools such as cDNA libraries are available. (4) There are tools available for the study of cell or organ physiology in Drosophila (see below). (5) P-elements, which are transposable DNA elements with known sequences, have been very useful in Drosophila for cloning and mapping purposes (see Chap. 1). C. A Genetic Approach

Assuming that there are genes in the Drosophila that protect cells and tissues from anoxic injury, the main idea here is to mutagenize the fly genome and develop a mutagenesis screen in order to identify flies that have lost these genes; that is, flies that have loss of function mutations. Then, these mutations are mapped in detail using marker chromosomes and small deficiencies and then cloned. The genes responsible for the abnormal phenotype are then ascertained in various ways, including the use of transgenic techniques and rescue experiments. This approach has two main advantages: (1) we start with a phenotype which is useful and which is relevant to the question asked; and (2) there is no, or little, bias in terms of the genes and molecules found and the genes that are found in the mutagenesis screen. Although we had no a priori reason to suspect that the genes of interest are exclusively located on the X chromosome, we first focused our screen for mutations on that chromosome for two reasons. The first is to limit our initial task. The main reason, however, is related to the strategy used for this study. An advantage of focusing on the X chromosome is that, using a specific cross (see below and Chap. 1), the screening for mutations on the X chromosome can be observed in the immediate next generation without the need for subsequent single-pair matings.

146

Haddad

There are at least three ways for mutagenizing the genome in the Drosophila: (1) x-rays, (2) ethylmethane sulfonate, and (3) P-elements insertional mutagenesis. There are advantages and disadvantages for each of these methods, but space in this minireview does not allow us to detail these. We have started with all of the three methods, but most of our results so far are obtained from x-ray mutagenesis. We therefore mutagenized (x-ray, 4000 rad) C-S, or wild-type males and crossed them to attached-X females [32]. By so doing, irradiated males (carrying mutations on all three chromosomes) will transmit their mutated X chromosome to the male offspring (normally, the male offspring inherits the X chromosome from the female parent). Therefore, by testing the first-generation male progeny, we could test for mutations irrespective whether they are recessive or dominant. In a specialized apparatus, more than 22,000 flies carrying mutagenized chromosomes have so far been screened in our laboratory [32]. Using a threshold for recovery time (close to the 96th percentile of the wild-type distribution), we have identified to date 10 mutants that have loss of function mutations (eight complementation groups). Although some of these mutants are weak alleles, most have profoundly altered distribution of recovery times after reoxygenation. The marked delay in recovery after anoxia displayed by these mutant flies suggested to us that they were much more sensitive to lack of O2 [32]. Indeed, when we studied some of these fly lines to determine whether they lose coordination at a faster rate, they indeed showed this phenotype as well. The behavioral testing, which showed delayed recovery from anoxia in the mutants, led us to believe that these mutations affected the CNS [32]. To further our understanding of these mutations, we directly examined their effect on CNS function. Identified neurons that can be studied electrophysiologically in Drosophila are those of the Giant Fiber system [32]. During reoxygenation, the wildtype flies started to respond by firing evoked potentials (neurons in CNS are stimulated and muscle action potentials recorded across several synapses, EP) after 2 min into recovery. However, flies with mutations had a much longer latency time to firing of the first EP and some mutant flies required more than 20 min for the first EP [32]. Mapping of the induced mutations was performed with X chromosomal markers and complementation tests [32]. Several markers were used, including y, cv, v, f, car, and su(f). These are markers that are apparent on the body of the fly such as body color, whether wing veins are normal, eye color, and shape of bristles. Complementation testing was done on several X-linked recessive mutations obtained. A number of these mutations were mapped and they are spread in a number of locations on the X chromosome. For example, one of them is on the right side of f, or forked, and another is between y and cv. One of the mutations has been refined in terms of the mapping and we have localized it to a rather narrow region, in between y and cv, using a number of deficiencies. Cloning of the first mutation is underway at present, taking advantage of the possibility that

Anoxia Tolerance

147

x-ray–induced mutations are usually sizable. Since this region has been totally sequenced by the University of California–Berkeley/European Drosophila Genome Project, we have taken a two-step approach to clone this first mutation. We first narrow the region of interest where the mutation lies using Southern analysis and genomic DNA from the mutant and wild-type flies. Subsequently, we will design flanking primers (based on the sequenced DNA) and polymerase chain reaction (PCR) to clone the open reading frames present in that sequence in the mutant Drosophila. In addition to this approach, we are using P-element jumping to produce more alleles and possibly help in the cloning of this mutation. The current genetic screen has not been saturated. However, we had to screen thousands of mutagenized flies to obtain the mutations described. This suggests that a limited number of genes could be mutated in Drosophila to produce similar phenotypes. Because these mutations profoundly disrupted the recovery from anoxia, we believe that this approach can be used effectively to dissect the genetic basis of resistance to anoxia and can help delineate the genes responsible for protection against low O2.

V.

Conclusions

That recovery from anoxia in flies and lack of tissue injury after anoxia are genetically controlled is hardly surprising. What we show in our example, however, is that questions such as what is the genetic basis for anoxia tolerance may now be tractable using these types of genetic models. Although we do not show in this chapter that we have cloned genes responsible for this anoxia tolerance, it is clear from our work that there are mutations that lessen the resistance to anoxia and that there are genes that control processes that are crucial for anoxic injury and survival. As detailed above, it is now a matter of time before we clone some of these genes, study the nature of the molecules, perform rescue experiments to ascertain the roles of these genes, and study how mutations in these genes result in altered cell function and in the phenotype observed. Clearly there is a risk to this type of approach. There are a number of reasons that render the potential risk a real one. For example: (1) X-ray mutagenesis generally produces rather large deletions as compared with chemical mutagenesis. However, it is often the case that a deletion or alteration in the genetic code occurs in a single nucleotide. This renders cloning much more difficult. (2) Jumping P-elements that are close to a gene for the purpose of mutagenizing that gene and cloning it may not be possible as these elements have ‘‘cold’’ and ‘‘hot’’ spots on the genome. (3) Since there are likely many genes that are responsible for the resistance to anoxia, some may not be as interesting as others in the sense that these have been cloned and their actions are well known. An example would be if a mutation occurs in a metabolic enzyme that has been well characterized,

148

Haddad

say lactate dehydrogenase. In this case, this approach would not have generated very exciting new knowledge. The importance of this approach, however, is that it is not biased in terms of molecules or genes to study and focus on. Also, this approach assumes that, in order to be productive, this research needs to proceed over a number of years, since the power of this research is that it should enable us to generate a genetic pathway or network responsible for the phenotype in question. Finally, the impetus for this type of approach has been enhanced for at least two reasons: (1) the evolutionary conservation of gene complements from yeast to humans and (2) the success that this type of approach has had in solving questions about normal biological behavior or disease processes such as cancer, aging, alcohol intoxication, and organ development. References 1. Haddad GG, Jiang C. O2 deprivation in the central nervous system: On mechanisms of neuronal response, differential sensitivity and injury. Prog Neurobiol 1993; 40: 277–318. 2. Fernandes JA, Lutz PL, Tannenbaum A, Todorov AT, Liebovitch L, Vertes R. Electroencephalogram activity in the anoxic turtle brain. Am J Physiol 1997; 273(3 Pt 2):R911–919. 3. Hochachka PW, Buck LT, Doll CJ, Land SC. Unifying theory of hypoxia tolerance: molecular/metabolic defense and rescue mechanisms for surviving oxygen lack. Proc Natl Acad Sci USA 1996; 93(18):9493–9498. 4. Lutz PL, Nilsson GE. Contrasting strategies for anoxic brain survival—glycolysis up or down. J Exp Biol 1997; 200:411–419. 5. Xia Y, Jiang C, Haddad GG. Oxidative and glycolytic pathways in rat (newborn and adult) and turtle brain: role during anoxia. Am J Physiol 1992; 262(31):R595– R603. 6. Young RSK, During MJ, Donnelly DF, Aquila WJ, Perry VL, Haddad GG. Effect of anoxia on excitatory amino acids in brain slices of rats and turtles: in vitro microdialysis. Am J Physiol 1993; 264(33):R716–R719. 7. Haddad GG, Donnelly DF. O2 Deprivation induces a major depolarization in brainstem neurons in the adult but not in the neonatal rat. J Physiol (Lond) 1990; 429: 411–428. 8. Singer D. Neonatal tolerance to hypoxia: a comparative-physiological approach. Comp Biochem Physiol A Mol Integr Physiol 1999; 123(3):221–234. 9. Ghai HS, Buck LT. Acute reduction in whole cell conductance in anoxic turtle brain. Am J Physiol 1999; 277(46):R887–R893. 10. Jiang C, Haddad GG. The effect of anoxia on intracellular and extracellular potassium activity in hypoglossal neurons in-vitro. J Neurophysiol 1991; 66:103–111. 11. Jiang C, Xia Y, Haddad GG. Role of ATP sensitive K⫹ channels during anoxia: major differences between rat (newborn, adult) and turtle neurons. J Physiol (Lond) 1992; 448:599–612.

Anoxia Tolerance

149

12. Jiang C, Haddad GG. Differential responses of neocortical neurons to glucose and/ or O2 deprivation in the human and rat. J Neurophysiol 1992; 68(6):2165–2173. 13. Jiang C, Haddad GG. Short periods of hypoxia activate a K⫹ current in central neurons. Brain Res 1993; 64:352–356. 14. Jiang C, Sigworth FJ, Haddad GG. O2 deprivation activates an ATP-inhibitable K⫹ channels in substantia nigra neurons. J Neurosci 1994; 14:5590–5602. 15. Hochachka PW, Land SC, Buck LT. Oxygen sensing and signal transduction in metabolic defense against hypoxia: lessons from vertebrate facultative anaerobes. Comp Biochem Physiol A Physiol 1997; 118(1):23–29. 16. Pek-Scott M, Lutz PL. ATP-sensitive K⫹ channel activation provides transient protection to the anoxic turtle brain. Am J Physiol 1998; 275(6 Pt 2):R2023–2027. 17. Bickler PE. Reduction of NMDA receptor activity in cerebrocortex of turtles (Chrysemys picta) during 6 wk of anoxia. Am J Physiol 1998; 275(1 Pt 2):R86–91. 18. Buck LT, Bickler PE. Role of adenosine in NMDA receptor modulation in the cerebral cortex of an anoxia-tolerant turtle (Chrysemys picta belli). J Exp Biol 1995; 198(7):1621–1628. 19. Buck LT, Bickler PE. Adenosine and anoxia reduce N-methyl-D-aspartate receptor open probability in turtle cerebrocortex. J Exp Biol 1998; 201(2):289–297. 20. Lutz PL, Kabler S. Release of adenosine and ATP in the brain of the freshwater turtle (Trachemys scripta) during long-term anoxia. Brain Res 1997; 769(2):281– 286. 21. Lutz PL, Manuel L. Maintenance of adenosine A1 receptor function during longterm anoxia in the turtle brain. Am J Physiol 1999; 276(45):R633–R636. 22. Milton SL, Lutz PL. Low extracellular dopamine levels are maintained in the anoxic turtle (Trachemys scripita) striatum. J Cereb Blood Flow Metab 1998; 18(7):803– 807. 23. Pek M, Lutz PL. Role for adenosine in channel arrest in the anoxic turtle brain. J Exp Biol 1997; 200(13):1913–1917. 24. Perez-Pinzon MA, Rosenthal M, Sick TJ, Lutz PL, Pablo J, Mash D. Downregulation of sodium channels during anoxia: a putative survival strategy of turtle brain. Am J Physiol 1992; 262(4 Pt 2):R712–715. 25. Hylland P, Milton S, Pek M, Nilsson GE, Lutz PL. Brain Na⫹ /K⫹-ATPase activity in two anoxia tolerant vertebrates: crucian carp and freshwater turtle. Neurosci Lett 1997; 235(1–2):89–92. 26. Xia Y, Haddad GG. Major differences in CNS sulfonylurea receptor distribution between the rat (newborn, adult) and turtle. J Comp Neurol 1991; 314:278–289. 27. Buck L, Espanol M, Litt L, Bickler P. Reversible decreases in ATP and PCr concentrations in anoxic turtle brain. Comp Biochem Physiol A Mol Integr Physiol 1998; 120(4):633–639. 28. Willmore WG, Storey KB. Antioxidant systems and anoxia tolerance in a freshwater turtle Trachemys scripta elegans. Mol Cell Biochem 1997; 170(1–2):177–185. 29. Willmore WG, Storey KB. Glutathione systems and anoxia tolerance in turtles. Am J Physiol 1997; 273(1 Pt 2):R219–225. 30. Perez-Pinzon MA, Rice ME. Seasonal- and temperature-dependent variation in CNS ascorbate and glutathione levels in anoxia-tolerant turtles. Brain Res 1995; 705(1– 2):45–52.

150

Haddad

31. Krishnan SN, Sun Y-A, Mohsenin A, Wyman RJ, Haddad GG. Behavioral and electrophysiologic responses of Drosophila melanogaster to prolonged periods of anoxia. J Insect Physiol 1997; 43(3):203–210. 32. Haddad GG, Sun Y-A, Wyman RJ, Xu T. Genetic basis of tolerance to O2 deprivation in Drosophila melanogaster. Proc Natl Acad Sci USA 1997; 94:10809–10812. 33. Haddad GG. Mechanisms of anoxia tolerance. A novel approach using a Drosophila model system. Adv Exp Med Biol 1998; 454:273–280. 34. Haddad GG. Enhancing our understanding of the molecular responses to hypoxia in mammals using Drosophila melanogaster. J Appl Physiol 2000; 88:1481–1487. 35. Ma E, Haddad GG. Anoxia regulates gene expression in the central nervous system of Drosophila melanogaster. Mol Brain Res 1997; 46:325–328. 36. Ma E, Haddad GG. A Drosophila Cdk5a-like molecule and its possible role in response to O2 deprivation. Biochem Biophys Res Commun 1999; 261:459–463. 37. Ma E, Haddad GG. Isolation and characterization of the hypoxia-inducible factor 1a in Drosophila Melanogaster. Mol Brain Res 1999; 73:11–16. 38. Ma E, Xu T, Haddad GG. Gene regulation by O2 deprivation: an anoxia-regulated novel gene in Drosophila melanogaster. Mol Brain Res 1999; 63:217–224. 39. Lutz PL, Nilsson GE, Perez-Pinzon MA. Anoxia tolerant animals from a neurobiological perspective. Comp Biochem Physiol B Biochem Mol Biol 1996; 113(1):3– 13. 40. Tischkau SA, Barnes JA, Lin F, Myers EM, Soucy JW, Meyer-Bernstein EL, Hurst WJ, Burgoon PW, Chen D, Sehgal A, Gillette MU. Oscillation and light induction of timeless of mRNA in the mammalian circadian clock. J Neurosci 1999; 19:RC15. 41. Emery P, So WV, Kaneko M, Hall JC, Rosbash M. CRY, a Drosophila clock and light-regulated cryptochrome, is a major contributor to circadian rhythm resetting and photosensitivity. Cell 1998; 95:669–679. 42. Rutila JE, Suri V, Le M, So WV, Rosbash M, Hall JC. Cycle is a second bHLHPAS clock protein essential for circadian rhythmicity and transcription of Drosophila period and timeless. Cell 1998; 93:805–814. 43. Allada R, White NE, So WV, Hall JC, Rosbash M. A mutant Drosophila homolog of mammalian clock disrupts circadian rhythms and transcription of period and timeless. Cell 1998; 93:791–804. 44. Lin YJ, Seroude L, Benzer S. Extended life-span and stress resistance in the Drosophila mutant methuselah. Science 1998; 282:943–960. 45. Moore MS, DeZarro J, LuK AY, Tully T, Singh CM, Heberlein U. Ethanol intoxication in Drosophila: genetic and pharmacological evidence for regulation by the cAMP signaling pathway. Cell 1998; 93:997–1007. 46. Metzger RJ, Krasnow MA. Genetic control of branching morphogenesis. Science 1999; 284:1635–1639. 47. Artavanis-Tsakonas S, Rand MD, Lake RJ. Notch signaling: cell fate control and signal integration in development. Science 1999; 284:770–776. 48. Su MT, Fujioka M, Goto T, Bodmer R. The Drosophila homeobox genes zfh-1 and even-skilled are required for cardiac-specific differentiation of a numb-dependent lineage decision. Development 1999; 126:3241–3251. 49. Park M, Lewis C, Turbay D, Chung A, Chen JN, Evans S, Breitbart RE, Fishman MC, Izumo S, Bodmer R. Differential rescue of visceral and cardiac defects in Dro-

Anoxia Tolerance

151

sophila by vertebrate tinman-related genes. Proc Natl Acad Sci USA 1998; 95:9366– 9371. 50. Gehring WJ, Ikeo K. Pax 6: mastering eye morphogenesis and eye evolution. Trends Genet 1999; 15:371–377. 51. Metzstein MM, Stanfield GM, Horvitz JR. Genetics of programmed cell death in C. elegans: past, present and future. Trends Genet 1998; 14:410–416.

7 Control of Oxygen Homeostasis by Hypoxia-Inducible Factor 1 Essential Roles in Embryogenesis, Physiology, and Disease Pathogenesis

GREGG L. SEMENZA The Johns Hopkins University School of Medicine Baltimore, Maryland

I.

Introduction

Hypoxia plays a clinically significant role in many different human diseases (reviewed in Ref. 1). Disorders associated with systemic hypoxia include altitude sickness, anemia, chronic obstructive pulmonary disease, congestive heart failure, congenital heart disease with right-to-left shunt, and high O2 –affinity hemoglobinopathy. Vascular disorders are associated with local hypoxia, and in these diseases, O2 delivery, energy substrate delivery, and metabolite removal are all affected. Local hypoxia plays a significant role in the pathophysiology of cerebral, coronary, renal, and peripheral ischemic vascular disease. Hypoxia is also a major pathogenetic factor in other vascular-related disease processes, including retinal and tumor neovascularization. In this chapter, evidence supporting the critical involvement of a single transcription factor in a wide variety of developmental, physiological, and pathological responses to hypoxia will be described. These results have been obtained from experiments utilizing tissue culture and animal model systems. In several cases, the investigations have also included clinical studies, thus providing a direct connection between fundamental molecular/biological processes and clinical medicine. 153

154

Semenza II. Hypoxia-Inducible Factor 1 Mediates Transcriptional Activation of Gene Expression in Response to Hypoxia

Hypoxia-inducible factor 1 (HIF-1) is expressed in most mammalian cells in response to reduced O2 availability [2–5]. HIF-1 is a basic helix–loop–helix– PAS (bHLH-PAS) transcription factor consisting of HIF-1α and HIF-1β subunits [6,7] (Fig. 1). HIF-1α expression is quantitatively regulated by cellular O2 concentration and determines the biological activity of HIF-1. In contrast, HIF-1β (also known as the aryl hydrocarbon nuclear translocator [ARNT]) can also dimerize with other bHLH-PAS proteins (reviewed in Ref. 8). The molecular mechanisms by which hypoxia is sensed and the signal is transduced leading to increased HIF-1α expression are not well understood (reviewed in Refs. 9 and 10). To activate transcription of target genes, HIF-1α dimerizes with HIF-1β and the heterodimer binds to DNA at sites represented by the consensus sequence 5′-RCGTG-3′ [11]. The HIF-1 binding site is present within a hypoxia response element, a cis-acting transcriptional regulatory sequence that can be located within 5′-flanking, 3′-flanking, or intervening sequences of target genes. The presence of an intact HIF-1 binding site is necessary for these elements to mediate transcriptional activation [11–13]. The number of target genes that are known to be activated by HIF-1 continues to grow rapidly and includes genes whose protein products are involved in

Figure 1 Structure of HIF-1. The stippled box indicates the bHLH domain and the striped box indicates the PAS domain, which together are required for dimerization of the HIF-1α and HIF-1β subunits and for subsequent DNA binding. A single human HIF1α isoform of 826 amino acids (aa) has been identified, whereas HIF-1β (ARNT) isoforms of 774 and 789 aa are generated by alternative RNA splicing. Functional domains of HIF1α shown are the amino- (N) and carboxyl (C) terminal transactivation domain (TAD), the transcriptional inhibitory domain (ID), and the proline/serine/threonine-rich protein stability domain (PSTD), with the first and last amino acid residue of each domain indicated.

Hypoxia-Inducible Factor 1

155

angiogenesis, energy metabolism, erythropoiesis, cell proliferation and viability, vasomotor responses, and vascular remodeling (Table 1). HIF-1 activates transcription of genes encoding proteins that can be classified into several groups: (1) those that increase O2 delivery to cells, such as erythropoietin (EPO) and vascular endothelial growth factor (VEGF), which stimulate erythropoiesis and angiogenesis, respectively, and nitric oxide synthase-2 and heme oxygenase-1, which synthesize the vasoactive molecules nitric oxide and carbon monoxide, respectively; (2) those involved in metabolic adaptation to hypoxia such as the glucose transporters, GLUT1 and GLUT3, and the glycolytic enzymes aldolase A and C, enolase-1 (ENO1), hexokinase-1 and hexokinase-2, lactate dehydroge-

Table 1 HIF-1 Target Genes. Gene Product Adenylate kinase 3 α1B-Adrenergic receptor Adrenomedullin Aldolase A Aldolase C Endothelin-1 Enolase 1 Erythropoietin (EPO) Glucose transporter-1 (GLUT1) Glucose transporter-3 (GLUT3) Glyceraldehyde phosphate dehydrogenase Heme oxygenase-1 Hexokinase-1 Hexokinase-2 Insulin-like growth factor-2 (IGF-2) IGF binding protein-1 IGF binding protein-2 IGF binding protein-3 Lactate dehydrogenase A Nitric oxide synthase-2 p21cip1 p35srj Phosphofructokinase L Phosphoglycerate kinase-1 Pyruvate kinase M Transferrin Transferrin receptor Vascular endothelial growth factor (VEGF) VEGF receptor FLT-1

References 14 15 16 17 17 18 17 19 14,17 17 17 20 17 17 21 22 21 21 17,23 24 25 26 17 17,23,25 17 27 28,29 17,23,25 30

156

Semenza

nase A, phosphofructokinase L, phosphoglycerate kinase-1, and pyruvate kinase M; and (3) other growth/survival factors, such as insulin-like growth factor-2 (IGF-2) and IGF binding proteins-1, -2, and -3. III. Expression of HIF-1: Inducers, Inhibitors, and Signal Transduction Pathways A. Regulation of HIF-1 Activity by O2 Concentration

The expression of HIF-1 DNA-binding activity and of HIF-1α and HIF-1β protein in human HeLa cells was quantified as a function of O2 concentration [31]. HIF-1α expression (as determined by immunoblot assay) paralleled HIF-1 DNAbinding activity (as determined by electrophoretic mobility-shift assay). There was an approximately twofold increase in HIF-1 expression as O2 concentration declined from 20 to 6% and then an approximately 10-fold increase between 6.0 and 0.5% O2. The response was maximal at 0.5% and half-maximal at 1.5–2.0% O2. The response to hypoxia was similar in the absence or presence of 1 mM KCN as an inhibitor of oxidative phosphorylation and O2 consumption [31]. The characteristics of the response curve are of particular interest because of their physiological relevance: In the heart, epicardial microvascular and myocardial Po2 measurements of 17 and 12 mm Hg (⬃2% O2) have been reported and occlusion of a distal branch of the left anterior descending artery for 1 min resulted in an O2 concentration of near zero in the ischemic core and an O2 gradient as distance from the core increased [32,33]. These data suggest that any reduction of tissue oxygenation in vivo would occur along the steep portion of the HIF-1 response curve, thus providing a mechanism for a graded transcriptional response to hypoxia. Using an isolated perfused and ventilated ferret lung preparation, HIF-1α expression was shown to increase dramatically at inspired O2 concentrations below 4% [5], as described for HeLa cells [31]. B. Pharmacological Inducers and Inhibitors of HIF-1 Activity

The expression of EPO mRNA was induced by exposure of Hep3B human hepatoblastoma cells to 1% O2, divalent cations such as CoCl2, or iron chelators such as desferrioxamine [34,35]. These same stimuli were also shown to induce HIF1 DNA-binding activity and the expression of reporter genes containing the EPO gene HRE [2,13,35–39]. These compounds were subsequently shown to induce both expression of HIF-1α protein and its transactivation function [6,40,41]. Although the mechanisms by which cobalt and desferrioxamine induce EPO expression have not been determined, the present data suggest that they do not act via the same mechanisms as in hypoxia, as described below. The hypoxia signal-transduction pathway may involve a protein kinase/ phosphatase cascade. The induction of HIF-1α protein and HIF-1 DNA binding

Hypoxia-Inducible Factor 1

157

activity was blocked in hypoxic Hep3B cells pretreated with genistein, a tyrosine kinase inhibitor, or sodium fluoride, a serine/threonine phosphatase inhibitor, and the serine/threonine kinase inhibitor 2-aminopurine had a partial inhibitory effect [42]. Cells transformed with the v-Src oncogene showed increased expression of HIF-1α protein, HIF-1 DNA binding activity, HIF-1–dependent reporter genes, and endogenous genes regulated by HIF-1 under nonhypoxic conditions and showed a superinduction under hypoxic conditions [43]. These results suggested that the constitutively active V-SRC tyrosine kinase phosphorylated the target(s) of an endogenous (genistein-inhibited) tyrosine kinase that was normally activated in response to hypoxia. However, the endogenous O2-regulated tyrosine kinase activity is not exclusively due to C-SRC, since c-Src ⫺/⫺ mouse embryo fibroblasts showed normal induction of HIF-1α protein and HIF-1 DNA binding activity in response to hypoxia [43,44]. C. Regulation of HIF-1 Activity by Activated Growth Factor Receptors

The complexity of gene regulation by HIF-1 was increased by the discovery that treatment of Hep3B cells with mersalyl, a cell-impermeant organomercurial compound, induced expression of HIF-1α protein and HIF-1 DNA binding activity [45]. Mersalyl induced expression of VEGF and ENO1 mRNA, as well as reporter genes containing the HRE from the VEGF or EPO gene, under nonhypoxic conditions. Induction of HIF-1 by mersalyl was observed in wild-type mouse embryo fibroblasts but not in mouse embryo fibroblasts lacking the insulin-like growth factor-1 receptor (IGF-1R). In contrast, HIF-1 was induced by hypoxia, CoCl2, or desferrioxamine in MEFs regardless of the presence or absence of IGF1R expression. The mitogen-activated protein (MAP) kinase kinase inhibitor PD098059 blocked the induction of HIF-1 by mersalyl but had no effect on the induction of HIF-1 by hypoxia in wild-type mouse embryo fibroblasts [45]. All three endogenous IGF-1R ligands (IGF-I, IGF-II, and insulin), as well as epidermal growth factor and basic fibroblast growth factor (FGF2), induced HIF-1α expression in human embryonic kidney 293 cells at concentrations of ⬍10 nM [21,46]. These results indicate that (1) physiological stimuli other than hypoxia can result in the expression of HIF-1 and downstream genes via mechanisms distinct from the hypoxia signal-transduction pathway; and (2) HIF-1 may play a wider role in the regulation of energy metabolism, cell proliferation, and cell viability than previously appreciated. D. Effect of Mitochondrial Electron Transport Chain Inhibitors

O2 can be converted to reactive oxygen species (ROS) such as superoxide ion, H2O2, or hydroxyl radical, and one model of O2 sensing proposes that the level of intracellular ROS, which would vary as a function of O2 concentration, deter-

158

Semenza

mines the magnitude of hypoxia signal transduction [47–49]. NADPH oxidoreductases convert O2 to superoxide. However, diphenylene iodonium (DPI), an inhibitor of NADPH oxidoreductases and other flavoproteins, blocks the induction of EPO mRNA and HIF-1 DNA binding activity in cells exposed to hypoxia [50]. This is the opposite result of what is predicted by the NADPH oxidase model, since inhibition of superoxide formation should mimic hypoxia. DPI blocks expression of HIF-1α protein, HIF-1 DNA binding activity, and transcription of a reporter gene containing the EPO HRE in cells exposed to hypoxia but not in cells exposed to desferrioxamine [51,52]. The induction of reporter gene expression by desferrioxamine is HIF-1 dependent, as it can be inhibited by cotransfection of an expression vector encoding a dominant negative form of HIF1α [51], which was previously shown to block hypoxia-induced reporter gene expression [12,19]. Thus, whereas hypoxia and desferrioxamine both induce EPO gene expression via HIF-1, they do so by different mechanisms. In addition to DPI, exposure of hypoxic Hep3B cells to rotenone (another electron transport chain (ETC) complex I inhibitor) or myxothiazol (complex III inhibitor) blocks induction of HIF-1α protein, HIF-1 DNA binding activity, and EPO mRNA expression [51,52]. Several lines of evidence indicate that the observed effects relate to their known inhibition of mitochondrial electron transport at the concentrations used rather than to nonspecific impairment of cellular metabolism: (1) these compounds have no effect on HIF-1 expression induced by CoCl2 or desferrioxamine; and (2) inhibitors of complex II (thenoyltrifluoroacetone) or complex IV (potassium cyanide and sodium azide) do not block induction of HIF-1. A major source of H2O2 production in nonhypoxic cells is mitochondrial ETC complex III [53]. Inhibitors of complex I or III should thus decrease H2O2 production, whereas inhibitors of complex IV would not affect this process. These data thus suggest that ROS production by mitochondria may be necessary for hypoxia signal transduction. This hypothesis is supported by studies which demonstrated that (1) mitochondrial production of H2O2 increased in response to hypoxia in the absence but not in the presence of DPI or rotenone; and (2) ρ0 Hep3B cells which lacked functional mitochondria expressed HIF-1 DNA binding activity in response to CoCl2 or desferrioxamine but not in response to hypoxia [52]. E. Effects of CO And NO on HIF-1 Activity

Both nitric oxide (NO) and carbon monoxide (CO) can inhibit hypoxia-inducible gene expression by interfering with HIF-1 DNA binding activity [54–56]. In bovine pulmonary aortic cells exposed to 1% O2 and 5% CO, hypoxia-induced HIF-1 DNA-binding is inhibited by CO without affecting HIF-1α protein expression, indicating an effect at the level of dimerization or DNA binding [55]. Carbon monoxide may directly bind to HIF-1 or may act indirectly via guanylate cyclase and cGMP-dependent protein kinase. In contrast, NO donors have been

Hypoxia-Inducible Factor 1

159

shown to inhibit HIF-1α protein expression [54], an effect that may be mediated by inhibition of electron transport complex I. Hypoxia-inducible transcription of the HO1 and NOS2 genes encoding heme oxygenase-1 and inducible nitric oxide synthase, which generate CO and NO, respectively, is dependent upon the presence of an intact HIF-1 binding site in cis and HIF-1 expression in trans [20, 24,57,58]. Taken together, these results suggest a potential mechanism for downregulation of HIF-1 activity in cells exposed by either an autocrine or paracrine mechanism to high levels of CO or NO. These effects may be of particular relevance in the responses of vascular endothelial and smooth muscle cells to acute and chronic hypoxia [59].

Figure 2 Signal transduction pathways leading to induction of HIF-1 activity. Hypoxiaand growth factor-induced pathways are shown. Horizontal pointed and blocked arrows represent stimulatory and inhibitory effects, respectively. (Adapted from Ref. 10.)

160

Semenza

In summary, the signal-transduction pathways leading to HIF-1 transcriptional activity appear to be complex and as yet undefined (Fig. 2). Furthermore, most of the experiments have been performed in Hep3B hepatoblastoma and other transformed cell lines and the extent to which these results can be generalized to specific primary cell types in vivo remains to be determined. Whereas Sections I–III have focused on the molecular mechanisms by which HIF-1 activity is induced leading to transcriptional activation of target genes, the remaining sections of this chapter will focus on studies investigating the consequences of HIF-1 activity with respect to essential developmental, physiological, and pathophysiological processes. IV. Generation and Analysis of HIF-1␣–Deficient Embryonic Stem Cells To establish definitively the role of HIF-1 in vivo, the mouse Hif1a gene encoding HIF-1α was inactivated by homologous recombination in embryonic stem (ES) cells [17,23,25]. Hif1a ⫺/⫺ ES cells, which completely lacked HIF-1α expression and HIF-1 DNA binding activity, demonstrated a profound loss of expression of hypoxia-inducible genes, including those encoding VEGF and 13 different glucose transporters and glycolytic enzymes required for the conversion of extracellular glucose to intracellular lactate, thus representing one of the most extensive examples of coordinate transcriptional regulation of a metabolic pathway that has been described in mammalian cells [17]. In addition, Hif1a ⫺/⫺ ES cells showed significantly reduced proliferation under hypoxic culture conditions [17]. V.

Role of HIF-1 in Embryogenesis

In order to investigate the role of HIF-1 in development, Hif1a ⫹/⫺ ES cells were injected into mouse blastocysts and the mutant allele was transmitted through the germline [17,23]. When male and female Hif1a ⫹/⫺ mice were mated, no Hif1a ⫺/⫺ mice were found among several hundred offspring. Analysis of timed matings revealed that development of Hif1a ⫺/⫺ embryos arrested by embryonic day 9.0 (E 9.0) and the mice died by E10.5 [17,23]. The greater severity of the embryonic defects in HIF-1α–deficient mice compared to ARNT (HIF-1β)–deficient mice [60,61] indicates that HIF-1α also dimerizes with other partners such as ARNT2 [62] or ARNT3 [63]. The gross morphology and vascular development of Hif1a ⫺/⫺ and Hif1a ⫹/⫹ embryos were indistinguishable at E8.5–E8.75, but by E9.25 there was a marked regression of blood vessels in the cephalic region and replacement by a smaller number of enlarged vascular structures in the mutant embryos [17]. Concomitant with the disruption of vascular development, massive cell death was observed within the cephalic mesenchyme.

Hypoxia-Inducible Factor 1

161

Similar vascular defects were observed in HIF-1α– and VEGF-deficient embryos, and VEGF mRNA expression was not induced by hypoxia in Hif1a ⫺/⫺ ES cells [17]. Surprisingly, however, Hif1a ⫺/⫺ embryos demonstrated increased VEGF mRNA expression compared to wild-type embryos [64]. In cultured cells, VEGF mRNA expression was induced by glucose deprivation independent of HIF-1α, thus providing a mechanism for increased VEGF mRNA expression in Hif1a ⫺/⫺ embryos in which the absence of adequate tissue perfusion resulted in both O2 and glucose deprivation [17,64]. Rather than being associated with VEGF deficiency, the vascular defects in Hif1a ⫺/⫺ embryos were spatially and temporally correlated with cell death. Prior to the onset of vascular or gross morphological defects, Hif1a ⫺/⫺ embryos manifested cell death, especially at the neurosomatic junction, which represents the site at which neural crest cells emigrate from the neural tube to populate the cephalic mesenchyme [64]. Neural crest cells are sensitive to a variety of teratogens, including hypoxia, perhaps because apoptosis is an essential aspect of their normal developmental program [65]. Neural crest cells make an essential contribution to the developing circulatory system [66]. In particular, neural crest cells are required not for the formation but rather for the persistence of aortic arch arteries [67]. Capillaries in the cephalic mesenchyme consist of endothelial cells coupled to pericytes via gap junctions as early as E9 of mouse development [68], which coincides with the time at which mesenchymal cell death and vascular regression occurred in Hif1a ⫺/⫺ embryos [17]. The loss of mesenchymal cell support of the vascular endothelium may thus represent a critical event leading to vascular regression and remodeling. In addition to the dramatic vascular regression, Hif1a ⫺/⫺ embryos manifested major defects in cardiac morphogenesis by E9.75 [17]. Unlike stage-matched Hif1a ⫹/⫹ embryos, in which the myocardium was consistently organized into two or three concentric cell layers, the myocardium of Hif1a ⫺/⫺ embryos was hyperplastic with as many as 12 cell layers being observed in some regions of the developing heart. The lumen of the heart tube was not patent in Hif1a ⫺/⫺ embryos in contrast to the patent lumen observed in Hif1a ⫹/⫹ embryos, indicating that the abnormal cellular proliferation resulted in a marked reduction in the size of the ventricular cavity and outflow tract. Thus, HIF-1α deficiency has profound effects on both cardiac and vascular development.

VI. Involvement of HIF-1 in the Pathophysiology of Hypoxic Pulmonary Hypertension The embryonic lethality of Hif1a ⫺/⫺ mice precluded investigation of the role of HIF-1 in later developmental processes or in postnatal physiology. In contrast, Hif1a ⫹/⫺ mice developed normally and were indistinguishable from Hif1a ⫹/⫹ littermates when maintained in room air. To determine the effect

162

Semenza

of partial HIF-1α deficiency on physiological responses to chronic hypoxia, 8week-old male Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice were exposed to 10% O2 for 0, 1, 2, 3, 4, 5, or 6 weeks [69]. The mice were weighed and blood was obtained for hematocrit before and after the hypoxic exposure. The mice were sacrificed, the heart was excised, and the mass of the right ventricle (RV) and left ventricle (LV) plus interventricular septum (LV⫹S) was determined. There was no difference in the hematocrit of Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice maintained under normoxic conditions. In contrast, there was a significant difference with respect to the development of polycythemia in Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice under hypoxic conditions (P ⫽ .025 by ANOVA [analysis of variance]). Compared to their wild-type littermates, Hif1a ⫹/⫺ mice showed a significantly delayed erythropoietic response (Fig. 3A). The differences between genotypes were most significant at 1 and 2 weeks. Thereafter, the differences gradually decreased such that at 5 and 6 weeks there was no difference in hematocrit between the two groups. There was no difference between Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice maintained in room air with respect to the RV/LV ⫹ S ratio (Fig. 3B). However, Hif1a ⫹/⫺ mice manifested impaired development of hypoxia-induced right ventricular hypertrophy (P ⫽ .0001 by ANOVA). Despite the small numbers in each group (n ⱕ 10), the differences between genotypes at weeks 1–5 were highly significant (P ⬍ .001 at 2, 3, and 5 weeks and P ⱕ .01 at 1 and 4 weeks by Student’s t-test). As in the case of the erythropoietic response, there was no significant difference between genotypes at 6 weeks. Thus, the development of both polycythemia and right ventricular hypertrophy, two well-documented responses to chronic hypoxia, were impaired but not eliminated in Hif1a ⫹/⫺ mice [69]. To determine whether the right ventricular hypertrophy observed in hypoxic mice was associated with pulmonary hypertension, right ventricular pres-

Figure 3 Impaired responses to chronic hypoxia in Hif1a ⫹/⫺ mice. Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice were exposed to room air (21% O2) or 10% O2 for 1–6 weeks prior to analysis of hematocrit (A), RV/LV⫹S ratio (B), and weight gain (C). (Bar graphs are adapted from Ref. 69.)

Hypoxia-Inducible Factor 1

163

sures were measured directly [69]. There was no significant difference in the mean right ventricular pressures of Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice under normoxic conditions (Fig. 4A). Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice were exposed to 10% O2 for 3 weeks and right ventricular pressures were measured during ventilation with 10% O2. Mean right ventricular pressure was increased in hypoxic Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice, but the degree of pulmonary hypertension was significantly greater in Hif1a ⫹/⫹ mice (18.36 ⫾ 1.88 vs 11.87 ⫾ 0.95 mm Hg, P ⫽ .003). The difference in mean right ventricular pressure between Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice could reflect altered vasomotor and/or vasoproliferative responses to chronic hypoxia. Since these represent dynamic and fixed changes, respectively, we sought to distinguish between them by exposing mice to 10% O2 for 3 weeks and then returning the mice to room air for 3 hr prior to measuring right ventricular pressures while the animals were ventilated with room air. The mean right ventricular pressures of the reoxygenated mice were not significantly different from those of the chronically hypoxic mice. These data suggested that partial deficiency of HIF-1α resulted in impaired hypoxiainduced vascular remodeling in Hif1a ⫹/⫺ mice. To investigate the effects of HIF-1α deficiency on remodeling of pulmonary arterioles, histological sections of lungs from Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice exposed to 10% O2 for 3 weeks were prepared for morphometric analysis [69]. The proportion of nonmuscularized (N), partially muscularized (P), and completely muscularized (C) pulmonary arterioles with an external diameter of ⱕ 100 µm in Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice (Fig. 4B) was significantly different by χ2 analysis (P ⫽ .00001). The wall thickness of completely muscularized pulmonary arterioles with a diameter of ⱕ 100 µm was also determined using two different methods. In both cases, Hif1a ⫹/⫺ mice demonstrated a significant (P ⬍ .001) reduction in wall thickness when compared to Hif1a ⫹/⫹

Figure 4 Impaired development of pulmonary hypertension and vascular remodeling in Hif1a ⫹/⫺ mice. (A) Right ventricular pressures. (B) Muscularization of small pulmonary arterioles. (C) Medial wall thickening in completely muscularized arterioles analyzed as a function of medial area or diameter. (Bar graphs are adapted from Ref. 69.)

164

Semenza

mice (Fig. 4C). These results indicate that not only did chronically hypoxic Hif1a ⫹/⫺ mice have fewer completely muscularized pulmonary arterioles, but the degree of muscularization in such vessels was reduced. The effect of O2 concentration and Hif1a genotype on body weight was also analyzed [69]. There was no significant difference in weight gain by Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice maintained under normoxic conditions for 6 weeks and there was no significant difference between Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice maintained under normoxic conditions with respect to mean body weight at the beginning or end of the study period. Both Hif1a ⫹/⫺ and Hif1a ⫹/⫹ mice lost weight when maintained at 10% O2 and, because weight loss was maximal after 1 week, data from the groups of mice subjected to hypoxia for 1–6 weeks were pooled to increase statistical power (Fig. 3C). Hif1a ⫹/⫺ mice lost a significantly greater percentage of their body weight than Hif1a ⫹/⫹ mice under hypoxic conditions (5.79 ⫾ 0.83% vs 2.99 ⫾ 0.88%, P ⫽ .02). Taken together, these results demonstrate that partial HIF-1α deficiency has significant effects on physiological responses to chronic hypoxia. Despite the presence of one normally functioning allele, Hif1a ⫹/⫺ mice were impaired in the development of polycythemia, right ventricular hypertrophy, pulmonary hypertension, and pulmonary vascular remodeling. Hif1a ⫹/⫺ mice also lost more weight than Hif1a ⫹/⫹ mice. HIF-1α expression increases exponentially as O2 concentration is decreased both in cultured cells [31] and in vivo [5] and levels of HIF-1α correlate with the expression of downstream target genes. These results suggest that the more severe the hypoxic stimulus, the greater the magnitude of HIF-1α expression, HIF-1 DNA binding activity, transcription of downstream genes, and ultimate physiological responses. The data presented above provide a definitive connection between HIF-1α expression and (patho)physiological responses to hypoxia in adult animals. Furthermore, the data are consistent with the hypothesis that inhibition of HIF-1 activity may prevent or ameliorate the development of pulmonary hypertension in patients with chronic lung disease.

VII. Involvement of HIF-1 in Ischemic Neovascularization When near-term fetal sheep were subjected to chronic anemia by daily isovolemic hemorrhage in utero, hematocrit and arterial O2 content decreased threefold within 1 week [70]. Concomitantly, cardiac output increased by 50% and the heart/body weight ratio increased by 30%. To maintain myocardial oxygenation under these circumstances, myocardial blood flow increased fivefold [70]. Capillary morphometry revealed increased capillary density and diameter, and decreased intercapillary distance, in the hearts of anemic fetuses [71]. VEGF protein and mRNA were increased 4.5- and 3.2-fold respectively, and HIF-1α protein expression was increased 3.8-fold in the hearts of anemic fetuses. There was thus

Hypoxia-Inducible Factor 1

165

a striking correlation between increased myocardial vascularization and expression of VEGF and HIF-1α. In contrast to the pulmonary response to chronic hypoxia described above, the response to myocardial hypoxia is clearly adaptive for the organism as, in its absence, cardiac output could not increase sufficiently to compensate for decreased blood O2-carrying capacity and/or myocardial ischemia would ensue (Fig. 5). Therapeutic angiogenesis represents a novel approach to the treatment of ischemic cardiovascular disease, and gene therapy trials involving a number of angiogenic growth factors, such as fibroblast growth factors and various VEGF isoforms, are currently underway (reviewed in Refs. 72–74). The use of HIF-1α gene therapy might have several theoretical advantages. First, HIF-1 will induce the expression of all VEGF isoforms rather than a single isoform as would occur with administration of a VEGF-expressing virus or administration of a recombi-

Figure 5 Cardiovascular responses to chronic anemia in near-term fetal sheep. The potential use of HIF-1α gene therapy for promoting neovascularization of ischemic tissue is also indicated.

166

Semenza

nant VEGF polypeptide. Second, HIF-1 may induce expression of other angiogenic factors that in concert with VEGF will result in better functional outcome. Third, HIF-1 may induce the expression of survival factors such as IGF-2 [21] that may promote myocardial cell survival under ischemic conditions. The clinical efficacy of any strategy of therapeutic angiogenesis remains to be demonstrated, but the possibility that expression of angiogenic factors, either directly or as a result of increased HIF-1α expression, may provide a nonsurgical means of inducing neovascularization in ischemic tissue is an exciting prospect. VIII. Involvement of HIF-1 in Retinal Vascularization Vascularization of the retina occurs in the first week after birth in mice, proceeding from the optic nerve to the periphery, a process that appears to be driven by VEGF expression [75]. HIF-1α expression in the mouse retina was developmentally regulated and showed temporal and spatial correlation with VEGF mRNA expression and retinal vascularization during the first week of postnatal life [76]. These results suggest that hypoxia in the avascular retina induces expression of HIF-1α which results in the transcriptional activation of the VEGF gene and subsequent vascularization as described above for the ischemic myocardium (see Fig. 5). HIF-1α expression also demonstrated temporal and spatial correlation with VEGF mRNA expression in a mouse model of retinal ischemia [76]. Exposure of neonatal mice to 75% O2 for 5 days resulted in decreased expression of HIF1α protein and VEGF mRNA. When the mice were returned to room air, there was a marked increased in the expression of HIF-1α protein and VEGF mRNA, primarily within the inner nuclear layer of the retina. Taken together, these data suggest that in the developing retina the level of HIF-1α expression is modulated by O2 concentration and determines the level of VEGF expression and, thus, the extent of retinal vascularization. These results may be of particular relevance with regard to the pathogenesis of diabetic retinopathy and the retinopathy of prematurity. In contrast to myocardial ischemia, in which neovascularization is a desired therapeutic endpoint, retinal neovascularization in these conditions is a pathological state which, if left untreated, results in blindness. IX. Involvement of HIF-1 in Cerebral Ischemia Following middle cerebral artery occlusion, increased HIF-1α mRNA expression was demonstrated in adult rat brain by Northern blot and in situ hybridization [77]. Expression was observed at 7.5 hr and further increased at 19–24 hr after the onset of ischemia and was localized to the penumbra, which is the viable

Hypoxia-Inducible Factor 1

167

brain tissue surrounding the area of infarction. Expression of HIF-1α mRNA showed temporal and spatial colocalization with mRNAs encoding glucose transporters and glycolytic enzymes (GLUT1, aldolase A, lactate dehydrogenase A, phosphofructokinase L, and pyruvate kinase M), suggesting that increased expression of HIF-1 target genes may contribute to tissue viability in the penumbra [77]. X.

Involvement of HIF-1 in Ischemic Preconditioning

Exposure to hypoxia (8% O2) for 3 hr has been shown to protect against cerebral infarction in 1-week-old rats subjected to combined hypoxia and ischemia (left common carotid artery ligation) 24 hr later [78]. By immunoblot assay, hypoxia alone was shown to induce HIF-1α and HIF-1β protein expression, whereas hypoxia/ischemia did not. By immunohistochemistry, hypoxia induced increased HIF-1α expression throughout the brain parenchyma, whereas hypoxia/ischemia resulted in decreased HIF-1α expression in brain tissue and increased HIF-1α expression within the cerebral vasculature [79]. The expression of HIF-1α mRNA and protein was also induced in the brains of neonatal rats following intraperitoneal injection of cobalt chloride or desferrioxamine, and, like hypoxia, prior cobalt or desferrioxamine administration provided protection against cerebral infarction following hypoxia/ischemia. Furthermore, the rank order of potency (hypoxia ⬎ cobalt ⬎ desferrioxamine) was similar for induction of HIF-1α expression and cerebral protection [79]. The role of HIF-1 in ischemic preconditioning in the heart has not been determined. However, the hearts of rats subjected to chronic systemic hypoxia by 3 weeks’ exposure to 10% O2 were protected against damage caused by prolonged ischemia [80]. Hypoxic adaptation/preconditioning has also been demonstrated in isolated cardiomyocyte preparations [81]. Recent data regarding the role of NOS2 gene expression in mediating ischemic preconditioning in the heart [82] is intriguing given data suggesting that HIF-1 mediates transcriptional activation of NOS2 gene expression in response to hypoxia [24,57,58]. A particularly appealing hypothesis is that induction of HIF-1 and nuclear factor-κB (NF-κB) results in a synergistic activation of NOS2 expression in hypoxic cardiomyocytes. Both of these factors are induced by hypoxia in cardiomyocytes and both can activate NOS2 expression [24,83–86]. XI. Involvement of HIF-1 in Hypoxia-Mediated Apoptosis In contrast to the results presented in the preceding sections, several studies have provided evidence that HIF-1 may in certain settings promote cell death in response to severe hypoxia and/or ischemia. In rodent embryonic stem cells and

168

Semenza

cultured cortical neurons, loss of HIF-1 activity was associated with increased cell survival under the experimental conditions employed [25,87]. The data were consistent with a model in which HIF-1α interacts with p53 and prevents its degradation, leading to increased p53-mediated apoptosis [88]. Analysis of transgenic or knockout animal models will be required to determine the effect of increased or decreased HIF-1 expression on cell survival in ischemic tissue both in the presence and in the absence of a prior preconditioning stimulus.

XII. Involvement of HIF-1 in Tumor Progression Neovascularization and increased glycolysis, two universal characteristics of solid tumors, represent adaptations to a hypoxic microenvironment that are correlated with tumor invasion, metastasis, and lethality (reviewed in Refs. 10 and 89). HIF-1 activates transcription of genes encoding VEGF as well as glucose transporters and glycolytic enzymes (see Table 1), suggesting that it may play an important role in tumor progression. HIF-1–deficient mouse C4 hepatoma cells demonstrated markedly reduced induction of VEGF mRNA in response to hypoxic culture conditions [12,90,91]. When injected into nude mice, C4 cells formed tumors at a reduced rate relative to wild-type Hepal cells, whereas partially revertant RB13 cells were intermediate with respect to levels of HIF-1 activity, VEGF mRNA expression, and rates of tumor growth and vascularization in vivo [43,92]. In a recent immunohistochemical study using a monoclonal antibody that is specific for HIF-1α, overexpression of HIF-1α (relative to the corresponding normal tissue) was demonstrated in common human cancers and their metastases, including breast, colon, lung, and prostate cancer, which represent the major causes of cancer mortality in the United States [93]. HIF-1α overexpression was detected in breast ductal carcinoma in situ and prostatic intraepithelial neoplasia, the earliest malignant lesions in these tissues, but was not observed in nonmalignant tumors such as fibroadenoma of the breast and uterine leiomyoma [93]. In brain tumors, HIF-1α expression was correlated with the degree of vascularization and tumor grade [94]. In the highly aggressive brain tumor glioblastoma multiforme, HIF-1α was highly expressed in viable cells surrounding necrotic regions, a pattern identical to that described for VEGF mRNA in these tumors [95]. In these studies, the induction of VEGF and HIF-1α was most marked in cells that were farthest away from a blood vessel and thus subject to the greatest degree of hypoxia. In contrast to glioblastoma multiforme, hemangioblastoma is a brain tumor which is highly vascularized and does not contain large areas of necrosis. In this tumor type, HIF-1α was also highly expressed, including within tumor cells immediately adjacent to blood vessels, suggesting a mechanism of induction unrelated to cellular O2 concentration [94]. Another highly

Hypoxia-Inducible Factor 1

169

vascularized tumor type, clear cell renal carcinoma, demonstrated a similar pattern [93]. Hemangioblastoma and clear cell renal carcinoma have in common the fact that they are characterized by loss of function mutations in the VHL gene encoding the von Hippel–Lindau (VHL) tumor suppressor protein (reviewed in Ref. 96). Remarkably, in renal carcinoma cell lines lacking VHL, expression of HIF1α (and the related protein HIF-2α [97–100]) was found to be dysregulated such that high levels of the protein were detected under both hypoxic and nonhypoxic conditions [101]. Recent structural and functional studies indicate that VHL functions as a ubiquitin ligase [102,103]. These results suggest that, in hemangioblastoma and clear cell carcinoma cells, the loss of VHL function prevents the ubiquitin-mediated proteasomal degradation of HIF-1α that normally occurs under nonhypoxic conditions [104–106], resulting in constitutive expression of the protein. HIF-1α has also been implicated in autocrine growth factor pathways that are required for tumor cell proliferation and survival [21,107]. As described above, stimulation of receptors for epidermal, fibroblast, or insulin-like growth factors results in expression of HIF-1α protein, HIF-1 DNA binding activity, and expression of VEGF and other downstream genes [21,46]. Remarkably, IGF-2 induces HIF-1α expression which is in turn required for IGF-2 gene expression [21]. Overexpression of IGF-2 is a common event in the progression of colon carcinoma and other common malignancies [108], and these data suggest that increased HIF-1α expression may contribute to this important mechanism for autocrine stimulation of tumor cell proliferation and survival. Taken together, these data that mutations that either activate oncogenes (such as IGF-2 or v-src) or inactivate tumor suppressor genes (such as VHL) result in the increased expression of HIF-1α and downstream target genes such as VEGF, glucose transporters, and glycolytic enzymes that mediate angiogenesis and metabolic adaptations which are essential aspects of tumor progression. The demonstration that HIF-1α overexpression occurs in the majority of common human cancers [93] suggests that it may play a critical role in tumor biology. If this is true, then inhibitors of HIF-1 activity might represent novel and efficacious cancer therapeutics.

XIII. Conclusions Recent studies suggest that HIF-1 plays an essential role in a remarkable number of developmental, physiological, and pathophysiological processes. An important goal of future studies is to determine whether genetic or pharmacological manipulation of HIF-1α expression in vivo may provide therapeutic benefit in one or more of the common causes of mortality in the U.S. population, including cancer,

170

Semenza

chronic lung disease, and ischemic cardiovascular disease. In the case of cancer and chronic lung disease, inhibition of HIF-1 activity would be the therapeutic goal, whereas in ischemic cardiovascular disease, augmentation of endogenous HIF-1 activity would be desired. Toward this end, the same paradigm of combining data from experiments in tissue culture and animal model systems with the results of clinical studies will be necessary in order to evaluate rigorously whether the advances in our understanding of fundamental physiological processes that have been described in this chapter can be translated into advances in patient care. References 1. Semenza GL. Transcriptional regulation by hypoxia-inducible factor: molecular mechanisms of O2 homeostasis. Trends Cardiovasc Med 1996; 6:151–157. 2. Wang GL, Semenza GL. General involvement of hypoxia-inducible factor 1 in transcriptional response to hypoxia. Proc Natl Acad Sci USA 1993; 90:4304–4308. 3. Wenger RH, Rolfs A, Marti HH, Guenet JL, Gassmann M. Nucleotide sequence, chromosomal assignment and mRNA expression of mouse hypoxia-inducible factor 1α. Biochem Biophys Res Commun 1996; 223:54–59. 4. Wiener CM, Booth G, Semenza GL. In vivo expression of mRNAs encoding hypoxia-inducible factor 1. Biochem Biophys Res Commun 1996; 225:485–488. 5. Yu AY, Frid MG, Shimoda LA, Wiener CM, Stenmark K, Semenza GL. Temporal, spatial and oxygen-regulated expression of hypoxia-inducible factor 1 in the lung. Am J Physiol 1998; 275:L818–L826. 6. Wang GL, Jiang B-H, Rue EA, Semenza GL. Hypoxia-inducible factor 1 is a basichelix-loop-helix-PAS heterodimer regulated by cellular O2 tension. Proc Natl Acad Sci USA 1995; 92:5510–5514. 7. Wang GL, Semenza GL. Purification and characterization of hypoxia-inducible factor 1. J Biol Chem 1995; 270:1230–1237. 8. Semenza GL. HIF-1: mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol 2000; 88:1474–1480. 9. Semenza GL. Perspectives on oxygen sensing. Cell 1999; 98:281–284. 10. Semenza GL. Regulation of mammalian O2 homeostasis by hypoxia-inducible factor 1. Annu Rev Cell Dev Biol 1999; 15:551–578. 11. Semenza GL, Jiang B-H, Leung SW, Passantino R, Concordet J-P, Maire P, Giallongo A. Hypoxia response elements in the aldolase A, enolase 1, and lactate dehydrogenase A gene promoters contain essential binding sites for hypoxia-inducible factor 1. J Biol Chem 1996; 271:32529–32537. 12. Forsythe JA, Jiang B-H, Iyer NV, Agani F, Leung SW, Koos RD, Semenza GL. Activation of vascular endothelial growth factor gene transcription by hypoxiainducible factor 1. Mol Cell Biol 1996; 16:4604–4613. 13. Semenza GL, Wang GL. A nuclear factor induced by hypoxia via de novo protein synthesis binds to the human erythropoietin gene enhancer at a site required for transcriptional activation. Mol Cell Biol 1992; 12:5447–5454.

Hypoxia-Inducible Factor 1

171

14. Wood SM, Wiesener MS, Yeates KM, Okada N, Pugh CW, Maxwell PH, Ratcliffe PJ. Selection and analysis of a mutant cell line defective in the hypoxia-inducible factor-1α subunit (HIF-1α): characterization of HIF-1α-dependent and independent hypoxia-inducible gene expression. J Biol Chem 1998; 273:8360–8358. 15. Eckhart AD, Yang N, Xin X, Faber JE. Characterization of the α1B-adrenergic receptor gene promoter region and hypoxia regulatory elements in vascular smooth muscle. Proc Natl Acad Sci USA 1997; 94:9487–9492. 16. Cormier-Regard S, Nguyen SV, Claycomb WC. Adrenomedullin gene expression is developmentally regulated and induced by hypoxia in rat ventricular cardiac myocytes. J Biol Chem 1998; 273:17787–17792. 17. Iyer NV, Kotch LE, Agani F, Leung SW, Laughner E, Wenger RH, Gassmann M, Gearhart JD, Lawler AM, Yu AY, Semenza GL. Cellular and developmental control of O2 homeostasis by hypoxia-inducible factor 1α. Genes Dev 1998; 12:149– 162. 18. Hu J, Discher DJ, Bishopric NH, Webster KA. Hypoxia regulates expression of the endothelin-1 gene through a proximal hypoxia-inducible factor-1 binding site on the antisense strand. Biochem Biophys Res Commun 1998; 245:894–899. 19. Jiang B-H, Rue E, Wang GL, Roe R, Semenza GL. Dimerization, DNA binding, and transactivation properties of hypoxia-inducible factor 1. J Biol Chem 1996; 271:17771–17778. 20. Lee PJ, Jiang B-H, Chin BY, Iyer NV, Alam J, Semenza GL, Choi AMK. Hypoxiainducible factor 1 mediates transcriptional activation of the heme oxygenase-1 gene in response to hypoxia. J Biol Chem 1997; 272:5375–5381. 21. Feldser D, Agani F, Iyer NV, Pak B, Ferreira G, Semenza GL. Reciprocal positive regulation of hypoxia-inducible factor 1α and insulin-like growth factor 2. Cancer Res 1999; 59:3915–3918. 22. Tazuke SI, Mazure NM, Sugawara J, Carland G, Faessen GH, Suen L-F, Irwin JC, Powell DR, Giaccia AJ, Giudice LC. Hypoxia stimulates insulin-like growth factor binding protein 1 (IGFBP-1) gene expression in HepG2 cells: a possible model for IGFBP-1 expression in fetal hypoxia. Proc Natl Acad Sci USA 1998; 95:10188– 10193. 23. Ryan HE, Lo J, Johnson RS. HIF-1α is required for solid tumor formation and embryonic vascularization. EMBO J 1998; 17:3005–3015. 24. Palmer LA, Semenza GL, Stoler MH, Johns RA. Hypoxia induces type II NOS gene expression in pulmonary artery endothelial cells via HIF-1. Am J Physiol 1998; 274:L212–L219. 25. Carmeliet P, Dor Y, Herbert J-M, Fukumura D, Brusselmans K, Dewerchin M, Neeman M, Bono F, Abramovitch R, Maxwell P, Koch CJ, Ratcliffe P, Moons L, Jain RK, Collen D, Keshet, E. Role of HIF-1α in hypoxia-mediated apoptosis, cell proliferation, and tumour angiogenesis. Nature 1998; 394:485–490. 26. Bhattacharya S, Michels CL, Leung M-K, Arany ZP, Kung AL, Livingston DM. Functional role of p35srj, a novel p300/CBP binding protein, during transactivation by HIF-1. Genes Dev 1999; 13:64–75. 27. Rolfs A, Kvietikova I, Gassmann M, Wenger RH. Oxygen-regulated transferrin expression is mediated by hypoxia-inducible factor-1. J Biol Chem 1997; 272: 20055–20062.

172

Semenza

28. Lok CN, Ponka P. Identification of a hypoxia response element in the transferrin receptor gene. J Biol Chem 1999; 274:24147–24152. 29. Tacchini L, Bianchi L, Bernelli-Zazzera A, Cairo G. Transferrin receptor induction by hypoxia: HIF-1-mediated transcriptional activation and cell-specific post-transcriptional regulation. J Biol Chem 1999; 274:24142–24146. 30. Gerber H-P, Condorelli F, Park J, Ferrara N. Differential transcriptional regulation of the two vascular endothelial growth factor receptor genes: Flt-1, but not Flk-1/ KDR, is upregulated by hypoxia. J Biol Chem 1997; 272:23659–23667. 31. Jiang B-H, Semenza GL, Bauer C, Marti HH. Hypoxia-inducible factor 1 levels vary exponentially over a physiologically relevant range of O2 tension. Am J Physiol 1996; 271:C1172–C1180. 32. Benzing H, Losse B, Schuchhardt S, Niederle N. Simultaneous measurement of regional blood flow and oxygen pressure in the dog myocardium during coronary occlusion or hypoxic hypoxia. Adv Exp Med Biol 1973; 37:541–546. 33. Rumsey WL, Pawlowski M, Lejavardi N, Wilson DF. Oxygen pressure distribution in the heart in vivo and evaluation of the ischemic ‘‘border zone.’’ Am J Physiol 1994; 266:H1676–H1680. 34. Goldberg MA, Dunning SP, Bunn HF. Regulation of the erythropoietin gene: evidence that the oxygen sensor is a heme protein. Science 1988; 242:1412–1415. 35. Wang GL, Semenza GL. Desferrioxamine induces erythropoietin gene expression and hypoxia-inducible factor 1 DNA binding activity: implications for models of hypoxia signal transduction. Blood 1993; 82:3610–3615. 36. Beck I, Ramirez S, Weinmann R, Caro J. Enhancer element at the 3′-flanking region controls transcriptional response to hypoxia in the human erythropoietin gene. J Biol Chem. 1991; 266:15563–15566. 37. Pugh CW, Tan CC, Jones RW, Ratcliffe PJ. Functional analysis of an oxygenregulated transcriptional enhancer lying 3′ to the mouse erythropoietin gene. Proc Natl Acad Sci USA 1991; 88:10553–10557. 38. Semenza GL, Nejfelt MK, Chi SM, Antonarakis SE. Hypoxia-inducible nuclear factors bind to an enhancer element located 3′ to the human erythropoietin gene. Proc Natl Acad Sci USA 1991; 88:5680–5684. 39. Wang GL, Semenza GL. Characterization of hypoxia-inducible factor 1 and regulation of DNA-binding activity by hypoxia. J Biol Chem 1993; 268:21513– 21518. 40. Jiang B-H, Zheng JZ, Leung SW, Roe R, Semenza GL. Transactivation and inhibitory domains of hypoxia-inducible factor 1α; modulation of transcriptional activity by oxygen tension. J Biol Chem 1997; 272:19253–19260. 41. Pugh CW, O’Rourke JF, Nagao M, Gleadle JM, Ratcliffe PJ. Activation of hypoxiainducible factor-1; definition of regulatory domains within the α subunit. J Biol Chem 1997; 272:11205–11214. 42. Wang GL, Jiang B-H, Semenza GL. Effect of protein kinase and phosphatase inhibitors on expression of hypoxia-inducible factor 1. Biochem Biophys Res Commun 1995; 216:669–675. 43. Jiang B-H, Agani F, Passaniti A, Semenza GL. V-SRC induces expression of hypoxia-inducible factor 1 (HIF-1) and transcription of genes encoding vascular endothelial growth factor and enolase 1: involvement of HIF-1 in tumor progression. Cancer Res 1997; 57:5328–5335.

Hypoxia-Inducible Factor 1

173

44. Gleadle JM, Ratcliffe PJ. Induction of hypoxia-inducible factor-1, erythropoietin, vascular endothelial growth factor, and glucose transporter-1 by hypoxia: evidence against a regulatory role for src kinase. Blood 1997; 89:503–509. 45. Agani F, Semenza GL. Mersalyl is a novel inducer of vascular endothelial growth factor gene expression and hypoxia-inducible factor 1 activity. Mol Pharmacol 1998; 54:749–754. 46. Zelzer E, Levy Y, Kahana C, Shilo B-Z, Rubinstein M, Cohen B. Insulin induces transcription of target genes through the hypoxia-inducible factor HIF-1α/ARNT. EMBO J 1998; 17:5085–5094. 47. Ehleben W, Bolling B, Merten E, Porwol T, Strohmaier AR, Acker H. Cytochromes and oxygen radicals and putative members of the oxygen sensing pathway. Respir Physiol 1998; 114:25–36. 48. Ehleben W, Porwol T, Fandrey J, Kummer W, Acker H. Cobalt and desferrioxamine reveal crucial members of the oxygen sensing pathway in HepG2 cells. Kidney Int 1997; 51:483–491. 49. Fandrey J, Frede S, Ehleben W, Porwol T, Acker H, Jelkmann W. Cobalt chloride and desferrioxamine antagonize the inhibition of erythropoietin production by reactive oxygen species. Kidney Int 1997; 51:492–496. 50. Gleadle JM, Ebert BL, Ratcliffe PJ. Diphenylene iodonium inhibits the induction of erythropoietin and other mammalian genes by hypoxia: implications for the mechanism of oxygen sensing. Eur J Biochem 1995; 234:92–99. 51. Agani F, Feldser D, Semenza GL. Unpublished data, 1999. 52. Chandel NS, Maltepe E, Goldwasser E, Mathieu CE, Simon MC, Schumacker PT. Mitochondrial reactive oxygen species trigger hypoxia-induced transcription. Proc Natl Acad Sci USA 1998; 95:11715–11720. 53. Turrens JF, Alexandre A, Lehninger AL. Ubisemiquinone is the electron donor for superoxide formation by complex III of heart mitochondria. Arch Biochem Biophys 1985; 237:408–414. 54. Huang LE, Willmore WG, Gu J, Goldberg MA, Bunn HF. Inhibition of hypoxiainducible factor 1 activation by carbon monoxide and nitric oxide: implications for oxygen sensing and signaling. J Biol Chem 1999; 274:9038–9044. 55. Liu Y, Christou H, Morita T, Laughner E, Semenza GL, Kourembanas S. Carbon monoxide and nitric oxide suppress the hypoxic induction of vascular endothelial growth factor gene via the 5′ enhancer. J Biol Chem 1998; 273:15257–15262. 56. Sogawa K, Numayama-Tsuruta K, Ema M, Abe M, Abe H, Fujii-Kuriyama Y. Inhibition of hypoxia-inducible factor 1 activity by nitric oxide donors in hypoxia. Proc Natl Acad Sci USA 1998; 95:7368–7373. 57. Melillo G, Musso T, Sica A, Taylor LS, Cox GW, Varesio L. A hypoxia-responsive element mediates a novel pathway of activation of the inducible nitric oxide synthase promoter. J Exp Med 1995; 182:1683–1693. 58. Melillo G, Taylor LS, Brooks A, Musso T, Cox GW, Varesio L. Functional requirement of the hypoxia-responsive element in the activation of the inducible nitric oxide synthase promoter by the iron chelator desferrioxamine. J Biol Chem 1997; 272:12236–12243. 59. Kourembanas S, Morita T, Liu Y, Christou H. Mechanisms by which oxygen regulates gene expression and cell-cell interaction in the vasculature. Kidney Int 1997; 51:438–443.

174

Semenza

60. Kozak KR, Abbott B, Hankinson O. ARNT-deficient mice and placental differentiation. Dev Biol 1997; 191:297–305. 61. Maltepe E, Schmidt JV, Baunoch D, Bradfield CA, Simon MC. Abnormal angiogenesis and responses to glucose and oxygen deprivation in mice lacking the protein ARNT. Nature 1997; 386:403–407. 62. Hirose K, Morita M, Ema M, Mimura J, Hamada H, Fujii H, Saijo Y, Gotoh O, Sogawa K, Fujii-Kuriyama Y. cDNA cloning and tissue-specific expression of a novel basic-helix-loop-helix/PAS factor (Arnt2) with close sequence similarity to the aryl hydrocarbon receptor nuclear translocator (Arnt). Mol Cell Biol 1996; 16: 1706–1713. 63. Takahata S, Sogawa K, Kobayashi A, Ema M, Mimura J, Ozaki N, Fujii-Kuriyama Y. Transcriptionally active heterodimer formation of an Arnt-like PAS protein, Arnt3, with HIF-1a, HLF, and Clock. Biochem Biophys Res Commun 1998; 248: 789–794, 1998. 64. Kotch LE, Iyer NV, Laughner E, Semenza GL. Defective vascularization of HIF1α–null embryos is not associated with VEGF deficiency but with mesenchymal cell death. Dev Biol 1999; 209:254–267. 65. Lumsden A, Sprawson N, Graham A. Segmental origin and migration of neural crest cells in the hindbrain region of the chick embryo. Development 1991; 113: 1281–1291. 66. Bergwerff M, Verberne ME, DeRuiter MC, Poelmann RE, Gittenberger-de Groot AC. Neural crest cell contribution to the developing circulatory system: implications for vascular morphology? Circ Res 1998; 82:221–231. 67. Waldo KL, Kumiski D, Kirby ML. Cardiac neural crest is essential for the persistence rather than the formation of an arch artery. Dev Dyn 1996; 205:281–292. 68. Bauer HC, Bauer H, Lametschwandtner A, Amberger A, Ruiz P, Steiner M. Neovascularization and the appearance of morphological characteristics of the bloodbrain barrier in the embryonic mouse central nervous system. Dev Brain Res 1993; 75:269–278. 69. Yu AY, Shimoda LA, Iyer NV, Huso DL, Sun X, McWilliams R, Beaty T, Sham JSK, Wiener CM, Sylvester JT, Semenza GL. Impaired physiological responses to chronic hypoxia in mice partially deficient for hypoxia-inducible factor 1α. J Clin Invest 1999; 103:691–696. 70. Davis LE, Hohimer AR. Hemodynamics and organ blood flow in fetal sheep subjected to chronic anemia. Am J Physiol 1991; 261:R1542–R1548. 71. Martin C, Yu AY, Jiang B-H, Davis L, Kimberly D, Hohimer AR, Semenza GL. Cardiac hypertrophy in chronically anemic fetal sheep: increased vascularization is associated with increased myocardial expression of vascular endothelial growth factor and hypoxia-inducible factor 1. Am J Obstet Gynecol 1998; 178:527– 534. 72. Ferrara N. Molecular and biological properties of vascular endothelial growth factor. J Mol Med 1999; 77:527–543. 73. Losordo DW, Vale PR, Isner JW. Gene therapy for myocardial angiogenesis. Am Heart J 1999; 138:132–141. 74. Sellke FW, Simons M. Angiogenesis in cardiovascular disease: current status and therapeutic potential. Drugs 1999; 58:391–396.

Hypoxia-Inducible Factor 1

175

75. Stone J, Itin A, Alon T, Pe’er J, Gnessin H, Chan-Ling T, Keshet E. Development of retinal vasculature is mediated by hypoxia-induced vascular endothelial growth factor (VEGF) expression by neuroglia. J Neurosci 1995; 15:4738–4747. 76. Ozaki H, Yu AY, Della N, Ozaki K, Luna JD, Yamada H, Hackett SF, Okamoto N, Zack DJ, Semenza GL, Campochiaro PA. Hypoxia-inducible factor 1α is increased in ischemic retina: temporal and spatial correlation with VEGF expression. Invest Ophthalmol Vis Sci 1999; 40:182–189. 77. Bergeron M, Yu AY, Solway KE, Semenza GL, Sharp FR. Induction of hypoxiainducible factor 1 (HIF-1) and its target genes following focal ischemia in rat brain. Eur J Neurosci 1999; 11:4159–4170. 78. Gidday JM, Fitzgibbons JC, Shah AR, Park TS. Neuroprotection from ischemic brain injury by hypoxic preconditioning in the neonatal rat. Neurosci Lett 1994; 168:221–224. 79. Bergeron M, Gidday JM, Yu AY, Semenza GL, Ferriero DM, Sharp FR. Hypoxiainducible factor 1 (HIF-1) role in hypoxia-induced ischemic tolerance in neonatal rat brain. Ann Neurol 2000; 48:285–296. 80. Tajima M, Katayose D, Bessho M, Iyohama S. Acute preconditioning and chronic hypoxia independently increase myocardial tolerance to ischemia. Cardiovasc Res 1994; 28:312–319. 81. Silverman HS, Wi S-k, Haigney MCP, Ocampo CJ, Stern MD. Myocyte adaptation of chronic hypoxia and development of tolerance to subsequent acute severe hypoxia. Circ Res 1997; 80:699–707. 82. Bolli R, Manchikalapudi S, Tang XL, Takano H, Qiu Y, Guo Y, Zhang Q, Jadoon AK. The protective effect of late preconditioning against myocardial stunning in conscious rabbits is mediated by nitric oxide synthase: evidence that nitric oxide acts as both a trigger and as a mediator of the late phase of ischemic preconditioning. Circ Res 1997; 81:1094–1107. 83. Das DK, Engelman RM, Maulik N. Oxygen free radical signaling in ischemic preconditioning. Ann NY Acad Sci 1999; 874:49–65. 84. Koong AC, Chen EY, Giaccia AJ. Hypoxia causes the activation of nuclear factor κB through the phosphorylation of IκBα on tyrosine residues. Cancer Res 1994; 54:1425–1430. 85. Taylor BS, de Vera ME, Ganster RW, Wang Q, Shapiro RA, Morris SM Jr, Billiar TR, Geller DA. Multiple NF-κB enhancer elements regulate cytokine induction of the human inducible nitric oxide synthase gene. J Biol Chem 1998; 273:15148– 15156. 86. Xie QW, Kashiwabara Y, Nathan C. Role of transcription factor NF-κB/Rel in induction of nitric oxide synthase. J Biol Chem 1994; 269:4705–4708. 87. Halterman MW, Miller CC, Federoff HJ. Hypoxia-inducible factor 1α mediates hypoxia-induced delayed neuronal death that involves p53. J Neurosci 1999; 19: 6618–6624. 88. An WG, Kanekal M, Simon MC, Maltepe E, Blagosklonny, Neckers LM. Stabilization of wild-type p53 by hypoxia-inducible factor 1α. Nature 1998; 392:405–408. 89. Dang CV, Semenza GL. Oncogenic alterations of metabolism. Trends Biochem Sci 1999; 24:68–72. 90. Salceda S, Beck I, Caro J. Absolute requirement of aryl hydrocarbon receptor nu-

176

91.

92.

93.

94.

95.

96. 97.

98.

99.

100.

101.

102.

103.

Semenza clear translocator protein for gene activation by hypoxia. Arch Biochem Biophys 1996; 334:389–394. Wood SM, Gleadle JM, Pugh CW, Hankinson O, Ratcliffe PJ. The role of the aryl hydrocarbon receptor nuclear translocator (ARNT) in hypoxic induction of gene expression: studies in ARNT-deficient cells. J Biol Chem 1996; 271:15117–15123. Maxwell PH, Dachs GU, Gleadle JM, Nicholls LG, Harris AL, Stratford IJ, Hankinson O, Pugh CW, Ratcliffe PJ. Hypoxia-inducible factor 1 modulates gene expression in solid tumors and influences both angiogenesis and tumor growth. Proc Natl Acad Sci USA 1997; 94:8104–8109. Zhong H, De Marzo AM, Laughner E, Lim M, Hilton DA, Zagzag D, Buechler P, Isaacs WB, Semenza GL, and Simons JW. Overexpression of hypoxia-inducible factor 1α in common human cancers and their metastases. Cancer Res 1999; 59: 5830–5835. Zagzag D, Zhong H, Scalzitti JM, Laughner E, Simons JW, Semenza GL. Immunolocalization of hypoxia-inducible factor 1 in brain tumors suggests two distinctive pathways to angiogenesis. Cancer 2000; 88:2606–2618. Shweiki D, Itin A, Soffer D, Keshet E. Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 1992; 359:843– 845. Kaelin WG Jr, Maher ER. The VHL tumor-suppressor gene paradigm. Trends Genet 1998; 14:423–426. Ema M, Taya S, Yokotani N, Sogawa K, Matsuda Y, Fujii-Kuriyama Y. A novel bHLH-PAS factor with close sequence similarity to hypoxia-inducible factor 1α regulates VEGF expression and is potentially involved in lung and vascular development. Proc Natl Acad Sci USA 1997; 94:4273–8. Flamme I, Frohlich T, von Reutern M, Kappel A, Damert A, Risau W. HRF, a putative basic helix-loop-helix-PAS-domain transcription factor is closely related to hypoxia-inducible factor 1α and developmentally expressed in blood vessels. Mech Dev 1997; 63:51–60. Hogenesch JB, Chan WK, Jackiw VH, Brown RC, Gu Y-Z, Perdew GH, Bradfield CA. Characterization of a subset of the basic helix-loop-helix-PAS superfamily that interacts with components of the dioxin signalling pathway. J Biol Chem 1997; 272:8581–8593. Tian H, McKnight SL, Russell DW. Endothelial PAS domain protein 1 (EPAS1), a transcription factor selectively expressed in endothelial cells. Genes Dev 1997; 11:72–82. Maxwell PH, MS Wiesener, G-W Chang, SC Clifford, EC Vaux, ME Cockman, CC Wykoff, CW Pugh, ER Maher, and PJ Ratcliffe. The tumor suppressor protein VHL targets hypoxia-inducible factors for oxygen-dependent proteolysis. Nature 399:271–275, 1999. Lisztwan J, Imbert G, Wirbelauer C, Gstaiger M, Krek W. The von Hippel-Lindau tumor suppressor protein is a component of an E3 ubiquitin-protein ligase activity. Genes Dev 1999; 13:1822–1833. Stebbins CE, Kaelin WG Jr, Pavletich NP. Structure of the VHL-elonginC-elonginB complex: implications for VHL tumor suppressor function. Science 1999; 284:455–460.

Hypoxia-Inducible Factor 1

177

104. Huang LE, Gu J, Schau M, Bunn HF. Regulation of hypoxia-inducible factor 1α is mediated by an O2-dependent degradation domain via the ubiquitin-proteasome pathway. Proc Natl Acad Sci USA 1998; 95:7987–7992. 105. Kallio PJ, Wilson WJ, O’Brien S, Makino Y, Poellinger L. Regulation of the hypoxia-inducible transcription factor 1α by the ubiquitin-proteasome system. J Biol Chem 1999; 274:6519–6525. 106. Salceda S, Caro J. Hypoxia-inducible factor 1α (HIF-1α) protein is rapidly degraded by the ubiquitin-proteasome system under normoxic conditions: its stabilization by hypoxia depends on redox-induced changes. J Biol Chem 1997; 272: 22642–22647. 107. Zhong H, Agani F, Baccala AA, Laughner E, Rioseco-Camacho N, Isaacs WB, Simons JW, Semenza GL. Increased expression of hypoxia inducible factor-1α in rat and human prostate cancer. Cancer Res 1998; 58:5280–5284. 108. Zhang L, Zhou W, Velculescu VE, Kern SE, Hruban RH, Hamilton S, Vogelstein B, Kinzler KW. Gene expression profiles in normal and cancer cells. Science 1997; 276:1268–1272.

Part Three CARDIOVASCULAR AND RED BLOOD CELL DEVELOPMENT

8 A Genetic Model for Cardiac Pattern Formation and Cell Fate Determination

WENDY K. LOCKWOOD, MARGARET LIU, MING-TSAN SU, and ROLF BODMER University of Michigan Ann Arbor, Michigan

I.

Introduction

The Drosophila heart shares many embryological and molecular similarities with those of vertebrates despite the morphological differences of the final products. For example, their mesodermal origin and initial assembly into a linear heart tube are comparable in many respects. Moreover, numerous gene functions have been identified that are utilized by both vertebrates and Drosophila for the specification and differentiation of the heart progenitor cells. Because of the simplicity in structure and the wealth of genetic tools available, the Drosophila heart has emerged as a pioneering model system for elucidating the molecular basis of heart development. Heart development studies in Drosophila began with the discovery of the homeobox gene tinman, which then led to the isolation of numerous tinman homologs in vertebrates (for review, see Bodmer, 1995; Harvey, 1996; Olson and Srivastava, 1996; Fishman and Chien, 1997; Fishman and Olson, 1998; Newman and Krieg, 1998; Tanaka et al., 1998; Bodmer and Frasch, 1999). tinman Gene functions are central to our understanding of the molecular basis of heart development and, as recent studies have shown, have also shed a spectacular amount of 179

180

Lockwood et al.

light on the molecular basis of human heart disease (Benson et al. 1998; Schott et al., 1998). A number of other gene functions have recently been identified in Drosophila that are important for heart development. We are now beginning to understand not only their individual functions in cardiac induction and cell type specification, but also their interactions and the genetic logic by which they act combinatorially to determine correctly and position the heart within the embryo. The outcome of these studies are an essential prerequisite designed to elucidate the genetic basis of cardiac morphogenesis and function.

II. Cardiac Morphogenesis of Insects and Vertebrates The heart of Drosophila is a simple linear tubular structure that forms at the dorsal midline of the embryo and pumps the hemolymph through the larval body cavity in an open circulatory system. Located in the middle of the heart is a simple valve. The heart in Drosophila consists of two major cell types: the inner two rows are contractile muscle cells (myocardial cells) which are flanked on each side by a row of pericardial cells. The myocardial cells are crescent shaped and face each other. They contact each other at the midline and form a central cavity, the lumen of the heart. Very much in contrast to the fly is the vertebrate heart, which is a complex, asymmetrically looped organ containing multiple valves, chambers, and tissue layers; that is, numerous specialized cell types (Harvey and Rosenthal, 1999). Despite the obvious morphological differences of the mature heart between insects and vertebrates, there are considerable similarities during the early embryonic induction events of cardiogenesis. Most interestingly, the embryonic origin of the insect heart and that of vertebrates are apparently equivalent: both originate from lateral (as opposed to medial) mesoderm that migrates most distally from the point of invagination during gastrulation, and both form initially as a linear heart tube (Fig. 1A and 2). Moreover, in both systems, the heart assembles from bilaterally symmetrical mesodermal progenitors that join together at the embryonic midline opposite to the central nervous system (CNS) and the gut (Fig. 2). Because of the reversal of the dorsal–ventral axis between vertebrates and invertebrates, the insect heart forms dorsally, whereas the vertebrate heart forms ventrally (Bodmer, 1995). Additional mechanisms must operate in vertebrates to limit the region of cardiogenesis to the anterior portion of the lateral plate mesoderm in the embryo. In contrast to vertebrates, the insect heart does not undergo any further dramatic morphogenetic events and remains a linear pulsating vessel along most of the trunk region of the embryo. Since the Drosophila heart contains a limited number of easily identifiable cell types, and many of the cardiogenic gene functions are apparently conserved in vertebrates, this insect model has

Cardiac Pattern Formation and Cell Fate Determination

181

Figure 1 Gastrulation and mesodermal subdivision in Drosophila. (A) Schematic cross sections of a Drosophila embryo at progressively later stages of embryogenesis. The mesoderm invaginates along the ventral midline during gastrulation (stage 8) and expresses tinman. The mesoderm migrates then dorsally in a monolayer of cell and maintains tinman expression only in the dorsal portion of the mesoderm in a dpp-dependent fashion (stage 10). Subsequently, the mesoderm subdivides into cardiac, visceral, and somatic mesoderm, and tinman expression is maintained exclusively in the cardiac progenitors and most of its progeny (stage 12). (B) Flow diagram of the major mesodermal specification events necessary for cardiac determination (see text). The genes encoding signaling factors, dpp, decapentaplegic, and wg, wingless, are genes encoding signaling factors, are secreted primarily from the ectoderm and affect patterning of the underlying mesoderm.

proven to be well suited for identifying and studying the genes and their interactions that specify cardiac patterning and determine the identity of individual cardiac cell types. III. The Cardiogenic Role of tinman and Its Vertebrate Homologs An elaborate but well-understood network of genes is responsible for determining and positioning the mesoderm in the blastoderm embryo of Drosophila (reviewed

182

Lockwood et al.

Figure 2 Linear heart tube assembly and tinman-like gene expression is similar in flies and vertebrates. (A,B) Dorsal view of tinman expression in the differentiating Drosophila heart before (A) and after (B) assembly into a linear tube. (C,D) Nkx.25 expression in the cardiac progenitors of a chick embryo before (C) and after (D) assembly into the linear heart tube. Arrowheads in (A) and (C) indicate direction of migration of the bilateral heart primordia. (Micrographs C and D are courtesy of T. Schultheiss.)

in St Johnston and Nu¨sslein-Volhard, 1992). A cascade of maternal gene activation culminates in the translocation of the rel-related transcription factor, encoded by the dorsal gene, into the nucleus along the ventral side of the embryo. Nuclear dorsal protein is necessary and sufficient to specify mesoderm and to activate the first zygotic mesoderm determinants, twist and snail, which code for basic helix-loop-helix and zinc finger–containing transcription factors, respectively. In the absence of twist function, no mesoderm forms. With the onset of gastrulation, the mesoderm then invaginates along the ventral midline and migrates dorsally in a monolayer of cells dorsally (see Fig. 1A). At this point, the first mesodermal subdivisions occur. First, the dorsal and ventral portions of the mesoderm become distinct in their myogenic potential (see Fig. 1A, middle panel): The dorsal mesoderm will give rise to the heart, visceral gut muscles, and dorsal skeletal muscles, whereas the ventral mesoderm generates the fat body (a liver equivalent) and the ventral skeletal muscles. Subsequently, the physical separation between these four major mesodermal derivatives of the embryo trunk region takes place (see Fig. 1A, right panel). The genetic basis of determining the cardiac mesoderm in Drosophila has recently been elucidated (see below; for review, see Bodmer et al., 1997; Bodmer and Frasch, 1999). A crucial determinant of heart in Drosophila is encoded by tinman, a NK2 class homeodomain-containing transcription factor (Bodmer et al., 1990). tinman Is first expressed uniformly in the presumptive mesoderm, and this expres-

Cardiac Pattern Formation and Cell Fate Determination

183

sion depends critically on twist, the determinant of all mesoderm (see Fig. 1B). After dorsal migration during gastrulation, tinman expression is limited to the dorsal portion of the mesoderm which gives rise to the heart, the gut muscles, and some dorsal skeletal muscles. It is this dorsally restricted expression of tinman that first distinguishes between dorsal and ventral mesodermal cell fates during development (see Fig. 1A, middle panel). Experimental evidence for this claim stems from the phenotype of tinman mutants: They are missing all dorsal mesodermal derivatives; that is, no progenitors for the heart or the visceral or the dorsal skeletal muscles are formed (Azpiazu and Frasch, 1993; Bodmer, 1993). Thus, tinman is crucial for subdividing the mesoderm and endowing the dorsal mesoderm with the competence to form heart, in addition to other dorsal mesodermal cell types (see also Bodmer et al.,1997; Bodmer and Frasch, 1999). Later, tinman is expressed transiently in a portion of the visceral mesoderm and permanently in the forming heart (see Fig. 1A, right panel). It is not known whether the cardiac progenitor-restricted expression tinman is essential for correct morphogenesis or function of the heart. Recently, individual enhancer elements of the tinman gene have been identified that distinguish expression in the entire early mesoderm, the dorsal mesoderm, and the myocardial and pericardial cells of the heart (Yin et al., 1997; Xu et al., 1998, Venkatesh et al., 2000). Since tinman is initially expressed in all mesodermal cells of the trunk region of the embryo (see Fig. 1A, left panel), additional patterning events must take place to restrict its expression dorsally. Indeed, the product of decapentaplegic (dpp), a secreted factor of the tumor growth factor-β (TGF-β) superfamily, acts as an inductive signal originating from the dorsal ectoderm to maintain tinman expression in the underlying dorsal mesoderm (see Fig. 1A, middle panel) (Staehling-Hampton et al., 1994; Frasch, 1995; reviewed in Venkatesh and Bodmer, 1995). As in tinman mutants, heart and visceral mesoderm formation is abolished in dpp mutant embryos. In contrast, when the dpp pathway is activated in more ventral regions, the tinman expression domain is also expanded ventrally (Frasch, 1995). Thus, TGF-β signaling in the appropriate (i.e., mesodermal) context is prerequisite for heart specification (see Fig. 1B). The further restriction of tinman expression to the heart progenitor cells in segmental clusters at the dorsal edge of the mesoderm must be due to additional patterning information (see Fig. 1A, left panel and Fig. 2A). The signal responsible is encoded by wingless, the original Wnt gene of Drosophila, which is also secreted by the overlaying ectoderm. As dpp, wingless function is absolutely necessary for cardiac mesoderm formation, but unlike dpp, is not required for the induction of the visceral mesoderm (see Fig. 1B) (Wu et al., 1995; Park et al., 1996, 1998a; reviewed in Bodmer et al., 1997; Bodmer and Frasch, 1999). wingless Is expressed orthogonally to dpp in the embryo and thereby seems to provide positional information to subdivide further the dorsal mesoderm and to contribute to the distinction between the heart and the visceral mesoderm primor-

184

Lockwood et al.

dia (for details, see Sec. IV). All the known components of the wingless pathway participate in transducing the cardiogenic wingless signal, with the exception of the GSK3β-kinase encoded by the shaggy/zeste-white3 gene, which not only has a cardiac-specific function downstream of wingless, but is also required for specifying the entire dorsal mesoderm (Park et al., 1996, 1998a). Until recently, the molecular basis for understanding cardiac induction and differentiation in vertebrate systems has been largely unknown due to the lack of suitable genetics and in vitro systems (Olson and Srivastava, 1996). For example, the myogenic genes that seem to be at the top of the hierarchy in skeletal myogenesis are not expressed in the developing heart. Thus, the importance of tinman in Drosophila prompted many laboratories to search for homologs of tinman in vertebrates. The first tinman homolog was isolated in the mouse (named Nkx.25 or Csx), and it has since been found in every vertebrate species investigated, including the chordate Amphioxus (Komuro and Izumo,1993; Lints et al., 1993; Tonissen et al., 1994; reviewed in Harvey 1996; Newman and Krieg, 1998; Tanaka et al., 1998; T.V. Venkatesh and R. Bodmer, unpublished data). Nkx.25 is initially expressed in the bilateral cardiac progenitors of the anterior plate mesoderm and in the pharyngeal endoderm (see Fig. 2C), and its is the earliest marker known for the cardiac lineage in vertebrates. As for tinman in Drosophila, the bilateral strands of Nkx.25-expressing heart progenitors fuse medially into a linear heart tube (see Fig. 2). In mice homozygous for a Nkx.25 knockout mutation, the early heart tube does form, and most contractile proteins are expressed, except for a ventriculespecific myosin light chain gene (Lyons et al., 1995; Tanaka et al., 1999). However, a number of transcription factors are not expressed in the developing heart, and the linear heart tube fails to undergo further differentiation and morphogenesis in Nkx.25 mutant hearts (Biben and Harvey, 1997; Zou et al., 1997; Tanaka et al., 1999). Moreover, in the Xenopus system, dominant-negative forms of Nkx.25 interfere strongly with heart formation (Fu et al., 1998; Grow and Krieg, 1998). Although the loss of function phenotype demonstrates that Nkx.25 is required for cardiac differentiation, the mutant defect does not result in the complete absence of heart and visceral mesoderm formation, in contrast to tinman loss of function mutants in Drosophila. This may be because several other tinman-related genes, in addition to Nkx.25, also show early cardiac expression and may thus have partially redundant function in vertebrate heart development (Evans et al., 1995; Lee et al., 1996; Reecy et al., 1997; Newman and Krieg, 1998; Tanaka et al., 1998). So far only two members, Nkx.23 and Nkx.25 have been reported in multiple species, whereas Nkx.26/Tix, Nkx.27, Nkx.28, and Nkx.29 have only been described in single species. Aside from the homeodomain, significant sequence similarities between tinman-related genes are also present in a N terminal 11–amino acid region, the TN domain, and a C terminal 20–amino

Cardiac Pattern Formation and Cell Fate Determination

185

acid stretch of hydrophobic residues, the NK2-specific domain (Bodmer, 1995; Harvey, 1996; Newman and Krieg, 1998; Tanaka et al., 1998). Intriguingly, tinman itself has apparently lost the NK2-specific domain during evolution, probably because tinman in the fly has diverged more from the postulated common ancestor than its vertebrate counterparts (for discussion, see Park et al., 1998b). Evidence that strongly supports an evolutionary relationship, but also indicates functional divergence, between dipteran tinman and its vertebrate homologs stems from experiments using vertebrate tinman transgenes expressed in tinman mutant flies. Intriguingly, when vertebrate tinman homologs from several different species are expressed in Drosophila embryos, they can readily substitute for Drosophila tinman function in visceral mesoderm formation but restore heart development only slightly or not at all (Park et al., 1998b; Ranganayakulu et al., 1998). The difference apparently is encoded in the N terminal region of Tinman. In a converse experiment, it will be interesting to see whether or not fly tinman knocked into the Nkx.25 locus can substitute for the loss of Nkx.25-function in cardiac morphogenesis. The clinical importance of tinman genes in heart development has recently been demonstrated by the discovery of multiple forms of congenital heart disease as a result of mutations in the human form of NKX2–5. These diseases include cardiac malformations, atrial (or ventricular) septal defects, and atrioventricular conduction delays (Benson et al., 1998; Schott et al., 1998). Interestingly, these abnormalities are observed in patients that are heterozygous for a variety of NKX.25 gene disruptions, suggesting that tinman-like functions may be required throughout life. It remains to be seen if heterozygous Nkx.25 mutant mice can serve as a useful animal model for certain forms of congenital heart disease. These observations led to the speculation that tinman and its homologs are not only required for determination of the cardiac mesoderm and differentiation of the early heart tube, but also for later aspects of cardiac differentiation, including the correct functioning of the mature heart (Schott et al., 1998).

IV. Cardiac Induction: Signals and Context An experimental approach to the problem of how cardiac competence and heart progenitors are specified is to determine which set of cues is required for that specification. This is easily addressed genetically in Drosophila by identifying genes which when mutated result in the loss of heart specification. However, even among this class of genes, some are more informative with respect to the specification process than others. For example, each gene required for heart specification requires a set of genes to activate or maintain its own expression. Thus, indirectly these ‘‘regulatory’’ genes will also affect heart specification by affect-

186

Lockwood et al.

ing a gene directly required for heart specification. We focus here on wingless, dpp, and tinman, three genes that appear to play direct and independent roles in cardiac mesoderm induction. A related problem is to understand how heart precursors are positioned spatially. This question is more difficult to address experimentally, since any gene required to position the heart is presumably also required to specify the heart. In principle, it seems plausible that the patterns of those genes required for heart specification must spatially limit the region in which heart precursors can be specified, since the lack of any required gene product is sufficient to prevent heart formation. Moreover, if each required gene product spatially restricts the position of heart precursors, then heart specification can only occur at the overlap of all required gene products. In this section, we summarize the evidence that wingless, dpp, and tinman are not only required for specification of the heart, but that their spatial patterns of gene expression also determine the precise embryonic position in which heart specification takes place. A. wingless, dpp, and tinman Are Required for Heart Cell Specification

As summarized in the previous section, wingless and dpp signaling and the transcription factor tinman are each essential for the specification of the cardiac mesoderm. Moreover, wingless and dpp, by signaling primarily from the ectoderm to the underlying mesoderm, are required more than once during embryonic development in order to specify the normal development of the heart (Frasch, 1995; W.K.L. and R.B., unpublished data). Importantly, however, the heart is not the only mesodermal cell type that requires all or a subset of these three gene products at one and/or another time during development. In the following section, we describe the spatial and temporal function of these three gene products in specifying and positioning the Drosophila heart. B. Spatial Intersect of wingless and dpp Correlates with the Position of Heart Cell Specification

The heart is specified by tinman-expressing mesodermal cells in dorsal segmented clusters where these cells are also exposed to signaling by wingless and dpp. Prospective heart precursors express tinman not only at early stages, when tinman is expressed uniformly in the trunk mesoderm, but also at late stages, when tinman expression is restricted exclusively to the heart. Heart precursors are specified dorsally in regions exposed to dpp at both early stages when dpp is expressed in a broad dorsal band and again later when dpp takes on a restricted expression pattern at the dorsal ectodermal edge (Fig. 3A). Finally, these heart precursors are specified along the anterior–posterior axis in each segment in a region that is

Cardiac Pattern Formation and Cell Fate Determination

187

exposed to wingless at both early stages when wingless is expressed in continuous stripes and again later when these stripes have been discontinuous (Fig. 3A). The patterns of wingless, dpp, and tinman at any one stage cannot explain the localization of heart precursors. At early stages, the patterns of wingless and dpp intersect in a line in every segment where the broad dorsal band of dpp expression intersects the continuous stripes of wingless. Many more cells are exposed to both wg and dpp at this stage than actually become heart, as heart precursors are actually specified only at the dorsal point of this line. In contrast, at later stages, after the patterns of wingless and dpp have transitioned, wingless and dpp do coincide at the point where heart precursors are specified. However, at this stage, there is also an additional ventrolateral point of wingless and dpp intersect in every segment just dorsal to the CNS (Fig. 3A). Thus, the patterns of wingless and dpp can not uniquely specify the region of heart cell specification at either early or late stages. However, the only mesodermal region exposed to both wingless and dpp at both early and late stages occurs in dorsal segmented clusters where the heart precursors normally are specified. C. No Other Known Gene Product Provides Independent Spatial Information

The coincidence of heart precursor specification in the mesoderm exposed to persistent wingless and dpp intersects does not exclude the possibility that other spatial cues are also required for specification of the heart. However, no other known genes appear to provide spatial information required for heart specification within the mesoderm independently of wingless or dpp. Many genes, such as zfh1, D-mef2, ladybird, DER, laminin, and numb are required for differentiation of heart precursors or their progeny but not for the initial specification of the cardiac mesoderm, and they are discussed in Section VI. Others, such as tinman and heartless, a Drosophila fibroblast growth factor (FGF) receptor (Shishido et al., 1993), are required for heart specification and are expressed at late stages in spatially informative patterns. However, both tinman and heartless appear to acquire their specific patterns as a result of the patterns of wingless and dpp (Frasch, 1995; Wu et al., 1995; Carmena et al., 1998), and so provide no independent spatial information within the mesoderm. Similarly, even-skipped (eve), hedgehog, and sloppy-paired are required for heart specification and are expressed in spatially restricted, informative patterns. However, these genes appear to affect heart specification indirectly via their roles in positioning, maintaining, or transmitting the wingless signal in each segment . Thus, although other spatial signals may also be required for the initial localization of heart precursor specification within the mesoderm, it appears that no known gene products do so independently of wingless and dpp.

188

Lockwood et al.

Cardiac Pattern Formation and Cell Fate Determination

189

D. Ectopic Heart Is Specified at Persistent Intersects of wingless and dpp

An important measure of whether a given spatial overlap of signals is required to localize heart precursor specification is to generate new patterns of overlap and assay for ectopic heart specification. Since wingless, dpp, and tinman are expressed not only in the region of heart, but also in other regions of the embryo, this can most easily be achieved by misexpressing one of the required genes in a region in which the other two genes are also present. At early stages, the mesoderm is exposed to a line of wingless and dpp intersect in each segment; heart precursors are specified by the subset of cells at the dorsal edge of this line. At later stages, after the patterns of wingless and dpp have transitioned, there are two points of wingless and dpp intersect per segment; heart is specified at the dorsal point of late wingless and dpp intersection but not at the later ventrolateral wingless and dpp intersect. This ventrolateral intersect differs from the dorsal intersect in that no dpp is present at that intersect at early stages. If heart specification requires both wingless and dpp at both early and late stages, then the misexpression of dpp at that intersect at early stages should result in the ectopic specification of heart precursors in those segmented positions. To test this, we misexpressed dpp uniformly in the mesoderm and assayed for heart markers, and showed that heart is indeed specified in these positions (see Figs. 3B; Fig. 4A,B)

Figure 3 The elaboration of the tinman (tin) expression pattern and specification of the heart. (A) Patterns of wingless (wg), dpp, and tin in wildtype embryos. The initiation of tin expression occurs independently of wg and dpp and, under the control of twist, is expressed uniformly in the trunk mesoderm through stage 9 (not shown). At this stage wg and dpp are expressed in their primary patterns in the overlying ectoderm. Only the tin-expressing cells exposed to dpp during stage 9 continue to express tin during stage 10. This pattern is further restricted during stage 11, when only the tin-expressing cells additionally exposed to wg during stage 9 continue to express tin. However, not all cells exposed to both wg and dpp during stage 9 express tin at stage 11, since tin is not expressed in a stripe in the dorsal mesoderm, but rather in two clusters of cells that fall along the stripe of wg/dpp intersect. By stage 11 the patterns of wg and dpp have transitioned into their secondary expression patterns. Only the dorsal cluster of tin-expressing cells, which both express tin and remain exposed to wg and dpp, continue to maintain tin at stage 12, and go on to form the heart. (B) Patterns of wg, dpp, and tin in embryos in which dpp is mis-expressed. In these embryos all tin-expressing cells are exposed to dpp during stage 9, and continue to express tin during stage 10. At stage 11, when the pattern of tin appears to require both wg and dpp, tin is expressed in continuous stripes. As a result, when the pattern of wg undergoes its normal transition, tin expressing cells are present below the ventrolateral wg/dpp intersect. In addition, as a result of dpp mis-expression this ventrolateral intersect is not only present at stage 11, as usual, but is also present at stage 9. Ectopic heart specification is observed in this region of persistent wg/dpp intersect (Fig. 4).

190

Lockwood et al.

Figure 4 Ectopic specification of cardiac cell fates by TGF-β factors. (A,B) Ectopic expression of tinman (arrowhead in B) in Drosophila as a consequence of mesodermal overxpression of dpp in fly embryos leads to evidence of ectopic heart formation. (C) Ectopic expression of Nkx.25 (around bead in C) as a consequence of bmp2 overxpression. (Micrograph C is courtesy of T. Schultheiss.)

(Yin et al., 1998; W.K.L. and R.B., unpublished data). A broader measure of the ability of the wingless and dpp to specify heart is to ask whether any region of the mesoderm, and not just the ventrolateral intersection of wingless and dpp, is capable of specifying heart ectopically. When wingless and dpp are expressed in combination, uniformly and persistently in the mesoderm, ectopic heart specification is indeed observed broadly throughout the mesoderm, demonstrating that every region of the mesoderm is capable of responding to wingless and dpp by specifying heart. However, even in these embryos, heart specification is not uniform but is distributed randomly throughout the mesoderm (W.K.L. and R.B., unpublished data). E. Ectopic Heart Is Specified in the Ectoderm When tinman Is Misexpressed at Ectodermal Intersects of wingless and dpp

If heart is specified as tinman-expressing cells exposed to both wingless and dpp, then it should be possible to generate ectopic heart not only by misexpressing wingless and dpp in the tinman-expressing mesoderm, but also by misexpressing tinman in wingless- and dpp-expressing regions of other germ layers. To test this,

Cardiac Pattern Formation and Cell Fate Determination

191

tinman was misexpressed at ectodermal wingless and dpp intersects in the embryo which normally generate the imaginal disk primordia for the adult legs and wings (Cohen et al., 1993). The misexpression of tinman in these ectodermal wingless and dpp intersects indeed results in the ectopic specification of cardiac cell fates (W.K.L. and R.B., unpublished data). Therefore, patterned signaling by wingless and dpp provide positional information, and the transcription factor encoded by tinman provides mesodermal context information, which in combination is sufficient to specify and position the Drosophila heart. F. Is wingless- and dpp-Like Signaling also Involved in Vertebrate Cardiogenesis?

The TGF-β homologs of dpp in vertebrates, encoded by Bmp2, Bmp4, and bmp7, are well known to play functionally exchangeable roles in dorsal–ventral patterning (Padgett et al., 1993; Francois and Bier, 1995; Holley et al., 1995). Similar to dpp in the fly, these genes are expressed in endodermal or ectodermal tissues that are closely apposed to Nkx.25 expressing primordia of the cardiac crescent in the anterior lateral plate mesoderm. In the chick, transplantation of Bmp2expressing endoderm to the anterior paraxial mesoderm located medial to the later plate mesoderm or implantation of Bmp2- or Bmp4- laden beads in that tissue can specifically induce Nkx.25 and other cardiac-specific marker gene expression (see Fig. 4C) (Schultheiss et al., 1995, 1997). Moreover, cocultures of anterior paraxial mesoderm in vitro with either BMP2 or BMP4 induces robust cardiac differentiation. In contrast, incubation of precardiac mesoderm with Noggin, an antagonist of BMP signaling, inhibits cardiac myogenesis. Therefore, in vertebrates, as in Drosophila, expression of tinman expression and heart specification seems to depend on localized expression of TGF-β factors. In contrast, it is not yet known whether or not any of the vertebrate Wnt genes play a definitive or direct role in cardiac induction. In conclusion, the detailed characterization of wingless, dpp, and tinman gene functions and their interactions provide a basic genetic framework for understanding how the Drosophila heart is specified and positioned and moreover continue to serve as a prototype model for studying cardiac induction in vertebrates. V.

pannier and Vertebrate GATA Factors

Three members of the GATA family in vertebrates, GATA4/5/6, are expressed in the developing heart and various endoderm-derived tissues in overlapping patterns (Laverriere et al., 1994; Morrisey et al., 1996, 1997). For example, GATA4 expression in the mouse, frog and chick is detected in the early precardiac mesoderm, as Nkx.25, and persists throughout development (Kelley et al., 1993; Heikinheimo et al., 1994). They all appear to be potent activators of cardiac-specific

192

Lockwood et al.

gene transcription (Grepin et al., 1994, 1997; Ip et al., 1994; Jiang and Evans, 1996), particularly in conjunction with Nkx.25 (Durocher and Nemer, 1997, 1998; Lee et al., 1998; Sepulveda et al., 1998). Although manipulation of GATA levels in embryonic stem cells affects cardiogenesis (Grepin et al., 1995, 1997; Jiang et al., 1998), an indispensable cardiac-specific requirement for any of the GATA factors during normal development has not yet been demonstrated (Kuo et al., 1997; Molkentin et al., 1997; Morrisey et al., 1998; Koutsourakis et al., 1999). In Drosophila, three GATA factors have been identified; they are encoded by pannier, serpent, and dGATAc (Abel et al., 1993; Ramain et al., 1993; Lin et al., 1995), but their cardiac expression or role during heart development had not been studied until recently. pannier Is not only expressed in the dorsal ectoderm, as dpp, but apparently also in the region of the cardiac mesoderm (Gajewski et al., 1999). Loss and gain of function experiments indicate that this GATA factor has an essential role in specifying cardiac cell fates in Drosophila. However, it remains to be determined if it acts at the level of induction of the cardiac mesoderm (see Sec. IV) or at the level of differentiation or distinction of individual cardiac cell types (see Sec. VI). VI. Cell Type Diversification: Lineages and Context Although the spatial patterns of wingless, dpp, and tinman appear to limit the cardiogenic region of the embryo to dorsal segmented clusters of the mesoderm, these patterns do not explain the distinction between individual heart-specific cell types, some of which begin to be established already early during cardiac induction. In this section, we discuss some of the mechanisms that generate cellular diversity within the cardiac mesoderm in Drosophila. A. Distinction Between Alternative Cell Fates Within the Cardiac Mesoderm

One mechanism for generating a diversity of cell types is by specifying alternative cell fates as a consequence of asymmetrical cell division. This involves intracellular determinants which are localized asymmetrically in the predivisional cell, and which segregate to only one of the two daughter cells. In the nervous system of Drosophila, the membrane-associated product of the gene numb has been shown to be necessary and sufficient for determining alternative daughter cell fates during asymmetrical cell divisions (Uemura et al., 1989; Rhyu et al., 1994; Brewster and Bodmer, 1995; Spana et al., 1995; Guo et al., 1996; Spana and Doe, 1996; for recent review, see Vervoort et al., 1997; Lu et al., 1998). In the daughter cell that inherits the Numb protein, the function of a transmembrane receptor protein encoded by Notch and a tropomodulin protein encoded by (sanpodo) spdo is

Cardiac Pattern Formation and Cell Fate Determination

193

inhibited, and as a consequence, this cell assumes a fate that is different from its sibling (Guo et al., 1996; Spana and Doe, 1996; Dye et al., 1998; Skeath and Doe, 1998). Some cells in the cardiac mesoderm also divide asymmetrically under the control of numb. One class of progenitors express the homeobox gene eve and gives rise to a subset of pericardial cells, the EPCs (Ruiz Gomez and Bate, 1997; Carmena et al., 1998; Park et al., 1998c). As in the nervous system, asymmetrically segregating Numb protein in Eve progenitor lineages antagonizes Notch and spdo activity and causes the distinction between the two daughter cells, a skeletal muscle founder and an EPC (Park et al., 1998c). Since the Numb pathway specifies alternative cell fates not only of neural, but also of cardiac lineages, it is necessary that appropriate context information is provided for the correct differentiation of the progeny (Fig. 5). Thus, the situation is analogous to the initial specification of the cardiac mesoderm: a universally used mechanism for distinguishing cellular fates, in this case numb-dependant asymmetrical cell division, must include specific clues from the molecular history of that cell (‘‘cardiac context’’) to interpret this distinction correctly (see below). This mechanism is not limited to Drosophila, since numb homologs have been identified also in vertebrates and shown to be localized asymmetrically during ventricular stem cell divisions in the developing brain (Verdi et al., 1996; Zhong et al., 1996,

Figure 5 Model of stepwise specification of cardiac mesoderm and of cardiac-specific cell types. Note that multiple layers of positional and context information are necessary for specification of a single cardiac cell type.

194

Lockwood et al.

1997). As in Drosophila, vertebrate numb is found in most tissues, including the developing heart and muscles. Thus, it is likely that numb-related genes also control cell fate decisions of certain cardiac lineages in vertebrates. B. zfh-1 Provides a Second Layer of Mesodermal Context for Cardiac Cell Type–Specific Differentiation

As the twist and tinman genes encode transcription factors conferring mesoderm specificity to the interpretation of positional information for implementing cardiac fates overall and in the right place, additional competence or context factors are likely to be required for the appropriate specification of individual cell types within the forming heart. A candidate for such a function is encoded by zfh-1, a zinc finger–containing and homeobox-containing gene, which like tinman is expressed throughout the early mesoderm and later in the forming heart (Lai et al., 1991). Within the developing heart, zfh-1 is specifically required for EPC differentiation; that is, EPCs are missing in zfh-1 mutants without affecting formation of the EPC siblings (Su et al., 1999). This failure to differentiate phenotype of zfh-1 mutants is distinct from that of numb loss or gain of function mutants in which one sibling is transformed to assume the same fate of the other (Park et al., 1998c). Interestingly, when zfh-1 and numb are mutant in the same embryo, neither of the two single mutant phenotypes is observed; that is, not only the EPCs, but also their siblings are missing. Thus, these epistasis experiments suggest that zfh-1 is needed for EPC differentiation independently of and in addition to the cell fate decision dictated by the Numb pathway (Su et al., 1999). Consistent with this interpretation is the finding that, in these double mutants, the EPC progenitor cells form normally and apparently divide symmetrically, as in numb single mutants in which twice the normal number of EPCs are produced. However, these prospective EPCs must lack sufficient information, here in the form of zfh-1, to go on differentiating with the EPC-specific cardiac cell type identity and thus presumably die. C. Combinatorial Integration of Position and Context at the Transcriptional Level

In addition to zfh-1, mesodermal eve function itself appears to be needed for correct EPC differentiation, possibly as a direct target of zfh-1 (Su et al., 1999). Since the EPC progenitors are part of the cardiac mesoderm, initiation of eve expression also depends on the convergent functions of dpp, wingless, and tinman at the dorsal edge of the mesoderm (see Sec. IV). Moreover, eve expression is not initiated in all heart progenitors, which is due in part to the inhibition by ladybird homeobox gene expression in heart precursors adjacent to the EPC progenitors (Jagla et al., 1997). This highly cell type–specific eve expression pattern in the mesoderm appears to be integrated on a 400–base pair enhancer element

Cardiac Pattern Formation and Cell Fate Determination

195

of the eve gene (Z. Han, M.-T. S., and R.B., unpublished data). Indeed, this enhancer contains an accumulation of multiple consensus binding sites for the Tinman, Twist, and Ladybird transcription factors as well as consensus response elements for the wingless and dpp signaling pathways, suggesting a direct combinatorial regulation by the convergence of these gene functions on this enhancer. Future studies hopefully will reveal how the genetic logic of generating specificity is encoded and integrated at the transcriptional level.

VII. Summary We propose a model of how the two mechanisms, that is, global specification of (cardiac) competence and cell type–specific differentiation, interface with each other (see Fig. 5). The experimental evidence suggests that specification of a particular cell type (that of EPCs in this case) is a stepwise process, and the sequential elements of specification are integrated on a defined enhancer element. First, twist activates tinman and zfh-1 specifically in the mesoderm. Then, dorsally restricted dpp signaling from the ectoderm maintains tinman expression in the underlying dorsal mesoderm (Frasch, 1995). Persistent exposure to the striped expression of wingless in the dorsal regions also exposed to dpp then maintains tinman expression in the even more restricted pattern and of the presumptive cardiac mesoderm at the dorsal mesodermal edge (Wu et al., 1995; W.K.L. and R.B., unpublished data). All three components seem to be necessary for the subsequent differentiation of cardiac cell types, since ectopic heart-specific gene expression only occurs at the intersection of persistent dpp, wingless, and tinman expression. With the initiation of the EPC progenitor lineage, Numb is localized asymmetrically in the progenitors and segregates to only one of the progeny. In the daughter cell that receives Numb, dorsal skeletal muscle founder differentiation is initiated in cooperation with the activity of the ras-dependent Drosophila epidermal growth factor (EGF) receptor pathway (Su et al., 1999). In the Numbdeficient daughter, in which Numb is inhibited by activity from the Notch pathway, EPC differentiation ensues aided by the transcription factor Zfh-1.

VIII. Conclusions A two-tiered context information system (first tinman then zfh-1) operates in conjunction with the global mechanisms of determining position (dpp and wingless) and of conferring alternative cell fates during asymmetrical lineages (numb) to specify tissue- and position-specific cell fates within the developing embryo. The interaction of these systems is necessary and prerequisite for the correct formation of the heart in Drosophila. The proposed mechanism of cardiac cell

196

Lockwood et al.

specification provides a useful and testable model for cardiogenesis in vertebrates. Acknowledgments We would like to thank Thomas Schultheiss for providing some of the micrographs used in this chapter. We are also much indebted to Krista Golden for preparing the figures. M.L. is supported by a Career Development Award from the American Heart Association. R.B. is an Established Investigator of the American Heart Association. This work is also supported by a grant from NHLBI, National Institutes of Health, to R.B. References Abel T, Michelson AM, Maniatis T. (1999). A Drosophila GATA family member that binds to Adh regulatory sequences is expressed in the developing fat body. Development 119:623–633. Azpiazu N, Frasch M. (1993). tinman and bagpipe: Two homeobox genes that determine cell fates in the dorsal mesoderm of Drosophila. Genes Dev 7:1325–1340. Azpiazu N, Lawrence PA, Vincent JP, Frasch M. (1996). Segmentation and specification of the Drosophila mesoderm. Genes Dev 10:3183–3194. Beiman M, Shilo BZ, Volk T. (1996). Heartless, a Drosophila FGF receptor homolog, is essential for cell migration and establishment of several mesodermal lineages. Genes Dev 10:2993–3002. Benson DW, Sharkey A, Fatkin D, Lang P, Basson CT, McDonough B, Strauss AW, Seidman JG, Seidman CE. (1998). Reduced penetrance, variable expressivity, and genetic heterogeneity of familial atrial septal defects. Circulation 97:2043–2048. Biben C, Harvey RP. (1997). Homeodomain factor Nkx-2–5 controls left/right asymmetric expression of bHLH gene eHAND during murine heart development. Genes Dev 11:1357–1369. Bodmer R. (1993). The gene tinman is required for specification of the heart and visceral muscles in Drosophila. Development 118:719–729. Bodmer R. (1995). Heart development in Drosophila and its relationship to vertebrate systems. TCM 5:21–28. Bodmer R, Frasch M. (1999). Genetic Determination of Drosophilia Heart Development. In: Rosenthal N, Harvey R, eds. Heart Development. San Diego: Academic Press, pp 65–90. Bodmer R, Golden K, Lockwood WK, Ocorr KA, Park M, Su M-T, Venkatesh T. (1997). Heart development in Drosophila. In: Wasserman P, ed. Advances in Developmental Biology. Greenwich, CT.: JAI Press, pp. 201–205. Bodmer R, Jan LY, Jan YN. (1990). A new homeobox-containing gene, msh-2, is transiently expressed early during mesoderm formation of Drosophila. Development 110:661–669.

Cardiac Pattern Formation and Cell Fate Determination

197

Brewster R., Bodmer R. (1995). Origin and specification of type II sensory neurons in Drosophila. Development 121:2923–2936. Carmena A, Gisselbrecht S, Harrison J, Jimenez F, Michelson AM. (1998a). Combinatorial signaling codes for the progressive determination of cell fates in the Drosophila embryonic mesoderm. Genes Dev 12:3910–3922. Carmena A, Murugasu-Oei B, Menon D, Jimenez F, Chia W. (1998b). inscuteable and numb mediate asymmetric muscle progenitor cell divisions during Drosophila myogenesis. Genes Dev 12:304–315. Cohen B, Simcox AA, Cohen SM. (1993). Allocation of the thoracic imaginal primordia in the Drosophila embryo. Development 117:597–608. Durocher D, Nemer M. (1998). Combinatorial interactions regulating cardiac transcription. Dev Genet 22:250–262. Durocher D, Charron F, Warren R, Schwartz RJ, Nemer M. (1997). The cardiac transcription factors Nkx2–5 and GATA-4 are mutual co-factors. EMBO J 16:5687– 5696. Dye CA, Lee J-K, Atkinson RC, Brewster R, Han P-L, Bellen HJ. (1998). The Drosophila sanpodo gene controls sibling cell fate and encodes a tropomodulin homolog, an actin/tropomyosin associated protein. Development 125:1845–1856. Evans SM, Yan W, Murillo MP, Ponce J, Papalopulu N. (1995). tinman, A Drosophila homeobox gene required for heart and visceral mesoderm specification, may be represented by a family of genes in vertebrates: XNkx-2.3, a second vertebrate homologue of tinman. Development 121:3889–3899. Fishman MC, Chien KR. (1997). Fashioning the vertebrate heart: earliest embryonic decisions. Development 124:2099–2117. Fishman MC, Olson EN. (1998). Parsing the heart: genetic modules for organ assembly. Cell 91:153–156. Francois V, Bier E. (1995). Xenopus chordin and Drosophila short gastrulation genes encode homologous proteins functioning in dorsal-ventral axis formation. Cell 80: 19–20. Frasch M. (1995). Induction of visceral and cardiac mesoderm by ectodermal Dpp in the early Drosophila embryo. Nature 374:464–467. Fu Y, Yan W, Mohun TJ, Evans SM. (1998). Vertebrate tinman homologues XNkx2–3 and XNkx2–5 are required for heart formation in a functionally redundant manner. Development 125:4439–4449. Gajewski K, Fossett N, Molkentin JD, Schulz RA. (1999). The zinc finger proteins Pannier and GATA4 function as cardiogenic factors in Drosophila. Development 126: 5679–5688. Gisselbrecht S, Skeath JB, Doe CQ, Michelson AM. (1996). heartless Encodes a fibroblast growth factor receptor (DFR1/DFGF-R2) involved in the directional migration of early mesodermal cells in the Drosophila embryo. Genes Dev 10:3003–3017. Grepin C, Robitaille L, Antay T, Nemer M. (1995). Inhibition of transcription factor GATA-4 expression blocks in vitro cardiac muscle differentiation. Mol Cell Biol 15:4095–4102. Grepin C, Dagnino L, Robitaille LL, Haberstroh LL, Antakly T, Nemer M. (1994). A hormone-encoding gene identifies a pathway for cardiac but not skeletal muscle gene transcription. Mol Cell Biol 14:3115–3129.

198

Lockwood et al.

Grepin C, Nemer G, Nemer M. (1997). Enhanced cardiogenesis in embryonic stem cells overexpressing the GATA-4 transcription factor. Development 124:2387–2395. Grow MW, Krieg PA. (1998). Tinman function is essential for vertebrate heart development: elimination of cardiac differentiation by dominant inhibitory mutants of the tinman-related genes, XNkx2–3 and XNkx2–5. Dev Biol 204:187–196. Guo M, Jan LY, Jan YN. (1996). Control of daughter cell fates during asymmetric cell division: Interaction of Numb and Notch. Neuron 17:27–24. Harvey RP. (1996). NK-2 homeobox genes and heart development. Dev Biol 178:203– 216. Harvey RP, Rosenthal N, eds. (1999). Heart Development. San Diego: Academic Press. Heikinheimo M, Scardrett JM, Wilson DB. (1994). Localization of the transcription factor GATA-4 to regions of the mouse embryo involved in cardiac development. Dev. Biol. 164:361–373. Holley SA, Jackson PD, Sasai Y, Lu B, De Robertis EM, Hoffmann FM, Ferguson EL. (1995). A conserved system for dorsal-ventral patterning in insects and vertebrates involving sog and chordin. Nature 376:249–253. Ip HS, Wilson DB, Heikinheimo M, Tang Z, Ting CN, Simon MC, Leiden JM, Parmacek MS. (1994). The GATA-4 transcription factor transactivates the cardiac musclespecific troponin C promoter-enhancer in non-muscle cells. Mol Cell Biol 14:7517– 7526. Jagla K, Frasch M, Jagla T, Dretzen G, Bellard F, Bellard M. (1997). ladybird, A new component of the cardiogenic pathway in Drosophila required for the diversification of heart precursors. Development 124:3471–3479. Jiang Y, Evans T. (1996). The Xenopus GATA-4/5/6 genes are associated with cardiac specification and can regulate cardiac-specific transcription during embryogenesis. Dev Biol 174:258–270. Jiang Y, Tarzami S, Burch BE, Evans T. (1998). A common role for each of the cGATA4/5/6 genes in the regulation of cardiac morphogenesis. Dev Genetics 22:263– 277. Kelley C, Blumberg H, Zon LI, Evans T. (1993). GATA-4 is a novel transcription factor expressed in endocardium of the developing heart. Development 118:817–827. Komuro I, Izumo S. (1993). Csx: A murine homeobox-containing gene specifically expressed in the developing heart. PNAS 90:8145–8149. Koutsourakis M, Langeveld A, Patient R, Beddington R, Grosveld F. (1999). The transcription factor GATA6 is essential for early extraembryonic development. Development 126:723–732. Kuo CT, Morrisey EE, Anandappa R, Sigrist K, Lu MM, Parmacek MS, Soudais C, Leiden JM. (1997). GATA-4 transcription factor is required for ventral morphogenesis and heart tube formation. Genes & Dev 11:1048–1060. Lai Z, Fortini ME, Rubin GM. (1991). The embryonic expression patterns of zfh-1 and zfh-2, two Drosophila genes encoding novel zinc-finger homeodomain proteins. Mech Dev 34:123–134. Laverriere AC, NacNeill C, Mueller C, Poelmann RE, Burch JBE, Evans T. (1994). GATA-4/5/6: a subfamily of three transcription factors expressed in developing heart and gut. J Biol Chem 69:23177–23184. Lawrence PA, Bodmer R, Vincent J-P. (1995). Segmental patterning of heart precursors in Drosophila. Development 121:4303–4308.

Cardiac Pattern Formation and Cell Fate Determination

199

Lee KH, Xu Q, Breitbart RE. (1996). A new tinman-related gene, nkx2.7, anticipates the expression of nkx2.5 and nkx2.3 in zebrafish heart and pharyngeal endoderm. Dev Biol 180:722–731. Lee Y, Shioi T, Kasahara H, Jobe SM, Wiese RJ, Markham BE, Izumo S. (1998). The cardiac tissue-restricted homeobox protein Csx/Nkx2.5 physically associates with the zinc finger protein GATA4 and cooperatively activates atrial natriuretic factor gene expression. Mol Cell Biol 18:3120–3129. Lin WH, Huang LH, Yeh JY, Hoheisel J, Lehrach H, Sun YH, Tsai SF. (1995). Expression of a Drosophila GATA transcription factor in multiple tissues in developing embryos. J Biol Chem 270:25150–25158. Lints TJ, Parsons LM, Hartley L, Lyons I, Harvey P. (1993). Nkx-2.5: a novel murine homeobox gene expressed in early heart progenitor cells and their myogenic descendants. Development 119:419–431. Lu B, Jan LY, Jan YN. (1998). Asymmetric cell division: lessons from flies and worms. Curr Opin Genet Dev 8:392–399. Lyons I, Parsons LM, Hartley L, Li R, Andrews JE, Robb L, Harvey RP. (1995). Myogenic and morphogenetic defects in the heart tubes of murine embryos lacking the homeo box gene Nkx2–5. Genes Dev 9:1654–1666. Molkentin JD, Lin Q, Duncan SA, Olson EN. (1997). Requirement of the transcription factor GATA-4 for heart tube formation and ventral morphogenesis. Genes & Dev 11:1061–1072. Morrisey EE, Ip HS, Lu MM, Parmacek MS. (1996). GATA-6: A zinc finger transcription factor that is expressed in multiple cell lineages derived from lateral mesoderm. Dev Biol 177:309–322. Morrisey EE, Ip HS, Tang Z, Lu MM, Parmacek MS. (1997). GATA-5: A transcriptional activator expressed in a novel temporally and spatially-restricted pattern during embryonic development. Dev Biol 183:21–36. Morrisey EE, Tang Z, Sigrist K, Lu MM, Jiang F, Ip HS, Parmacek MS. (1998). GATA6 regulates HNF4 and is required for differentiation of visceral endoderm in the mouse embryo. Genes & Dev 12:3579–3590. Newman CS, Krieg PA. (1998). tinman-related genes expressed during heart development in Xenopus. Dev Genet 22:230–238. Olson EN, Srivastava D. (1996). Molecular pathways controlling heart development. Science 272:671–676. Padgett RW, Wozney JM, Gelbart WM. (1993). Human BMP sequences can confer normal dorsal-ventral patterning in the Drosophila embryo. Proc Natl Acad Sci USA 90: 2905–2909. Park M, Wu X, Golden K, Axelrod JD, Bodmer R. (1996). The wingless signaling pathway is directly involved in Drosophila heart development. Dev Biol 177:104–116. Park M, Venkatesh TV, Bodmer R. (1998a). A dual role for the zeste-white3/shaggy– encoded kinase in mesoderm and heart development of Drosophila. Dev Genetics 22:201–211. Park M, Lewis C, Turbay D, Chung A, Chen JN, Evans S, Breitbart RE, Fishman MC, Izumo S, Bodmer R. (1998b). Differential rescue of visceral and cardiac defects in Drosophila by vertebrate tinman-related genes [published erratum appears in Proc Natl Acad Sci USA 1998 Oct 27;95(22):13348]. Proc Natl Acad Sci USA 95:9366– 9371.

200

Lockwood et al.

Park M, Yaich LE, Bodmer R. (1998c). Mesodermal cell fate decisions in Drosophila are under the control of the lineage genes numb, Notch, and sanpodo. Mech Dev 75: 117–126. Ramain P, Heitzler P, Haenlin M, Simpson P. (1993). pannier, A negative regulator of achaete and scute in Drosophila, encodes a zinc finger protein with homolgy to the vertebrate transcription factor GATA-1. Development 119:1277–1291. Ranganayakulu G, Elliott DA, Harvey RP, Olson EN. (1998). Divergent roles for NK-2 class homeobox genes in cardiogenesis in flies and mice. Development 125:3037– 3048. Reecy JM, Yamada M, Cummings K, Schwartz RJ. (1997). Chicken Nkx-2.8: a novel homeobox gene expressed in early heart progenitor cells and the pharyngeal pouch2 and -3 endoderm. Dev Biol 188:295–311. Rhyu MS, Jan LY, Jan YN. (1994). Asymmetric distribution of Numb protein during division of the sensory organ precursor cell confers distinct fates to daughter cells. Cell 76:477–491. Riechmann V, Irion U, Wilson R, Grosskortenhaus R, Leptin M. (1997). Control of cell fates and segmentation in the Drosophila mesoderm. Development 124:2915–2922. Ruiz Gomez M, Bate M. (1997). Segregation of myogenic lineages in Drosophila requires numb. Development 124:4857–4866. St Johnston RD, Nu¨sslein-Volhard C. (1992). The origin of pattern and polarity in the Drosophila embryo. Cell 68:201–219. Schott JJ, Benson DW, Basson CT, Pease W, Silberbach GM, Moak JP, Maron BJ, Seidman CE, Seidman JG. (1998). Congenital heart disease caused by mutations in the transcription factor NKX2–5 [see comments]. Science 281:108–111. Schultheiss TM, Xydas S, Lassar AB. (1995). Induction of avian cardiac myogenesis by anterior endoderm. Development 121:4203–4214. Schultheiss TM, Burch JB, Lassar AB. (1997). A role for bone morphogenetic proteins in the induction of cardiac myogenesis. Genes Dev 11:451–462. Sepulveda JL, Belaguli N, Nigam V, Chen CY, Nemer M, Schwartz RJ. (1998). GATA4 and Nkx-2.5 coactivate Nkx-2 DNA binding targets: Role for regulating early cardiac gene expression. Mol Cell Biol 18:3405–3415. Shishido E, Higashijima S-I, Emori Y, Saigo K. (1993). Two FGF-receptor homologues of Drosophila: one is expressed in mesodermal primordium in early embryos. Development 117:751–761. Skeath J, Doe CQ. (1998). Sanpodo and Notch act in opposition to Numb to distinguish sibling neuron fates in the Drosophila CNS. Development 125:1857–1865. Spana EP, Doe CQ. (1996). Numb antagonizes Notch signalling to specify sibling neuron cell fates. Neuron 17:21–26. Spana EP, Kopczynski C, Goodman CS, Doe CQ. (1995). Asymmetric localization of Numb autonomously determines sibling neuron identity in the Drosophila CNS. Development 121:3489–3494. Staehling-Hampton K, Hoffmann FM, Baylies MK, Rushton E, Bate M. (1994). dpp induces mesodermal gene expression in Drosophila. Nature 372:783–786. Su M-T, Fujioka M, Goto T, Bodmer R. (1999). The Drosophila homeobox genes zfh-1 and evenskipped are required for cardiac-specific differentiation of a numbdependent lineage decision. Development 126:3241–3251.

Cardiac Pattern Formation and Cell Fate Determination

201

Tanaka M, Kasahara H, Bartunkova S, Schinke M, Inagaki H, Lee Y, Lyons GE, Izumo S. (1998). Vertebrate homologs of tinman and bagpipe: roles of the homeobox genes in cardiovascular development. Dev Genet 22:239–249. Tanaka M, Chen Z, Bartunkova S, Yamasaki N, Izumo S. (1999). The cardiac homeobox gene Csx/Nkx2.5 lies genetically upstream of multiple genes essential for heart development. Development 126:1269–1280. Tonissen KF, Drysdale TA, Limts TJ, Harvey RP, Krieg PA. (1994). XNkx2.5: a Xenopus homologue of Nkx-2.5 and tinman. Evidence for a conserved role in cardiac development. Dev Biol 162(1):325–328. Uemura T, Shepherd S, Ackerman L, Jan LY, Jan YN. (1989). numb, A gene required in determination of cell fate during sensory organ formation in Drosophila embryos. Cell 58:349–360. Venkatesh TV, Bodmer R. (1995). How many signals does it take? Bioessays 17:754– 757. Venkatesh TV, Park M, Ocorr KA, Nemaceck J, Golden K, Wemple M, Bodmer R. (2000). Cardiac enhancer activity of the homeobox gene tinman depends on CREB consensus binding sites in Drosophila. Genesis 26:55–66. Verdi JM, Schmandt R, Bashirullah A, Jacob S, Salvino R, Craig CG, Program AE, Lipshitz HD, McGlade CJ. (1996). Mammalian Numb is an evolutionarily conserved signaling adaptor protein that specifies cell fate. Curr Biol 6:1134–1145. Vervoort M, Dambly-Chaudiere C, Ghysen A. (1997). Cell fate determination in Drosophila. Curr Opin Neurol 7:21–28. Wu X, Golden K, Bodmer R. (1995). Heart development in Drosophila requires the segment polarity gene wingless. Dev Biol 169:619–628. Xu X, Yin Z, Hudson JB, Ferguson EL, Frasch M. (1998). Smad proteins act in combination with synergistic and antagonistic regulators to target dpp responses to the Drosophila mesoderm. Genes Dev 12:2354–2370. Yin Z, Frasch M. (1998). Regulation and function of tinman during dorsal mesoderm induction and heart specification in Drosophila. Dev Genet 22:187–200. Yin Z, Xu X-L, Frasch M. (1997). Regulation of the Twist target gene tinman by modular cis-regulatory elements during early mesoderm development. Development 124: 4971–4982. Zhong W, Feder JN, Jiang MM, Jan LY, Jan YN. (1996). Asymmetric localization of a mammalian Numb homolog during mouse cortical neurogenesis. Neuron 17:43– 53. Zhong W, Jiang MM, Weinmaster G, Jan LY, Jan YN. (1997). Differential expression of mammalian Numb, Numblike and Notch1 suggests distinct roles during mouse cortical neurogenesis. Development 124:1887–1897. Zou Y, Evans S, Chen J, Kou H-C, Harvey RP, Chien KR. (1997). CARP, a ankyrin repeat protein, is downstream in the Nkx2–5 homeobox gene pathway. Development 124: 793–804.

9 Cardiac Development in Vertebrates

EMILY C. WALSH and DIDIER Y. STAINIER University of California, San Francisco San Francisco, California

I.

Introduction

In this chapter, we examine the process of vertebrate heart formation through the lens of genetic analysis. The vertebrate heart is initially composed of two cell layers: an outer muscular myocardium and an inner endocardium that connects the heart to the vasculature. Much attention has been focused toward understanding not only the initial induction of these two cell types, but also the subsequent morphogenesis that results in the formation of a multichambered, beating organ. As the construction of the vertebrate heart is an intricate process, genetic analysis has been particularly effective in its study. In fact, some of the most illuminating work in the field of heart development has employed genetic techniques, including forward screens for heart defects in zebrafish and targeted gene disruption in the mouse. One of the most powerful approaches to understanding any biological problem is a genetic screen. For this reason, the zebrafish is an incredibly powerful system for studying heart development. Screening for embryonic lethal mutations in zebrafish is facilitated by its relatively short generation time, external fertilization, and transparent early embryonic development. Not only can one identify novel genes in cardiac development through forward genetic screens in the fish, 203

204

Walsh and Stainier

but one can also, in theory, delineate entire signaling pathways through modifier screens. Furthermore, through cell transplantation experiments, which are easily carried out in zebrafish, one can examine the behavior of wild-type cells in a mutant background (and vice versa). Thus, one can determine in which cells a gene is required. This technique is especially helpful prior to the molecular isolation of a locus defined by a mutation, as it provides insight into the cell autonomy of a particular mutant phenotype. For instance, if wild-type cells are able to differentiate into normal myocardial cells in a mutant background, then the gene that is mutated is required cell autonomously in the myocardial lineage. In comparison, the mouse system allows for whole embryo and tissuespecific gene disruption. Through these gene-targeting techniques, one can study the roles that broadly expressed genes play specifically in the cardiac progenitors. Additionally, the mouse is the only genetic system suitable for studying late steps in heart development like septation, as this process does not occur in the simple, two-chambered zebrafish heart. Although work in these two systems has revealed many cellular and molecular underpinnings to the process of heart development, it has left many questions unanswered. First, we will briefly summarize what is known about the cellular movements underlying heart tube morphogenesis. Next, we will examine the role that genetic analysis has played in further defining these movements and the molecules required. At the end of each section, we will point out future directions for genetic inquiry into the intricate process of heart formation.

II. Heart Morphogenesis A. Gastrulation

Fate mapping studies in several model organisms have defined the heart precursors at early stages in development ((Stainier et al., 1993; Tam et al., 1997; Warga and Nu¨sslein-Volhard, 1999) (Fig. 1). At the onset of gastrulation in zebrafish, the bilateral domains of cardiac precursors are located 90 degrees from the future dorsal side (Stainier et al., 1993). The murine equivalents of these cells are located distal and lateral to the primitive streak as gastrulation begins. What happens next to these fields of cells differs depending on the topography of the particular organism. In zebrafish, the heart precursors involute early during gastrulation and join with the rest of the mesoderm in its convergence toward the developing axis (see Fig. 1). At the end of gastrulation, the myocardial anlagen are located in the anterior lateral plate mesoderm where they begin to show signs of differentiation. The myocardial-specific homeodomain protein, Nkx2.5, is expressed in these cells at this time, along with its more widely expressed homolog, Nkx2.7 (Lee

Cardiac Development in Vertebrates

205

Figure 1 Location of the cardiac precursors at the onset and end of gastrulation. At the onset of gastrulation in the zebrafish, the bilateral domains of cardiac precursors are located 90 degrees from the future dorsal side (top left). The murine equivalents of these cells are located distal and lateral to the node as gastrulation begins (top right). In both organisms, these cells involute early during gastrulation and join the rest of the mesoderm as it converges toward the developing axis. At the end of gastrulation, the bilateral myocardial anlagen are located in the anterior lateral plate mesoderm (bottom left and right).

et al., 1996). Also at this time, gata genes [4–6] are even more broadly expressed throughout much of the anteroposterior extent of the lateral plate mesoderm (Reiter et al., 1999). As mouse gastrulation begins at embryonic day 6.5 (E6.5), the heart precursors also ingress with other mesodermal derivatives into the primitive streak (Tam et al., 1997). They then leave the streak, migrate anterolaterally, and at the end of gastrulation can be found in the anterior lateral plate mesoderm, like their zebrafish counterparts (see Fig. 1). At approximately E7.5, one sees Nkx2.5 ex-

206

Walsh and Stainier

pression in the heart primordia (Lints et al., 1993). Murine gata genes are expressed slightly earlier (E7) throughout the headfold mesoderm that contains the cardiac precursors (Heikinheimo et al., 1994; Morrisey et al., 1997). B. Secondary Convergence and Heart Tube Assembly

The heart precursors are now set to undergo a second convergence from the lateral plate mesoderm toward the midline where they will join to form a single tube. Some of the most interesting mutations that perturb heart development affect this secondary convergence and result in the formation of two physically separate cardiac structures, a phenotype commonly referred to as cardia bifida. What makes these cardia bifida mutations so striking is that many seem to affect the process in a cell nonautonomous fashion. That is to say, these genes are not required in the cardiac mesoderm but rather in another tissue that is somehow necessary for the proper morphogenesis of the heart precursors. Study of these mutations has revealed a critical role for endodermal derivatives in heart fusion and will be discussed at length later. First, we will focus on the movements that elaborate the definitive heart tube. These movements have been extensively documented in zebrafish using both molecular and genetic analysis (Yelon et al., 1999) (Fig. 2). At 19 hr postfertilization (hpf), the two fields of myocardial precursors merge posteriorly to form a horseshoe-shaped structure. The horseshoe transforms into a ring as anterior cells migrate medially to close the circle. Next, the cardiac ring telescopes out to form a tube. The ventricular end of the heart tube assembles first, followed by the atrial end, which subsequently assumes a left-sided position. This leftward bias of heart assembly is the embryo’s first morphological break in bilateral symmetry. At around 19 hpf, the endocardial cells also converge at the midline of the developing zebrafish embryo. There they form a sheet that underlies the myocardial precursors (E. Perens and D. Y. Stainier, unpublished data). As the myocardium telescopes out, the endocardium follows, lining the inside of the developing tube. Subsequently, the endocardium connects seamlessly with the endothelium lining the dorsal aorta. Intriguingly, the endocardium is not required during early cardiac development. Work on the zebrafish cloche mutation, which perturbs endocardial induction, revealed that myocardial differentiation and fusion proceed without consequence (Stainier et al., 1995). However, myocardial function is defective in cloche mutants. Murine myocardial movements are mainly similar. Cardiac fusion begins between E7.5 and E8. As the left and right endoderm sheets undergo ventral folding at the midline, they bring the bilateral myocardial and endocardial precursors into proximity (Kuo et al., 1997). The requirement for this endoderm-

Cardiac Development in Vertebrates

207

Figure 2 Secondary convergence and cardiac assembly. In the zebrafish, secondary convergence begins at 19 hr postfertilization (hpf) (left column, all views dorsal with posterior at the top unless indicated otherwise). The two fields of myocardial precursors merge posteriorly to form a horseshoe-shaped structure. The horseshoe transforms into a ring as anterior cells migrate medially to close the circle. Next, the cardiac ring telescopes out to form a tube. The ventricular end of the heart tube assembles first, followed by the atrial end. Note the overt left–right asymmetry of the zebrafish heart at the end of fusion. By 19 hpf, the endocardial cells have converged at the midline. There they form a sheet that underlies the myocardial precursors. As the myocardium telescopes, the endocardium follows, lining the inside of the developing tube. In the mouse, fusion begins around E7.5 (right column, ventral views, anterior to the top). At this time, ventral folding brings the bilateral myocardial and endocardial precursors in proximity. At the midline, these precursors fuse to form the definitive heart tube. As in the zebrafish, the ventricular end of the murine heart tube assembles first, followed by the atrial end. (Zebrafish panels adapted from Yelon et al. [1999], and unpublished data of E. Perens and D. Y. Stainier, mouse panels adapted from DeRuiter et al. [1992].]

208

Walsh and Stainier

mediated motion in cardiac fusion has been shown in mouse chimeras lacking the gata4 gene in the endoderm, as detailed below (Narita et al., 1997). Once near the midline, the bilateral heart precursors fuse to form the definitive heart tube. Several studies have addressed the exact manner by which the myocardial precursors form a single tube around the endocardium (DeRuiter et al., 1992; Vira´gh et al., 1989). As in zebrafish, the ventricular end of the murine heart tube assembles first, followed by the atrial end (see Fig. 2). Also as in zebrafish, mouse embryos lacking endocardium such as the flk-1 mutant mice assemble the heart tube apparently normally. C. Looping Morphogenesis

During the process of dextral looping, the heart tube undergoes a dramatic threedimensional reorganization. This looping event is conserved among all vertebrate organisms studied to date (Olson and Srivastava, 1996). In vertebrates with lungs, this rearrangement helps to align the outflow tract over the left and right ventricles such that unoxygenated blood is sent to the lungs and oxygenated blood is returned to the body. However, this cannot be the only reason for the conservation of looping among vertebrates as some, like the zebrafish, do not have lungs or four-chambered hearts. Nonetheless, these less complex hearts undergo the same rightward bending. This fact leads to the hypothesis that looping is a prerequisite for some other yet unidentified aspect of heart function. At the end of tube assembly, the vertebrate heart is asymmetrically aligned, with the atrium to the left of the midline (as seen in Fig. 2). At 33 hpf in zebrafish, the heart bends back to the right, bringing the ventricle into a position ventral to the atrium, thus completing looping morphogenesis in the zebrafish embryo. Murine looping is similar. At approximately E9.5, the heart begins a sweeping dextral motion that positions the future right and left ventricles side-by-side and ventral to the atria. Many advances in our understanding of left-right asymmetry and looping morphogenesis have been made in recent years and are covered in Chapter 10. D. Valve Formation

Valves form at three sites along the anteroposterior axis of the developing heart: the outflow tract, atrioventricular boundary (AVB), and sinus venosus. Once formed, these largely acellular structures serve to prevent retrograde blood flow through the heart. However, at early stages, these valve development sites are hotbeds of cell signaling, migration, and proliferation. At the three sites of valve initiation, a conversation occurs between the myocardium and endocardium (Fig. 3). In vitro tissue recombination experiments in the chick show that a subset of endocardial cells at the AVB are competent to respond to a signal provided by the overlying myocardium (reviewed in Eisen-

Cardiac Development in Vertebrates

209

Figure 3 Valve formation. Valves form at multiple locations along the anteroposterior axis of the developing heart (top panel illustrates two such positions, outflow tract and atrioventricular boundary). At these sites, endocardial cells undergo an epithelial to mesenchymal transition and migrate into the extracellular matrix between the myocardium and endocardium (bottom panels A and B). There, the cells form ‘‘endocardial cushions’’ that are later remodeled to create flap-like structures (bottom panel C). The final products of this process are fibrous, largely acellular structures (not shown) that serve to prevent retrograde blood flow through the heart.

berg and Markwald, 1995). In response to this signal, the endocardial cells undergo an endothelial to mesenchymal transition. They then migrate into the extracellular matrix between the endocardium and myocardium (also called the cardiac jelly). Once there, these cells form an endocardial cushion that is later remodeled into a valve. Although we understand the cellular movements of valve formation, we know little of the molecular underpinnings of this process. Nonetheless, there is a growing number of mouse and zebrafish mutations that perturb this process.

210

Walsh and Stainier

Analysis of these mutations should begin to reveal the molecular pathways leading to valve formation. Zebrafish valve induction is initiated at about 36 hpf. At this time, we see the first molecular indications of AVB differentiation: restricted Bmp4 and Notch1b expression (Walsh and Stainier, unpublished data; Westin and Lardelli, 1997). At 48 hpf, we see another sign of molecular differentiation: the expression of the extracellular matrix protein Fibulin-1 (Zhang et al., 1997). It is at this time that the embryo becomes dependent on its valves to prevent retrograde blood flow. A number of zebrafish mutations have been identified that perturb valve formation, but in this volume, we will focus our attention on the best studied of these, the jekyll mutation. In the mouse, valve formation is initiated about E9.5. However, the process is much more complex, as it will lead to the formation of a four-chambered structure rather than a simple two-chambered one. This process is known as septation. E. Septation

Although zebrafish valve morphogenesis is complete after the development of valves at the outflow tract, sinus venosus, and atrioventricular boundary, murine valvulogenesis continues past this point (Fig. 4). Beginning at E9.5, endocardial cushions grow in from opposing sides of the AVB, thereby bisecting this structure. This central AV cushion then experiences a second phase of growth to bisect the atrium partially. Meanwhile, the outflow tract cushions are undergoing growth to contribute to the ventricular septum. However, the bulk of the atrial and ventricular septa are fashioned from fenestrated myocardial outgrowths between the future left and right chambers (Olson and Srivastava, 1996). At E13.5, this complex process is complete. In recent years, many genes have been implicated in this process based on phenotypes of targeted mutations. These genes encode a range of molecules from transcription factors to signaling molecules to ras regulators. Yet, much remains to be understood. It will be particularly interesting to determine the role neural crest cells play in this process, as many of the implicated genes are expressed in both the neural crest and cardiac cells that contribute to the heart’s outflow tract and septal elements. F. Further Morphogenesis

The process of vertebrate heart development is much more complex than simple chamber and valve formation. For instance, a conduction system must be developed to regulate heart excitation. This process has been reviewed recently (Gourdie et al., 1999; Moorman et al., 1998; Moorman et al., 1997) and will not

Cardiac Development in Vertebrates

211

Figure 4 Two views of the process of murine cardiac septation. The top row illustrates the process via a coronal (or longitudinal) section through the ventricle, viewed ventrally. The bottom row shows septation via an oblique cross section through the heart. At E9.5 in the mouse, endocardial cushions grow in from opposing sides of the AVB, thereby bisecting this structure (B and F). This central AV cushion grows to bisect the atrium partially (G). Around the same time, the outflow tract cushions are undergoing growth to contribute to the ventricular septum (D). However, the bulk of the atrial and ventricular septa are fashioned from fenestrated myocardial outgrowths between the future left and right chambers (C, D and G, H).

be addressed here. We will, however, describe recent work in the mouse that has opened the door to some of the genetic underpinnings of two other steps in heart development: trabeculation of the ventricle and epicardium formation. At about E9.5 in the mouse and 4 dpf in the zebrafish, myocardial cells in the ventricle begin to make muscular outgrowths into the cardiac jelly (Fig. 5). These trabeculae are thought to be important for the optimal performance of the ventricle. Interestingly, a number of mouse mutations that perturb septation in the heart also affect the process of ventricular trabeculation. Mutations leading to septation and/or trabeculation phenotypes will be addressed concurrently in this chapter. The process of epicardial development is similarly complex. Beginning at E9, cells migrate from the septum transversum to the heart and then spread along the entire surface of the myocardium in a stereotypic fashion (Hiruma and Hirakow, 1989; Vira´gh and Challice, 1981). Some cells delaminate from the epicar-

212

Walsh and Stainier

Figure 5 Trabeculation and epicardial development. Top panel, H&E–stained section through the ventricle of an embryonic chick heart. Bottom panel, graphic illustration of a similar section. Beginning at E9, mesothelial cells migrate from the septum transversum to the heart and then spread along the entire surface of the myocardium to form the epicardium. Some of these epicardial cells delaminate into the subepicardial space to form the coronary vasculature. The epicardium and subepicardial vasculature are indicated in the bottom panel. At about E9.5 in the mouse and 4 dpf in the zebrafish, myocardial cells in the ventricle begin to make muscular outgrowths into the cardiac jelly. These trabeculae are also indicated in the bottom panel. (Section kindly provided by J. Bristow.)

Cardiac Development in Vertebrates

213

dium into the subepicardial space to form the coronary vasculature (Moore et al., 1999). We will discuss five genes required for the proper formation of the epicardium and coronary vessels. III. Genetic Loci Involved in Cardiac Differentiation Very little is known about the genetic basis of early cardiac differentiation. The zebrafish swirl mutation is the only reported mutation that perturbs myocardial induction from the earliest stages of myocardial marker expression. In contrast, there are three known mutations that result in defects in endocardial induction: two that lead to a lack of endocardium, cloche (zebrafish) and flk-1 (mouse), and one that results in excess endocardial induction, flt-1 (mouse). A. Myocardial Differentiation: swirl

The swirl mutation was originally identified on the basis of its dorsoventral patterning deficits (Mullins et al., 1996). Mutant embryos exhibit expansion of dorsal structures, such as somites and notochord, and reduction of ventral fates, such as blood and pronephros. Loss of swirl function also results in a reduction or absence of myocardial precursors as assessed by nkx2.5 expression (Kishimoto et al., 1997). This is the first report of a single gene mutation that affects myocardial induction in such a profound manner. swirl encodes a zebrafish Bone morphogenetic protein 2 (Bmp2b) ortholog (Kishimoto et al., 1997; Nguyen et al., 1998). Consistent with the swirl phenotype, work in Xenopus implicates Bmps in dorsoventral patterning of the mesoderm and ectoderm (Harland and Gerhart, 1997; Holley and Ferguson, 1997; Sasai and De Robertis, 1997). Additionally, Bmps are thought to be important in cardiogenesis. Bmp2-soaked beads are sufficient to induce myocardial marker expression in chick anterior paraxial mesoderm and application of Noggin, a Bmp antagonist, leads to the loss of myocardial gene expression in precardiac tissue (Schultheiss et al., 1997). Analysis of the swirl mutant further demonstrates that Bmp2 is not only sufficient to induce myocardial fates, it is also necessary, although how directly Bmp2 is involved in this process is unclear. Notably, murine Bmp2 disruption does not have as severe an effect on cardiac development. The murine mutation will be discussed later in the section regarding secondary convergence. B. Endocardial Differentiation: cloche, flk-1, and flt-1

The zebrafish cloche mutation results in defects in blood and endothelial and endocardial development (Stainier et al., 1995). The endocardial defect is cell autonomous, indicating that Cloche is required in the endocardial precursors for

214

Walsh and Stainier

their development. Interestingly, the cloche defect can be rescued by overexpression of the endothelium-expressed transcription factor Scl/Tal1 (Liao et al., 1998). However, cloche is not linked to the scl locus, showing that Cloche is not Scl. Thus, Cloche appears to act upstream of Scl in endothelial and blood development (Liao et al., 1998). Like the cloche mutation, targeted disruption of the murine flk-1 locus results in failure of vasculogenesis, endocardial development, and blood-island formation (Shalaby et al., 1995). flk-1 encodes a tyrosine kinase receptor for the vascular endothelial growth factor (VEGF) protein. Expression of flk-1 can be detected from very early stages of endothelial development. Low levels of this expression, as well as that of other early endothelial markers, is present in homozygous mutant embryos. This result leads to the hypothesis that flk-1 mutant embryos do form early endothelial precursors. However, expression of a later endothelial marker, tie-2, is absent, suggesting that if present, the early precursors do not mature in flk-1 mutant embryos. Targeted disruption of another VEGF receptor, flt-1, causes a very different phenotype (Fong et al., 1995). Loss of flt-1 does not prevent endothelial cell differentiation, rather this loss causes a lack of organization of the developing vascular channels. Additionally, one does not see an absence of endocardium in flt-1 mutant embryos. Instead, the endocardium over proliferates, and like the rest of the vasculature, fails to organize into a tubular endothelium. One model for the disparity between the loss of function phenotypes for these two receptors is that they have different affinities for VEGF in vivo. As such, they may be activated at different levels of the ligand and signal for distinct cellular responses. C. Future Directions

Early myocardial development has been a particularly difficult process to study. So far, mutations in genes hypothesized to act early in cardiac differentiation (e.g., see Nkx2.5, mef2c, gata4, and faust in Sec. IV.A and B and V.A and C, respectively) have not resulted in a complete loss of myocardial tissue as is the case with the tinman mutation in the fly (see below). Given this observation, single-gene mutational approaches to analysis of heart induction may not be entirely revealing. Double-mutant analysis may reveal interactions between genes in this process; however direct screening for defects in early myocardial differentiation could be the most rewarding approach. Forward genetic in situ hybridization screens for early cardiac markers should allow relatively quick identification of loci that are essential for myocardial differentiation. This type of approach is wholly feasible in the zebrafish (Alexander et al., 1998). Furthermore, the pioneering work of Hopkins (Amsterdam and Hopkins, 1999) and her associates in the field of viral insertional mutagenesis should help to speed the cloning of insertional mutations that affect these processes.

Cardiac Development in Vertebrates

215

Endocardial differentiation is also not well understood. Particularly, there is no report of a gene required specifically in endocardial induction and not overall endothelial differentiation. The three genes presented here are required for both processes. Based on analysis of these mutations, one model is that endocardial cells are simply endothelial cells that lie in close apposition to the myocardium. However, after the endocardium assembles into a tube, endocardial cells do express genes not expressed by the rest of the endothelium [notch1b, for instance (Westin and Lardelli, 1997)]. This observation raises many questions about the differences between endocardial and noncardiac endothelial cell fate. Are the cardiac and noncardiac endothelial lineages related? Experimentally speaking, can a single labeled cell give rise to cells in both tissues? If so, at what point do the endocardial cells know they are different from other endothelial cells; that is, when is the endocardial fate specified? It will also be interesting to understand how similar cardiac and noncardiac endothelia are with respect to the cellular mechanisms underlying tissue organization. For instance, during assembly, are the endocardial cells responding to the same cues as their vessel-making counterparts? Forward genetics in the zebrafish and detailed lineage analysis constitute one approach to look into these unresolved issues. IV. Loci Required for Chamber Differentiation Proper ventricular and atrial morphogenesis is essential for the effective functioning of the heart. Our knowledge of the genes involved in the process of chamber differentiation is still in its infancy. In the past 4 years, a number of zebrafish and mouse mutations that cause chamber-specific defects have been identified. Some of these reduce the number of cells in a single chamber, whereas others seem to perturb chamber differentiation. A. myocyte enhancer factor 2c (mef2c)

Mef2c is a transcription factor that directly regulates cardiac and skeletal muscle genes (Olson et al., 1995). Expression of mef2c in the myocardial lineages begins while the heart precursors are located in bilateral fields (Lin et al., 1997). This expression persists uniformly throughout the definitive heart tube. Loss of mef2c results in a multitude of cardiac defects. Mutant embryos do not form a right ventricle as judged by heart morphology and defects in dhand expression (Lin et al., 1997). mef2c ⫺/⫺ hearts do not loop and show defects in ventricular trabeculation and endocardial organization. In addition, although the atrioventricular boundary is demarcated in mef2c-negative hearts, endocardial cushions do not form there. mef2c is also required in other tissues in the developing mouse embryo. Noncardiac defects in mef2c mutant embryos include a lack of smooth muscle

216

Walsh and Stainier

cell differentiation and endothelial disorganization, which is perhaps secondary to the smooth muscle defects. B. nkx2.5

Two groups have created different targeted disruptions of the NK domain transcription factor gene, nkx2.5. Lyons et al. (1995) report a complete absence of looping and a drastic reduction of mlc-2v expression in nkx2.5 mutant hearts. Meanwhile Tanaka et al. (1999) report that mlc-2v is only partially reduced and looping is initiated normally. What the two groups agree upon is that loss of nkx2.5 perturbs ventricular differentiation, particularly trabeculation. Also, expression of heart markers such as atrial natriuretic factor (anf) and brain natriuretic peptide (bnp) is more strongly perturbed in the ventricle of mutant embryos than in the atrium (Tanaka et al., 1999). Additionally, as in the mef2c mutant, there is an absence of chamber septation and atrioventricular boundary/endocardial cushion formation. A priori, one might have predicted a much more severe effect from a null mutation in nkx2.5. nkx2.5 is one of the earliest myocardial specific markers in the developing embryo and loss of the nkx2.5 homolog, tinman, in Drosophila melanogaster results in the complete absence of a heart. Yet, the murine heart does develop to the point of looping. One model to explain the reduced severity of the mouse mutant phenotype as compared to the tinman phenotype is the functional overlap of Nkx2.5 with other existing vertebrate tinman homologs. C. dhand and ehand

The bHLH transcription factors, eHand and dHand, are required for proper ventricular development. The corresponding genes were originally isolated by a number of groups focusing on the role of bHLH proteins in development (Cross et al., 1995; Hollenberg et al., 1995; Srivastava et al., 1995) dhand (also known as thing2/hed/hand2) and ehand (thing1/hxt/hand1) are coexpressed throughout the prefusion cardiac crescent at E7.5 in the mouse (Srivastava et al., 1995). Later, these two genes are expressed in a complementary fashion. At E9.5, dhand expression is restricted to the conotruncus and right ventricle, whereas ehand is expressed in the conotruncus and left ventricle. Loss of dhand in the mouse results in a loss of right ventricular tissue and death at E10.5 (Srivastava et al., 1997). Trabeculation of the remaining ventricle is reduced; however, cardiac jelly deposition and endocardial cushion formation proceed normally. In addition to the heart phenotype, dhand mutant embryos show defects in aortic arch development and branchial arch morphogenesis (other sites of dhand expression). Interestingly, a mutation in the zebrafish dhand ortholog has been recently identified. Initial results suggest that loss of dhand in the zebrafish results in an

Cardiac Development in Vertebrates

217

earlier and more severe failure of myocardial differentiation (D. Yelon and D. Y. Stainier, unpublished data). Analysis of this mutation is ongoing and may provide insight into an earlier role for dhand in heart development. Two groups have performed targeted gene disruption of the ehand locus. Riley et al. (1998) report that ehand-null embryos arrest from trophoblast defects at E7.5 just as heart fusion is about to begin. However, wild-type tetraploid extraembryonic tissue can rescue the trophoblast defects. Such tetraploid-rescued embryos survive to E10.5. These embryos have unlooped hearts that appear to lack ventricular differentiation. Ventricular trabeculation and atrioventricular boundary formation fail to occur. Moreover, ehand-negative hearts lack myosin light chain-2v (mlc-2v) expression, another indicator of ventricular maturation. However, other markers of cardiac fate like gata4, nkx2.5, mef2c, and myosin light chain-2a are expressed normally. The second group, Firulli et al. (1998) showed that loss of eHand function results in incomplete atrial fusion, failure of looping morphogenesis, and a reduction in dhand-expressing tissue. In contrast to Riley et al. (1998), they showed that mlc-2v is expressed, suggesting that ventricular chamber specification has occurred normally. Further analysis will be required to resolve these two disparate findings and determine the exact role that ehand plays in heart development. D. pandora and lonely atrium

Originally identified in two large-scale genetic screens, neither of these two zebrafish mutations, pandora and lonely atrium, that perturb ventricular development has been cloned. Among the many defects in pandora mutant embryos are the reduction of the ventricle to a small stub and the dilation of the atrium (Stainier et al., 1996). On the other hand, lonely atrium mutant hearts seem to lack a ventricle entirely, whereas the rest of the embryo remains unaffected (Chen et al., 1996; this mutation unfortunately appears to have been lost). The cloning of pandora should add to our knowledge of ventricular differentiation. E. coup-tfII and raldh2

Chick ovalbumin upstream promoter-transcription factor II is an orphan nuclear receptor expressed in the atrial myocardium and sinus venosus of the developing heart. Consistent with this expression pattern, loss of coup-tfII results in a lack of atrial differentiation at E9.5 (Pereira et al., 1999). Other aspects of heart development, like ventricular trabeculation, seem normal in most mutant embryos. Recent results suggest that the retinoic acid synthesis enzyme, RALDH2, is also required for atrial development (Niederreither et al., 1999). Loss of raldh2 results in a dilated cardiac tube that fails to loop and appears ventricular in nature. Molecular analysis of this phenotype has not yet been completed but will be required to confirm the absence of atrial tissue.

218

Walsh and Stainier F. Future Directions

One intriguing aspect of the mef2c and nkx2.5 mutations is that they perturb the development of a single chamber, yet are not expressed in a chamber-restricted fashion. Both mef2c and nkx2.5 show expression in both the ventricle and atrium of the developing heart. Further analysis of these mutations may reveal partnerships between these chamber nonspecific factors and other chamber-specific proteins during chamber development. Double-mutant analysis may help to elucidate these partnerships. For instance, does a mef2c ⫺/⫺; dhand ⫺/⫺ embryo have more severe defects in ventricular development than either single mutant alone? This analysis may also help to uncover functional overlap in the process of ventricular differentiation. The coup-tfII and raldh2 mutant phenotypes are striking because very few reported mutations result in atrium-specific heart defects. This observation remains to be explained. One model is that the atrial fate is the default state of the heart. Therefore, any mutation affecting the atrium would affect the whole heart. Genetic screens in the zebrafish lend support to this model. In a recent screen for chamber-specific defects, 20 mutations were identified as being ventricle defective, whereas only a single mutation affected the atrium (Alexander et al., 1998). Hopefully, the cloning of that atrium-defective mutation and the further analysis of all mutations in this class will deepen our understanding of atrial development.

V.

Genes Involved in Secondary Convergence

Post-gastrulation, the bilateral heart primordia come to rest beside the midline in the anterior lateral plate. However, their trip is far from over. One of the fastest growing classes of mutant phenotypes seen in both mouse and zebrafish studies affects this next step in heart formation. These mutants lack midline migration, which results in the formation of two bilateral hearts. As such, they are referred to as cardia bifida mutants. Cardia bifida mutants can be split into three classes. The first class exhibits primary defects in endodermal development, whereas the second class is heart autonomous. The third class is less straightforward. These mutations are either uncloned or correspond to genes with expression in both heart and endoderm for which cell autonomy studies have not been completed. In this section, we will also address the murine bmp2 and smad5 mutations. These two mutations lead to ventral folding defects similar to other mutations in this class. However, bmp2-null and smad5-null mice exhibit some degree of

Cardiac Development in Vertebrates

219

heart tube fusion. As such they may prove to be helpful in further analysis of the tissue requirements for cardiac fusion. A. Endodermal Defects

gata4

gata4 encodes a zinc finger transcription factor thought to act in cardiac specification and yolk sac differentiation. Two groups have reported the targeted disruption of gata4 in the mouse (Kuo et al., 1997; Molkentin et al., 1997). Both groups note cardia bifida in gata4-null embryos, as well as severe defects in ventral folding morphogenesis. As gata4 is expressed in both precardiac mesoderm and extraembryonic endoderm, it was unclear where it was required for proper heart tube fusion. Narita et al. (1997) examined the autonomy of the heart phenotype through chimera analysis. They reported that wild-type visceral and/or anterior definitive endoderm restored cardiac fusion in gata4 ⫺/⫺ embryos, thereby providing evidence for the role of endodermal tissues in heart tube fusion. one-eyed pinhead and casanova

The one-eyed pinhead and casanova mutations were identified in large-scale screens for developmental defects in the zebrafish (Chen et al., 1996; Hammerschmidt et al., 1996). The zebrafish one-eyed pinhead gene encodes a member of the Cripto/FRL-1/Cryptic family of epidermal growth factor (EGF) repeat– containing proteins (Zhang et al., 1998). As an extracellular, membrane-attached protein, Oep is required in a cell-autonomous fashion for Nodal signal transduction (Gritsman et al., 1999; Stra¨hle et al., 1997). Loss of maternal and zygotic function of this protein results in severe defects in mesendodermal involution and induction. In such so-called MZoep mutants, only tail mesoderm is formed. However, loss of zygotic oep alone results in a less drastic phenotype. Zoep ⫺/⫺ embryos show defects in heart fusion, endoderm formation, floorplate induction, and prechordal plate formation. The cell autonomy of the oep heart defect is unknown. Endoderm is missing in these mutants, yet oep is required in mesoderm derivatives as demonstrated clearly by the MZoep phenotype. Further analysis is needed to determine whether Oep is required in the cardiac mesoderm during heart fusion. However, work from Peyrie´ras et al. (1998) has begun to answer this question. Injection of an activated version of taramA (an orphan type I receptor for a transforming growth factor-β [TGF-β] superfamily ligand) autonomously rescues endoderm formation in oep mutants (Peyrie´ras et al., 1998). Rescuing endoderm formation in this fashion also rescues cardiac fusion, suggesting that the heart defect in oep is nonautonomous.

220

Walsh and Stainier

casanova mutant embryos also show cardia bifida. Examination of the mutant phenotype revealed an additional defect in endodermal development as judged by axial, sox17, gata5, fkd2, and fkd7 expression analysis (Alexander and Stainier, 1999). The gene responsible for the casanova mutant phenotype remains elusive. However, based on cell transplantation experiments, the cas gene product is required cell autonomously for endoderm development (Alexander and Stainier, 1999). The cell autonomy of the cardiac defect is unknown, but cloning of the cas locus should aid efforts in this regard. Given the apparent ability of endoderm to rescue heart fusion in oep mutants, it is likely that the endoderm defect is also primary in cas mutants. B. Heart/Mesodermal Defects

mesP1

mesP1 encodes a basic helix–loop–helix protein expressed in a subset of mesodermal cells at the onset of gastrulation. Lineage analysis using the Cre-LoxP recombination system revealed that these mesP1-expressing cells contribute to the myocardium later in development, as well as to extraembryonic mesoderm (Saga et al., 1999). Loss of mesP1 results in a range of heart phenotypes from a single heart tube to complete cardia bifida (Saga, 1998; Saga et al., 1999). In the extreme case of bifid hearts, the two resulting hearts display disorganized myocardium and endocardium. These phenotypes suggest a role for mesP1 in heart patterning in addition to migration of the bilateral heart fields, although this disorganization could result from delayed fusion. C. Undetermined or Dual Cell Autonomy

fibronectin, furin, and hrs

fibronectin-null mice have variable defects in heart tube fusion (George et al., 1993; 1997). This widely expressed gene is also required in notochord, somite, yolk sac, and amnion, as well as embryonic and extraembryonic vascular development. As it is so broadly required, tissue-specific knockouts will be critical for a more complete understanding of the role of fibronectin in heart fusion. furin encodes a serine protease expressed in extraembryonic endoderm and mesoderm, as well as in precardiac mesoderm at E7.5 (Roebroek et al., 1998). By E8.5, its expression is localized to the definitive heart tube, notochord, definitive gut tube, and lateral plate mesoderm. Loss of furin results in severe defects in ventral folding as well as cardia bifida. As in the case of the fibronectin-null

Cardiac Development in Vertebrates

221

phenotype, furin’s broad expression makes determination of autonomy important. One aspect of the furin mutant phenotype raises the possibility that this gene is required in the heart precursors themselves. furin ⫺/⫺ mice have fewer myocardial precursors that express lowered levels of nkx2.5, suggesting some degree of heart autonomous activity by the furin gene product. In addition to the heart phenotype, other defects in furin-negative mice include a kinked neural tube and defects in the yolk sac vasculature. Also, furinnegative embryos do not complete chorioallantoic fusion. This defect is present in the hrs mutant discussed next and the vcam-1 and α4-integrin mutants that will be discussed in Section X.A. hrs encodes a ubiquitously expressed, FYVE double-zinc finger domain protein believed to act in ligand-induced receptor endocytosis (Komada and Soriano, 1999). In addition to cardia bifida, hrs ⫺/⫺ mice do not form a foregut, lack fusion of the chorion and allantois, and have wavy neural tubes and head folds that fail to close. One phenotype that sets hrs mutants apart from other mutants discussed in this section is the health of the bifid hearts. Unlike furin- and gata4-negative hearts, hrs mutant hearts were always beating when isolated. This indicates that Hrs may play less of a role in myocyte differentiation than Furin or GATA4, although further molecular analysis will be important to test this hypothesis. Taken with the observation that hrs-null embryos have high levels of apoptosis in definitive endoderm, it is likely that Hrs activity is primarily required in the endodermal lineage. faust

faust is the zebrafish gata5 ortholog (Reiter et al., 1999). During gastrulation, faust/gata5 is expressed in both mesodermal and endodermal precursors. Later its expression is restricted to the lateral plate mesoderm from which the heart precursors arise. Like murine furin mutants, faust ⫺/⫺ embryos have fewer myocardial precursors (Reiter et al., 1999). Moreover, the heart anlage express lower levels of myocardial markers like nkx2.5, cardiac myosin light chain genes 1 and 2, tropomyosin, cardiac troponin T, and ventricular myosin heavy chain 1. In addition to the heart defects, gut tube morphogenesis is disrupted in faust/gata5 mutant embryos (as assessed by the expression of the endodermal markers fkd2 and axial). As faust mutant embryos show dramatic defects in both cardiac and endodermal differentiation, and as faust is expressed in the precursors of both tissues, it is likely to be autonomously required in both tissues.

222

Walsh and Stainier bonnie and clyde, miles apart, and natter

Several other zebrafish mutations cause cardia bifida; namely, bonnie and clyde, miles apart, and natter (Chen et al., 1996; Stainier et al., 1996). Work is ongoing to clone these genes and determine whether they function in the endoderm, the precardiac mesoderm, or both. D. Other Ventral Folding Defects

bone morphogenetic protein 2 (bmp2) and smad5

Murine mutations in these two Bmp signal pathway members also display defects in ventral folding and heart assembly. Both bmp2- and smad5-negative hearts, although at least partially fused, remain exocoelomic (Chang et al., 1999; Zhang and Bradley, 1996). Further analysis will be required to understand the relationship between these two genes and the other genes discussed in this section. Such work may provide a deeper understanding of the role of ventral folding morphogenesis in cardiac fusion in mouse. Interestingly, mutations in both bmp2 and smad5 have been identified in zebrafish. The zebrafish bmp2 mutation, swirl, was discussed earlier; swirl mutant embryos have reduced or absent nkx2.5 expression. Whether the smad5 mutation, somitabun, has a heart phenotype has not yet been determined. What has been ascertained is that the swirl and somitabun loci interact genetically (Hild et al., 1999). This fact, taken with the proposed role of Smad5 as a Bmp-specific downstream effector (Suzuki et al., 1997), suggests that the similarity of the phenotypes of the targeted mouse mutations is not coincidental. E. Future Directions

In the case of cardia bifida mutations, we now know much more about the molecules involved than the cellular interactions underlying the process. It remains to be determined exactly how the endoderm is involved in the process of cardiac fusion. Does the endoderm provide signals or substrates for migration? Or is it the motile force behind the fusion? Furthermore, do the bmp2 and smad5 mutations tell us more about the actual requirement for ventral folding in cardiac fusion? Why do the hearts in these mutants fuse while other ventral morphogenesis mutants prevent heart tube formation? Investigation is ongoing into the identity of the remaining uncloned zebrafish cardia bifida mutations. This work should provide us with insight not only into endoderm formation, but also heart autonomous factors in this process. As the cardia bifida phenotype is one of the most common identified in zebrafish screens, it is clear that there are numerous genes involved in this complex event.

Cardiac Development in Vertebrates

223

VI. Loci Required in Tube Assembly Very little is known about the complex process by which the bilateral heart primordia join to form the definitive heart tube. Until recently, we have been quite limited in our access to this problem. Before 1997, there were no reported mutations that perturbed heart assembly. At that time, Radice et al. reported that Ncadherin-null mice exhibited defects in the postassembly integrity of the myocardium. It is only now, from forward genetic screening in zebrafish, that a mutation has been identified, heart-and-soul, that truly affects heart tube assembly (Stainier et al., 1996).

A. heart-and-soul

The heart-and-soul (has) mutation was identified in the Boston-based screen for embryonic mutations. heart-and-soul mutants properly initiate heart tube fusion. However, has mutant heart development halts as wild-type hearts are undergoing DV to AP tilting and elongation (S. A. Horne and D. Y. Stainier, unpublished data) (see Fig. 3). Eventually the preatrial tissue folds back to cover the ventricle, giving the heart an inside-out appearance. Additional defects in has-negative embryos include dorsal body curvature and disruption of the retinal pigmented epithelium. The heart-and-soul gene has not yet been isolated.

B. N-cadherin

After initial heart assembly is complete, mutants for this calcium-dependent cell– cell adhesion molecule lose myocardial integrity (Radice et al., 1997). This results in a discontinuous myocardium in N-cadherin-null embryos. Interestingly, Ncadherin ⫺/⫺ myocytes continue to aggregate loosely outside the endocardium, suggesting that there is a residual cell–cell adhesion pathway at work in these cells. Moreover, in vitro culturing of N-cadherin-null myocytes reveals that they are competent to form beating tissue. As such, the defect in these mutant hearts seems to affect specifically the continued integrity of the myocardial tissue after heart assembly.

C. Future Directions

The identification of the heart-and-soul locus should provide a valuable entry point for the molecular study of heart tube assembly. However, this work will be greatly enhanced by further screening in zebrafish to identify additional mutations that perturb the assembly process.

224

Walsh and Stainier VII. Genes Implicated in Looping

Much work has been devoted to understanding left–right axis designation in vertebrates. In recent years, we have seen several exciting inroads made toward the understanding of left–right patterning as well as of heart autonomous components in the interpretation of left–right pattern. However, because of the sheer volume of this recent work exploring left–right asymmetry, this topic will be covered in chapter 10. VIII. Loci Involved in Valve Formation A. Murine Atrioventricular Valve Mutants

Four mouse mutations have been reported to abrogate valve development at the atrioventricular boundary (AVB). Mutations in these four genes, versican, has2, bmp4, and vinculin, provide insight into the process of valve formation as each allows unique access to the problem. versican (cspg2)

The hdf (heart defect) mutation described by Yamamura et al. (1997) was identified as a recessive lethal insertion of a lacZ containing transgene. Later cloning efforts by Mjaatvedt et al. (1998) showed that the hdf locus encodes the chondroitin sulfate proteoglycan, Versican (Cspg2). Mice homozygous for the insertion die around day E10.5. This timing is comparable to that of other atrioventricular valve mutants. Since the transgene that caused the mutation contained the lacZ gene, expression analysis of the locus was possible prior to cloning. Xgal staining was found throughout the anteroposterior axis of the heart, with heaviest staining being in the ventricles and outflow tract. This expression is consistent with the finding that conotruncal and right ventricular development is also perturbed in hdf mutants. In addition to the failed ventricular differentiation, hdf mutants lack proliferation of matrix and cellular invasion at the AVB. Surprisingly, hdf mutant endocardial cells were competent to invade in an in vitro system (Yamamura et al., 1997). This result suggests that some component of the collagen gel matrix in the in vitro system was able to complement the loss of versican. hyaluronic acid synthase 2 (has2)

Hyaluronic acid (HA) is a high molecular weight glucosaminoglycan polymer synthesized by a family of membrane-integral proteins. As a component of the extracellular matrix, HA binds water and creates a voluminous, highly structured

Cardiac Development in Vertebrates

225

gel. In addition to this architectural role, HA is also implicated in cell adhesion and migration (Camenisch et al., 1999). Has2 is the earliest acting of the three mammalian HA synthases (Has1– 3) (Camenisch et al., 1999). Loss of the has2 gene results in severe defects in cardiac development that closely resemble those of the hdf mice discussed above (Camenisch et al., 1999). has2-null mice lack cardiac jelly production, ventricular trabeculation, and endocardial cushion formation. has2-negative hearts also have severe reductions in outflow tract and right ventricular tissue. However, in contrast to the hdf phenotype, has2 mutant AVB explants cannot invade a collagen matrix in an in vitro assay (Camenisch et al., 1999). This defect is rescued by transfection of the wild-type gene into the explants or by the addition of exogenous HA to the culture medium/collagen gel matrix. This observation suggests that the has2-negative endocardial cells at the atrioventricular boundary are competent to invade the cardiac jelly but require HA for migration to occur. vinculin

Inactivation of the vinculin locus by targeted gene disruption results in embryonic death around E10 (Xu et al., 1998). Similar to other valve mutations, blood accumulates in the vinculin mutant heart, as there is no valve to prevent retrograde blood flow. Additionally, vinculin mutants are 30–40% smaller than similarly staged siblings and have grossly attenuated neural development (presumably in part because of vinculin’s role in other tissues). vinculin mutants differ from both versican and has2 mutants by the relative abundance of extracellular matrix deposited at the vinculin mutant atrioventricular boundary. Based on vinculin’s role at focal adhesion plaques, which allow cellular interaction with extracellular matrix during migration, it seems likely that the mutant endocardial cells are simply unable to undergo migration into the nascent cushions. bone morphogenetic protein 4 (bmp4)

Like its paralog bmp2 discussed earlier, bmp4 is a member of the TGF-β superfamily of signaling molecules. Based on its expression pattern, Bmp4 could act as the proposed myocardial signal initiating endocardial migration (Jones et al., 1991; E. C. Walsh and D. Y. Stainier, unpublished data). Bmp4 is initially expressed throughout the myocardium. Its expression then becomes restricted to the myocardium overlying sites of valve formation before the initiation of endocardial migration. Depending on genetic background, very few bmp4 mutant embryos survive to day E10.5 (Winnier et al., 1995). Many homozygous mutant embryos show drastic retardation starting at the egg cylinder stage, and most seem to be resorbed sometime between days E9.5 and E11.5. Those few that do progress past the egg

226

Walsh and Stainier

cylinder stage have severe defects in endocardial cushion formation. Although matrix seems to be deposited at normal levels, the cushions remain acellular. However, this observation warrants further investigation. The authors point out that it is unclear whether these heart defects result from a direct requirement of Bmp4 in the heart, or whether these embryos are simply delayed and unhealthy. Unquestionably a tissue-specific knockout of the bmp4 gene is called for here, as the authors suggest. This will be critical for ascertaining whether Bmp4 is the myocardial signal for endocardial to mesenchymal transition, or whether it behaves in a more complicated fashion at the valve. B. Future Directions for Mouse Studies

Although the existing analysis of these mouse mutations has afforded a molecular peek into the process of valve formation, many questions remain. For instance, does loss of versican reduce the expression of other matrix components and how? Does loss of has2 affect the matrix in the same way? Why are versican ⫺/⫺ cushion explants rescued in a collagen gel matrix assay? Why are has2 mutant explants not similarly rescued? Are these matrix molecules required for signal presentation, as a migrational thoroughfare, or both? There are also several unanswered details with regard to the vinculin and bmp4 mutant phenotypes. Specifically, is vinculin required within the endocardial cells that migrate into the matrix as our hypothesis predicts? Is Bmp4 truly required for valve formation? If so, does Bmp4 signaling in an autocrine or paracrine fashion at the valve site? Even when all these queries are addressed, one overarching question remains: What is the hierarchical relationship between vinculin, Bmp4, versican, and Hyaluronic Acid at the developing valve? Careful examination of the ‘‘molecular epistasis’’ in the relevant mutant mice will certainly provide insight into the above question. Specifically, expression analysis in each of the four mutant backgrounds with each of the other required genes should shed light on the process. For instance, if bmp4 expression were perturbed in versican mutant embryos, one would surmise that bmp4 lies downstream of versican in the molecular pathway leading to endocardial cushion formation. C. Zebrafish Valve Mutants

jekyll

Studies in the zebrafish have also resulted in the identification of mutations that disrupt valve development. One of these is the jekyll mutation. jekyll was identified in the Boston-based screen for recessive lethal zebrafish mutations both for its heart phenotype (Stainier et al., 1996) and its complete lack of Alcian staining in its developing branchial arches and fin buds (Neuhauss et al., 1996).

Cardiac Development in Vertebrates

227

Like one of the classes of mouse phenotype described above, jekyll mutants do not elaborate a healthy extracellular matrix. Efforts to clone jekyll are ongoing and when complete should provide further insight into the molecular basis of valve development. IX. Loci Involved in Cardiac Septation and Further Morphogenesis The next steps in mouse cardiac development are the most intricate: septation of the ventricle and atrium, trabeculation of the ventricle, and outflow tract development. One reason for this complexity is the involvement of neural crest cells in one of these events, the development of the outflow tract. Recently, M. L. Kirby reviewed the contribution of neural crest cells to the heart and great vessels (in Harvey and Rosenthal, 1999). As such we will attempt to limit our discussion of these later events to heart intrinsic components of the process and/or point out where confusion about autonomy persists. A. Ventricular Septal Defects

RXRα and transcription enhancer factor-1 (tef-1)

The RXRα gene encodes a 9-cis retinoic acid X receptor that signals either as a homodimer or as a heterodimer with RAR-type retinoic acid receptors (Gruber et al., 1996). Targeted disruption of the RXRα locus leads to embryonic lethality of homozygous mutants between E13.5 and E17.5 (Gruber et al., 1996; Kastner et al., 1994; Sucov et al., 1994). The main defect in RXRα homozygotes is a reduction of ventricular wall mass. This primary defect is thought to result in the secondary hypoplasia of the muscular component of the interventricular septum that leads to a ventricular septal defect. Additionally, Gruber et al. also report disorganized ventricular trabeculae in 67% of homozygous mutants. Recent studies of the cell autonomy of the RXRα mutant heart phenotype suggest that the receptor is not required in the ventricular myocardium but rather in some neighboring tissue that is necessary for ventricular development (Chen et al., 1998; Tran and Sucov, 1998). Other heart anomalies present in RXRα mutant embryos may reveal a retinoid requirement in the cardiac neural crest. Two groups observed defects in outflow tract and conotruncal development in a majority of mutant hearts (Gruber et al., 1996; Kastner et al., 1994). These are both areas where cardiac neural crest cells are thought to play important morphogenetic roles. Further work is required to determine the autonomy of these defects, as there does not seem to be widespread perturbation of neural crest development in RXRα-null embryos (Gruber et al., 1996). One should note, however, that Sucov et al. (1994) do not report such a high incidence of these defects.

228

Walsh and Stainier

Like the RXRα mutation, loss of the transcription factor tef-1 results in ventricular defects. tef-1–null embryos die at about E11.5 and 12.5 of development with a reduction in ventricular wall proliferation, fewer trabeculae, and a severe defect in the formation of the interventricular septum (Chen et al., 1994). However, most other noncardiac tissues appeared to be normal by histological examination. The only exception was a dilation of the fourth brain ventricle, which the authors hypothesize to be a secondary effect of the heart defect. tolloid-like–1 (tll1)

tll1 encodes a BMP-1–like metalloprotease that has been shown to cleave the BMP inhibitor Chordin (Blader et al., 1997; Piccolo et al., 1997). tll1-negative hearts do not complete interventricular septum formation (Clark et al., 1999). As tll1 is expressed heavily in the interventricular septum it is likely that this defect is heart autonomous. Atrial septal defects were also found in 12 of 16 observed mutants. Based on tll1 expression in the atrial endocardium, this defect is likely autonomous to that tissue as well. Additionally, tll1 mutant embryos exhibit perturbations in outflow tract positioning, namely double outlet right ventricle and double inlet left ventricle defects, which may result from incomplete looping morphogenesis. NF-ATc

NF-ATc is a calcineurin-regulated transcription factor. When activated by dephosphorylation, it is able to translocate to the nucleus and drive transcription. Immunohistochemical analysis reveals NF-ATc expression in the endocardium and thymus of the developing embryo (Ranger et al., 1998). Expression of NFATc can be seen throughout the anteroposterior extent of the endocardium. However, nuclear staining of NF-ATc is seen only in the endocardial cells that overlie the developing cushions (de la Pompa et al., 1998; Ranger et al., 1998). This observation suggests that NF-ATc is only active at sites of valve initiation. Two groups, De la Pompa et al. and Ranger et al., have recently reported targeted disruptions of the NF-ATc locus. While both report cardiac defects as the cause of death of NF-ATc-deficient embryos at E14.5, the two groups report different primary defects in homozygous mutant hearts. De la Pompa et al. report that NF-ATc mutants do not form mitral and tricuspid valves at the atrioventricular boundary. Endocardial cushions do form, but valve flaps are never elaborated. Mutants also lack a completed ventricular septum. Additionally, De la Pompa et al. describe ventricular chamber size reduction, ventricular wall hypertrophy, and underdeveloped semilunar valves in some mutant embryos. Ranger et al. also report the presence of ventricular septal defects in their NF-ATc-mutant embryos. However, Ranger et al. do not describe any defects in valve flap articulation at the AVB and outflow tract. For that matter, they also

Cardiac Development in Vertebrates

229

do not report ventricular wall hypertrophy. While it is difficult to resolve the differences in phenotype reported by these investigators, the data are nonetheless intriguing and illustrate an important role of NF-ATc in ventricular septal development. Transforming Growth Factor– β2 (TGF-β2)

TGF-β2 is a widely expressed gene. Hence, TGF-β2–deficient mice have multiple developmental defects affecting many different tissues. These tissues include the heart, lung, craniofacial cartilage, ear, limb, urogenital tract, and the skeleton, all of which are areas of epithelial to mesenchymal communication. TGF-β2 is a member of the TGF-β superfamily of secreted signaling molecules. Many members of this protein family have been implicated in epithelial to mesenchymal transition and signaling events, not unlike those that occur in valve initiation and elaboration. In fact, much work has focused on the roles of TGF-β2 and TGF-β3 in the process of endocardial cushion formation (Boyer et al., 1999a, 1999b; Brown et al., 1996, 1999; Nakajima et al., 1998; Ramsdell and Markwald, 1997). However, the targeted disruption of TGF-β2 marks the first genetic data describing TGF-β involvement in this process. Like the NF-ATc–null mice discussed above, loss of TGF-β2 results in a ventricular septal defect in a majority of mutants analyzed (94%) (Sanford et al., 1997). Additional cardiac defects noted include double-outlet right ventricle (19% penetrance) and double-inlet left ventricle (25% penetrance). Other phenotypes noted were a reduced level of trabeculation in null animals, atrial septal defects, and enlarged ventricular chambers, none of which appears to be fully penetrant. FK506 Binding Protein 12 (FKBP12)

FKBP12, or FK506 binding protein 12, is a cis-trans prolyl isomerase originally identified by its ability to bind to the immunosuppressant FK506 (a calcineurin inhibitor). FKBP12 has been suggested to act in a number of signaling pathways, including those of TGF-β superfamily members and the inositol 1,4,5-trisphosphate receptor (IP3R). Little is known about the exact role of FKBP12 in the TGF-β pathway; however, it has been shown that although FKBP12 is bound to IP3R, it can bind the phosphatase calcineurin (Cameron et al., 1997). Based on this result, one model is that FKBP12 acts to link the calcineurin phosphatase with other protein contacts, thereby regulating the target’s phosphorylation state. FKBP12 is expressed at high levels throughout the myocardium, endocardium, and septal elements of the developing heart. The majority of FKBP12mutant mice die between E14.5 and birth with heart defects, including ventricular septal defects and thinned ventricular walls (Shou et al., 1998).

230

Walsh and Stainier Future Directions

The analysis of these six mutations suggests many interesting lines of further investigation. Particularly engaging are the NF-ATc–null, TGF-β–null, and FKBP12-null phenotypes. It is possible that the resemblance of these three phenotypes is coincidence; however, interconnections between these molecules do exist, and therefore further analysis could prove illuminating. De la Pompa et al. (1998) show that FK506 treatment of embryos, which inhibits calcineurin activity, prevents the nuclear localization of NF-ATc in endocardial cells. As mentioned above, FKBP12 was identified by its ability to bind FK506. In addition, FKBP12 has also been shown to bind type I TGF-β receptors, although these results are somewhat controversial (Charng et al., 1996; Chen et al., 1997; Wang et al., 1996). As FKBP12 seems to interact with both calcineurin and TGF-β receptors, it could represent a linchpin that connects the two signaling pathways, thus explaining the similarity between the NF-ATc–negative, TGF-β–negative, and FKBP12-negative embryos. Conversely, it is possible that the similarities between the null phenotypes are simply superficial. What is needed to differentiate between these two models is further molecular analysis of these phenotypes. This analysis should be directed to answer questions such as: Does NF-ATc undergo nuclear localization in FKBP12-null and TGF-β2-null hearts? Likewise, it could be informative to attempt rescue of each of the mutations with the other putative pathway members (i.e., can loss of FKBP12 be rescued by expression of a constitutively nuclear NF-ATc construct in the ventricular septum?). B. Trabeculation Defects

neuregulin and Its Receptors erbB2, erbB3 and erbB4

Neuregulin is an EGF family member that is thought to signal through ErbB2/ B3 or ErbB2/B4 tyrosine kinase receptor heterodimers (Meyer and Birchmeier, 1995). In contrast to the first four mutant phenotypes discussed in this section, neuregulin, erbB2, erbB3, and erbB4 mutations do not result in ventricular septal defects. Loss of neuregulin, erbB2, or erbB4 results in ventricular trabeculation defects and hypoplastic endocardial cushions at E10.5 (Gassmann et al., 1995; Lee et al., 1995; Meyer and Birchmeier, 1995). erbB3-negative hearts display only mild defects in trabeculation; however, endocardial cushion development is affected (Erickson et al., 1997). The erbB3 ⫺/⫺ cushions have reduced mesenchyme and are thinner than wild-type cushions at E9.5. The differences between these phenotypes may be explained by the observed expression patterns of neuregulin, erbB2, erbB3, and erbB4. The neuregulin gene is expressed by the endocardium (Erickson et al., 1997; Meyer and Birchmeier, 1995). From there, it is thought to provide a signal required for

Cardiac Development in Vertebrates

231

trabeculation to the myocardium, where erbB2 and erbB4 are expressed (Gassmann et al., 1995; Lee et al., 1996). Neuregulin is also believed to promote the development of the endocardial cushion where erbB3 is expressed (presumably through an ErbB2/B3 complex [Erickson et al., 1997]). Future Directions

Future exploration of the neuregulin pathway in ventricular trabeculation should focus on both upstream and downstream events in this process. For instance, how do chamber specification signals feed into the neuregulin pathway? Do they simply modulate the level of neuregulin expressed so that it is higher in the ventricle where trabeculation occurs? Also, what are the differences between the molecular and cellular cues imparted by ErbB2/B4 signaling versus ErbB2/B3? The observation that these four mutations lead to endocardial cushion defects should be investigated further. Particularly interesting are the potential interactions between neuregulin signaling and hyaluronic acid activity suggested by Camenisch et al. (1999) and reviewed in Section VIII on the has2-null phenotype. C. Cushion Overgrowth

Neurofibromatosis 1 (Nf1)

The Nf1 gene encodes a ras regulator and is expressed highly in the endocardial cushions of the developing heart beginning at E11.5 (Lakkis and Epstein, 1998), as well as, at a lower level, in the myocardium (Huynh et al., 1994; Lakkis and Epstein, 1998). This expression continues through E13.5 when the first defects in valvulogenesis are apparent. Nf1-mutant hearts exhibit an overabundance of endocardial cushion tissue at the atrioventricular boundary and a lack of mature valve leaflet formation. In addition, mutant hearts exhibit double-outlet right ventricles and variable myocardial compact layer thinning. Initially thought to represent a neural crest–based defect in valve development, the Nf1-null phenotype now appears to have a heart-intrinsic cause (Lakkis and Epstein, 1998). Based on BrdU, TUNEL, and in vitro culture analyses, it appears that the reasons for this cushion overgrowth are threefold. First, there is increased proliferation of mesenchymal cells in the cushions. Second, there is reduced cell death in cushion mesenchyme. Last, Nf1⫺/⫺ AVB explants invade collagen gel matrix at much higher rates than the wild type. This invasion is blocked by transfection with a dominant-negative ras construct and mimicked by transfection with a constitutively active ras. This effect is observed in tissue cultures from E10.5 at a time before cardiac neural crest cells reach the heart. As such, this experiment reveals a heart-autonomous requirement for Nf1 in endocardial cushion development.

232

Walsh and Stainier Future Directions

Although the cushion overgrowth defect is heart autonomous, the presence of double-outlet right ventricles may indicate a second requirement for Nf1 in cardiac neural crest. Tissue-specific targeting of the Nf1 gene should resolve this issue. X.

Loci Involved in Epicardial Development

The process of epicardium formation requires many distinct cellular behaviors. Proepicardial cells migrate to the heart, and cover the myocardium in a process that likely requires cell–cell adhesion modulation. A subset of epicardial cells then delaminates, enters the subepicardial space, and forms blood vessels. In 1995, with the creation of α4-integrin and vcam-1 knockout mice, we caught a first glimpse of the genetic elements required in epicardial development. In 1999, three other mutations were shown to perturb this intricate process. A. Mutations α4-integrin and vascular cell adhesion molecule-1 (vcam-1)

The similarity between the these two knockout phenotypes is probably more than coincidence, as α4-integrin and VCAM-1 (Vascular Cell Adhesion Molecule-1) have been suggested to interact in skeletal myogenesis and leukocyte/endothelial adhesion. Their complementary expression patterns in the heart are also suggestive of some kind of interaction (α4-integrin is expressed in the epicardium, whereas vcam-1 is expressed by the underlying myocardium). Loss of α4-integrin results in defects in the maintenance of this epicardial tissue and a lack of coronary vessel formation in addition to defects in allantois/ chorion fusion (Yang et al., 1995). As observed at E10.5, the epicardium is initially formed in the α4-integrin mutant. This structure, however, is not maintained. At E11.5, there is no epicardial layer visible via sectioning of the heart and coronary vessels are also absent. vcam-1 encodes a cell adhesion molecule originally isolated for its ability to mediate the adhesion of white blood cells to umbilical blood vessels. Similar to the α4-integrin mutation, loss of vcam-1 leads to the failure of epicardial development, subepicardial coronary vessel formation, and allantois/chorion fusion (Kwee et al., 1995). One interesting observation is that epicardial development is perturbed at a much earlier stage in vcam-1–deficient animals than in α4integrin mutants. Although α4-integrin seems to be involved in maintaining the epicardium after its migration, vcam-1 is required for that initial movement around the heart. The phenotypic differences between mutations in these two putative partners are not understood.

Cardiac Development in Vertebrates

233

The epicardial and coronary vessel defects are not the only heart defects found in vcam-1–null embryos. Additionally, vcam-1 ⫺/⫺ hearts display ventricular septal defects and reduction of the ventricular compact layer. These additional defects in the vcam-1–null heart may be explained by the relatively broad expression of vcam-1 compared to α4-integrin. Whereas α4-integrin is exclusively expressed in the epicardium, vcam-1 is expressed not only in the underlying myocardium, but also in the interventricular septum, where it may be interacting with another integrin family member to foster ventricular septal development. Wilms’ tumor (wt-1)

Originally identified as a frequently deleted locus in Wilms’ tumor patients, the molecular function of this gene is largely unknown. Loss of the wt-1 gene, which is expressed in the epicardium and underlying precoronary vessel mesenchyme, leads to disruption of epicardial integrity and failure of coronary vasculogenesis (Moore et al., 1999). The epicardial disruption is variable along the anteroposterior axis of the heart, with most severe defects being noted over the ventricle. Interestingly, the defects in the subepicardial coronary vessels were not similarly variable. All subepicardial vessel formation was perturbed, not just that overlying the ventricle. Other defects in wt-1 mutant mice include hypoplastic development of the adrenal glands, spleen, and kidney. erythropoietin and erythropoietin receptor

The last two mutations in this class, erythropoietin (epo) and erythropoietin receptor (epor) represent another matched pair. Erythropoietin and its receptor were originally identified as playing a role in megakaryocyte erythropoiesis. Surprisingly, loss of either the erythropoietin cytokine or its receptor results in multiple cardiac defects. Specifically, mutant mice have defective epicardial layers, defects in coronary vasculogenesis, and a reduction in the proliferation of the ventricular compact zone (Wu et al., 1999). Additionally, epo-null and epor-null mice are reported to exhibit interventricular septal defects; however, the nature of these perturbations is not discussed. epor is expressed in the epicardium, endocardium, and endocardial cushions of the developing heart. Based on ventricular compact zone defects in epornull mice, one might have predicted that epor expression would be also found in the myocardium. However, chimera studies suggest that epor is required cell nonautonomously in the ventricle during compact zone proliferation (Wu et al., 1999). This result suggests that EPOR signaling, in the endocardium or epicardium, triggers the release of a second signal required for myocardial proliferation. In contrast, the epicardial defect in epo mutant and epor mutant mice may be cell autonomous, as epor is expressed in the epicardium. Compared to the

234

Walsh and Stainier

wild type, at E13.5, the epo or epor ⫺/⫺ epicardium is not tightly apposed to the myocardium. Although the epicardium forms in mutant hearts, it appears to have fewer cells than the wild type. Perhaps secondary to the epicardial defect, epo- and epor-negative hearts show gross disorganization of the subepicardial coronary vessels. Chimeric analysis did not address these aspects of the mutant phenotype. B. Future Directions

One intriguing observation is that many of these mutations that perturb epicardial development also have effects on ventricular differentiation. Loss of vcam-1, epo, or epor results in defects in the ventricular compact zone and interventricular septum. Also, wt-1 mutants have thinned ventricular myocardium. The cell autonomy of these defects is not yet known. One attractive model is that these genes are required primarily for epicardial development and that the ventricular defects are secondary to the loss of epicardium. The absence of reported ventricular defects in α4-integrin mutant hearts seems to dispute this model. However, the epicardium does form initially in this mutant and is perhaps present long enough to foster ventricular differentiation. The observation certainly warrants further investigation into the autonomy of these defects, as well as close examination of the epicardial layer in other mutants with thinned myocardium and ventricular septal defects; for example, the RXRα mutant. It will also be important to explore the ‘‘molecular epistasis’’ between WT1 activity, EPOR signaling, and Vcam-1/α4-integrin interaction. Is wt-1 required for epor or α4-integrin expression? Or is it instead downstream of one or both of these genes? This line of investigation should help to elaborate our understanding of the molecular hierarchy underlying epicardial development. Future investigation should also be directed toward understanding the differential requirement of vcam-1 and α4-integrin in epicardial development. Is there another integrin that interacts with Vcam-1 during the epicardium’s initial migration? Moreover, is there another integrin involved in interventricular septal development? Expression analysis of other integrin family members may provide clues to these questions. Finally, none of these mutations seems to affect the induction of epicardial or coronary vessel cell fates. Perhaps future investigation will reveal the molecular pathway that leads to the initial specification of these tissues before they migrate from the septum transversum to the heart. XI. Conclusions In this chapter, we have presented 46 mutations in zebrafish or mouse genes that lead to defects in cardiac development. Although this number represents

Cardiac Development in Vertebrates

235

an impressive first step in our understanding of the genes involved in cardiac development, it is important to note that this work represents the tip of a much larger iceberg. In some of the steps of cardiac development described here, there is only one mutation that leads to disruption of wild-type processes. Furthermore, in no case is an entire signaling pathway revealed in a specific morphogenetic event. These two observations indicate that none of the steps leading to heart formation has been studied to genetic saturation.

Acknowledgments We thank members of our laboratory, especially Sally Horne, Holly Field, Ann Wehman, and Erik Kupperman, for comments on the manuscript, John McDonald for sharing unpublished data, and Jim Bristow for help with Figure 5.

References Alexander J, Stainier DY (1999). Curr Biol 9:1147–1157. Alexander J, Stainier DY, Yelon D (1998). Dev Genet 22:288–299. Amsterdam A, Hopkins N (1999). Methods Cell Biol 60:87–98. Blader P, Rastegar S, Fischer N, Stra¨hle U (1997). Science 278:1937–1940. Boyer AS, Ayerinskas II, Vincent EB, McKinney LA, Weeks DL, Runyan RB (1999a). Dev Biol 208:530–545. Boyer AS, Erickson CP, Runyan RB (1999b). Dev Dyn 214:81–91. Brown CB, Boyer AS, Runyan RB, Barnett JV (1996). Dev Biol 174:248–257. Brown CB, Boyer AS, Runyan RB, Barnett JV (1999). Science 283:2080–2082. Camenisch TD, Spicer AP, Brehm-Gibson T, Augustine ML, Kubalak S, Klewer SE, McDonald JA (2000). J.C.I. Cameron AM, Nucifora FC, Jr, Fung ET, Livingston DJ, Aldape RA, Ross CA, Snyder SH (1997). Jo Biol Chem 272:27582–27588. Chang H, Huylebroeck D, Verschueren K, Guo Q, Matzuk MM, Zwijsen A (1999). Development 126:1631–1642. Charng MJ, Kinnunen P, Hawker J, Brand T, Schneider MD (1996). Jo Biol Chem 271: 22941–22944. Chen J, Kubalak SW, Chien KR (1998). Development 125:1943–1949. Chen JN, Haffter P, Odenthal J, Vogelsang E, Brand M, van Eeden FJ, Furutani-Seiki M, Granato M, Hammerschmidt M, Heisenberg CP, Jiang YJ, Kane DA, Kelsh RN, Mullins MC, Nu¨sslein-Volhard C (1996). Development 123:293–302. Chen YG, Liu F, Massagu J (1997). EMBO J 16:3866–3876. Chen Z, Friedrich GA, Soriano P (1994). Genes Dev 8:2293–2301. Clark TG, Conway SJ, Scott IC, Labosky PA, Winnier G, Bundy J, Hogan BL, Greenspan DS (1999). Development 126:2631–2642.

236

Walsh and Stainier

Cross JC, Flannery ML, Blanar MA, Steingrimsson E, Jenkins NA, Copeland NG, Rutter WJ, Werb Z (1995). Development 121:2513–2523. de la Pompa JL, Timmerman LA, Takimoto H, Yoshida H, Elia AJ, Samper E, Potter J, Wakeham A, Marengere L, Langille BL, Crabtree GR, Mak TW (1998). Nature 392:182–186. DeRuiter MC, Poelmann RE, VanderPlas-de Vries I, Mentink MM, Gittenberger-de Groot AC (1992). Anat Embryol 185:461–473. Eisenberg LM, Markwald RR (1995). Circ Res 77:1–6. Erickson SL, O’Shea KS, Ghaboosi N, Loverro L, Frantz G, Bauer M, Lu LH, Moore MW (1997). Development 124:4999–5011. Firulli AB, McFadden DG, Lin Q, Srivastava D, Olson EN (1998). Nat Genet 18:266– 70. Fong GH, Rossant J, Gertsenstein M, Breitman ML (1995). Nature 376:66–70. Gassmann M, Casagranda F, Orioli D, Simon H, Lai C, Klein R, Lemke G (1995). Nature 378:390–394. George EL, Baldwin HS, Hynes RO (1997). Blood 90:3073–3081. George EL, Georges-Labouesse EN, Patel-King R S, Rayburn H, Hynes RO (1993). Development 119:1079–1091. Gourdie RG, Kubalak S, Mikawa T (1999). Trends Cardiovasc Med 9:18–26. Gritsman K, Zhang J, Cheng S, Heckscher E, Talbot WS, Schier AF (1999). Cell 97:121– 132. Gruber PJ, Kubalak SW, Pexieder T, Sucov HM, Evans RM, Chien KR (1996). Clin Invest 98:1332–1343. Hammerschmidt M, Pelegri F, Mullins MC, Kane DA, Brand M, van Eeden FJ, FurutaniSeiki M, Granato M, Haffter P, Heisenberg CP, Jiang YJ, Kelsh RN, Odenthal J, Warga RM, Nu¨sslein-Volhard C (1996). Development 123:143–151. Harland R, Gerhart J (1997). Annu Rev Cell Dev Biol 13:611–667. Harvey RP, Rosenthal N (1999). Heart Development, San Diego: Academic Press, pp 179–189, 530. Heikinheimo M, Scandrett JM, Wilson DB (1994). Dev Biol 164:361–373. Hild M, Dick A, Rauch GJ, Meier A, Bouwmeester T, Haffter P, Hammerschmidt M (1999). Development 126:2149–2159. Hiruma T, Hirakow R (1989). Am J Anat 184:129–138. Hollenberg SM, Sternglanz R, Cheng PF, Weintraub H (1995). Mol Cell Biol 15:3813– 3822. Holley SA, Ferguson EL (1997). Bioessays 19:281–284. Huynh DP, Nechiporuk T, Pulst SM (1994). Dev Biol 161:538–551. Jones CM, Lyons KM, Hogan BL (1991). Development 111:531–542. Kastner P, Grondona JM, Mark M, Gansmuller A, LeMeur M, Decimo D, Vonesch JL, Dolle´ P, Chambon P (1994). Cell 78:987–1003. Kishimoto Y, Lee KH, Zon L, Hammerschmidt M, Schulte-Merker S (1997). Development 124:4457–4466. Komada M, Soriano P (1999). Genes Dev 13:1475–1485. Kuo CT, Morrisey EE, Anandappa R, Sigrist K, Lu MM, Parmacek MS, Soudais C, Leiden JM (1997). Genes Dev 11:1048–1060. Kwee L, Baldwin HS, Shen HM, Stewart CL, Buck C, Buck CA, Labow MA (1995). Development 121:489–503.

Cardiac Development in Vertebrates

237

Lakkis MM, Epstein JA (1998). Development 125:4359–4367. Lee KF, Simon H, Chen H, Bates B, Hung MC, Hauser C (1995). Nature 378:394–398. Lee KH, Xu Q, Breitbart RE (1996). Dev Biol 180:722–731. Liao EC, Paw BH, Oates AC, Pratt SJ, Postlethwait JH, Zon LI (1998). Genes Dev 12: 621–6. Lin Q, Schwarz J, Bucana C, Olson EN (1997). Science 276:1404–1407. Lints TJ, Parsons LM, Hartley L, Lyons I, Harvey RP (1993). Development 119:419– 431. Lyons I, Parsons LM, Hartley L, Li R, Andrews JE, Robb L, Harvey RP (1995). Genes Dev 9:1654–1666. Meyer D, Birchmeier C (1995). Nature 378:386–90. Mjaatvedt CH, Yamamura H, Capehart AA, Turner D, Markwald RR (1998). Dev Biol 202:56–66. Molkentin JD, Lin Q, Duncan SA, Olson EN (1997). Genes Dev 11:1061–1072. Moore AW, McInnes L, Kreidberg J, Hastie ND, Schedl A (1999). Development 126: 1845–1857. Moorman AF, de Jong F, Denyn MM, Lamers WH (1998). Circ Res 82:629–644. Moorman A F, de Jong F, Lamers WH (1997). Pacing Clin Electrophysiol 20:2087–2092. Morrisey EE, Ip HS, Tang Z, Lu MM, Parmacek MS (1997). Dev Biol 183:21–36. Mullins MC, Hammerschmidt M, Kane DA, Odenthal J, Brand M, van Eeden FJ, FurutaniSeiki M, Granato M, Haffter P, Heisenberg CP, Jiang YJ, Kelsh RN, Nu¨ssleinVolhard C (1996). Development 123:81–93. Nakajima Y, Yamagishi T, Nakamura H, Markwald RR, Krug EL (1998). Ann NY Acad Sci 857:272–275. Narita N, Bielinska M, Wilson DB (1997). Dev Biol 189:270–274. Neuhauss SC, Solnica-Krezel L, Schier AF, Zwartkruis F, Stemple DL, Malicki J, Abdelilah S, Stainier DY, Driever W (1996). Development 123:357–367. Nguyen VH, Schmid B, Trout J, Connors SA, Ekker M, Mullins MC (1998). Dev Biol 199:93–110. Niederreither K, Subbarayan V, Dolle´ P, Chambon P (1999). Nat Genet 21:444–448. Olson EN, Perry M, Schulz RA (1995). Dev Biol 172:2–14. Olson EN, Srivastava D (1996). Science 272:671–676. Pereira FA, Qiu Y, Zhou G, Tsai MJ, Tsai SY (1999). Genes Dev 13:1037–49. Peyrie´ras N, Stra¨hle U, Rosa F (1998). Curr Biol 8:783–786. Piccolo S, Agius E, Lu B, Goodman S, Dale L, De Robertis EM (1997). Cell 91:407– 416. Radice GL, Rayburn H, Matsunami H, Knudsen K A, Takeichi M, Hynes RO (1997). Dev Biol 181:64–78. Ramsdell AF, Markwald RR (1997). Dev Biol 188:64–74. Ranger AM, Grusby MJ, Hodge MR, Gravallese EM, de la Brousse FC, Hoey T, Mickanin C, Baldwin HS, Glimcher LH (1998). Nature 392:186–90. Reiter JF, Alexander J, Rodaway A, Yelon D, Patient R, Holder N, Stainier DY (1999). Genes Dev 13:2983–2995. Riley P, Anson-Cartwright L, Cross JC (1998). Nat Genet 18:271–275. Roebroek AJ, Umans L, Pauli IG, Robertson EJ, van Leuven F, Van de Ven WJ, Constam DB (1998). Development 125:4863–4876. Saga Y (1998). Mech Dev 75:53–66.

238

Walsh and Stainier

Saga Y, Miyagawa-Tomita S, Takagi A, Kitajima S, Miyazaki J, Inoue T (1999). Development 126:3437–3447. Sanford LP, Ormsby I, Gittenberger-de Groot AC, Sariola H, Friedman R, Boivin GP, Cardell EL, Doetschman T (1997). Development 124:2659–2670. Sasai Y, De Robertis EM (1997). Dev Biol 182:5–20. Schultheiss TM, Burch JB, Lassar AB (1997). Genes Dev 11:451–462. Shalaby F, Rossant J, Yamaguchi TP, Gertsenstein M, Wu XF, Breitman ML, Schuh AC (1995). Nature 376:62–66. Shou W, Aghdasi B, Armstrong DL, Guo Q, Bao S, Charng MJ, Mathews LM, Schneider MD, Hamilton SL, Matzuk MM (1998). Nature 391:489–492. Srivastava D, Cserjesi P, Olson EN (1995). Science 270:1995–1999. Srivastava D, Thomas T, Lin Q, Kirby ML, Brown D, Olson EN (1997). Nat Genet 16: 154–160. Stainier DY, Fouquet B, Chen JN, Warren KS, Weinstein BM, Meiler SE, Mohideen MA, Neuhauss SC, Solnica-Krezel L, Schier AF, Zwartkruis F, Stemple DL, Malicki J, Driever W, Fishman MC (1996). Development 123:285–92. Stainier DY, Lee RK, Fishman MC (1993). Development 119:31–40. Stainier DY, Weinstein BM, Detrich HW, 3rd, Zon LI, Fishman MC (1995). Development 121:3141–3150. Stra¨hle U, Jesuthasan S, Blader P, Garcia-Villalba P, Hatta K, Ingham PW (1997). Genes Func 1:131–148. Sucov HM, Dyson E, Gumeringer CL, Price J, Chien KR, Evans RM (1994). Genes Dev 8:1007–1018. Suzuki A, Chang C, Yingling JM, Wang XF, Hemmati-Brivanlou A (1997). Dev Biol 184:402–405. Tam PP, Parameswaran M, Kinder SJ, Weinberger RP (1997). Development 124:1631– 1642. Tanaka M, Chen Z, Bartunkova S, Yamasaki N, Izumo S (1999). Development 126:1269– 1280. Tran CM, Sucov HM (1998). Development 125:1951–1956. Vira´gh S, Challice CE (1981). Anat Rec 201:157–168. Vira´gh S, Szabo´ E, Challice CE (1989). J Mol Cell Cardiol 21:123–37. Wang T, Li BY, Danielson PD, Shah PC, Rockwell S, Lechleider RJ, Martin J, Manganaro T, Donahoe PK (1996). Cell 86: 435–444. Warga RM, Nu¨sslein-Volhard C (1999). Development 126:827–838. Westin J, Lardelli M (1997). Dev Genes Evol 207:51–63. Winnier G, Blessing M, Labosky PA, Hogan BL (1995). Genes Dev 9:2105–2116. Wu H, Lee SH, Gao J, Liu X, Iruela-Arispe ML (1999). Development 126:3597–3605. Xu W, Baribault H, Adamson ED (1998). Development 125:327–337. Yamamura H, Zhang M, Markwald RR, Mjaatvedt CH (1997). Dev Biol 186: 58–72. Yang JT, Rayburn H, Hynes RO (1995). Development 121:549–560. Yelon D, Horne SA, Stainier DY (1999). Dev Biol 214:23–37. Zhang H, Bradley A (1996). Development 122: 2977–2986. Zhang H-Y, Lardelli M, Ekblom P (1997). Dev Genes Evol 207:340–351. Zhang J, Talbot WS, Schier AF (1998). Cell 92:241–251.

10 Genetic Defects Resulting in Abnormalities of Vertebrate Left–Right Asymmetry Provide Insight Into the Underlying Developmental Mechanisms

MARTINA BRUECKNER Yale School of Medicine New Haven, Connecticut

I.

Left–Right Asymmetry Is a Unique Feature of Vertebrate Development

Humans and all other vertebrates externally appear, to be bilaterally symmetrical. The left and right arm are mirror images of each other, as are the legs and the left and right halves of the trunk. Underlying the external symmetry, however, is extensive left–right asymmetry of the internal organs. The thoracic organs, in particular, have a high degree of left–right asymmetry. The three basic body axes along which the vertebrate body plan is organized are anteroposterior (AP), dorsoventral (DV), and left–right (LR), and among these, the LR axis is unique in several ways. Left–right is the dependent axis, being determined by the previously established AP and DV axes. In addition, LR is a binary decision instead of a continuous gradient like the AP and DV axes. Finally, the handedness of left–right asymmetry is consistent across all vertebrates. This means that the organism not only needs to have a mechanism by which it creates asymmetry along the LR axis, but that there also has to be a mechanism that consistently aligns that asymmetry to the preexisting AP and DV axes. Why did the vertebrate organism evolve this additional level of spatial complexity? Perhaps it permitted the development of the more complex cardiopulmonary system that is necessary 239

240

Brueckner

to support a larger organism dependent on active circulation instead of diffusion. This notion is supported by the high incidence of complex, often lethal cardiac defects in mice and humans with defective development of LR asymmetry. Study of genetic defects in left–right development in mice and humans have lead to significant insights into the molecular mechanism underlying this fascinating and unique step in vertebrate development. II. There Is Extensive Anatomical Left–Right Asymmetry in Normal Heart and Lungs The most striking of the visceral left–right asymmetries is observed in the thoracic organs. The normal alignment of organs along the vertebrate left–right axis is referred to as situs solitus (Fig. 1a), and determination of situs is the starting point for the evaluation of all normal and abnormal cardiovascular anatomy. In humans with situs solitus, the right lung has three lobes, whereas the left lung has two. The trachea gives rise to anatomically distinct left and right mainstem bronchi. The left mainstem bronchus courses underneath the pulmonary artery, and is thus called the hyparterial bronchus; the right mainstem bronchus is shorter and courses above the pulmonary artery as the eparterial bronchus.

Figure 1 Left–right asymmetry of the thoracic and visceral organs. (a) Situs solitus: The cardiac apex points to the left, the right lung is trilobed, the left bilobed. The stomach is on the left, liver is on the right, and there is a solitary, left-sided spleen. (b,c) Isomerism syndromes. There is failure to break bilateral symmetry and the organization of one or more organ systems approaches symmetry. (b) Right isomerism: There are two right atria, both lungs are trilobed, and the spleen is absent. (c) Left isomerism: There are two left atria, both lungs are bilobed, and there are multiple spleens. (d) Situs inversus: This is a perfect mirror image of situs solitus, and the relationships between structures remains unaltered from that observed in situs solitus.

Genetic Defects Resulting in Abnormalities

241

The arterial and venous systems are also highly LR asymmetrical. The inferior and superior venae cavae (IVC and SVC) are to the right of the spine, whereas the descending aorta is on the left. Both right and left pulmonary veins join the atrium on the left. The most striking intrathoracic LR asymmetry, however, is the heart itself. It is located almost entirely within the left side of the chest cavity, with the cardiac apex pointing to the left. The left and right atria are distinct structures that are clearly identifiable by the morphology of the atrial appendages: The right is broad-based and thumb-like, whereas the left is a long narrow structure. The IVC and SVC drain into the right atrium, and the sinus node is located at the confluence of the SVC and the right atrial appendage. Internally, the right atrium is characterized by the presence of the eustachian valve and crista terminalis. The left atrium receives the pulmonary veins and is characterized by the relative absence of distinct trabeculated regions, reflecting that it is derived largely from embryonic venous structures. Visible left–right asymmetry of the heart continues in the positioning of the ventricles, with the smooth left ventricle on the left of the trabeculated right ventricle. III. Pathology Resulting from Abnormal Development of Left–Right Asymmetry When there is failure of the normal development of asymmetry along the leftright body axis, a variety of outcomes are possible (see Fig. 1). If there is an established asymmetry, but it is opposite from the normal, the result is situs inversus totalis (see Fig. 1d): a complete mirror image of organs along the left right axis resulting in normal relationships between the left–right position of the organs. If there is failure to establish any concordant asymmetry of organs along the left–right axis, then heterotaxy, defined as any arrangement of organs across the left–right axis differing from complete situs solitus or complete situs inversus results. The most extreme form of heterotaxy is isomerism (see Fig.1 B,C): In this case, there is apparent failure to break bilateral symmetry in the developing embryo resulting in two right body sides, as in right atrial isomerism, or two left body sides, as in left atrial isomerism. IV. Cardiovascular Malformations Are Associated with Heterotaxia The spectrum of cardiovascular malformations associated with heterotaxia demonstrates tremendous heterogeneity. The spectrum is similar between human cases of heterotaxy and mouse mutations that result in abnormal development of left–right asymmetry. The clinical severity of cardiovascular defects associated with heterotaxia ranges from asymptomatic to complex life-threatening malfor-

242

Brueckner

mations. Although discrete entities are described, it is important to recognize the significant amount of overlap between them, emphasizing that heterotaxia represents a spectrum of lateralization disturbances. A. Situs Inversus

The incidence of situs inversus has been estimated to range from 1/2500 to 1/ 20,000 living persons; with 1/7000 to 1/8000 being the most common estimate from mass radiographic surveys in adults [1]. The exact incidence of congenital heart disease in situs inversus is not known but appears to be similar to the general population. B. Isolated Dextrocardia (Dextrocardia with Situs Solitus)

Dextrocardia is defined by the displacement of the cardiac apex to the right. Intracardiac pathology exists in 90% of patients with isolated dextrocardia. The incidence of isolated dextrocardia has been reported to range from 1/7500 to 1/ 29,000 [2]. The cardiovascular malformations associated with isolated dextrocardia include atrioventricular discordance situs (l-loop ventricles) (50%), single ventricle (25%), and ventricular septal defect (60%). Fifty percent have atrioventricular and ventriculoarterial discordance, or l-transposition of the great arteries (corrected transposition), and 10% have d-transposition of the great arteries (complete transposition). Sixty percent have pulmonary stenosis or atresia and 8% have a right-sided aortic arch [3]. C. Isolated Levocardia (Levocardia with Situs Inversus)

Isolated levocardia is rare, but the incidence of intracardiac pathology approaches 100%. Of 65 cases of isolated levocardia reviewed by Van Praagh, 9 had situs inversus and 56 had situs ambiguous and thus actually represented heterotaxia. All nine cases of isolated levocardia had d-ventricular loops and d-malposition of the aorta. Associated lesions include pulmonary stenosis or atresia (89%), atrioventricular canal (67%), and total anomalous pulmonary venous return (44%) (4). D. Right Isomerism

Right isomerism represents bilaterally symmetrical right sides: Both lungs are trilobed, there are two right atrial appendages and bilateral sinus nodes, and the spleen, a left-sided structure, is absent. The cardiovascular malformations associated with right isomerism include atrioventricular septal defect, which is frequently unbalanced (85%); total anomalous pulmonary venous return (70%); transposition of the great arteries (80%); pulmonary stenosis/atresia (80%); dextrocardia (40%); atrial septal defect/common atrium (90%); persistence of a left

Genetic Defects Resulting in Abnormalities

243

superior vena cava (50%); and single ventricle (50%) [5]. Bilateral sinus nodes are frequently present. E. Left Isomerism

In general, the cardiovascular malformations with left isomerism are often less severe than those observed in association with right atrial isomerism, and include atrioventricular septal defects (40%), which are frequently unbalanced; dextrocardia (40%); and atrial septal defect/common atrium (80%). Transposition (30%) and pulmonary stenosis/atresia (30%) occurs less frequently than in right isomerism. Persistent left superior vena cava (40%), interruption of the suprarenal portion of the inferior vena cava with azygous continuation (70%), partial anomalous pulmonary venous return (40%), and left ventricular outflow tract obstruction (40%) are frequently seen in left isomerism [6]. Since the sinus node is a right atrial structure, and thereby absent in left isomerism, an ectopic atrial focus may be present with an abnormal P-wave axis on the electrocardiogram. V.

Development of Left–Right Asymmetry

A. Morphological Development of Left–Right Asymmetry Begins at the Node and Proceeds in a Cephalocaudal Direction

The first visible asymmetry in the developing vertebrate embryo is the subtle leftward movement of the cardiac inflow called ‘‘cardiac jogging’’ [7]. This is rapidly followed by pronounced rightward bending of the primitive heart tube. This event, called ‘‘cardiac looping,’’ is the first grossly obvious organismal asymmetry, and it occurs at stage 44 of Xenopus development, postconceptual day 8.5 in the mouse and day 23 of human gestation. Cardiac looping is essential to normal cardiac development: It establishes all the normal relationships between vessels, developing valvular structures, and cardiac chambers. When there is defective cardiac looping, a range of complex intracardiac defects result. Following the formation of the cardiac loop, the entire embryo turns in a helical fashion, placing the ‘‘tail’’ to the right of the head. Left–right asymmetrical development then continues to proceed in a cephalad to caudal fashion. The foregut derivatives, including the lung, become asymmetrical at day 10.5, finally the midgut and hindgut rotate to complete the development of obvious internal asymmetry. B. Molecular Left–Right Asymmetries Develop Before Morphological Asymmetry

Molecular asymmetry is established well before there is visible embryonic asymmetry, and analysis of the hierarchy of molecular asymmetries in the mouse, chick, and Xenopus has permitted the elucidation of a molecular pathway of left–

244

Brueckner

right development [8]. For the purpose of this chapter, the focus is on genetic evidence obtained from the study of mammalian systems; in particular the mouse and human. In the mouse, the first molecular asymmetry arises when the nodal gene, initially expressed throughout the node, becomes restricted to the left side of the node [9,10]. Nodal and lefty-2 are expressed in the left lateral plate mesoderm, and lefty-1 is expressed in the left side of the neural floorplate where it may function as part of a midline barrier [9–12]. Sonic hedgehog expression at the midline may be required to maintain these expression patterns [13,14]. The homeobox genes PitX2 [15] and Nkx.32 [16] have opposite asymmetrical expression patterns in the left and right lateral plate mesoderm, respectively, which may regulate asymmetrical expression of downstream effector genes. Overall, asymmetrical gene expression begins at the node and then proceeds laterally and caudally . It is important to note, however, that asymmetrical expression of a

Figure 2 Theoretical model for the generation of left–right asymmetry. (a) An asymmetrical structure is made. Failure to create this asymmetry in the first place results in persistence of bilateral symmetry. (b) The asymmetrical structure is aligned relative to the AP and DV axes. This produces cellular asymmetry (shown in a broadgray band). When the alignment is defective, it results in random asymmetry. (c) The cellular asymmetry is signaled to the developing organism. Failure of this step again generates persistent symmetry.

Genetic Defects Resulting in Abnormalities

245

signaling molecule like nodal cannot arise de novo, and requires preexisting LR positional information. C. Theoretical Considerations that Apply to the Development of Handed Asymmetry

The left–right axis is the last of the embryonic axes to be established. It is therefore defined by the AP and DV axes. The embryo must thus be able to both create asymmetry and align that asymmetry relative to the AP and DV axes (Fig. 2). It was first proposed by Wilhemi [17] that the organism had an underlying mechanism to generate random asymmetry. Brown and Wolpert then hypothesized that this asymmetry is biased in a consistent direction by orienting a handed molecule or macromolecular structure relative to the AP and DV axes [18]. Left and right isomerism are the likely phenotypes resulting from such persistence of bilateral symmetry. In contrast, the inability of the embryo to orient left–right asymmetry relative to the AP and DV axes should result in randomly oriented asymmetry. A defect in this step would therefore result in a distribution of phenotypes, including complete situs solitus and complete situs inversus. This hypothesis relies upon the existence of an asymmetrical molecule or macromolecular structure. This molecule or structure must be able to align consistently to the AP and DV axes; subsequently, it must provide a method by which to communicate the resulting cellular asymmetry to the rest of the organism. Once local asymmetry is established, asymmetrical signals could be communicated to developing organs by coopting well-established signaling pathways that are used in multiple other developmental steps. VI. Genetic Defects Affecting Left–Right Development Provide Insight into the Molecular Pathway That Creates Organismal Asymmetry Naturally occurring genetic defects in the development of left–right asymmetry have been identified in humans [19] and mice [20]. In addition, targeted mutagenesis of genes known or suspected to be important in this pathway because of, for example, their asymmetrical mRNA expression patterns, has generated a series of mutations that have been essential to our understanding of the mechanism of specifying and communicating left–right positional information. A. The Node and the Midline Are the Signaling Centers from Which Left–Right Positional Information Originate

The embryonic node is a major organizing center in primitive streak–stage embryos that regulates pattern formation (Fig. 3). It is the mammalian equivalent

246

Brueckner

Figure 3 The mouse node. (a) schematic drawing of the mouse node. The ectoderm is shown in dark gray, endoderm in black, and mesoderm in light gray. (b) Electron micrograph at 500⫻ showing the triangular shape of the mouse node (outlined in black). (c) Each of the central (pit) cells has a single cilium.

of Hensen’s node in the chick and the Spemann organizer in Xenopus. The mouse node forms at 7.0–7.5 days of embryonic development and is located at the distal tip of the embryo. At the node, the ectoderm (shown in yellow in Fig. 3) and endoderm (shown in pink in Fig. 3) make contact with each other. The proliferative ectodermal layer is separated from the nonproliferative endodermal layer by a basement membrane. The ‘‘organizer’’ function is contained within the endodermal cells: When they are transplanted to different parts of the embryo, they are able to induce a secondary axis. It is thought that positional information is acquired by the mesoderm as it contacts the node during gastrulation. Many genes with fundamental developmental roles are expressed in the node, and defects in genes essential to node formation are usually lethal and result in a failure of gastrulation. Interestingly, however, compound heterozygous mutations in two of these genes produce discernible defects in left–right development [9].

Genetic Defects Resulting in Abnormalities

247

B. Hnf-3␤ and nodal Are Essential to Node Development and Left– Right Development

Hnf-3β is a member of the Hnf-3/fork head family of transcription factors. It is expressed in the node, notochord, and floorplate [21] in addition to the visceral and definitive endoderm. Hnf-3β ⫺/⫺ mice are completely missing the node and notochord, have severe abnormalities of DV patterning, and die before day 11.5 of development. When the early lethality of the hnf-3β ⫺/⫺ mice is partially rescued by normal visceral endoderm in tetraploid embryo aggregation experiments, the resulting embryos demonstrate a loss of left–right asymmetry. This is demonstrated by the bilateral or absent expression of the molecular markers nodal and lefty, which are normally expressed in the left lateral plate mesoderm [9]. Nodal Is a member of the TGF-β family of secreted growth factors, which, like hnf-3β, is expressed in the mouse node at the time of gastrulation [22]. Mice that are nodal ⫺/⫺ also fail to develop a node, but their phenotype is more severe than that of the hnf-3β ⫺/⫺ animals. The nodal ⫺/⫺ mutants fail to gastrulate and form no embryonic mesoderm at all. While both nodal ⫹/⫺ and hnf-3β ⫹/⫺ mice appear phenotypically normal, the compound heterozygous nodal ⫹/⫺/hnf-3β ⫹/⫺ mice express nodal bilaterally, and a significant number of them fail to turn and develop abnormal positioning of the heart and viscera along the left–right axis [9]. Both nodal and hnf-3β are essential to forming a normal node, and mutations in both produce abnormal development of left–right asymmetry. Although the embryos do not develop far enough to evaluate the morphological asymmetry of the heart and lungs, the bilateral or absent expression of normally lateralized markers suggest that these mutations lead to a failure to generate asymmetry. The node itself is formed prior to any evidence of morphological or molecular asymmetry, and absence of asymmetry in mouse mutants with an absent or abnormal node suggests that the initial asymmetry is formed here. C. Defects in Node Monociliary Structure and Function Demonstrate that the Node Monocilium Is the Fundamental Asymmetrical Structure

Cells of the embryonic node each contain a single cilium on their ventral side [23]. Evidence for the direct involvement of node monocilia in LR patterning came from analysis of mutations in several members of the microtubule-dependent motor protein families dynein and kinesin. Kinesins and dyneins are large families of directed microtubule-dependent motors that are required for diverse intracellular transport functions, with the kinesins providing movement toward the plus end of the microtubule, whereas the dyneins move toward the minus

248

Brueckner

end. In addition, the axonemal dyneins provide the force required to bend cilia and flagellae. Targeted mutagenesis of two members of the heterotrimeric kinesin family KIF3A [24,25] and KIF3B [27] result in similar phenotypes, including midgestation lethality and multiple severe developmental abnormalities. Notably, both mutants display randomization of cardiac looping and abnormal LR expression patterns of genes such as lefty-2, implicating heterotrimeric kinesin in LR development. A striking feature of kif3A and kif3B homozygous mutants is that they have no node monocilia. Thus, heterotrimeric kinesin is essential both for assembly of node cilia and normal LR development. The recessive mouse iv (inversus viscerum) mutation, first described at the Jackson laboratories in 1959 [20], results in random asymmetry. Fifty percent of liveborn mice homozygous for the iv mutation have situs solitus, the other 50% have complete situs inversus. Expression of many of the asymmetrically expressed genes is perturbed [9,10]; however, unlike the mice with KIF3A and KIF3B mutations described above, the abnormalities in iv/iv mice are entirely limited to defective left–right development [27]. The iv mutant phenotype is due to a missense mutation in the motor domain of an axonemal dynein heavy chain gene named left–right dynein (lrd) [28]. lrd mRNA is expressed in the developing embryo in the ventral node cells (see Fig. 2b) which, as previously described, each possess a monocilium. These monocilia have a 9 ⫹ 0 microtubule configuration which sets them apart from other cilia, such as the cilia found on cells lining the oviduct, which have a 9 ⫹ 2 microtubule arrangement. Despite this, however, normal node cilia are motile, rotating in a vortical manner. The movement of the node monocilia is able to produce directional, leftward flow of the fluid surrounding the mouse node, called the ‘‘nodal flow’’ [26]. In mice homozygous for defects in the left–right dynein gene, the node monocilia are present and appear morphologically normal, but they are rigid and immotile [27]. This observation suggests that movement of node monocilia is required for the establishment of left–right asymmetry. Further evidence for the involvement of dynein in the determination of left– right asymmetry comes from humans with Kartagener’s syndrome. Kartagener’s syndrome is a genetic disorder defined by bronchiectasis, chronic sinusitis, and a roughly 50% incidence of situs inversus (mirror-image organ positioning) [19]. Like the mice with lrd mutations, patients with Kartagener’s syndrome have no other developmental abnormalities outside of random left–right asymmetry. Unlike lrd mutations, however, Afzelius observed that patients diagnosed with Kartagener’s syndrome had immotile respiratory cilia and sperm flagella [29]. The ciliary and flagellar immotility in these patients was due to absent or defective outer dynein arms. Kartagener’s syndrome is genetically heterogeneous with predominantly recessive inheritance [30]. Recent evidence suggests that defects in components of the dynein complex, including dynein intermediate chains, may

Genetic Defects Resulting in Abnormalities

249

be responsible for Kartagener’s syndrome [31]. Therefore, both the mouse iv mutation and human Kartagener’s syndrome are due to defects in dynein, and it is quite possible that dynein-driven motility of the node monocilia is also abnormal in patients with Kartagener’s syndrome. The KIF3A, KIF3B, and lrd mutations indicate that directional motility of the node monocilia is responsible for creating and aligning LR asymmetries [32]. Thus, asymmetry is initially provided by node monocilia. Note that both basal bodies and cilia are inherently highly chiral structures. Handedness is thus provided to the organism by the chirality of the complex of basal body and cilium. Assembly of the node monocilia requires heterotrimeric kinesin KIF3. KIF3A and KIF3B mutations thus produce the absence of the initial asymmetrical structure and result in preservation of symmetry, as manifested by bilateral or absent expression of normally lateralized molecular markers. The macromolecular chirality of the cilium is then converted to organismal asymmetry by lrd-driven ciliary motility. Mutations in ciliary function, such as the iv mouse or human Kartagener’s syndrome, impair the ability of the handedness provided by the node monocilium to be correctly aligned to the AP and DV axes. Therefore, their predominant phenotype is random asymmetry. D. Defects in the Lateral Signaling Pathway Result in Preservation of Symmetry

In order for the developing organs to interpret the asymmetry set up at the node by the node monocilia, the initial asymmetrical information must be communicated over relatively large distances and in the correct temporal order. This cascade of gene expression hypothetically regulates asymmetrical organ morphogenesis by leading to differential cell proliferation, migration, adhesion, and/or apoptosis on the left and right sides of the embryo. Mother Nature is very economical, recycling signaling molecules for use in multiple developmental pathways; for example, nodal signaling is used both in mesoderm formation and LR development [22]. The different developmental roles of this single gene are controlled by separate transcriptional regulatory elements, one of which is specific to expression at the node and the other to later expression as a component of the downstream left-rided signaling cascade [33]. Although the structure or soluble factor that is moved by the node monocilia remains an enigma, several of the downstream signaling molecules are well characterized. Fgf8 May Be Involved in Signaling Left–Right Information from the Node

Fibroblast growth factor-8 (Fgf8) is a secreted growth factor that is expressed in the developing embryo at the posterior border of the node and along the primitive streak. Although homozygosity for a null allele in Fgf 8 results in failure to gastrulate and early embryonic lethality, a hypomorphic allele that produces a

250

Brueckner

greater than 50% reduction in Fgf 8 mRNA produces mice with abnormal direction of heart looping and right pulmonary isomerism [14]. Expression of nodal and PitX2 that is normally found in the left lateral plate at day 8 was absent. Implanting an Fgf 8-soaked bead in the right lateral plate mesoderm of a normal embryo resulted in ectopic nodal expression. These data imply that Fgf 8 is required to induce lateral plate nodal expression on the left, and that Fgf 8 signaling is essential for the developing heart and lung to acquire ‘‘left’’ identity. It is possible that Fgf 8 itself is the soluble factor moved to the left by the node monocilia, and that Fgf8 protein activates a receptor on the left side of the node which triggers the downstream ‘‘left’’ signaling cascade. Activin2B Receptor and the Transcription Factor PitX2 Are Required to Signal Asymmetrical Information to the Developing Lungs

The activin2B receptor (ActIIRB) is a cell surface serine/threonine kinase receptor that interacts in vitro with activins, and it has been shown to affect Vg1 signaling in Xenopus. The range of ligands that interact with the ActIIRB in vivo has not been well characterized, but it may include many members of the TGFβ family of secreted growth factors, including some implicated in left–right development such as nodal and Vg1. ActRIIB is expressed symmetrically in a broad range of tissues during early development, including the epiblast of the pregastrulation embryo and the embryonic ectoderm during gastrulation. Targeted mutagenesis of ActIIRB results in postnatal lethality and right isomerism of the lungs and atria [34]. In addition, there was frequently observed splenic hypoplasia. Interestingly, situs inversus was never observed, along with consistently normal left–right positioning of the gut. PitX2 is a bicoid-related transcription factor that is asymmetrically expressed in the left lateral plate mesoderm of the day 8 mouse embryo. In the chick and Xenopus, ectopic PitX2 expression in the right lateral plate can reverse cardiac looping and alter the left–right development of the gut [15]. Targeted mutagenesis of PitX2 in mice results in embryonic lethality and right lung isomerism [35,36]. The embryos begin to turn normally in a counterclockwise direction, but fail to complete turning the most caudal part of the embryo. Interestingly, the heart loop begins to form normally and is always directed to the right. This implies that PitX2 is a signaling molecule specifically for lung left–right positional information, but it does that it does not communicate left–right information to the straight heart tube or the gut. The right pulmonary isomerism phenotype is striking and shared by the ActIIRB ⫺/⫺ and PitX2 ⫺/⫺ mice. This suggests that PitX2 may be a downstream target of signaling mediated by the ActIIRB, and that it provides essential ‘‘leftness’’ information to the developing lung. Notably, neither of these genes determines heart or gut handedness. Fgf8 and nodal may be involved in signaling

Genetic Defects Resulting in Abnormalities

251

broad left–right information from the node and subsequently connect to separate, organ-specific signaling cascades to the heart and gut. Defects in all of these downstream signaling molecules produce preserved symmetry as manifested by right pulmonary isomerism and the symmetrical absence of expression of molecular markers normally lateralized to the left. VII. Conclusions Careful analysis of the ever-increasing number of identified genetic defects in the mouse and human that perturb the development of handed asymmetry during embryogenesis has begun to reveal a pathway that includes cytoskeletal proteins, secreted growth factors, and transcription factors as outlined in Fig. 4. Left–

Figure 4 The beginnings of a molecular pathway that determines left–right asymmetry in the mouse. Nodal and Hnf3-β contribute to development of the node, whereas KIF3A, KIF3B and lrd contribute to dynein-driven directional ciliary motility. The asymmetry is possibly signalled via fgf8 to produce asymmetrically expression of the genes nodal, lefty2, Pitx2, and Nkx.32. Lefty-1 may function as a midline barrier to prevent the diffusion of left-lateralized gene products to the right.

252

Brueckner

right asymmetry is generated at the node at the onset of gastrulation. The handed molecule that is aligned relative to the AP and DV axes to create the first break in bilateral symmetry is the node monocilium. The resultant cellular asymmetry is then aligned with the AP and DV asymmetry of the node, and the cellular asymmetry is converted to organismal asymmetry by left–right dynein-driven directional motility of the node monocilium. Well-conserved signaling pathways subsequently communicate the asymmetry defined in the peri-nodal region to the organs of the developing embryo. Failure of creation or communication of the initial asymmetry results in persistence of symmetry, such as that observed in the human isomerism syndromes: The two body halves are relatively symmetrical, with; for example, two left atria and spleens on both sides of the abdomen. In contrast, failure to align the asymmetry relative to the AP and DV axes results in random asymmetry: fifty-percent of affected patients with Kartagener’s syndrome and 50% of mice with defects in left–right dynein are situs solitus and the other 50% are situs inversus. Many more genes are most likely to be involved in this pathway, and we have just begun to appreciate the complexity involved in establishing the left– right axis. For instance, the identity of the molecule or structure that is moved by the node cilia remains unknown. Hopefully, the creation and analysis of additional genetic defects in the mouse that affect the development of left–right asymmetry will lead to a complete understanding of this fundamental step in vertebrate development.

References 1. Squarcia U, Ritter DG, Kincaid OW. Dextrocardia: angiographic study and classification. Am J Cardiol 1973; 32:965–977. 2. Gutgesell HP. Cardiac malposition and heterotaxy. In: Garson A, Bricker JT, McNamara DG, eds. The Science and Practice of Pediatric Cardiology. Philadelphia: Lea and Febiger, 1990, pp 1280–1303. 3. Lev M, Liberthson RR, Eckner FAO, Arcilla RA. Pathologic anatomy of dextrocardia and its clinical implications. Circulation 1968; 37:979–999. 4. Van Praagh R, Weinberg PM, Matsouka R, Van Praagh S. Malpositions of the heart. In: Adams FC, Emmanouilides GC, eds. Moss’ Heart Disease in Infants, Children, and Adolescents. Baltimore; Williams & Wilkins, 1983, pp 422–458. 5. Van Mierop LHS. Asplenia and polysplenia syndrome. Birth Defects: Original Article Series 1972; 8:74–82. 6. Peoples WM, Moller JH, Edwards JE. Polysplenia: a review of 146 cases. Pediatr Cardio 1983; 4:129–137. 7. Chen JN, van Eeden FJ, Warren KS, Chin A, Nusslein-Volhard C, Haffter P, Fishman MC. Left-right pattern of cardiac BMP4 may drive asymmetry of the heart in zebrafish. Development 1997; 124:4373–82.

Genetic Defects Resulting in Abnormalities

253

8. Levin M, Johnson RL, Stern CD, Kuehn M, Tabin C. A molecular pathway determining left-right asymmetry in chick embryogenesis. Cell 1995; 82:803–814. 9. Collignon J, Varlet I, Robertson E. Relationship of asymmetricnodal expression and the direction of embryonic turning. Nature 1996; 381:155–158. 10. Lowe LA, Supp DM, Sampath K, Yokoyama T, Wright CVE, Potter SS, Overbeek P, Kuehn MR. Conserved left-right asymmetryof nodal expression and alterations in murine situs inversus. Nature 1996; 381:158–161. 11. Meno C, Saijoh Y, Fijii H, Ikeda M, Yokoyama T, Yokoyama M, Toyoda Y, Hamada H. Left-right asymmetric expression of the Tgf-β family member lefty in mouse embryos. Nature 1996; 381:151–155. 12. Meno C, Shimono A, Saijoh Y, Yashiro K, Mochida K, Ohishi S, Noji S, Kondoh H, Hamada H. Lefty-1 is required for left-right determination as a regulator of lefty2 and nodal. Cell 1998; 94:287–297. 13. Izraeli S, Lowe LA, Bertness VL, Good DJ, Dorward DW, Kirsch IR, Kuehn MR. The SIL gene is required for mouse embryonic axial development and left-right specification. Nature 1999; 399:691–694. 14. Meyers EN, Martin GR. Differences in left-right axis pathways in the mouse and chick: functions of fgf8 and shh. Science 1999; 285:403–406. 15. Yoshioka H, Meno C, Koshiba K, Sugihara M, Itoh H, Ishimaru Y, Inoue T, Ohuchi H, Semina EV, Murray JC, Hamada H, Noji S. Pitx2, a bicoid-type homeobox gene, is involved in the lefty signalling pathway in determination of left-right asymmetry. Cell 1998; 94:299–305. 16. Schneider A, Mijalski T, Schlange T, Dai WL, Overbeek P, Arnold HH, Brand T. The homeobox gene Nkx3.2 is a target of left-right signaling and is expressed on oposite sides in chick and mouse embryos. Curr Biol 1999; 9:911–914. 17. Wilhelmi H. Experimentelle Untersuchungen ueber situs inversus viscerum. Arch Entwicklungsmech 1921; 48:517–532. 18. Brown NA, Wolpert L. The development of handedness in left/right asymmetry. Development 1990; 109:1–9. 19. Eliasson R, Mossberg B, Cammer P, Afzelius B. The immotile-cilia syndrome: a congenital ciliary abnormality as an etiologic factor in chronic airway infections and male sterility. N Engl J Med 1977; 297:1–6. 20. Hummel KP, Chapman DB. Visceral inversion and associated anomalies in the mouse. J Hered 1959; 50:9–13. 21. Ang SL, Rossant J. HNF-3 beta is essential for node and notochord formation in mouse development. Cell 1994; 78:561–574. 22. Zhou X, Sasaki H, Lowe L, Hogan BL, Kuehn MR. Nodal is a novel TGF-β–like gene expressed in the mouse node during gastrulation. Nature 1993; 361:543–547. 23. Sulik K, Dehart DB, Iangaki T, Carson JL, Vrablic T, Gesteland K, Schoenwolf GC. Morphogenesis of the murine node and notochordal plate. Dev Dyn Am 1994; 201:260–278. 24. Takeda S, Yonekawa Y, Tanaka Y, Okada Y, Nonaka S, Hirokawa N. Left-right asymmetryand the kinesin superfamily protein KIF3A: New insights in determination of laterality and mesoderm induction by kif3A ⫺/⫺ mice. J Cell Biol 1999; 145:825–836. 25. Marszalek JR, Ruiz-Lozano P, Roberts E, Chien KR, Goldstein LS. Situs inversus

254

26.

27.

28. 29. 30. 31.

32. 33.

34.

35.

36.

Brueckner and embryonic ciliary morphogenesis defects in mouse mutants lacking the KIF3A subunit of kinesin-II. Proc Natl Acad Sci USA 1999; 96:5043–5048. Nonaka S, Tanaka Y, Okada Y, Takeda S, Harada A, Kanai Y, Kido M, Hirokawa N. Randomization of left-right asymmetry due to loss of nodal cilia generating leftward flow of extraembryonic fluid in mice lacking the KIF3B motor protein. Cell 95:829–837. Supp DM, Brueckner M, McGrath J, Corrales J, Witte DP, Kuehn M, Lowe L, Potter SS. Targeted deletion of the ATP binding domain of left-right dynein confirms its role in specifying development of left-right asymmetries. Development 1999; 126: 5495–5504. Supp DM, Witte DP, Potter SS, Brueckner M. Mutation of an axonemal dynein gene affects left-right asymmetry in inversus viscerum mice. Nature 1997; 389:963–966. Afzelius, BA. A human syndrome caused by immotile cilia. Science 1976; 193:317– 319. Sturgess JM, Thompson MW, Czegledy-Nagy E, Turner JA. Genetic aspects of immotile cilia syndrome. Am J Med Genet 1986; 25:149–60. Pennarun G, Escudier E, Chapelin C, Bridoux AM, Cacheux V, Roger G, Clement A, Goossens M, Amselem S, Duriez B. Loss-of-function mutations in a human gene related to Chlamydomonas reinhardtii dynein IC78 result in primary ciliary dyskinesia. Am J Med Genet 1999; 65:1508–1519. Supp DM, Potter SS, Brueckner M. Molecular motors: the driving force behind mammalian left-right development. Trends Cell Biol 2000; 10:41–45. DP, Robertson EJ. Asymmetric and node-specific nodal expression patterns are controlled by two distinct cis-acting regulatory elements. Genes Dev 1999; 13:1575– 1588. Oh SP, Li E. The signalling pathway mediated by the type-IIB activin receptor controls axial patterning and lateral asymmetry in the mouse. Genes Dev 1997; 11: 1357–1369. Lu M, Pressman C, Dyer R, Johnson RL, Martin JF. Function of Rieger syndrome gene in left-right asymmetry and craniofacial development. Nature 1999; 401:276– 278. Lin CR, Kioussi C, O’Connell S, Briata P, Szeto D, Liu F, Izpisua-Belmonte JC, Rosenfeld MG. Pitx2 regulates lung asymmetry, cardiac positioning and pituitary and tooth morphogenesis. Nature 1999; 401:279–282.

11 Zebrafish A Developmental and Genetic Model for Hematopoiesis and Hematopoietic Disorders

SCOTT D. MARTY, M. ERNEST DODD, and SHUO LIN Medical College of Georgia Augusta, Georgia

I.

Introduction to Hematopoiesis

Developmental biology has made impressive strides over the past decade to unravel the mysteries surrounding vertebrate development. Much of the progress has been generated through intensive studies of a few model organisms. A new player in the field of vertebrate development is the zebrafish, which provides a unique opportunity for investigators to combine genetic and embryonic approaches into the studies of developmental biology. Since this volume deals with cardiorespiratory biology, we will focus our discussion on blood formation, specifically red blood cell development. To understand how this process can be thoroughly studied using zebrafish, we will start with an introduction to vertebrate hematopoiesis and zebrafish development. We will then detail specific techniques that are often utilized in zebrafish studies and conclude with a discussion of a number of zebrafish hematopoietic mutations, some of which share similarities to human diseases. A. Review of Vertebrate Hematopoiesis

Hematopoiesis is the process by which blood cells of multiple lineages are continuously produced throughout the life of an individual [1]. This process begins 255

256

Marty et al.

with a small number of multipotent stem cells known as hematopoietic stem cells (HSCs). The HSCs are a unique population of cells in that they have the ability to self-renew and give rise to immature progenitor cells. This small population of HSCs gives rise to the multiple lineages of blood cells (erythroid cells, B-lymphoid cells, T-lymphoid cells, megakaryocytes, macrophages, platelet cells, neutrophils, and mast cells) through progressive commitment to a single lineage. Commitment of the HSCs to a single lineage involves extensive cell signaling and transcriptional control, which will be discussed later in the chapter. Vertebrate hematopoiesis occurs in successive waves, and the anatomical location of the HSCs changes throughout development. In mammals, the first blood cells arise in the blood islands of the yolk sac. These cells are derived from the ventral mesoderm and are referred to as the embryonic red blood cell progenitors, as they are restricted to the erythroid lineage. This stage of hematopoiesis is termed ‘‘primitive’’ hematopoiesis. Primitive erythroid cells express embryonic globins and remain nucleated. The next wave of hematopoiesis occurs sometime during midgestation in the fetal liver. These cells are derived from the dorsal mesoderm and are capable of giving rise to pluripotent hematopoietic stem cells which are capable of producing the complete set of hematopoietic lineages. This wave of hematopoiesis is referred to as ‘‘definitive’’ hematopoiesis. Definitive or fetal/adult stem cells remain at rest in G0 for an extended period of time. Upon expression of adult globins, the erythroid progeny expel their nuclei during maturation [2]. The differences between the primitive and definitive erythroid cells have led to the formation of a hypothesis that primitive and definitive hematopoietic stem cells originate independently from different embryonic sites. Although these waves of hematopoiesis occur in different anatomical locations, it has been shown that many of the same genes are needed for both waves to proceed. The final site of hematopoiesis resides in the bone marrow where blood cell formation begins shortly before birth and continues throughout the life of the vertebrate. B. Factors Required for Hematopoiesis

The following questions are often asked in the studies of hematopoiesis: 1. What is the embryonic origin of HSCs? 2. What are the early embryonic patterning events responsible for the induction of embryonic blood? 3. What are the key cell-autonomous regulatory genes that function downstream of the cytokine network to control proliferation and differentiation of progenitor cells [2]? To understand how an embryonic tissue becomes hematopoietic during development and to uncover the factors required for this process, model systems such as Xenopus, the mouse, and the zebrafish have been used extensively. Prior to primitive hematopoiesis, the establishment of the hematopoietic program begins with the induction and specification of ventral mesoderm to a

Zebrafish: A Developmental and Genetic Model

257

hematopoietic fate. Molecular studies have uncovered a number of factors that induce mesoderm, namely, members of the transforming growth factor-β (TGF-β) superfamily, fibroblast growth factors, and Wnt proteins [1]. Of these factors, bone morphogenic protein 4 (BMP4), a TGF-β family member, has been implicated in ventral mesoderm induction [3–6]. The role of BMP4 was revealed through molecular studies in which BMP4’s activity was removed or added. When BMP4’s function was removed, normal mesoderm axis formation was disrupted, resulting in a dorsalized embryo accompanied by a decrease in globin gene expression. The expression of globin genes is often used as specific markers to determine the extent of erythropoiesis. Thus, a decrease of globin gene expression is reminiscent of a decrease in blood development. When BMP4’s function was added, globin gene expression was increased, and the embryo was ventralized. These two experiments establish BMPs as candidate factors involved in early mesoderm induction. In addition to inducing mesoderm, BMP4 has been shown to play a role in other embryonic processes such as gastrulation and organ formation [7]. The ability of one factor to play a role in multiple processes is not uncommon for critical hematopoietic factors. Once ventral mesoderm is induced, it is further specified to a hematopoietic fate. Exactly how this occurs has been a central issue of developmental hematopoiesis studies. A loss of a factor required for this process should affect all blood cell lineages. One such candidate is the T-lymphoid acute leukemia oncoprotein (Tal-1). This factor is also known as stem cell leukemia (SCL), which we will use in this text (Fig. 1). SCL is a basic helix–loop–helix transcription factor expressed in embryonic and extraembryonic mesoderm as well as other sites [8]. Expression of this factor is induced by BMP4 and is lost in the ventral region of Xenopus embryos devoid of BMP4 [9]. These results suggest that SCL is under the control or downstream of BMP4. Owing to the temporal and spatial expression and the proposed relationship with BMP4, SCL is regarded as an early marker of hematopoietic tissue. Further specifying SCL as a factor involved in early hematopoiesis, its expression was also found in erythroid, mast, and megakaryocytic cells. The expression pattern mimics that of another erythroid-specific marker, GATA1, which will be discussed later [10]. Shevdasani et al. have shown that when SCL is deleted in mice, all hematopoietic lineages, including embryonic erythrocytes, adult red cells, myeloid cells, megakaryocytes, mast cells, and both T- and B-lymphoid cells are affected [11]. Since the lineages being affected are located in the site of embryonic hematopoiesis and all lineages subsequent to HSC expansion are affected, SCL has been widely accepted as a factor involved in early hematopoiesis. Much like BMP4, SCL has been found in cell populations outside those involved in hematopoiesis, such as the vascular endothelium [8] and the developing brain [12]. Specifically, SCL has been shown to play an essential role in vascular development by mouse gene knockout studies [13].

258

Marty et al.

Figure 1 Genetics of hematopoietic development. The specification and differentiation to a hematopoietic fate is shown from top to bottom in this figure. The events necessary for the process (boxed) are often under the control of specific factors (bold). The function of these factors were revealed through a series of genetic studies. Disruption at the genetic level of these factors often results in a mutant phenotype. Studies in zebrafish have assisted in supporting the roles of these factors in hematopoiesis. Through genetic and transgenic approaches, scientists have generated mutants expressing defects in the hematopoietic lineage. Shown in italics are some of the zebrafish hematopoietic mutants characterized to date, which are discussed in this chapter.

Hematopoiesis, like most developmental pathways, is extremely complex and requires many factors acting in concert to function properly. These factors do not always function in a sequential order, but they often work together simultaneously and in the same locations. This idea led investigators to search for a protein that may assist SCL in carrying out its function. A potential candidate was soon revealed. Rbtn-2, another T-lymphoid acute leukemia oncoprotein, was found in the nucleus of erythroid cells, whereas a loss of this gene in mice resulted in a similar phenotype as seen in SCL mutants. When Rbtn-2 function was re-

Zebrafish: A Developmental and Genetic Model

259

moved, embryonic erythropoiesis was severely depleted to levels similar to the SCL model [14]. Thus, these two factors are regarded as essential factors in the specification of early embryonic hematopoiesis (see Fig. 1). To establish fully a hematopoietic system, the pool of HSCs must undergo both expansion and maintenance of the different hematopoietic cell lineages created from the original pool of HSCs. Several candidates have been implicated in these two processes, and two of these factors are discussed here. GATA2 is a member of the GATA family of transcription factors, all of which bind a specific DNA motif T/A (GATA) A/G in the promoters or enhancers of many genes and contain a conserved DNA binding region in two zinc fingers. GATA2 mRNA levels were found to be highest in hematopoietic progenitor cells while also correlating with terminal differentiation of these cells. Because of its specific expression in progenitor cells, GATA2 was proposed to play a role in the expansion and maintenance of HSCs. The idea that GATA2 is an early factor in hematopoiesis was strengthened by the observation that GATA2 is expressed in the earliest sites of hematopoiesis in Xenopus [15]. The definitive proof came from the studies showing that GATA2 null mice can generate hematopoietic progenitors, but fail to expand and/or maintain HSCs [16]. These results placed GATA2’s function downstream of SCL and Rbtn-2, such that it plays a role after the specification of HSCs to a hematopoietic fate. c-myb Is a proto-oncogene product expressed in high levels in hematopoietic progenitors, and its expression decreases as cells differentiate. Like GATA2, c-myb expression matches the time and place already specified for HSC expansion and maintenance. When c-myb was deleted in mice, impairment in the expansion and/or maintenance of the HSCs or progenitor cells was observed. Further molecular studies revealed that c-myb upregulates the expression of proliferationpromoting genes and inhibits the transcription of differentiation-promoting genes [17]. Taken together, both GATA2 and c-myb regulate the second step of hematopoiesis concerning the expansion/maintenance of HSCs or progenitor cells. The next step is to induce the cells to select a specific cell lineage and then mature in that lineage. Extensive research was performed to identify and characterize factors which regulate the process of lineage specification. A handful of factors that fit the description have been found, including PU.1, bmi-1, AML-1, GATA1, EKLF, NF-E2, Ikaros, and Pax-5. We will limit our discussion to one factor, GATA1 and its role in the specification of the erythroid lineage. GATA1 was the first GATA family member isolated, and it binds to the T/ A (GATA) A/G cis-acting elements in the promoters and enhancers of erythroidspecific genes to activate terminal differentiation [18,19]. GATA1 was found to be expressed in the ventral mesoderm of the yolk sac and fetal liver in mice, suggesting a possible role in both primitive and definitive erythropoiesis. Specifically, GATA1 was identified in erythroid, mast, eosinophilic, and megakaryo-

260

Marty et al.

cytic cells [20,21], and removal of GATA1 function in mice led to an ablation of primitive erythropoiesis. Although definitive erythropoiesis was initiated, maturation failed to occur [22]. The initiation of definitive hematopoiesis was accompanied by an increase in GATA2 expression. Since GATA1 is required for normal differentiation of erythroid cells and other GATA-binding factors cannot compensate for loss of GATA1 function [23], it has been established that some factors may cooperate in early, but not late, hematopoiesis/erythropoiesis. Upon GATA1 overexpression, an increase in erythroid cells along with a decrease in myeloid cells was observed. This revealed that some transcription factors increase a certain lineage at the expense of another lineage. In this case, GATA1 was found to be able to increase the expansion of the erythroid lineage while discouraging maturation of the myeloid lineage [24]. GATA1, like many other factors, does not act alone. A multiple zinc finger protein, friend of GATA1 (FOG), was identified to interact with GATA1. FOG expression was found to overlap that of GATA1 in both erythroid and megakaryocytes. It synergizes with GATA1 in activating transcription in hematopoietic-specific regulatory regions in cell culture. When FOG was deleted in mice, a partial blockage of erythroid maturation was observed, demonstrating the synergistic effect of FOG and GATA1 in erythropoiesis [25]. To conclude this section on factors involved in hematopoiesis, one should consider that these factors might have several functions at different stages and in different processes of embryological development. A schematic diagram of these factors and their hierarchy to specify and induce differentiation to a hematopoietic fate is shown in Figure 1. Although critical roles have been established for these factors through mouse genetic studies, it is likely that many more factors are yet to be identified. As discussed below, identification of novel hematopoietic genes through genetic screens and expression studies in zebrafish will enhance our understanding of the genetic mechanisms underlying developmental hematopoiesis.

II. Background Information on Zebrafish Up to this point, we have primarily discussed the genetics of vertebrate hematopoiesis using the mouse as a model system. The main reason for the popularity of the mouse stems from a type of reverse genetics where investigators have the ability to introduce a mutation, known as knockout technology, into a specific gene thought to play a role in hematopoiesis. Knockout technology has led to the characterization of many of the factors involved in vertebrate hematopoiesis, as discussed earlier in this chapter. With this in mind, the mouse system will continue to be a useful model to study vertebrate hematopoiesis through reverse

Zebrafish: A Developmental and Genetic Model

261

genetics. In the last decade, the zebrafish has risen to the front as an important model implementing a forward genetic approach to study vertebrate development. A. Overall Development and Useful Characteristics

Although mouse knockout technology has led to many important discoveries, the process is confined to studying only previously identified genes. Another restraint of the mouse system lies in the small litter size and internal development of the mouse embryos. Restraints such as these are absent in the zebrafish. The usefulness of the zebrafish was first recognized by Streisinger about 20 years ago at the University of Oregon [26]. Streisinger analyzed these freshwater fish and noted several characteristics he felt would facilitate embryological and developmental experimentation. The first obvious characteristic is their small size. Adult zebrafish are about 1 in. in length (Fig. 2A), which allows for large numbers to be maintained in a small area. Also, unlike the mouse, female zebrafish can lay up to several hundred eggs per mating. The immense production of eggs allows investigators to produce large amounts of embryos for visual and genetic screening in a short amount of time, overcoming the limitations of mice due to the mouse embryo’s intrauterine growth and small litter sizes [27]. The zebrafish embryos themselves are fertilized externally, which is an important attribute exploited by researchers, as we will see later in the development of haploid embryo. Zebrafish also allow for the freezing of sperm for later analysis and for maintaining certain lines. Embryos are transparent, revealing the simple organization of the zebrafish’s organ systems, lending themselves as excellent candidates for lineage tracing and transplant experiments [28]. The clarity of the embryo also provides easy access to all developmental stages, which in turn allows for easy observation of cell movement and formation of domains during gastrulation. Early development of the zebrafish embryos occurs extremely rapidly. A beating heart and circulating erythrocytes become visible by 24 hr postfertilization (hpf). At the same time, somites, brain, eyes, ears, neural tube, and notochord are also well formed. In the following days, other organs become fully differentiated and embryos begin to swim around 6 days post fertilization (dpf) [29]. The ease at which zebrafish can be observed also allows for time-lapse studies to monitor cell division and patterning. In addition to its phenotypic qualities, the zebrafish has a number of genotypic qualities that make it a good model to study development. The chromosomal structure of the zebrafish is similar to the human’s [30]. The syntenic structure between the two allows for positional cloning of zebrafish genes utilizing information from the Human Genome Project [31]. This cross talk between species has been termed ‘‘genome ping-ponging.’’ Thus, the genetic information of one organism can be used to facilitate an understanding of another.

262

Marty et al.

Figure 2 Transgenic analysis of zebrafish. (A) An adult wild type adult zebrafish. Through transgenic analysis, specific populations of cells can be marked and traced in live zebrafish throughout development. (B) A 20 hpf transgenic zebrafish carrying the GFP reporter gene driven by the putative GATA-1 promoter. The arrow indicates GFP expression in the intermediate cell mass. (C) A six day transgenic zebrafish carrying the GFP reporter gene driven by the putative rag-1 promoter. The arrow indicates GFP expression in the thymus.

Although the zebrafish model has many advantages, it does have some drawbacks. Unlike the mouse model, isolation of embryonic stem cells for the purpose of gene knockout is not yet possible in the zebrafish. The homologous recombination machinery is present, but investigators have yet to isolate or culture the embryonic stem cells. Genetic redundancy in the zebrafish may also complicate things. Since the zebrafish appears to have an extra genome duplication, it is possible that some of the homologous genes have assumed only a subset of function that is seen in the mouse or human. Nevertheless, the ability to exploit the genetics and embryology of the zebrafish model has already led to important discoveries concerning the hematopoietic development of vertebrates. Many of these discoveries will be covered in the sections discussed below.

Zebrafish: A Developmental and Genetic Model

263

B. Hematopoietic Development in Zebrafish

Many of the same factors discussed earlier also play a role in zebrafish hematopoiesis. Investigators have studied the expression of these factors and subsequently used them as molecular markers in zebrafish developmental stages and in different adult tissues. These studies have been key in uncovering the hematopoietic system in the zebrafish and providing insight into vertebrate hematopoietic development. Zebrafish hematopoietic development occurs in successive waves of primitive and definitive hematopoiesis in a manner similar to higher vertebrates. Hematopoiesis begins with the induction and specification of ventral mesoderm to a hematopoietic fate before the start of primitive hematopoiesis. As in other vertebrates, specific factors have been suggested to play a role in the induction process. The discovery of GATA2 and BMP2/4 expression at the site of ventral mesoderm induction at about 6 hpf is the first of the similarities [32]. Further importance of BMP members in the induction of ventral mesoderm was elucidated by two other groups [33,34]. Kishimoto’s group demonstrated that a mutation in the bmp2A gene resulted in decreased embryonic blood production, whereas overexpression of bmp2A RNA led to an increase in embryonic blood development. These experiments established a genetic role of BMP-2 in blood cell formation. As development continues, embryonic blood cells begin to commit to a specific lineage between 12 and 18 hpf in the zebrafish. At this time, hematopoietic cells begin to express GATA1 [35]. Unlike the higher vertebrates, zebrafish do not contain yolk sac blood islands but rather embryonic hematopoiesis occurs in an intraembryonic region of the tail bud called the intermediate cell mass (ICM) [36]. The ICM is formed by the convergence of two paraxial stripes of mesoderm that arise during gastrulation. Cells in the ICM appear as round, proerythroblast-like cells and express a number of hematopoietic markers, including Ikaros, GATA2, LMO2, GATA1, c-myb, globin, and SCL. The expression of the early markers provides evidence of early primitive hematopoiesis, whereas the later markers reveal lineage commitment within the ICM. Thus, the ICM has been characterized as an early site of both primitive and definitive hematopoiesis [27]. Examination of the differential expression of several hematopoietic transcription factors in the ICM suggests that LMO2 and GATA2, but not GATA1, c-myb, or any globins are present in the posterior region of the ICM. This led investigators to believe that this region of the ICM was giving rise to an undifferentiated hematopoietic stem cell population. Cells of the ICM enter circulation between 24 and 30 hpf, marking the first anterior migration of the blood cells, another characteristic of zebrafish hematopoiesis [35]. During this time, the ICM disappears, but expression of hematopoietic markers persists for another 3 days in a region posterior to the ICM and ventral to the axial vein [37,38]. The heart begins to beat at approximately 24

264

Marty et al.

hpf, and primitive blood cells accumulate in a region below the dorsal aorta [35]. It was believed that a few pluripotent hematopoietic cells exist at locations other than the ICM; namely, the ventral wall of the aorta and in the region posterior to the ICM [37,38]. It was also suggested that heart endocardiatic cells serve as transient hematopoietic sites [39]. These data supported the idea of multiple embryonic locations of pluripotent hematopoietic cells. However, Willet et al. was not able to find any hematopoietic progenitor cells in the heart [37]. From the regions of the dorsal aorta and tail, hematopoietic progenitors migrate to an area in the pronephros, the zebrafish kidney. The pronephros of zebrafish gives rise to both the head and trunk kidney present in the adult. Thus, it serves as an analogous site to the bone marrow in higher vertebrates, the site of adult hematopoiesis. Although the kidney is well developed at 72 hpf, the first hematopoietic cells do not appear until 96 hpf. At this time, erythrocytes, lymphocytes, and granulocytes are all detectable, establishing the kidney as the site of immature blood cell differentiation. In the zebrafish, some mature as well as immature red blood cells have also been found in the spleen; however, it is considered as a recycling site of blood cells rather than a site of blood cell production. The kidney’s hematopoietic role is further supported by the expression of multiple hematopoietic markers such as GATA1, c-myb, GATA2, and SCL in the tissue comprising both the head and trunk kidneys. III. Techniques Used to Study Zebrafish Development The similarities in the development of the zebrafish and other vertebrates suggest that the mechanisms underlying vertebrate development are shared by many animal systems. From research conducted thus far, many developmentally important genes have been identified, but gaps are still present. In order to fill these gaps, investigators have begun to employ genetic and transgenic-based analyses to identify new genes and their respective roles in zebrafish development. A. Genetic Analysis

Owing to the ease of manipulation and the ability to view the early development of the zebrafish embryo, the generation of mutants has become a popular way to investigate development of the zebrafish. To date, saturation mutagenesis screens for embryonic lethal mutations have uncovered many genes responsible for controlling development of vertebrates. Solnica-Krezel et al. stated three requirements to make saturation mutagenesis screens feasible (40). First, a mutagen has to be available that is both efficient and known to cause predominant intragenic lesions. This minimizes the number of genomes to be screened in order to reach saturation levels. Second, efficient transmission of the mutation to subsequent generations has to be feasible without seasonal variations and in a reason-

Zebrafish: A Developmental and Genetic Model

265

ably short amount of time. And finally, convenient selection or screening procedures for phenotypes has to be available. Saturation mutagenesis screens have already been carried out in Drosophila (41) and C. elegans (42). Forward mutagenesis screens in mice were found to be very difficult, because identification of embryonic lethal mutations was complicated owing to the intrauterine mode of embryogenesis. This is not a problem in zebrafish owing to their external fertilization. To date, chemical mutagens have been the mutagens of choice for the zebrafish. Recently, insertional mutagenesis using pseudotyped retroviral vectors has also been utilized to induce mutations in the genome. Once the mutations have been introduced, the next step lies in the identification and cloning of the mutated gene. Generation of Mutants N-Ethyl-N-Nitrosourea

One way to induce mutations within the zebrafish genome is through the use of the alkylating agent N-ethyl-N-nitrosourea (ENU). Grunwald and Streisinger first showed that zebrafish spermatozoa could be mutagenized by ENU in vitro [43]. The F1 individuals produced by their method frequently tended to be mosaic for the induced mutation. It was later discovered that by mutagenizing premeiotic germ cells, specifically the spermatogonia of male zebrafish, the mutations were found to be nonmosaic heterozygotes. The mutagenesis is normally carried out by placing male fish into a solution containing ENU. The treated males are then crossed to a nontreated female, giving rise to the F1 progeny. In order to measure the efficiency of the methodology, Solnica-Krezel et al. used females containing one or several pigment mutations to determine the frequency of mutagenesis at a specific loci (40). Thus, if the F1 progeny exhibited a change in pigment, it was thought that the change in phenotype would be indicative of an induced mutation by ENU. They also checked the frequency of embryonic mutations by visual inspection of the progeny of the mutagenized families by what is known as an F2 screen. The F2 screen is a type of visual screen where the progeny obtained from the mutagenized F2 progeny are inbred. The offspring from this cross are then visually inspected at 50–100⫻ magnification for any phenotypic defects or for embryonic lethality. Gamma Irradiation

Gamma irradiation of zebrafish sperm is another way to generate zebrafish mutants. ENU treatment and gamma irradiation are both referred to as a forward genetic approaches to identify genes involved in vertebrate development. Forward genetics is the term used when one first identifies a phenotype and then attempts to identify the gene responsible as opposed to reverse genetics, where one already knows the gene and attempts to find the function by altering or induc-

266

Marty et al.

ing a mutation in that gene. Gamma irradiation techniques have been previously studied in Drosophila [44–46], the mouse [47–49], and the madakafish [50,51]. Walker et al. was the first to use gamma irradiation on zebrafish pregonial germ cells [52]. Gamma rays were later shown to be an efficient mutagen that can produce a specific locus mutational rate of about 1:100 [53]. The mechanism of gamma ray mutagenesis often results in a chromosomal deletion or translocation, whereas ENU causes single base point mutations. Because a high frequency of deletions can be induced by gamma irradiation, it is possible to identify deletion alleles within a specific gene. This is normally carried out by first collecting sperm from male zebrafish. The sperm is then subjected to ultraviolet irradiation. Mutagenized females are then squeezed to collect their eggs. The irradiated sperm is added to the eggs and activated in an aqueous solution. The irradiated sperm contain an inactive genome and do not contribute to the offspring’s genome, but they are still capable of fertilizing the eggs. The fertilized eggs are then considered haploid. These haploid embryos are capable of completing embryogenesis, yet they arrest at the larval stage. Any visible phenotypes prior to this stage can therefore be recognized in a single generation. Fritz et al. further utilized a multiplex polymerase chain reaction (PCR) process to identify mutations affecting already cloned genes in the zebrafish [54]. Multiplex PCR involves the use of many primer pairs specific to previously cloned genes in the zebrafish to amplify genomic DNA taken from each haploid offspring. If a deletion occurs in a region containing the gene of interest, there would be no PCR product amplified from the DNA. When using this method, one should be aware that a deletion may affect many genes in the region, and a phenotype observed needs to be confirmed by examining additional mutants containing additional mutations in the specific gene. Insertional Mutagenesis

Generation of mutants via chemical mutagenesis can produce large amounts of mutants, but cloning of the mutated genes is often difficult. Identification of the mutated genes often requires positional cloning methods, such as those discussed later, which are very expensive and laborious owing to the size of the zebrafish genome (⬃1.7 ⫻ 10 9 bp). This led investigators to look for an improved way to generate mutants with subsequent fast and effective ways to clone the mutated gene. It had already been shown that integration of proviral DNA into mice can disrupt essential genes [55], so the next step was to attempt this in the zebrafish. Burns et al. developed a murine leukemia virus–derived vector capable of efficiently introducing foreign genes into both mammalian and nonmammalian cells [56]. The vector Burns developed contained a Moloney murine leukemia virusbased genome (MoMLV) surrounded by an envelope containing the glycoprotein of the vesicular stomatitis virus (VSV). The construct of the VSV glycoprotein allows for a broad range of infection, characteristic of the VSV upon entry into

Zebrafish: A Developmental and Genetic Model

267

the cell. Lin and colleagues showed that the proviral vector can indeed integrate into the zebrafish genome and transmit through the germline by injecting the virus into blastula-stage zebrafish embryos [57]. Higher titers of viral stocks have been produced and used to generate insertion-induced mutations [58]. Because known sequences of the proviral insert can be used to amplify the region surrounding the insert, the proviral vector sequence serves as a molecular tag to assist in isolating the flanking genomic fragments containing the disrupted gene [59,60]. Cloning of the Mutated Genes

Once a mutation is generated, the task becomes centered on the identification of the altered gene. As mentioned above, if a mutation is caused by a DNA insertion, cloning of the mutated gene is relatively easy, because the inserted DNA serves as a molecular tag. However, most of the existing zebrafish mutations are induced by ENU. The identification of the genes responsible for these mutations often relies on either a candidate gene or a positional cloning approach. Candidate Gene Approach

Expression patterns, determined by in situ analysis, can provide clues as to which candidate genes might overlap the mutated gene’s expected function [61]. For instance, it is reasonable to assume that a phenotype affecting hematopoiesis is caused by a mutation in a gene that is exclusively or highly expressed in blood cells. Once a pool of candidates is obtained, the first step is to utilize linkage analysis to test for colocalization of the candidate gene to the mutated locus. If the candidate maps to the mutation, it can be used to isolate the corresponding genomic locus sequence from the mutant locus to determine if a mutation is indeed present in that gene. A further extensive linkage analysis is then performed to confirm that no recombination has occurred between the phenotype and mutation. This would establish a complete linkage between the identified mutation and the noted phenotype. The final step in confirming a candidate gene would consist of phenotypic rescue upon germline transgenic expression of the candidate gene [62]. One drawback to this approach is the fact that many zebrafish genes remain to be isolated. To rectify this problem, a genomics initiative is producing 100,000 ESTs and their mapping information from zebrafish. This should further facilitate the identification of genes responsible for mutant phenotype through a candidate gene approach. Positional Cloning

The positional cloning approach is achieved through mapping the mutated locus with tightly linked markers, such as simple sequence length polymorphisms (SSLPs), random amplified polymorphic DNA (RAPD), and amplified fragment length polymorphism (AFLPs) as well as cloned genes. The markers are DNA

268

Marty et al.

sequences which are physically linked to the mutation as defined by infrequent recombination between the two positions over many meioses. Since the recombination frequency and the distance between markers are directly correlated, a closer marker would be more tightly linked. Therefore, a more tightly linked marker would recombine with the mutation at a lower frequency than a more distant marker. The zebrafish system has an advantage in developing a highresolution map because of its fecundity which allows for the analysis of recombinations of many meioses. The mutant strain is mated with another mapping strain, and their progeny are analyzed for linkage between these markers and the mutation. Once linkage within 1 centimorgan, roughly 600 kilobases, is established, the next step is screening large insert genomic libraries contained in yeast artificial chromosomes (YACs), bacterial artificial chromosomes (BACs), or P1 artificial chromosomes (PACs). After obtaining positive clones, the crucial interval containing the mutated locus is better defined through sequence walking. Then a candidate gene can be identified by analyzing the sequence or by screening a cDNA library with the positive clone from the large insert library. Again, in situ analysis can support a gene’s candidacy if its expression overlaps the expected mutated gene’s function. Once a mutation is found within the candidate gene by direct sequence comparison of the mutated allele and the corresponding wildtype locus, a phenotypic rescue would support the identification of the mutated gene [62]. B. Transgenic Analysis

As with most animal models, transgenesis remains a highly important method for studying various aspects of development. In order to produce transgenic lines in the zebrafish, retroviral infection or DNA microinjection can be used. Although retroviral vectors have been shown to provide a higher rate of germline integration, the cloning capacity of the vector for exogenous DNA is small. Therefore, they have been used more for insertional mutagenesis rather than transgenic analysis. The most commonly used method for transgenesis is introducing naked DNA into the embryo by microinjection. The exogenous DNA can be in the form of circular plasmids, linearized DNA fragments, BACs, or PACs [63]. Although plasmid vectors often provide enough sequence for proper transgenic expression, the larger regions contained in BACs or PACs are sometimes needed to obtain proper expression of the transgene [64]. As with other animal models, transgenics is used to study the effect of exogenous DNA on the development of the embryo. By using different promoters, the exogenous DNA can be ubiquitously overexpressed or can be expressed in the correct temporal and spatial manner as determined by development. The exogenous DNA itself can consist of various configurations including wild-type sequences or dominant negative forms of a gene of interest. One advantage for transgenesis in the zebrafish lies in its transparent

Zebrafish: A Developmental and Genetic Model

269

body which allows for in vivo analysis of the transgene expression by using the green fluorescent protein (GFP) reporter gene (Fig. 2B,C). For example, the development of blood cells can be followed from the appearance of the earliest hematopoietic progenitors to the adult stage. The continuing advancement of transgenic technology in the zebrafish provides for an exciting look at vertebrate development. IV. Hematopoietic Disorders in Zebrafish Many factors necessary for proper hematopoietic development in vertebrates have now been identified. The identification of these factors has often come through analyzing the phenotypic changes associated with induced mutations and/or knockouts. The following section will discuss some of the hematopoietic mutations in zebrafish. A. Prehematopoietic Mutations

In order to better understand hematopoiesis, the earliest embryo should be thought of as a group of undifferentiated cells with no predetermined fate. To establish the body axis, synergistic and antagonistic interactions of different signals are required. Once the axis patterns are established, the cells begin to differentiate into specific lineages. For example, the hematopoietic lineages are derived from the ventral mesoderm, which is induced by ventralizing factors like the bone morphogenetic proteins (BMPs) and Wnts. The definition of the boundaries of the ventral mesoderm is a direct result of the antagonism of dorsalizing factors like Chordin, Noggin, and Follistatin with the aforementioned ventralizing factors [33]. Mutations in some of these early embryonic patterning genes would tend to disrupt not only the body plan but also downstream events such as hematopoiesis. Swirl

One example of a prehematopoietic mutation affecting both body plan and hematopoiesis is the zebrafish mutant swirl. This embryonic lethal mutant was generated through an ENU mutagenesis screen. Swirl Mutants have a highly dorsalized phenotype with expanded notochord and somites, whereas lacking ventral structures such as blood and nephros. The swirl mutation is caused by either of two changes in the zBMP2A coding sequence: (1) a lengthening of the amino acid sequence by six amino acids, or (2) by altering the protein’s maturation domain. Both of these apparently lead to a nonfunctional BMP2A. The mutation can be rescued by injection of wild-type zBMP2A mRNA, is non–cell autonomous, and has no effect on viability and fertility as shown by the ability of two rescued mutants to produce offspring. In addition, no maternal effect of the swirl mutation

270

Marty et al.

was found in mesoderm induction. Gene expression analysis of the swirl mutants using the hematopoietic marker GATA1 and the cardiac markers Nkx.25 and Nkx.27 shows a defect in the formation of both blood and cardiac cells. These studies indicate that BMP2A is absolutely required for hematopoiesis and development of ventral structures but has no role in premidblastula development, viability, or fertility [33]. Spadetail

The zebrafish mutant spadetail was originally discovered through a haploid genetic screen as a mutant which failed to form somites in the trunk [65]. Additional spadetail alleles were identified through large-scale ENU mutagenesis [66] and gamma-irradiation screens [67]. It was later discovered that spadetail mutants also have defects in primitive and definitive hematopoiesis [38]. Spadetail has been shown to have a mutation in a T-box transcription factor necessary for directing convergence of the lateral mesoderm, thus disrupting normal trunk mesoderm formation. Although all the alleles result in the loss of functional protein, some do produce nonfunctional mRNA transcripts which can be detected for expression analysis. Convergence of the lateral plate mesoderm is necessary for the formation of the ICM, the site of early zebrafish hematopoiesis. Through dominant negative receptor experiments, spadetail has been shown to be dependent upon fibroblast growth factor (FGF) signaling for maintenance but not activation [67]. Downstream of spadetail is the paraxial protocadherin (PAPC), a cell adhesion molecule necessary for proper mesodermal convergence. Embryos expressing a dominant negative form of PAPC produced the spadetail phenotype [68]. In spadetail mutants, SCL expression is also greatly reduced, suggesting a loss of the stem cell population [69]. Analysis of LMO2 and GATA2 expression shows that hematopoietic progenitors are present in similar levels as compared to wild type. However, LMO2 and GATA2 patterning are disrupted owing to the absence of the convergence of the paraxial mesoderm in early development. A decrease of GATA1 and c-myb expression levels further correlates with the lack of blood phenotype of the spadetail mutants [38]. Taken together, the disruption of the somite patterning in the trunk and the lack of blood production have led investigators to propose that spadetail acts as a synergistic signal necessary for proper spatial patterning of both dorsal somites and ventral blood tissues [70]. Cloche

The zebrafish mutant cloche, named for its bell-shaped heart, was originally identified as a spontaneous mutant. Cloche Has a normal body plan but lacks embryonic erythrocytes and endothelial derived endocardium [71]. Gene expression analysis showed that hematopoietic markers, including LMO2 and GATA2, are

Zebrafish: A Developmental and Genetic Model

271

reduced in cloche mutants as compared to wild type [38]. Expression of flk-1, a marker for endothelial cells, is only found in the lower trunk and tail region. Interestingly, GATA1 expression is only detectable in the flk-1–positive cells [72]. The expression of c-myb, a marker for definitive hematopoiesis, is either lost or severely reduced at 24 hpf and is completely absent by 48 hpf [38]. The expression of SCL is also essentially absent in cloche mutants. The cloche phenotype can be rescued through forced expression of SCL, leading to the conclusion that the cloche gene functions upstream of SCL [39]. By cell transplantation experiments, the cloche mutation has been shown to be cell autonomous with respect to the defect of endocardium. However, mutant cells placed in a wildtype host reveal that the mutation is non–cell autonomous for hematopoietic differentiation by their expression of GATA1. Cloche Was also found to be cell autonomous for hematopoietic proliferation, since mutant cells did not contribute to the mature red blood cells [73]. The cloche mutation supports the hypothesis of common progenitors shared between endothelial and hematopoietic lineages. This is consistent with the observation that the ventral marginal zone of the zebrafish is multipotent and can give rise to both endothelial and hematopoietic cells as well as other mesodermal lineages [74]. B. Hematopoietic Mutations

Bloodless Mutants

The zebrafish mutant moonshine was generated in an ENU large-scale chemical mutagenesis screen and has a bloodless phenotype. The mutation acquired its name because it produces an increased number of iridophores that reflect light in the dark microscopic field [75]. Moonshine appears to be defective in the signals needed for proper proliferation of embryonic proerythroblasts and differentiation of definitive erythrocytes. The moonshine mutation was initially correlated to eight different alleles with variable blood defects. In addition, all have defects in fin morphology, pigment cell proliferation, and migration, whereas the melanocytes appear normal. Analysis of GATA2 expression in the moonshine mutants reveals wild-type quantities and patterning, whereas GATA1 expression is reduced depending on the strength of the allele. The weaker moonshine alleles are able to produce roughly 50–100 circulating blood cells by day 5 compared to wild-type levels of 1000–3000. Morphological analysis of these embryos reveals that the circulating blood in the moonshine mutants consists of proerythroblasts in contrast to mature circulating erythrocytes in wild type embryos. The proerythroblasts found in moonshine mutants also lack hemoglobin expression as shown by o-dianisidine staining. The phenotype of the moonshine mutants appears to resemble the mouse models W, a null c-kit mutation, and Steel, a null c-kit ligand mutation. Both of these mutations produce defects in hematopoiesis, melano-

272

Marty et al.

cytes, and germ cells [76]. Although a c-kit mutation has been identified in the zebrafish [77], the molecular nature of the moonshine mutation has yet to be determined. Mutants with Severe Blood Deficiency

The next set of mutants shows phenotypes associated with defects at later stages of hematopoiesis resulting in severe anemia. The mutants discussed here and in the following sections were named after different wines owing to their deficiencies in blood cell formation. They were all generated by ENU mutagenesis. Chablis (two alleles), frascati (four alleles), merlot (two alleles), retsina (two alleles), thunderbird (one allele), riesling (one allele), cabernet (one allele), and grenache (one allele) all have normal initial blood cell counts but start to show a drastic reduction in the circulating erythroid cells at embryonic day 5. The blood is pigmented normally and hemoglobin expression levels correlate with the numbers of erythroid cells. Except for thunderbird (tbr ty118b ), all of these recessive mutations are embryonically lethal within the first 2 weeks of larval life. Here we concentrate on two of these mutants, frascati and retsina, which differ slightly in their erythroid morphological defects. Both express wild-type levels of GATA1 and GATA2 at 1 dpf, showing normal primitive hematopoietic differentiation and proliferation. Depending on the strength of the allele, the mutants begin to display defects in erythropoiesis shortly thereafter. The stronger frascati alleles (frs tg221 and frs tq223, exhibit a phenotype upon hatching, whereas the weaker alleles (frs tg280a and frs tm130d ) exhibit a defect around 4 dpf of development. At 2 dpf of development, the frs tq223 mutants express some hemoglobin and have larger blood cells with basophilic cytoplasms and fragmented nuclei that are characteristic of erythroblasts. This is very different from 2 dpf wild-type embryos, which typically have circulating cells consisting of more mature erythrocytes with more condensed nuclei. Day 3 pf retsina embryos differ in another way, as they tend to have larger cells with large open nuclei that are reminiscent of immature proerythroblasts. The retsina embryos, like the frascati mutants, also have a reduced hemoglobin expression level. Rather than a direct defect in globin expression, the phenotypes of the mutations seem to suggest a defect in erythroid cell differentiation and proliferation which results in a reduction of hemoglobin expression [76]. Hypochromic Mutants

The next set of mutations, chardonnay (one allele), chianti (one allele), pinotage (one allele), sauternes (two alleles), weibherbst (two alleles), and zinfandel (one allele), exhibits hypochromic anemia phenotypes that are likely due to defects in hemoglobin production. Defects in hemoglobin production can stem from a

Zebrafish: A Developmental and Genetic Model

273

decrease in heme, globin, or iron, possibly due to insufficient diet, defective iron transport, or blood loss [78]. Both recessive and embryonically lethal, the two weibherbst alleles, weh th238 and weh tp85c, first begin to exhibit hypochromic blood around the hatching stage. As the mutant embryos develop, they show a 50% reduction in blood count. At 2 dpf, the blood cells morphologically resemble the proerythroblasts with large open nuclei compared to the wild-type erythroblasts with condensing nuclei. Early o-dianisidine staining reveals a loss of hemoglobin expression compared to wild-type expression levels. At 2 dpf, the zinfandel mutants have oval, larger, paler blood cells with less condensed nuclei. By 3 dpf, the dominant viable mutants show hypochromic erythroid cells and a lower blood count accompanied with almost no detectable hemoglobin expression [76]. The zinfandel mutation has been linked to the globin gene locus and possibly resembles thalassemia in humans [78]. The most well-defined hypochromic mutation is sauternes, which represents an animal model of congenital sideroblastic anemia (CSA). The sauternes phenotype is caused by a mutation in the erythroid specific isoform of δ-aminolevulinate synthase (ALAS2), the first of eight enzymes in the heme biosynthetic pathway. Injection of a wild-type ALAS2 cDNA expression plasmid allowed for phenotypic rescue. Humans with mutations in ALAS2 develop CSA and tend to have similar phenotypes to that of the sauternes zebrafish mutant. Two alleles of the sauternes have been identified, the lethal sau tb223 allele and the viable sau ty121 allele. At about 33 hpf of development, the blood is visibly hypochromic but has a wild-type blood count. By 48–72 hpf, the red blood cell numbers drop to that of 25–50% of wild type. At 50 hpf, the blood cells are morphologically immature compared to wild type. Hemoglobin expression is undetectable at 28 hpf but can be detected at very low levels later in development. Expression of the three alphaglobins, αe1, αe2, αe3, and the three beta-globins, βe1, βe2, βe3, are initiated normally. However, unlike wild-type expression, βe2 globin expression persists after 72 hpf in the mutants. This mutant has a defect in heme synthesis, which leads to abnormal regulation of globin expression. This observation supports that heme can indeed influence the expression of globin genes. At 72 hpf, GATA1 expression in the mutants also persists in circulating blood cells and in the posterior region of the tail long after its downregulation in wild-type embryos. The sauternes adults (sau ty121 ) are anemic but have five times more immature erythroid cells in circulation and two times more erythroid precursors in the adult kidney as compared to wild type. These observations suggest that a defect in the heme pathway can delay differentiation of erythroid cells. Although the phenotypes for the zebrafish sauternes mutation and human CSA do resemble each other, there are some differences which may be due to divergence of the two species [78]. Therefore, the sauternes model does provide a reasonably accurate animal model to study CSA.

274

Marty et al. Photosensitive Mutants

Porphyrias are a group of inherited or acquired disorders with defects in the heme biosynthetic pathway. Yquem and freixenet are two zebrafish mutants which have defects later in the heme biosynthesis pathway, leading to porphyria rather than CSA. Both mutants have autofluorescent blood cells that are rapidly photoablated upon exposure to light. The most prevalent porphyria is porphyria cutanea tarda (PCT), which is characterized by photosensitive skin and excessive excretion of uroporphyrin and 7-decarboxylate porphyrin in the urine. This autosomal dominant human disease can be inherited or acquired and has reduced uroporphyrinogen decarboxylase (UROD) activity. Uroporhyrinogen is the fifth enzyme in the heme biosynthetic pathway causing the decarboxylation of the four acetates of uroporphyrinogen I and III to form coproporphyrinogen I and III, respectively. Hepatoerythropoietic porphyria (HEP) is an autosomal recessive human disease with both copies of UROD mutated resulting in a more severe phenotype. The yquem mutant has been shown to have a missense mutation in UROD causing enzyme deficiency. The mutant phenotype can be rescued by expressing a wildtype UROD cDNA transgene construct. Although the homozygous yquem mutant provides the first true animal model for HEP, the heterozygotes can be used as models for PCT, since the UROD activity is only about 67% of wild type [79]. The yquem mutant zebrafish should be useful for testing therapeutic drugs and studying pathogenesis of the disease. C. Hemophilia Model

Hemostasis, the physiological maintenance of the blood as a fluid in the circulatory system, is perturbed in hemophilia owing to reduced coagulation [80]. The current zebrafish hemophilia model is not the result of a genetic alteration; however, it could eventually lead to genetic studies of the in vivo events that lead to hemostasis. Blood coagulation occurs as the result of a chain of enzymatic reactions: Coagulation factor VIII activates factor X that cleaves prothrombin to yield thrombin, which then cleaves fibrinogen into the fibrin fibers [81]. Hemophilia is often caused by a deficiency of factor VIII leading to adhesion and lysis of thrombocytes upon contact with damaged tissues [80]. Studies have been done using a technique called thrombelastography to show that freshwater fish and humans have a similar coagulatory system [82]. In the zebrafish hemophilia model, copper chloride is used to induce hemophilia characterized by reduced coagulation [83]. Originally done in Orechromis mossambicus, copper was shown to either cause defective synthesis of coagulation factor VIII or the synthesis of a defective coagulation factor VIII, therefore leading to reduced factor X activity and less coagulation [80]. In this experimental model, the synthetic disease mirrors the natural mutation and provides an excellent background to test possible avenues of correcting the disease.

Zebrafish: A Developmental and Genetic Model V.

275

Conclusion

The zebrafish model system, although relatively young, has already proven to be a powerful model to study vertebrate hematopoietic development. Mutagenesis of the zebrafish genome has produced important insights into genes required for the development of cardiac, vascular and hematopoietic systems, including disease models of CSA and HEP. These studies have also provided evidence that these genes are highly conserved between zebrafish and human. By combining molecular, genetic, and transgenic-based analysis, researchers have begun to develop new screening assays that should further unravel the molecular steps leading to the formation of blood cells. Thus, studies in zebrafish should ultimately help us understand how hematopoietic development occurs in higher vertebrates. References 1. Orkin SH. Hematopoiesis: how does it happen? Curr Opin Cell Biol 1995; 7(6): 870–877. 2. Evans T. Developmental biology of hematopoiesis. Hematol Oncol Clin North Am 1997; 11(6):1115–1147. 3. Maeno M, Ong RC, Suzuki A, Ueno N, Kung HF. A truncated bone morphogenetic protein 4 receptor alters the fate of ventral mesoderm to dorsal mesoderm: roles of animal pole tissue in the development of ventral mesoderm. Proc Natl Acad Sci USA 1994; 91(22):10260–10264. 4. Amaya E, Musci TJ, Kirschner MW. Expression of a dominant negative mutant of the FGF receptor disrupts mesoderm formation in Xenopus embryos. Cell 1991; 66(2):257–270. 5. Dale L, Howes G, Price BM, Smith JC. Bone morphogenetic protein 4: a ventralizing factor in early Xenopus development. Development 1992; 115(2):573–585. 6. Graff JM, Thies RS, Song JJ, Celeste AJ, Melton DA. Studies with a Xenopus BMP receptor suggest that ventral mesoderm-inducing signals override dorsal signals in vivo. Cell 1994; 79(1):169–179. 7. Suzuki A, Thies RS, Yamaji N, Song JJ, Wozney JM, Murakami K, Ueno N. A truncated bone morphogenetic protein receptor affects dorsal-ventral patterning in the early Xenopus embryo. Proc Natl Acad Sci USA 1994; 91(22):10255–10259. 8. Kallianpur AR, Jordan JE, Brandt SJ. The SCL/TAL-1 gene is expressed in progenitors of both the hematopoietic and vascular systems during embryogenesis. Blood 1994; 83(5):1200-1208. 9. Mead PE, Kelley CM, Hahn PS, Piedad O, Zon LI. SCL specifies hematopoietic mesoderm in Xenopus embryos. Development 1998; 125(14):2611–2620. 10. Visvader J, Begley CG, Adams JM. Differential expression of the LYL, SCL and E2A helix-loop-helix genes within the hemopoietic system. Oncogene 1991; 6(2): 187–194. 11. Shivdasani RA, Mayer EL, Orkin SH. Absence of blood formation in mice lacking the T-cell leukaemia oncoprotein tal-1/SCL. Nature 1995; 373(6513):432–434.

276

Marty et al.

12. Green AR, Lints T, Visvader J, Harvey R, Begley CG. SCL is coexpressed with GATA-1 in hemopoietic cells but is also expressed in developing brain. Oncogene 1992; 7(4):653–660. 13. Visvader JE, Fujiwara Y, Orkin SH. Unsuspected role for the T-cell leukemia protein SCL/tal-1 in vascular development. Genes Dev 1998; 12(4):473–479. 14. Valge-Archer VE, Osada H, Warren AJ, Forster A, Li J, Baer R, Rabbitts TH. The LIM protein RBTN2 and the basic helixloop-helix protein TAL1 are present in a complex in erythroid cells. Proc Natl Acad Sci USA 1994; 91(18):8617–8621. 15. Walmsley ME, Guille MJ, Bertwistle D, Smith JC, Pizzey JA, Patient RK. Negative control of Xenopus GATA-2 by activin and noggin with eventual expression in precursors of the ventral blood islands. Development 1994; 120(9):2519–2529. 16. Tsai FY, Keller G, Kuo FC, Weiss M, Chen J, Rosenblatt M, Alt FW, Orkin SH. An early haematopoietic defect in mice lacking the transcription factor GATA-2. Nature 1994; 371(6494):221–226. 17. Weston KM. The myb genes. Semin Cancer Biol 1990; 1(6):371–382. 18. Evans T, Felsenfeld G. The erythroid-specific transcription factor Eryf1: a new finger protein. Cell 1989; 58(5):877–885. 19. Tsai SF, Martin DI, Zon LI, D’Andrea AD, Wong GG, Orkin SH. Cloning of cDNA for the major DNA-binding protein of the erythroid lineage through expression in mammalian cells. Nature 1989; 339(6224):446–451. 20. Kelley C, Yee K, Harland R, Zon LI. Ventral expression of GATA-1 and GATA2 in the Xenopus embryo defines induction of hematopoietic mesoderm. Dev Biol 1994; 165(1):193–205. 21. Zhang C, Evans T. Differential regulation of the two GATA-1 genes during Xenopus development. J Biol Chem 1994; 269(1):478–484. 22. Weiss MJ, Keller G, Orkin SH. Novel insights into erythroid development revealed through in vitro differentiation of GATA-1 embryonic stem cells. Genes Dev 1994; 8(10):1184–1197. 23. Pevny L, Simon MC, Robertson E, Klein WH, Tsai SF, D’Agati V, Orkin SH, Constantini F. Erythroid differentiation in chimaeric mice blocked by a targeted mutation in the gene for transcription factor GATA-1. Nature 1991; 349(6306):257–260. 24. Kulessa H, Frampton J, Graf T. GATA-1 reprograms avian myelomonocytic cell lines into eosinophils, thromboblasts, and erythroblasts. Genes Dev 1995; 9(10): 1250–1262. 25. Tsang AP, Fujiwara Y, Hom DB, Orkin SH. Failure of megakaryopoiesis and arrested erythropoiesis in mice lacking the GATA-1 transcriptional cofactor FOG. Genes Dev 1998; 12(8):1176–1188. 26. Streisinger G, Walker C, Dower N, Knauber D, Singer F. Production of clones of homozygous diploid zebra fish (Brachydanio rerio). Nature 1981; 291(5813):293– 296. 27. Bahary N, Zon LI. Use of the zebrafish (Danio rerio) to define hematopoiesis. Stem Cells 1998; 16(2):89–98. 28. Kimmel CB, Ballard WW, Kimmel SR, Ullmann B, Schilling TF. Stages of embryonic development of the zebrafish. Dev Dyn 1995; 203(3):253–310. 29. Nusslein-Volhard C. Of flies and fishes. Science 1994; 266(5185):572–574. 30. Postlethwait JH, Yan YL, Gates MA, Horne S, Amores A, Brownlie A, Donovan

Zebrafish: A Developmental and Genetic Model

31. 32. 33.

34.

35.

36. 37. 38.

39.

40. 41. 42. 43. 44.

45.

46.

277

A, Egan ES, Force A, Gong Z, Goutel C, Fritz A, Kelsh R, Knapik E, Liao E, Paw B, Ransom D, Singer A, Thomson M, Abduljabbar TS, Yelick P, Beier D, Joly JS Larhammar D, Rosa F, Westerfield M, Zon L, Johnson SL, Talbot WS. Vertebrate genome evolution and the zebrafish gene map. Nat Genet 1998; 18(4):345–349. Zon LI. Zebrafish: a new model for human disease. Genome Res 1999; 9(2):99– 100. Nikaido M, Tada M, Saji T, Ueno N. Conservation of BMP signaling in zebrafish mesoderm patterning. Mech Dev 1997; 61(1–2):75–88. Kishimoto Y, Lee KH, Zon L, Hammerschmidt M, Schulte-Merker S. The molecular nature of zebrafish swirl: BMP2 function is essential during early dorsoventral patterning. Development 1997; 124(22):4457–4466. Nguyen VH, Schmid B, Trout J, Connors SA, Ekker M, Mullins MC. Ventral and lateral regions of the zebrafish gastrula, including the neural crest progenitors, are established by a bmp2b/swirl pathway of genes. Dev Biol 1998; 199(1):93–110. Detrich HW, III, Kieran MW, Chan FY, Barone LM, Yee K, Rundstadler JA, Pratt S, Ransom D, Zon LI. Intraembryonic hematopoietic cell migration during vertebrate development. Proc Natl Acad Sci USA 1995; 92(23):10713–10717. Al-Adhami MA, and Kunz YW. Ontogenesis of haematopoietic sites in brachydanio rerio. Dev Growth Differ 1977; 19:171–179. Willett CE, Cortes A, Zuasti A, Zapata AG. Early hematopoiesis and developing lymphoid organs in the zebrafish. Dev Dyn 1999; 214(4):323–336. Thompson MA, Ransom DG, Pratt SJ, MacLennan H, Kieran MW, Detrich III HW, Vail B, Huber TL, Paw B, Brownlie AJ, Oates AC, Fritz A, Gates MA, Amores A, Bahary N, Talbot WS, Her H, Beier DR, Postlethwait JH, Zon LI. The cloche and spadetail genes differentially affect hematopoiesis and vasculogenesis. Dev Biol 1998; 197(2):248–269. Liao EC, Paw BH, Oates AC, Pratt SJ, Postlethwait JH, Zon LI. SCL/Tal-1 transcription factor acts downstream of cloche to specify hematopoietic and vascular progenitors in zebrafish. Genes Dev 1998; 12(5):621–626. Solnica-Krezel L, Schier AF, Driever W. Efficient recovery of ENU-induced mutations from the zebrafish germline. Genetics 1994; 136(4):1401–1420. Nusslein-Volhard C, Wieschaus E. Mutations affecting segment number and polarity in Drosophila. Nature 1980; 287(5785):795–801. Hirsh D, Vanderslice R. Temperature-sensitive developmental mutants of Caenorhabditis elegans. Dev Biol 1976; 49(1):220–235. Grunwald DJ, Streisinger G. Induction of recessive lethal and specific locus mutations in the zebrafish with ethyl nitrosourea. Genet Res 1992; 59(2):103–116. Chia W, Savakis C, Karp R, Pelham H, Ashburner M. Mutation of the Adh gene of Drosophila melanogaster containing an internal tandem duplication. J Mol Biol 1985; 186(4):679–688. Kelley MR, Mims IP, Farnet CM, Dicharry SA, Lee WR. Molecular analysis of Xray–induced alcohol dehydrogenase (ADH) null mutations in Drosophila melanogaster. Genetics 1985; 109(2):365–377. Pastink A, Schalet AP, Vreeken C, Paradi E, Eeken JC. The nature of radiationinduced mutations at the white locus of Drosophila melanogaster. Mutat Res 1987; 177(1):101–115.

278

Marty et al.

47. van Buul PP, Leonard A, Goudzwaard JH. Dose-effect relationship for X-ray–induced reciprocal translocations in mouse spermatogonia following pretreatment with 3-aminobenzamide. Mutat Res 1990; 232(2):273–280. 48. Tease C, Fisher G. Cytogenetic and genetic studies of radiation-induced chromosome damage in mouse oocytes. I. Numerical and structural chromosome anomalies in metaphase II oocytes, pre- and post-implantation embryos. Mutat Res 1996; 349(1):145–153. 49. Tease C, Fisher G. Cytogenetic and genetic studies of radiation-induced chromosome damage in mouse oocytes. II. Induced chromosome loss and dominant visible mutations. Mutat Res 1996; 349(1):155–162. 50. Kubota Y, Shimada A, Shima A. Detection of gamma-ray–induced DNA damages in malformed dominant lethal embryos of the Japanese medaka (Oryzias latipes) using AP-PCR fingerprinting. Mutat Res 1992; 283(4):263–270. 51. Shima A, Shimada A. The Japanese medaka, Oryzias latipes, as a new model organism for studying environmental germ-cell mutagenesis. Environ Health Perspect 1994; 102(Suppl 12):33–35. 52. Walker, C. S. Induction of mutations by gamma-rays in pregonial germ cells of zebrafish embryos. Genetics 1983; 103:125–136. 53. Kimmel, C. B. Genetics and early development of zebrafish. Trends Genet 1989; 5(8):283–288. 54. Fritz A, Rozowski M, Walker C, Westerfield M. Identification of selected gammaray induced deficiencies in zebrafish using multiplex polymerase chain reaction. Genetics 1996; 144(4):1735–1745. 55. Meisler MH. Insertional mutation of ‘classical’ and novel genes in transgenic mice. Trends Genet 1992; 8(10):341–344. 56. Burns JC, Friedmann T, Driever W, Burrascano M, Yee JK. Vesicular stomatitis virus G glycoprotein pseudotyped retroviral vectors: concentration to very high titer and efficient gene transfer into mammalian and nonmammalian cells [see comments]. Proc Natl Acad Sci USA 1993; 90(17):8033–8037. 57. Lin S, Gaiano N, Culp P, Burns JC, Friedmann T, Yee JK, Hopkins N. Integration and germ-line transmission of a pseudotyped retroviral vector in zebrafish. Science 1994; 265(5172):666–669. 58. Gaiano N, Allende M, Amsterdam A, Kawakami K, Hopkins N. Highly efficient germ-line transmission of proviral insertions in zebrafish. Proc Natl Acad Sci USA 1996; 93(15):7777–7782. 59. Allende ML, Amsterdam A, Becker T, Kawakami K, Gaiano N, Hopkins N. Insertional mutagenesis in zebrafish identifies two novel genes, pescadillo and dead eye, essential for embryonic development [see comments]. Genes Dev 1996; 10(24): 3141–3155. 60. Amsterdam A, Burgess S, Golling G, Chen W, Sun Z, Townsend K, Farrington S, Haldi M, Hopkins N. A large-scale insertional mutagenesis screen in zebrafish. Genes Dev 1999; 13(20):2713–2724. 61. Gong Z. Zebrafish expressed sequence tags and their applications. Methods Cell Biol 1999; 60:213–233. 62. Talbot WS, Schier AF. Positional cloning of mutated zebrafish genes. Methods Cell Biol, 1999; 60:259–286.

Zebrafish: A Developmental and Genetic Model

279

63. Jessen JR, Meng A, McFarlane RJ, Paw BH, Zon LI, Smith GR, Lin S. Modification of bacterial artificial chromosomes through chi-stimulated homologous recombination and its application in zebrafish transgenesis. Proc Natl Acad Sci USA 1998; 95(9):5121–5126. 64. Jessen JR, Willett CE, Lin S. Artificial chromosome transgenesis reveals long-distance negative regulation of rag1 in zebrafish [letter]. Nat Genet 1999; 23(1):15– 16. 65. Kimmel CB, Kane DA, Walker C, Warga RM, Rothman MB. A mutation that changes cell movement and cell fate in the zebrafish embryo. Nature 1989; 337(6205):358–362. 66. Henion PD, Raible DW, Beattie CE, Stoesser KL, Weston JA, Eisen JS. Screen for mutations affecting development of zebrafish neural crest. Dev Genet 1996; 18(1): 11–17. 67. Griffin KJ, Amacher SL, Kimmel CB, Kimelman D. Molecular identification of spadetail: regulation of zebrafish trunk and tail mesoderm formation by T-box genes. Development 1998; 125(17):3379–3388. 68. Yamamoto A, Amacher SL, Kim SH, Geissert D, Kimmel CB, De Robertis EM. Zebrafish paraxial protocadherin is a downstream target of spadetail involved in morphogenesis of gastrula mesoderm. Development 1998; 125(17):3389–3397. 69. Oates AC, Brownlie A, Pratt SJ, Irvine DV, Liao EC, Paw BH, Dorian KJ, Johnson SL, Postlethwait JH, Zon LI, Wilks AF. Gene duplication of zebrafish JAK2 homologs is accompanied by divergent embryonic expression patterns: only jak2a is expressed during erythropoiesis. Blood 1999; 94(8):2622–2636. 70. Ho RK, Kane DA. Cell-autonomous action of zebrafish spt-1 mutation in specific mesodermal precursors. Nature 1990; 348(6303):728–730. 71. Stainier DY, Weinstein BM, Detrich HW, III, Zon LI, Fishman MC. Cloche, an early acting zebrafish gene, is required by both the endothelial and hematopoietic lineages. Development 1995; 121(10):3141–3150. 72. Liao W, Bisgrove BW, Sawyer H, Hug B, Bell B, Peters K, Grunwald DJ, Stanier DY. The zebrafish gene cloche acts upstream of a flk-1 homologue to regulate endothelial cell differentiation. Development 1997; 124(2):381–389. 73. Parker L, Stainier DY. Cell-autonomous and non-autonomous requirements for the zebrafish gene cloche in hematopoiesis. Development 1999; 126(12):2643–2651. 74. Stainier DY, Lee RK, Fishman MC. Cardiovascular development in the zebrafish. I. Myocardial fate map and heart tube formation. Development 1993; 119(1):31– 40. 75. Kelsh RN, Brand M, Jiang YJ, Heisenberg CP, Lin S, Haffter P, Odenthal J, Mullins MC, van Eeden FJ, Furutani-Seiki M, Granato M, Hammerschmidt M, Kane DA, Warga RM, Beuchle D, Vogelsang L, Nusslein-Volhard C. Zebrafish pigmentation mutations and the processes of neural crest development. Development 1996; 123: 369–389. 76. Ransom DG, Haffter P, Odenthal J, Brownlie A, Vogelsang E, Kelsh RN, Brand M, van Eeden FJM, Furutani-Seiki M, Granato M, Hammerschmidt M, Heisenberg CP, Jiang Y, Kane D, Mullins MC, Nusslein-Volhard C. Characterization of zebrafish mutants with defects in embryonic hematopoiesis. Development 1996; 123:311– 319.

280

Marty et al.

77. Parichy DM, Rawls JF, Pratt SJ, Whitfield TT, and Johnson SL. Zebrafish sparse corresponds to an orthologue of c-kit and is required for the morphogenesis of a subpopulation of melanocytes, but is not essential for hematopoiesis or primordial germ cell development. Development 1999; 126 (15):3425–3436. 78. Brownlie A, Donovan A, Pratt SJ, Paw BH, Oates AC, Brugnara C, Witkowska HE, Sassa S, and Zon L. Positional cloning of the zebrafish sauternes gene: a model for congenital sideroblastic anaemia. Nat Genet 1998; 20(3):244–250. 79. Wang H, Long Q, Marty SD, Sassa S, Lin S. A zebrafish model for hepatoerythropoietic porphyria. Nat Genet 1998; 20(3):239–243. 80. Nussey G, Van Vuren JH, Du Preez HH. Effect of copper on blood coagulation of Oreochromis mossambicus (Cichlidae). Comp Biochem Physiol C Pharmacol Toxicol Endocrinol 1995; 111(3):359–367. 81. Jagadeeswaran P, Liu YC, Eddy CA. Effects of hirudin (thrombin specific inhibitor) in zebrafish embryos: a developmental role for thrombin. Blood Cells Mol Dis 1997; 23(3):410–414. 82. van Vliet KJ, Smit GL, Pieterse JJ, Schoonbee HJ, Van Vuren JH. Thrombelastographic diagnosis of blood coagulation in two freshwater fish species. Comp Biochem Physiol A 1985; 82(1):19–21. 83. Jagadeeswaran P, Liu YC. A hemophilia model in zebrafish: analysis of hemostasis. Blood Cells Mol Dis 1997; 23(1):52–57.

12 Evolving Paradigms in Vasculogenesis and Angiogenesis

JOSEPH A. MADRI Yale University School of Medicine New Haven, Connecticut

I.

Introduction

Vasculogenesis, the de novo formation of blood vessels, and angiogenesis, the formation of new blood vessels from those pre-existing, are crucial processes in embryonic development, postnatal growth and development, inflammation, wound healing/repair, and tumor growth and metastasis. Vasculogenesis and angiogenesis have been recognized as necessary key elements of all these processes for some time. However, only recently, with the advent of rapid advances in protein chemistry, molecular biology, genetics, immunology and cell biology and our abilities to integrate emerging concepts and methods of these disciplines, have we been able to begin to elucidate the basic underlying mechanisms at work during these processees. Prior to the development and use of targeted disruption and overexpression techniques to elucidate the roles of selected receptors and adhesion molecules, their ligands, proteases, and their inhibitors in the processes of vasculogenesis and angiogenesis, numerous studies had been published by a myriad of investigators demonstrating the importance of selected receptors, adhesion molecules, their ligands, proteases, and their inhibitors in these processes [1,2]. Many of these 281

282

Madri

studies employed the use of a variety of tissue culture/organ culture methods which permitted the formation and assessment of in vitro angiogenesis (tube formation) [3–9] . Using these culture methods, investigators assessed the relative importance of various extracellular matrix (ECM) components by altering the composition and/or organization of the ECM; adding antibodies directed against ECM components; interfering with ECM synthesis, secretion, and/or assembly; and degradation [7,8,10] . Similar approaches have been taken in the investigation of several adhesion receptors, growth factor receptors, proteases, and protease inhibitors [11–14]. The advent of transfection technology provided scientists with the tools to investigate the importance of these molecules in the processes of in vitro vessel formation stabilization and involution by expressing genes of choice in endothelial cells or decreasing specific gene expression using antisense technology [15,16]. In this chapter, selected studies focusing on and pertinent to the processes of vasculogenesis and angiogenesis and illustrating the application of targeted disruption and overexpression approaches will be discussed and evaluated for their impact on furthering the basic understanding of selected disease processes. II. Targeted Disruption and Overexpression Approaches Used to Study Vasculogenesis and Angiogenesis The pioneering work of Cappechi et al. [17–21] led to the use of homologous recombination techniques in a variety of areas and eventually to their use in the investigation of the processes of vasculogenesis and angiogenesis (for reviews, see Refs. 22–27). A. Receptors and Their Ligands

VEGFR/VEGF Family

Over the past several years, the importance of the vascular endothelial growth factor (VEGF) family (including VEGF-A, VEGF-B, VEGF-C, VEGF-D, and placenta growth factor, their cognate receptors VEGFR1–4) has been increasingly appreciated in the control of vasculogenesis, angiogenesis, vessel survival, and maintenance [24, 27] . flk-1 (VEGFR2) null mutants display an absence of mature endothelial cells and hematopoietic precursors and exhibit embryonic lethality at embryonic days (E) E8.5–9.5 [28,29] . Similarly, VEGF-A null mutants exhibit early embryonic lethality, which is consistent with the concept that VEGF and VEGFR2 are essential for the differentiation of mesoderm into hemangioblasts and ultimately endothelial cells [30,31] . In contrast, flt-1 (VEGFR1) null mutants, although exhibiting embryonic lethality at E8.5–9.5, exhibit increased

Vasculogenesis and Angiogenesis

283

numbers of endothelial cells and secondarily a disorganized vasculature [32,33]. Additional studies illustrating that the cytoplasmic domain of flt-1 is not essential for angiogenesis suggest that this member of the VEGFR family may function as a VEGF sink or reservoir and/or may be involved in some other, as yet undefined, signaling pathway(s) during the processes of vasculogenesis and angiogenesis [34] . What these and the myriad of other studies in this area have indicated is that the VEGF/VEGFR families are essential for and/or modulate a wide range of vascular functions, including vasculogenesis (VEGF and VEGFR2), angiogenesis remodeling (VEGF and VEGFR1), lymphangiogenesis (VEGFR3 and VEGF-C), mobilization of marrow-derived endothelial precursors, vascular permeability, endothelial cell migration, and survival [24,28,29,31–33,35–39]. Relevance of Genetic Approaches to Developmental, Adaptive, and Disease States: Current Status and Future Prospects

The rapid accrual of information regarding the VEGF/VEGFR families’ roles as major modulators of the vasculature has led to the proposed use of recombinant VEGF ligands and ligand cDNAs as therapeutic agents in patients diagnosed with both peripheral and cardiac ischemia due to vascular occlusive disease [38,40] . Using a variety of animal models, investigators have documented angiogenesis, increased blood flow, and functional improvement, and ongoing human trials likewise show promising results in the short term. The long-term benefit of such therapeutic interventions is unknown and many questions remain unanswered. Given the dynamic state of the vasculature, it is unclear if such VEGF-induced beds will have the degree of permanence required to ensure a long-term beneficial result. Since maintenance of a vascular bed requires complex interactions among several growth factors, extracellular matrix components, and endothelial and mural cells (pericyte/smooth muscle cell), it is unclear if VEGF treatment alone will be sufficient to accomplish this. In addition, given that VEGF is a potent permeability factor, the potential of detrimental effects due to changes in vascular permeability should not be overlooked. Several examples of this complexity are apparent in selected in vitro and in vivo models. Using a rodent model of chronic sublethal hypoxia which can mimic chronic hypoxia in some premature newborns, Ment et al. [41] have demonstrated increases in tissue VEGF levels in the cerebrum which correlate with significant, persistent permeability changes in the cerebral microvasculature. Additionally, robust angiogenesis was noted in the cerebral hemispheres of the rodent pups at later time points during the hypoxic insult, also correlating with increased VEGF levels. Finally, when these rodent pups were placed in a normoxic environment, vessel density in the cerebral hemispheres was noted to decrease to normoxic levels, suggesting that the expression levels of VEGF also mediated maintenance/survival of the recently formed vessels. In more recent

284

Madri

studies, Ogunshola et al. have demonstrated multiple spatiotemporal-regulated roles of VEGF during post-natal cerebral neurovascular development in rodents [42]. Specifically, neuronal VEGF expression was noted in band-like patterns progressing from upper to lower cortical levels; apparently preceding angiogenesis from meningeal vessels. This was followed by a decrease of neuronal expression and the appearance of glial VEGF expression at a time when cerebral angiogenesis is complete and stabilized. In contrast, during chronic sublethal hypoxia, the progressive laminar localization pattern was lost and neuronal VEGF expression appeared to be diffuse and sustained throughout the cortex over time; apparently inducing angiogenesis throughout all the cortical layers during the hypoxic period. In addition, glial VEGF expression was increased [42]. Data accrued from these and other in vivo models employing modulation of oxygen tension are consistent with the concept that cell-specific VEGF and VEGFR levels play critical roles in regulating vascular behavior. Although useful in illustrating the importance of particular cells, ligands, and receptors in physiological and pathophysiological processes, in vivo models have limitations in providing insights into underlying signaling pathways and mechanisms. The use of in vitro models can effectively overcome several of the limitations noted in in vivo models. For example, in a recent study, Ilan et al. demonstrated that VEGF functioned as a survival factor only in the presence of sufficient expression of VEGFR1 and 2 in human umbilical vein endothelial cell cultures (HUVEC) by activating the mitogen-activated protein kinase (MAPK) [43]. However, caution must be taken in making broad-based conclusions from such in vitro data. Illustrating the importance of cell type, passage number, and culture conditions in determining what pathways may be involved in inducing VEGF-mediated cell survival, Gerber et al. observed that the Akt/PKB pathway was activated during VEGF-mediated cell survival in primary cultures of human endothelial cells [44]. Furthermore, it has become apparent that VEGF receptors are expressed on a variety of nonendothelial cells and that other receptors are capable of interacting with VEGF family members, allowing for the possibility of activation of yet undefined signaling pathways and/or additional modulation of VEGF–VEGFR signaling [45–47]. An intriguing example of this complexity is illustrated in the recent work of Soker et al. In these studies, the investigators found that neuropilin-1, a semaphorin/collapsin receptor known to mediate neuronal guidance, binds VEGF165 and when expressed in cells along with VEGFR2 enhances VEGF165 binding to VEGFR1. Such molecules may participate in regulating the binding of selected VEGF isoforms to VEGFR1 and thus influence angiogenesis in autocrine and/or paracrine manner. These in vitro and in vivo studies illustrate the complexity of VEGF actions, the variable results obtained using different in vivo and in vitro model systems and the potential difficulties in using VEGF agonists and antagonists as therapeutic agents.

Vasculogenesis and Angiogenesis

285

Tie/Angiopoietin Family

Tie1 and Tie2 receptors were discovered as orphan tyrosine kinase receptors [48]. Subsequent studies elucidated the ligands for Tie2 to be the angiopoietins [49– 51]. Tie2-deficient animals were shown to undergo normal development of the early VEGF-mediated stages of vasculogenesis, but remodeling and stabilization of the developing vasculature were found to be compromised and embryonic lethality was noted at E9.5–10.5 [48,49]. With the discovery and elucidation of the roles of angiopoietin-1 and angiopoietin-2 (and later angiopoietin-3 and angiopoietin-4) came the finding that they belong to a growth factor family which comprises both activators and inhibitors. Specifically, murine embryos deficient in angiopoietin-1 behaved similarly to Tie2-deficient embryos [22], resulting in disruptions in endothelial cell–supporting (or mural cell) interactions, suggesting that angiopoietin-1 engagement (activation) of Tie2 was necessary for vascular stabilization [24]. In contrast, although angiopoietin-2 was found to bind to Tie2, it did not appear to activate the receptor. Additionally, transgenic mice overexpressing angiopoietin-2 were found to exhibit an embryonic lethal phenotype closely resembling either angiopoietin-1– deficient or Tie2-deficient embryos, which is consistent with a blocking function [24,50]. This concept of receptor activation and blockade is supported by in vivo localization studies in which angiopoietin-1 was found to be widely expressed in adult tissues, whereas angiopoietin-2 was found to have a restricted expression, present at sites of vascular remodeling [50]. Furthermore, the colocalization of angiopoietin-2 and VEGF correlated with foci of vessel sprouting, whereas the expression of angiopoietin-2 without VEGF colocalization correlated with areas of vessel regression [50]. In contrast to our understanding of the roles of Tie2, angiopoietin-1, and angiopoietin-2 in the development, maintenance, and remodeling of the vasculature, our current understanding of the role(s) of the Tie1 receptor is considerably more deficient. Although no ligand has been identified as a Tie1 ligand to date, there is a growing body of evidence suggesting that the Tie1 receptor is involved in signaling in hematopoietic precursor and endothelial cell populations. Investigators demonstrated that in hematopoietic precursor cells expressing Tie1, the Tie1 undergoes a cytokine mediated proteolytic clip (mediated by an undefined metalloproteinase), releasing the Tie1 extracellular domain into the cell culture media en bloc [52]. In a follow-up study, they reported that the remaining transmembrane and cytoplasmic domain of the Tie1 receptor also survives intact and remains associated with the membrane fraction of the cells [53]. Similar cleavages have been documented for other receptors [54,55], and the released extracellular domains have been postulated to function as soluble binders of specific ligands either reducing the effective ligand concentration or serving a ‘‘presenters’’ of the ligand to uncleaved or related receptors [56]. Recently, McCarthy

286

Madri

et al. demonstrated a cytokine-induced, metalloproteinase-mediated cleavage of endothelial cell Tie1 in a HUVEC culture model and the presence of the extracellular and cytoplasmic domains of Tie1 in human pulmonary and placental tissues [57]. In this model, the investigators also demonstrated that the cytoplasmic domain of Tie1 associated with a number of undefined phosphoproteins; suggesting that the generation of the cytoplasmic domain of Tie1 may be involved in signal transduction via binding to adapter and signaling proteins. These studies (and those mentioned in the previous section regarding VEGFs and VEGFRs) suggest a complex cross-talk among members of specific receptor families and members of their ligand family(ies), affecting a tightly controlled balance between vessel sprouting and growth, vessel stabilization/maintenance, and vessel regression. Relevance of Genetic Approaches to Developmental, Adaptive, and Disease States: Current Status and Future Prospects

In contrast to the growing number of preclinical and clinical studies investigating the usefulness of anti-VEGF, anti–Flk-1, soluble Flt-1 receptor, inhibitors of Flk1 and Flt-1, VEGF antisense oligonucleotides, and ribozymes targeting VEGF receptors [58], preclinical and clinical trials centered around the Tie1 and Tie2 receptors and the angiopoietins are less well developed owing in large part to our incomplete understanding of the roles these receptors and ligands play in the processes of vasculogenesis and angiogenesis [22]. Insight into the importance of this family of receptors and their ligands is illustrated in the work of Vikkula et al. [59]. In this study, the investigators have described linkage of an inherited venous malformation in two families to the short arm of chromosome 9. These malformations are characterized by vascular structures having dilated, serpiginous channels and variable smooth muscle investiture. Segregating with the dominantly inherited venous malformation in the two families was a missense mutation resulting in an arginine to tryptophan substitution at residue 849 in the kinase domain of the Tie2 receptor tyrosine kinase. This mutation was found to be an activating mutation in Tie2 (causing increased phosphorylation activity of Tie2) and is postulated to result in a defect in the recruitment of mural cells. A more complete understanding of the roles of Tie1, Tie2, and the angiopoietins in developmental and adult vasculogenesis and angiogenesis will provide a solid basis for designing, testing, and implementing therapeutic agents based on these receptors and ligands. Eph/Ephrin Family

The Eph receptors and their ligands, the ephrins, comprise the largest subfamily of receptor tyrosine kinases. Unlike the ligands of the VEGF and Tie receptors,

Vasculogenesis and Angiogenesis

287

the ephrins must be membrane tethered to activate their cognate receptors [60]. In addition, receptor activation appears to require clustering/multimerization of the ligands, facilitated by membrane tethering, for receptor activation on adjacent cells. Recently, selected members of the ephrin family have been implicated in the development of the vascular system. Ephrin A1 (B61) was initially identified as a tumor necrosis factor-β (TNFα)–, interleukin-1β (IL-1β)–, and lipopolysaccharide (LPS)–inducible endothelial cell product [61,62] which was also found to be upregulated during in vitro capillary tube formation [63]. Later experiments demonstrated that ephrin A1 is an angiogenic factor [64]. In other studies, ephrin B1 ligand was found to support human renal microvascular endothelial cell (HRMEC), but not HUVEC, organization, whereas ephrin A1 had the opposite cell-type preferences [65]. Recently, both ephrin B1 and ephrin B2 were demonstrated to induce sprouting angiogenesis in vitro to an extent similar to that observed with VEGF or angiopoietin-1 [66]. Interestingly, a glutathione-S-transferase (GST)–Tie2 fusion protein was able to induce ephrin B1 phosphorylation in an in vitro kinase assay [66]; suggesting an interesting cross talk between these two distinct signaling pathways. In the adult, arteries and veins are morphologically and functionally distinct. These differences have been assumed to be due to different physiological parameters such as blood flow, blood pressure, blood oxygenation, and/or shear stress. In contrast, in the developing embryo, the primary capillary plexus appears to be homogeneous and arteries and veins are virtually indistinguishable. A recent study of Wang et al. (1998) [67] suggests that early embryonic arterial and venous endothelial cells are indeed molecularly distinct. Homozygous ephrin B2 knockout mice were found to exhibit growth retardation at day 10 and lethality around day 11, with multiple cardiovascular anomalies, including an abnormal, enlarged yolk sac vasculature [67]. Further analysis of the vascular system of these mice revealed that ephrin B2 expression (as detected by the lacZ reporter gene) was restricted to arteries, but not to veins, whereas ephrin B2 receptor (EphB4) expression was reciprocally observed in veins but not in arteries [67]. Although vasculogenesis occurred normally in the ephrin B2 knockout mice, angiogenesis was disrupted. Vascular remodeling of both arteries and veins into large and small branches was severely affected; suggesting that reciprocal ephrin/Eph– mediated interactions between veins and arteries are necessary for proper angiogenic remodeling of the initial capillary plexus. In addition, no sprouting angiogenesis of the neuroectoderm was detected in ephrin B2 knockout mice. These observations may provide insights into the molecular guidance of blood vessel ingrowth, specifically that vascular sprouts expressing ephrin B2 interact with EphB expressed on ectodermal cells, and that these interactions modulate, in part, sprouting angiogenesis. Additionally, vessel maturation was impaired in the ephrin B2 knockout mice; specifically periendothelial cells appeared to be poorly

288

Madri

associated with the endothelial cell layer [67] similar to the phenotype reported for angiopoietin-1 [49] or its receptor, Tie2 knockout mice [48] and in the yolk sac vasculature of murine conceptuses cultured in 20 mM D-glucose or harvested from diabetic mothers [68,69]. The notion of a single Eph–ephrin pair reciprocally expressed on arteries or veins has recently been challenged by Adams et al. (1999) [66], who demonstrated the presence of ephrin B1 ligand on both arteries and veins. In addition, two additional EphB receptors, EphB2 and EphB3, were found to be expressed in the yolk sac and the embryo vasculature. Embryos lacking both EphB2 and EphB3 displayed vascular defects that mimicked the ephrin B1 phenotype, with the exception of normal sprouting angiogenesis [66]. The expression of multiple Ephs/ephrins in vascular endothelial cells and in surrounding tissues is consistent with the notion that ephrin–Eph interactions are not restricted to the arterial– venous boundary, but possibly occur throughout the vasculature between endothelial cells of the same vessel type or at sites of endothelial–mesenchymal interaction. Because they are membrane tethered, ephrin B ligands may themselves become tyrosine phosphorylated upon Eph receptor engagement, and thus may activate signaling events within the expressing cell [70,71]. This potential bidirectional signaling renders ligand/receptor definitions blurry and Eph/ephrin signaling very complex. Although the traditional concept for receptor tyrosine kinase signaling has implicated ligand-induced receptor dimerization, receptor transphosphorylation, and initiation of signal transduction, ephrin B1 oligomerization has been shown to be required for EphB receptor activation [60]. Indeed, Stein et al. (1998) [72] suggested that EphB1 and EphB2 can distinguish between different ephrin B1 oligomers, allowing for the formation of alternative EphB1 and EphB2 signaling complexes. In these studies, although both dimeric and multimeric ephrin B1 were able to induce EphB1 and EphB2 phosphorylation, multimeric ephrin B1 promoted endothelial organization into capillary-like structures, whereas dimeric ephrin B1 did not [72]. Similarly, multimeric ephrin B1 was more effective in promoting endothelial and P19 cell attachment to fibronectin. Recently, HuynhDo et al. (1999) [73] demonstrated that ephrin B1–induced endothelial cell attachment is mediated by αvβ3 but not αvβ5 or α5β1 integrins. Further, ephrin B1 did not induce any changes in αvβ3 surface expression, suggesting the potential of inside-out signaling affecting the αvβ3 activation state. These studies are consistent with the notion that in addition to Eph–ephrin bidirectional and unidirectional signaling, these molecules are capable of interactions with other receptor tyrosine kinases and nontyrosine kinase receptors. Further investigations are required to elucidate these signaling pathways in the context of vasculogenesis/ angiogenesis.

Vasculogenesis and Angiogenesis

289

Relevance of Genetic Approaches to Developmental, Adaptive, and Disease States: Current Status and Future Prospects

Similar to the Tie receptors and their ligands, our current understanding of the roles and importance of the ephrins and their receptors as potential therapeutic targets are not well developed owing again in large part to our incomplete understanding of the roles these receptors and ligands play in the processes of vasculogenesis and angiogenesis. Transforming Growth Factor Receptor-β (TGF-βR)/Transforming Growth Factor-β Family

A large number of publications have appeared over the past several years which have elucidated the complex phenotypes of various TGF-β isoform and TGF-β receptor deficient mice [74–79]. The TGF-β-1 1 null mutation in mice was found to cause partial embryonic lethality and an inflammatory response and early death in surviving mice. This complex phenotype has been found to be dependent upon the genetic background of the conceptus [77]. The embyronic lethality is associated with defective yolk sac vasculogenesis, whereas the inflammation and wasting are consistent with a role for TGF-β1 in the regulation of the immune system [78]. Additional studies have shown that transforming growth factor type II receptor–deficient mice were found to be embryonic lethal at E10.5, exhibiting defects in yolk sac vasculogenesis and hematopoiesis, similar to the phenotype noted in the TGF-β1–deficient mice [76], and that TGF-β isoforms elicit isoform-specific activities [75]. These and many other studies have led to the understanding that cellular and tissue response(s) to TGF-β isoform engagement are modulated at several levels. Particular TGF-β isoforms have been shown to affect specific cells in distinct ways, influencing proliferation, migration, matrix synthesis, and breakdown [80,81]. In addition to modulation by isoform composition and concentration, responses to TGF-β ligands appear to be modulated by expression levels of TGF-β receptors. Different cell types are known to express varying profiles of TGF-β receptors [82], and cells cultured under different conditions have been shown to exhibit variability in the expression of their TGF-βR ratios, leading to altered responsiveness to TGF-β isoforms resulting in differential proliferative, migratory and extracellular matrix synthetic patterns [15,16]. TGF-β–induced signaling has also been found to be modulated at the postreceptor level via SMAD proteins which function to transduce signals from the cell membrane to the nucleus [83]. In this section, we will discuss recent work describing the role of the TGF-βR III in endocardial cushion development [84], the role of endoglin in vascular remodeling [79] and the role of SMAD-5 in the process of vascularization as examined in SMAD-5–deficient mice [85].

290

Madri

TGF-β Type III Receptor

Luminal endothelial cells that line the atrioventricular (AV) cushion and outflow tract regions of the developing heart participate in the formation of the cardiac valves and membranous septa by undergoing an epithelial to mesenchymal transformation. TGF-β isoforms have been shown to be necessary for AV cushion development to occur. Recently, TGF-β2 and TGF-β3 were found to have separate and sequential activities during this epithelial to mesenchymal transformation in the developing avian heart [86]. Specifically, investigators found that TGFβ2 appears to mediate initial endothelial cell–cell separation, whereas the TGFβ3 isoform appears to mediate the changes that enable the migration of the cushion endothelial cells into the underlying cardiac jelly. The distinct effects of these TGF-β isoforms appears to be mediated by differential receptor expression in the developing heart. Whereas TGF-β type II receptor is expressed in all endothelia in the heart, TGF-β type III receptor appears to be expressed on the endothelial cells overlying the AV cushion; suggestive of a role for TGF-β type III receptor (which has a high affinity for TGF-β2) in mediating the epithelial to mesenchymal transformation noted during AV cushion development in the developing heart. This concept is supported by the observation that the ability of normally unresponsive endothelial cells infected with TGF-β type III receptor to form mesenchyme in response to TGF-β2 is significantly increased [86]. Whether TGF-β type III receptor directly or indirectly transduces signals in response to TGF-β2 is currently unknown. However, modulation of expression levels of this receptor is known to change the cellular responsiveness to TGF-β2 and the expression levels of TGF-β type I and II receptors [15,16]; suggesting the possibility of more than one mechanism of control of TGF-β–mediated signal transduction during AV cushion and septal development. Relevance of Genetic Approaches to Developmental, Adaptive, and Disease States: Current Status and Future Prospects

Currently, it is unknown whether altered expression levels or mutations in the TGF-β type III receptor cause specific cardiac defects. However, recent studies have demonstrated a link between human AV cushion defects and a region of chromosome 1 near to the TGF-β type III receptor coding region [87]. In addition, animal studies utilizing the db/db mouse model of diabetes have demonstrated a correlation between the upregulation of TGF-β type II receptor mRNA, protein, and activity and hyperglycemia and the development of diabetic nephropathy in adult mice [88]. Maternal diabetes mellitus is associated with an increased incidence of congenital abnormalities as well as embryonic and perinatal lethality. In particular, a wide range of cardiovascular abnormalities have been noted in children of diabetic mothers and in the offspring of diabetic animals. Murine conceptuses harvested from diabetic mothers and conceptuses harvested from

Vasculogenesis and Angiogenesis

291

normal mothers and cultured in 20 mM D-glucose exhibit failure of AV cushion formation [69] (Fig. 1). Thus, in light of the human and animal studies demonstrating changes in TGF-βR profiles causing changes in TGF-β responsiveness, the linking of AV cushion defects to a region of chromosome 1 near the coding region of TGF-β type III receptor and studies demonstrating hyperglycemia affecting TGF-β type II receptor expression, it is tempting to speculate that hyperglycemic insult in utero could affect TGF-βR and/or TGF-β isoform expression, ultimately affecting AV cushion development, and resulting in embryonic lethality or congenital cardiac abnormalities. Endoglin and SMAD5

Endoglin, a TGF-β binding glycoprotein exhibiting high affinities for TGF-β1 and TGF-β3, is found on endothelial cell surfaces and is thought to participate in the modulation of TGF-β signaling [79,89]. Loss of function mutations in this gene result in an autosomal dominant vascular disease, hereditary hemorrhagic telangiectasia (HHT), which is characterized by multisystemic vascular dysplasia and recurrent hemorrhages [90]. Recent studies utilizing targeted inactivation of endoglin (Eng) followed by analysis of Eng ⫹/⫹, Eng ⫹/-, and Eng -/- mice has resulted in a better-defined understanding of the role of endoglin during the processes of vasculogenesis and angiogenesis [79]. These studies revealed that, although vasculogenesis (initial formation of hemangioblasts form mesoderm and subsequent differentiation of angioblasts and blood islands) was not impaired in the Eng -/- conceptuses, angiogenesis (differential growth and sprouting of endothelial tubes, arborization and mural cell recruitment) was, resulting in embryonic lethality at day 11.5 postconception (PC) from defective vascular development. Namely, initial formation of the vasculature from mesoderm appeared to be normal with the formation of the primary capillary plexus. However, there was a failure of remodeling which resulted in the persistence of immature vascular plexi and poor vascular smooth muscle development and investiture around the endothelial lumina (see Fig. 2). As mentioned above, TGF-β–induced signaling has been found to be modulated at the postreceptor level via SMAD proteins (proteins of 42–60 kD having two regions of homology at their amino and carboxy termini—Mad-homology domains MH1 and MH2—which are connected via a proline-rich linker region) and which function to transduce signals from the cell membrane to the nucleus [83]. Since their discovery, targeted disruptions of individual SMAD proteins have been carried out resulting in a variety of phenotypes. Recently, two groups have reported on vascular defects associated with SMAD5 knockout mice [85, 91]. In both instances, the investigators reported that the SMAD5 -/- mice exhibited embryonic lethality between day 9.5 and 11.5 postconception and that the conceptuses exhibited abnormal yolk sac vasculature. Further investigation of the

292

Madri

Figure 1 Light micrographs of 1-micron sections of developing hearts from: a representative cultured normoglycemic 9.5 day PC embryo culture (A,C); a representative cultured hyperglycemic 9.5 day PC embryo culture (B,D); a representative embryo at 9.5 days PC harvested from a control, normoglycemic mother; and a representative affected embryo at 9.5 days PC harvested from a hyperglycemic mother. Although the development of the epicardium, myocardium, and endocardium appears to be normal in all instances (A,B,E,F), upon further examination of the hearts, we noted a failure of endocardial cushion development in the embryos cultured in hyperglycemic media (20 mM D-glucose, B,D) and in the affected embryos harvested from hyperglycemic mothers (F,H). Specifically, as illustrated in panels D and H, there is a failure of migration of endocardial cells overlying putative cushion areas into the cardiac jelly (CJ). Normal cushion development at this stage of development is illustrated in panels C and C, inset and G and G, inset, which consists of endocardial cell transformation into mesenchymal cells which are observed migrating into the cardiac jelly (arrows). In contrast, following either in vitro or in vivo hyperglycemic insult, there is arrest of cushion development (D and D, inset, and H and H, inset) which consists of a failure of mesenchymal of transformation and imigration of endocardial cells into the cardiac jelly (CJ). Scale bars ⫽ 100 µ. (From Ref. 69.)

Vasculogenesis and Angiogenesis

293

yolk sac vasculature abnormalities revealed persistence of the primary capillary plexus, failure of vessel aborization, and diminished recruitment of pericytes/ smooth muscle cells. This phenotype is reminiscent of that observed in BMP4, TGF-β1, TGF-β type II receptor, ephrin B2, and tissue factor knockout mice and in murine conceptuses cultured in high-glucose media or harvested from diabetic mothers (Fig. 2) [48,49,59,67,69]. This phenotype is consistent with a disruption of mesoderm development and is also observed in several mice exhibiting targeted disruptions of fibronectin, α5 and α4 integrins, and vascular cell adhesion molecule-1 (VCAM-1) [48,69,92–96]. Relevance of Genetic Approaches to Developmental, Adaptive, and Disease States: Current Status and Future Prospects

Over the past several years, there has been considerable interest in the potential use of TGF-β isoforms and TGF-β antagonists in several clinical arenas. However, initial enthusiasm has been dampened somewhat by the appreciation of the complexity and the contextual nature of the signaling pathways involved in TGFβ–mediated signal transduction [16,83,97,98]. The age of the individual, location, extent, and cause of the injury all appear to play roles in modulating TGF-β– mediated responses. Thus, the use of TGF-β isoforms and TGF-β antagonists awaits further elucidation of the signal transduction pathways involved and the use of such reagents in defined animal models and clinical settings which are modulated by specific TGF-β isoform levels and/or TGF-βR expression. B. Adhesion Molecules and Their Ligands

The vascular system is the first organ system to develop in the embryo and is critical for normal organogenesis. The organization of mesodermal cells into endothelial and hematopoietic cells and into a complex vascular system is mediated by a series of specific cell factor, cell extracellular matrix, and cell–cell interactions (Fig. 3). In this section, the contributions of two cell adhesion molecules, VE-cadherin and platelet-endothelial cell adhesion molecule-1 (PECAM-1) (CD31), will be discussed. VE-Cadherin

VE-cadherin is a member of the cadherin family of transmembrane adhesion proteins which are involved in homotypic cell–cell interactions and whose cytoplasmic domains are dynamically associated with α-catenin, β-catenin and plakoglobin, which, in turn, mediate interactions with the actin cytoskeleton. VEcadherin has been detected in day 7.5 PC. conceptuses in the mesodermal aggregates which will eventually form the blood islands and has been shown to be expressed in all developing and adult vasculature. VE-cadherin is thought to promote junctional assembly and participates in the control of vascular permeability [99–101]. In addition, cadherins can serve as reservoirs for catenins, modulating

Figure 2 Light (A,B) and electron microscopic (C,D) analyses of the developing yolk sac vasculature at day 9.5 PC in normoglycemic (A,C) and hyperglycemic (B,D) culture conditions. (A) High-power representative light micrograph of the terminal yolk sac vasculature capillaries observed in a conceptus harvested at day 7.5 PC and cultured for 48 hr (now a day 9.5 PC conceptus) in normoglycemic conditions. Note the vessel lined by flattened endothelial cells (arrowheads) containing two circulating blood cells in the lumen, bordered by the endodermal cells on one aspect (double arrows) and flattened, elongated mesenchymal cells (pericytes) and mesothelial cells on the other aspect (arrows) in close contact with the abluminal surfaces of the endothelium. Scale bar ⫽ 20 µ. (B) High-power representative light micrograph of the terminal yolk sac vasculature capillaries observed in a conceptus harvested at day 7.5 PC and cultured for 48 hr (now a day 9.5 PC conceptus) in hyperglycemic conditions. Note the vessel lined by flattened endothelial cells (arrowheads), bordered by the endodermal cells on one aspect (double arrows) and plump, rounded, mesenchymal cells (pericytes) and mesothelial cells on the other aspect (arrows) in loose contact with the abluminal surfaces of the endothelium. A blood cell is noted in the lumen. Scale bar ⫽ 20 µ. (C) Representative transmission electron micrograph of the terminal yolk sac vasculature capillaries observed in a conceptus harvested at day 7.5 PC and cultured for 48 hr (now a day 9.5 PC conceptus) in normoglycemic conditions. Note the vessel lined by flattened endothelial cells (arrowheads) bordering the lumen, and the flattened, elongated mesenchymal cells (pericytes) in close contact with the abluminal surfaces of the endothelium. Scale bar ⫽ 7.3 m. (D) Representative transmission electron micrograph of the terminal yolk sac vasculature capillaries observed in a conceptus harvested at day 7.5 PC and cultured for 48 hr (now a day 9.5 PC conceptus) in hyperglycemic conditions. Note the vessel lined by flattened endothelial cells (arrowheads) bordering the lumen, and the plump, rounded, mesenchymal cells (pericytes) in loose contact with the abluminal surfaces of the endothelium. Scale bar ⫽ 7.3 m. (From Ref. 69.)

Vasculogenesis and Angiogenesis

295

Figure 3 General scheme of the temporal and spatial relationships of selected cell adhesion molecules—CAMs (PECAM-1); substrate adhesion molecules—SAMs (β1 and β3 integrins); and junction-associated molecules—JAMs (ZO-1) during angiogenesis. Also illustrated are the effects of selected inhibitors of cell–cell and cell–matrix interactions on the angiogenesis process.

their pool sizes, localizations, degradation, and translocation into the nucleus [102,103]. Previously, it was found that disruption of VE-cadherin in embryonic stem cells resulted in impaired vasculogenesis [104]. In recent studies, VE-cadherin homozygous null mutants have been shown to be embryonic lethal, exhibiting vasculogenic defects in the yolk sac and embryo [105]. Specifically, at day 8.5 PC. dorsal aortae were incompletely formed, posterior cardinal veins were lacking, and the endocardial endothelial cells failed to form a lumen. Although yolk sac blood islands formed, they did not undergo branching and failed to establish a primary vascular plexus. These data suggested that VE-cadherin is not required for endothelial homophilic adhesion (sorting) but is required for vascular branching and subsequent morphogenesis. Given the role of cadherins as modulators of catenin pools, VE-cadherin deficiency may have its effects on the vasculature by altering catenin pools, levels, localizations, and possibly signaling within the endothelial cells. In a more recent study, potential signaling mechanisms which involve VEcadherin were investigated. Specifically, Carmeliet et al. [106] postulated that, since cadherins are known to participate in signaling by clustering signaling molecules and growth factor receptors [107], VE-cadherin may perform similar functions in endothelial cells. In these studies, VE-cadherin–deficient embryos were observed to undergo normal initial vasculogenesis. However, at later stages of development (day 8.5 PC.) defects in lumen formation and sprouting angiogen-

296

Madri

esis were noted as was increased endothelial cell apoptosis. The increased endothelial apoptosis was determined to be due to an inability of VE-cadherin null cells to mount a VEGF-A–mediated survival response via a VE-cadherin/βcatenin/PI3-kinase/VEGFR2 complex. Thus, VE-cadherin appears not to be necessary during initial vasculogenesis, but rather it is necessary for the later stages of angiogenesis, regulating endothelial cell apoptosis by allowing complex formation of selected signaling proteins and activation of antiapoptotic pathways. PECAM-1

Expression of PECAM-1 (CD31), a 130-kD member of the Ig superfamily, begins early in development at the stage of hemangioblast formation and persists through adulthood [108]. PECAM-1 is expressed on endothelial cells, platelets, polymorphonuclear leukocytes, monocytes, and T and B lymphocytes [108–110]. PECAM-1 is thought to serve several functions, including that of an adhesion molecule mediating cell–cell adhesion between adjacent endothelial cells and between endothelial cells and polymorphonuclear leukocytes, platelets, monocytes, and lymphocytes [108]. PECAM-1 has also been shown to be a modulator of in vitro and in vivo angiogenesis, endothelial cell migration, and polymorphonuclear leukocyte transmigration. Furthermore, PECAM-1 localization and phosphorylation states are known to change dynamically during vasculogenesis and angiogenesis and can be affected by integrin engagement, hypoxia, hyperglycemia, and osmolarity [13,42,68,69,109–115]. In addition to its adhesive functions, PECAM-1 has been shown to be a participant in signaling pathways [68,108,114,116]. Following engagement of PECAM-1, integrin affinity changes have been noted on platelets and lymphocytes [117–121] and changes in intracellular calcium localization have been documented [122]. PECAM-1–mediated signaling is thought to occur, in part, via its cytoplasmic immunoregulatory tyrosine inhibition motif (ITIM) domain [108,114,115]. This domain is known to mediate binding of signaling and adapter molecules having one or tandem SH2 domains when the tyrosine residues residing in the PECAM-1 ITIM domain are phosphorylated [108,114,115]. Specifically, the phosphatase SHP-2 has been shown to bind to tyrosine phosphorylated PECAM-1. Recently, we have found that in addition to its interactions with SHP2, PECAM-1 can serve as a reservoir for and a modulator of β-catenin, binding tyrosine phosphorylated β-catenin, and if PECAM-1 is itself tyrosine phosphorylated, bringing β-catenin into close proximity to SHP-2 facilitating dephosphorylation of the bound β-catenin [123]. These findings, coupled with what currently is known about endothelial cell–cell interactions, the cellular components at junctional sites, and the close proximity of PECAM-1 to these sites are consistent with the possibility of interactions among PECAM-1 and components associated with junctional complexes (Fig. 4).

Vasculogenesis and Angiogenesis

297

Figure 4 Working model of the dynamic PECAM-1/β-catenin/SHP-2 interactions in endothelia. PECAM-1 is a sink/reservoir for tyrosine phosphorylated β-catenin. Tyrosine phosphorylation of β-catenin (β-cat) occurs upon VEGF stimulation, c-src activity or as a result of ECM–cell interactions. The tyrosine phosphorylated β-catenin then binds to the cytoplasmic domain of PECAM-1 sequestering it and keeping it available for subsequent dephosphorylation by SHP-2. PECAM-1 is a modulator of β-catenin tyrosine phosphorylation. During endothelial cell attachment, spreading, migration and differentiation, there is modulation of the phosphorylation state of the PECAM-1 ITIM domain by protein tyrosine kinases (PTK) and protein tyrosine phosphatases (PTP), thus regulating the binding of the tyrosine phosphatase SHP-2. Bound, activated SHP-2 may then dephosphorylate the PECAM-1–bound tyrosine phosphorylated β-catenin, making it available for interactions with VE-cadherin and the formation of adherens junctions and/or translocation to the nucleus and interaction with LEF/Tcf, leading to changes in gene expression. (Drawn from data presented in Ref. 123.)

These observations prompted the generation of PECAM-1–deficient mice [124]. PECAM-1 null animals surprisingly were found to be viable and exhibited an abnormal transit of polymorphonuclear leukocytes across vascular basement membranes as their only demonstrable phenotype [124]. Other recent investigations using vital microscopy techniques have revealed delays in leukocyte transmigration in PECAM-1–deficient mice [125]. A plausible explanation accounting for the discrepancy between inhibition of in vitro angiogenesis and in vivo adult angiogenesis by PECAM-1 antibodies and/or soluble PECAM-1 peptides [13] and the viable phenotype observed in the PECAM-1–deficient mice is redundancy of function during early vascular development. This phenomenon has been suggested by investigators in their explanations of the resultant phenotypes gener-

298

Madri

ated from integrin and VE-cadherin knockout mice [105,126]. The use of conditional knockouts and ‘‘knockout-knockin’’ approaches, allowing the controlled expression or wild-type PECAM-1 or expression of mutagenized and truncated PECAM-1 constructs will perhaps permit a better understanding of the roles this molecule plays in vasculogenesis and angiogenesis during development and vessel maintenance and adaptive and reactive angiogenesis in the adult. Relevance of Genetic Approaches to Developmental, Adaptive, and Disease States: Current Status and Future Prospects

The demonstrated importance of VE-cadherin and PECAM-1 as modulators of in vivo and in vitro angiogenesis and catenin pool dynamics has not yet been translated into clinical applications. With a more complete understanding of their expression profiles, alternative splicing in the case of PECAM-1, regulation of catenin, kinase and phosphatase binding, and subsequent signaling, rationales for the manipulation of VE-cadherin and PECAM-1 expression may be forthcoming. C. Matrix Metalloproteases

Precise, controlled, specific breakdown of the extracellular matrix is a common essential component of reproduction, embryonic development, morphogenesis, remodeling, and angiogenesis. Matrix metalloproteinases (MMPs, matrixins: members of the metzincin families) are known to play important roles in these processes [127,128]. MMP induction is tightly controlled at the transcriptional level, whereas activation is controlled by proteolytic cleavage of MMP precursors and interactions with endogenous inhibitors. Angiogenesis is considered to be an invasive process which involves endothelial cell activation, protease induction and activation, ECM synthesis, migration, proliferation, differentiation (tube formation), remodeling, stabilization, and sometimes involution [1,2,24,27]. Matrix Metalloprotease-2 (MMP-2) and Membrane-Type Matrix Metalloproteinase-1 (MT-1 MMP).

Several MMPs have been shown to be expressed by endothelial cells and are required for angiogenesis. MMP-2 and MT-1 MMP are perhaps the most studied, and both appear to be induced during the early phases of angiogenesis, as evidenced by in vitro, organ culture and in vivo evidence accrued in several models [14,129]. Both enzymes are produced as zymogens, and their activation is thought to be tightly regulated at the cell surface. Microvascular endothelial cells cultured on collagen type I and other ECM coatings express low, constitutive levels of pro–MMP-2 with no appreciable activation. However, microvascular endothelial cells or microvascular vessel fragments placed in three-dimensional collagen type I gels exhibit robust coordinate induction of both pro–MMP-2 and MT-1 MMP with significant activation of MMP-2 by MT-1 MMP. The process by which

Vasculogenesis and Angiogenesis

299

activation occurs is complex, and is thought to involve interactions between MTMMP, latent MMP-2, and tissue inhibitor of metalloproteinase-2 (TIMP-2) [130– 133]. A current hypothesis is that TIMP-2 functions as a bridge molecule between MT-MMP and MMP-2 by occupying the active site of MT-MMP via its amino terminal domain and binding the hemopexin domain of MMP-2 in a noninhibitory interaction involving the TIMP-2 carboxy terminal domain [134]. Thus, TIMP2 plays a dual role by both inhibiting MT-MMP activity and by bringing latent MMP-2 into close proximity with other nonoccupied MT-MMP molecules, thus stimulating cleavage of MMP-2 to the active form (Fig. 5). Functional implications of this ternary complex of proteases and inhibitor are multiple levels of control for proteolysis through altering the production of one or more of the components of the ternary complex, as well as the capability of cells to localize extracellular matrix proteolysis to specific domains of the cell surface, such as the leading edge (invadopodia) of a migrating cell [135]. Additionally, there is evidence that in selected cell types, cell surface receptors other than MT-MMPs, such as αvβ3 integrin, may bind MMP-2 via the MMP-2 hemopexin domain and thus function to catalyze its activation by presenting latent MMP-2 to adjacent MT-MMPs (Fig. 5) [129,136,137]. The link between extracellular matrix attachment and the proteolytic phenotype of a cell has been documented for several cell types. In rat microvascular endothelial cells, upregulation of MMP-2 and MT1-MMP occurs when cells are cultured on or within a collagen type I gel but not when grown on or within laminin-rich matrix (Matrigel Collaborative Research, Bedford, MA) [14]. Similar findings have been reported for fibroblastoid cells, which activate MMP-2 (consistent with upregulation of MT1-MMP) when cultured on type I collagen gels but not on fibronectin, collagen IV, or Matrigel [138]. These data suggest that particular matrix adhesion receptors may be responsible for initiating the signaling events that result in MMP production. Integrins are the major class of receptors that mediate adhesion to the extracellular matrix. These receptors are heterodimers composed of alpha and beta subunits, the combination of which provides each integrin with a unique range of specificities for a variety of matrix molecules [139,140]. Integrin engagement by extracellular matrix molecules has been associated with protease induction in a variety of cell types. The induction and activation of MMPs following integrin-mediated adhesion to selected extracellular matrix components has been shown to be cell specific. In synovial fibroblasts, α5β1 binding to the RGD sequence of fibronectin signals an induction of MMP-1; in contrast, α4β1 binding to the CS1 sequence of fibronectin suppresses MMP-1 transcription [141]. In other studies, we have shown that murine antigen-specific T-cell clones increase production of MMP-2 upon engagement of α4β1 via VCAM-1 or the CS1 peptide of fibronectin, whereas engagement of α5β1 does not induce MMP production [142,143]. Since rat microvascular endothelial cells are induced to express and activate MMP-2 and MT1-MMP during culture in type I collagen, but not in Matrigel, it seems reason-

300

Madri

Figure 5 Schematic representation of MT1-MMP/TIMP-2/pro-MMP-2 ternary complex formation and pro-MMP-2/α/β3 complex formation (A); Subsequent activation of tethered pro-MMP-2 by MT1-MMP and proteolysis of extracellular matrix by MMP-2 and MT1-MMP (B); Inhibition of MMP-2 and MT1-MMP when high concentrations of TIMP-2 are present (C). (From Ref. 129.)

Vasculogenesis and Angiogenesis

301

able that α2β1 integrin could play a role in initiating this process. However, given that the induction of MMP-2 and MT1-MMP occurs when the cells are cultured on or within a gel composed of type I collagen but does not occur when cells are grown on planar, thin coatings of type I collagen, the induction may not simply be the result of a specific ligand–receptor interaction but could involve mechanosensory force transduction [14,129,144]. In further attempts to elucidate the control of MMP-2 and MT-1 MMP expression, investigators have begun to analyze the noncoding promoter regions of these two genes. It was found that the gene encoding MMP-2 differs greatly from most other MMPs in that it lacks a TATA box, has no AP1 sites within the promoter region [145–147], and has regulatory features commonly associated with constitutively transcribed housekeeping genes. Thus, the induction of MMP2 when endothelial cells are undergoing angiogenesis in three-dimensional collagen matrices may be very distinct from the signaling pathways identified for other MMPs. Interestingly, characterization of the rat MMP-2 promoter activity in mesangial cells defined strong enhancer activity associated with a region of DNA that binds a transcription factor known as YB-1 [146,148]. Although this region also is critical for controlling transcription of MMP-2 in rat endothelial cells, a role for this factor in the collagen gel induction of MMP-2 transcription has not yet been confirmed (T. Haas and J. A. Madri, unpublished observations). In that MT-MMPs appear to be requisite in the activation of MMP-2, the transcriptional control of MT1-MMP might be critical in determining the extent of cell surface tethering and/or activation of MMP-2. The MT1-MMP gene has been found to be unresponsive to multiple growth factors such as epidermal growth factor (EGF), (basic fibroblast growth factor (bFGF), and TGF-β1 [149] and the rodent MT1-MMP gene is unresponsive to PMA treatment, which implies distinct species-specific regulatory elements in the promoter for MT1-MMP. Analysis of the murine MT1-MMP promoter region shows that this gene lacks a TATA box, similar to the MMP-2 gene, and that it also is unresponsive to growth factor stimuli such as TGF-β1 [144]; T. Haas and J. A. Madri, unpublished observations). A guanine cytosine (GC)-rich box present in a region of the promoter that is required for transcription activity contains overlapping consensus binding sites for the transcription factors Sp1 and Egr1. Increases in Egr1 protein were found to correlate with increased transcription of MT1-MMP in rat endothelial cells grown within a type I collagen matrix [144]. Fitting with this potential role for Egr1 in regulating MT1-MMP transcription, endothelial cell induction of Egr1 is known to be stimulated by potential initiators of angiogenesis such as wounding, mechanical stress, and fluid shear stress [150,151]. Thus, despite the coordinate upregulation of MMP-2 and MT1-MMP in a variety of cell types, it appears that distinct signaling pathways are stimulated in order to activate transcription of these two uniquely regulated genes. This dual control provides cells with the flexibility to dissociate the transcription of MMP-2 from MT1-MMP,

302

Madri

as has been observed in murine antigen-specific T-cell clones that constitutively produce MT1-MMP but induce transcription of MMP-2 only following engagement of the α4β1 integrin with its ligand [143]. Relevance of Genetic Approaches to Developmental, Adaptive, and Disease States

Current Status and Future Prospects. Because of their central importance in invasive processes, MMP inhibitors have been a topic of considerable clinical interest [128]. Many in vitro and in vivo animal studies have demonstrated the efficacy of a variety of MMP inhibitors [14,143]. However, owing to the relative lack of specificity, inability to target local tissue/organ sites, functional redundancy, and effects on normal tissue/organs, unwanted side effects and loss of clinical effectiveness have dampened enthusiasm somewhat. A potentially useful tool for further investigations is the recently described MMP-2–deficient mouse strain [152]. Although these animals develop normally without any apparent detectable anatomical abnormalities, tumor-induced angiogenesis in these animals is partially suppressed [152,153]. These findings suggest an intriguing possibility worthy of further investigation: that the events required for angiogenesis in the adult, such as matrix metalloproteinase induction, expression, and activity, may be different than those required for the initiation of embryonic vasculogenesis/angiogenesis and can be exploited in the future. III. Conclusions A major therapeutic issue in the angiogenesis field is the development and implementation of strategies and agents which will selectively modulate the upregulation, maintenance, and/or regression of new vessel formation in particular tissues/organs. Such reagents would enable the medical community to optimize vascularization at cutaneous graft sites, or eliminate vascularization following corneal injury and during tumor growth and metastasis. Thus, our understanding of underlying mechanisms in the area of angiogenesis should allow continued progress in our ability to beneficially modulate this critical aspect of the wound healing process. Acknowledgment This work was supported in part by USPHS Grants R37-HL28373, PO1KD38979, and RO1-HL51018. References 1. Clark RAF, ed. The Molecular and Cellular Biology of Wound Repair. New York: Plenum Press, 1996.

Vasculogenesis and Angiogenesis

303

2. Bicknell R, Lewis CE, Ferrara N, eds. Tumour Angiogenesis. Oxford, UK: Oxford University Press, 1997. 3. Nicosia RF, Madri JA. The microvascular extracellular matrix: Developmental changes during angiogenesis in the aortic ring plasma clot model. Am J Pathol 1987; 128:78–90. 4. Nicosia RF, Bonanno E, Smith M. Fibronectin promotes the elongation of microvessels during angiogenesis in vitro. J Cell Physiol 1993; 154:654–661. 5. Vernon RB, Sage EH. Between molecules and morphology: extracellular matrix and creation of vascular form. Am J Pathol 1995; 147:873–883. 6. Vernon RB, Sage EH. Contraction of fibrillar type I collagen by endothelial cells: a study in vitro. J Cell Biochem 1996; 60:185–197. 7. Madri JA, Williams SK. Capillary endothelial cell cultures: Phenotypic modulation by matrix components. J Cell Biol 1983; 97:153–165. 8. Madri JA, Pratt BM, Tucker AM. Phenotypic modulation of endothelial cells by transforming growth factor b depends upon the composition and organization of the extracellular matrix. J Cell Biol 1988; 106:1375–1384. 9. Montesano R, Orci L. Phorbol esters induce angiogenesis in vitro from large vessel endothelial cells. J Cell Physiol 1987; 130:284–291. 10. Ingber DE, Prusty D, Frangioni JV, Crague EJ, Lechene C, Schwartz M. Control of intracellular pH and growth by fibronectin in capillary endothelial cells. J Cell Biol 1990; 110:1803–1811. 11. Davis GE, Camarillo CW. Regulation of endothelial cell morphogenesis by integrins, mechanical forces and matrix guidance pathways. Exp Cell Res 1995; 216: 113–123. 12. Davis GE, Camarillo CW. An alpha2 beta1 integrin dependent pinocytic mechanism involving intracellular vacuole formation and coalescence regulates capillary lumen and tube formation in three-dimensional collagen matrix. Exp Z Cell Res 1996; 224:39–51. 13. DeLisser HM, Christofidou-Solomidou M, Strieter RM, Burdick MD, Robinson CS, Wexler RS, Kerr JS, Garlanda C, Merwin JR, Madri JA, Albelda SM. Involvement of endothelial PECAM-1/CD31 in angiogenesis. Am J Pathol 1997; 151:671– 677. 14. Haas TL, Davis SJ, Madri JA. Three dimensional type I collagen lattices induce coordinate expression of matrix metalloproteinases MT1-MMP and MMP-2 in microvascular endothelial cells. J Biol Chem 1998; 273:3604–3610. 15. Sankar S, Mahooti-Brooks N, Centrella M, McCarthy TL, Madri JA. Expression of transforming growth factor beta type III receptor in vascular endothelial cells increases their responsiveness to transforming growth factor b2. J Biol Chem 1995; 270:13567–13572. 16. Sankar S, Mahooti-Brooks N, Bensen L, Centrella M, McCarthy TL, Madri JA. Modulation of transforming growth factor beta receptor expression in microvascular endothelial cells during in vitro angiogenesis. J Clin Invest 1996; 97:1436–1446. 17. Capecchi MR. Altering the genome by homologous recombination. Science 1989; 244:1288–1292. 18. Mansour SL, Thomas KR, Capecchi MR. Disruption of the proto-oncogene int-2 in mouse embryo-derived stem cells: A geneeral strategy for targeting mutations to non-selectable genes. Nature 1988; 336:348–352.

304

Madri

19. Mansour SL, Thomas KR, Deng C, Capecchi MR. Introduction of a lacZ reporter gene into the mouse int-2 locus by homologous recombination. Proc Natl Acad Sci USA 1990; 87:7688–7692. 20. Thomas KR, Capecchi MR. Targeted disruption of the murine int-1 proto-ocnogene resulting in severe abnormalities in midbrain and creebellar development. Nature 1990; 346:847–850. 21. Wong EA, Caqpecchi MR. Homologous recombination between coinjected DNA sequences peaks in early to mid-S phase. Mol Cell Biol 1987; 7:2294–2295. 22. Gale NW, Yancopoulos GD. Growth factors acting via endothelial cell-specific receptor tyrosine kinases: VEGFs, angiopoietins, and ephrins in vascular development. Genes Dev 1999; 16:1055–1066. 23. Holder N, Klein R. Eph receptors and ephrins: effectors of morphogenesis. Development 1999; 126:2033–2044. 24. Hanahan D. Signaling vascular morphogenesis and maintenance. Science 1997; 277:48–50. 25. Folkman J, D’Amore PA. Blood vessel formation: What is its molecular basis? Cell 1996; 87:1153–1155. 26. Risau W, Lemmon V. Changes in the vascular extracellular matrix during embryonic vasculogenesis and angiogenesis. Dev Biol 1988; 125:441–450. 27. Risau W. Mechanisms of angiogenesis. Nature 1997; 386:671–674. 28. Shalaby F, Rossant J, Yamaguchi TP, Gertenstein M, Wu X, Breitman ML, Schuh AC. Failure of blood-island formation and vasculogenesis in Flk-1–deficient mice. Nature 1995; 376:62–66. 29. Shalaby F, Ho J, Stanford WL, Fischer K, Schuh AC, Schwartz L, Bernstein A, Rossant J. A requirement for Flk1 in primitive and definitive hematopoiesis and vasculogenesis. Cell 1997; 89:981–990. 30. Carmeliet P, Ferreira V, Breier G, Pollefeyt S, Kieckens L, Gertsenstein M, Fahrig M, Vandenhoeck A, Harpal K, Eberhardt C, Declercq C, Pawling J, Moons L, Collen D, Risau W, Nagy A. Abnormal blood vessel development and lethality in embryos lacking a single VEGF allele. Nature 1996; 380:435–439. 31. Ferrara N, Carver-Moore K, Chen H, Dowd M, Lu L, O’Shea KS, Powell-Braxton L, Hillan KJ, Moore MW. Heterozygous embryonic lethality induced by targeted inactivation of the VEGF gene. Nature 1996; 380:439–442. 32. Fong G, Rossant J, Gertenstein M, Breitman ML. Role of the Flt-1 receptor tyrosine kinase in regulating the assembly of vascular endothelium. Nature 1995; 376: 66–70. 33. Fong G, Zhang L, Bryce D, Peng J. Increased hemangioblast commitment, not vascular disorganization, is the primary defect in flt-1 knock-out mice. Development 1999; 126:3015–3025. 34. Hiratuska S, Minowa O, Kuno J, Noda T, Shibuya M. Flt-1 lacking the tyrosine kinase domain is sufficient for normal development and angiogenesis in mice. Proc Natl Acad Sci USA 1998; 95:9349–9354. 35. Dumont DJ, Jussila L, Taipale J, Lymboussaki A, Mustonen T, Pajusola K, Breitman M, Alitalo K. Cardiovascular failure in mouse embryos deficient in VEGF receptor-3. Science 1998; 282:946–949. 36. Dvorak HF, Nagy JA, Feng D, Brown LF, Dvorak AM. Vascular permeability

Vasculogenesis and Angiogenesis

37.

38. 39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

305

factor/vascular endothelial growth factor and the significance of microvascular hypermeability in angiogenesis. Curr Top Microbiol Immunol 1999; 237:97– 132. Asahara A, Takahashi T, Masuda H, Kalka C, Chen D, Iwaguro H, Inai Y, Silver M, Isner JM. VEGF contributes to postnatal neovascularization by mobilizing bone marrow-derived endothelial progenitor cells. EMBO J. 1999; 18:3964–3972. Isner JM, Asahara T. Angiogenesis and vasculogenesis as therapeutic strategies for postnatal neovascularization. J Clin Invest 1999; 103:1231–1236. Alon T, Hemo I, Pe’er J, Stone J, Keshet E. Vascular endothelial growth factor as a survival factor for newly formed retinal vessels and ahs implications for retinopathy or prematurity. Nat Med 1995; 1:1024–1028. Asahara T, Takahashi T, Masuda H, Kalka C, Chen D, Iwaguro H, Inai Y, Silver M, Isner JM. VEGF contributes to postnatal neovascularization by mobilizing bone marrow-derived endothelial progenitor cells. EMBO J 1999; 18:3964–3972. Ment LR, Stewart WB, Fronc R, Seashore C, Mahooti S, Scaramuzzino D, Madri JA. Vascular endothelial growth factor mediates reactive angiogenesis in the postnatal developing brain. Dev Brain Res 1997; 100:52–61. Ogunshola O, Stewart WB, Mihalcik V, Solli T, Madri JA, Ment LR. Neuronal VEGF mediates angiogenesis in postnatal developing rat brain. Dev Brain Res 2000; 119:139–153. Ilan N, Mahooti S, Madri JA. Distinct signal transduction pathways are utilized during the tube formation and survival phases of in vitro angiogenesis. J Cell Sci 1998; 111:3621–3631. Gerber HP, McMurtrey A, Kowalski J, Yan M, Keyt BA, Dixit V, Ferrara N. Vascular endothelial growth factor regulates endothelial cell survival through the phosphoatidylinositol 3′-kinase/Akt signal transduction pathway: requirement for Flk1/KDR activation. J Biol Chem 1999; 273:30336–30346. Waltenberger J, Claesson-Welsh L, Siegbahn A, Shibuya M, Heldin C-H. Different signal transduction properties of KDR and Flt-1, two receptors for vascular endothelial growth factor. J Biol Chem 1994; 269:26988–26995. Wilting J, Eichmann A, Christ B. Expression of the avain VEGF receptor homologues Quek1 and Quek2 in blood-vascular and lymphatic endothelial and nonendothelial cells during quail embryonic development. Cell Tissue Res 1997; 288: 207–223. Soker S, Takashima S, Miao HQ, Neufeld G, Klagsbrun M. Neuropilin-1 is expressed by endothelial and tumor cells as an isoform-specific receptor of vascular endothelial growth factor. Cell 1999; 92:735–745. Sato TN, Tozawa Y, Deutsch U, Wolburg-Buchholz K, Fujiwara Y, Gendron-Maguire M, Gridley T, Wolburg H, Risau W, Qin Y. Distinct role of the receptor tyrosine kinases Tie-1 and Tie-2 in blood vessel formation. Nature 1995; 376: 70–74. Suri C, Jones PF, Patan S, Bartunkova S, Maisonpierre PC, Davis S, Sato TN, Yancopoulos GD. Requisite role of angiopoietin-1, a ligand for the TIE2 receptor, during embryonic angiogenesis. Cell 1996; 87:1171–1180. Maisonpoerre PC, Suri C, PF Jones, Bartunkova S, Wiegand SJ, Radziejewski C, Compton D, McClain S, Aldrich TN, Papadopoulos N, Daly TJ, Davis S, Sato TN,

306

51.

52.

53.

54. 55.

56. 57.

58.

59.

60.

61.

62.

63.

64.

Madri Yancopoulos GD. Angiopoietin-2, a natural antagonist for Tie2 that disrupts in vivo angiogenesis. Science 1997; 277:55–60. Davis S, Aldrich TH, Jones PF, Acheson A, Compton DL, Jain V, Ryan TE, Bruno J, Radziejewski C, Maisonpierre PC, Yancopoulos GD. Isolation of angiopoietin1, a ligand for the TIE2 receptor, by secretion-trap expression cloning. Cell 1996; 87:1161–1169. Yabkowitz R, Meyer S, Yanagihara D, Brankow D, Staley T, Elliot G, Hu S, Ratzkin B. Regulation of tie receptor expression on human endothelial cells by protein kinase C-mediated release of soluble tie. Blood 1997; 90:706–715. Yabkowitz R, Meyer S, Black T, Elliott G, Merewether LA, Yamane HK. Inflammatory cytokines and vascular endothelial growth factor stimulate the release of soluble Tie receptor from human endothelial cells via metalloprotease activation. Blood 1999; 93:1969–1979. Vecchi M, Carpenter G. Constitutive proteolysis of the ErbB-4 receptor tyrosine kinase by a unique, sequential mechanism. J Cell Biol 1997; 139:995–1003. Cabrera N, Diaz-Rodriguez E, Becker E, Martin-Zanca D, Pandiella A. TrkA receptor ectodomain cleavage generates a tyrosine-phosphorylated cell–associated fragment. J Cell Biol 1996; 132:427–436. Ilan N, Madri JA. New paradigms of signaling in the vasculature: ephrins and metalloproteases. Curr Opin in Biotech 1999; 10:536–540. McCarthy MJ, Burrows R, Bell SC, Christie G, Bell PRF, Brindle NPJ. Potential roles of metalloprotease mediated ectodomain cleavage in signaling by the endothelial receptor tyrosine kinase tie-1. Lab. Invest. 1999; 79:889–895. Shawver LK, Lipson KE, Fong AT, McMahon G, Plowman GD, Strawn LM. Receptor tyrosine kinases as targets for inhibition of angiogenesis. DDT 1997; 2:50–63. Vikkula M, Boon LM, Carraway III KL, Calvert JT, Diamont AJ, Goumnerov B, Pasyk KA, Marchuk DA, Warman ML, Cantley LC, Muliken JB, Olsen BR. Vascular dysmorphogenesis caused by an activating mutation in the receptor tyrosine kinase TIE2. Cell 1996; 87:1181–1190. Davis S, Gale NW, Aldrich TH, Maisonpierre PC, Lhotak V, Pawson T, Goldfarb M, Yancopoulos GD. Ligands for Eph-related receptor tyrosine kinases that require membrane attachment or clustering for activity. Science 1994; 266:816–819. Holzman LB, Marks RM, Dixit VM. A novel immediate-early response gene of endothelium is induced by cytokines and encodes a secreted protein. Mol Cell Biol 1990; 10:5830–5838. Shao H, Pandey A, O‘Shea KS, Seldin M, Dixit VM. Characterization of B61, the ligand for the Eck receptor protein-tyrosine kinase. J Biol Chem 1995; 270:5636– 5641. Sarma V, Wolf FW, Marks RM, Shows TB, Dixit VM. Clonning of a novel tumor necrosis factor alpha-induced primary response gene that is differentially expressed in development and capillary tube-like formation in vitro. J Immunol 1992; 148: 3302–3312. Pandey A, Shao H, Marks RM, Polverini PJ, Dixit VM. Role of B61, the ligand for the Eck receptor tyrosine kinase, in TNF-α–induced angiogenesis. Science 1995; 268:567–569.

Vasculogenesis and Angiogenesis

307

65. Daniel TO, Stein E, Cerretti DP, St. John PL, Robert B, Abrahamson DR. ELK and LERK-2 in developing kidney and microvascular endothelial assembly. Kidney International 1996; 50(57):S-73–S-81. 66. Adams RH, Wilkinson GA, Weiss C, Diella F, Gale NW, Deutsch U, Risau W, Klein R. Roles of ephrinB ligands and EphB receptors in cardiovascular development: demarcation of arterial/venous domains, vascular morphogenesis, and sprouting angiogenesis. Genes Dev 1999; 13:295–306. 67. Wang HU, Chen Z, Anderson DJ. Molecular distinction and angiogenic interaction between embryonic arteries and veins revealed by ephrin-B2 and its receptor EphB4. Cell 1998; 93:741–753. 68. Pinter E, Barreuther M, Lu TT, Imhof B, Madri JA. PECAM-1/CD31 tyrosine phosphorylation state changes during vasculogenesis. Am J Pathol 1997; 150: 1523–1530. 69. Pinter E, Mahooti S, Wang Y, Imhof BA, Madri JA. Hyperglycemia-induced vasculopathy in the murine vitelline vasculature: correlation with PECAM-1/CD31 tyrosine phosphorylation state. Am J Pathol 1999; 154:1367–1379. 70. Bruckner K, Pasquale EB, Klein R. Tyrosine phosphorylation of transmembrane ligands for Eph receptor. Science 1997; 275:1640–1643. 71. Holland SJ, Gale NW, Mbamalu G, Yancopoulos GD, Henkemeyer M, Pawson T. Bidirectional signalling through the Eph-family receptor Nuk and its transmembrane ligands. Nature 1996; 383:722–725. 72. Stein E, Lane AA, Cerretti DP, Schoeckimann HO, Schroff AD, Van Etten RL, Daniel TO. Eph receptors discriminate specific ligand oligomers to determine alternative signaling complexes, attachment and assembly responses. Genes Dev 1998b; 12:667–678. 73. Huynh-Do U, Stein E, Lane AA, Liu H, Cerretti DP, Daniel TO. Surface densities of ephrin-B1 determine EphB1-coupled activation of cell attachement through αvβ3 and α5β1 integrins. EMBO J 1999; 18:2165–2173. 74. Xu T, Bianco P, Fisher LW, Longenecker G, Smith E, Goldstein S, Bonadio J, Boskey A, Heegaard AM, Sommer B, Satomura K, Dominguez P, Zhao C, Kulkarni AB, Robey PG, Young MF. Trageted disruption of the biglycan gene leads to an osteoporosis-like phenotype in mice. Nat Genet 1998; 20:78–82. 75. Letterio JJ, Roberts AB. Transforming growth factor-beta1–deficient mice: identification of isoform-specific activities in vivo. J Leukoc Biol 1996; 59:769– 74. 76. Oshima M, Oshima H, Taketo MM. TGF-beta receptor type II deficiency results in defects of yolk sac hematopoiesis and vasculogenesis. Dev Biol 1996; 179:297– 302. 77. Bonyadi M, Rusholme SA, Cousins FM, Su HC, Biron CA, Farrall M, Akhurst RJ. Mapping of a major genetic modifier of embryonic lethality in TGF beta 1 knockout mice. Nat Genet 1997; 15:207–211. 78. Kulkarni AB, Huh CG, Becker D, Geiser A, Lyght M, Flanders KC, Roberts AB, Sporn MB. Transforming growth factor beta 1 null mutation in mice causes excessive inflammatory response and early death. Proc Natl Acad Sci USA 1993; 90: 770–774. 79. Li DY, Sorensen LK, Brooke BS, Urness LD, Davis EC, Taylor DG, Boak BB,

308

80. 81. 82.

83. 84.

85.

86.

87.

88.

89.

90.

91.

92. 93.

94.

Madri Wendel DP. Defective angiogenesis in mice lacking endoglin. Science 1999; 284: 1534–1537. Madri JA, Kocher O, Merwin JR, Basson CT, Bell L. The interactions of vascular cells with transforming growth factor β. NY Acad Sci 1990; 593:243–258. Madri JA, Bell L, Merwin JR. Modulation of vascular cell behavior by transforming growth factors beta. Molecular Reprod Dev 1992; 32:121–126. Merwin JR, Newman W, Beall D, Tucker A, Madri JA. Vascular cells respond differentially to transforming growth factors-beta1 and beta2 in vitro. Am J Pathol 1991; 138:37–51. Heldin C-H, Miyazano K, ten Dijke P. TGF-β signalling from cell membrane to nucleus through SMAD proteins. Nature 1997; 390:465–471. Brown CB, Boyer AS, Runyan RB, Barnett JV. Requirement of type III TGF-β receptor for endocardial cell transformation in the heart. Science 1999; 283:2080– 2082. Yang X, Castilla LH, Xu X, Li C, Gotay J, Weinstein M, Liu P, Deng C-X. Angiogenesis defects and mesenchymal apoptosis in mice lacking SMAD5. Development 1999; 126:1571–1580. Boyer AS, Ayerinskas I, Vincent EB, McKinney LA, Weeks DL, Runyan RB. TGFbeta2 and TGFbeta3 have separate and sequential activities during epithelialmesenchymal cell transformation in the embyronic heart. Dev Biol 1999; 208:530– 545. Sheffield VC, Pierpoint ME, Nishimura D, Beck JS, Burns TL, Berg MA, Stone EM, Patil SR, Lauer RM. Identification of a complex congenital heart defect susceptibility locus by using DNA pooling and shared segment analysis. Hum Mol Genet 1997; 6:117–121. Cohen MP, Sharma K, Guo J, Eltayeb BO, Ziyadeh FN. The renal TGF beta system in the db/db mouse model of diabetic nephropathy. Exp Nephrol 1998; 6:226– 233. Cheifetz S, Bellon T, Cales C, Vera S, Bernabeu C, Massague J, Letarte M. Endoglin is a component of the transforming growth factor-beta receptor system in human endothelial cells. J Biol Chem 1992; 267:19027–19030. McAllister KA, Grogg KM, Johnson DW, Gallione CJ, Baldwin MA, Jackson CE, Helmbold EA, Markel DS, McKinnon WC, Murrell J, McCornmick MK, PericakVance MA, Heutink P, Oostra BA, Haitjema T, Westerman CJJ, Porteous ME, Guttmacher AE, Letarte M, Marchuk DA. Endoglin, a TGF-b binding protein of endothelial cells, is the gene for hereditary haemorrhagic telangiectasia type 1. Nat Genet 1994; 8:345–351. Chang H, Huylebroeck D, Verschueren K, Guo Q, Matzuk MM, Zwijsen A. Smad5 knockout mice die at mid-gestation due to multiple embryonic and extraembryonic defects. Development 1999; 126:1631–1642. Yang JT, Rayburn H, Hynes RO. Embryonic mesodermal defects in α5 integrin– deficient mice. Development 1993; 119:1093–1105. Yang JT, Rayburn H, Hynes RO. Cell adhesion events mediated by α4 integrins are essential in placental and cardiac development. Development 1995; 121:549– 560. George EL, Georges-Labouesse EN, Patel-King RS, Rayburn H, Hynes RO. De-

Vasculogenesis and Angiogenesis

95.

96.

97.

98.

99.

100.

101. 102. 103.

104.

105.

106.

107.

309

fects in mesoderm, neural tube and vascular development in mouse embryos lacking fibronectin. Development 1993; 119:1079–1091. Kwee L, Baldwin HS, Shen HM, Stewart CL, Buck C, Buck CA, Labow MA. Defective development of the embryonic and extraembryonic circulatory systems in vascular cell adhesion molecule (VCAM-1) deficient mice. Development 1995; 121:489–503. Carmeliet P, Mackman N, Moons L, Luther T, Gressens P, Van Vlaenderen I, Demunck H, Kasper M, Breier G, Evrad P, Muller M, Risau W, Edgington T, Collen D. Role of tissue factor in embryonic blood vessel development. Nature 1996; 383:73–75. Toti P, Buonocore G, Tanganelli P, Catella AM, Palmeri ML, Vatti R, Seemayer TA. Bronchopulmonary dysplasia of the premature baby: an immunohistochemical study. Pediatr Pulmonol 1997; 24:22–28. Wu L, Xia YP, Roth SI, Gruskin E, Mustoe TA. Transforming growth factor-beta1 fails to stimulate wound healing and impairs its signal transduction in an aged ischemic ulcer model: importance of oxygen and age. Am J Pathol 1999; 154:301– 309. Breier G, Albrecht U, Sterrer S, Risau W. Expression of vascular endothelial growth factor during embryonic angiogenesis and endothelial cell differentiation. Development 1992; 114:521–532. Navarro P, Ruco L, Dejana E. Differential localization of VE- and N-cadherins in human endothelial cells: VE-cadherin competes with N-cadherin for junctional localization. J Cell Biol 1998; 140:1475–1484. Huber O, Bierkamp C, Kemler R. Cadherins and catenins in development. Curr Opin Cell Biol 1996; 8:685–691. Lampugnani MG, Dejana E. Interendothelial junctions: structure, signaling and function. Curr Opin Cell Biol 1997; 9:674–682. Barth AIM, Nathke IS, Nelson WJ. Cadherins, catenins and APC protein: interplay between cytoskeletal complexes and signaling pathways. Curr Opin Cell Biol 1997; 9:683–690. Vittet D, Buchou T, Schweitzer A, Dejana E, Huber P. Targeted null-mutation in the vascular endothelial-cadherin gene impairs the organization of vascularlike sturctures in embryoid bodies. Proc Natl Acad Sci USA 1997; 94:6273– 6278. Gory-Faure S, Prandini M-H, Pointu H, Rouliot V, Pignot-Paintrand I, Vernet M, Huber P. Role of vascular endothelial-cadherin in vascular morphogenesis. Development 1999; 126:2093–2102. Carmeliet P, Lampugnani M-G, Moons L, Breviario F, Compernolle V, Bono F, Balconi G, Spagnuolo R, Oosthuyse B, Dewerchin M, Zanetti A, Angellilo A, Maattot V, Nuyens D, Lutgens E, Clotman F, de Ruiter MC, Gittenberger-de Groot A, Poelmann R, Lupu F, Herbert J-M, Collen D, Dejana E. Targeted deficiency or cytosolic truncation of the VE-cadherin gene in mice impairs VEGF-mediated endothelial survival and angiogenesis. Cell 1999; 98:147–157. Hoschuetzky H, Aberle H, Kemler R. Beta-catenin mediates the interaction of the cadherin-catenin complex with epidermal growth factor receptor. J Cell Biol 1994; 127:1375–1380.

310

Madri

108. Newman PJ. Switched at birth: A new family for PECAM-1. J Clin Invest 1999; 103:5–9. 109. Chosay JG, Fisher MA, Farhood A, Ready KA, Dunn CJ, Jaeschke H. Role of Pecam-1 (Cd31) in neutrophil transmigration in murine models of liver and peritoneal inflammation. Am J Physiol Gastrointest Liver Physiol 1998; 3 7:G776–G782. 110. Vaporciyan AA, DeLisser HM, Yan HC, Mendiguren II, Thom SR, Jones ML, Ward PA, Albelda SM. Involvement of platelet-endothelial cell adhesion molecule1 in neutrophil recruitment in vivo. Science 1993; 262:1580–1582. 111. Bogen S, Pak J, Garifallou M, Deng X, Muller WA. Monoclonal antibody to murine PECAM-1 (CD31) blocks acute inflammation in vivo. J Exp Med 1994; 179:1059– 1064. 112. Elias CG, Spellberg JP, Karantamir B, Lin CH, Wang YJ, McKenna PJ, Muller WA, Zukowski MM, Andrew DP. Ligation of Cd31/Pecam-1 modulates the function of lymphocytes, monocytes and neutrophils. Eur J Immunol 1998; 28:1948– 1958. 113. Kalra VK, Shen Y, Sultana C, Rattan V. Hypoxia induces PECAM-1 phosphorylation and transendothelial migration of monocytes. Am J Physiol 1996; 271:H2025– H2034. 114. Lu TT, Yan LG, Madri JA. Integrin engagement mediates tyrosine dephosphorylation on platelet-endothelial cell adhesion molecule 1. Proc Natl Acad Sci USA 1996; 93:11808–11813. 115. Lu TT, Barreuther M, Davis S, Madri JA. Platelet endothelial cell adhesion molecule-1 is phosphorylatable by c-Src, binds Src-Src homology 2 domain, and exhibits immunoreceptor tyrosine-based activation motif-like properties. J Biol Chem 1997; 272:14442–14446. 116. Kim CS, Wang T, Madri JA. Platelet endothelial cell adhesion molecule-1 expression modulates endothelial cell migration in vitro. Lab Invest 1998; 78:583– 590. 117. Bogen SA, Baldwin HS, Watkins SC, Albelda SM, Abbas AK. Association of murine CD31 with transmigrating lymphocytes following antigenic stimulation. Am J Pathol 1992; 141:843–54. 118. Varon D, Jackson DE, Shenkman B, Dardik R, Tamarin I, Savion N, Newman PJ. Platelet/endothelial cell adhesion molecule-1 serves as a costimulatory agonist receptor that modulates integrin-dependent adhesion and aggregation of human platelets. Blood 1998; 91:500–507. 119. Rainger GE, Buckley C, Simmons DL, Nash GB. Cross-talk between cell adhesion molecules regulates the migration velocity of neutrophils. Curr Biol 1997; 7:316– 325. 120. Pellegatta F, Chierchia SL, Zocchi MR. Functional association of platelet endothelial cell adhesion molecule-1 and phosphoinositide 3-kinase in human neutrophils. J Biol Chem 1998; 273:27768–27771. 121. Sagawa K, Kimura T, Swieter M, Siraganian RP. The protein-tyrosine phosphatase Shp-2 associates with tyrosine-phosphorylated adhesion molecule Pecam-1 (Cd31). J Biol Chem 1997; 272:31086–31091. 122. Gurubhagavatula I, Amrani Y, Pratico D, Ruberg FL, Albelda SM, Panettieri RA. Engagement of human Pecam-1 (Cd31) on human endothelial cells increases intra-

Vasculogenesis and Angiogenesis

123.

124.

125.

126. 127. 128. 129.

130.

131.

132.

133.

134.

135.

136.

137.

311

cellular calcium ion concentration and stimulates prostacyclin release. J of Clin Invest 1998; 101:212–222. Ilan N, Mahooti S, Rimm DL, Madri JA. PECAM-1 (CD31) functions as a reservoir for and a modulator of tyrosine phosphorylated beta catenin. J. Cell Sci. 1999; 112: 3005–3014. Duncan GS, Andrew DP, Takimoto H, Kaufman SA, Yoshida H, Spellberg J, Luis de la Pompa J, Elia A, Wakeham A, Karan-Tamir B, Muller WA, Senaldi G, Zukowski MM, Mak TW. Genetic evidence for functional redundancy of platelet/ endothelial cell adhesion molecule-1 (PECAM-1): CD31-deficient mice reveal PECAM-1–dependent and PECAM-1–independent functions. J Immunol 1999; 162:3022–3030. Thompson RD, Larbi KY, Harrison V, Duncan GS, Mak TW, Nourshargh S. IL1B–induced leukocyte responses in cremasteric venules of mice deficient in PECAM-1. FASEB J 1999; 13:A670. Hynes RO. Targeted mutations in cell adhesion genes: What have we learned from them? Dev Biol 1996; 180:402–412. Nagase H, Woessner JF. Matrix metalloproteinases. J Biol Chem 1999; 274:21491– 21494. Stetler-Stevenson WG. Matrix metalloproteinases in angiogenesis: a moving target for therapeutic intervention. J Clin Invest 1999; 103:1237–1241. Haas TL, Madri JA. Extracellular matrix-driven matrix metalloproteinase production in endothelial cells: implications for angiogenesis. Trends Cardiovasc Med 1999; 9:70–77. Cao J, Sato H, Takino T, Seiki M. The C-terminal region of membrane type matrix metalloproteinase is a functional transmembrane domain required for pro-gelatinase A activation. J Biol Chem 1995; 270:801–805. Pei D, Weiss SJ. Transmembrane-deletion mutants of the membrane-type matrix metalloproteinase-1 process progelatinase A and express intrinsic matrix-degrading activity. J Biol Chem 1996; 271:9135–9140. Strongin AY, Collier I, Bannikov G, Marmer BL, Grant GA, Goldberg GI. Mechanism of cell surface activation of 72-kDa type IV collagenase. J Biol Chem 1995; 270:5331–5338. Corcoran ML, Emmert-Buck MR, McClanahan JL, Pelina-Parker M, Stetler-Stevenson WG. TIMP-2 mediates cell surface binding of MMP-2. Adv Exp Med Biol 1996; 389:295–304. Zucker S, Drews M, Conner C, Foda HD, DeClerck YA, Langley KE, Bahou WF, Docherty AJP, Cao J. Tissue inhibitor of metalloproteinase-2 (TIMP-2) binds to the catalytic domain of the cell surface receptor, membrane type 1-matrix metalloproteinase 1 (MT1-MMP). J Biol Chem 1998; 273:1216–1222. Madri JA, Graesser D. Cell migration in the immune system: the evolving interrelated roles of adhesion molecules and proteinases. Dev Immunol 2000, In Press. Brooks PC, Stromblad S, Sanders LC, et al. Localization of matrix metalloproteinase MMP-2 to the surface of invasive cells by interaction with integrin alphaV beta3. Cell 1996; 85:683–693. Brooks PC, Silletti S, von Schalscha TL, Friedlander M, Cheresh DA. Disruption

312

138.

139. 140. 141.

142. 143.

144.

145. 146.

147. 148.

149.

150. 151.

152.

153.

Madri of angiogenesis by PEX, a noncatalytic metalloproteinase fragment with integrin binding activity. Cell 1998; 92:391–400. Azzam HS, Thompson EW. Collagen-induced activation of the Mr 72,000 type IV collagenase in normal and malignant human fibroblastoid cells. Cancer Res 1992; 52:4540–4544. Hynes RO. Integrins: versatility, modulation and signaling in cell adhesion. Cell 1992; 69:11–25. Stromblad S, Cheresh DA. Integrins, angiogenesis and vascular cell survival. Chem Biol 1996; 3:881–885. Huhtala P, Humphries MJ, McCarthy JB, Tremble PM, Werb Z, Damsky CH. Cooperative signalling by alpha5 beta1 and alpha4 beta1 integrins regulates metalloproteinase gene expression in fibroblasts adhering to fibronectin. J Cell Biol 1995; 129:867–879. Romanic AM, Madri JA. The induction of 72-kD gelatinase in T cells upon adhesion to endothelial cells is VCAM-1 dependent. J Cell Biol 1994; 125:1165–1178. Graesser D, Mahooti S, Haas T, Davis S, Clark R, Madri JA. The interrelationship of α4 integrin and matrix metalloproteinase-2 in the pathogenesis of experimental autoimmune encephalomyelitis. Lab Invest 1998; 78:1445–1458. Haas TL, Stitelman D, Davis SJ, Apte SS, Madri JA. Transcriptional upregulation of the membrane type matrix metalloproteinase-1 in response to extracellular cues involves Egr-1. J Biol Chem 1999; 274:22679–22685. Huhtala P, Chow LT, Tryggvason K. Structure of the human type IV collagenase gene. J Biol Chem 1990; 265:11077–11082. Harendza S, Polluck AS, Mertens PR, Lovett DH. Tissue-specific enhancer-promoter interactions regulate high level constitutive expression of matrix metalloproteinase 2 by glomerular mesangial cells. J Biol Chem 1995; 270:18786–18796. Corcoran ML, Hewitt RE, Kleiner Jr DE, Stetler-Stevenson WG. MMP-2: expression, activation and inhibition. Enzyme Protein 1996; 49:7–19. Mertens PR, Alfonso-Jaume A, Steinmann K, Lovett DH. A synergistic interaction of transcription factors AP-2 and YB-1 regulates gelatinase A enhancer–dependent transcription. J Biol Chem 1998; 273:32957–32965. Lohi L, Lehti K, Westermarck J, Kahari VM, Keski-Oja J. Regulation of membrane-type matrix metalloproteinase-1 expression by growth factors and phorbol 12-myristate 13-acetate. Eur J Biochem 1996; 239:239–247. Khachigian LM, Linder V, Williams AJ, Collins T. Egr-1-induced endothelial gene expression: A common theme in vascular injury. Science 1996; 271:1427–1431. Schwachtgen JL, Houston P, Campbel IC, Sukhatme V, Braddock M. Fluid shear stress activation of egr-1 transcription in cultured human endothelial and epithelial cells is mediated via the extracellular signal-related kinase 1/2 mitogen-activated protein kinase pathway. J Clin Invest 1998; 101:2540–2549. Itoh T, Ikeda T, Gomi H, Nakao S, Suzuki T, Itohara S. Unaltered secretion of beta-amyloid precursor protein in gelatinase A (matrix metalloproteinase 2)–deficient mice. J Biol Chem 1997; 272:22389–22392. Itoh T, Tanioka M, Yoshida H, Yoshioka T, Nishimoto H, Itohara S. Reduced angiogenesis and tumor progression in gelatinase A–deficient mice. Cancer Res 1998; 58:1048–1051.

13 Mouse Genetic Models in Atherosclerosis and Lipid Metabolism Research

KAREN REUE University of California, Los Angeles, and Veterans Administration Greater Los Angeles Healthcare System Los Angeles, California

I.

Introduction

The study of genetic variants has been instrumental in the elucidation of many biological processes and biochemical pathways. In the field of atherosclerosis and lipid metabolism, a classic example is the identification of the low-density lipoprotein (LDL) receptor in familial hypercholesterolemia patients in which a genetic defect in the synthesis of the receptor results in the characteristic high LDL cholesterol levels and premature atherosclerosis [1]. In the general population, however, atherosclerosis is not a monogenic disorder as seen in familial hypercholesterolemia but rather a consequence of the interaction of numerous genetic and environmental factors, each exerting a positive or negative influence on susceptibility to the disease. Atherosclerosis can be considered to result from chronic injury to the arterial wall that is often initiated by hypercholesterolemia [2]. Cholesterol-rich lipoproteins derived from the plasma may become modified through oxidation and subsequently retained in the arterial wall. The accumulation of lipoproteins triggers an inflammatory response leading to endothelial cell activation and attraction of blood monocytes, which may pass into the artery wall and differentiate into macrophages. The macrophages express receptors that recognize modified lipoproteins and accumulate lipoprotein-derived cholesterol 313

314

Reue

in the cytoplasm to produce ‘‘foam cells’’ that are characteristic of the early atheromatous lesion, known as a fatty streak. Progression of the lesion involves proliferation of arterial smooth muscle cells and death of foam cells with deposition of their cholesterol contents extracellularly. These events produce a more advanced lesion in which a cholesterol-rich core is covered by a fibrous cap composed of smooth muscle cells. Eventually, the surface of the lesion may rupture to expose the subendothelial layer and potentially initiate a thrombotic event leading to tissue ischemia such as an acute myocardial infarction. Among the factors that influence the progression of atherosclerosis, plasma levels of lipoproteins are widely accepted as playing a major role. Elevated levels of LDL and intermediate-density lipoprotein (IDL) are associated with an increased risk for atherosclerosis, which is related to the role of these particles in cholesterol deposition in the artery wall. High levels of lipoprotein (a) Lp(a), an abnormal lipoprotein in which the major protein of LDL is complexed to another protein known as apolipoprotein (a) Apo(a), are also associated with increased risk. In contrast, high-density lipoprotein (HDL) levels are protective against atherosclerosis. This property is attributed largely to the role of this lipoprotein in ‘‘reverse cholesterol transport,’’ a process whereby cholesterol is removed from peripheral cells and transported to the liver where it can be metabolized. Thus, an individual’s susceptibility to atherosclerosis depends upon the interplay of multiple factors that each exert a positive or negative influence on the susceptibility to the disease. Owing to the complexity of atherogenesis, genetic studies in humans have been limited, and many important insights have been gained from studying genetic variants in experimental animal models. Animal models offer valuable advantages for studies of genetic factors in atherosclerosis, including the ability to control and manipulate both environmental factors such as diet and genetic factors through selective breeding. Genetic variants in atherosclerosis and lipid metabolism have been identified in numerous laboratory animal species, including nonhuman primates (rhesus, squirrel, cynomolgus, and African green monkeys, baboon, and marmoset), dog, pig, rabbit, avian species (pigeons, quail, and chickens), rat and mouse (reviewed in Refs. 3–7). Each of these model systems offers particular advantages and limitations, such that no one model can be considered the perfect model for human atherosclerosis. However, these models allow the investigation of specific components in the process of atherogenesis, providing new insights that may eventually be applied in the diagnosis or treatment of atherosclerosis in humans. This chapter will focus on mouse models of atherosclerosis, as the extensive repertoire of tools available for genetic analysis in the mouse has made it possible to investigate aspects of lipid metabolism that cannot yet be approached in other species. Additional information on mouse models of atherosclerosis can be found in several excellent reviews that have appeared in recent years [6,8–12].

Mouse Genetic Models in Atherosclerosis

315

II. Atherosclerosis in the Mouse The first reports of the mouse as a genetic model for atherosclerosis appeared more than 30 years ago. In these initial studies, it was demonstrated that some strains develop early atheromatous lesions when fed diets containing very high concentrations of cholesterol (5%) and fat (30%) supplemented with cholic acid (2%) [13] or in combination with irradiation [14]. These early diets produced high mortality among mice, and were subsequently modified to reduce the concentrations of cholesterol (1.25%), fat (15%), and cholic acid (0.5%). The modified atherogenic diet greatly improved survival and made it possible to compare numerous mouse strains for their susceptibility to atherosclerosis. The standard assay for atherosclerosis, pioneered by Paigen, involves microscopic quantification of the cross-sectional area of fatty streak lesions in serial thin sections prepared from the aortic sinus and valve area. Using this assay, Paigen and others demonstrated that lesion formation is reproducible within a given inbred strain and that inbred strains differ in their susceptibility to diet-induced atherosclerosis [15,16]. Although the aortic sinus assay has been most useful for studies of dietinduced atherosclerosis in inbred strains, some genetically engineered mouse strains develop more extensive atherosclerosis that can be assayed en face by removing the entire descending aorta, dissecting it longitudinally, and staining for lipids to detect lesions [17]. Studies of diet-induced atherosclerosis in inbred mouse strains by several investigators have revealed genetic variation in many factors associated with atherosclerosis, including fatty streak lesion area, plasma lipoprotein levels and composition, lipid accumulation in coronary arteries, arterial wall calcification, and expression of inflammatory genes [15,18,19]. Susceptibility to aortic lesion development in some strains is inversely related to HDL levels, as seen in humans. For example, strains DBA, C57BL/10, and C57BL/6 all have relatively low HDL cholesterol levels and high lesion scores, whereas BALB/c and C3H mice typically maintain higher HDL cholesterol levels and are relatively resistant to aortic lesions (Fig. 1). However, other strains such as MRL and NZB have extremely high HDL cholesterol levels, but they still develop moderate lesions, indicating that other factors determine atherosclerosis susceptibility in these strains. Likewise, lipid accumulation in the coronary artery and aorta wall calcification show a large degree of variation among strains and fail to show a strict correlation with lesion scores (Fig. 1). Thus, it appears that a genetic approach will be among the most fruitful strategies to dissect out specific factors involved in the complex process of atherogenesis. In recent years, two general approaches have been employed to characterize specific genetic factors that give rise to differences in atherosclerosis-associated phenotypes in the mouse. One approach can be thought of as a ‘‘gene → phenotype’’ approach where specific genes are manipulated in the mouse germline to

316

Reue

Figure 1 Genetic variation among inbred mouse strains in atherosclerosis-associated phenotypes. Values presented are from female mice fed an atherogenic diet (15% fat, 1.25% cholesterol, 0.5% cholic acid) for 15 weeks. Aortic lesions were assayed as lipid staining areas in cross sections of the aortic sinus. Values for coronary artery lipid accumulation and arterial wall calcification are given as the percentage of mice exhibiting the trait. (Adapted from data presented in Ref. 19.)

investigate the role of individual proteins in lipid metabolism and atherosclerosis. A second approach can be considered a ‘‘phenotype → gene’’ approach where naturally occurring genetic variations are exploited to identify novel genes involved in lipid metabolism. In the following sections, we will illustrate these approaches and their application to the study of atherosclerosis in mouse models. III. Tools for Genetic Analysis in the Mouse The mouse has played a major role in genetic studies performed since the rediscovery of Mendel’s laws in 1900. The inbred mouse strains that are widely used in genetic research today were developed through selective breeding of ‘fancy’ mice chosen for genetic variation in coat color and other features [20]. Inbred mouse strains provide an endless source of ‘‘clones’’ owing to genetic homozygosity at every locus. As a consequence, studies performed with a particular inbred strain are cumulative. Literally hundreds of inbred strains are available, each representing a unique combination of alleles and source of independent genetic variation/mutation. The relative ease of selective breeding in mice has led to the development of additional tools to facilitate genetic analysis, including consomic, congenic, and recombinant inbred strains [20]. In addition, a dense genetic linkage map having more than 10,000 markers has been constructed to aid in mapping novel genes that underlie mutant phenotypes and complex phenotypes, such as atherosclerosis susceptibility. In addition to the tools for genetic analysis, techniques for planned modification of the mouse germline are now routinely practiced in numerous laboratories. Several clever experimental strategies have been developed to produce a range of genetic alterations. These include stable integration of an exogenous gene construct directing the expression of a gene of interest in a spatially and/

Mouse Genetic Models in Atherosclerosis

317

or temporally controlled manner, inactivation of an endogenous gene in a generalized or tissue-specific manner, replacement of a normal gene with a mutant version, inducible activation/suppression of transgene expression, and cross breeding transgenic and knockout models to produce new genetic combinations. Using this arsenal of powerful tools, specific genetic alterations can be generated and the physiological consequences studied in vivo. Detailed discussion of the methodology for these techniques is beyond the scope of this chapter, and can be found in several excellent reference works [21–23]. Resources for genetic analysis in the mouse, including features of inbred and spontaneous mutant strains, genetic and physical maps of the mouse genome, and induced mutant (i.e., transgenic and gene knockout) strains are readily accessible in electronic databases (Table 1). In the following sections, we will describe both naturally occurring genetic variants and genetically engineered mutant mouse models that have contributed to our understanding of atherosclerosis and lipid metabolism.

Table 1 Electronic Resources for Mouse Genetics Mouse Genome Database (MGD) Information on inbred strains, mutant strains, genetic and physical maps of the mouse genome, genetic markers and polymorphisms, mammalian homology. Transgenic/Targeted Mutation Database (TBASE) Catalog of thousands of induced mouse mutants including transgenic and knock-out models produced by the worldwide research community. Also includes a citation database, glossary of genetic terms, and experimental protocols. Induced Mutant Resource Database Mouse strains available commercially from the Jackson Laboratory having transgenes, targeted genes, or chemically induced mutations. The Whole Mouse Catalog Information on the use of rodents in biomedical research including genome informatics, anatomy and physiology, lab animal suppliers, transgene and targeted mutant animal facilities, and links to hundreds of related sites.

www.informatics.jax.org

tbase.jax.org

www.jax.org/resources/documents/imr

www.rodentia.com/wmc/

318

Reue IV. Induced Mouse Mutants in Atherosclerosis Research

Over the last 15 years, genes have been isolated for most of the major known proteins involved in lipoprotein metabolism. These include genes for the protein constituents of lipoprotein particles (apolipoproteins), genes for cell surface receptors that recognize native or modified lipoproteins, and enzymes that function in the metabolism of lipoprotein lipids. These genes are likely candidates as factors that will influence atherogenesis and as such have been the initial targets for modification in the mouse germline via transgenesis. Transgenic and gene knockout technology has been enthusiastically implemented in the field of atherosclerosis and lipid metabolism, resulting in an extensive repertoire of genetically modified models (Table 2). Some key examples of mouse models that have been generated to address questions in several aspects of lipid metabolism are described below. For the sake of the discussion, these examples have been categorized as follows: modulation of atherosclerosis susceptibility in transgene or gene Table 2

Genetically Engineered Mouse Models for Proteins in Lipid Metabolism Component

Apolipoproteins

Lipoprotein receptors

Lipoprotein remodeling enzymes

A-I A-II A-IV B CI CII CIII CIV E Apo(a) LDL receptor LDL receptor–related protein VLDL receptor ApoE receptor 2 Scavenger receptor Lipoprotein lipase Hepatic lipase Cholesterol ester transfer protein Lecithin cholesterol acyl transferase

Transgenic model

Knockout model

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹

⫹ ⫹ ⫹ ⫹a ⫹

⫹ ⫹ ⫹ ⫹ ⫹

⫹ —b ⫹ ⫹a ⫹ ⫹ ⫹ ⫹ ⫹ —b ⫹

A ⫹ in transgenic or knockout column indicates that an induced mutant mouse model has been generated for the corresponding protein. a The knockout model for this protein showed embryonic lethality. b Not applicable; no mouse homolog exists.

Mouse Genetic Models in Atherosclerosis

319

knockout models, ‘‘humanization’’ of the mouse lipoprotein profile by transgene expression, murine models of human lipid disorders, investigation of protein function in vivo, protein structure/function studies, and analysis of gene regulatory elements. Classification of a particular model into one of these groups does not imply that it is the only feature that can be addressed with the model, as several models could actually fit into more than a single category. Owing to the ever-increasing number of these models, this discussion is not comprehensive, but illustrates some of the most widely used models and some of the interesting experimental strategies that have been utilized.

A. Modulation of Atherosclerosis Susceptibility in Transgene or Gene Knockout Models Mouse Models with Increased Susceptibility to Atherosclerosis

One of the most important achievements in atherosclerosis research resulting from application of transgenic/knockout techniques is the generation of mouse strains that develop atherosclerosis spontaneously and more extensively than standard inbred strains. As described in an earlier section, some inbred mouse strains are susceptible to atherosclerosis, but the production of lesions requires feeding an unphysiological diet containing high concentrations of cholesterol and cholic acid. In many strains, this diet produces a lipoprotein profile similar to that in humans with substantial concentrations of LDL/very low density lipoprotein (VLDL) cholesterol, but can be toxic when fed to mice for long periods of time. Furthermore, the early fatty streak lesions that develop in the aortic root do not faithfully mimic the range of lesion severity or distribution seen in humans, although susceptible strains maintained on the atherogenic diet for prolonged periods (8–9 months) may develop more advanced fibrofatty lesions and lesions in additional locations [24,25]. In 1992, two independent groups demonstrated that inactivation of the mouse apoE gene leads to severe hypercholesterolemia and spontaneous atherosclerosis which more closely resembles human atheromas than the diet-induced models [26,27]. ApoE is a constituent of VLDL and HDL, and as a ligand for both the LDL receptor and chylomicron receptor, plays a central role in lipoprotein clearance from the plasma. ApoE-deficient mice develop elevated cholesterol levels on a chow diet of ⬃500 mg/dL compared to 60–80 mg/dL for wild-type mice. The increased cholesterol levels in these mice can be attributed to high levels of βVLDL, a cholesterol-rich VLDL particle that is thought to arise as a remnant after metabolism of intestinal derived chylomicrons. The cholesterol levels are elevated further to 2000 mg/dL when ApoE knockout mice are fed a ‘‘Western-type’’ diet, which has a similar composition to an average American diet with 21% fat and 0.15% cholesterol [27]. The elevated levels of intestinally

320

Reue

derived cholesterol-rich lipoproteins result from impaired clearance of these particles in ApoE-deficient animals. Associated with the high cholesterol levels, ApoE-deficient mice develop spontaneous aortic lesions that progress from fatty streak to intermediate and advanced lesions typical of human lesion pathology [28]. In chow-fed ApoE knockout mice, fatty streak lesions are apparent in the aorta by 12 weeks of age and develop into fibroproliferative lesions by 20 weeks. More advanced lesions having medial necrosis and occasional aneurysm also form. In analogy with human atherosclerosis, lesions are widespread, occurring not only at the base of the aorta but additionally in the proximal coronaries and all major branch points of vessels from the aorta [17,28]. Atherosclerosis is accelerated in ApoE-deficient mice by feeding the Western-type diet, and lesion size is increased fourfold. However, even in this extremely useful atherosclerosis-susceptible mouse model, it is important to note that a significant difference in lesion pathology exists when compared to humans. In humans, plaque rupture leading to heart attack is a fairly common outcome of lesion development, but plaque rupture is not observed in the mouse. Since factors contributing to plaque rupture are not fully understood, it is not clear why this species difference exists, but it has been suggested that the lack of rupture is related to the smaller diameter and greater surface tension of the mouse aorta which might prevent rupture [6]. Other genetic manipulations that increase atherosclerosis susceptibility include disruption of the LDL receptor gene and overexpression of a transgene encoding human apolipoprotein B (ApoB). The LDL receptor binds to apolipoproteins B and E on the surface of lipoprotein particles and mediates their clearance from the plasma. Mice carrying an LDL receptor gene knockout exhibit a twofold increase in cholesterol levels resulting from increased LDL and IDL concentrations [29]. When fed an atherogenic diet containing 1.25% cholesterol, 7.5% fat, and 0.5% cholic acid, LDL receptor knockout mice develop atherosclerotic lesions in the aorta and coronary arteries, as well as subcutaneous xanthomas, a typical manifestation of LDL receptor deficiency in human patients. Transgenic mice have also been generated which express high levels of human apoB, the primary protein component of LDL [30]. These mice synthesize transgene-encoded ApoB in the liver which is processed into both forms of ApoB protein—full-length ApoB-100 and the truncated ApoB-48 form that arises through editing of the human ApoB mRNA by the endogenous ApoB editing enzyme (ApoB mRNA editing catalytic polypeptide-1, Apobec-1). The ApoB transgenic mice have modestly elevated LDL cholesterol levels on a chow diet, which increase severalfold when fed the cholesterol/cholic acid atherogenic diet. These animals do not develop atherosclerosis on the chow diet, but they exhibit 11-fold greater aortic lesion scores than nontransgenic controls when fed the atherogenic diet.

Mouse Genetic Models in Atherosclerosis

321

Although both the LDL receptor knockout and ApoB transgenic models described above exhibit increased atherosclerosis, these mice develop lesions only when fed a diet containing cholesterol and cholic acid. Furthermore, unlike human patients lacking functional LDL receptor, the LDL receptor–deficient mice exhibit mild rather than marked hypercholesterolemia. One possible explanation for the lower LDL levels in mice compared to humans lacking LDL receptor function is that the human liver produces only the full-length ApoB isoform (ApoB-100), but the mouse liver also produces the truncated ApoB-48 form; ApoB-48, but not ApoB-100, is present on ApoE-containing lipoproteins that can be cleared efficiently by receptors other than the LDL receptor. To test this possibility, LDL receptor knockout mice were bred to mice carrying a knockout mutation in the Apobec-1 gene encoding the ApoB mRNA editing enzyme [31]. The resulting double knockout mice have markedly increased ApoB-100 and LDL cholesterol levels, similar to human familial hypercholesterolemia patients. Additionally, these mice spontaneously develop extensive atherosclerosis on a chow diet with lesions being detected in the aorta, intercostal arteries and at points of vessel branching, and occasionally in the renal and iliac arteries. As with the ApoE knockout model, cholesterol levels and atherosclerosis were exacerbated by feeding a Western-type diet. Both the ApoE and LDL receptor/Apobec-1 knockout mouse strains provide useful models of spontaneous atherosclerosis, but some important differences exist in the etiology of the hypercholesterolemia, and possibly atherogenesis, in the two. As described earlier, the elevated cholesterol in the plasma of ApoE knockout mice occurs in the VLDL fraction, which contains almost exclusively the ApoB-48 isoform of ApoB. This lipid profile is very different from that typically found in hyperlipidemic humans. In contrast, the LDL receptor/ Apobec-1 knockouts exhibit elevated LDL cholesterol levels, and corresponding elevated ApoB-100 concentrations, which resembles that of human hyperlipidemia. Another difference between the ApoE knockout mice and humans is the lack of ApoE secretion in the macrophage foam cells present in atherosclerotic lesions, a condition which may independently promote atherogenesis in the ApoE knockout mice. Evidence for the latter point comes from studies demonstrating inhibition of diet-induced atherosclerosis in transgenic mice expressing ApoE in the arterial wall [32], and increased atherosclerosis in wild-type mice which have been irradiated and reconstituted with bone marrow macrophages derived from ApoE knockout mice [33]. The atherosclerosis-susceptible mouse models described above were produced by disruption of gene function. There are additional models in which overexpression of a particular protein leads to increased atherosclerosis. One of these is a transgenic model that overexpresses mouse ApoA-II. ApoA-II is the second most abundant protein constituent of HDL, and the transgenic mice exhibit ele-

322

Reue

vated HDL cholesterol levels, and they unexpectedly developed spontaneous aortic lesions on a chow diet [34]. Interestingly, the increased atherogenesis is significant only in male mice, which have aortic lesion scores 18-fold higher than the nontransgenic controls. A likely explanation for the increased atherosclerosis in the presence of elevated HDL levels is that the transgenic HDL particles have an unusual composition with an increased ratio of ApoA-II to ApoA-I and a larger particle size. Thus, these studies demonstrated that not only concentration, but also composition, of HDL appear to be important determinants of atherosclerosis. Mouse Models with Increased Resistance to Atherosclerosis

The development of mouse models for atherosclerosis susceptibility has made it possible to investigate the potential antiatherogenic effect of various components of the lipid transport system and vascular wall. One notable example is ApoAI, the major protein constituent of HDL. In human epidemiological studies, levels of ApoA-I and HDL cholesterol have repeatedly been associated with a reduced risk for atherosclerosis, which some attribute to a role for this lipoprotein in the removal of cholesterol from peripheral tissues, including cells of the artery wall. To test whether elevated ApoA-I levels can directly lead to a reduction in atherosclerosis, transgenic mice expressing high levels of human ApoA-I were generated on the C57BL/6 strain, which is susceptible to diet-induced atherosclerosis [35]. The human ApoA-I transgenic mice fed a diet containing cholesterol and dairy butter fat were completely protected from aortic lesions compared to C57BL/6 control mice; when fed a diet containing cholesterol and cocoa butter fat, the ApoA-I transgenics developed sevenfold fewer lesions than controls. Interestingly, despite the robust protective effect of ApoA-I, mice carrying an ApoA-I gene knockout do not exhibit an increased susceptibility to atherosclerosis even when fed the cholesterol/cholic acid atherogenic diet [36]. These results suggest that a deficiency in ApoA-I and low HDL levels alone are not sufficient to predispose to atherosclerosis. Subsequent to the studies of ApoA-I transgenics on the C57BL/6 susceptible background, the ApoA-I transgene was bred onto the ApoE-deficient model of spontaneous atherosclerosis [37,38]. When maintained on a low-fat, low-cholesterol diet, animals expressing the ApoA-I transgene had elevated HDL levels and a significant reduction in lesions with nearly absent lesions at 4 months of age and only small fatty streak lesions at 8 months with no evidence of fibrous plaques as seen in the ApoE-deficient controls. These results illustrate that ApoAI and HDL have a protective effect against lesion development even under conditions in which non-HDL lipoproteins are elevated. These studies also lend support to the proposed activities of HDL in reverse cholesterol transport and/or protection of the arterial wall against atherogenic stimuli.

Mouse Genetic Models in Atherosclerosis

323

Additional ApoE-deficient models with reduced susceptibility to atherosclerosis have been produced by modulation of gene expression in macrophages present in atherosclerotic lesions. It has been unclear whether macrophages have an antiatherogenic or proatherogenic role in the artery wall, as macrophages entering the subendothelial space from the circulation are thought to scavenge and eliminate atherogenic lipoprotein species, but they also are among the cell types that accumulate cholesterol and contribute to lesion progression upon their lysis. At least three mouse models have been informative concerning this issue. In one case, mice carrying a spontaneous mutation in the Op gene encoding the macrophage colony-stimulating factor (MCSF) were cross bred with ApoE knockout mice [39]. The MCSF-deficient/ApoE-deficient mice exhibit a two- to threefold elevation in plasma cholesterol levels compared to ApoE knockout mice, yet they are less susceptible to lesion formation with lesion scores fourto sixfold lower. These results are noteworthy, because they suggest that the macrophage may have a proatherogenic role, since in their absence, the effect of hypercholesterolemia on atherogenesis in ApoE-deficient mice was dramatically blunted. Lesions were also significantly reduced in ApoE knockout mice after bone marrow reconstitution with macrophages from wild-type mice [40]. In contrast to the MCSF-deficient ApoE-deficient mice, however, the bone marrow recipients had plasma ApoE levels of ⬃10% normal due to synthesis and secretion by the transplanted macrophages. The macrophage-derived ApoE was apparently adequate to associate with plasma lipoproteins and promote receptor-mediated clearance. The antiatherogenic effect of ApoE expression in macrophages was confirmed in transgenic mice expressing elevated levels of human ApoE in macrophages in the arterial wall [32]. B. Humanization of the Mouse Lipid Profile by Transgene Expression

One of the primary differences in lipid physiology between mice and humans is the distribution of circulating lipoproteins in the two species. Whereas humans carry ⬃75% of their cholesterol in LDL, C57BL/6 mice fed a standard chow diet carry nearly all circulating cholesterol in HDL [16]. Feeding mice diets containing 7.5–30.0% fat increases the non-HDL fraction to resemble more closely the human levels, but unlike humans, the mouse accumulates cholesterol-rich VLDL particles to a higher level than LDL [41]. Thus, one strategy to produce a mouse model that more closely resembles the human lipoprotein profile has been to generate mice that express higher levels of LDL. This has been achieved by producing mice that carry transgenes for human ApoB-100 [42,43]. Owing to the large size of the ApoB gene, transgenic mice were produced using phagemid P1 DNA clones to express human ApoB at levels similar to those of normolipidemic humans. When maintained on a chow diet, these animals showed up to a

324

Reue

fourfold increase in LDL cholesterol levels compared to nontransgenics with an overall lipoprotein profile that resembles that of humans. In addition to the native lipoprotein profile, mice differ from humans in lacking two proteins involved in lipid metabolism, Apo(a) and cholesterol ester transfer protein (CETP). Apo(a) is a large protein that closely resembles plasminogen, containing single or multiple copies of several plasminogen-like protein domains. Apo(a) is present in human plasma on LDL particles attached to ApoB through a disulfide bond to produce a lipoprotein complex known as Lp(a). In humans, elevated levels and polymorphic isoforms of Apo(a) and Lp(a) are correlated with increased risk for atherosclerosis and associated conditions [44]. To allow the further examination of the role of Apo(a) in lipid metabolism and atherosclerosis, transgenic mice expressing the human apoa gene were produced [45]. These mice showed plasma Apo(a) levels similar to the median levels in humans but did not form Lp(a) particles, indicating that human Apo(a) cannot complex with mouse ApoB. Nevertheless, the Apo(a) transgenic mice fed an atherogenic diet exhibited a twofold increase in incidence of fatty streak lesions, and Apo(a) was detected within the aortic lesions. These results led to the novel suggestion that Apo(a) itself (as opposed to Lp(a)) may have a direct proatherogenic effect. As an extension of this work, mice carrying the human ApoB transgene were crossbred to Apo(a) transgenic mice to investigate whether Lp(a) particles would form [42,43]. The double transgenic mice did indeed form high levels of Lp(a) demonstrating efficient assembly of this particle from the human transgenic proteins. Like Apo(a), CETP has been implicated in human atherosclerosis but is not expressed in the mouse. CETP is a plasma protein that facilitates the exchange of triglycerides in VLDL for cholesterol esters in HDL [46]. CETP is thought to promote net transport of cholesterol from plasma to the liver, as the cholesterol esters in HDL are exchanged to VLDL and subsequently cleared in the liver. In agreement with this role for CETP, human CETP activity is inversely correlated with HDL cholesterol levels. Transgenic mouse strains have been generated using both human and cynomolgus monkey CETP genes. Mice expressing high levels of monkey CETP on a C57BL/6 background had reduced HDL cholesterol levels and size and ApoB containing lipoproteins of increased size [47]. Additionally, the monkey CETP transgenic mice fed the cholesterol/cholic acid diet had increased aortic lesion scores. The human CETP transgenic mice also showed reduced HDL cholesterol levels and smaller HDL size, but these effects were quite modest, and only about 20% of CETP was associated with HDL compared to 100% in humans [48]. It was subsequently determined that introduction of a human ApoA-I transgene into the human CETP transgenic mice results in 100% association of CETP with HDL and more pronounced reductions in HDL levels and size [49].

Mouse Genetic Models in Atherosclerosis

325

C. Murine Models of Human Lipoprotein Disorders

The impetus for some efforts toward humanization of lipid metabolism in the mouse is the possibility of producing mouse models that mimic human lipid disorders to study disease processes better and develop therapeutic interventions. As discussed earlier, humans and mice differ in several factors that influence lipid metabolism, so few models that exactly replicate human disorders have been produced. Nevertheless, even models with apparent species differences may be useful to address specific questions and may provide novel insights into some aspects of lipid metabolism disorders. Some of the genetically altered mouse models that approximate human lipid disorders are briefly described below. Familial hypercholesterolemia is characterized by hypercholesterolemia and premature coronary artery disease as a result of mutations in the LDL receptor gene. The production of LDL receptor–deficient mice via gene knockout technology was expected to provide a comparable phenotype in the mouse. However, it was determined that LDL receptor knockout mice had a much milder increase in cholesterol levels than their human counterparts [29], apparently due to the lack of high levels of ApoB-100 in mice. This difference was subsequently overcome by incorporating a second knockout mutation in the Apobec-1 gene such that the LDL receptor/Apobec-1 double knockout appears to be a useful model of familial hypercholesterolemia [31]. Lipoprotein lipase deficiency (type I hyperlipoproteinemia) has been widely studied in humans [50]. These patients have severely elevated chylomicron and VLDL levels resulting from lack of lipoprotein lipase (LPL) activity, which normally hydrolyzes chylomicron and VLDL triglycerides. LPL hydrolysis of lipoprotein triglycerides also promotes the exchange of lipids and surface proteins among lipoprotein classes, such that LPL-deficient patients also have reduced LDL and HDL levels. LPL-deficient mice were generated by gene knockout technology [51,52] and were shown to have threefold increases in triglyceride levels, sevenfold elevations in VLDL levels and reduced HDL levels at birth. Unlike the human condition, however, LPL deficiency is lethal in the mouse within 18–48 hr of birth due to cyanosis. Capillaries of the knockout mice become engorged with chylomicrons that presumably prevent red cell contact with the endothelium for oxygen exchange. Potential reasons for the much greater severity of the LPL-deficient phenotype in mouse compared to humans have been suggested, including the difference in type of LPL mutation (i.e., a null allele in the mouse vs point mutations that may allow small amounts of residual LPL activity in human patients), the greater fat load in suckling mice compared to newborn humans, and the action of additional triglyceride lipases in humans that may not be acting in the mouse [51].

326

Reue

A potential mouse model for hypobetalipoproteinemia has been described in which, as in the human condition, plasma ApoB levels are diminished by ⬃70%. This phenotype was produced in mice heterozygous for a knockout in the ApoB allele [53] and in an independent model consisting of mice homozygous for a mutation in the apoB gene that results in synthesis of a truncated ApoB protein [54]. In the heterozygous knockout mice, the animals were also found to have reduced susceptibility to diet-induced hypercholesterolemia due to the lack of a diet-induced increase in ApoB-containing lipoproteins that normally occurs in mice eating a high-fat diet [53]. Mice carrying the truncated apoB allele also exhibited features not present in human hypobetalipoproteinemic patients, most notably nervous system developmental abnormalities [54]. Type III hyperlipoproteinemia is a disorder resulting in increased levels of cholesterol and triglycerides in the VLDL and IDL fractions. Individuals with mutant forms of apoE that do not bind cellular receptors and are not cleared normally are predisposed to this condition. The expression of two mutant apoE alleles in the mouse results in a phenotype similar to type III patients with elevated VLDL/LDL cholesterol levels [55,56]. Mice expressing the ApoE3-Leiden variant and fed the cholesterol/cholic acid diet also developed atherosclerotic lesions in several locations, many of which progressed to fibrous lesions [55]. In mice as in humans, the ApoE3-Leiden isoform acts as a dominant trait leading to the expression of hyperlipoproteinemia and may represent a useful model in which to study factors influencing the metabolism of remnant VLDL particles. Transgenic mice overexpressing two of the four known C apolipoproteins appear to be models of primary hypertriglyceridemia in humans. Mouse strains carrying the human gene for either ApoC-II or ApoC-III develop hypertriglyceridemia with an accumulation of triglyceride-rich VLDL particles having an increased ratio of the ApoC protein to ApoE [57,58]. Metabolic studies revealed that VLDL accumulation results from delayed clearance, perhaps due to less access to cell surface lipases or receptors [58,59]. These studies suggest that ApoC levels may influence triglyceride levels in humans, in which hypertriglyceridemia is common, affecting up to one third of males. Finally, a mouse model with features of familial combined hyperlipidemia (FCHL) has been generated by cross breeding three induced mutant strains [60]. FCHL is considered to be the most common genetic dyslipidemia predisposing to premature atherosclerosis, affecting 1–2% of the general population and 15– 20% of patients with premature coronary artery disease [61]. FCHL patients exhibit a variable hyperlipidemia characterized by elevated cholesterol levels, elevated triglyceride levels, or both. Typically, FCHL patients exhibit increased ApoB and VLDL levels and decreased HDL levels. To generate a mouse model that exhibits a similar lipoprotein profile to that seen in FCHL, previously characterized strains were interbred. The ApoC-III transgene was bred onto an LDL receptor knockout background, resulting in mice with elevations in plasma ApoB

Mouse Genetic Models in Atherosclerosis

327

levels, VLDL and IDL-LDL cholesterol levels, and VLDL triglyceride levels. However, these mice did not display the reduced HDL levels typical in FCHL, and a human CETP transgene was subsequently introduced. The resulting LDL receptor–deficient/apoC-III transgenic/CETP transgenic mice exhibited a reduction in HDL levels and an overall lipoprotein profile resembling that of FCHL. These studies further implicate ApoC-III in hypertriglyceridemias, perhaps including FCHL. D. Investigation of Protein Function In Vivo

The numerous mouse models that have been generated in the field of lipid metabolism and atherosclerosis have involved manipulation of genes encoding proteins with unknown or only partially understood physiological function. Among the models already discussed here, most revealed novel or unexpected implications regarding the function of the protein under investigation. For example, it was well established years ago through biochemical means that ApoC-II functions as a cofactor for LPL. Furthermore, ApoC-II deficiency in humans was known to produce hypertriglyceridemia identical to LPL deficiency [50]. However, it was not until this protein was overexpressed in a transgenic mouse that investigators became aware that elevated ApoC-II levels can also result in hypertriglyceridemia, suggesting that this protein may have a role in controlling triglyceride levels [58]. Likewise, neither the occurrence of spontaneous atherosclerosis in ApoA-II transgenic mice [34] nor protection against diet-induced atherosclerosis in ApoA-IV transgenic mice [62,63] could be foreseen based on the previously known roles of these proteins. Other examples of unexpected phenotypes include the finding in ApoA-I knockout mice that ApoA-I plays an essential role in adrenal gland physiology, being required for cholesterol ester accumulation in steroidogenic cells and for normal adrenal steroid production [64]. The relationship between LDL receptor deficiency and protection against lethal endotoxemia [65] was also unexpected. Among the most clever use of mouse models to dissect the biological function of a particular protein is the investigation of the relative contributions of two lipoprotein receptors in the clearance of ApoE-rich lipoproteins. As described earlier, the LDL receptor plays a central role in cholesterol homeostasis through binding and clearance of ApoB- and ApoE-containing lipoproteins. Studies of lipoprotein metabolism in LDL receptor knockout mice and ApoE knockout mice suggested the existence of a second receptor active in the clearance of ApoEcontaining lipoproteins [66]. A candidate for the second receptor was the structurally related protein known as the LDL receptor–related protein (LRP). To address this issue, LRP knockout mice were generated, but unfortunately were not viable [67]. In a tour de force application of genetic modulation of gene expression in the mouse, the investigators overcame this obstacle by overexpressing the receptor-

328

Reue

associated protein (RAP) known to inhibit LRP activity. RAP expression was achieved in both wild-type and in LDL receptor knockout mice by adenoviral delivery of the RAP gene to liver for acute, high-level expression. A comparison of the resulting lipoprotein metabolism in the two mouse strains revealed that wild-type mice expressing RAP had only minor changes in remnant lipoprotein uptake, whereas LDL receptor–deficient mice had greatly diminished remnant clearance [68]. These results elegantly demonstrated that remnants are cleared by both the LDL and LRP receptor pathways. Another variation of the standard overexpression/knockout approach to investigate physiological function is to express transgenes in a tissue-restricted fashion on either wild-type or knockout backgrounds. As described in an earlier section, LPL knockout mice exhibit neonatal hypertriglyceridemia and lethality and are also characterized by diminished adipose tissue mass [51]. The primary tissue sites for LPL expression are adipose and skeletal muscle. To evaluate the role of LPL specifically in skeletal muscle, LPL knockout mice were crossbred to transgenics expressing human LPL driven by a muscle-specific promoter. The resulting animals expressed LPL exclusively in muscle, which was adequate to rescue these mice from death, restore normal adipose tissue mass, and effectively normalize triglyceride levels [51,69]. In addition to LPL, the enzyme hormone–sensitive lipase (HSL) is an important lipase in adipose tissue. HSL is the rate-limiting intracellular enzyme for lipolysis of endogenous triglycerides and also hydrolyzes cholesterol esters [70]. Several studies have implicated HSL in cholesterol ester turnover in macrophages [71,72], a process that could be relevant in macrophage-derived foam cells in atherosclerotic lesions. To investigate a possible beneficial effect of increased HSL expression in macrophages, transgenic mice expressing an HSL transgene driven by a macrophage-specific promoter/enhancer were produced on a C57BL/ 6 background [73]. After eating the cholesterol/cholic acid diet for several weeks, these mice exhibited sevenfold increased HSL activity in macrophages, but paradoxically exhibited greater susceptibility to aortic lesion development. Although the mechanism of increased atherogenesis in this model is not understood, these results point to a role for some other lipase besides HSL in cholesterol ester hydrolysis in cholesterol-loaded macrophages. E. Protein Structure/Function Studies

Genetically modified mice have also allowed investigations of protein structure/ function relationships in vivo. One example is the use of a modified human apoB transgene to identify the precise cysteine residue that forms a disulfide linkage with Apo(a) to produce the Lp(a) complex. In a clever exploitation of yeast to achieve high-efficiency homologous recombination to produce a desired mutation, a yeast artificial chromosome containing the human apoB gene was recom-

Mouse Genetic Models in Atherosclerosis

329

bined in yeast to produce a site-directed mutation changing the candidate cysteine residue to glycine [74]. The resulting mutant human apoB gene was used to construct a transgenic mouse strain. It was determined that the mutant apoB could not bind Apo(a), demonstrating that the binding occurs through the specific cysteine residue targeted. Other examples of structure/function studies in vivo include the studies of ApoE isoforms on hyperlipidemia as described in an earlier section [55,56]. F. Mapping Gene Regulatory Elements In Vivo

In addition to their use in studying structural genes for lipid metabolism proteins, transgenic mouse models have been utilized to delineate regulatory elements that govern tissue-specific expression of these genes. Of particular interest are the regulatory regions that control two clusters of apolipoprotein genes, the ApoAI/C-III/A-IV cluster and the ApoE/C-I/C-IV/C-II cluster. The genes for ApoAI and ApoC-III are located adjacent to each other in convergent transcriptional orientation, and are expressed in liver and intestine. The apoA-IV gene is located downstream of apoC-III in divergent transcriptional orientation, and is expressed primarily in intestine. Numerous lines of transgenic mice have been generated using different portions of the human apoA-I/C-III/A-IV gene cluster, and the expression pattern of the genes in mice has allowed mapping of pertinent transcriptional elements [75]. Liver expression of the apoA-I and apoC-III genes occurred when very small segments (⬍300 bp) of 5′ flanking DNA were included in transgene constructs, but intestinal expression was initially elusive. Eventually, constructs containing large portions of gene-flanking sequences revealed that intestinal expression of all three genes is controlled by a common element located in the region between the apoC-III and apoA-IV genes, more than 6 kb beyond the 3′ end of the apoA-I gene [75]. In the 50-kb apoE/C-I/C-IV/C-II gene cluster, all genes are transcribed in the same orientation. apoE Is expressed at very high levels in liver, and at lower levels in a broad range of tissues. Initial transgenic mice generated using up to 30 kb of sequence from the 5′ flanking region of the apoE gene failed to produce expression in liver, but gave high expression in kidney. Subsequent work with numerous transgenic mouse lines localized an element 11-kb 3′ of the apoE gene and several kilobases downstream of the adjacent apoC-I gene that drives liver expression and also silences kidney expression [76,77]. Further studies precisely narrowed the hepatic control region to ⬃300 bp [78], and this control element has been a useful tool to direct expression of heterologous mouse transgenes in a liver-specific manner. Transgenic mice have also been employed to identify regulatory elements that mediate changes in transcription levels in response to factors such as diet. One example is the detection of a cholesterol response element that activates

330

Reue

transcription of the CETP gene in response to dietary cholesterol [79]. It is established that both CETP mRNA levels and plasma protein levels increase in humans in response to a high-fat/high-cholesterol diet. Studies in transgenic mice carrying the human CETP coding sequences and several kilobases of native 5′ and 3′ flanking sequences revealed the presence of a cis-acting element that induces transgene transcription fivefold in response to cholesterol. The identification and analysis of such gene regulatory elements may provide new targets for modulation of gene expression levels.

V.

Identification of Novel Genes in Atherosclerosis and Lipid Metabolism

This chapter has focused thus far on the use of directed genetic alterations in known genes to produce an altered phenotype. Beyond testing the effects of known genes, the mouse model affords an opportunity to identify novel genes and phenotypes that is unequaled in any other animal model. As described near the beginning of the chapter, hundreds of inbred mouse strains are available, each carrying a different combination of gene alleles and exhibiting a different phenotypic array (see, e.g., Fig. 1). Additionally, many mutant mouse strains carrying spontaneous, single-gene mutations that affect lipid metabolism have been described (reviewed in Ref. 80). With the current tools for high-resolution genetic mapping in experimental mouse crosses, as well as recent physical maps of the mouse genome [81,82], it is feasible to identify specific genes underlying both single-gene mutations and complex traits [83,84]. Three genetic loci that are actively being pursued and likely to reveal novel genes with significant roles in lipid metabolism are Hyplip1 (hyperlipidemia), cld (combined lipase deficiency), and fld (fatty liver dystrophy). The Hyplip1 phenotype was discovered in a survey of recombinant congenic strains derived from two parental inbred strains, C57BL/10 and C3H [85]. One of the strains had plasma triglyceride levels that were dramatically elevated 10- to 30-fold and cholesterol levels that were 30–60% higher than the parental strains. The hyperlipidemia in these mice is exacerbated as the animals age, and metabolic studies revealed that the elevated triglycerides are associated with increased ApoB levels and VLDL secretion rates. These features are reminiscent of FCHL, the most common genetic dyslipidemia underlying premature coronary artery disease in humans. Experimental crosses performed between the Hyplip1 strain and the two parental strains revealed a pattern consistent with segregation of a single-gene and codominant expression of the hyperlipidemia trait. Using offspring of a large genetic cross and typing them for genetic markers spanning the genome, the Hyplip1 gene was localized to mouse chromosome 3. Although no genes for human FCHL have yet been identified, it is significant that the mouse Hyplip1

Mouse Genetic Models in Atherosclerosis

331

locus on chromosome 3 coincides with an FCHL locus mapped in a Finnish genetic isolate by independent investigators at nearly the same time [86]. The mouse model is now being used to investigate further the mechanism for hyperlipidemia and to map the gene to higher resolution in preparation for its isolation via a positional cloning approach. The ability to produce large genetic crosses in the mouse for this purpose may accelerate the identification of the mouse Hyplip1 gene, which can then be used to isolate the corresponding human gene. The cld and fld mutations are spontaneous, recessive mutations that confer abnormalities in triglyceride metabolism. These mutations were identified several years ago, but their isolation using molecular genetic approaches has only recently become practicable. Both mutations produce phenotypes that have similarities and differences with known human disorders. The cld mutation results in a combined deficiency of the two key enzymes required for circulating lipid metabolism, LPL and hepatic lipase (HL) (see Ref. 80 and references therein). The mutant animals develop a severe hypertriglyceridemia leading to cyanosis and neonatal death, analogous to the LPL knockout mice discussed earlier [51]. Biochemical studies have demonstrated that the cld mutation does not result from impaired rates of lipase synthesis but from production of enzymatically inactive protein that is not secreted. Unlike the LPL knockout mouse and human LPL deficiency, the cld mutation does not reside within the LPL, HL, or ApoC-II structural genes but rather in a trans-acting locus that maps to chromosome 17 within the t complex. The t complex is a unique region comprising about a third of mouse chromosome 17 that is characterized by a series of inversions. The presence of inversions has important practical consequences, since the inversions exclude the t region from normal pairing and recombination with noninverted, wild-type DNA, and thus make genetic mapping of the mutation difficult to achieve by construction of standard mouse crosses. However, several deletion mutations have been discovered and mapped in this region, and it has been possible to localize the cld mutation with respect to these deletions to a region of a manageable size for positional cloning. At present, the putative cld locus has been isolated on a physical contig composed of large-insert clones, and efforts are underway to identify the gene (M. Doolittle, personal communication). The fld mutation gives rise to a pleiotropic phenotype affecting lipid metabolism in several tissues (reviewed in Ref. 80). At birth the mice appear normal, but as they begin to suckle, they develop an enlarged, pale fatty liver and hypertriglyceridemia associated with reduced plasma LPL activity. Accompanying the lipid accumulation in the fatty liver are pronounced alterations in protein expression patterns, including greater than 10-fold elevations in apolipoproteins A-IV and C-II and a 6-fold reduction in hepatic lipase [87]. Between 2 and 3 weeks of age, the fatty liver, hypertriglyceridemia, and associated changes in hepatic gene expression spontaneously resolve, but additional manifestations are apparent. The fld mice develop a neuropathy characterized by a whole body tremor,

332

Reue

which has been attributed to abnormal myelin formation in peripheral nerves. Additionally, we have determined that adult fld mice exhibit nearly one third lower body mass than their wild-type littermates, part of which can be attributed to vastly diminished fat pad mass. Morphologically, adipocytes in the fld mice appear immature with reduced lipid accumulation and fail to respond to insulin normally leading to insulin resistance in these animals [88,89]. These features bear resemblance to a group of rare human disorders known as lipodystrophies, which are characterized by reduced adipose tissue stores, insulin resistance, fatty liver, and hypertriglyceridemia [90]. We have recently determined that fld mice also have increased susceptibility to diet-induced fatty lesion formation [89], a feature that has not been investigated systematically in human lipodystrophy patients. Using a positional cloning approach, we have recently isolated the fld gene and found that it encodes a novel nuclear protein [91,92]. Studies are in progress to define further the physiological function of the fld protein and to determine whether polymorphisms in the homologous human gene are associated with lipid metabolism disorders, including lipodystrophy, coronary artery disease, and insulin resistance. The examples above represent only the tip of the iceberg of naturally occurring genetic mutations/variations in the mouse that may reveal novel genes important in atherosclerosis and lipid metabolism. Numerous other single-gene mutations affecting various aspects of lipid metabolism (cellular cholesterol homeostasis, fatty acid metabolism, serum lipoprotein levels, and tissue lipid composition) have been described and await identification [80]. But perhaps even more pertinent to human disease are genetic variations that contribute to complex traits such as hypercholesterolemia and atherosclerosis susceptibility. Mice provide a unique resource in which to identify genetic variation in these traits without the confounding problems of environmental factors or genetic heterogeneity and to dissect out the underlying genes by methods that are similar to those which have been successful for single-gene mutations [83,84]. An example of a naturally occurring genetic variation among mouse strains in the response of their lipoprotein cholesterol levels to consuming an atherogenic diet is shown in Figure 2. Three strains of mice were analyzed: C57BL/6J (the original ‘‘atherosclerosis-susceptible’’ inbred mouse strain), C57BL/6ByJ (closely related genetically to C57BL/6J), and CAST/EiJ (a strain derived from Mus castaneus, a species distinct from the common laboratory strains including C57BL/6). Plasma lipoproteins from mice of the three strains fed the atherogenic diet were fractionated by size to distinguish VLDL (largest), LDL/IDL (intermediate), and HDL (smallest) particles. As discussed in an earlier section, chowfed mice of all strains have total cholesterol levels of 60–85 mg/dL, most of which is distributed in HDL. When fed the atherogenic diet, there is typically an increase in total cholesterol levels to 200–300 mg/dL, and a shift in lipoprotein profile with the majority of cholesterol being carried in VLDL and LDL, and a

Mouse Genetic Models in Atherosclerosis

333

Figure 2 Cholesterol distribution in plasma lipoproteins from three mouse strains fed the atherogenic diet. C57BL/6J, C57BL/6ByJ, and CAST/EiJ mice were fed the atherogenic diet for 2 weeks and plasma lipoproteins fractionated by fast performance liquid chromatography as described [79]. Cholesterol measurements were performed in the resulting fractions. The positions at which VLDL, IDL/LDL and HDL elute are indicated.

reduction in HDL cholesterol. This response is exemplified by C57BL/6J (see Fig. 2). In contrast, C57BL/6ByJ mice exhibit a blunted response to the diet, with total cholesterol levels remaining under 100 mg/dL and with corresponding lower levels of VLDL/LDL cholesterol. On the other side of the coin, the CAST/ EiJ mice increase total cholesterol levels to 500 mg/dL with 95% of the cholesterol being in VLDL/LDL fractions. The reduced dietary response in C57BL/ 6ByJ mice is associated with a 10-fold reduction in aortic lesions compared to C57BL/6J , but surprisingly the extreme hypercholesterolemia in CAST/EiJ mice is not associated with susceptibility to atherosclerosis (M. Mehrabian, personal communication). We have recently mapped a locus in C57BL/6ByJ mice that cosegregates with the reduced dietary response, and efforts will now be directed at isolating the underlying gene [93]. Genes underlying variations in many aspects of lipid metabolism in the mouse will likely provide insight into the role of specific factors in normal and disease states, and may eventually identify targets for therapeutic intervention in humans.

334

Reue References

1. Tolleshaug H, Hobgood KK, Brown MS, Goldstein JL. The LDL receptor locus in familial hypercholesterolemia: multiple mutations disrupt transport and processing of a membrane receptor. Cell 1983; 32:941–951. 2. Ross R. The pathogenesis of atherosclerosis: a perspective for the 1990s. Nature 1993; 362:801–809. 3. Clarkson TB. Nonhuman primate models of atherosclerosis. Lab Anim Sci 1998; 48:569–572. 4. Brousseau ME, JMH. Transgenic rabbits as models for atherosclerosis research. J Lipid Res 1999; 40:365–375. 5. St. Clair RW. The contribution of avian models to our understanding of atherosclerosis and their promise for the future. Lab Anim Sci 1998; 48:565–568. 6. Smith JD. Mouse models of atherosclerosis. Lab Anim Sci 1998; 48:573–579. 7. Reue K, Warden C, Lusis AJ. Animal models for genetic lipoprotein disorders. Curr Opin Lipidol 1990; 1:143–150. 8. Paigen B, Plump AS, Rubin EM. The mouse as a model for human cardiovascular disease and hyperlipidemia. Curr Opin Lipidol 1994; 2:258–264. 9. Rubin EM, Smith DJ. Atherosclerosis in mice: getting to the heart of a polygenic disorder. Trends Genet 1994; 10:199–203. 10. Shih DM, Welch C, Lusis AJ. New insights into atherosclerosis from studies with mouse models. Mol Med Today 1995; 1:364–372. 11. Paigen B. Genetics of responsiveness to high-fat and high-cholesterol diets in the mouse. Am J Clin Nutr 1995; 62:458S–462S. 12. Breslow JL. Mouse models of atherosclerosis. Science 1996; 272:685–688. 13. Thompson JS. Atheromata in an inbred strain of mice. J Atheroscler Res 1969; 10: 113–122. 14. Vesselinovitch D, Wissler RW. Experimental production of atherosclerosis in mice. 2. Effects of atherogenic and high-fat diets on vascular changes in chronically and acutely irradiated mice. J Atheroscler Res 1968; 8:497–523. 15. Paigen B, Morrow A, Brandon C, Mitchell D, Holmes P. Variation in susceptibility to atherosclerosis among inbred strains of mice. Atherosclerosis 1985; 57:65–73. 16. Paigen B, Mitchell D, Reue K, Morrow A, Lusis AJ, LeBoeuf RC. Ath-l, a gene determining atherosclerosis susceptibility and high density lipoprotein levels in mice. Proc Natl Acad Sci USA 1987; 84:3763–3767. 17. Palinski W, Ord VA, Plump AS, Breslow JL, Steinberg D, Witztum JL. ApoEdeficient mice are a model of lipoprotein oxidation in atherogenesis. Demonstration of oxidation-specific epitopes in lesions and high titers of autoantibodies to malondialdehyde-lysine in serum. Arterioscler Thromb 1994; 14:605–616. 18. Qiao JH, Xie PZ, Fishbein MC, et al. Pathology of atheromatous lesions in inbred and genetically engineered mice. Genetic determination of arterial calcification. Arterioscler Thromb 1994; 14:1480–1497. 19. Liao F, Andalibi A, Qiao JH, Allayee H, Fogelman AM, Lusis AJ. Genetic evidence for a common pathway mediating oxidative stress, inflammatory gene induction, and aortic fatty streak formation in mice. J Clin Invest 1994; 94:877–884.

Mouse Genetic Models in Atherosclerosis

335

20. Silver LM. Mouse Genetics. Concepts and Applications. New York: Oxford University Press, 1995. 21. Hogan B, Constantini F, Lacy E. Manipulating the Mouse Embryo: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Laboratory, 1986. 22. Wasserman PM, DePamphilis ML. Guide to Techniques in Mouse Development. San Diego: Academic Press, 1993. 23. Joyner ALe. Gene Targeting. A Practical Approach. New York: Oxford University Press, 1993. 24. Paigen B, Ishida BY, Verstuyft J, Winters RB, Albee D. Atherosclerosis susceptibility differences among progenitors of recombinant inbred strains of mice. Arteriosclerosis 1990; 10:316–323. 25. Stewart-Phillips J, Lough J. Pathology of atherosclerosis in cholesterol-fed, susceptible mice. Atherosclerosis 1991; 90:211–218. 26. Zhang SH, Reddick RL, Piedrahita JA, Maeda N. Spontaneous hypercholesterolemia and arterial lesions in mice lacking apolipoprotein E. Science 1992; 258:468–471. 27. Plump AS, Sith JD, Hayek T, et al. Severe hypercholesterolemia and atherosclerosis in apolipoprotein E–deficient mice created by homologous recombination in ES cells. Cell 1992; 71:343–353. 28. Nakashima Y, Plump AS, Raines ES, Breslow JL, Ross R. ApoE–deficient mice develop lesions of all phases of atherosclerosis throughout the arterial tree. Arterioscler Thromb 1994; 14:133–140. 29. Ishibashi S, Goldstein JL, Brown MS, Herz J, Burns DK. Massive xanthomatosis and atherosclerosis in cholesterol-fed low density lipoprotein receptor-negative mice. J Clin Invest 1994; 93:1885–1893. 30. Purcell-Huynh DA, Farese RV, Jr, Johnson DF, et al. Transgenic mice expressing high levels of human apolipoprotein B develop severe atherosclerotic lesions in response to a high-fat diet. J Clin Invest 1995; 95:2246–2257. 31. Powell-Braxton L, Venient M, Latvala RD, et al. A mouse model of human familial hypercholesterolemia: markedly elevated low density lipoprotein cholesterol levels and severe atherosclerosis on a low-fat chow diet. Nature Med 1998; 4:934–938. 32. Shimano H, Ohsuga J, Shimada M, et al. Inhibition of diet-induced atheroma formation in transgenic mice expressing apolipoprotein E in the arterial wall. J Clin Invest 1995; 95:469–476. 33. Fazio S, Babee VR, Murray AB, et al. Increased atherosclerosis in mice reconstituted with apolipoprotein E null macrophages. Proc Natl Acad Sci USA 1997; 94:4647– 4652. 34. Warden CH, Hedrick CC, Qiao J-H, Castellani LW, Lusis AJ. Atherosclerosis in transgenic mice overexpressing apolipoprotein A-II. Science 1993; 261:469–472. 35. Rubin EM, Krauss R, Spangler E, Verstuyft S, Clift S. Inhibition of early atherogenesis in transgenic mice by human apolipoprotein A-I. Nature 1991; 353:265–267. 36. Li H, Reddick RL, Maeda N. Lack of apoA-I is not associated with increased susceptibility to atheroclerosis in mice. Arterioscler Thromb 1993; 13:1814–1821. 37. Plump AS, Scott CJ, Breslow JL. Human apolipoprotein A-I gene expression increases high density lipoprotein and suppresses atherosclerosis in the apolipoprotein E-deficient mouse. Proc Natl Acad Sci USA 1994; 91:9607–9611. 38. Paszty C, Maeda N, Verstuyft J, Rubin EM. Apolipoprotein AI transgene corrects

336

39.

40. 41.

42.

43.

44. 45.

46. 47.

48.

49.

50.

51.

52.

53.

Reue apolipoprotein E deficiency-induced atherosclerosis in mice. J Clin Invest 1994; 94: 899–903. Smith JD, Trogan E, Ginsberg M, Grigaux C, Tian J, Miyata M. Decreased atherosclerosis in mice deficient in both macrophage colony-stimulating factor (op) and apolipoprotein E. Proc Natl Acad Sci USA 1995; 92:8264–8268. Linton MF, Atkinson JB, Fazio S. Prevention of atherosclerosis in apolipoprotein E–deficient mice by bone marrow transplantation. Science 1995; 267:1034–1037. Ishida BY, Blanche PJ, Nichols AV, Yashar M, Paigen B. Effects of atherogenic diet consumption on lipoproteins in mouse strains C57BL/6 and C3H. J Lipid Res 1991; 32:559–568. Linton MF, Farese RV, Jr, Chiesa G, et al. Transgenic mice expressing high plasma concentrations of human apolipoprotein B100 and lipoprotein(a). J Clin Invest 1993; 92:3029–3037. Callow MJ, Stoltzfus LJ, Lawn RM, Rubin EM. Expression of human apolipoprotein B and assembly of lipoprotein(a) in transgenic mice. Proc Natl Acad Sci USA 1994; 91:2130–2134. Scanu AM, Fless GM. Lipoprotein(a). Heterogeneity and biological relevance. J Clin Invest 1990; 85:1709–1715. Lawn RM, Wade DP, Hammer RE, Chiesa G, Verstuyft JG, Rubin EM. Atherogenesis in transgenic mice expressing human apolipoprotein(a). Nature 1992; 360:670– 672. Tall AR. Plasma lipid transfer proteins. J Lipid Res 1986; 27:359–365. Marotti KR, Castle CK, Murray RW, Rehberg EF, Polites HG, Melchior GW. The role of cholesteryl ester transfer protein in primate apolipoprotein A-I metabolism. Insights from studies with transgenic mice. Arterioscler Thromb 1992; 12:736–744. Agellon LB, Walsh A, Hayek T, et al. Reduced high density lipoprotein cholesterol in human cholesterol ester transfer protein transgenic mice. J Biol Chem 1991; 266: 10796–10801. Hayek T, Chajek-Shaul T, Walsh A, et al. An interaction between the human cholesteryl ester transfer (CETP) and apolipoprotein A-I genes in transgenic mice results in a profound CETP-mediated depression of high density lipoprotein cholesterol levels. J Clin Invest 1992; 90:505–510. Brunzell JD. Familial lipoprotein lipase deficiency and other causes of chylomicronemia syndromes. In: Scriver RS, Beaudet AL, Sly WS, Valli D, eds. The Metabolic Basis of Inherited Disease. New York: McGraw-Hill, 1985:1913–1932. Weinstock PH, Bisgaier CL, Aalto-Setala K, et al. Severe hypertriglyceridemia, reduced high density lipoprotein, and neonatal death in lipoprotein lipase knockout mice. J Clin Invest 1995; 96:2555–2568. Coleman T, Seip RL, Gimble JM, Lee D, Maeda N, Semenkovich CF. COOH-terminal disruption of lipoprotein lipase in mice is lethal in homozygotes, but heterozygotes have elevated triglycerides and impaired enzyme activity. J Biol Chem 1995; 270:12518–12525. Farese RV, Jr, Ruland SL, Flynn LM, Stokowski RP, Young SG. Knockout of the mouse apolipoprotein B gene results in embryonic lethality in homozygotes and protection against diet-induced hypercholesterolemia in heterozygotes. Proc Natl Acad Sci USA 1995; 92:1774–1778.

Mouse Genetic Models in Atherosclerosis

337

54. Homanics GE, Smith TJ, Zhang SH, Lee D, Young SG, Maeda N. Targeted modification of the apolipoprotein B gene results in hypobetalipoproteinemia and development abnormalities in mice. Proc Natl Acad Sci USA 1993; 90:2389–2393. 55. van Vlijmen BJM, van den Maagdenberg AMJM, Gijbels MJJ, et al. Diet-induced hyperlipoproteinemia and atherosclerosis in apolipoprotein E3-Leiden transgenic mice. J Clin Invest 1994; 93:1403–1410. 56. Fazio S, Lee Y, Sheng X, Rall SC, Jr. Type III hyperlipoproteinemic phenotype in transgenic mice expressing dysfunctional apolipoprotein. J Clin Invest 1993; 92: 1497–1503. 57. Ito Y, Azrolan N, O’Connell A, Walsh A, Breslow JL. Hypertriglyceridemia as a result of human apolipoprotein C-III gene expression in transgenic mice. Science 1990; 249:790–793. 58. Shachter NS, Hayek T, Leff T, et al. Overexpression of apolipoprotein C-II causes hypertriglyceridemia in transgenic mice. J Clin Invest 1994; 93:1683–1690. 59. Aalto-Setala K, Fisher EA, Chen X, et al. Mechanism of hypertriglyceridemia in human apoC-III transgenic mice: diminished VLDL fractionaral catabolic rate associated with increased apoC-III and reduced apoE on the particles. J Clin Invest 1992; 90:1889–1900. 60. Masucci-Magoulas L, Goldberg IJ, Bisgaier CL, et al. A mouse model with features of familial combined hyperlipidemia. Science 1997; 275:391–394. 61. Goldstein JL, Schrott HG, Hazzard WR, Bierman EL, Motulsky AG. Hyperlipidemia in coronary heart disease II. Genetic analysis of lipid levels in 176 families and delineation of a new inherited disorder, combined hyperlipidemia. J Clin Invest 1973; 52:1544–1568. 62. Duverger N, Tremp G, Caillaud J-M, et al. Protection against atherogenesis in mice mediated by human apolipoprotein A-IV. Science 1996; 273:966–968. 63. Cohen RD, Castellani LW, Qiao J-H, Van Lenten BJ, Lusis AJ, Reue K. Reduced aortic lesions and elevated high density lipoprotein levels in transgenic mice overexpressing mouse apolipoprotein A-IV. J Clin Invest 1997; 99:1906–1916. 64. Plump AS, Erickson SK, Weng W, Partin JS, Breslow JL, Williams DL. Apolipoprotein A-I is required for cholesteryl ester accumulation in steroidogenic cells and for normal adrenal steroid production. J Clin Invest 1996; 97:2660–2671. 65. Netea MG, Demacker PNM, Kullberg BJ, et al. Low-density lipoprotein receptor– deficient mice are protected against lethal endotoxemia and severe Gram-negative infections. J Clin Invest 1996; 97:1366–1372. 66. Ishibashi S, Herz J, Maeda N, Goldstein JL, Brown MS. The two-receptor model of lipoprotein clearance: tests of the hypothesis in ‘‘knockout’’ mice lacking the low density lipoprotein receptor, apolipoprotein E, or both proteins. Proc Natl Acad Sci USA 1994; 91:4431–4435. 67. Herz J, Clouthier DE, Hammer RE. LDL receptor–related protein internalizes and degrades uPA-PAI-1 complexes and is essential for embryo implantation. Cell 1992; 71:411–421. 68. Willnow TE, Sheng Z, Ishibashi S, Herz J. Inhibition of hepatic chylomicron remnant uptake by gene transfer of a receptor antagonist. Science 1994; 264:1471–1474. 69. Levak-Frank S, Weinstock PH, Hayek T, et al. Induced mutant mice expressing lipoprotein lipase exclusively in muscle have subnormal triglycerides yet reduced

338

70.

71. 72.

73. 74.

75.

76.

77.

78. 79.

80. 81. 82. 83. 84. 85. 86.

Reue high density lipoprotein cholesterol levels in plasma. J Biol Chem 1997; 272:17182– 17190. Stra˚lfors P, Bjo¨rgell P, Belfrage P. Hormonal regulation of hormone-sensitive lipase in intact adipocytes: identification of phosphorylated sites and effects on the phosphorylation of lipolytic hormones and insulin. Proc Natl Acad Sci USA 1984; 81: 3317–3321. Khoo JC, Reue K, Steinberg D, Schotz MC. Expression of hormone-sensitive lipase mRNA in macrophages. J Lipid Res 1993; 34:1969–1974. Escary J-L, Choy HA, Reue K, Schotz MC. Hormone-sensitive lipase overexpression increases cholesteryl ester hydrolysis in macrophage foam cells. Arterioscler Thromb Vasc Biol 1998; 18:991–998. Escary J-L, Choy HA, Reue K, et al. Paradoxical effect on atherosclerosis of hormonesensitive lipase overexpression in macrophages. J Lipid Res 1999; 40:397–404. McCormick SPA, Ng JK, Taylor S, Flynn LM, Hammer RE, Young SG. Mutagenesis of the human apolipoprotein B gene in a yeast artificial chromosome reveals the site of attachment for apolipoprotein(a). Proc Natl Acad Sci USA 1995; 92:10147– 10151. Walsh A, Azrolan N, Wang K, Marcigliano A, O’Connell A, Breslow JL. Intestinal expression of the human apoA-I gene in transgenic mice is controlled by a DNA region 3′ to the gene in the promoter of the adjacent convergently transcribed apoCIII gene. J Lipid Res 1993; 34:617–623. Simonet WS, Bucay N, Pitas RE, Lauer SJ, Taylor JM. Multiple tissue-specific elements control the apolipoprotein E/C-I gene locus in transgenic mice. J Biol Chem 1991; 265:8651–8654. Shachter N, Zhu Y, Walsh A, Breslow JL, Smith JD. Localization of a liver-specific enhancer in the apolipoprotein E/C-I/C-II gene locus. J Lipid Res 1993; 34:1699– 1707. Dang Q, Walker D, Taylor S, et al. Structure of the hepatic control region of the human apolipoprotein E/C-I gene locus. J Biol Chem 1995; 270:22577–22585. Jiang XC, Agellon LB, Walsh A, Breslow JL, AT. Dietary cholesterol increases transcription of the human cholesteryl ester transfer protein gene in transgenic mice. Dependence on natural flanking sequences. J Clin Invest 1992; 90:1290–1295. Reue K, Doolittle MH. Naturally occurring mutations in mice affecting lipid transport and metabolism. J Lipid Res 1996; 37:1387–1405. Van Etten WJ, Steen RG, Nguyen H, et al. Radiation hybrid map of the mouse genome. Nature Genet 1999; 22:384–387. Nusbaum C, Slonim DK, Harris KL, et al. A YAC-based physical map of the mouse genome. Nature Genet 1999; 22:388–393. Avner P. Complex traits and polygenic inheritance in the mouse. Methods: A companion to Methods in Enzymology 1998; 14:191–198. Boehm T. Positional cloning and gene identification. Methods: A Companion to Methods in Enzymology 1998; 14:152–158. Castellani LW, Weinreb A, Bodnar J, et al. Mapping of a gene for combined hyperlipidaemia in a mutant mouse strain. Nature Genet 1998; 18:374–377. Pajukanta P, Nuotio I, Terwilliger JD, et al. Linkage of familial combined hyperlipidaemia to chromosome 1q21-q23. Nature Genet 1998; 18:369–373.

Mouse Genetic Models in Atherosclerosis

339

87. Rehnmark S, Giometti CS, Slavin BG, Doolittle MH, Reue K. The fatty liver dystrophy mouse: microvesicular steatosis associated with reduced fatty acid oxidation and altered expression levels of peroxisome proliferator-regulated proteins. J Lipid Res 1998; 39:2209–2217. 88. Klingenspor M, Xu P, Cohen RD, Welch C, Reue K. Altered gene expression pattern in the fatty liver dystrophy mouse reveals impaired insulin-mediated cytoskeleton dynamics. J Biol Chem 1999; 274:23078–23084. 89. Reue K, Xu P, Wang X-P, Slavin BG. Adipose tissue deficiency, insulin resistance, and increased atherosclerosis result from mutation in the mouse fatty liver dystrophy ( fld) gene. J Lipid Res 2000; 41:1067–1076. 90. Siep M, Trygstad O. Generalized lipodystrophy, congenital and acquired (lipoatrophy). Acta Paediatr Scand 1996; 413(suppl):2–28. 91. Pe´terfy M, Phan J, Oswell GM, Xu P, Reue K. Genetic, physical and transcript map of the fld region on mouse chromosome 12. Genomics 1999; 62:436–444. 92. Pe´terfy M, Phan J, Xu P, Reue K. Lipodystrophy in the fld mouse results from mutation in a novel nuclear protein, lipin. Nature Genet. (In press). 93. Mouzeyan A, Choi J-H, Allayee H, et al. A gene conferring resistance to diet-induced hypercholesterolemia and atherosclerosis on mouse chromosome 2. J Lipid Res 2000; 41:573–582.

Part Four MODULATORS IN CARDIORESPIRATORY BIOLOGY

14 Apoptosis in Development and Disease

TIMOTHY S. ZHENG Biogen, Incorporated Cambridge, Massachusetts

ALEX K.-K. KUAN and STE´PHANE HUNOT Yale University School of Medicine New Haven, Connecticut

RICHARD A. FLAVELL Yale University School of Medicine and Howard Hughes Medical Institute New Haven, Connecticut

I.

Introduction

Much under appreciated until recently, physiological mechanisms of cell death are required throughout the development and adulthood of all multicellular organisms [1,2]. From the formation of xylem in plants to the generation of the interdigital space in vertebrates, well-orchestrated cell death events are critical for the removal of superfluous cells as a vital part of tissue sculpting. Equally importantly, physiological cell death enables the elimination of unwanted extra cells to maintain cellular homeostasis in developed adults, as well as to rid the organisms of harmful cells or cells with serious cellular or genomic damage that are potentially dangerous. Morphologically recognized as apoptosis, physiological cell death is characterized by a distinct set of morphological and biochemical features, including chromatin condensation, internucleosomal DNA fragmentation, and, perhaps most importantly, cell surface alterations that signal for the rapid recognition and engulfment of apoptotic cells by neighboring phagocytic cells, thus avoiding the induction of inflammation [3]. As for many other biological processes, genetic models, especially that of Caenorhabditis elegans (see Chap. 2), have made seminal contributions to our understanding of the evolutionarily well-conserved genetic and biochemical path341

342

Zheng et al.

ways underlying physiological cell death. In this chapter, we mainly focus on the role of apoptosis in the vertebrate, in particular as revealed by various murine transgenic model systems. A. Historical Perspective

Although the actual term apoptosis was not introduced until the early 1970s by Kerr and colleagues to define a type of cell death distinct from necrosis based on unique morphological characteristics [4], the presence of physiological cell death was first documented during amphibian development in frog embryos in 1842 [5]. However, it was not until 1951 when Glu¨cksmann provided the first comprehensive review on physiological cell death in which he pointed out that naturally occurring cell death had been observed in almost all tissues in vertebrates and concluded that ‘‘there can be no doubt that cell deaths occur regularly at certain developmental stages of all vertebrate embryos’’ [6]. In their landmark experiment that led to the ‘‘birth’’ of apoptosis, Kerr et al. observed that after ligation of the portal vein, liver cells underwent two different types of cell death that can be defined morphologically [4]. Whereas dying hepatocytes in areas immediately surrounding the vessel exhibited classic necrotic features including cellular swelling, mitochondria damage, and cytoplasmic membrane rupturing, liver cells in the peripheral areas underwent slow ischemic death, and instead of swelling up, those dying cells actually shrunk with blebbed, yet intact cytoplasmic membranes. This result successfully delineated the two distinct death mechanisms in a single cell type. Eight years later, in an attempt to understand the mechanism of apoptosis, Wyllie demonstrated for the first time that apoptosis could be induced experimentally in isolated cells in vitro. By exposing thymocytes to glucocorticoid, he also discovered that cellular DNA was being fragmented to generate a ‘‘ladder’’ of DNA bands during apoptotic death, indicating the activation of an endogenous endonuclease. Based on their study and other previous observations, Wyllie and his colleagues hypothesized that apoptosis is the common form of cell death under physiological conditions, such as naturally occurring death during development and homeostasis, whereas necrosis occurs only in response to pathological conditions such as injury. According to their model, the morphological and biochemical features of apoptosis, such as cell shrinkage, degradation of genetic material, and removal by phagocytosis, make it more congruous with the subtle removal of cells during sculpting of developing tissues than the far more violent necrosis [3]. B. Apoptotic Pathways: An Overview

The above-mentioned pioneering work of Kerr, Wyllie, and others led to subsequent demonstrations that apoptosis could be induced in vitro in almost all cell types when given appropriate stimuli; providing further proof that each cell had

Apoptosis in Development and Disease

343

the inherent capacity to trigger its own death machinery [7]. Early mechanistic studies of apoptosis, however, mainly focused on the effects of various pharmacological reagents on apoptosis and often produced inconsistent results on the relevance of calcium influx, cAMP, and protein phosphorylation to apoptosis [8]. The breakthrough in understanding cell death mechanisms came from genetic studies in C. elegans by Horvitz and his colleagues over the last decade. The clearly defined and highly consistent lineage commitment during C. elegans development made it possible to trace the fate of each single cell, and it was found that an appreciable number of cells are eliminated shortly after their generation. Importantly, these deaths occur in a highly reproducible manner in that the same cells die at the same exact time for every animal, indicating the existence of a precisely controlled death mechanism. Direct microscopic observations of these transparent organisms revealed that these dying cells all exhibited the diagnostic characteristics of apoptosis, providing indisputable evidence that apoptosis is indeed the preferred form of physiological cell death during development of multicellular organisms [9]. Further genetic screening and molecular analysis have uncovered more than a dozen genes whose mutation affects programmed cell death (PCD) in C. elegans. In particular, two genes, ced-3 and ced-4, were found to be absolutely required in an autonomous fashion for PCD to occur, as functional mutations in either gene allow the survival of those cells that are destined to die [10]. In contrast, another gene, ced-9, was identified as the only gene whose activity was necessary and sufficient to inhibit cell death in C. elegans, as ced-9 inactivation causes the death of cells that normally survive and overexpression of wild-type ced-9 results in the rescue of cells that otherwise die. Perhaps most interestingly, mutations in ced-3 and ced-4 can completely block the ectopic death caused by the lack of ced-9 function, suggesting the existence of a ced-3– and ced-4–mediated cell death pathway in C. elegans that can be negatively regulated by ced-9 [11]. The evolutionary conservation of the death pathway between nematodes and mammals became evident when the cloning of ced-3 and ced-9 revealed that they indeed encoded proteins homologous to the known mammalian proteins interleukin-1α converting enzyme (ICE) and Bcl-2, respectively [12,13]. With expected greater complexity in mammals, other mammalian homologs of ced-3 and ced-9 were subsequently discovered and together they form the ced-3/caspase family and Bcl-2 protein family [14,15]. More recently, Apaf-1 (apoptotic protease-activating factor-1), a candidate mammalian homolog of ced-4, was identified biochemically by its ability to activate one of the mammalian caspases, caspase-9; thus filling in the last piece of the conservation puzzle [16]. To date, it is well accepted that diverse apoptotic stimuli converge into this common apoptotic pathway that eventually leads to the execution of physiological cell death (Fig. 1).

344

Zheng et al.

Figure 1 The apoptotic pathway. Diverse apoptotic stimuli converge into a common apoptosis pathway with evolutionarily conserved families of cysteine proteases (ced-3/ caspase) and their adaptors (ced-4/Apaf-1) as central executioners. The conservation is also reflected by the presence of both positive (EGL1/Bax) and negative (ced-9/Bcl-2) regulators of apoptosis that belong to the ced-9/Bcl-2 superfamily. C. Caspases and the Caspase Activation Cascade

Central to the apoptotic pathway, is a family of cysteinyl aspartate–specific proteinases known as caspases. Similar to many other intracellular proteases, all caspases are synthesized de novo as proenzymes containing three domains (N terminal prodomain, large P20 subunit, and small P10 subunit) with little, if any, catalytic activity. In response to apoptotic stimulation, these otherwise latent proteases become activated and in turn proteolytically cleave a number of cellular proteins whose degradation lead to eventual cell death. Importantly, activation of caspases usually requires proteolytic cleavages between domains and reassociation of the large and small subunits, and the proteolytic cleavages required for the generation of active enzymes occur at sites that contain caspase consensus sequences; strongly implying that caspases are activated autocatalytically and/ or by other caspases in a sequential manner. It is postulated that such a cascade mechanism would not only allow simultaneous regulation of the activities of multiple caspases through stringent control over the initiation of upstream caspase(s), but also provide a positive-feedback mechanism to amplify the activation process and thus ensure the rapid elimination of the ‘‘doomed’’ cell [14]. The presence of such a hierarchy of caspases consisting of initiators and effectors was supported by the demonstration that following Fas signaling, caspase-1–like activity is rapidly induced followed by a slower and gradual increase

Apoptosis in Development and Disease

345

in caspase-3–like activity. Although the exact identity of caspases responsible for these activities is still uncertain, since multiple caspases possess similar enzymatic specificities, studies from both substrate specificity and prodomain function of various mammalian caspases strongly suggest a stepwise mechanism for caspase activation during apoptosis where long prodomain caspases such as caspase8 and caspase-9 are initiator caspases which undergo autoactivation, and activated initiator caspases subsequently activate downstream effector caspases that carry out the destruction of cellular components through proteolytic degradation of various target proteins (Fig. 2). Initiation of Caspase Cascade

Although the mechanistic details of how diverse stimuli converge into one common apoptotic machinery remains poorly understood, two well-studied apoptosis

Figure 2 The caspase cascade. Evidence from caspases substrate specificity, prodomain function, and how caspases are activated all point to the presence of a caspase hierarchy in which upstream caspases (initiator caspases) are directly activated in response to apoptotic stimuli. These activated initiator caspases in turn proteolytically process downstream caspases (effector caspases) whose activation leads to the final destruction of apoptotic cells through cleaving various cellular targets.

346

Zheng et al.

signaling pathways support the overall scheme that in response to their corresponding apoptotic signals, a limited number of adaptors mediate the recruitment and subsequent activation of their respective initiator caspases [14]. Perhaps the best-characterized apoptotic pathway, signaling of the death receptor Fas relies upon the stepwise formation of the death-inducing signaling complex (DISC) following the trimerization of Fas receptor. One central component of this complex is the apoptotic adaptor molecule FADD, which is rapidly recruited to the Fas receptor through protein–protein interactions mediated between the death domain (DD) motifs present both in the cytoplasmic tail of Fas and in the C terminus of FADD. Following its recruitment to the DISC complex, through its N terminal death effector domain (DED), FADD further engages initiator caspases such as procaspase-8 and/or procaspase-10 via DED–DED interactions. This induced proximity of procaspase-8 and/or procaspase-10 proteins, via a yet unknown mechanism, results in their self-processing and activation and ultimately leads to initiation of the whole caspase cascade [7]. A second pathway capable of triggering the activation of caspase cascade requires cytochrome c, a component of the mitochondrial electron transport chain. During apoptosis induced by many stimuli, cytochrome c is released from the intermembrane space of the mitochondria into the cytosol in a poorly understood process that can be regulated both positively and negatively by members of the Bcl-2 family. Once in the cytosol, cytochrome c binds to the apoptotic adaptor Apaf-1 and, with the help of ATP, recruits procaspase-9 through CARD–CARD interaction to form an effector caspase-activating complex, also known as the apoptosome. It is within this apoptosome complex where Apaf-1 potentiates initiator procaspase-9 to undergo autocatalysis or transcatalysis to generate active caspase-9 that is capable of activating effector phase caspases such as caspase-3 [18]. Taken together, emerging evidence supports the model of oligomerizationinduced autocatalysis or transcatalysis of procaspases. According to this model, procaspases are normally present in cells with low but detectable enzymatic activity, but in a conformation or a complex that prevents their autocatalysis or transcatalysis. Upon apoptotic stimulation, recruitment by adaptors leads to one or more of the following outcomes that subsequently results in autocatalysis or transcatalysis: dissociation of inhibitors, change of procaspase conformation, and aggregation of procaspases. Regulation of Caspase Cascade

Since the caspase cascade is a self-amplifying process, it is believed that most, if not all, regulation of the caspase cascade resides at the level of initiation. In fact, the very requirement for stepwise assembly of various components to form caspase-activating complexes such as DISC and the apoptosome implicates the

Apoptosis in Development and Disease

347

existence of multiple safeguarding mechanisms that exert stringent controls at various levels over this cellular switch between life and death. The most-studied regulatory mechanism of caspase activation is that provided by members of the Bcl-2 superfamily [15]. Originally identified as the gene activated in non-Hodgkin’s follicular lymphomas due to chromosomal translocation, bcl-2 was later found to be a negative regulator of cell death, and the subsequent revelation that bcl-2 is the mammalian homolog of ced-9 further established it as one of the central components of an evolutionarily conserved cell death pathway. To date, at least 15 Bcl-2 family members have been identified in mammalian cells and interestingly, although many possess antiapoptotic activity as Bcl-2 does, other members can antagonize Bcl-2–like functions and are in fact ‘‘death inducers.’’ Two mutually nonexclusive mechanisms by which Bcl-2 family members regulate the activation of caspases have been proposed [15]. In a model analogous to biochemical studies in the C. elegans system where Ced-9 can directly bind to Ced-4, antiapoptotic Bcl-2 proteins such as Bcl-xL have been shown directly to bind to Apaf-1. Such binding might perturb the ability of Apaf-1 to bind to the CARD domain of procaspase-9 and thus inhibits the activation of caspase9. According to this model, proapoptotic members of the Bcl-2 family, such as Bax and Bad, function, at least in part, by heterodimerizing with prosurvival members like Bcl-2 and thereby titrating out their antiapoptotic activity. The second model, although not disputing the ability of Bcl-2 and Bcl-xL directly to inhibit Apaf-1–mediated caspase-9 activation in cytosol, argues that many Bcl-2 family members localize predominantly in the outer membrane of the mitochondria, and contends that their physiological function is likely to either facilitate or block the release of cytochrome c and other factors capable of inducing apoptosis from the mitochondria. In support of this viewpoint, Bcl-2 and Bcl-xL have been shown to inhibit the release of cytochrome c in vitro independent of caspase activity; reaffirming the notion that Bcl-2 and its relatives act upstream of caspase activation. More recently, both Bax and Bid were found to translocate into the mitochondria upon apoptotic stimulation and subsequently mediate the release of cytochrome c in a Bcl-2–inhibitable fashion. Although very little is known about the molecular basis for how Bcl-2 family members antagonize each other in mediating the release of cytochrome c, crystallographic analysis of Bcl-xL and other in vitro studies seem to favor the hypothesis that several Bcl-2 family members are in fact pore-forming proteins that are capable of regulating the transport of small molecules across the membrane. A second family of proteins that regulates caspase activation is a group of endogenous caspase inhibitors known as inhibitor of apoptosis proteins (IAPs), a family of proteins characterized by the presence of at least one baculovirus IAP repeat (BIR) domain [19]. Interestingly, several IAPs, including c-IAP1, c-IAP2, and x-linked IAP (XIAP), also contain a conserved RING finger domain

348

Zheng et al.

at their C termini that is also present in the viral IAP. Although the RING domain is required for the ability of viral IAP to suppress apoptosis in insect cells, neuronal apoptosis inhibitory protein (NAIP), and surviving, both lacking RING domains, inhibit apoptosis in mammalian cells. Despite the uncertainty about the functional importance of the RING domain, recent studies indicate that the BIR domain contributes to IAPs’ antiapoptotic activity by mediating direct interactions between IAPs and caspases. Specifically, XIAP can potently and specifically inhibit active caspase-3 and caspase-7 in a BIR-dependent manner, whereas having no effect on active caspases 1, 6, or 8. It has also been shown that both c-IAP1 and c-IAP2 can selectively suppress the activity of caspase-3 and caspase-7, albeit with a much lower potency. There seems to be a compensatory mechanism for c-IAPs, however, since they are capable of ‘‘compensating’’ their low direct caspase-inhibitory activity with the induction of prosurvival nuclear factor-κB (NF-κB) activity through their RING finger domain. Taken together, although the antiapoptotic activities of IAPs are well established, the underlying mechanisms involved are far from clear. Since the biological significance of inhibiting activated effector caspases is still unknown, it is possible that the in vivo function of IAPs is to inhibit the residual proteolytic activity of initiator procaspases such as procaspase-8 or procaspase-9. It must be noted that, in addition to regulation by Bcl-2 family proteins and IAPs, most recent studies indicate that caspase activation can also be regulated through other mechanisms such as phosphorylation, nitrosylation, and compartmentalization of caspases. In studying the mechanism by which p21-Ras inhibits apoptosis, it was shown that p21-Ras activated Akt kinase which could phosphorylate procaspase-9. More importantly, phosphorylation of caspase-9 by Akt reduced its proteolytic activity, indicating the existence of a previously unknown means of regulating caspase activation [20]. Similarly, it has been demonstrated that caspase activity can be negatively regulated through nitrosylation, and Fas signaling induces denitrosylation of caspase-3 [21]. To further complicate the matter, a number of studies have demonstrated that upon apoptotic stimulation, several caspases, including caspases 2, 3, 7, and 9 undergo intracellular translocation, although the potential functional significance of such translocations remains to be elucidated [22–24].

II. Apoptosis in Vertebrate Development A. Overview

Despite the early recognition of the existence of physiological cell death during vertebrate development, its importance in development was not fully realized until recently. Thanks to the various genetic models in C. elegans, Drosophila, and mice, it is now evident that physiological cell death in the form of apoptosis is critical for deleting structures that are no longer needed, controlling cell num-

Apoptosis in Development and Disease

349

bers, as well as molding tissue structures [1]. Studies using these model systems further suggest that the necessity of using apoptosis as a crucial mechanism to regulate the development of tissue and organs has increased during evolution, as defects in cell death result in much more severe phenotypes in Drosophila and mice than in the nematode C. elegans [25]. B. Apoptosis and Neural Development

The nervous system is one of the exemplary organs where cells are overproduced during development and apoptosis is subsequently employed as the mechanism for cell number adjustment. It has long been recognized that neurogenesis produces many more neurons in a given structure during the maturation of the vertebrate nervous system than exist in the adult organism [26]. It is therefore interesting that developmental defects were observed in several genetic models where critical members of the apoptotic pathway were inactivated. Caspase-3 and Caspase-9 Are Required for Developmental Apoptosis of Neurons

In line with the hypothesis that caspase-9 activates caspase-3 in a linear fashion, both caspase-3 and caspase-9 null mutant mice exhibited essentially the same developmental defects in the nervous system. Most homozygous caspase-3 and caspase-9 knockout mice die perinatally with severe brain malformations, including multiple indentations of the cerebrum and periventricular ectopic cell masses [27–29]. Close examination of the mutant embryonic tissues revealed a drastic reduction in the number of pyknotic cells in the neuroepithelial progenitor population around the periproliferative zone. Such an observation came as a surprise, as it was previously thought that neuronal cell death in developing vertebrates occurred at a much later stage and played an essential role in matching neurons and their targets to ensure the proper formation of synaptic connections in postmitotic regions. Thus, caspase-3 and caspase-9 knockout models not only confirmed the importance of cell death during neuronal development, but also revealed the existence of early cell death in the progenitor population that is not target-driven (see discussion below). The Bcl-2 Superfamily Has Both Proapoptotic and Antiapoptotic Effects

Similarly, transgenic and gene-targeting approaches have also confirmed the involvement of Bcl-2 family members in mediating apoptosis during nervous system development. Transgenic mice overexpressing Bcl-2 exhibited somewhat enlarged brain (12% increase in weight) and a 40–50% increase in the number of neurons in the facial nucleus [30]. In contrast, Bcl-2 null mutations caused fulminant lymphoid apoptosis, polycystic kidneys, and hypopigmentation without

350

Zheng et al.

any obvious cell death abnormality in the nervous system [31]. The major cell death suppressor in the nervous system appears to be Bcl-xL, as its null mutations caused extensive apoptosis of postmitotic neurons in the embryonic nervous system [32]. In contrast, naturally occurring neuronal death and apoptosis induced by the withdrawal of trophic factors were greatly reduced in mice lacking the proapoptotic gene Bax [33,34]. Furthermore, the Bax deficiency was able to prevent the increased apoptosis of postmitotic neurons in animals carrying a Bcl-xL null mutation, demonstrating an intracellular balance between proapoptotic and antiapoptotic effects within the Bcl-2 protein family [35]. The above-mentioned gene-targeting studies of caspases and Bcl-2 family proteins together indicate that the cell death pathway is remarkably conserved between the nematode and mammals. Another testimony to the evolutionarily conserved mechanism is provided by the recent genetic analysis of the interaction between caspase-3 and Bcl-xL in mammals. In C. elegans, the loss of function mutation of ced-3 (homolog of caspase-3) suppresses the ectopic cell deaths caused by the loss of function mutation of ced-9 (homolog of Bcl-xL) [11]. We therefore asked whether the null mutation of caspase-3 would decrease the ectopic neuronal apoptosis in Bcl-xL –deficient mice as predicted by an epistatic interaction of these two genes. Indeed, the aberrant neuronal apoptosis due to the Bcl-xL deficiency was abolished by the additional caspase-3 deficiency and, in fact, the neuronal phenotype of the caspase-3 and Bcl-xL double deficiency was literally indistinguishable from the caspase-3 single null mutation [36]. Apoptotic Defects in Founders and Postmitotic Neurons Have Distinct Consequences

Based on the various phenotypes of the Bax, Bcl-xL, caspase-3, and caspase-9 null mutations, it appears that the normal function of Bcl-xL is to inhibit the apoptotic effect of caspase-3 in the postmitotic neuronal population. Thus, Bcl-x deficiency causes increased apoptosis of postmitotic neurons [32], which is prevented by the additional absence of caspase-3 [36]. Bax appears to modulate the antiapoptotic effects of Bcl-xL; the null mutation of Bax thus reduces the normally occurring developmental death of postmitotic neurons without affecting the global formation of the nervous system [33,34]. Intriguingly, current studies indicate that caspase-3 and caspase-9 may have a preferential effect on early founder cells, which is not shared by Bax and Bcl-xL. The preferential effect of caspases on neuronal progenitor cells is best illustrated by cerebral malformations caused by null mutations of caspase-3 or caspase-9 (Fig. 3A–D). The entire cell population of the vertebrate telencephalon is generated in the ventricular zones and migrates into distinct cerebral areas [37]. Recent studies indicate apoptosis within the proliferative ventricular zones, which may regulate the size of the progenitor pool [38]. Conceivably, a reduction in the

Apoptosis in Development and Disease

351

Figure 3 Caspase-9 regulates the size of neuroepithelial progenitor pool. Gene targeting of caspase-9 typically produces an expanded and convoluted cerebral cortex with an increased number of neurons (compare A and C in wild-type to B and D in mutant mice). These abnormalities suggest that the caspase-9 null mutation primarily affects apoptosis of the founder cells in the ventricular zone (E,F).

352

Zheng et al.

apoptosis of founder cells in the ventricular zone would produce a large supernumerary progeny leading to a large cortical surface and thus multiple indentations and convolutions of the cerebrum, as typically observed in caspase-3 and caspase9 null mutations (Fig. 3E,F). These cortical malformations are similar to the rat spontaneous mutation, telencephalic internal structural heterotopia (TISH), and the human periventricular heterotopia anomaly [39,40]. Taken together, these results suggest that programmed cell death is not only important for adjusting the final number of neurons to match their target fields but is also essential for establishing the initial size of the progenitor pool in the developing nervous system. C. Apoptosis and Development of the Immune System

The daunting task faced by the immune system to fend off a colossal array of potential pathogens beckons a fundamental involvement of apoptosis in its development and homeostasis maintenance. By relying on a random gene rearrangement mechanism to generate receptor-bearing B and T cells capable of recognizing virtually all antigens, the immune system inadvertently produces a vast number of nonfunctional and potentially self-reactive lymphocytes that have to be eliminated. In fact, it has been estimated that a large majority of lymphocytes (⬎90%) generated in primary lymphoid organs undergo apoptosis as a result of either negative selection or unsuccessful positive selection [41,42]. Various studies using T-cell receptor (TCR) or immunoglobulin (Ig) transgenic models have definitively established the occurrence of both positive (survival of T and B cells bearing functional T and B receptors) and negative selection (elimination of self-reactive T and B cells) in vivo. It was further established that antigen receptor–mediated signaling were responsible for both selection events. Interestingly, members of the Bcl-2 family and caspase family seem to have different involvement during lymphoid development. Developmental regulated expression patterns of both Bcl-2 and Bcl-xL suggest their role in positive, but not negative selection of lymphoid cells. In supporting this hypothesis, Bcl-2 transgenic expression was shown to enhance positive selection of thymocytes, but failed to block the negative selection of both T and B lineage. On the other hand, Bcl-2– and Bcl-xL –deficient mice both exhibited massive apoptosis of lymphoid cells, confirming their critical role in mediating lymphocyte survival in vivo [31,32]. In contrast, although no single caspase-deficient mice to date have shown any defect in lymphoid development [25], caspase activity has been detected in various in vitro and in vivo models of negative selection. Transgenic expression of the pan-caspase viral inhibitor P35 in the thymus was reported to block negative selection in the F5 TCR transgenic model system, suggesting that caspase activation is a critical event in the elimination of self-reactive lymphocytes.

Apoptosis in Development and Disease

353

D. Apoptosis and Cardiac Development

One of the major surprises revealed by these apoptosis genetic models was the requirement of apoptosis during cardiac development. Mice lacking caspase-8, the upstream caspase primarily involved in death receptor–mediated apoptosis, died in utero around embryonic day 11.5 (E11.5), exhibiting impaired formation of cardiac muscles [43]. Importantly, a nearly identical developmental defect was also observed in embryos carrying a null mutation in FADD [44], the apoptotic adaptor molecule bridging death receptors and caspase-8; providing further evidence that apoptosis mediated through death receptor is crucial for proper cardiac development, a notion suggested by the constitutive expression of Fas and DR5 in cardiac muscles. Although it is currently not known how caspase-8 participates in cardiac muscle development, a recent study has demonstrated, for the first time, the existence of apoptotic cardiomyocytes during chicken development that is associated with caspase activation and its participation in remodeling and shortening of the cardiac outflow tract; a critical process for the proper connection between the ventricular chambers and the appropriate arterial trunks [45]. III. Apoptosis and Human Diseases In addition to its critical involvement during vertebrate development, apoptosis is also the principal mechanism by which cellular homeostasis is maintained, and its importance is perhaps best exemplified by the devastating consequences resulting from the dysregulation of apoptosis. Genetic alterations in this pathway that result in the survival of cells normally destined for destruction may underlie both tumorigenesis, such as in bcl-2 translocation-associated follicular lymphomas and fatal lymphoproliferation caused by Fas mutation in the lpr/lpr mouse mutant strain. On the other hand, defects that promote excessive cell death are thought to contribute to the pathogenesis of many neurological disorders ranging from acute ischemia to chronic neurodegenerative diseases such as Alzheimer’s disease as well as various viral infections and immunodeficiencies, including AIDS [2]. A. Neurodegenerative Diseases

Human neurodegenerative disorders are characterized by the progressive loss of selective neuronal populations. Although the mechanisms by which these neurons degenerate remain elusive, much progress has been made in the past decade in the elucidation of the molecular events leading to neuronal cell death under such pathological conditions. In particular, increasing evidence has raised the possibility that degenerating neurons may ultimately die by apoptosis. Primarily based on the TUNEL method, as well as dyes such as acridine orange or YOYO-1, in

354

Zheng et al.

situ detection of condensed neuronal nuclei with apparent DNA cleavage and fragmentation features has been reported in postmortem brain tissue from patients with a variety of neurodegenerative diseases, including Alzheimer’s disease (AD) [46–51], Parkinson’s disease (PD) [52], Huntington’s disease (HD) [46,53], and amyotrophic lateral sclerosis (ALS) [53–55]. Furthermore, Anglade et al. were able to show chromatin condensation and nucleus fragmentation profiles using electron microscopy in nigral dopaminergic neurons from patients with Parkinson’s disease; confirming that apoptosis represents a main feature of cell death in neurodegenerative disorders [56]. Given the obvious limitations of postmortem studies and the possibility of false-positive staining of necrotic nuclei by the TUNEL method [53], a more mechanistic approach for study of the role of apoptosis in human neurodegenerative diseases has been made possible by the development of several murine transgenic models for such diseases together with apoptotic genetic models in which genes implicated in the regulation of the apoptotic program are either overexpressed or inactivated. Among them, members of the Bcl-2–related proteins family and those of the caspase group of proteases have attracted most attention because of their key roles in neurogenesis. Not surprisingly, Bcl-2 has been shown to protect against diverse cytotoxic insults, and its overexpression in the brain has been shown to confer neuroprotection of spinal cord motor neurons in an animal model of familial amyotrophic lateral sclerosis (FALS); that is, mice carrying mutations in the gene encoding Cu/Zn superoxide dismutase (SOD) enzyme [57]. Interestingly, upregulation of Bax and downregulation of Bcl-2 expression have been reported in spinal cord motor neurons from patients with ALS, suggesting that the altered Bax/Bcl-2 ratio in these neurons could be critical for apoptosis to occur [58]. However, it is likely that Bcl-2 overexpression in this model can only delay the onset of motor neuron disease without affecting the progression of the disease. This may be an indication that Bcl-2 on its own is insufficient to prevent neuronal loss and that other elements of the apoptotic machinery are likely to be involved in nerve cell death as well. Indeed, Friendland et al. [59] reported that expression of a dominant negative of ICE (caspase-1) in neurons of mutant SOD mice was able to slow the symptomatic progression of the disease and delay mortality. Overexpression of Bcl-2 in transgenic mice has also been shown to confer neuroprotection in an animal model of PD; that is, mice treated by the neurotoxic 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) [60]. Increased bcl-2 gene expression has been reported in dopaminergic neurons from patients with PD; maybe reflecting an attempt for these neurons to escape from death [61]. Thus, it seems highly likely that Bcl-2 is not sufficient to prevent nerve cell death in PD. More recently, in vitro experiments have suggested that caspases might also play a role in MPTP-induced apoptosis in dopaminergic neurons [62], and consistent with this hypothesis, Klevenyi et al. [63] showed that

Apoptosis in Development and Disease

355

transgenic mice expressing a dominant negative mutant of caspase-1 were less sensitive to MPTP intoxication than wild-type animals. However, it is not entirely clear if neuroprotection observed in this model is due to inhibition of a caspase1–dependent apoptotic pathway or simply alterations in the inflammatory microenvironment resulting from deficiency in interleukin-1α (IL-1α) and IL-β production. One of the most interesting developments has been the recent realization that, in a number of inherited neurodegenerative disorders, the gene products involved in disease progression are direct substrates of caspases. For example, in the autosomal dominant neurodegenerative disorder Huntington’s disease (HD) that is characterized by a mutation in the huntington gene which leads to excessive CAG/polyglutamine repeats, evidence has been reported that caspases can cleave mutated Huntington protein and lead to the formation of proteinous aggregates of the cleaved fragment containing the polyglutamine tract that confer neurotoxicity [64,65]. This suggests that caspase cleavage of polyglutamine-containing proteins could represent a key step in the pathogenesis of HD. It is further demonstrated that in a mouse model of HD (mice expressing a variant of the Huntington protein containing an expanded polyglutamine repeat), intracerebroventricular administration of the caspase inhibitor zVAD and transgenic expression of a dominant negative caspase-1 both resulted in delayed onset of the disease and symptoms [66]. Once again, although this study provided in vivo evidence for a role of caspase-1 in HD, it is still unclear whether caspase-1– mediated pathophysiology in this model reflects a caspase-1–dependent apoptotic pathway. Indeed, one cannot rule out the possibility that endogenous production of mature IL-1β could trigger nerve cell death either directly, as previously described, or through induction of gliosis, since IL-1β has been reported to be a strong activator of astrocyte proliferation both in vitro and in vivo [67,68]. Nonetheless, inhibition of caspase-1 clearly remains a promising strategy to treat HD patients and perhaps other neurodegenerative diseases such as ALS. In addition to processing of Huntingtin by caspase-1, cleavage of the β-amyloid precursor protein (β-APP) by caspase-3 has also been implicated in the pathogenesis of AD. In both sporadic and familial AD, deposition of amyloid β-peptide (Aβ) generated by proteolytic processing of β-APP in senile plaques represents a common histopathological feature in this disease. Furthermore, Aβ-induced neurotoxicity has extensively been shown both in vitro and in vivo [69]. Gervais et al. has now showed that caspase-3 can efficiently cleave β-APP which leads to elevated Aβ formation [70]. Consistent with this hypothesis, caspase-3 colocalizes with Aβ in senile plaques, and its level is increased in dying pyramidal neurons in the brain of patients with AD. Altogether, increasing evidence suggests a twofold role (both the initiator and executor) for caspases in the pathogenesis of neurodegenerative diseases: Through the cleavage of key cellular targets, cas-

356

Zheng et al.

pases contribute to the initiation of a vicious cycle where a primary insult inducing the death by apoptosis in few neurons could eventually lead to the progressive but discrete cell loss over a period of several years. B. Apoptosis in Cardiac Disease

Using the TUNEL method together with histochemistry, apoptosis of cardiomyocytes and nonmyocytes have been identified in a number of cardiac diseases, including ischemia, reperfusion, Chagasic myocarditis, acute myocardial infarction, and arrhythmogenic cardiomyopathy [71]. Although the precise contribution of apoptosis to the progression of these diseases in vivo is difficult to assess, studies with various animal models have suggested that apoptosis is an important factor in their pathogenesis. For example, transgenic expression of Gαq, a receptor-mediated G protein–signaling molecule, resulted in initial hypertrophy followed by apoptotic cardiomyocytic death, indicating that robust apoptosis can directly result in fulminant heart failure [72]. Perhaps most importantly, release of cytochrome c from mitochondria and activation of caspase-3 have recently been demonstrated in explanted hearts from patients with end-stage cardiomyopathy, providing direct evidence supporting the role of apoptosis in continued loss of myocytes in vivo [73]. Through the development of various in vitro culture systems, several stimuli have been shown to induce apoptosis of cardiomyocytes: (1) stress conditions such as ischemia-induced activation of the JNK pathway and reactive oxygen radicals; (2) proinflammatory cytokines such as tumor necrosis factor (TNF-α); and (3) nitric oxide (NO) production as a result of iNOS induction. Although the involvement of these mediators in vivo remains to be established, these studies offer clues to the mechanistic basis of apoptosis in cardiocytes that might lay the foundation for the development of novel therapeutic regimens. C. Apoptosis and Autoimmune Diseases

In addition to shaping the lymphocyte repertoire during development, apoptosis is also responsible for ‘‘downsizing’’ the expanded lymphoid populations following antigenic encounter to ensure cellular homeostasis, avoid bystander activation, and decrease the level of toxic cytokines. Autoimmune diseases characterized by tissue destruction resulting from self-reactive T and/or B cells can arise from either defective negative selection during development of the immune system or breakdown of peripheral tolerance [74]. The best-studied example of failure to maintain peripheral tolerance is provided by the two naturally occurring autoimmune mouse models, lpr and gld, that carry autosomal recessive mutations in the death receptor Fas and its ligand, FasL, respectively. It is now clear that elimination of excessive lymphocytes after

Apoptosis in Development and Disease

357

immune responses requires a process called activation-induced cell death (AICD) of lymphocytes in which lymphocytes become Fas-expressing cells and are ultimately killed by Fas-mediated apoptosis. Thus, mutations in either Fas or Fas ligand result in an inability of the immune system to execute AICD properly and eventually lead to various lymphoproliferative syndromes that often have fatal consequences [75]. These two animal models not only provided us with an important tool in studying the apoptosis pathway, but also were instrumental in our understanding of the underlying mechanism behind a rare human disease called autoimmune lymphoproliferative disease (ALPS). Primarily affecting children, ALPS patients accumulate an unusual population of CD4,CD8 double-negative T cells and often develop autoimmune manifestations such as hemolytic anemia and thrombocytopenia due to circulating antibodies against their own red blood cells and platelets, respectively. Since these abnormalities resemble those observed in lpr and gld mice, it was soon revealed that these patients contain various mutations in the Fas apoptotic pathway, including Fas ligand, Fas, and caspase-10 [76]. At the other end of the autoimmune disease are the tissues/cells that are targets of the self-reactive T and B cells. Increasing evidence suggests that the destruction of these target cell populations also takes the form of apoptosis. For example, in autoimmune diabetes, insulin-secreting β cells can be killed by antigen-specific T cells by apoptosis [77]. Similarly, oligodendrocytes in multiple sclerosis [78] and thyrocytes in Hashimoto’s thyroiditis have all been shown to die by apoptosis [79]. These studies thus provide new insights into the mechanistic basis of pathogenesis of autoimmune diseases and may potentially offer new therapeutic opportunities through preventing the destruction of target tissues. D. Apoptosis and Cancer

A large body of evidence suggests that breakdown of the control of apoptosis may be a key step in the development of cancer [80]. In fact, the first mammalian apoptosis-regulating gene, bcl-2, was isolated as a chromosomal translocation break gene associated with lymphoma. Elevated expression level of bcl-2 due to the translocation results in B-cell follicular lymphoma by inhibiting cell death rather than promoting cell proliferation. In addition, many of the previously identified tumor-suppressor genes, such as p53 and myc, all exhibit apoptosis-inducing activity; further suggesting that these molecules inhibit tumorigenesis, at least in part, through promoting apoptosis of cancerous precursors [81]. In support of this view, both bcl-2–overexpressing transgenic and p53-deficient mice have spontaneous neoplastic hyperplasia associated with a defective response to apoptotic stimuli [82,83], indicating that inhibition of apoptosis, particularly in combination with dysregulated cell cycle control, can promote neoplastic transformation.

358

Zheng et al.

It has also become clear that inhibition of apoptosis can influence the response of tumor cells to anticancer therapy. This is most apparent for metastatic tumors with a decreased ability to undergo apoptosis [84]. Although the precise molecular basis for such an increase in the threshold for apoptosis is not clear, the apoptotic machinery in these cells is almost certainly perturbed, possibly by the elevated Bcl-2 expression level and/or defective caspase activation. Thus, advancement in our understanding of the regulation of apoptotic pathways will likely provide additional arsenals for anticancer treatment. E. Other Diseases

In addition to above-mentioned diseases, the dysfunctional control of apoptosis has also been suggested to contribute to the pathogenesis of a growing list of many other human diseases. Similar to the gradual loss of neurons in neurodegenerative disorders, acute neuronal death due to excitotoxicity following stroke, ischemia, or epileptic episodes has also been associated with apoptosis [85]. In an entirely different setting, there is also growing evidence that apoptosis of osteoblasts contributes significantly to the bone loss in osteoporosis, and it was recently shown that parathyroid hormone (PTH) treatment of osteoporosis increases bone formation by preventing osteoblastic apoptosis [86]. Finally, during various bacterial and viral infections ranging from HIV to bacterial meningitis, cell death in the infected cell populations also takes the form of apoptosis [87]. In fact, strategies that either enhance the death of infected T cells [88] or prevent healthy cells (including T cells, macrophages, and dendritic cells) from apoptosis are both under intensive investigation [89]. Although there is still a long way to go before our growing knowledge of the apoptotic pathway can translate into clinical applications, the recent successful demonstration of the use of a caspase inhibitor as a neuroprotective agent against acute bacterial meningitis in an animal model once again underscores the promise of combating human diseases using apoptosis-modulating strategies [90]. IV. Conclusions Half a century has passed since Glu¨cksmann’s seminal recognition that physiological cell death is a normal part of vertebrate development. Today, it has become an ineluctable fact that life is a delicate balance between cell proliferation and cell death. As the primary mechanism of physiological cell death, apoptosis not only plays a critical role during development for the proper formation of various organs/tissues, but also is intimately involved in the control of cellular homeostasis throughout the life span of the organism. As a consequence, breakdown of such homeostasis due to dysregulation of the apoptotic pathway inevitably results in altered cell physiology and contributes significantly to the pathogen-

Apoptosis in Development and Disease

359

esis of a wide range of human diseases ranging from cancer formation to viral infections. Although much remains unknown, great progress has been made in understanding the underlying mechanism by which the apoptotic pathway is regulated and the potential therapeutic applications therein. It is conceivable that within the foreseeable future therapeutic strategies based on modulation of apoptosis will provide new windows of opportunities for the treatment of various human diseases. References 1. Vaux DL, Korsmeyer SJ. Cell death in development. Cell 1999; 22:245–254. 2. Thompson CB. Apoptosis in the pathogenesis and treatment of disease. Science 1995; 267:1455–1462. 3. Wyllie AH, Kerr JFR, Currie AR. Cell Death: the significance of apoptosis. Int Rev Cytol 1980; 68:251–306. 4. Kerr JFR, Willie AH, Currie AR. Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer 1972; 26:239–257. 5. Vogt C. Untersuchungen u¨ber die Entwicklungsgeschichte der Geburtshelferkroete (Alytes obstetricians). Solothurn, Switzerland: Jent und Gassman, 1842. 6. Glu¨cksmann A. Cell deaths in normal vertebrate ontogeny. Biol Rev Cambridge Philos Soc 1951; 26:59–86. 7. Raff MC, Barres BA, Burne JF, Coles HF, Ishizaki Y, Jacobson MD. Programmed cell death and the control of cell survival: lessons from the nervous system. Science 1993; 262:695–700. 8. Ellis RE, Yuan J, Horvitz HR. Mechanisms and functions of cell death. Annu Rev Cell Biol 1991; 4:663–698. 9. Ellis RE, Horvitz HR. Genetic control of programmed cell death in the nematode Caenorhabditis elegans. Cell 1986; 44:817–829. 10. Yuan J, Horvitz HR. The Caenorhabditis elegans genes ced-3 and ced-4 act cellautonomously to cause programmed cell death. Dev Biol 1990; 138:33–41. 11. Hengartner MO, Ellis RE, Horvitz HR. C. elegans gene ced-9 protects cells from programmed cell death. Nature 1992; 356:494–499. 12. Yuan J, Shaham S, Ledoux S, Ellis HM, Horvitz HR. The C. elegans cell death gene ced-3 encodes a protein similar to the mammalian interleukin-1 beta converting enzyme. Cell 1993; 75:641–652. 13. Hengartner MO, Horvitz HR. C. elegans cell death gene ced-9 encodes a functional homolog of mammalian proto-oncogene bcl-2. Cell 1994; 76:665–676. 14. Thornberry NA, Lazebnik Y. Caspases: enemies within. Science 1998; 281:1312– 1316. 15. Adams JM, Cory S. The Bcl-2 protein family: arbiters of cell survival. Science 1998; 281:1322–1326. 16. Zou H, Henzel WJ, Liu X, Lutzchg A, Wang X. Apaf-1, a human protein homologous to c. elegans CED-4, participates in cytochrome c–dependent activation of caspase-3. Cell 1997; 90:405–413.

360

Zheng et al.

17. Ashkenzi A, Dixit VM. Death receptors: signaling and modulation. Science 1998; 281:1305–1308. 18. Green DR, Reed JC. Mitochondria and apoptosis. Science 1998; 281:1309–1312. 19. Deveraux QL, Reed JC. IAP family proteins—suppressors of apoptosis. Genes Dev 1999; 13:239–252. 20. Cardone MH, Roy N, Stennicke HR, Salvesen GS, Franke TF, Stanbridge E, Frisch S, Reed JC. Regulation of cell death protease caspase-9 by phosphorylation. Science 1998; 282:1318–1321. 21. Mannick JB, Hausladen A, Liu L, Hess DT, Zeng M, Miao QX, Kane LS, Gow AJ, Stamler JS. Fas-induced caspase denitrosylation. Science 1999; 284:651–654. 22. Chandler JM, Cohen GM, MacFarlane M. Different subcellular distribution of caspase-3 and caspase-7 following Fas-induced apoptosis in mouse liver. J Biol Chem 1998; 273:10815–10818. 23. Mancini M, Nicholson DW, Roy S, Thornberry NA, Peterson EP, Casciola-Rosen LA, Rosen A. The caspase-3 precursor has a cytosolic and mitochondrial distribution: implications for apoptotic signaling. J Cell Biol 1998; 140:1485–1495. 24. Susin SA, Lorenzo HK, Zamzami N, Marzo I, Brenner C, Larochette N, Prevost M-C, Alzari PM, Kroemer G. Mitochondrial release of caspase-2 and -9 during the apoptotic process. J Exp Med 1999; 189:381–393. 25. Zheng TS, Hunot S, Kuida K, Flavell RA. Caspase knockouts: matters of life and death. Cell Death Differ 1999; 6:1043-1053. 26. Oppenheim, RW. Cell death during development of the nervous system. Annu Rev Neurosci 1991; 14:453–501. 27. Kuida K, Zheng TS, Na S, Kuan C, Yang D, Karasuyama H, Rakic P, Flavell RA. Decreased apoptosis in the brain and premature lethality in CPP32-deficient mice. Nature 1996; 384:368–372. 28. Kuida K, Haydar TF, Kuan CY, Gu Y, Taya C, Karasuyama H, Su MSS, Rakic P, Flavell RA. Reduced apoptosis and cytochrome c–mediated caspase activation in mice lacking caspase 9. Cell 1998; 94: 325–337. 29. Hakem R, Hakem A, Duncan GS, Henderson JT, Woo M, Soengas MS, Elia A, de la Pompa JL, Kagi D, Khoo W, Potter J, Yoshida R, Kaufman SA, Lowe SW, Penninger JM, Mak TW. Differential requirement for caspase 9 in apoptotic pathways in vivo. Cell 1998; 94:339–352. 30. Martinou JC, Dubois-Dauphin M, Staple JK, Rodriguez I, Frankowski H, Missotten M, Albertini P, Talabot D, Catsicas S, Pietra C, Huarte J. Overexpression of Bcl-2 in transgenic mice protects neurons from naturally occurring cell death and experimental ischemia. Neuron 1994; 13:1017–1030. 31. Veis DJ, Sorenson CM, Shutter JR, Korsmeyer SJ. Bcl-2–deficient mice demonstrate fulminant lymphoid apoptosis, polycystic kidneys, and hypopigmented hair. Cell 1993; 75:229–240. 32. Motoyama N, Wang F, Roth KA, Sawa H, Nakayama KI, Nakayama K, Negishi I, Senju S, Zhang Q, Fuiji S, Loh DY. Massive cell death of immature hematopoietic cells and neurons in Bcl-x–deficient mice. Science 1995; 267:1506–1510. 33. Deckwerth TL, Elliott JL, Knudson CM, Johnson EM Jr, Snider WD, Korsmeyer SJ. BAX is required for neuronal death after trophic factor deprivation and during development. Neuron 1996; 17:401–411.

Apoptosis in Development and Disease

361

34. White FA, Keller-Peck CR, Knudson CM, Korsemeyer SJ, Snider WD. Widespread elimination of naturally occurring neuronal death in Bax-deficient mice. J Neurosci 1998; 18:1428–1439. 35. Shindler KS, Latham CB, Roth KA. Bax deficiency prevents the increased cell death of immature neurons in bcl-x–deficient mice. J Neurosci 1997; 17:3112–3119. 36. Roth KA, Kuan CY, Haydar TF, D’Sa-Eipper C, Shindler KS, Zheng TS, Kuida K, Flavell RA, Rakic P. Epistatic and independent functions of caspase-3 and Bcl-xL in developmental programmed cell death. Proc Natl Acad Sci USA 2000; 97:466–471. 37. Rakic P. Specification of cerebral cortical areas. Science 1988; 241:170–176. 38. Thomaidou D, Mione MC, Cavanagh JF, Parnavelas JG. Apoptosis and its relation to the cell cycle in the developing cerebral cortex. J Neurosci 1997; 17:1075–1085. 39. Eksioglu YZ, Scheffer IE, Cardness P, Knoll J, Dimario F, Ramsby G, Berg M, Kamuro K, Berkovic SF, Duyk M, Parisi J, Huttenlocher PR, Walsh CA. Periventricular heterotopia: an X-linked dominant epilepsy locus causing aberrant cerebral cortical development. Neuron 1996; 16:77–87. 40. Lee KS, Schottler F, Collins JL, Lanzino G, Couture D, Rao A, Hiramatsu KI, Goto Y, Hong SC, Caber H, Yamamoto H, Chen ZF, Bertram E, Berr S, Omary R, Scrable H, Jackson T, Goble J, Eisenman L. A genetic animal model of human neocortical heterotopia associated with seizures. J Neurosci 1997; 17:6236–6242. 41. von Boehmer H. Positive selection of lymphocytes. Cell 1994; 76:219–228. 42. Nossal GJV. Negative selection of lymphocytes. Cell 1994; 79:229–239. 43. Varfolomeev EE, Schuchmann M, Luria V, Chiannilkulchai N, Beckmann JS, Mett IL, Rebrikov D, Brodianski VM, Kemper OC, Kollet O, Lapidot T, Soffer D, Sobe T, Avraham KB, Goncharov T, Holtmann H, Lonai P, Wallach D. Targeted disruption of the mouse caspase-8 gene ablates cell death induction by the TNF receptors, Fas/Apo1, and DR3 and is lethal prenatally. Immunity 1998; 9:267–276. 44. Zhang J, Cado D, Chen A, Kabra NH, Winoto A. Fas-mediated apoptosis and activation-induced T-cell proliferation are defective in mice lacking FADD/Mort1. Nature 1998; 392:296–300. 45. Watanabe M, Choudhry A, Berlan M, Singal A, Siwik E, Mohr S, Fisher SA. Developmental remodeling and shortening of the cardiac outflow tract involves myocyte programmed cell death. Development 1998; 125:3809–3820. 46. Dragunow M, Faull RL, Lawlor P, Beilharz EJ, Singleton K, Walker EB, Mee E. In situ evidence for DNA fragmentation in Huntington’s disease striatum and Alzheimer’s disease temporal lobes. Neuroreport 1995; 6:1053–1057. 47. Sugaya K, Reeves M, McKinney M. Topographic associations between DNA fragmentation and Alzheimer’s disease neuropathology in the hippocampus. Neurochem Int 1997; 31:275–281. 48. Lassmann H, Bancher H, Breitschopf H, Wegiel M, Bobinski K, Jellinger K, Wisniewski HM. Cell death in Alzheimer’s disease evaluated by DNA fragmentation in situ. Acta Neuropathol 1995; 89:35–41. 49. Smale G, Nichols NR, Brady DR, Finch CE, Horton WE Jr. Evidence for apoptotic cell death in Alzheimer’s disease. Exp Neurol 1995; 133:225–330. 50. Li WP, Chan WY, Lai DT, Yew J. Terminal dUTP nick end labeling (TUNEL) positive cells in the different regions of the brain in normal aging and Alzheimer patients. J Mol Neurosci 1997; 8:75–82.

362

Zheng et al.

51. Tatton WG, Olanow CW. Apoptosis in neurodegenerative diseases: the role of mitochondria. Biochem Biophys Acta 1999; 1410:195–213. 52. Mochizuki H, Goto K, Mori H, Mizuno Y. Histochemical detection of apoptosis in Parkinson’s disease. J Neurol Sci 1996; 137:120–123. 53. Thomas LB, Gates DJ, Richfield EK, TF OB, Schweitzer JB, Steindler DA. DNA end labeling (TUNEL) in Huntington’s disease and other neuropathological conditions. Exp Neurol 1995; 133:265–272. 54. Yoshiyama Y, Yamada T, Asanuma K, Asahi T. Apoptosis related antigen, Le(Y) and nick-end labeling are positive in spinal motor neurons in amyotrophic lateral sclerosis. Acta Neuropathol 1994; 88:207–211. 55. Troost D, Aten J, Morsink F, de Jong JM. Apoptosis in amyotrophic lateral sclerosis is not restricted to motor neurons. Bcl-2 expression is increased in unaffected postcentral gyrus. Neuropathol Appl Neurobiol 1995; 21:498–504. 56. Anglade P, Vyas S, Javoy-Agid F, Herrero MT, Michel PP, Marquez J, MouattPrigent A, Ruberg M, Hirsch EC, Agid Y. Apoptosis and autophagy in nigral neurons of patients with Parkinson’s disease. Histol Histopathol 1997; 12:25–37. 57. Kostic V, Jackson-Lewis V, de Bilbao F, Dubois-Dauphin M, Przedborski S. Bcl-2: prolonging life in a transgenic mouse model of familial amyotrophic lateral sclerosis. Science 1997; 277:559–562. 58. Mu X, He J, Anderson DW, Trojanowski JQ, Springer JE. Altered expression of bcl-2 and bax mRNA in amyotrophic lateral sclerosis spinal cord motor neurons. Ann Neurol 1996; 40:379–386. 59. Friedlander RM, Brown RH, Gagliardini V, Wang J, Yuan J. Inhibition of ICE slows ALS in mice. Nature 1997; 388:31. 60. Offen D, Beart PM, Cheung NS, Pascoe CJ, Hochman A, Gorodin S, Melamed E, Bernard R, Bernard O. Transgenic mice expressing human Bcl-2 in their neurons are resistant to 6-hydroxydopamine and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine neurotoxicity. Proc Natl Acad Sci USA 1998; 95:5789–5794. 61. Mogi M, Harada M, Kondo T, Mizuno Y, Narabayashi H, Riederer P, Nagatsu T. Bcl-2 protein is increased in the brain from parkinsonian patients. Neurosci Lett 1996; 215:137–139. 62. Dodel RC, Du Y, Bales KR, Ling ZD, Carvey PM, Paul SM. Peptide inhibitors of caspase-3-like protease attenuate 1-methyl-4-phenylpyridinium–induced toxicity of cultured fetal rat mesencephalic dopamine neurons. Neuroscience 1998; 86:701– 707. 63. Klevenyi P, Andreassen O, Ferrante RJ, Schleicher JR Jr, Friedlander M, Beal MF. Transgenic mice expressing a dominant negative mutant interleukin-1β converting enzyme show resistance to MPTP neurotoxicity. Neuroreport 1999; 10:635–638. 64. Wellington CL, Ellerby LM, Hackam AS, Margolis RL, Trifiro MA, Singaraja R, McCutcheon K, Salvesen GS, Propp SS, Bromm M, Rowland KJ, Zhang T, Rasper D, Roy S, Thornberry N, Pinsky L, Kakizuka A, Ross CA, Nicholson DW, Bredesen DA, Hayden MR. Caspase cleavage of gene products associated with triplet expansion disorders generates truncated fragments containing the polyglutamine tract. J Biol Chem 1998; 273:9158–9167. 65. Ona VO, Li M, Vonsattel JPG, Andrews LJ, Khan SQ, Chung WM, Frey AS, Menon AS, Li X-J, Stieg PE, Yuan J, Penney JB, Young AB, Cha J-HJ, Friedlander RM.

Apoptosis in Development and Disease

66.

67. 68. 69. 70.

71. 72.

73.

74. 75.

76.

77. 78.

79.

80.

363

Inhibition of caspase-1 slows disease progression in a mouse model of Huntington’s disease. Nature 1999; 399:263–267. Friedlander RM, Gagliardini V, Rotello RJ, Yuan J. Functional role of interleukin1β (IL-1β) in IL-1β-converting enzyme–mediated apoptosis. J Exp Med 1997; 184: 717–724. Michel PP, Ruberg M, Agid Y. Rescue of mesencephalic dopamine neurons by anticancer drug cytosine arabinoside. J Neurochem 1997; 69:1499–1507. Lee SC, Dickson DW, Brosnan CF. Interleukine 1, nitric oxide and reactive astrocytes. Brain Behav Immun 1995; 9:345–354. Nishimoto I, Okamoto T, Giambarella U, Iwatsubo T. Apoptosis in neurodegenerative diseases. Adv Pharmacol 1997; 41:337–368. Gervais FG, Xu D, Robertson GS, Vaillancourt JP, Zhu Y, Huang JQ, LeBlanc A, Smith D, Rigby M, Shearman MS, Clarke EE, Zheng H, Van Der Ploeg LHT, Ruffolo SC, Thornberry NA, Xanthoudakis S, Zamboni RJ, Roy S, Nicholson DW. Involvement of caspases in proteolytic cleavage of Alzheimer’s amyloid-β precursor protein and amyloidogenic Aβ peptide formation. Cell 1999; 97:395–406. Feuerstein GZ. Apoptosis in cardiac diseases—new opportunities for novel therapeutics for heart diseases. Cardiovasc Drugs Ther 1999; 13:289–294. Adams JW, Sakata Y, Davis MG, Sah VP, Wang Y, Liggett SB, Chien KR, Brown JH, Dorn GW II. Enhanced Galphaq signaling: a common pathway mediates cardiac hypertrophy and apoptotic heart failure. Proc Natl Acad Sci USA 1998; 95:10140– 10145. Narula J, Pandey P, Arbustini E, Haider N, Narula N, Kolodgie FD, Dal Bello B, Semigran MJ, Bielsa-Masdeu A, Dec GW, Israels S, Ballester M, Virmani R, Saxena S, Kharbanda S. Apoptosis in heart failure: release of cytochrome c from mitochondria and activation of caspase-3 in human cardiomyopathy. Proc Natl Acad Sci USA 1999; 96:8144–8149. Van Parijs L, Abbas AK. Homeostasis and self-tolerance in the immune system: turning lymphocytes off. Science 1998; 280:243–248. Nagata S. Human autoimmune lymphoproliferative syndrome, a defect in the apoptosis-inducing Fas receptor: a lesson from the mouse model. J Hum Genet 1998; 43:2–8. Straus SE, Sneller M, Lenardo MJ, Puck JM, Strober W. An inherited disorder of lymphocyte apoptosis: the autoimmune lymphoproliferative syndrome. Ann Intern Med 1999; 130:591–601. Chervonsky AV, Wang Y, Wong FS, Visintin I, Flavell RA, Janeway CA Jr, Matis LA. The role of Fas in autoimmune diabetes. Cell 1997; 89:17–24. D’Souza SD, Bonetti B, Balasingam V, Cashman NR, Barker PA, Troutt AB, Raine CS, Antel JP. Multiple sclerosis: Fas signaling in oligodendrocyte cell death. J Exp Med 1996; 184:2361–2370. Giordano C, Stassi G, De Maria R, Todaro M, Richiusa P, Papoff G, Ruberti G, Bagnasco M, Testi R, Galluzzo A. Potential involvement of Fas and its ligand in the pathogenesis of Hashimoto’s thyroiditis. Science 1997; 275:960–963. Wyllie AH, Bellamy CO, Bubb VJ, Clarke AR, Corbet S, Curtis L, Harrison DJ, Hooper ML, Toft N, Webb S, Bird CC. Apoptosis and carcinogenesis. Br J Cancer 1999; 80 Suppl 1:34–37.

364

Zheng et al.

81. Canman CE, Kastan MB. Induction of apoptosis by tumor suppressor genes and oncogenes. Semin Cancer Biol 1995; 6:17–25. 82. Strasser A, Harris AW, Bath ML, Cory S. Novel primitive lymphoid tumours induced in transgenic mice by cooperation between myc and bcl-2. Nature 1990; 348: 331–3. 83. Harvey M, McArthur MJ, Montgomery CA Jr, Butel JS, Bradley A, Donehower LA. Spontaneous and carcinogen-induced tumorigenesis in p53-deficient mice. Nat Genet 1993; 5:225–229. 84. Hoffman B, Liebermann DA. Molecular controls of apoptosis: Differentiation/ growth arrest primary response genes, proto-oncogenes and tumor suppressor gene as positive and negative modulators. Oncogene 1994; 9:1807–1812. 85. Savitz SI, Rosenbaum DM. Apoptosis in neurological disease. Neurosurgery 1998; 42:555–572. 86. Jilka RL, Weinstein RS, Bellido T, Roberson P, Parfitt AM, Manolagas SC. Increased bone formation by prevention of osteoblast apoptosis with parathyroid hormone. J Clin Invest 1999; 104:439–446. 87. Moss JE, Aliprantis AO, Zychlinsky A. The regulation of apoptosis by microbial pathogens. Int Rev Cytol 1999; 187:203–259. 88. Vocero-Akbani AM, Heyden NV, Lissy NA, Ratner L, Dowdy SF. Killing HIVinfected cells by transduction with an HIV protease–activated caspase-3 protein. Nat Med 1999; 5:29–33. 89. Goldberg B, Stricker RB. Apoptosis and HIV infection: T-cells fiddle while the immune system burns. Immunol Lett 1999; 70:5–8. 90. Braun JS, Novak R, Herzog KH, Bodner SM, Cleveland JL, Tuomanen EI. Neuroprotection by a caspase inhibitor in acute bacterial meningitis. Nat Med 1999; 5: 298–302.

15 Fly Clocks RAVI ALLADA

MICHAEL ROSBASH

Northwestern University Evanston, Illinois

Brandeis University and Howard Hughes Medical Institute Waltham, Massachusetts

I.

Introduction

As a result of the combination of classic genetics, modern molecular biology, and a remarkably precise behavioral assay, studies in Drosophila have generated an explosion in our understanding of the basic molecular mechanisms of circadian rhythms. Furthermore, the identification of Drosophila clock genes has led to the isolation of the mammalian counterparts, including some from humans. It is probably only a matter of time before alterations in the structure or expression of these genes is intimately linked with diseases thought to be caused by clock defects. This chapter outlines the monumental contributions of forward genetics in mice as well as in flies to our understanding of the molecular basis of circadian rhythmicity. Circadian rhythms are classically defined by a few core properties. First, they proceed under constant environmental conditions. Thus, daily behaviors, such as the sleep–wake cycle, may proceed in the absence of exposure to the solar cycle and thus are driven by an endogenous mechanism. When placed in an underground cellar isolated from all environmental or temporal cues, experimental subjects continue to experience a daily sleep–wake cycle [1]. Similarly 365

366

Allada and Rosbash

Figure 1 Actogram or activity plot from wild-type and period mutant fruit flies. For wild-type flies on the top actogram, 2 days of data are plotted on each line. Each dark horizontal bar reflects locomotor activity. Darker areas with more bars indicate times of higher activity; lighter areas with fewer bars indicate times of lower activity. Transitions from high activity to low activity, called activity offsets, are prominent. In constant darkness (DD) conditions, these offsets occur slightly later and later each subjective day, indicating a periodicity of slightly greater than 24 hr. Under 12-hr light–12-hr dark entrainment (LD), these offsets synchronize to the 24-hr LD cycle. In the actogram of per L mutant and per S mutant, activity offset occurs later and later for per L and earlier and earlier for per S in DD conditions, reflecting their internal periodicities. The actogram of per O mutant reveals no apparent rhythmicity.

Fly Clocks

367

in the fruit fly (Fig. 1), a circadian rhythm of locomotor activity persists under constant darkness (DD) conditions. Second, rhythmic phenomena display a periodicity of approximately, but not exactly, 24 hr; hence the term circadian (L. circa ⫽ approximate, dia ⫽ day). As a consequence, humans generally, but not exclusively, exhibit a sleep–wake rhythm which slightly exceeds 24 hr when isolated from temporal cues. These individuals will gradually go to sleep and wake up later and later with respect to external objective time. In fruit flies (Fig. 1), bouts of high activity also drift later and later with respect to external time, reflecting a slightly long-period clock. This consistent imprecision of clocks may contribute to the difficulty many people experience crawling out of bed on Monday morning after allowing themselves to follow their long-period clocks over the weekend. Third, environmental signals are capable of resetting the clock. Light is probably the most dominant, and internal clocks are synchronized by the 24-hr solar cycle. As a result, most individuals generally wake up and go to sleep at about the same time every day. If fruit flies are maintained in 12-hr light–12-hr dark cycles (LD; see Fig. 1), their locomotor activity rhythm becomes synchronized to this external cycle. In addition, activity changes precede the environmental transitions, reflecting an anticipatory role of the clock. Keeping time rather than simply responding to external stimuli is crucial to fitness and survival throughout biology, as comparable clocks have been identified even in photosynthetic cyanobacteria. Finally, the periodicity of clocks is largely invariant over a physiological range of temperatures. In stark contrast, the behavior of most proteins are dramatically influenced by temperature. This temperature compensation may be considered a necessary adaptation for poikilotherms to prevent ambient temperature changes from causing dramatic shifts in the phase of internal clocks. However, cultured rhythmic tissues from homeotherms are also temperature compensated [2]. Temperature compensation, therefore, appears to be a fundamental clock property; perhaps reflecting the benefits of a robust homeostatic system to defend against many environmental perturbations.

II. Circadian Rhythms in Human Health and Disease Given the ubiquitous nature of circadian clocks, circadian biologists are puzzled that the role of rhythms in human health and disease appears to be underappreciated. Circadian clocks greatly influence the rates of myocardial infarction, which occurs far more commonly in the morning than at other times of the day [3]. This predictable temporal pattern of variation has implications for the dosing of cardiac medications [4]. Blood pressure also varies significantly with the time

368

Allada and Rosbash

of day, which has implications for the clinical diagnosis of hypertension [5]. Finally, airway pressure in asthmatics appears to increase significantly during the night, leading to the institution of nocturnal dosing of antiasthmatic medications [6]. In addition to the circadian influence on diseases, certain disorders, called circadian dysrhythmias, are thought to be due to primary defects in the central clock mechanism. The most obvious manifestations of a disturbed circadian clock are sleep disorders. People with delayed sleep phase syndrome (DSPS) have great difficulty going to sleep at night and difficulty waking in the morning. People with advance sleep phase syndrome (ASPS) have the opposite problem; that is, difficulty staying awake in the evening with consequent early morning awakening. These patients presumably manifest long-period clock and short-period endogenous clocks, respectively. Recently, families have been identified with an inherited form of ASPS [7]. Jet lag results from the limited ability of circadian clocks to reset rapidly to new external time. Rapid jet travel across several time zones and the consequent desynchronization of internal clocks from external time results in the pathology of this disorder. Furthermore, the typically long (⬎24 hr) endogenous human periodicity explains why most people experience less jet lag when traveling westward. In this direction, we have to lengthen our internal period to adapt to the new environmental cycle. Traveling eastward requires a shortening of our already ⬎24-hr period and is therefore more difficult. This subjective impression of jet travelers has been confirmed more objectively by analyzing retrospectively the performance of professional baseball teams as a function of their direction of travel [8]. A less obvious and somewhat more controversial notion is that clock defects lead to pathological mood states, such as bipolar disorder and seasonal affective disorder. Seasonality, or changes in behavior with the season, is thought to be mediated in part by circadian clocks; they are capable of telling the season by sensing changes in day length. In many people, extreme seasonality can lead to ‘‘winter depression’’ or seasonal affective disorder. In fact, the therapeutic efficacy of light for winter depression is correlated with its ability to advance the clock [9]. In addition, bipolar disorder, a far more common disease, may also be due to disrupted clocks. It is intriguing in this context that the mood stabilizer lithium (whose mechanism of action remains unknown) also has clock-resetting properties at therapeutic doses [10].

III. A Mammalian Clock: The Suprachiasmatic Nucleus Much of the circadian work during the last three decades has been devoted to the identification and analysis of the anatomical substrate of the mammalian

Fly Clocks

369

clock, the suprachiasmatic nucleus (SCN). This pacemaker locus consists of a large collection of hypothalamic neurons (approximately 10,000) located just above the optic chiasm. This region was first identified as a direct target of retinal projections and its function was confirmed in lesion experiments [11,12]. An elegant series of lesion and transplantation rescue experiments solidified the role of the SCN not only in being required for rhythmicity but also in being sufficient to determine rhythmic properties. One of the key experiments resulted from the serendipitous discovery of a spontaneously mutant short-period hamster named tau [13]. In lesion/transplantation experiments, the period of the rescued animal reflected the genotype of the donor tissue, virtually proving that the SCN was the seat of the mammalian clock [14]. The discovery of tau marked the beginning of clock genetics in mammals. However, the absence of genetic markers in the hamster has hindered cloning efforts. The SCN not only mediates rhythmic behaviors but also manifests rhythmic physiological properties. For example, this group of neurons exhibits 24-hr metabolic rhythms in intact animals as measured by 2-deoxyglucose uptake [15]. In dissociated cell culture, SCN neurons exhibit circadian regulation of spontaneous activity even when electrically isolated. Rhythmicity is therefore a property of individual SCN neurons [16]. Despite the work of many SCN biologists, little was known about the molecular underpinnings of these intracellular rhythms.

IV. Circadian Rhythms Are Influenced by Genes: Behavioral Genetics and the period Mutants The application of classic genetics in model organisms such as Drosophila led to the first major insights into the molecular basis of circadian rhythms. During the late 1960s, Konopka, a graduate student in Benzer’s laboratory at the California Institute of Technology, initiated genetic screens in Drosophila to find mutants with altered or missing rhythms [17]. It had been observed by fruit fly biologists, who routinely select young virgin female flies for mating crosses, that adult flies tend to emerge from their pupal case at dawn. This rhythm persists even under constant dark conditions, and is thus circadian [18]. Like pioneers earlier in the century who had screened mutagenized flies for altered eye colors or other morphological phenotypes, Konopka and Benzer devised a simple screen to identify adults that emerged or eclosed at the ‘‘wrong’’ time; that is, at times when wild-type flies do not normally emerge. They identified three mutant strains (see Fig. 1): One had a long period of 29 hr under constant darkness conditions, a second had a short period of 19 hr, and a third had no measurable rhythmic features. All three mutant phenotypes mapped to a single locus on the X chromosome, termed period (per). Interestingly, these mutants did not have any other obvious defect in behavior and physiology. Therefore, the period gene appeared

370

Allada and Rosbash

to be dedicated to clock function. The success of this very simple approach was due in part to the precision of the eclosion assay, which could easily detect the difference between strains with 24- or 29-hr periods (approximately 20% difference). Few behavioral assays have this magnitude of precision. Second, the short generation time (10 days) as well as abundant genetic tools in Drosophila made it an ideal organism for a classic genetic attack on this enigmatic behavior. Subsequent genetic screens have assayed the circadian locomotor activity of individual flies. This assay has since been miniaturized and automated for high-throughput analysis. Single flies are placed into small transparent glass tubes and inserted into monitors, which have infrared beams perpendicular to the length of the tube. The infrared light is not visible to the adult fly to prevent a light influence on clock functions. Each time the fly crosses the infrared beam, one activity unit is measured. Activity levels as a function of time are then analyzed to produce a period measurement.

V.

Circadian Clocks Consist of Circadian Transcriptional Feedback Loops: The period Gene

Prior to the era of clock gene cloning, circadian rhythms were largely viewed as an emergent property of neuronal populations [19]. The cloning and analysis of the period gene reformulated the circadian clock as an intracellular transcriptional feedback loop at its center. Konopka and Benzer were clearly ahead of their time, as the requisite molecular tools to clone the period gene did not emerge until the early 1980s. The cloning and sequencing of per revealed that the period-altering per mutants contained missense mutations only. The arrhythmic per allele contained a premature stop codon, a gratifying result considering the clockless phenotype [20,21]. However, neither the cloning nor the sequencing of per led to any immediate functional insight, as it was a pioneer protein with no obvious relatives. During these early days, it was proposed that period may be a cell surface molecule; consistent with notions of coupled oscillator models [22,23]. But subsequent noncircadian work identified a substantial homology with the basic helix–loop–helix transcription factor, single-minded (sim). However, per itself does not contain any canonical DNA-binding domains, which made the relationship with transcription factors less than certain [24]. This confusion was resolved by the observations that (1) per RNA cycles with a 24-hr period, (2) missense mutants which cause short- and long-period phenotypes alter in a parallel manner the period of the RNA cycling, and (3) the arrhythmic per allele abolishes per RNA cycling [25]. In addition, transient induction of per results in stable shifts in the phase of the clock [26]. Taken together, these observations indicate that per protein regulates the levels of its own RNA.

Fly Clocks

371

This feedback appears to be largely transcriptional. Genomic regions 5′ to the per coding region, the putative per promoter, mediate cycling of a reporter RNA, indicating that per RNA oscillations are at least in part mediated by transcriptional control [27]. In addition, PER levels, its phosphorylation state and nuclear localization all undergo rhythmic changes [28–31]. The general view is that per transcription rises during the day leading to rises in per RNA. PER then undergoes a set of posttranscriptional events, which results in translocation into the nucleus and inhibition of its own transcription (Fig. 2). This idea of a perbased circadian transcriptional feedback loop is the central organizing principle for the field of molecular rhythms. A large fraction of research during the last decade in flies and more recently in mammals has been devoted toward advancing our understanding of this molecular cycle.

VI. Molecular Mechanisms of Timekeeping Delays in the Feedback Loop: The timeless Gene The cloning of the second circadian rhythm gene in Drosophila, the timeless (tim) gene was not reported until a full decade after per; a testament to the difficulty of the problem. Nonetheless, tim was initially identified by classic behavioral genetics as a recessive arrhythmic mutant [32]. To this day, timeless remains somewhat enigmatic owing to a lack of substantial sequence similarity with any well-studied gene family. However, timeless RNA and protein cycle with 24-hr periods like per. Arrhythmic tim mutants abolish both per and tim RNA cycling, and arrhythmic per mutants abolish RNA cycling from both genes, indicating reciprocal feedback regulation [32,33]. Taken together with similar observations from other organisms, rhythmic expression and feedback regulation appear to be general features of clock genes. To generate a free-running oscillation, there must be one or more delays between the synthesis of RNA and proteins and feedback onto transcription of genes. Delays are crucial to keeping time and are distinguishing characteristics of circadian feedback systems. In the absence of delays, a feedback system does not oscillate and ultimately comes to a static equilibrium. Therefore, a major question for circadian biologists is what are the molecular mechanisms that delay period protein’s negative feedback on its own transcription. The positional cloning of tim, allowed the first molecular forays into the delay problem. Initial work on PER focused on a region of sequence similarity with Drosophila singleminded (sim) and the aryl hydrocarbon nuclear translocator (ARNT), known as the PAS (per-ARNT-sim) domain. In fact, the PAS domain mediates homotypic and heterotypic PAS–PAS interactions in vitro and is capable of coimmunoprecipitating PAS proteins from fly head extracts, indicating a potential role in protein dimerization [34,35]. The in vivo relevance of the PAS domain to period

372

Allada and Rosbash

Figure 2 The fly circadian clock. The circadian rhythm genes period (per) and timeless (tim) are rhythmically transcribed by the CLOCK (CLK) and CYCLE (CYC) bHLH-PAS transcription factors. CLK and CYC associate with each other and bind to rhythmically regulated E-boxes (CACGTG) in the per and tim promoters. Q, polyglutamine tracts thought to be important for activation function and deleted in the Jrk allele of Clock. PER and TIM proteins are able to physically associate with each other, become progressively phosphorylated and translocate to the nucleus to inhibit activation by the CLK–CYC complex. PER and TIM proteins become degraded, thus reinitiating a cycle. DOUBLETIME (DBT), a kinase, appears to play a crucial role in PER phosphorylation and degradation. Light is perceived by CRYPTOCHROME (CRY) which associates with TIM, resulting in degradation of both proteins. In the early night (ZT 15), degradation of TIM slows TIM accumulation, thus delaying the clock. ZT, Zeitgeber time where experiment is done under 12-hr light: 12-hr dark conditions. ZT 0, lights on and ZT 12, lights off. In the late night (ZT 21), degradation of TIM speeds TIM degradation, thus advancing the clock.

Fly Clocks

373

function is buttressed by the identification of a missense mutation in the PAS domain of a period-lengthening allele of per (per L ). Therefore, a (PAS-containing?) PER-interacting factor may impose a delay on PER-mediated feedback. timeless not only behaves like period in its temporal pattern of expression, but it also appears to be intimately intertwined with period both genetically and biochemically. In addition to its effects on per RNA, tim is also required for nuclear localization of PER protein [36]. Coexpression of PER and TIM in Drosophila tissue culture cells results in the mutually dependent nuclear localization of these two proteins [37]. They interact strongly in the yeast two-hybrid assay, and their contact is seriously disrupted by the per L PAS mutation [38]. These in vitro interactions are supported by both genetic and biochemical experiments in vivo. An allele-specific suppressor of per L has been identified in tim, and although TIM does not contain a PAS domain, it represents the most potent in vivo partner of PER in coimmunoprecipitation experiments [39,40]. Thus, non-PAS proteins can directly interact with PAS-containing proteins. A required interaction between PER and TIM for nuclear entry may impose an additional temporal delay between the appearance of PER and transcriptional feedback.

VII. How Do Molecular Clocks Become Synchronized to External Light–Dark Cycles? Circadian clocks do not simply free-run in contrived constant darkness conditions but must function in natural light–dark environments. In addition, they must synchronize or entrain to the external environment. How does the clock, that is, the PER–TIM cycle, become synchronized to the light–dark cycle? A significant step forward came with the observation that TIM protein levels, unlike those of PER, are suppressed within minutes of exposure to light [40–43]. Therefore, reduction in TIM protein levels may explain in part how the molecular cycle adjusts to changes in the external light regimen. In addition, the TIM light response appears to explain the clock response to brief light pulses (see Fig. 2). For decades circadian biologists used the resetting of the clock phase by short light pulses (minutes) as an experimental paradigm to assess both the state of the clock and its photoreceptive pathways [44]. The administration of light pulses at different times of day (in otherwise constant darkness) causes a phase advance during late night, phase delays during early night, and no changes during subjective day. This phenomenon correlates well with TIM protein levels and their response to light. When TIM levels are falling during the advance zone, a light pulse would hasten the disappearance of TIM and advance the clock. When TIM levels are rising during the delay zone, a light pulse would slow the appearance of TIM and delay the clock. Finally, when TIM levels are very low, light pulses

374

Allada and Rosbash

cause little or no phase shifts. Thus, TIM appears to play a role in the running of the clock as well as in responding to light. VIII. Did Nature Create Different Clocks for Different Organisms?: The Search for Mammalian Clock Gene Homologs One of the most striking observations from the last two decades of molecular biology is the high degree of evolutionary conservation in gene structure and function even between unicellular organisms and humans. Consider the Drosophila eyeless transcription factor, which is required for proper eye development [45]. Remarkably, ectopic expression of eyeless leads to the formation of ectopic eyes. Mutations in the human homolog Pax6 lead to aniridia, the absence of the iris muscle which controls pupil size [46]. Despite the obvious anatomical differences between the structure of the Drosophila and human eyes, the ectopic overexpression of the human homolog in flies also results in the ectopic development of Drosophila eyes [47]. Therefore, eyeless and Pax6 not only share structural similarity at the sequence level and parallel functional roles in visual organ development, these functions can be transferred between these divergent species. Given this and other similar observations, it was widely believed that there should be a mammalian ortholog of per and that its identification would serve as a molecular entree into the study of mammalian circadian rhythms. However, despite attempts by several laboratories, no such molecules were convincingly identified. Perhaps early developmental pathways or housekeeping (e.g., basal transcription or splicing) functions are conserved, but complex brain functions such as circadian rhythms may have arisen independently between the simple fly and complex mammals. IX. Konopka and Benzer in the Mouse: Mammalian Forward Genetics Given the dearth of progress in the cloning of mammalian homologs, a forward behavioral genetic approach in mice was initiated. Inspired by the work of Konopka and Benzer, mutagenized populations of mice were screened for circadian rhythm phenotypes as assayed by rhythmic wheel-running behavior. Such a screen had obvious technical obstacles to overcome; primarily the relatively poor state of the mouse as a useful organism for forward genetics. In addition to the long generation times and maintenance costs, the density of genetic markers, especially in relation to Drosophila, made positional cloning of induced single– base pair mutations especially difficult. Despite these challenges, there were many arguments in favor of screening for rhythmic phenotypes in mice. The

Fly Clocks

375

wheel-running assay is very precise, facilitating the detection of subtle phenotypes. Furthermore, it had been observed that nearly every circadian rhythm mutant in Drosophila exhibited a semidominant phenotype [17]. Assuming a similar situation in the mouse, an F1 screen for dominant phenotypes would eliminate the need for additional backcrosses to homozygose-induced mutations. Finally, the advance of the Human Genome Project and the advent of transgenic technologies in mammals facilitated the identification and functional analysis of candidate clones once mutant phenotypes were finely mapped. These advantages proved instrumental in the identification of a mutant, Clock. Heterozygotes displayed a slightly long period and homozygotes had even longer periods grading into arrhythmicity [48]. A variety of defects in general physiology could lead to arrhythmic behavior, but the slight period alterations of the heterozygote suggested that this mutation may have disrupted a central timekeeping component [49]. The subsequent effort to map and clone Clock revealed that it contained a PAS domain, placing it in the same family as per [50,51]. However, Clock also contained a basic helix–loop–helix domain (bHLH), previously shown in other DNA-binding transcription factors to play a role in direct protein–DNA contacts and protein–protein dimerization. As expected, Clock RNA is expressed in the SCN, but its levels do not oscillate with circadian time [50]. Although Clock was not a true ortholog of per, their PAS domains indicated family ties and raised the possibility that the fly and mammalian circadian systems were more well-conserved than suggested by the failure of previous cloning efforts.

X.

Nature Does Not Create New Ways of Making Clocks: The Genome Projects and the Cloning of New Evolutionarily Conserved period Genes

More than 10 years after the cloning of Drosophila per, two independent groups reported the cloning of a putative human ortholog (hPer). Using a series of degenerate oligonucleotides complementary to the PAS domain of Drosophila per (dper), hPer was isolated by polymerase chain reaction (PCR) [52]. As part of the Human Genome Project, another group discovered hPer while sequencing cDNAs from chromosome 17 [53]. Several previously identified genes had PAS domains and therefore limited homology to Drosophila per. However, this putative human per had homology with dper thoughout its full length [52,53]. The analysis of mouse homologs revealed that mouse per (mPer) not only is expressed in the SCN, but its transcript also oscillates with a 24-hr period in constant darkness [52,53]. Interestingly, mPer cycling in this nocturnal (night-active) animal is antiphase to per cycling in the diurnal (day-active) Drosophila. Furthermore, changes in the lighting regimen which shifted the behavioral circadian rhythm

376

Allada and Rosbash

similarly shifted the RNA cycle phase [52,53]. Thus, mammalian per not only had structural similarity with Drosophila per, but also executed similar functions based on its expression profile. Genome database searches identified two additional mammalian per homologs, mPer2 and mPer3, both of which cycle in phase with mPer1 in the SCN [54–56]. Further study distinguished the three mammalian per homologs from fly per and from each other. The most notable difference is that mPer1 and mPer2, but not mPer3 or dper, are rapidly induced by short light pulses [54,57,58]. However, this response is circadianly gated. As mentioned above, the ability of light to phase shift or reset circadian clocks is dependent on the circadian time at which the light pulse is administered. In other words, the circadian clock itself regulates the efficacy of light-induced resetting. Therefore, per induction by light correlates well with the ability of light to phase shift the clock, implicating this molecular response in light-mediated phase shifts. However, there is currently no genetic evidence which demonstrates a requirement of per induction for light-mediated clock resetting. Recent genetic evidence suggests that an autoregulatory feedback loop underlies the mammalian clock. Deletion of a portion of the PAS domain of mPer2 in vivo results in short periods grading into arrhythmicity, indicating a functional role in the central pacemaker [59]. In addition, the cycling levels and amplitude of mPer1 and mPer2 RNAs is substantially reduced, suggesting the presence of feedback regulation. Thus, the notion of feedback loops, established in Drosophila, can be extended to mammals.

XI. Getting Cyc’d about Clocks: Flies and Mammals Share Circadian Transcription Factors Given the penchant of transcription factors to heterodimerize, it was thought that the period PAS domain may represent an interface for protein–protein contacts with a transcriptional activator. Unlike the other PAS family members, the period gene does not contain a canonical DNA-binding basic helix–loop–helix domain. Thus, the heterodimerization model remained speculative. Studies of the Drosophila period promoter identified a 69–base pair circadian enhancer which requires the presence of an E-box (CACGTG) to confer high-amplitude cycling to a reporter gene. Such E-boxes are known binding sites of bHLH-PAS transcription factors [60]. Thus, it was proposed that the PER-PAS domain may serve as a protein–protein contact to inhibit the activity of an undefined bHLH-PAS transcription factor. Concurrent with the mapping and cloning of Clock, rhythm mutagenesis screens in flies identified three novel complementation groups distinct from per and tim. One group was initially defined by a mutant named Jrk. This semidominant mutant displayed a slight period alteration with substantial arrhyth-

Fly Clocks

377

micity as a heterozygote. Jrk flies were completely arrhythmic as homozygotes [61]. In addition, per and tim transcription, RNA levels, and protein levels were pegged at trough levels in homozygous Jrk flies [61]. A second group, called cycle, shared many similarities with Jrk. As a homozygote, cycle also abolished locomotor rhythms and severely reduced per and tim transcription, RNA, and protein levels [62]. Given their phenotype, the two genes appeared to encode important positive components for the transcripton of clock genes, perhaps even bHLH-PAS transcription factors. A search of the ever-expanding expressed sequence tag (EST) database for new members of the Drosophila bHLH-PAS fam-

Figure 3 The mammalian circadian clock. CLOCK and the mammalian homolog of BMAL1 activate transcription from E-boxes. The per, cry, and tim (in part) homologs are rhythmically expressed. Line with AAAAAA indicates polyA⫹ mRNA. Various combinations of PER, TIM, and CRY homologs can feedback and inhibit transcription of the CLOCK–BMAL complex. CRY appears to be the most potent inhibitor.

378

Allada and Rosbash

ily identified two novel members: one, Drosophila Clock (dClk), displayed striking homology with mouse Clock and chromosomally colocalized with Jrk. The sequencing of dClk in Jrk identified a premature stop codon removing much of the glutamine-rich C terminus presumably involved in transcriptional activation [61]. This truncated form is predicted to behave as a dominant negative transcription factor, which is consistent with the dominant phenotype observed in flies. The other novel bHLH-PAS gene showed striking homology with the mammalian gene BMAL1 and colocalized with cycle. Sequencing also revealed a premature stop codon in cycle mutant flies, which is consistent with its null phenotype. A search for the partner of mouse CLOCK using a two hybrid approach also identified BMAL1 [63]. Furthermore, these proteins interact and activate from E-boxes in the mPer1 promoter. Drosophila Clock and cycle were also cloned by homology and were shown to interact with each other, bind to functionally relevant E boxes (CACGTG) in the per and tim promoters, and transcriptionally activate per and tim [64]. Thus, the CLK : CYC heterodimer appears to be the positive transcription factor that transcribes per and tim. In addition, coexpression of per and tim modestly inhibit (about two- to threefold) the activity of the CLK : CYC complex in cell culture [64]. Consistent with feedback inhibition, PER and TIM proteins coimmunoprecipitate with dCLK in fly head extracts at times of falling transcription rates [65]. In fact, a mammalian tim homolog has also been identified and may also participate with PER homologs to feedback on CLOCK-BMAL1–mediated transcription (Fig. 3) [66,67]. These complementary studies in flies and mice revealed the central transcriptional components of the circadian feedback loop and solidified the evolutionary connection between flies and mice. They also demonstrate the growing synergy between classic genetic methods and genome projects, which appear to have replaced to some extent the traditional role of individual molecular biologists.

XII. The Postranslational Clock: The doubletime Kinase Although the current view of circadian clocks has focused on the role of transcription, there is accumulating evidence that posttranscriptional mechanisms also contribute and may even be the predominant means of maintaining self-sustaining oscillations. They also may contribute to the delays that are so important to the molecular feedback cycle. In Drosophila, promoterless per transgenes are capable of rescuing both RNA cycling and behavioral rhythms in a null per genetic background [68]. Nuclear run-on analysis in these flies clearly demonstrated an absence of transcriptional cycling, implying a reliance on posttranscriptional regulation [69]. In addition, there is direct evidence that posttranslational control regulation is also important. For example, antibodies to PER and TIM proteins

Fly Clocks

379

reveal that they undergo progressive changes in mobility with respect to time of day [30,40–43]. These mobility changes appear to be due to changes in the phosphorylation state, as they are reversed by phosphatase treatment. The time of the lowest mobility (most phosphorylation) immediately precedes the disappearance of the protein, suggesting that phosphorylation may be a signal for activation of the proteolytic machinery. Furthermore, short- and long-period alleles of per appear to affect the magnitude and timing of these phosphorylation changes, suggesting, but certainly not proving, that phosphorylation may be a functional element of the clock [30]. Similar rhythmic changes in the phosphorylation state have been observed in the product of the Neurospora circadian rhythm gene, frequency [70]. However, genetic evidence demonstrating the importance of these phosphorylation changes in PER was lacking until the identification of the relevant kinase, the doubletime gene. Like per, tim, Clock, and cyc, doubletime was originally identified in behavioral screens [71]. Short-period, arrhythmic, and long-period alleles as well as a homozygous lethal allele were identified. In fact, the lethal doubletime allele remains the only clock gene mutant with a developmental phenotype, suggesting that there may be more circadian gold to be mined among mutants with more pleiotropic phenotypes. The doubletime (dbt) gene encodes a serine–threonine kinase highly homologous with the mammalian casein kinase I-epsilon [72]. Short- and long-period alleles shift both per and tim RNA and protein profiles, including their phosphorylation states [71]. In the null homozygous lethal mutant, PER levels were examined in homozygous viable larvae (only the adults are dead). High PER levels in these larvae indicate that doubletime may regulate PER stability, which is consistent with previous correlations between phosphorylation state and protein disappearance [71]. In addition, doubletime protein (DBT) physically associates with PER in cotransfection experiments [72]. Thus, DBT appears to be involved in PER phosphorylation, which may be required for PER degradation. Although genetic and biochemical data circumstantially link dbt with per, there is no direct evidence that per is a phosphorylation target of doubletime. Nonetheless, forward genetics in the fruit fly and the subsequent cloning of doubletime has provided the first molecular handle on the role of phosphorylation in the circadian clock. As day follows night, the cloning of this fly rhythm gene will be followed by knockout of the mammalian counterpart to address its potential role in rhythms.

XIII. A Clockwork Blue: The Blue Light–Sensitive cryptochrome Genes and the Eyes of the Clock Mammals and insects appear to utilize physiologically distinct photoreceptive pathways for circadian entrainment and phase shifting. In mammals, eyes are

380

Allada and Rosbash

essential, but circadian photoreception does not appear to require the presence of retinal rods or cone photoreceptor cells [73,74]. However, the action spectrum of phase-shifting light correlates well with those of vitamin A–based opsin photoreceptors [75]. Thus, it has been proposed that an unconventional opsin may mediate the phase-shifting effects of light. These action spectra, however, do not exclude a role for multiple photoreceptors, or multiple types of photoreceptors, that is, nonopsins, which in combination produce this action spectrum. In fact, it may even be the case that rod and cone–based photoreception is sufficient for phase shifting, but its lack of requirement may reflect functional redundancy in the light input pathway. Differences between the action spectrum of rodless and coneless mice and their wild-type counterparts may reveal the relative roles and photopigments in each compartment. Indeed, recent data suggest that nonopsins participate in phase shifting. For example, light applied to the back of the knee of human subjects can cause phase shifts [76]. The authors proposed a hemebased mechanism. However, these striking results have not been reproduced in humans and other mammals and therefore remain controversial [77]. In addition, a class of blue light photoreceptors, called cryptochromes, have been implicated as potential photoreceptors. Although lacking repair activity, cryptochromes are homologous to blue light–sensitive DNA repair enzymes [78]. Nonetheless, they mediate blue light–sensitive photoresponses in plants, such as phototropism [79]. In mammals, cryptochromes are found both in the SCN and in the retinal ganglion and inner nuclear cell layers; the latter are parts of the retina remaining in rodless, coneless mice [80]. Intriguingly, the expression of one of these genes, mCry1, is under circadian control in the SCN [80]. Given their potential function and expression pattern, it was proposed that cryptochromes may be responsible for light-mediated phase shifting. In vivo deletion of the mCry2 gene results in slightly lengthened periods, and, in contrast to prediction, increased photosensitivity [81]. Nonetheless, these mice do display a reduced light induction of mPer1 RNA [81]. In addition, the double knockout of both mCry1 and mCry2 is completely arrhythmic, suggesting a role in the central pacemaker [82]. Consistent with these genetic data, both mCRY1 and mCRY2 are able to repress CLOCKBMAL1–mediated transcriptional activation strongly when cotransfected in cultured cell lines [83]. The current data therefore support a role of mCry in the transcriptional feedback loop itself, but do not rule out the possibility that Cry may also function in photoreception. For example, light induction of per RNAs might be mediated by cryptochromes. Thus, bona fide mammalian circadian photoreceptors remain unknown and an object of current investigations. On this issue of circadian photoreception, the mammalian picture is blurred but the fruit fly appears much clearer. In Drosophila, a difference in the phaseshifting action spectrum with mammals suggested the presence of a non–retinalbased photopigment [84,85]. These beliefs were buttressed by experiments demonstrating that depletion of vitamin A from flies did not abolish their phaseshift response, implicating a photoreceptor more blue shifted than mammalian

Fly Clocks

381

photoreceptors. Using a novel luciferase-based screen, a mutation in a Drosophila cryptochrome gene (dcry b ) was identified which displayed impaired entrainment and abolished phase responses to light pulses [86]. Furthermore, overexpression of cry resulted in increased circadian photosensitivity [87]. Similar to plant cryptochromes, dCRY protein is degraded within minutes in response to exposing flies to light (87). In apparent contrast to mammals, these alterations in cry function had no apparent effect on free-running periodicity; placing cry within the input pathway and firmly outside the central pacemaker. Despite crippled cry function, dcry b flies are still able to entrain to 12-hr light, 12-hr dark cycles. This entrainment can be impaired but not completely abolished by mutations in a phospholipase C protein involved in the visual (opsin)–based phototransduction cascade, indicating a second phototransduction pathway [86]. Therefore, as in mammals, vitamin A–based photoreception may also be involved in Drosophila circadian photoreception. The remaining entrainment capability implies that there may even be a third circadian phototransduction pathway. Finally, dcry itself seems to be under robust circadian control at the RNA level, but such control is low or absent at the protein level [87]. RNA cycling may reflect the importance of newly synthesized dCRY protein in phase shifting. Alternatively, the peak of RNA during the morning may ensure sufficient dCRY protein during the daylight hours when it is degraded. RNA cycling may even be an epiphenomenon, a reflection of common transcription factors which contribute both to expression in circadian tissues and to circadian regulation itself. CRY may also have light-independent functions on circadian pacemaker gene macrometabolism in flies, not unlike the mammalian model [86]. Despite an absence of period differences between flies with and without dcry, there are light-independent alterations in per and tim expression under temperature cycling constant darkness conditions. Indeed, dCRY directly interacts with TIM in both cultured cells and in yeast two-hybrid assay, although the latter is dramatically dependent on light exposure [88]. Therefore, dCRY appears to be biochemically intimate with pacemaker components but not required for their function, at least under laboratory conditions. The intimate association with TIM may merely reflect a role in phase shifting. In sum, forward genetics has identified a strong fruit fly candidate for a circadian photoreceptor. Further studies will be required to see if the close ties between flies and mammals in the structure and function of clock components breaks down in the cryptochrome family.

XIV. Clocks Are Everywhere! Since the discovery of the suprachiasmatic nucleus, lesion and transplantation experiments have supported a model in which the circadian kingdom is a dictatorship: the SCN directing otherwise passive peripheral tissues. The more recent

382

Allada and Rosbash

identification of clock genes and the analysis of their spatial and temporal expression patterns reveals that the role of the SCN is more akin to the conductor of an orchestra: keeping the far-flung organs in synchrony with each other. First, the expression of clock genes in Drosophila is not restricted to a handful of clockrelated neurons [89]. Several nonneural tissues also rhythmically express clock genes, suggesting circadianly related functions [90]. Flies containing per promoter transgenes fused to a luciferase reporter (per-luc) exhibit cycling bioluminescence when fed on luciferin-containing media [91]. Surprisingly, peripheral tissues from these per-luc flies autonomously cycle in culture. In fact, these cultured oscillators autonomously entrain to light–dark cycles, indicating the presence of independent photoreceptive pathways. Some of these concepts extend to mammals. Like the fly rhythm genes, mClock is transcribed in several nonneural tissues [50]. In addition, rat per mRNAs cycle in these peripheral tissues like the heart and lungs. Oscillatory molecular cycling is therefore not exclusive to the SCN [92]. In fact, circadian oscillation of the per genes is evident in a rat-1 fibroblast cell line after serum shock [93]. This immortalized cell line has been maintained in culture for over 25 years, yet retains much of the circadian machinery. Thus, fruit fly genetics has led to a reformulation of our view not only of molecular mechanisms but also of organism-level organization of circadian systems.

XV. Clocks and Outputs How do core pacemaker elements biochemically link with output behaviors, such as the sleep–wake cycle? Several putative genes in circadian output pathways have been identified based on the cycling of their expression at the RNA level. In fact, virtually the entire genome is under circadian control in photosynthetic cyanobacteria [94]. In Drosophila, random sampling of transcripts suggests a much smaller but substantial fraction of cycling transcripts, probably between 1% and 5% [95]. Cycling at the RNA level has implicated transcription factors in mediating output. As Clock encodes the first bona fide transcription factor identified in the circadian pacemaker, it has been proposed to mediate the cycling of these putative output RNAs in both flies and mammals. In mammals, the mRNA from the vasopressin gene is circadianly regulated [96]. The vasopressin promoter contains a putative CLOCK E-box binding site, CACGTG [97]. In homozygous Clock mutant mice, vasopressin mRNA levels are reduced. Furthermore, in tissue culture experiments, CLOCK activates vasopressin gene expression in an E-box–dependent manner. These data are consistent with the hypothesis that CLOCK directly binds to E-boxes in the vasopressin promoter and mediates its rhythmic transcription. However, the sufficiency or requirement of this promoter E-box has not been demonstrated in vivo, and the tissue culture

Fly Clocks

383

experiments do not prove that the activation of CLOCK on the vasopressin promoter is biochemically direct. In addition to transcriptional mechanisms, posttranscriptional mechanisms may be important for mediating rhythmic output gene expression. One Drosophila output gene, lark, appears to be circadianly regulated by a posttranscriptional mechanism. lark Was initially identified as a mutant with a late phase of eclosion [98]. However, the absence of a circadian locomotor aberration suggested that the central pacemaker is unaffected in these mutants, placing lark on an output pathway. Unlike other output genes, however, LARK protein does cycle but lark RNA does not [99]. Therefore, posttranscriptional mechanisms must underly this circadian cycling. In fact, lark itself encodes an RNA-binding protein, suggesting that a downstream RNA targets in the eclosion output pathway may be nontranscriptionally regulated [100]. One such candidate is the gene crg-1 whose mRNA cycles in phase with per and tim [101]. However, at the level of transcription, it does not cycle, suggesting that the mRNA cycling is due to changes in stability not synthesis [69]. It remains to be seen if crg-1 is a direct target of LARK. A theoretical concern with the transcriptional output model is that high amplitude transcriptional cycling may be attenuated by the stability of output mRNA and protein products. Most reporter genes that demonstrate RNA cycling when fused to circadian promoters do not demonstrate cycling at the protein level, reflecting protein stability [102]. Thus, posttranscriptional and even posttranslational mechanisms may be the best way to ensure that pacemaker oscillations are transferred out of the clock. These putative posttranslational changes in output have yet to be identified. However, mutations in protein kinase A abolish locomotor rhythms but not eclosion rhythms, implicating phosphorylation changes in a subset of output pathways [103]. What is the function of cycling gene expression in mammalian peripheral tissues? There are a variety of clock-driven physiological and pathological processes. It remains to be seen, of course, how pacemaker proteins link to diseaserelated pathways in these tissues. The identification of pacemaker genes in flies and mammals as well as their downstream targets should further the understanding of clock-controlled processes, such as the sleep–wake cycle.

XVI. An Ode to Forward Genetics and Gene Discovery The strength of the forward genetic approach is its ability to identify genes in the absence of any hypothesis or notion of how a system operates. The experience of forward genetics in flies and in mice demonstrates the successful application of this approach. However, this experience can be extended especially now with an n ⫽ 6 in flies, and suggests an even more important fruit of this genetic labor.

384

Allada and Rosbash

At the outset of Konopka and Benzer’s endeavor, one might have hypothesized that some fraction of mutants would only indirectly affect clock function but would not participate directly in the clock mechanism. Put more experimentally, one might clone clock genes with no direct biochemical relationship one to another. One might also have hypothesized that another fraction of mutants would affect some general function like transcription, which participates only indirectly in circadian rhythm generation by participating in nearly all processes required in a living cell. Thirty years after Konopka and Benzer’s original work, the geneticists now have hard evidence to counter this pessimistic view; we are now six for six and counting. In each case, further functional analysis of the cloned rhythm genes demonstrated that they encode central components that directly interact in biochemically meaningful fashions. In addition, none of the genes encodes a general component, such as RNA polymerase II. Are circadian geneticists simply lucky? A tenable argument perhaps after the discovery and analysis of period but less plausible after timeless. After analysis of several genes in both flies and mammals, this seems unlikely. By screening for rhythm phenotypes in adult flies, the probability of identifying a mutation that affects the clock only very indirectly appears to be low. Perhaps such a mutant probably affects other developmental processes and therefore would not live to be assayed. Similarly, mutations in genes such as RNA polymerase II, which may play a direct role but also have more general functions, would be so essential for orchestrating the intricacies of early development that a mutant would be unlikely to live to adulthood. It is therefore no accident that the genetics selected for central clock components are largely dedicated to the assayed phenotype. With the completion of the full Drosophila genomic sequence, the forward genetic approach should be even more powerful as an arm of genome-wide functional analysis.

Acknowledgments We thank Joan Rutila, Patrick Emery, and Vipin Suri for helpful comments and assistance with figures. R.A. is supported by a Burroughs Wellcome Fund Career Award in the Biomedical Sciences.

References 1. Aschoff J, Wever R. Spontanperiodik des menschen bei ausschluss aller zeitgeber. Natwissen 1962; 49:337–342. 2. Ruby NF, Heller HC. Temperature sensitivity of the suprachiasmatic nucleus of ground squirrels and rats in vitro. J Biol Rhythms 1996; 11(2):126–136.

Fly Clocks

385

3. Muller JE. Circadian variation and triggering of acute coronary events. Am Heart J 1999; 137(4):S1–S8. 4. Anwar YA, White WB. Chronotherapeutics for cardiovascular disease. Drugs 1998; 55(5):631–643. 5. Uzu T, Fujii T, Nishimura M, Kuroda S, Nakamura S, Inenaga T, et al. Determinants of circadian blood pressure rhythm in essential hypertension. Am J Hypertens 1999; 12(1):35–39. 6. Martin RJ. Nocturnal asthma and the use of theophylline. Clin Exp Allergy 1998; 28(3):64–70. 7. Jones CR, Campbell SS, Zone SE, Cooper F, DeSano A, Murphy PJ, et al. Familial advanced sleep-phase syndrome: a short-period circadian rhythm variant in humans. Nat Med 1999; 5(9):1062–1065. 8. Recht LD, Lew RA, Schwartz WJ. Baseball teams beaten by jet lag. Nature 1995; 377(6550):583. 9. Lewy AJ, Bauer VK, Cutler NL, Sack RL, Ahmed S, Thomas KH, et al. Morning vs. evening light treatment of patients with winter depression. Arch Gen Psychiatry 1998; 55(10):890–896. 10. Johnsson A, Pflug B, Engelmann W, Klemke W. Effect of lithium carbonate on circadian periodicity in humans. Pharmakopsychiatr Neuropsychopharmakol 1979; 12:423–425. 11. Moore RY, Eichler VB. Loss of a circadian adrenal corticosterone rhythm following suprachiasmatic lesions in the rat. Brain Res 1972; 42(1):201–206. 12. Stephan FK, Zucker I. Circadian rhythms in drinking behavior and locomotor activity of rats are eliminated by hypothalamic lesions. Proc Natl Acad Sci USA 1972; 69:1583–1586. 13. Ralph MR, Menaker M. A mutation of the circadian system in golden hamsters. Science 1988; 241:1225–1227. 14. Ralph MR, Foster RG, Davis FC, Menaker M. Transplanted suprachiasmatic nucleus determines circadian period. Science 1990; 247:975–978. 15. Schwartz WJ, Gainer H. Suprachiasmatic nucleus: use of a 14C-labeled deoxyglucose uptake as a functional marker. Science 1977; 197(4308):1089–1091. 16. Welsh DK, Logothetis DE, Meister M, Reppert SM. Individual neurons dissociated from rat suprachiasmatic nucleus express independently phased circadian firing rhythms. Neuron 1995; 14:697–706. 17. Konopka RJ, Benzer S. Clock mutants of Drosophila melanogaster. Proc Natl Acad Sci USA 1971; 68:2112–2116. 18. Pittendrigh CS. On temperature independence in the clock-system controlling emergence time in Drosophila. Proc Natl Acad Sci USA 1954; 40:1018–1029. 19. Klevecz RR, Bolen JL. Dynamic analysis of period mutants in Drosophila: how a precise circadian clock might emerge from a tissue composed of chaotic cellular oscillators. In: Young MW, ed. Molecular Genetics of Biological Rhythms. New York: Marcel Dekker, 1993:221–253. 20. Yu Q, Jacquier AC, Citri Y, Hamblen M, Hall JC, Rosbash M. Molecular mapping of point mutations in the period gene that stop or speed up biological clocks in Drosophila melanogaster. Proc Natl Acad Sci USA 1987; 84:784–788. 21. Baylies MK, Bargiello TA, Jackson FR, Young MW. Changes in abundance and

386

22. 23. 24.

25. 26. 27. 28.

29.

30. 31.

32.

33.

34. 35.

36.

37. 38.

Allada and Rosbash structure of the per gene product can alter periodicity of the Drosophila clock. Nature 1987; 326:390–392. Jackson RF, Bargiello TA, Yun S-H, Young MW. Product of the per locus of Drosophila shares homology with proteoglycans. Nature 1986; 320:185–188. Reddy P, Jacquier AC, Abovich N, Petersen G, Rosbash M. The period clock locus of D. melanogaster codes for a proteoglycan. Cell 1986; 46:53–61. Crews ST, Thomas JB, Goodman CS. The Drosophila single-minded gene encodes a nuclear protein with sequence similarity to the per gene product. Cell 1988; 52: 143–152. Hardin PE, Hall JC, Rosbash M. Feedback of the Drosophila period gene product on circadian cycling of its messenger RNA levels. Nature 1990; 343:536–540. Edery I, Rutila JE, Rosbash M. Phase shifting of the circadian clock by induction of the Drosophila period protein. Science 1994; 263:237–240. Hardin PE, Hall JC, Rosbash M. Circadian oscillations in period gene mRNA levels are transcriptionally regulated. Proc Natl Acad Sci USA 1992; 89:11711–11715. Siwicki KK, Eastman C, Petersen G, Rosbash M, Hall JC. Antibodies to the period gene product of Drosophila reveal diverse tissue distribution and rhythmic changes in the visual system. Neuron 1988; 1:141–150. Zerr DM, Hall JC, Rosbash M, Siwicki KK. Circadian fluctuations of period protein immunoreactivity in the CNS and the visual system of Drosophila. J Neurosci 1990; 10:2749–2762. Edery I, Zwiebel LJ, Dembinska ME, Rosbash M. Temporal phosphorylation of the Drosophila period protein. Proc Natl Acad Sci USA 1994; 91(6):2260–2264. Curtin K, Huang ZJ, Rosbash M. Temporally regulated nuclear entry of the Drosophila period protein contributes to the circadian clock. Neuron 1995; 14:365– 372. Sehgal A, Price JL, Man B, Young MW. Loss of circadian behavioral rhythms and per RNA oscillations in the Drosophila mutant timeless. Science 1994; 263:1603– 1606. Sehgal A, Rothenfluh-Hilfiker A, Hunter-Ensor M, Chen Y, Myers M, Young MW. Circadian oscillations and autoregulation of timeless RNA. Science 1995; 270:808– 810. Huang ZJ, Edery I, Rosbash M. PAS is a dimerization domain common to Drosophila Period and several transcription factors. Nature 1993; 364:259–262. Lindebro MC, Poellinger L, Whitelaw ML. Protein-protein interaction via PAS domains: role of the PAS domain in positive and negative regulation of the bHLH/ PAS dioxin receptor-Arnt transcription factor complex. EMBO J 1995; 14(14): 3528–3539. Vosshall LB, Price JL, Sehgal A, Saez L, Young MW. Specific block in nuclear localization of period protein by a second clock mutation, timeless. Science 1994; 263:1606–1609. Saez L, Young MW. Regulation of nuclear entry of the Drosophila clock proteins PERIOD and TIMELESS. Neuron 1996; 17:911–920. Gekakis N, Saez L, Delahaye-Brown A-M, Myers MP, Sehgal A, Young MW, et al. Isolation of timeless by PER protein interactions: defective interaction between timeless protein and long-period mutant PER L. Science 1995; 270:811–815.

Fly Clocks

387

39. Rutila JE, Zeng H, Le M, Curtin KD, Hall JC, Rosbash M. The tim SL mutant of the Drosophila rhythm gene timeless manifests allele-specific interactions with period gene mutants. Neuron 1996; 17:921–929. 40. Zeng H, Qian Z, Myers MP, Rosbash M. A light-entrainment mechanism for the Drosophila circadian clock. Nature 1996; 380:129–135. 41. Hunter-Ensor M, Ousley A, Sehgal A. Regulation of the Drosophila protein timeless suggests a mechanism for resetting the circadian clock by light. Cell 1996; 84: 677–685. 42. Lee C, Parikh V, Itsukaichi T, Bae K, Edery I. Resetting the Drosophila clock by photic regulation of PER and a PER-TIM complex. Science 1996; 271:1740–1744. 43. Myers MP, Wager-Smith K, Rothenfluh-Hilfiker A, Young MW. Light-induced degradation of TIMELESS and entrainment of the Drosophila circadian clock. Science 1996; 271:1736–1740. 44. Pittendrigh CS. Circadian rhythms and the circadian organization of living systems. Cold Spring Harbor Symp Quant Biol 1960; 25:159–182. 45. Quiring R, Walldorf U, Kloter U, Gehring WJ. Homology of the eyeless gene of Drosophila to the Small eye gene in mice and aniridia in humans. Science 1994; 265(5173):785–789. 46. Jordan T, Hanson I, Zaletayev D, Hodgson S, Prosser J, Seawright A, et al. The human PAX6 gene is mutated in two patients with aniridia. Nat Genet 1992; 1(5): 328–332. 47. Halder G, Callaerts P, Gehring WJ. Induction of ectopic eyes by targeted expression of the eyeless gene in Drosophila. Science 1995; 267(5205):1788–1792. 48. Vitaterna MH, King DP, Chang A-M, Kornhauser JM, Lowrey PL, McDonald JD, et al. Mutagenesis and mapping of a mouse gene, Clock, essential for circadian behavior. Science 1994; 264:719–725. 49. King DP, Vitaterna MH, Chang A-M, Dove WF, Pinto LH, Turek FW, et al. The mouse clock mutation behaves as an antimorph and maps within the W 19H deletion, distal of kit. Genetics 1997; 146:1049–1060. 50. King DP, Zhao Y, Sangoram AM, Wilsbacher LD, Tanaka M, Antoch MP, et al. Positional Cloning of the mouse circadian clock gene. Cell 1997; 89:641–653. 51. Antoch MP, Song E-J, Chang A-M, Vitaterna MH, Zhao Y, Wilsbacher LD, et al. Functional identification of the mouse circadian clock gene by transgenic BAC rescue. Cell 1997; 89:655–667. 52. Tei H, Okamura H, Shigeyoshi Y, Fukuhara C, Ozawa R, Hirose M, et al. Circadian oscillation of a mammalian homologue of the Drosophila period gene. Nature 1997; 389(6650):512–516. 53. Sun ZS, Albrecht U, Zhuchenko O, Bailey J, Eichele G, Lee CC. RIGUI, a putative mammalian ortholog of the Drosophila period gene. Cell 1997; 90:1003–1011. 54. Shearman LP, Zylka MJ, Weaver DR, Kolakowski LF Jr, Reppert SM. Two period homologs: circadian expression and photic regulation in the suprachiasmatic nuclei. Neuron 1997; 19:1261–1269. 55. Zylka MJ, Shearman LP, Weaver DR, Reppert SM. Three period homologs in mammals: differential light responses in the suprachiasmatic circadian clock and oscillating transcripts outside of brain. Neuron 1998; 20:1103–1110. 56. Takumi T, Taguchi K, Miyake S, Sakakida Y, Takashima N, Matsubara C, et al.

388

57.

58.

59.

60. 61.

62.

63.

64.

65.

66. 67.

68.

69. 70.

71.

Allada and Rosbash A light-independent oscillatory gene mPer3 in mouse SCN and OVLT. EMBO 1998; 17(16):4753–4759. Albrecht U, Sun ZS, Eichele G, Lee CC. A differential response of two putative mammalian circadian regulators, mper1 and mper2, to light. Cell 1997; 91:1055– 1064. Shigeyoshi Y, Taguchi K, Yamamoto S, Takekida S, Yan L, Tei H, et al. Lightinduced resetting of a mammalian circadian clock is associated with rapid induction of the mPer1 transcript. Cell 1997; 91:1043–1053. Zheng B, Larkin DW, Albrecht U, Sun ZS, Sage M, Eichele G, et al. The mPer2 gene encodes a functional component of the mammalian circadian clock. Nature 1999; 400:169–173. Hao H, Allen DL, Hardin PE. A circadian enhancer mediates PER-dependent mRNA cycling in Drosophila melanogaster. Mol Cell Biol 1997; 17(7):3687–3693. Allada R, White NE, So WV, Hall JC, Rosbash M. A mutant Drosophila homolog of mammalian Clock disrupts circadian rhythms and transcription of period and timeless. Cell 1998; 93:791–804. Rutila JE, Suri V, Le M, So WV, Rosbash M, Hall JC. CYCLE is a second bHLHPAS protein essential for circadian transcription of Drosophila period and timeless. Cell 1998; 93:805–814. Gekakis N, Staknis D, Nguyen HB, Davis CF, Wilsbacher LD, King DP, et al. Role of the CLOCK protein in the mammalian circadian mechanism. Science 1998; 280:1564–1569. Darlington TK, Wager-Smith K, Ceriani MF, Staknis D, Gekakis N, Steeves TDL, et al. Closing the circadian loop: CLOCK-induced transcription of its own inhibitors per and tim. Science 1998; 280:1599–1603. Lee C, Bae K, Edery I. The Drosophila CLOCK protein undergoes daily rhythms in abundance, phosphorylation and interactions with the PER-TIM complex. Neuron 1998. Zylka MJ, Shearman LP, Levine JD, Jin X, Weaver DR, Reppert SM. Molecular analysis of mammalian timeless. Neuron 1998; 21(5):1115–1122. Sangoram AM, Saez L, Antoch MP, Gekakis N, Staknis D, Whiteley A, et al. Mammalian circadian autoregulatory loop: a timeless ortholog and mPer1 interact and negatively regulate CLOCK-BMAL1–induced transcription. Neuron 1998; 21(5):1101–1113. Frisch B, Hardin PE, Hamblen-Coyle MJ, Rosbash M, Hall JC. A promoterless DNA fragment from the period locus rescues behavioral rhythmicity and mediates cyclical gene expression in a restricted subset of the Drosophila nervous system. Neuron 1994; 12:555–570. So WV, Rosbash M. Post-transcriptional regulation contributes to Drosophila clock gene mRNA cycling. EMBO J 1997; 16(23):7146–7155. Garceau NY, Liu Y, Loros JJ, Dunlap JC. Alternative initiation of translation and time-specific phosphorylation yield multiple forms of the essential clock protein FREQUENCY. Cell 1997; 89(3):469–76. Price JL, Blau J, Rothenfluh-Hilfiker A, Abodeely M, Kloss B, Young MW. doubletime Is a novel Drosophila clock gene that regulates PERIOD protein accumulation. Cell 1998; 94:83–95.

Fly Clocks

389

72. Kloss B, Price JL, Saez L, Blau J, Rothenfluh-Hilfiker A, Wesley CS, et al. The Drosophila clock gene double-time encodes a protein closely related to human casein kinase Ie. Cell 1998; 94:97–107. 73. Lucas RJ, Freedman MS, Munoz M, Garcia-Fernandez JM, Foster RG. Regulation of the mammalian pineal by non-rod, non-cone, ocular photoreceptors. Science 1999; 284(5413):505–507. 74. Freedman MS, Lucas RJ, Soni B, von Schantz M, Munoz M, David-Gray Z, et al. Regulation of mammalian circadian behavior by non-rod, non-cone, ocular photoreceptors. Science 1999; 284(5413):502–504. 75. Takahashi JS, DeCoursey PJ, Bauman L, Menaker M. Spectral sensitivity of a novel photoreceptive system mediating entrainment of mammalian circadian rhythms. Nature 1984; 308:186–188. 76. Campbell SS, Murphy PJ. Extraocular circadian phototransduction in humans. Science 1998; 279:396–399. 77. Yamazaki S, Goto M, Menaker M. No evidence for extraocular photoreceptors in the circadian system of the Syrian hamster. J Biol Rhythms 1999; 14(3):197–201. 78. Cashmore AR, Jarillo JA, Wu YJ, Liu D. Cryptochromes: blue light receptors for plants and animals. Science 1999; 284(5415):760–765. 79. Ahmad M, Jarillo JA, Smirnova O, Cashmore AR. Cryptochrome blue-light photoreceptors of Arabidopsis implicated in phototropism. Nature 1998; 392:720–723. 80. Miyamoto Y, Sancar A. Vitamin B2–based blue-light photoreceptors in the retinohypothalamic tract as the photoactive pigments for setting the circadian clock in mammals. Proc Natl Acad Sci USA 1998:6097–6102. 81. Thresher RJ, Vitaterna MH, Miyamoto Y, Kazantsev A, Hsu DS, Petit C, et al. Role of mouse cryptochrome blue-light photoreceptor in circadian photoresponses. Science 1998; 282:1490–1494. 82. van der Horst GT, Muijtjens M, Kobayashi K, Takano R, Kanno S, Takao M, et al. Mammalian Cry1 and Cry2 are essential for maintenance of circadian rhythms. Nature 1999; 398(6728):627–630. 83. Kume K, Zylka MJ, Sriram S, Shearman LP, Weaver DR, Jin X, et al. mCRY1 and mCRY2 Are essential components of the negative limb of the circadian clock feedback loop. Cell 1999; 98(2):193–205. 84. Zimmerman WF, Goldsmith TH. Photosensitivity of the circadian rhythm and of visual receptors in carotenoid-depleted Drosophila. Science 1971; 171:1167–1168. 85. Suri V, Qian Z, Hall JC, Rosbash M. Evidence that the TIM light response is relevant to light-induced phase shifts in Drosophila melanogaster. Neuron 1998; 21: 225–234. 86. Stanewsky R, Kaneko M, Emery P, Beretta M, Wager-Smith K, Kay SA, et al. The cry b mutation identifies cryptochrome as a circadian photoreceptor in Drosophila. Cell 1998; 95:681–692. 87. Emery P, So WV, Kaneko M, Hall JC, Rosbash M. CRY, a Drosophila clock and light-regulated cryptochrome, is a major contributor to circadian rhythm resetting and photosensitivity. Cell 1998; 95:669–679. 88. Ceriani MF, Darlington TK, Staknis D, Mas P, Petti AA, Weitz CJ, et al. Lightdependent sequestration of TIMELESS by CRYPTOCHROME. Science 1999; 285(5427):553–556.

390

Allada and Rosbash

89. Liu X, Lorenz L, Yu Q, Hall JC, Rosbash M. Spatial and temporal expression of the period gene in Drosophila melanogaster. Genes Dev 1988; 2:228–238. 90. Hardin PE. Analysis of period mRNA cycling in Drosophila head and body tissues indicates that body oscillators behave differently from head oscillators. Mol Cell Biol 1994; 14:7211–7218. 91. Plautz JD, Kaneko M, Hall JC, Kay SA. Independent photoreceptive circadian clocks throughout Drosophila. Science 1997; 278:1632–1635. 92. Sakamoto K, Nagase T, Fukui H, Horikawa K, Okada T, Tanaka H, et al. Multitissue circadian expression of rat period homolog (rPer2) mRNA is governed by the mammalian circadian clock, the suprachiasmatic nucleus in the brain. J Biol Chem 1998; 273(42):27039–27042. 93. Balsalobre A, Damiola F, Schibler U. Immortalized rat fibroblasts contain a circadian clock. Cell 1998; 93:929–937. 94. Liu Y, Tsinoremas NF, Johnson CH, Lebedeva NV, Golden SS, Ishiura M, et al. Circadian orchestration of gene expression in cyanobacteria. Genes Dev 1995; 9: 1469–1478. 95. Van Gelder RN, Bae H, Palazzolo MJ, Krasnow MA. Extent and character of circadian gene expression in Drosophila melanogaster: identification of twenty oscillating mRNAs in the fly head. Curr Biol 1995; 5(12):1424–1436. 96. Uhl GR, Reppert SM. Suprachiasmatic nucleus vasopressin messenger RNA: circadian variation in normal and Brattleboro rats. Science 1986; 232:390–393. 97. Jin X, Shearman LP, Weaver DR, Zylka MJ, de Vries GJ, Reppert SM. A molecular mechanism regulating rhythmic output from the suprachiasmatic circadian clock. Cell 1998; 96(1):57–68. 98. Newby LM, Jackson FR. A new biological rhythm mutant of Drosophila melanogaster that identifies a gene with an essential embryonic function. Genetics 1993; 135:1077–1090. 99. McNeil GP, Zhang X, Genova G, Jackson FR. A molecular rhythm mediating circadian clock output in Drosophila. Neuron 1999; 20(2):297–303. 100. Newby LM, Jackson FR. Regulation of a specific circadian clock output pathway by lark, a putative RNA-binding protein with repressor activity. J Neurobiol 1996; (311):117–128. 101. Rouyer F, Rachidi M, Pikielny C, Rosbash M. A new clock gene regulated by the circadian clock in the Drosophila head. EMBO J 1997:3944–3954. 102. Stanewsky R, Jamison CF, Plautz JD, Kay SA, Hall JC. Multiple circadian-regulated elements contribute to cycling period gene expression in Drosophila. EMBO J 1997; 16(16):5006–5018. 103. Majercak J, Kalderon D, Edery I. Drosophila melanogaster deficient in protein kinase A manifests behavior-specific arrhythmia but normal clock function. Mol Cell Biol 1997; 17(10):5915–5922.

AUTHOR INDEX

Italic numbers give the page on which the complete reference is listed.

A Aalto-Setala K, 325, 326, 336, 337 Abbas AK, 296, 310, 356, 363 Abbott B, 160, 174 Abdelilah S, 226, 237 Abduljabbar TS, 261, 277 Abe H, 107, 115, 131, 158, 173 Abe K, 24, 32 Abe M, 158, 173 Abel T, 192, 196 Aberle H, 295, 309 Abodeely M, 379, 388 Abovich N, 370, 386 Abrahamson DR, 287, 307 Abramovitch R, 106, 130, 155, 160, 168, 171 Acheson A, 103, 127, 285, 306 Acker H, 119, 120, 136, 157, 169, 173 Ackerman L, 192, 201 Acquaviva AM, 106, 130 Adams JM, 257, 275, 343, 359 Adams JW, 356, 363 Adamson ED, 225, 238 Adams RH, 287, 307 Affolter M, 67, 70, 82, 85, 92, 96, 97, 123, 125 Afzelius B, 245, 248, 253

Agani F, 147, 154, 155, 157, 158, 160, 161, 166, 168, 169, 170, 171, 172, 173, 177 Agellon LB, 324, 330, 336, 338 Aghdasi B, 229, 238 Agid Y, 355, 363 Agius E, 217, 237 Ahmed S, 368, 385 Ahrlund-Richter L, 50, 58 Akhurst RJ, 289, 307 Al-Adhami MA, 263, 277 Alami R, 50, 58 Alam J, 155, 159, 171 Albee D, 319, 335 Albelda SM, 282, 296, 303, 310 Alberola-Ila J, 49, 57 Albertini P, 350, 360 Albertson DG, 61, 79 Albrecht U, 293, 309, 375, 376, 387, 388 Aldape RA, 225, 235 Aldrich TH, 102, 103, 127, 128, 285, 287, 306 Aldrich TN, 285, 305 Alexander CR, 102, 127 Alexander J, 63, 81, 205, 214, 218, 220, 221, 235, 237 Alexandre A, 158, 173

391

392 Alfonso-Jaume A, 301, 312 Aliprantis AO, 358, 364 Alitalo K, 283, 304 Allada R, 145, 150, 377, 388 Allayee H, 316, 333, 334, 339 Allende M, 267, 278 Allen DL, 376, 388 Alonso JL, 73, 87 Alon T, 166, 175, 283, 305 Altaba A, 62, 68, 80 Alt FW, 69, 84, 259, 276 Alvarado M, 15, 19 Alward WL, 72, 87 Alzari PM, 348, 360 Amacher SL, 270, 279 Amaya E, 257, 275 Amberger A, 161, 174 Ambiru S, 16, 20 Amemiya CT, 41, 43, 55, 56 Amores A, 261, 276 Amrani Y, 296, 310 Amselem S, 249, 254 Amsterdam A, 214, 235, 267, 278 Anandappa R, 69, 85, 192, 198, 206, 219, 236 Anasuma K, 354, 362 Andalibi A, 316, 334 Andang M, 50, 58 Anderson DJ, 104, 128, 287, 288, 293, 307 Anderson DW, 354, 362 Anderson MG, 93, 124 Andreassen O, 354, 362 Andrew DJ, 66, 81, 93, 123 Andrew DP, 296, 297, 310, 311 Andrews JE, 63, 81, 184, 199, 216, 237 Andrews LJ, 355, 362 Angellilo A, 295, 309 Anglade P, 354, 362 Ang SL, 62, 68, 80, 247, 253 Anson-Cartwright L, 217, 237 Ansorge W, 12, 19 Antakly T, 192, 197 Antay T, 192, 197 Antel JP, 357, 363 Antin PB, 72, 87

Author Index Antoch MP, 45, 375, 378, 387, 388, 5656 Antonarakis SE, 156, 172 Antos C, 70, 85 Anwar YA, 367, 385 An WG, 168, 175 Appella E, 69, 84 Apte SS, 301, 312 Aquila WJ, 140, 148 Arany Z, 105, 115, 117, 129, 135 Arany ZP, 155, 171 Aravind L, 25, 32, 105, 129 Arbustini E, 356, 363 Arcilla RA, 242, 252 Arececi RJ, 69, 85 Armstrong B, 24, 32 Armstrong DL, 229, 238 Arnheiter H, 46, 47, 57 Arnold HH, 245, 253 Arnold RS, 120, 136 Artavanis-Tsakonas S, 3, 13, 18, 19, 145, 150 Asahara A, 283, 305 Asahara T, 104, 128 Asahi T, 354, 362 Asa SL, 40, 54 Aschoff J, 365, 384 Ashburner M, 266, 277 Ashkenzi A, 343, 360 Ashwroth A, 72, 87 As WG, 106, 131 Aten J, 354, 362 Atkinson JB, 323, 336 Atkinson RC, 193, 197 Aubin J, 71, 86 Augustine ML, 225, 235 Avery L, 23, 32 Avner P, 330, 338 Avraham KB, 353, 361 Axelrod JD, 183, 184, 199 Ayerinskas I, 290, 308 Ayerinskas II, 229, 235 Ayling C, 45, 56 Ayrall AM, 47, 57 Aza-Blanc P, 111, 133 Azpiazu N, 183, 196

Author Index

393

Azrolan N, 326, 329, 337, 338 Azzam HJS, 299, 312 Azzaria M, 62, 80

B Babee VR, 321, 335 Baccala AA, 169, 177 Bache A, 40, 54 Bacon N, 107, 108, 131 Bacon NC, 109, 120, 131 Bae H, 382, 390 Bae K, 373, 378, 387, 388 Baer R, 259, 276 Baeuerle PA, 111, 132 Bagnasco M, 357, 363 Bahary N, 261, 263, 276, 277 Bahou WF, 299, 311 Bailey A, 12, 19 Bailey J, 375, 376, 387 Balasingam V, 357, 363 Balconi G, 295, 309 Baldwin HS, 228, 230, 232, 236, 237, 293, 296, 309, 310 Baldwin MA, 291, 308 Bales KR, 354, 362 Ballard WW, 261, 276 Ball CA, 25, 32 Ballester M, 356, 363 Balsalobre A, 382, 390 Baltimore D, 36, 53 Bancher H, 354, 361 Bannikov G, 299, 311 Bao S, 229, 238 Barber JR, 49, 57 Barbosa JA, 35, 36, 52 Bargiello TA, 370, 385, 386 Bargmann CI, 23, 32 Baribault H, 225, 238 Barker PA, 357, 363 Barnes JA, 145, 150 Barnett JV, 289, 308 Barone LM, 263, 277 Barres BA, 343, 346, 359 Barreuther M, 288, 296, 307, 310 Barstead R, 30, 32

Barth AIM, 295, 309 Bartlett SM, 106, 109, 120, 130, 131 Bartunkova S, 103, 127, 179, 184, 201, 216, 238, 285, 305 Bashirullah A, 193, 201 Basler K, 2, 8, 18, 73, 88, 96, 111, 125, 132 Bassett DE, 25, 33 Basson CT, 180, 185, 196, 200, 289, 308 Basson M, 24, 32 Bate M, 183, 192, 200 Bates B, 231, 237 Bath ML, 357, 364 Batunkova S, 285, 305 Batzer MA, 41, 55 Bauer C, 120, 136, 156, 170, 172 Bauer H, 161, 174 Bauer HC, 161, 174 Bauer M, 230, 236 Bauer VK, 368, 385 Bauman L, 380, 389 Baunoch D, 160, 174 Baylies MK, 183, 200, 370, 385 Beach D, 111, 132 Beal MF, 354, 362 Beart PM, 354, 362 Beattie CE, 270, 279 Beaty T, 162, 163, 174 Beaudenon S, 112, 134 Becker D, 289, 307 Becker E, 285, 306 Becker T, 267, 278 Beck I, 156, 168, 172, 175 Beck JS, 290, 308 Beck L Jr, 101, 126 Beckmann JS, 353, 361 Beddington R, 192, 198 Beddington RS, 60, 72, 78, 79, 87 Bedell MA, 51, 58 Bedford FK, 72, 87 Beer-Romero P, 111, 133 Begley CG, 257, 275, 276 Behrends S, 122, 137 Behringer RR, 39, 54 Beier D, 261, 277

394 Beier DR, 263, 277 Beilharz EJ, 354, 361 Beiman M, 183, 196 Beitel G, 26, 32 Belaguli N, 192, 200 Belfrage P, 328, 338 Bellamy CO, 357, 363 Bellard F, 194, 198 Bellard M, 194, 198 Bell B, 271, 279 Bellen HJ, 193, 197 Bellido T, 358, 364 Bell L, 289, 308 Bellon T, 291, 308 Bell PRF, 286, 306 Bell SC, 286, 306 Bellusci S, 73, 74, 75, 76, 88, 89, 99, 126 Belluscki S, 99, 126 Belting HG, 43, 56 Bel-Vialar S, 37, 53 Benarous RA, 113, 134 Benes V, 12, 19 Benichou S, 113, 134 Benient M, 321, 335 Bensen L, 282, 303 Benson DW, 180, 185, 196, 200 Bentley KL, 41, 42, 55 Benzer S, 145, 150, 369, 385 Benzing H, 156, 170, 172 Berard J, 71, 86 Berberich MA, 67, 82 Bercovich B, 111, 133 Beretta M, 381, 389 Berg CM, 44, 56 Berg DE, 44, 56 Bergerib N, 166, 167, 175 Berg M, 352, 361 Berg MA, 290, 308 Bergwerff M, 161, 174 Berkovic SF, 352, 361 Berlan M, 353, 361 Bernabeau C, 291, 308 Bernacki SH, 71, 86 Bernard O, 354, 362 Bernard R, 354, 362

Author Index Bernelli-Zazzera A, 155, 172 Bernfield M, 67, 82 Bernstein A, 282, 304 Berr S, 352, 361 Bertness VL, 244, 253 Bertram E, 352, 361 Bertwistle D, 259, 276 Bessho M, 167, 175 Beuchle D, 271, 279 Bhargava J, 41, 42, 43, 55, 56 Bhattacharya S, 155, 171 Bianchi L, 155, 172 Bianco P, 289, 307 Biben C, 184, 196 Bickler P, 141, 149 Bickler PE, 141, 149 Bickler PI, 141, 149 Bicknell R, 281, 303 Bieberich CJ, 40, 54 Biehs B, 96, 125 Bielinska M, 69, 85, 208, 219, 237 Bielsa-Masdeu A, 356, 363 Bier E, 96, 125, 191, 197 Bierkamp C, 293, 309 Bierke-Nelson D, 72, 87 Bierman EL, 326, 337 Biggs J, 16, 20 Biliar TR, 167, 175 Bingle CD, 69, 83 Birchmeier C, 230, 237 Bird CC, 357, 363 Bird J, 3, 18 Birnberg NC, 46, 47, 57 Biron CA, 289, 307 Birren B, 41, 55 Bisgaier CL, 325, 326, 336, 337 Bisgrove BW, 271, 279 Bishopirc NH, 155, 171 Bitgood MJ, 73, 88 Bitoun P, 72, 87 Bjorgell P, 328, 338 Black T, 285, 306 Blader P, 214, 219, 235, 238 Blagosklonny MV, 106, 131, 168, 175 Blair R, 104, 128 Blanar MA, 216, 236

Author Index Blanchard KL, 106, 130 Blanche PJ, 323, 336 Blatt EN, 69, 84 Blau J, 379, 388, 389 Blessing M, 75, 88, 225, 238 Blom van Assendelft G, 40, 54 Blumberg H, 191, 198 Blumenthal T, 27, 33 Boak BB, 289, 307 Bobinski K, 354, 361 Bodmer R, 145, 150, 151, 179, 180, 182, 183, 184, 185, 187, 193, 194, 195, 196, 197, 198, 199, 200, 201 Bodnar J, 330, 338 Bodner SM, 358, 364 Boehm T, 330, 338 Bogen S, 296, 310 Bogen SA, 296, 310 Bogue CW, 70, 73, 86, 87 Boguski MR, 25, 33 Bohinski RJ, 69, 83 Boivin GP, 229, 238 Boland P, 102, 127 Bolen JL, 370, 385 Bollekens JA, 41, 43, 55, 56 Bolli R, 167, 175 Bolon B, 76, 89 Bonadio J, 289, 307 Bonetti B, 357, 363 Bono F, 106, 130, 155, 160, 168, 171, 295, 309 Bonyadi M, 289, 307 Boon LM, 286, 293, 306 Booth G, 154, 170 Bork P, 25, 33 Boskey A, 289, 307 Bostein D, 25, 32 Botterir FM, 36, 53 Bouhassira EE, 50, 58 Bour SP, 113, 134 Bouwmeester T, 222, 236 Bowerman B, 61, 79 Boyer AS, 229, 235, 289, 290, 308 Braddock M, 301, 312 Bradfield CA, 160, 169, 174, 176 Bradley A, 222, 238, 357, 364

395 Bradshaw MS, 41, 43, 55, 56 Brady DR, 354, 361 Brand AH, 6, 18 Brand M, 213, 219, 236, 237, 271, 272, 279 Brandon C, 315, 334 Brand T, 217, 219, 222, 230, 235, 245, 253 Brandt SJ, 257, 275 Braun JS, 358, 364 Bray S, 2, 17 Bredesen DA, 355, 362 Brehm-Gibson T, 225, 235 Breier G, 102, 126, 282, 293, 304, 309 Breindl M, 40, 54 Breitbart RE, 63, 81, 145, 150, 184, 185, 199, 204, 237 Breitman M, 283, 304 Breitman ML, 101, 102, 126, 127, 214, 236, 238, 283, 304 Breitschopf H, 354, 361 Brenner C, 348, 360 Brenner DA, 40, 54 Brenner S, 22, 32, 33 Brent R, 31, 32 Breslow JL, 314, 315, 320, 322, 326, 327, 329, 330, 334, 335, 337, 338 Brewster R, 193, 194, 197 Briata P, 72, 87, 250, 254 Bridoux AM, 249, 254 Brietman ML, 282, 304 Brindle NPJ, 286, 306 Bringas P, 71, 76, 86, 89, 101, 126 Brinster RL, 39, 46, 47, 54, 57 Britton S, 50, 58 Brocard J, 47, 57 Brodianski VM, 353, 361 Brody JS, 76, 89 Brody LS, 69, 84 Brody SL, 69, 84 Bromm M, 355, 362 Bronson RT, 45, 56 Bronson SK, 38, 40, 53 Bronwlie AJ, 263, 277 Brook A, 2, 17, 18 Brooke BS, 289, 307

396 Brooks A, 106, 130 Brooks PC, 299, 311 Brosnan CF, 355, 363 Brousseau D, 27, 32 Brousseau ME, 314, 334 Brown A, 72, 87 Brown CB, 229, 235, 289, 308 Brown D, 216, 238 Browne CP, 69, 84 Brown JH, 356, 363 Brown LF, 283, 304 Brownlie A, 261, 270, 272, 273, 276, 279, 280 Brown MS, 313, 320, 327, 334, 335, 337 Brown NA, 245, 253 Brown RC, 169, 176 Brown RH, 354, 362 Brownstein DG, 3, 18 Bruckner K, 288, 307 Brueckner M, 248, 254 Bruggemann M, 45, 56 Brugnara C, 273, 280 Bruno J, 103, 127, 285, 306 Bruno MD, 70, 72, 85 Brunzell JD, 325, 336 Brusselmans K, 106, 130, 155, 160, 168, 171 Bryant PJ, 2, 15, 18, 19 Bryce D, 283, 304 Brychzy A, 115, 135 Bubb VJ, 357, 363 Bucana C, 215, 237 Bucay N, 76, 89, 329, 338 Buchou T, 295, 309 Buck C, 232, 236, 293, 309 Buck CA, 232, 236, 293, 309 Buckley C, 296, 310 Buck LT, 140, 141, 144, 148, 149 Buechler P, 168, 169, 176 Bundy J, 228, 235 Bunn HF, 105, 106, 107, 108, 111, 118, 119, 129, 130, 131, 136, 156, 158, 169, 170, 172, 173, 177 Buonocore G, 293, 309 Burch BE, 192, 198

Author Index Burch JB, 69, 85, 180, 185, 200, 213, 238 Burch JBE, 194, 198 Burdick MD, 282, 303 Burgess S, 267, 278 Burgoon PW, 145, 150 Burke DT, 41, 55 Burke PV, 120, 137 Burne JF, 343, 346, 359 Burns DK, 320, 335 Burns JC, 266, 267, 278 Burns TL, 290, 308 Burrascano M, 266, 278 Burrows R, 286, 306 Burstyn JN, 121, 137 Busgen T, 115, 135 Butel JS, 357, 364 Byers S, 111, 133 Byrne GW, 48, 57

C Caber H, 352, 361 Cabrera N, 285, 306 Cacheux V, 249, 254 Cado D, 353, 361 Cagan RL, 6, 18 Caillaud JM, 327, 337 Cairo G, 155, 172 Cales C, 291, 308 Callaerts P, 374, 387 Calliani G, 15, 19 Callow MJ, 323, 336 Calvert JT, 286, 293, 306 Camaioni A, 39, 54 Camarillo CW, 282, 303 Camenisch TD, 225, 235 Cameron AM, 225, 235 Cammer P, 245, 248, 253 Campbel IC, 301, 312 Campbell KH, 36, 53 Campbell SS, 368, 380, 385, 389 Camper SA, 72, 87 Campochiaro PA, 166, 175 Canman CE, 357, 364 Cantley LC, 286, 293, 306

Author Index Cao J, 299, 311 Capdevila J, 73, 87 Capecchi MR, 35, 53, 71, 86, 282, 303, 304 Capehart AA, 224, 237 Carcangiu ML, 3, 18 Cardell EL, 229, 238 Cardone MH, 348, 360 Cardoso WV, 70, 76, 78, 86, 89, 99, 126 Carland G, 155, 171 Carle GF, 41, 55 Carmeliet P, 102, 126, 155, 160, 168, 171, 282, 293, 295, 304, 309 Carmena A, 187, 193, 197 Carndess P, 352, 361 Carneliet P, 106, 130 Caro J, 109, 118, 119, 131, 135, 136, 156, 168, 169, 172, 175, 177 Carpenter G, 285, 306 Carraway KL, 286, 293, 306 Carrero P, 117, 135 Carr JL, 41, 42, 43, 55, 56 Carson JL, 247, 253 Carthew RW, 115, 135 Carver-Moore K, 282, 304 Carver-Moore MW, 102, 103, 126 Carvey PM, 354, 362 Casagrande F, 230, 236 Casanova J, 66, 81, 93, 124 Casciola-Rosen LA, 348, 360 Casci T, 2, 17, 95, 125 Cashman NR, 357, 363 Cashmore AR, 380, 389 Castellani LW, 322, 327, 330, 335, 337, 338 Castilla LH, 289, 308 Castle CK, 324, 336 Catella AM, 293, 309 Catsicas S, 350, 360 Cchaplin DD, 44, 56 Celeste AJ, 257, 275 Celli G, 76, 89 Centrella M, 282, 303 Ceriani MF, 378, 381, 388, 389 Cerretti DP, 104, 128, 287, 288, 307

397 Certel K, 93, 124 Certel SJ, 93, 124 Chacko J, 96, 125 Chajek-Shaul T, 324, 336 Cha JHJ, 355, 362 Chalfie M, 23, 32 Challice CE, 208, 210, 238 Chambon P, 47, 57, 215, 217, 227, 236, 237 Chandel NS, 120, 136, 158, 173 Chandler JM, 348, 360 Chang AM, 45, 56, 375, 387 Chang C, 222, 238 Change S, 46, 47, 57 Chang FY, 263, 277 Chang GW, 113, 117, 135, 169, 176 Chang H, 222, 235, 293, 308 Chang HC, 13, 19 Chan-Ling T, 166, 175 Chan WK, 169, 176 Chan WY, 354, 361 Chapelin C, 249, 254 Chaplin DD, 44, 56 Chapman DB, 245, 253 Charng MJ, 229, 230, 235, 238 Charron F, 70, 85, 192, 197 Cheifetz S, 291, 308 Chen A, 112, 134, 353, 361 Chen C, 41, 55 Chen CK, 67, 82, 96, 125 Chen CY, 192, 200 Chen D, 145, 150, 283, 305 Cheng JF, 45, 56 Cheng PF, 216, 236 CHeng S, 219, 236 Chen H, 231, 237, 282, 304 Chen J, 24, 32, 69, 84, 184, 201, 227, 235, 259, 276 Chen JD, 47, 57 Chen JN, 63, 81, 145, 150, 216, 217, 219, 222, 235, 238, 243, 252 Chen W, 3, 18, 63, 81, 109, 131, 267, 278 Chen WS, 62, 68, 80, 83 Chen X, 326, 337 Chen Y, 111, 132, 371, 386

398 Chen YG, 230, 235 Chen Z, 179, 184, 201, 216, 228, 235, 238, 287, 288, 293, 307 Chen ZF, 104, 128, 352, 361 Chen ZJ, 115, 117, 135 Cherbas L, 47, 57 Cherbas P, 47, 57 Cheresh DA, 299, 311, 312 Cherry JM, 25, 32 Chervitz SA, 25, 32 Chervonsky AV, 357, 363 Cheung NS, 354, 362 Chia F, 72, 87 Chiang C, 71, 74, 86 Chiannilkulchai N, 353, 361 Chiaur DS, 111, 133 Chia W, 187, 193, 197, 266, 277 Chi CL, 78, 89 Chien KR, 179, 184, 197, 201, 227, 235, 236, 238, 248, 253, 356, 363 Chierchia SL, 296, 310 Chiesa G, 323, 336 Chin A, 243, 252 Chin BY, 155, 159, 171 Chinoy MR, 71, 86 Chisaka O, 71, 86 Chisholm AD, 104, 128 Chi SM, 156, 172 Chiu CH, 43, 56 Chi Y, 112, 134 Choi AMK, 155, 159, 171 Choi JH, 333, 339 Chosay JG, 296, 310 Choudhry A, 353, 361 Chow LT, 301, 312 Choy HA, 328, 338 Christ B, 284, 305 Christie G, 286, 306 Christofidou-Solomidou M, 282, 303 Christofori G, 104, 128 Christou H, 158, 159, 173 Chuang ML, 104, 128 Chu CY, 111, 133 Chung A, 145, 150, 185, 199 Chung AB, 120, 136 Chung JN, 185, 199

Author Index Chung MA, 62, 80 Chung WM, 355, 362 Ciechanover A, 111, 132, 133 Cilley RE, 71, 86 Citri Y, 370, 385 Claesson-Welsh L, 284, 305 Clapoff S, 43, 56 Clarke AR, 357, 363 Clarke EE, 355, 363 Clark R, 299, 302, 312 Clark RAF, 281, 302 Clark S, 26, 32 Clarkson TB, 314, 334 Clark TG, 228, 235 Claycomb WC, 155, 171 Claypool J, 112, 134 Cleaver O, 101, 126 Clegg CH, 41, 55 Clement A, 249, 254 Cleveland JL, 358, 364 Clevidence DE, 69, 83 Clifford SC, 113, 117, 135, 169, 176 Clift S, 322, 335 Clore GM, 69, 84 Clotman F, 295, 309 Clouthier DE, 327, 337 Clover R, 24, 32 Cockell M, 40, 54 Cockman ME, 113, 117, 135, 169, 176 Cohen B, 157, 169, 173, 191, 197 Cohen GM, 348, 360 Cohen MP, 290, 308 Cohen RD, 327, 332, 337, 339 Cohen SM, 12, 19, 191, 197 Colas P, 31, 32 Coleman T, 326, 336 Coles HF, 343, 346, 359 Cole T, 63, 81 Collen D, 102, 106, 127, 130, 155, 160, 168, 171, 282, 293, 295, 304, 309 Collier I, 299, 311 Collignon J, 246, 247, 248, 253 Collins JL, 352, 361 Collins T, 301, 312 Colman A, 36, 39, 53 Compernolle V, 295, 309

Author Index Compton D, 102, 127, 285, 305 Compton DL, 103, 127, 285, 306 Conaway JW, 112, 134 Conaway RC, 112, 134 Concordet JP, 106, 130, 154, 156, 170 Condorelli F, 155, 172 Conner C, 299, 311 Connors SA, 213, 237, 263, 277 Conrad MN, 112, 134 Conradt B, 24, 32 Constam DB, 220, 237 Constantini F, 35, 52, 69, 84, 317, 335 Conway SJ, 228, 235 Cooke NE, 40, 54 Cook GP, 45, 56 Cooper F, 368, 385 Copeland NG, 51, 58, 102, 127, 216, 236 Corbert S, 357, 363 Corcoran ML, 299, 301, 311, 312 Cormier-Regard S, 155, 171 Corps EM, 45, 56 Corrales J, 248, 254 Correll CC, 111, 112, 113, 132, 134 Cortes A, 263, 264, 277 Cory S, 343, 357, 359, 364 Costa RH, 68, 69, 83, 84 Cousins FM, 289, 307 Couture D, 352, 361 Cox GW, 106, 130, 159, 173 Crabtree GR, 228, 236 Crague EJ, 282, 303 Craig CG, 193, 201 Craig KL, 113, 134 Crews S, 66, 82, 93, 123 Crews ST, 105, 109, 129, 131, 370, 386 Cross JC, 216, 217, 236, 237 Crowe DL, 76, 89, 101, 126 Cserjesi P, 216, 238 Cui C, 43, 55 Culp P, 267, 278 Cummings K, 184, 200 Cunniff K, 69, 84 Curran T, 45, 56 Currie AR, 341, 342, 359 Curtin K, 371, 386

399 Curtin KD, 373, 387 Curtis L, 357, 363 Cutler NL, 368, 385 Czegledy-Nagy E, 248, 254 Czyzk-Krzeska MF, 106, 130

D Dabert P, 42, 55 Dachs GU, 168, 176 D’Agati V, 69, 84, 260, 276 Dagnino L, 192, 197 Dai WL, 245, 253 Dal Bello B, 356, 363 Dale L, 217, 237, 257, 275 Daly TJ, 285, 305 Dambly-Chaudiere C, 192, 201 Damert A, 169, 176 Damiola F, 382, 390 D’Amore PA, 101, 103, 126, 127, 282, 304 Damsky CH, 299, 312 D’Andrea AD, 259, 276 Dang CV, 168, 175 Dang Q, 329, 338 Danielson PD, 230, 238 Daniel TO, 104, 128, 287, 288, 307 Danilenko DM, 76, 89, 99, 126 Dardik R, 296, 310 Darling S, 43, 56 Darlington TK, 378, 381, 388, 389 Darnell JE, 62, 68, 80, 83 Das DK, 167, 175 Das P, 2, 17 Datson NA, 72, 87 Dausman J, 39, 54 David-Gray Z, 380, 389 Davidson EH, 2, 18 Davis AM, 75, 88 Davis CF, 378, 388 Davis EC, 289, 307 Davis FC, 369, 385 Davis GE, 282, 303 Davis L, 164, 174 Davis LE, 164, 174 Davis MG, 356, 363

400 Davis S, 102, 103, 127, 128, 285, 287, 296, 299, 302, 305, 306, 310, 312 Davis SJ, 282, 301, 303, 312 Davis SR, 50, 57 Dawson A, 37, 53 de Bilbao F, 354, 362 Dec GW, 356, 363 Decimo D, 227, 236 Deckwerth TL, 350, 360 DeClerck YA, 299, 311 Declercq C, 102, 127, 282, 304 DeCoursey PJ, 380, 389 De Felice M, 71, 86 Degenstein L, 76, 89 DeGregori J, 111, 133 Dehart DB, 247, 253 Dejana E, 293, 295, 309 de Jonge HR, 121, 137 de Jong F, 210, 237 de Jong JM, 354, 362 de Jong PJ, 41, 55 de la Brousse FC, 228, 237 de la Cruz AF, 8, 19 Delahaye-Brown AM, 373, 386 de la Pompa JL, 228, 236, 349, 360 DeLisser HM, 282, 296, 303, 310 Della N, 166, 175 Demacker PNM, 327, 337 De Maria R, 357, 363 De Marzo AM, 168, 169, 176 de Matteis F, 119, 136 de Maximy AA, 76, 89, 99, 126 Dembinska ME, 371, 386 Dembski M, 2, 18 Demunck H, 293, 309 Deng C, 282, 304 Deng CX, 289, 308 Deng X, 296, 310 Denson LA, 73, 87 Denyn MM, 210, 237 DePamphilis ML, 317, 335 Deppe U, 63, 81 De Robertis EM, 191, 198, 213, 217, 237, 238 DeRose M, 76, 89 DeRose ML, 76, 89

Author Index DeRuiter MC, 161, 174, 207, 208, 236 de Ruiter MC, 295, 309 DeSano A, 368, 385 de Sauvage FJ, 2, 17 Descombes P, 40, 55 Deshaies RJ, 111, 112, 113, 132, 134 De Strooper B, 27, 32 Detrick HW, 206, 213, 238, 263, 270, 277, 279 Deutsch U, 285, 287, 293, 305, 307 de Vera ME, 167, 175 Deveraux QL, 346, 360 de Vries GJ, 383, 390 Dewerchin M, 155, 160, 168, 171, 295, 309 Dewershin M, 106, 130 DeZarro J, 145, 150 Diamont AJ, 286, 293, 306 Diaz-Rodriguez E, 285, 306 Dicharry SA, 266, 277 Dick A, 222, 236 Dickson DW, 355, 363 Diella F, 287, 307 Dierks EA, 121, 137 Di Lauro R, 69, 71, 83, 86 Dimario F, 352, 361 Ding Q, 74, 88, 99, 126 Discher DJ, 155, 171 Ditta GS, 105, 118, 129, 130, 136 Dixit V, 284, 305 Dixit VM, 287, 306, 343, 360 Docherty AJP, 299, 311 Dodel RC, 354, 362 Doe CQ, 192, 193, 197, 200 Doetschman T, 35, 53, 229, 238 Dolci S, 39, 54 Dolinkski K, 25, 32 Doll CJ, 140, 148 Dolle P, 215, 217, 227, 236, 237 Domeier ME, 62, 79 Dominguez P, 289, 307 Donahoe PK, 230, 238 Donehower LA, 357, 364 Dong X, 115, 135 Donnelly DF, 140, 148 Donnelly ET, 45, 56

Author Index Donovan A, 261, 273, 276, 280 Doolittle MH, 330, 331, 338, 339 Dorian KJ, 270, 279 Dorman JB, 62, 79 Dorn GW, 356, 363 Dorward DW, 244, 253 Dor Y, 106, 130, 155, 160, 168, 171 Doughty ML, 45, 56 Dove WF, 375, 387 Do VM, 111, 133 Dowd M, 282, 304 Dowdy SF, 358, 364 Dower N, 261, 276 Dowland LK, 50, 57 Doyle TG, 27, 32 Dragunow M, 354, 361 Draper BW, 61, 79 Dretzen G, 194, 198 Drews M, 299, 311 Driever W, 216, 226, 237, 238, 263, 266, 277, 278 Drucker DJ, 69, 84 Drum H, 71, 86 Drysdale TA, 63, 81, 179, 184, 201 D-Sa-Eipper C, 350, 361 D’Souza SD, 357, 363 Dubois-Dauphin M, 350, 354, 360, 362 Duboule D, 70, 86 Ducker K, 92, 97, 123 Dufort D, 62, 80 Dumont DJ, 102, 127, 283, 304 Duncan GS, 297, 311, 349, 360 Duncan SA, 69, 85, 192, 199, 219, 237 Dunlap JC, 379, 388 Dunn CJ, 296, 310 Dunning SP, 118, 136, 156, 170, 172 Dunn NR, 75, 76, 88, 89 Du Preez HH, 274, 280 Durand H, 113, 134 Duriez B, 249, 254 During MJ, 140, 148 Durocher D, 70, 85, 192, 197 Dushay MS, 2, 18 Duverger N, 327, 337 Du W, 2, 18 Du Y, 354, 362

401 Duyk M, 352, 361 Dvorak AM, 283, 304 Dvorak HF, 283, 304 Dwight SS, 25, 32 Dye CA, 193, 197 Dyer R, 72, 87, 250, 254 Dynia DW, 70, 86 Dynlacht BD, 2, 18 Dyson E, 227, 238 Dyson N, 2, 18 Dziadek M, 70, 86 Dzierzak E, 69, 84

E Eastman C, 371, 386 Eastman DS, 3, 18 Easwaran V, 111, 133 Eaton BA, 61, 79 Eberhardt C, 102, 127, 282, 304 Ebert BL, 106, 107, 130, 131, 158, 173 Echelard Y, 69, 83 Eckhart AD, 155, 171 Eckner FAO, 242, 252 Eckner R, 115, 117, 135 Eddy CA, 274, 280 Edery I, 105, 129, 370, 371, 373, 378, 383, 386, 387, 388, 390 Edgar BA, 2, 8, 18, 19 Edington T, 293, 309 Edwards JE, 243, 252 Eeken JC, 266, 277 Egan ES, 261, 277 Eggan K, 39, 54 Ehleben W, 157, 169, 173 Eichele G, 375, 376, 387, 388 Eichler VB, 369, 385 Eichmann A, 284, 305 Eisenberg LM, 208, 236 Eisen JS, 270, 279 Eisenman L, 352, 361 Ekbolm P, 222, 238 Ekker M, 213, 237, 263, 277 Eksioglu YZ, 352, 361 Eldon ED, 2, 18 Elgin SC, 40, 54

402

Author Index

Elia A, 297, 311, 349, 360 Elia AJ, 228, 236 Elias CG, 296, 310 Eliasson R, 245, 248, 253 Elledge SJ, 111, 112, 113, 132, 133, 134 Ellerby LM, 355, 362 Elliott DA, 63, 80, 185, 200 Elliott G, 285, 306 Elliott JL, 350, 360 Ellisen LW, 3, 18 Ellis RE, 343, 359 Eltayeb BO, 290, 308 Ema M, 107, 115, 131, 158, 160, 169, 173, 174, 176 Emery P, 145, 150, 381, 389 Emmert-Buck MR, 299, 311 Emori Y, 187, 200 Emoto H, 76, 89, 99, 126 Engel JD, 69, 84 Engelmann W, 368, 385 Engelman RM, 167, 175 Enver T, 72, 87 Epstein DJ, 69, 83 Epstein JA, 231, 237 Erickson CP, 229, 235 Erickson RP, 49, 57 Erickson SK, 327, 337 Erickson SL, 230, 236 Escary JL, 328, 338 Escudier E, 249, 254 Espanol M, 141, 149 Euskirchen G, 23, 32 Evans RM, 36, 46, 47, 53, 57, 227, 236, 238 Evans S, 145, 150, 184, 185, 199, 201 Evans SM, 184, 197 Evans T, 69, 85, 191, 192, 194, 198, 256, 259, 260, 275, 276 Evrad P, 293, 309

F Fabry ME, 50, 58 Faessen GH, 155, 171 Fahrig M, 102, 127, 282, 304 Fandrey J, 119, 136, 157, 169, 173

Fang HM, 106, 130 Fan H, 36, 53 Farace MG, 39, 54 Farese RV, 320, 323, 326, 335, 336 Farhood A, 296, 310 Farnet CM, 266, 277 Farrall M, 289, 307 Farrington S, 267, 278 Fatkin D, 180, 185, 196 Faull RL, 354, 361 Fazio S, 321, 323, 326, 335, 336, 337 Fazio VM, 39, 54 Fber JE, 155, 171 Feder JN, 193, 201 Federoff HJ, 168, 175 Federspeil MJ, 37, 53 Feil R, 47, 57 Fei X, 3, 18 Feldman RM, 111, 112, 132, 134 Feldser D, 147, 155, 158, 166, 169, 171, 173 Felsenfeld G, 69, 84, 259, 276 Feng D, 283, 304 Feng YQ, 50, 58 Fenton HJH, 119, 120, 136 Ferguson EL, 183, 191, 198, 201, 213, 236 Fernandes JA, 140, 144, 148 Ferrante RJ, 354, 362 Ferrara N, 102, 103, 126, 155, 165, 172, 174, 281, 282, 284, 303, 304, 305 Ferreira G, 147, 155, 166, 169, 171 Ferreira V, 102, 126, 282, 304 Feuerstein GZ, 356, 363 Fijii H, 244, 253 Fiknnegan D, 37, 53 Finch CE, 354, 361 Fire A, 24, 32, 63, 80, 81 Firth JD, 106, 130 Firulli AB, 217, 236 Fischer HM, 118, 136 Fischer K, 282, 304 Fischer N, 214, 235 Fishbein MC, 316, 334 Fisher EA, 326, 337

Author Index Fisher G, 266, 278 Fisher LW, 289, 307 Fisher MA, 296, 310 Fisher SA, 353, 361 Fishman MC, 63, 81, 145, 151, 179, 185, 197, 199, 206, 213, 216, 238, 243, 252, 270, 279 Fitzgibbons JC, 167, 175 Flamme I, 102, 127, 169, 176 Flanders KC, 289, 307 Flannery ML, 216, 236 Flavell RA, 349, 350, 357, 360, 361, 363 Flax JD, 48, 57 Fleming RJ, 13, 19 Fless GM, 324, 336 Flynn LM, 326, 329, 336, 338 Foda HD, 299, 311 Foerster J, 121, 137 Fogelman AM, 316, 334 Folkman J, 103, 104, 105, 127, 128, 282, 304 Fong AT, 286, 306 Fong G, 283, 304 Fong GH, 102, 127, 214, 236 Force A, 261, 277 Forman BM, 47, 57 Forster A, 259, 276 Forsythe JA, 154, 168, 170 Fortini ME, 194, 198 Fossett N, 192, 197 Foster RG, 369, 380, 385, 389 Fouquet B, 216, 238 Fraf T, 260, 276 Frampton J, 260, 276 Francastel C, 40, 54 Francois V, 191, 197 Frangioni JV, 282, 303 Franke TF, 348, 360 Frankowski H, 350, 360 Franks RG, 105, 129 Frantz G, 230, 236 Frasch M, 179, 180, 182, 183, 186, 190, 194, 195, 196, 197, 198, 201 Frede S, 119, 136, 157, 169, 173 Freedman MS, 380, 389

403 Freeman M, 2, 17, 95, 125 Freemont PS, 111, 132 Freidman R, 229, 238 Freije JM, 16, 20 Frey AS, 355, 362 Frid MG, 154, 156, 170 Friedlander M, 299, 311, 354, 362 Friedlander RM, 354, 355, 362, 363 Friedmann T, 266, 267, 278 Friedrich GA, 38, 51, 54, 228, 235 Frisch B, 378, 388 Frisch S, 348, 360 Fritz A, 261, 263, 266, 277, 278 Frohlich T, 169, 176 Fronc R, 283, 305 Fuchs E, 76, 89 Fuchs SY, 112, 134 Fuiji S, 350, 360 Fuji H, 160, 174 Fujii-Kuriyama Y, 107, 115, 131, 158, 169, 173, 176 Fujii T, 368, 385 Fujimori KE, 72, 87 Fujioka M, 145, 150, 194, 195, 200 Fujiwara M, 76, 89 Fujiwara Y, 69, 84, 257, 260, 276, 285, 293, 305 Fukuhara C, 375, 376, 387 Fukui H, 382, 390 Fukumoto R, 3, 18 Fukumura D, 106, 130, 155, 160, 168, 171 Fukushige T, 62, 64, 79, 81 Fukuyama M, 64, 81 Fung ET, 225, 235 Fung J, 45, 56 Furnari BA, 106, 130 Furutani-Seiki M, 213, 217, 219, 222, 235, 236, 237, 271, 272, 279 Furuta Y, 73, 88, 99, 126 Fu Y, 184, 197

G Gabay L, 66, 67, 82, 92, 93, 95, 96, 97, 101, 123, 124, 125

404 Gage PJ, 72, 87 Gagliardini V, 354, 355, 362, 363 Gainer H, 369, 385 Gaino N, 267, 278 Gajewski K, 192, 197 Galaktionov K, 111, 132 Gale NW, 102, 103, 126, 128, 282, 287, 288, 304, 306, 307 Gallaher N, 111, 132 Gallegos ME, 62, 80 Gallione CJ, 291, 308 Galluzzo A, 357, 363 Galova M, 112, 134 Galson DL, 106, 130, 131 Gansmuller A, 227, 236 Ganster RW, 167, 175 Ganut SJ, 70, 86 Gao J, 225, 238 Garceau NY, 379, 388 Garcia-Fernandez JM, 380, 389 Garcia-Villalba P, 219, 238 Gardner DP, 48, 57 Gardner RL, 59, 79 Garifallou M, 296, 310 Garlanda C, 282, 303 Garnes J, 41, 55 Gasser SM, 40, 54 Gassmann M, 106, 120, 130, 136, 154, 155, 160, 161, 170, 171, 230, 236 Gateff E, 15, 19 Gates DJ, 354, 362 Gates MA, 261, 263, 276, 277 Gearhart JD, 41, 55, 155, 160, 161, 171 Gehring WJ, 15, 19, 70, 85, 97, 125, 145, 151, 374, 387 Geiser A, 289, 307 Geissert D, 270, 279 Gekakis N, 373, 378, 386, 388 Gelbart WM, 191, 199 Geller DA, 167, 175 Gendron-Maguire M, 285, 293, 305 Genova G, 383, 390 George EL, 220, 230, 236, 293, 308 George SE, 104, 128 Georges-Labouesse NE, 220, 236, 293, 308

Author Index Gerber HP, 155, 172, 284, 305 Gerhart J, 213, 236 Gertsenstein M, 101, 102, 126, 127, 214, 236, 238, 282, 283, 304 Gervais FG, 355, 363 Gesteland K, 247, 253 Ghaboosi N, 230, 236 Ghai HS, 140, 141, 144, 148 Ghysen A, 192, 201 Giaccia AJ, 104, 128, 155, 167, 171, 175 Giallongo A, 106, 130, 154, 156, 170 Giambarella U, 355, 363 Gibbs AH, 119, 136 Gidday JM, 167, 175 Gijbels MJJ, 326, 337 Gilbert DJ, 102, 127 Gilhooley H, 37, 53 Gillan A, 40, 54 Gilles-Gonzalez MA, 118, 136 Gillette MU, 145, 150 Gilman JG, 50, 58 Gilthorpe JD, 43, 55 Gimble JM, 326, 336 Gingrich JR, 46, 47, 57 Ginsberg M, 323, 336 Giometti CS, 331, 339 Giordano C, 357, 363 Gisselbrecht S, 187, 192, 193, 197 Gittenberger-de Groot A, 295, 309 Gittenberger-de Groot AC, 161, 174, 207, 208, 229, 236, 238 Gittline JD, 69, 83 Giudice LC, 155, 171 Glazer L, 66, 67, 82, 92, 93, 97, 101, 123, 124 Gleadle JM, 105, 106, 107, 108, 119, 120, 129, 130, 131, 157, 158, 168, 173, 176 Glick BaP, 36, 53 Glimcher LH, 228, 237 Glucksmann A, 342, 359 Glzer L, 66, 82 Gnessin H, 166, 175 Goble J, 352, 361 Goff SC, 69, 84

Author Index Goff SP, 35, 53 Goldberg B, 358, 364 Goldberg GI, 299, 311 Goldberg MA, 118, 136, 158, 173 Goldberg U, 326, 337 Goldbert MA, 156, 170, 172 Golden K, 179, 180, 182, 183, 184, 187, 195, 196, 199, 201 Golden SS, 382, 390 Goldfarb M, 76, 88, 103, 128, 287, 306 Goldsmith TH, 380, 389 Goldstein JL, 313, 320, 326, 327, 334, 335, 337 Goldstein LS, 248, 253 Goldstein S, 289, 307 Goldwasser E, 120, 136, 158, 173 Golic KG, 8, 9, 18 Golling G, 267, 278 Gomez C, 112, 134 Gomi H, 302, 312 Go MJ, 3, 18 Goncharov T, 353, 361 Gong Z, 261, 267, 277, 278 Gonzalez G, 118, 136 Good DJ, 244, 253 Goodman CS, 192, 193, 200, 370, 386 Goodman RH, 111, 132 Goodman S, 217, 237 Goossens M, 249, 254 Gordon JW, 34, 35, 36, 39, 40, 52 Gorey-Faure S, 295, 309 Gorodin S, 354, 362 Goszczynski B, 62, 79, 80 Gotay J, 289, 308 Gotoh O, 160, 174 Goto K, 354, 362 Goto M, 380, 389 Goto S, 93, 124 Goto T, 145, 150, 194, 195, 200 Goto Y, 352, 361 Gottlieb S, 62, 80 Goudzwaard JH, 266, 278 Goumnerov B, 286, 293, 306 Gourdie RG, 210, 236 Goutel C, 261, 277 Gow AJ, 348, 360

405 Goyenechea B, 45, 56 Graeber TG, 104, 128 Graesser D, 299, 302, 311, 312 Graff JM, 257, 275 Graham A, 70, 85, 86, 161, 174 Granato M, 213, 217, 219, 222, 235, 236, 237, 271, 272, 279 Grant GA, 299, 311 Grau E, 62, 68, 80 Gravallese EM, 228, 237 Greaves DR, 40, 54 Green AR, 257, 276 Green DR, 346, 360 Green SA, 45, 56 Greenspan DS, 228, 235 Greenwald I, 27, 32, 33 Gregg RG, 35, 53 Grepin C, 192, 197, 198 Gressens P, 293, 309 Gridley T, 285, 293, 305 Grieder NC, 67, 82, 96, 125 Griendling KK, 120, 136 Griffin KJ, 270, 279 Griffiths JA, 102, 127 Grigaux C, 323, 336 Grindley J, 76, 89, 99, 126 Grindley JC, 74, 88 Grishok A, 29, 33 Gritsman K, 219, 236 Grogg KM, 291, 308 Grondona JM, 227, 236 Gronenborn AM, 69, 84 Groppe J, 92, 97, 123, 125 Grosfeld J, 104, 128 Gross I, 70, 86 Grosskortenhaus R, 192, 200 Gross-Mesilaty S, 111, 133 Grosveld F, 40, 54, 192, 198 Grosveld FG, 69, 84 Grotkopp D, 40, 54 Groudine M, 40, 54 Grow MW, 184, 198 Gruber PJ, 227, 236 Grunwald DJ, 271, 279 Grunwald JD, 265, 277 Grusby MJ, 228, 237

406

Author Index

Gruskin E, 293, 309 Gstaiger M, 113, 135, 169, 176 Guenet JL, 154, 170 Guerrero I, 73, 87 Guichard A, 96, 125 Guillemin K, 66, 81, 92, 93, 95, 97, 98, 123, 124 Guille MJ, 259, 276 Gu J, 107, 111, 131, 158, 169, 173, 177 Gumeringer CL, 227, 238 Guo J, 290, 308 Guo L, 76, 89 Guo M, 193, 194, 198 Guo Q, 222, 229, 235, 238, 293, 308 Guo Y, 167, 175 Gurdon JB, 36, 39, 53 Gurubhagavatula I, 296, 310 Gutgesell HP, 242, 252 Guttmacher AE, 291, 308 Gu YZ, 169, 176

H Haas AL, 115, 135 Haas C, 27, 32 Haase K, 36, 53 Haas T, 299, 302, 312 Haas TL, 282, 298, 303, 311 Haberstroh LL, 192, 197 Hackam AS, 355, 362 Hackett BP, 69, 83, 84 Hackett SF, 166, 175 Hacohen N, 66, 78, 81, 82, 89, 93, 94, 95, 98, 124 Haddad GG, 139, 140, 141, 143, 144, 146, 148, 149, 150 Ha E, 63, 81 Haenlin M, 192, 200 Haerry T, 67, 82, 96, 125 Hafen E, 2, 18, 92, 97, 123 Haffter P, 213, 217, 219, 222, 235, 236, 237, 243, 252, 271, 272, 279 Hagman JR, 38, 40, 53 Hahn PS, 257, 275 Haider N, 356, 363

Haigney MCP, 167, 175 Haitjema T, 291, 308 Hakem A, 349, 360 Hakem R, 349, 360 Halder G, 374, 387 Haldi M, 267, 278 Haley L, 104, 128 Hall JC, 145, 150, 370, 371, 373, 377, 378, 380, 381, 382, 383, 385, 386, 387, 388, 389, 390 Hall K, 40, 54 Halterman MW, 168, 175 Hamada H, 160, 174, 244, 253 Hamblen-Coyle MJ, 378, 388 Hamblen M, 370, 385 Hamilos DL, 69, 84 Hamilton S, 169, 177 Hamilton SL, 229, 238 Hammer RE, 46, 47, 57, 324, 327, 329, 336, 337, 338 Hammerschmidt M, 2, 17, 213, 217, 219, 222, 235, 236, 237, 263, 271, 272, 277, 279 Hanada H, 60, 78, 79 Hanahan D, 104, 128, 282, 304 Hankinson O, 105, 129, 160, 168, 174, 176 Han M, 26, 27, 32, 33 Han PL, 193, 197 Hanscombe O, 40, 54 Hanson I, 374, 387 Hao H, 376, 388 Harada A, 248, 254 Harden N, 104, 128 Hardiman G, 24, 32 Hardin J, 104, 128 Hardin PE, 370, 376, 378, 382, 386, 388, 390 Harendza S, 301, 312 Harland R, 213, 236, 260, 276 Harley L, 184, 199 Harpal K, 62, 80, 102, 127, 282, 304 Harper JW, 111, 112, 113, 132, 133, 134 Harris Al, 168, 176 Harris AW, 357, 364

Author Index Harris KL, 330, 338 Harris MA, 25, 32 Harrison DJ, 357, 363 Harrison J, 187, 193, 197 Harrison V, 297, 311 Harteneck C, 121, 137 Hartley L, 63, 80, 81, 184, 199, 206, 216, 237 Harvey M, 357, 364 Harvey P, 184, 199 Harvey R, 257, 276 Harvey RP, 63, 80, 81, 179, 180, 183, 184, 185, 196, 198, 199, 200, 201, 206, 216, 227, 236, 237 Hashimoto G, 76, 89, 99, 126 Hastie ND, 213, 233, 237 Hatakeyama S, 111, 113, 133 Hatta K, 219, 238 Hattori K, 111, 113, 133 Haun C, 63, 81 Hauser C, 231, 237 Hausladen A, 348, 360 Hawker J, 230, 235 Hawkins MG, 64, 81 Hayashi S, 93, 98, 124, 125 Haydar TF, 349, 350, 360, 361 Hayden MR, 355, 362 Hayek T, 319, 324, 326, 328, 335, 336, 337 Hazendonk E, 24, 28, 30, 32 Hazzard WR, 326, 337 Heberlein U, 73, 87, 145, 150 Hebert JL, 69, 84 Heckscher E, 219, 236 Hedgecock EM, 112, 134 Hedrick CC, 322, 335 Heegaard AM, 289, 307 Heid PJ, 64, 81 Heikinheimo M, 69, 85, 191, 198, 206, 236 Heintz N, 42, 45, 55, 56 Heisenberg CP, 213, 217, 219, 222, 235, 236, 237, 271, 272, 279 Heitzler P, 192, 200 He J, 354, 362 Heldin CH, 284, 289, 305, 308

407 Helinski DR, 118, 136 Heller HC, 367, 384 Helmbold EA, 291, 308 Helsinki DR, 105, 129, 130 Hemmati-Brivanlou A, 222, 238 Hemo I, 283, 305 Henderson JT, 349, 360 Henderson R, 73, 75, 88, 99, 126 Hengartner MO, 26, 32, 343, 359 Henion PD, 270, 279 Henkemeyer M, 288, 307 Henzel WJ, 343, 359 Herbert JM, 106, 130, 155, 160, 168, 171, 295, 309 Her H, 263, 277 Hershko A, 111, 132, 133 Herskowitz I, 49, 57 Hersperger E, 16, 19, 20 Herz J, 320, 327, 328, 335, 337 Herzog KH, 358, 364 Hess DT, 348, 360 Heutink P, 291, 308 Hewitt RE, 301, 312 He WW, 112, 134 He X, 111, 133 Heyden NV, 358, 364 Hibberd MG, 104, 128 Hiemisch H, 69, 84 Higashijima SI, 187, 200 Hild M, 222, 236 Hillan KJ, 282, 304 Hill RJ, 64, 81 Hilton DA, 168, 169, 176 Hing H, 104, 128 Hinkula J, 50, 58 Hirakow R, 211, 236 Hiramatsu KI, 352, 361 Hiratuska S, 283, 304 Hirokawa N, 248, 253, 254 Hiromi Y, 67, 82, 95, 99, 125 Hirose K, 160, 174 Hirose M, 375, 376, 387 Hirose S, 98, 125 Hirota K, 107, 115, 131 Hirsh D, 265, 277 Hiruma T, 211, 236

408 Hoan BL, 69, 83 Hobbs AJ, 121, 137 Hobgod KK, 313, 334 Hochachka PW, 140, 141, 148, 149 Hoch M, 67, 83 Hochman A, 354, 362 Hochstrasser M, 111, 132, 133 Hodge MR, 228, 237 Hodges M, 111, 132 Hodgson S, 374, 387 Hoey T, 228, 237 Hoffman B, 358, 364 Hoffmann FM, 183, 191, 198, 200 Hofmann RM, 111, 133 Hogan B, 317, 335 Hogan BB, 35, 52 Hogan BL, 60, 62, 65, 68, 73, 74, 75, 76, 78, 79, 80, 81, 88, 89, 92, 99, 123, 126, 225, 228, 235, 236, 238, 247, 253 Hogan BLM, 70, 86 Hogenesch JB, 169, 176 Hoheisel J, 192, 199 Hohimer AR, 164, 174 Ho J, 282, 304 Holash J, 102, 127 Holder N, 205, 221, 237, 282, 304 Holland PWH, 70, 86 Holland SJ, 288, 307 Hollenberg MS, 216, 236 Holley SA, 191, 198, 213, 236 Holmes P, 315, 334 Holmgren L, 104, 128 Holmgren R, 74, 88 Holtmann H, 353, 361 Holzman LB, 287, 306 Homanics GE, 326, 337 Hong SC, 352, 361 Hoodless P, 62, 68, 80 Hooper ML, 35, 53, 357, 363 Hope IA, 62, 80 Hopkins N, 214, 235, 267, 278 Horikawa K, 382, 390 Ho RK, 270, 279 Horn DB, 260, 276 Horner MA, 62, 79

Author Index Horne S, 261, 276 Horne SA, 222, 238 Horton WE, 354, 361 Horvitz HR, 24, 25, 26, 28, 32, 343, 359 Horvitz JR, 145, 151 Hoschuetzky H, 295, 309 Hotchkiss G, 50, 58 Hough C, 15, 19 Houseman DE, 104, 128 Houston P, 301, 312 Ho VT, 106, 130 Howard K, 96, 125 Howes G, 257, 275 Howley PM, 112, 134 Hrada M, 354, 362 Hruban RH, 169, 177 Hsu DS, 380, 389 Huang JQ, 355, 363 Huang LE, 105, 106, 107, 111, 115, 117, 129, 131, 135, 158, 169, 173, 177 Huang LH, 192, 199 Huang T, 102, 127 Huang ZJ, 105, 129, 371, 386 Huarte J, 350, 360 Huber O, 293, 309 Huber P, 295, 309 Huber TL, 263, 277 Hudson JB, 183, 201 Hug B, 271, 279 Hughes SH, 37, 53 Huh CG, 289, 307 Huhtala P, 299, 301, 312 Huibregste JM, 112, 133, 134 Hui CC, 74, 88, 99, 126 Hu J, 155, 171 Hummel KP, 245, 253 Humphries MJ, 299, 312 Humphries S, 45, 56 Hung MC, 231, 237 Hunot S, 349, 360 Hunter-Ensor M, 371, 373, 386, 387 Hurst WJ, 145, 150 Hu S, 105, 109, 129, 131 Huso DL, 162, 163, 174

Author Index

409

Huttenlocher PR, 352, 361 Huylebroeck D, 222, 235, 293, 308 Huynh-Do U, 104, 128, 288, 307 Huynh DP, 231, 236 Hwang JK, 43, 56 Hyashi S, 98, 125 Hylland P, 141, 149 Hynes RO, 206, 217, 220, 222, 230, 236, 237, 238, 293, 298, 299, 308, 311, 312

I Iangaki T, 247, 253 Igarashi H, 73, 88 Ignarro LJ, 121, 137 Ikawa M, 43, 55 Ikeda K, 69, 72, 83 Ikeda M, 244, 253 Ikeda T, 302, 312 Ikeo K, 145, 151 Ikeya T, 98, 125 Ilan N, 284, 285, 296, 305, 306, 311 Iliopoulos O, 113, 135 Imbert G, 113, 135, 169, 176 Imhof B, 288, 307 Imhof BA, 288, 291, 293, 307 Inagaki H, 179, 184, 201 Inai Y, 283, 305 Inazu T, 73, 87 Inenanga T, 368, 385 Ingber DE, 282, 303 Ingham PW, 73, 87, 219, 238 Inoue T, 220, 238, 244, 253 Inoue Y, 73, 87 Ioannou PA, 41, 55 Ioliopoulos O, 112, 134 Ip HS, 69, 70, 84, 85, 191, 192, 198, 199, 206, 237 Irion U, 192, 200 Iruela-Arispe ML, 225, 233, 238 Irvine DV, 270, 279 Irwin JC, 155, 171 Isaac DD, 65, 81, 93, 123 Isaacs WB, 168, 169, 176, 177 Ishibashi S, 320, 327, 328, 335, 337

Ishida BY, 319, 323, 335, 336 Ishida N, 111, 113, 133 Ishii Y, 73, 88 Ishimaru Y, 244, 253 Ishiura M, 382, 390 Ishizaki Y, 343, 346, 359 Isner JM, 103, 104, 127, 128 Isner JN, 283, 305 Isner JW, 165, 174 Israels S, 356, 363 Itasaki N, 37, 53 Itin A, 105, 128, 129, 166, 168, 175, 176 Itohara S, 302, 312 Itoh H, 244, 253 Itoh N, 76, 89, 99, 126 Itoh T, 302, 312 Ito Y, 326, 337 Itsukaichi T, 373, 387 Iwaguro H, 283, 305 Iwatsubo T, 355, 363 Iyer NV, 147, 154, 155, 159, 160, 161, 162, 163, 166, 168, 169, 170, 171, 174 Iyohama S, 167, 175 Izpiusa-Belmonte JC, 72, 87, 250, 254 Izraeli S, 244, 253 Izumo S, 179, 184, 185, 192, 198, 199, 201, 216, 238

J Jacinto M, 45, 56 Jackie VH, 169, 176 Jackle H, 62, 68, 80, 83 Jackson CE, 291, 308 Jackson DE, 296, 310 Jackson FR, 370, 383, 385, 390 Jackson-Grusby l, 39, 54 Jackson-Lewis V, 354, 362 Jackson PD, 191, 198 Jackson PK, 111, 113, 132, 134 Jackson RF, 370, 386 Jackson T, 352, 361 Jacks T, 104, 128 Jacob S, 193, 201

410 Jacobs HC, 70, 73, 86, 87 Jacobson MD, 343, 346, 359 Jacquier AC, 370, 385, 386 Jadoon AK, 167, 175 Jaenisch R, 35, 36, 39, 53, 54 Jaeschke H, 296, 310 Jagadeeswaran P, 274, 280 Jagla K, 194, 198 Jagla T, 194, 198 Jahner D, 36, 53 Jain RK, 106, 130, 155, 160, 168, 171 Jain V, 103, 127, 285, 306 Jamison CF, 383, 390 Jamry I, 44, 56 Janeway CA, 357, 363 Jan LY, 182, 192, 193, 194, 196, 198, 199, 200, 201 Jansen G, 24, 28, 30, 32 Jan YN, 182, 192, 193, 194, 196, 198, 199, 200, 201 Jarecki J, 67, 82, 107, 131 Jarillo JA, 380, 389 Jarvis LA, 78, 89 Jasin M, 40, 54 Javoy-Agid Y, 354, 362 Jay G, 40, 54 Jeannotte L, 71, 86 Jelkmann W, 119, 136, 157, 173 Jellinger K, 354, 361 Jenkins NA, 51, 58, 102, 127, 216, 236 Jentsch S, 115, 135 Jessell TM, 62, 68, 80, 83 Jessen JR, 42, 55, 268, 279 Jesuthasan S, 219, 238 Jetten AM, 71, 86 Jiang BH, 105, 106, 107, 118, 129, 130, 131, 135, 154, 155, 156, 157, 159, 164, 168, 170, 170, 171, 172, 174 Jiang C, 113, 135, 139, 140, 141, 144, 148, 149 Jiang F, 192, 199 Jiang GH, 105, 129 Jiang J, 115, 117, 135 Jiang MM, 193, 201 Jiang XC, 330, 338

Author Index Jiang Y, 69, 85, 192, 198, 272, 279 Jiang YJ, 213, 217, 219, 222, 235, 236, 237, 271, 279 Jiang Z, 47, 57 Jian J, 111, 132 Jilka RL, 358, 364 Jimenez F, 187, 193, 197 Jin X, 378, 380, 383, 388, 389, 390 Jobe SM, 192, 199 Johnson A, 368, 385 Johnson CD, 24, 32 Johnson CH, 382, 390 Johnson DF, 320, 335 Johnson DW, 291, 308 Johnson E, 67, 82, 107, 131 Johnson EM, 350, 360 Johnson RI, 250, 254 Johnson RL, 72, 87, 244, 253 Johnson RS, 155, 160, 171 Johnson SL, 261, 270, 272, 277, 279, 280 Johnson WA, 93, 124 Johns RA, 155, 159, 167, 171 Johnston LA, 8, 19 Johnston M, 111, 132 Joly JS, 261, 277 Jones CM, 225, 236 Jones CR, 368, 385 Jones ML, 296, 310 Jones P, 103, 127 Jones PF, 102, 103, 127, 285, 305, 306 Jones RW, 156, 172 Jordan JE, 257, 275 Jordan T, 374, 387 Joutel A, 2, 17 Joyner AL, 69, 74, 83, 88, 317, 335 Jungermann K, 119, 136 Jurgens G, 62, 68, 80 Jussila L, 102, 127, 283, 304 Justice RW, 15, 19

K Kabler S, 141, 149 Kabra NH, 353, 361

Author Index Kaelin WG, 112, 113, 115, 134, 135, 169, 176 Kaestner KH, 62, 68, 69, 80, 84 Kagi D, 349, 360 Kahana C, 111, 133, 157, 169, 173 Kahari VM, 301, 312 Kakizuka A, 355, 362 Kalb JM, 62, 79, 80 Kalderon D, 111, 132, 383, 390 Kalka C, 283, 305 Kallianpur AR, 257, 275 Kallio PJ, 111, 117, 132, 135, 169, 177 Kalra VK, 296, 310 Kamura T, 112, 134 Kamuro K, 352, 361 Kanai Y, 248, 254 Kane D, 272, 279 Kane DA, 213, 217, 219, 222, 235, 236, 237, 270, 271, 279 Kanekal M, 106, 131, 168, 175 Kaneko M, 381, 382, 389, 390 Kane LS, 348, 360 Kanno S, 380, 389 Kaplan KB, 111, 132 Kappel A, 169, 176 Kappen C, 48, 57 Karantamir B, 296, 297, 310, 311 Karasuyama H, 349, 360 Karim FD, 13, 19 Karis A, 69, 84 Karpen SJ, 73, 87 Karp R, 266, 277 Kasahara H, 179, 184, 192, 199, 201 Kasai T, 39, 54 Kasai Y, 105, 129 Kasper M, 293, 309 Kassis JA, 98, 125 Kastan MB, 357, 364 Kastner P, 227, 236 Katayose D, 167, 175 Katoh-Fukui Y, 72, 87 Kato Y, 111, 133 Katz J, 69, 84 Kaufmann E, 68, 83 Kaufman SA, 297, 311, 349, 360 Kawakami K, 267, 278

411 Kay SA, 381, 382, 383, 389, 390 Kazantsev A, 380, 389 Keller G, 69, 84, 259, 260, 276 Keller-Peck CR, 350, 361 Kelley C, 191, 198, 260, 276 Kelley CM, 257, 275 Kelley MR, 266, 277 Kelsh R, 261, 277 Kelsh RN, 213, 217, 219, 222, 235, 236, 237, 271, 272, 279 Kemler R, 293, 295, 309 Kemper OC, 353, 361 Keng VW, 72, 73, 87 Kenyon C, 62, 79 Kern SE, 169, 177 Kerr JFR, 341, 342, 359 Kerr JS, 282, 303 Keshet E, 105, 106, 128, 129, 130, 155, 160, 166, 168, 171, 175, 176, 283, 305 Keski-Oja J, 301, 312 Keyt BA, 284, 305 Khachigian LM, 301, 312 Khan SQ, 355, 362 Kharbanda S, 356, 363 Khoo JC, 328, 338 Khoo W, 349, 360 Kido M, 248, 254 Kieckens L, 102, 126, 282, 304 Kieran MW, 263, 277 Kietzmann T, 119, 136 Kiger L, 118, 136 Kikuchi A, 111, 113, 133 Kimberly D, 164, 174 Kimble J, 62, 79 Kim CS, 296, 310 Kimelman D, 270, 279 Kimmel CB, 261, 266, 270, 276, 278, 279 Kimmel SR, 261, 276 Kim SH, 270, 279 Kim SK, 62, 80 Kim SY, 111, 133 Kim UJ, 41, 55 Kimura S, 69, 71, 84, 86 Kimura T, 296, 310 Kim Y, 63, 80

412 Kincaid OW, 242, 252 Kind AJ, 36, 53 Kinder SJ, 204, 205, 238 King AA, 69, 85 King DP, 45, 56, 375, 378, 387, 388 King J, 24, 32 King JA, 75, 88 Kingsley DM, 75, 88 Kinnunen P, 230, 235 Kinzler KW, 111, 133, 169, 177 Kioussi C, 72, 87, 250, 254 Kipreos ET, 112, 113, 134 Kirby ML, 161, 174, 216, 238 Kirsch IR, 244, 253 Kirschner MW, 257, 275 Kishikawa H, 39, 54 Kishimoto Y, 213, 236, 263, 277 Kiss I, 15, 19 Kitagawa M, 111, 113, 133 Kitajima K, 51, 53 Kitamura K, 72, 87 Klagsbrun M, 103, 104, 105, 128, 284, 305 Klambt C, 66, 67, 82, 93, 124 Klausner RD, 113, 135 Kleiner DE, 301, 312 Klein R, 230, 236, 287, 288, 307 Klein WH, 260, 276 Klemke W, 368, 385 Klevecz RR, 370, 385 Klevenyi P, 354, 362 Klewer SE, 225, 235 Klien R, 282, 304 Klingenspor M, 332, 339 Kloss B, 379, 388, 389 Kloter U, 97, 125, 374, 387 Kluckman KD, 38, 40, 53 Knapik E, 261, 277 Knauber D, 261, 276 Knighton D, 105, 128 Knochel W, 68, 83 Knoll J, 352, 361 Knowles HJ, 69, 84 Knudsen KA, 222, 237 Knudson AG, 15, 19 Knudson CM, 350, 360, 361

Author Index Kobayashi A, 160, 174 Kobayashi K, 380, 389 Kobayashi R, 111, 132 Koch CJ, 104, 106, 128, 130, 155, 160, 168, 171 Kocher O, 289, 308 Koelle MR, 30, 33 Koepp DM, 112, 134 Koesling D, 121, 122, 137 Koga Y, 76, 89 Kolakowski LF, 376, 387 Kollet O, 353, 361 Kollias G, 40, 54 Kolodgie FD, 356, 363 Komada M, 221, 236 Komuro I, 184, 198 Kondoh H, 60, 78, 79, 244, 253 Kondo T, 354, 362 Konopka RJ, 369, 385 Koong AC, 167, 175 Koonin EV, 25, 32, 33 Koos RD, 154, 168, 170 Kopchick J, 43, 55 Kopczynski C, 192, 193, 200 Kopp EB, 2, 18 Kornberg TB, 111, 133 Kornhauser JM, 375, 387 Korsemeyer SJ, 26, 32, 350, 361 Korsmeyer SJ, 341, 350, 359, 360 Korving JP, 75, 88 Koshiba K, 244, 253 Kostic V, 354, 362 Kotch LE, 155, 160, 161, 171, 174 Kothary R, 43, 56 Kou HC, 184, 201 Kourembanas S, 158, 159, 173 Koutsourakis M, 192, 198 Kowalski J, 284, 305 Kozak KR, 160, 174 Kramer S, 67, 82, 95, 99, 125 Krasnow M, 66, 82, 94, 124 Krasnow MA, 65, 66, 67, 78, 81, 82, 89, 91, 92, 93, 94, 95, 97, 98, 99, 101, 102, 103, 107, 123, 124, 131, 145, 150, 382, 390 Krauss R, 322, 335

Author Index Kreck W, 112, 134 Kreidberg J, 213, 233, 237 Kreig PA, 63, 81 Krek W, 113, 135, 169, 176 Krempen K, 40, 54 Krieg C, 63, 81 Krieg PA, 72, 87, 101, 126, 179, 184, 198, 199, 201 Krishnan BR, 44, 56 Krishnan SN, 143, 144, 150 Krishna S, 2, 17 Kroemer G, 348, 360 Kroisel PM, 41, 55 Kroll SL, 119, 136 Krug EL, 229, 237 Krumlauf R, 37, 53, 70, 85, 86 Krummel TM, 71, 86 Kuan C, 349, 360 Kuan CY, 350, 361 Kubalak S, 210, 225, 235, 236 Kubalak SW, 227, 235, 236 Kubota Y, 266, 278 Kuehn M, 244, 248, 253, 254 Kuehn MR, 244, 247, 248, 253 Kuida K, 349, 350, 360, 361 Kulessa H, 260, 276 Kuliszewski M, 76, 89 Kulkarni AB, 289, 307 Kullberg BJ, 327, 337 Kume K, 380, 389 Kumiski D, 161, 174 Kummer W, 157, 169, 173 Kung AL, 155, 171 Kung HF, 257, 275 Kuno J, 283, 304 Kunz YW, 263, 277 Kuo CT, 69, 85, 192, 198, 206, 219, 236 Kuo FC, 69, 84, 259, 276 Kuriyama M, 73, 87 Kuroda S, 368, 385 Kuttner F, 62, 68, 80 Kvietikova I, 106, 120, 130, 136, 155, 171 Kwast KE, 120, 137 Kwee L, 232, 236, 293, 309

413 L LaBanca F, 44, 56 Labosky PA, 75, 88, 225, 228, 235, 238 Labouesse M, 62, 79 Labow MA, 232, 236, 293, 309 Lacey E, 35, 52 Lacy E, 317, 335 Laget MP, 111, 133 Lai C, 230, 236 Lai DT, 354, 361 Lai E, 68, 83 Lai Z, 194, 198 Lake RJ, 145, 150 Lakkis MM, 231, 237 Lamb BT, 41, 55 Lambeth JD, 120, 136 Lambie EJ, 62, 79 Lamers WH, 210, 237 Lametschwandtner A, 161, 174 Lampugnani MG, 295, 309 Lander LE, 112, 134 Land SC, 140, 141, 148, 149 Lane AA, 104, 128, 288, 307 Lane WS, 112, 134 Laney JD, 111, 133 Langeveld A, 192, 198 Langille BL, 228, 236 Langley KE, 299, 311 Lang P, 180, 185, 196 Lanzino G, 352, 361 Lapidot T, 353, 361 Larbi KY, 297, 311 Lardelli M, 210, 215, 222, 238 Larhammar D, 261, 277 Larkin DW, 376, 388 LaRochelle WJ, 76, 89 Larochette N, 348, 360 LaRosa M, 97, 125 Larsson S, 50, 58 Lasman H, 354, 361 Lassar AB, 63, 81, 180, 185, 200, 213, 238 Lassegue B, 120, 136 Lathan CB, 350, 361 Latres E, 111, 133

414 Latvala RD, 321, 335 Lauer RM, 290, 308 Lauer SJ, 329, 338 Laughner E, 155, 158, 160, 161, 168, 169, 171, 173, 174, 176, 177 Lau KK, 62, 79 Laverriere AC, 69, 85, 194, 198 Laverty T, 12, 13, 19 Lavitrano M, 39, 54 Lawler AM, 155, 160, 161, 171 Lawlor P, 354, 361 Lawn RM, 323, 324, 336 Lawrence PA, 183, 194, 196, 198 Lawson EE, 106, 130 Lawson KA, 59, 75, 79, 88 Lazebnik Y, 343, 344, 346, 359 Lazzaro D, 71, 86 Lebeche D, 76, 78, 89, 99, 126 Lebedeva NV, 382, 390 LeBlanc A, 355, 363 LeBoueuf RC, 315, 334 Lechene C, 282, 303 Lechleider RJ, 230, 238 Leclerq B, 50, 57 Leder P, 48, 57 Ledoux S, 343, 359 Lee C, 373, 378, 387, 388 Lee CC, 375, 376, 387, 388 Lee D, 326, 336, 337 Lee JK, 193, 197 Lee KF, 231, 237 Lee KH, 63, 81, 184, 199, 204, 213, 236, 237, 263, 277 Lee KS, 352, 361 Lee L, 62, 80 Lee M, 76, 89, 101, 126 Lee MK, 27, 32 Lee PJ, 155, 159, 171 Lee RK, 204, 238, 271, 279 Lee S, 113, 135 Lee SC, 355, 363 Lee SH, 225, 238 Lee SM, 48, 57 Lee T, 66, 82, 93, 94, 124 Lee WR, 266, 277 Lee Y, 179, 184, 192, 199, 201, 326, 337

Author Index Leff T, 326, 337 Lehner CF, 2, 18 Lehninger Al, 158, 173 Lehrach H, 192, 199 Lehti K, 301, 312 Leiden JM, 69, 85, 191, 192, 198, 206, 219, 236 Lei L, 71, 74, 86 Leinwand M, 76, 89, 101, 126 Leivtan D, 27, 32 Lejavardi N, 156, 170, 172 Le M, 373, 387 LeMeur M, 47, 57, 227, 236 Lemieux M, 71, 86 Lemke G, 230, 236 Lemmon V, 282, 304 Lenardo MJ, 357, 363 Leonard A, 266, 278 Leonard MW, 69, 84 Leone G, 111, 133 Lepourry L, 50, 57 Leptin M, 192, 200 Letarte M, 291, 308 Letterio JJ, 289, 307 Leung MK, 155, 171 Leung SW, 106, 107, 130, 131, 154, 155, 156, 160, 161, 168, 170, 171, 172 Levak-Frank S, 328, 337 Levine JD, 378, 388 Levin M, 244, 253 Lev M, 242, 252 Levy AP, 113, 135 Levy Y, 157, 169, 173 Lewis C, 145, 150, 185, 199 Lewis CE, 281, 303 Lewis PM, 71, 74, 87 Lew RA, 368, 385 Lewy AJ, 368, 385 Leysens NJ, 72, 87 Lhotak V, 103, 128, 287, 306 L’Huillier PJ, 50, 57 Liakopoulos D, 115, 135 Liao E, 261, 277 Liao EC, 214, 237, 263, 270, 277, 279 Liao W, 271, 279

Author Index Liao X, 76, 89, 99, 126 Liapis H, 69, 84 Liberthson RR, 242, 252 Li BY, 230, 238 Li C, 289, 308 Li DY, 289, 307 Li E, 35, 53, 249, 254 Liebermann DA, 358, 364 Liebhaber SA, 40, 54 Liebovitch L, 140, 144, 148 Li FN, 111, 132 Liggett SB, 45, 56, 356, 363 Li H, 322, 335 Li J, 259, 276 Li M, 355, 362 Lim L, 69, 83, 84, 104, 128 Lim M, 168, 169, 176 Limts TJ, 179, 184, 201 Lin CH, 296, 310 Lin CR, 72, 87, 250, 254 Lin CS, 69, 84 Lindebro MC, 371, 386 Lindenbaum MH, 69, 84 Linder V, 301, 312 Lindquist S, 8, 9, 18 Lin F, 145, 150 Ling ZD, 354, 362 Lin K, 62, 79 Lin M, 115, 135 Lin Q, 69, 85, 192, 199, 215, 216, 217, 219, 236, 237, 238 Lin S, 42, 55, 267, 268, 271, 274, 278, 279, 280 Linton MF, 323, 336 Lints T, 257, 276 Lints TJ, 63, 80, 81, 184, 199, 206, 237 Lin WH, 192, 199 Lin YJ, 145, 150 Liotta LA, 16, 19, 20 Lipshitz HD, 193, 201 Lipson KE, 286, 306 Li Q, 41, 55 Li R, 63, 81, 184, 199, 216, 237 Li S, 115, 135 Lissy NA, 358, 364 Lisztwan J, 113, 135, 169, 176

415 Litingtung Y, 71, 74, 86 Litt L, 141, 149 Littlewood TD, 47, 57 Litvin O, 68, 83 Liu C, 111, 133 Liu F, 72, 87, 230, 235, 250, 254 Liu H, 288, 307 Liu J, 74, 76, 88, 89, 99, 126 Liu L, 348, 360 Liu P, 289, 308 Liu X, 225, 238, 343, 359, 382, 390 Liu XL, 24, 32 Liu Y, 69, 84, 158, 159, 173, 379, 382, 388, 390 Liu YC, 274, 280 Livingston DJ, 225, 235 Livingston DM, 105, 129, 155, 171 Li WP, 354, 361 Li X, 43, 56 Li XJ, 355, 362 Li Y, 37, 53 Llimargas M, 66, 81, 93, 98, 124, 125 Lockwood WK, 179, 180, 182, 183, 196 Logothetis DE, 369, 385 Loh DY, 350, 360 Lohi L, 301, 312 Lois AF, 118, 136 Lo J, 155, 160, 171 Lok CN, 155, 172 Lonai P, 353, 361 Lonergan KM, 113, 135 Longenecker G, 289, 307 Longmore W, 69, 83 Long Q, 274, 280 Lorenz JN, 45, 56 Lorenz L, 382, 390 Lorenzo HK, 348, 360 Loros JJ, 379, 388 Losordo DW, 104, 128, 165, 174 Losse B, 156, 170, 172 Lough J, 319, 335 Lou LJ, 70, 86 Loverro L, 230, 236 Lovett DH, 301, 312 Lowe L, 247, 248, 253, 254

416

Author Index

Lowe LA, 244, 248, 253 Lowery PL, 375, 387 Lowe SW, 104, 128, 349, 360 Lu B, 191, 192, 198, 199, 217, 237 Lucas RJ, 380, 389 Lufkin T, 43, 56 Luis de la Pompa J, 297, 311 LuK AY, 145, 150 Lu L, 282, 304 Lu LH, 230, 236 Lu M, 250, 254 Lu MF, 72, 87 Lu MM, 69, 70, 84, 85, 191, 192, 198, 199, 206, 219, 236, 237 Lumsden A, 161, 174 Luna JD, 166, 175 Lupu F, 295, 309 Luria V, 353, 361 Luse DS, 69, 83, 84 Lusis AJ, 314, 316, 322, 327, 334, 335, 337 Lutachg A, 343, 359 Lutgens E, 295, 309 Luther T, 293, 309 Lu TT, 288, 296, 307, 310 Lutz PL, 140, 141, 144, 148, 149, 150 Lu X, 28, 32 Luyckx VA, 50, 57 Lyapina S, 112, 134 Lyapina SA, 112, 113, 134 Lyght M, 289, 307 Lymboussaki A, 102, 127, 283, 304 Lyons GE, 179, 184, 201 Lyons I, 63, 80, 81, 184, 199, 206, 216, 237 Lyons KM, 225, 236

M Maattot V, 295, 309 MacArthur CA, 69, 85 MacFarlane M, 348, 360 Mackem S, 76, 89 Mackman N, 293, 309 MacLennan H, 263, 277 MacNeill C, 69, 85, 194, 198

Madri JA, 282, 283, 284, 285, 288, 289, 291, 293, 296, 298, 299, 301, 302, 303, 305, 306, 307, 308, 310, 311, 312 Ma E, 143, 150 Maeda N, 35, 38, 40, 53, 319, 322, 326, 327, 335, 336, 337 Maeno M, 257, 275 Magdaleno SM, 45, 56 Maher ER, 113, 117, 135, 169, 176 Mahooti-Brooks N, 282, 303 Mahooti S, 283, 284, 288, 291, 293, 296, 299, 302, 305, 307, 311, 312 Maisonpierre PC, 102, 103, 127, 128, 285, 287, 305, 306 Majercak J, 383, 390 Makhinson M, 45, 56 Makino Y, 111, 117, 132, 135, 169, 177 Mak TW, 228, 236, 297, 311, 349, 360 Malicki J, 216, 226, 237, 238 Malkewitz J, 121, 137 Malpel S, 76, 78, 89 Maltepe E, 106, 120, 131, 136, 158, 160, 168, 173, 174, 175 Maly FE, 120, 136 Man B, 371, 386 Manchikalapudi S, 167, 175 Mancini M, 348, 360 Mancino V, 41, 55 Manganaro T, 230, 238 Mango SE, 62, 79 Maniatis T, 192, 196 Mannick JB, 348, 360 Manning G, 66, 81, 93, 95, 98, 124 Manning GE, 92, 101, 103, 123 Mann R, 36, 53 Manolagas SC, 358, 364 Mansour SL, 282, 303, 304 Manuel L, 141, 149 Marchuk DA, 286, 291, 293, 306, 308 Marcigliano A, 329, 338 Marden MC, 118, 136 Marengere L, 228, 236 Margolis RL, 355, 362 Margottin F, 113, 134 Marie P, 106, 130, 154, 156, 170

Author Index Markel DS, 291, 308 Marker PC, 75, 88 Markham BE, 192, 199 Mark M, 227, 236 Marks RM, 287, 306 Markwald RR, 208, 224, 229, 236, 237, 238 Marletta MA, 121, 137 Marmer BL, 299, 311 Marmorato A, 50, 58 Maron BJ, 180, 185, 200 Marotti KR, 324, 336 Marszalek JR, 248, 253 Marti HH, 120, 136, 154, 156, 170, 170, 172 Martin A, 71, 86 Martin C, 164, 174 Martin DI, 40, 54, 259, 276 Martin GR, 76, 78, 88, 89, 244, 253 Martin J, 230, 238 Martin JF, 72, 87, 250, 254 Martinou JC, 350, 360 Martin RJ, 368, 385 Martin-Zanca D, 285, 306 Marty SD, 274, 280 Marty T, 96, 125 Marzo I, 348, 360 Mascrez B, 47, 57 Mash D, 141, 149 Mason I, 76, 89 Mas P, 381, 389 Massague J, 75, 88, 230, 235, 291, 308 Mass RL, 2, 17 Masucci-Magoulas L, 326, 337 Masuda H, 283, 305 Mather C, 37, 53 Mathews LM, 229, 238 Mathieu CE, 120, 136, 158, 173 Matis LA, 357, 363 Matsouka R, 242, 252 Matsubara C, 376, 387 Matsuda Y, 169, 176 Matsui D, 76, 89 Matsumoto M, 111, 113, 133 Matsunami H, 222, 237 Matthews B, 105, 129

417 Matzuk MM, 222, 229, 235, 238, 293, 308 Maulik N, 167, 175 Maxwell P, 106, 130, 155, 160, 168, 171 Maxwell PH, 105, 113, 117, 129, 135, 155, 168, 169, 171, 176 Mayer B, 122, 137 Mayer EL, 257, 275 Mazure NM, 155, 171 Mbamalu G, 288, 307 McAllister KA, 291, 308 McArthur MJ, 357, 364 McCague L, 24, 32 McCarthy JB, 299, 312 McCarthy MJ, 286, 306 McCarthy TL, 282, 303 McClain J, 102, 127 McClain S, 285, 305 McClanahan JL, 299, 311 McClure MH, 73, 87 McCormick F, 111, 133 McCormick SPA, 329, 338 McCormick MK, 291, 308 McCutcheon K, 355, 362 McDonald DM, 102, 127 McDonald JA, 225, 235 McDonald JD, 375, 387 McDonald NJ, 16, 20 McDonough B, 180, 185, 196 McFadden DG, 217, 236 McFarlane RJ, 42, 55, 268, 279 McGhee JD, 62, 64, 79, 80, 81 McGinnis W, 15, 19, 70, 85 McGlade CJ, 193, 201 McGrath J, 3, 18, 248, 254 McInnes L, 213, 233, 237 McKenna PJ, 296, 310 McKeon F, 113, 134 McKeown M, 47, 57, 104, 128 McKinney LA, 229, 235, 290, 308 McKinney M, 354, 361 McKinnon WC, 291, 308 McKnight SL, 169, 176 McMahon AP, 2, 17, 48, 57, 69, 71, 73, 74, 83, 87, 88 McMahon G, 286, 306

418 McMurtrey A, 284, 305 McNeil GP, 383, 390 McVey JH, 70, 86 McWhir J, 36, 53 McWilliams R, 162, 163, 174 Mead PE, 257, 275 Mechler BM, 15, 19 Medzhitov R, 2, 18 Mee E, 354, 361 Meier A, 222, 236 Meier E, 46, 47, 57 Meiler SE, 216, 238 Meisler MH, 266, 278 Meister M, 369, 385 Melamed E, 354, 362 Melchior GW, 324, 336 Melillo G, 106, 130, 159, 167, 173 Mello C, 24, 32 Mello CC, 29, 33, 61, 79 Melton DA, 257, 275 Melton DW, 35, 53 Menaker M, 369, 380, 385, 389 Meneses JJ, 59, 79 Meng A, 42, 55, 268, 279 Meno C, 60, 78, 79, 244, 253 Menon AS, 355, 362 Mentink MM, 207, 208, 236 Ment LR, 283, 284, 305 Mercer B, 70, 85 Mercier JC, 50, 57 Merewether LA, 285, 306 Merlino G, 76, 89 Mertens PR, 301, 312 Merwin JR, 282, 289, 303, 308 Metcalfe T, 98, 125 Methot N, 111, 132 Mett IL, 353, 361 Metzger D, 47, 57 Metzger RJ, 65, 78, 81, 92, 98, 123, 145, 150 Metzstein MM, 25, 26, 32, 145, 151 Meyer-Bernstein EL, 145, 150 Meyer BJ, 27, 33 Meyer D, 230, 237 Meyer S, 285, 306 Meyers EN, 244, 253

Author Index Miano JM, 70, 85 Miao HQ, 284, 305 Miao QX, 348, 360 Michel JJ, 112, 134 Michel PP, 355, 363 Michels CL, 155, 171 Michelson AM, 192, 197 Mickanin C, 228, 237 Mihalcik V, 284, 305 Mijalski T, 245, 253 Mikawa T, 210, 236 Milan M, 12, 19 Milhorn DE, 106, 130 Miller AD, 36, 53 Miller CC, 168, 175 Miller LM, 62, 80 Millhorn DE, 106, 130 Milton SL, 141, 149 Mims IP, 266, 277 Mimura J, 107, 115, 131, 160, 174 Min H, 76, 89, 99, 126 Minoo P, 71, 86 Minowada G, 78, 89 Minowa O, 283, 304 Miranda B, 76, 89, 99, 126 Missotten M, 350, 360 Mitchell D, 315, 334 Mitchell R, 37, 53 Miura H, 72, 87 Miyagawa-Tomita S, 72, 87, 220, 238 Miyake S, 376, 387 Miyamoto Y, 380, 389 Miyata M, 323, 336 Miyazaki J, 220, 238 Miyazaki M, 16, 20 Miyazano K, 289, 308 Mizuno Y, 354, 362 Mizutani Y, 37, 53 Mjaatvedt CH, 224, 237, 238 Moak JP, 180, 185, 200 Mochida K, 60, 78, 79, 244, 253 Mochizuki H, 354, 362 Model P, 42, 55 Mogi M, 354, 362 Mohideen MA, 216, 238 Mohler J, 69, 83

Author Index Mohr S, 25, 32, 353, 361 Mohsenin A, 143, 144, 150 Mohun TJ, 184, 197 Molkentin JD, 69, 70, 85, 192, 197, 199, 219, 237 Mollard R, 70, 86 Moller JH, 243, 252 Mombaerts P, 39, 54 Momoi MY, 73, 88 Momoi T, 73, 88 Monaghan AP, 62, 68, 80 Moncada S, 76, 89, 99, 126 Monson EK, 105, 129, 130 Montagne J, 97, 125 Montell DJ, 66, 82, 93, 94, 124 Montesano R, 282, 303 Montgomery CA, 357, 364 Moons D, 62, 79 Moons L, 102, 106, 127, 130, 155, 160, 168, 171, 282, 293, 295, 304, 309 Moore AW, 213, 233, 237 Moore MS, 145, 150 Moore MW, 230, 236, 282, 304 Moore RY, 369, 385 Moorman AF, 210, 237 Mo R, 74, 88, 99, 126 Moran LA, 43, 56 Moreadith RW, 48, 57 Moreland RJ, 112, 134 Mori H, 354, 362 Morisseau BA, 62, 80 Morita M, 160, 174 Morita T, 158, 159, 173 Morrisey EE, 69, 70, 84, 85, 191, 192, 198, 199, 206, 219, 236, 237 Morris SM, 167, 175 Morrow A, 315, 334 Morsink F, 354, 362 Mosher J, 66, 82, 93, 123 Mossberg B, 245, 248, 253 Moss JE, 358, 364 Motegi Y, 72, 87 Motoyama J, 74, 88, 99, 126 Motoyama N, 350, 360 Motulsky AG, 326, 337 Moulder G, 30, 32

419 Mouzeyan A, 333, 339 Moxley M, 69, 83 Moynahan ME, 40, 54 Mueller C, 69, 85, 194, 198 Muijtjens M, 380, 389 Mukasa T, 73, 88 Muliken JB, 286, 293, 306 Muller JE, 367, 385 Muller M, 293, 309 Muller U, 36, 38, 48, 51, 53 Muller WA, 296, 297, 310, 311 Mulligan EL, 24, 32 Mulligan R, 36, 53 Mulligan RC, 36, 53 Mullins MC, 213, 217, 219, 222, 235, 236, 237, 263, 271, 272, 277, 279 Munoz M, 380, 389 Murakami K, 257, 275 Muramatsu T, 37, 53 Murillo MP, 184, 197 Murone M, 2, 17 Murphy PJ, 368, 380, 385, 389 Murray AB, 321, 335 Murray JC, 244, 253 Murray RW, 324, 336 Murrell J, 291, 308 Musci TJ, 257, 275 Mushegian AR, 25, 33 Musso T, 106, 130, 159, 173 Mustoe TA, 293, 309 Mustonen T, 102, 127, 283, 304 Mu X, 354, 362 Myers EM, 145, 150 Myers M, 371, 386 Myers MP, 373, 386, 387 Myint Z, 72, 73, 87

N Nagao M, 107, 111, 131 Nagase H, 298, 311 Nagase T, 382, 390 Nagata S, 357, 363 Nagatsu T, 354, 362 Nagel RL, 50, 58 Nagy A, 102, 127, 282, 304

420 Nagy JA, 283, 304 Nakagawa K, 16, 20 Nakahara Y, 72, 87 Nakajima N, 16, 20 Nakajima Y, 229, 237 Nakamichi I, 111, 113, 133 Nakamura H, 229, 237 Nakamura R, 119, 136 Nakamura S, 368, 385 Nakanishi T, 43, 55 Nakao S, 302, 312 Nakashima Y, 320, 335 Nakatake Y, 76, 89, 99, 126 Nakayama K, 111, 113, 133, 350, 360 Nakayama KI, 350, 360 Nambu J, 105, 129 Nambu JR, 109, 131 Narabayashi H, 354, 362 Narita N, 69, 85, 208, 219, 237 Narula J, 356, 363 Narula N, 356, 363 Na S, 349, 360 Nash GB, 296, 310 Nasmyth F, 112, 134 Nasmyth K, 112, 134 Nathke IS, 295, 309 Navarro P, 293, 309 Nechiporuk T, 231, 236 Neckers LM, 106, 131, 168, 175 Neeman M, 105, 106, 129, 130, 155, 160, 168, 171 Negishi I, 350, 360 Nejfelt MK, 156, 172 Nellen D, 96, 125 Nelson WJ, 295, 309 Nemaceck J, 183, 184, 201 Nemer G, 192, 198 Nemer M, 70, 85, 192, 197, 198, 200 Nervi C, 71, 86 Netea MG, 327, 337 Neuberger MS, 45, 56 Neubuser A, 78, 89 Neufeld G, 284, 305 Neufield TP, 8, 19 Neuhauss SC, 216, 226, 237, 238 Nevins JR, 111, 133

Author Index Newby LM, 383, 390 Newman CS, 72, 87, 179, 184, 199 Newman PJ, 296, 310 Newman W, 289, 308 Ng JK, 329, 338 Nguyen H, 330, 338 Nguyen HB, 378, 388 Nguyen HQ, 76, 89 Nguyen SV, 155, 171 Nguyen VH, 213, 237 Nicholls LG, 168, 176 Nichols AV, 323, 336 Nichols NR, 354, 361 Nicholson DW, 348, 355, 360, 362, 363 Nicholson IC, 45, 56 Nicosia RF, 282, 303 Niederle N, 156, 170, 172 Niederreither K, 215, 217, 237 Nielsen HC, 71, 86 Nigam V, 192, 200 Nikaido M, 263, 277 Nilsson GE, 141, 144, 149, 150 Nirenberg M, 63, 80 Nishimoto H, 302, 312 Nishimoto I, 355, 363 Nishimura D, 290, 308 Nishimura M, 368, 385 Nisson GE, 140, 148 No D, 46, 47, 57 Noda T, 283, 304 Nogawa H, 76, 89 Noguchi PD, 98, 125 Noguchi T, 72, 73, 87 Noji S, 60, 78, 79, 244, 253 Nojyo Y, 72, 87 Nonaka S, 248, 253, 254 Norris ML, 106, 130 Nossal GJV, 353, 361 Noteborn M, 46, 47, 57 Nourshargh S, 297, 311 Novak R, 358, 364 Nuber U, 112, 133 Nucifora FC, 225, 235 Numayama-Tsurtua K, 158, 173 Nuotio I, 331, 338 Nusbaum C, 330, 338

Author Index

421

Nussbaumer U, 96, 125 Nusse R, 2, 17 Nussey G, 274, 280 Nusslein-Volhard C, 192, 200, 204, 213, 217, 219, 222, 235, 236, 237, 238, 243, 252, 261, 265, 271, 272, 276, 277, 279 Nuyens D, 295, 309 Nystrom G, 66, 82, 93, 123

O Oates AC, 214, 237, 263, 270, 273, 277, 277, 279, 280 O’Brien S, 169, 177 Ocampo CJ, 167, 175 O’Connell A, 326, 329, 337, 338 O’Connell S, 72, 87, 250, 254 Ocorr KA, 179, 180, 182, 183, 184, 196, 201 Oda H, 98, 125 O’Dell TJ, 45, 56 Odenthal J, 213, 217, 219, 222, 235, 236, 237, 271, 272, 279 Odenwald WF, 98, 125 O’Farrell PH, 120, 137 Offen D, 354, 362 Ogg S, 27, 33, 62, 80 Ogldberg MA, 113, 135 Ogunshola O, 284, 305 Ogura T, 106, 131 Ohishi S, 60, 78, 79, 244, 253 Ohlmeyer JT, 111, 132 Ohmori Y, 37, 53 Oh SP, 249, 254 Ohsuga J, 321, 335 Ohta T, 112, 134 Ohuchi H, 72, 76, 87, 89, 244, 253 Oikawa T, 75, 88 Okabe M, 39, 43, 54, 55, 67, 82 Okada N, 155, 171 Okada T, 382, 390 Okada Y, 248, 253, 254 Okamoto K, 117, 135 Okamoto N, 166, 175 Okamoto T, 355, 363

Okamura H, 375, 376, 387 Okkema PG, 62, 63, 79, 80, 81 Okumura J, 37, 53 Olanow CW, 354, 362 Oldmixon EH, 102, 127 Olsen BR, 286, 293, 306 Olson EN, 63, 69, 70, 80, 85, 179, 184, 185, 192, 197, 199, 200, 208, 215, 216, 217, 219, 236, 237, 238 Olson MV, 41, 55 Omary R, 352, 361 Omichinski JG, 69, 84 Ona VO, 355, 362 Ong RC, 257, 275 Oosthuyse B, 295, 309 Oostra BA, 291, 308 Oppenheim RW, 349, 360 Orci L, 282, 303 Ord VA, 315, 334 Orikn SH, 257, 275 Orimer J, 70, 86 Orioli D, 230, 236 Orkin SH, 69, 84, 85, 255, 257, 275, 276 Ormsby I, 229, 238 Ornitz DM, 48, 57 O’Rourke JF, 109, 131 Orr-Weaver TL, 2, 18 Osda H, 259, 276 O’Shea KS, 230, 236, 282, 287, 304, 306 Oshima H, 289, 307 Oshima RG, 40, 55 Oshiro T, 66, 82, 93, 123 Osmanian C, 104, 128 Oswell GM, 332, 339 Ousley A, 373, 387 Overbeek P, 244, 245, 248, 253 Overdier DG, 69, 83 Ozaki H, 166, 175 Ozawa R, 375, 376, 387

P Pablo J, 141, 149 Padgett RW, 2, 17, 191, 199 Pagano M, 111, 132, 133 Pahl HL, 111, 132

422 Paigen B, 314, 315, 319, 323, 334, 335, 336 Pajukanta P, 331, 338 Pajusola K, 102, 127, 283, 304 Pak B, 147, 155, 166, 169, 171 Pak J, 296, 310 Palazzolo MJ, 382, 390 Palinski W, 315, 334 Palmeri ML, 293, 309 Palmer LA, 155, 159, 167, 171 Palmiter RD, 39, 46, 47, 54, 57 Pan D, 12, 19 Pandey A, 287, 306 Pandey P, 356, 363 Pandiella A, 285, 306 Pandolfi PP, 69, 84 Panettieri RA, 296, 310 Pankratz MJ, 67, 83 Pan ZQ, 112, 134 Papadopoulos N, 102, 127, 285, 305 Papalopulu N, 70, 85, 86, 184, 197 Papoff G, 357, 363 Paradi E, 266, 277 Paradis S, 62, 80 Parameswaran M, 204, 205, 238 Parangi S, 104, 128 Parfitt AM, 358, 364 Parichy DM, 272, 280 Pariente F, 73, 87 Parikh V, 373, 387 Parisi J, 352, 361 Parker L, 271, 279 Park J, 155, 172 Park M, 145, 150, 179, 180, 182, 183, 184, 185, 196, 199, 201 Park N, 193, 200 Park TS, 167, 175 Park WY, 76, 89, 99, 126 Parlow AF, 3, 18 Parmacek MS, 69, 70, 84, 85, 191, 192, 198, 199, 206, 219, 236, 237 Parsons LM, 63, 80, 81, 184, 199, 206, 216, 237 Partin JS, 327, 337 Pascoe CJ, 354, 362 Pascoe S, 72, 87

Author Index Pasquale EB, 288, 307 Passaniti A, 157, 172 Passantino R, 106, 130, 154, 156, 170 Pastnik A, 266, 277 Pasyk KA, 286, 293, 306 Paszty C, 322, 335 Patan S, 285, 305 Patel-King RS, 220, 236, 293, 308 Patient R, 192, 198, 205, 221, 237 Patient RK, 259, 276 Patil SR, 290, 308 Patterson GI, 62, 80 Patton EE, 112, 134 Pauli IG, 220, 237 Paul SM, 354, 362 Paulson KE, 69, 83 Pause A, 113, 115, 135 Pavletich NP, 115, 135, 169, 176 Paw B, 261, 277 Paw BH, 42, 55, 214, 237, 268, 270, 273, 279, 280 Pawling J, 102, 127, 282, 304 Pawlowski M, 156, 170, 172 Pawson T, 103, 128, 287, 288, 306, 307 Pearlman JD, 104, 128 Pease W, 180, 185, 200 Pedersen RA, 59, 79 Peel D, 15, 19 Pe’er J, 166, 175, 283, 305 Pei D, 299, 311 Pek M, 141, 149 Pek-Scott M, 141, 149 Pelegri F, 219, 236 Pelham H, 266, 277 Pelina-Parker M, 299, 311 Pellegatta F, 296, 310 Pelletier GJ, 69, 84 Peng J, 283, 304 Pennarun G, 249, 254 Penney JB, 355, 362 Penninger JM, 349, 360 Peoples WM, 243, 252 Pepicelli CV, 71, 74, 87 Perdew GH, 169, 176 Pereira FA, 217, 237 Perez-Pinzon MA, 141, 142, 144, 149

Author Index Pericak-Vance MA, 291, 308 Perkins D, 74, 88 Perlmutter RM, 49, 57 Perrimon N, 2, 6, 17, 18 Perry AC, 39, 54 Perry M, 208, 237 Perry MD, 43, 56 Perry VL, 140, 148 Pertuz MF, 118, 136 Peterfy M, 332, 339 Petersen G, 370, 371, 386 Peters JM, 111, 132 Peters K, 76, 89, 99, 126, 271, 279 Peterson EP, 348, 360 Peterson KR, 41, 55 Peterson RS, 69, 83 Petit C, 380, 389 Petti AA, 381, 389 Pevny L, 69, 84, 260, 276 Pexieder T, 227, 236 Peyrieras N, 217, 237 Pezioso VR, 68, 83 Pflug B, 368, 385 Phan J, 332, 339 Piccolo S, 217, 237 Pickart CM, 111, 133 Pieczek A, 104, 128 Piedad O, 257, 275 Piedrahita JA, 319, 335 Pierpont ME, 290, 308 Pieterse JJ, 274, 280 Pietra C, 350, 360 Pignot-Paintrand I, 295, 309 Pikielny C, 383, 390 Pinsky L, 355, 362 Pinter E, 288, 291, 293, 307 Pinto LH, 45, 56, 375, 387 Pitas RE, 329, 338 Pitot HC, 15, 19 Pittendrigh CS, 369, 373, 385, 387 Pizzey JA, 259, 276 Placzek M, 99, 126 Plaehn EG, 38, 40, 53 Plasterk RHA, 24, 28, 30, 32 Platt KA, 74, 88 Plautz JD, 382, 383, 390

423 Plotkin DJ, 35, 36, 52 Plowman GD, 286, 306 Plump AS, 314, 315, 319, 320, 322, 327, 334, 335, 337 Poellinger L, 107, 111, 115, 117, 131, 132, 135, 169, 177, 371, 386 Poelmann R, 295, 309 Poelmann RE, 69, 85, 161, 174, 194, 198, 207, 208, 236 Pointu H, 295, 309 Polakis P, 111, 133 Polites HG, 324, 336 Pollefeyt S, 102, 126, 282, 304 Pollock RA, 40, 54 Polluck AS, 301, 312 Polveriini PJ, 287, 306 Ponce J, 184, 197 Ponka P, 155, 172 Ponting CP, 105, 129 Popov AV, 45, 56 Porcella A, 69, 83 Porteous ME, 291, 308 Porwol T, 157, 169, 173 Postlethwait JH, 214, 237, 261, 263, 270, 276, 277, 279 Post M, 74, 76, 88, 89, 99, 126 Potter CJ, 15, 19 Potter J, 228, 236, 349, 360 Potter SS, 244, 248, 249, 253, 254 Poulton J, 106, 130 Powell-Braxton L, 282, 304, 321, 335 Powell DR, 155, 171 Poyart CC, 118, 136 Poyton RO, 105, 106, 108, 118, 119, 120, 129, 137 Prandini MH, 295, 309 Prasher DC, 23, 32 Pratico D, 296, 310 Pratt BM, 282, 303 Pratt S, 263, 277 Pratt SJ, 214, 237, 263, 270, 272, 273, 277, 277, 279, 280 Pressman C, 72, 87, 250, 254 Prevost MC, 348, 360 Prezioso VR, 62, 68, 80, 83 Price BM, 257, 275

424

Author Index

Price J, 227, 238 Price JL, 371, 373, 379, 386, 388, 389 Price M, 71, 86 Priess JR, 27, 33, 61, 64, 79, 81 Probst I, 119, 136 Program AE, 193, 201 Propp SS, 355, 362 Prosser J, 374, 387 Prusty D, 282, 303 Przedborski S, 354, 362 Puck JM, 357, 363 Pugh CW, 105, 106, 107, 109, 111, 113, 117, 120, 129, 130, 131, 135, 155, 156, 168, 169, 171, 172, 176 Pulst SM, 231, 236 Purcell-Huynh DA, 320, 335

Q Qian Z, 373, 380, 387, 389 Qiao JH, 316, 322, 327, 334, 335, 337 Qin Y, 285, 293, 305 Quintin S, 62, 79 Quiring R, 374, 387 Qui Y, 167, 175, 217, 237

R Rabbitts TH, 259, 276 Rachidi M, 383, 390 Radice GL, 222, 237 Radziejewski C, 103, 127, 285, 305, 306 Raff MC, 343, 346, 359 Raftery LA, 2, 17 Raible DW, 270, 279 Raine CS, 357, 363 Raines ES, 320, 335 Rainger GE, 296, 310 Rakic P, 349, 350, 360, 361 Rall SC, 326, 337 Ralph MR, 369, 385 Ramain P, 192, 200 Ramirez S, 156, 172 Ramirez-Weber FA, 111, 133 Ramsby G, 352, 361 Ramsdell AF, 229, 237

Rancourt DE, 71, 86 Rand MD, 145, 150 Ranganayakulu G, 63, 80, 185, 200 Ranger AM, 228, 237 Ransom D, 261, 263, 277 Ransom DG, 272, 279 Rao A, 352, 361 Rao Y, 104, 128 Raska I, 15, 19 Rasper D, 355, 362 Rastegar S, 214, 235 Ratcliffe P, 106, 130, 155, 160, 168, 171 Ratcliffe PJ, 105, 106, 107, 108, 109, 113, 117, 119, 120, 129, 130, 131, 135, 155, 156, 157, 158, 168, 169, 171, 172, 173, 176 Ratner L, 358, 364 Rattan V, 296, 310 Rauch GJ, 222, 236 Rawls JF, 272, 280 Rayburn H, 206, 217, 220, 222, 236, 237, 238, 293, 308 Razvi S, 104, 128 Read RD, 6, 18 Ready KA, 296, 310 Reardon B, 24, 32 Rebay I, 13, 19 Rebrikov D, 353, 361 Recht LD, 368, 385 Reddick RL, 319, 322, 335 Reddy P, 370, 386 Reecy JM, 184, 200 Reed JC, 346, 348, 360 Reeves M, 354, 361 Rehberg EF, 324, 336 Rehm J, 12, 19 Rehnmark S, 331, 339 Reich A, 95, 125 Reichman-Fried M, 66, 67, 82, 92, 93, 97, 99, 101, 123, 124 Reinstein E, 111, 133 Reiter JF, 205, 221, 237 Reiter R, 72, 87 Ren Y, 106, 131 Reppert SM, 369, 376, 378, 383, 385, 387, 388, 390

Author Index Reth M, 47, 57 Reue K, 314, 315, 327, 328, 330, 331, 332, 334, 337, 338, 339 Reuter R, 61, 68, 79 Reynolds TC, 3, 18 Rhyu MS, 192, 200 Rice ME, 142, 149 Richardson C, 40, 54 Richard V, 113, 134 Richfield EK, 354, 362 Richiusa P, 357, 363 Riddle DL, 27, 33 Rideout WM, 39, 54 Riechmann V, 192, 200 Riederer P, 354, 362 Riesterer C, 47, 57 Rigby M, 355, 363 Rigby PW, 43, 55 Riley O, 217, 237 Rimm DL, 296, 311 Ring BD, 76, 89 Rioseco-Camacho N, 169, 177 Rippe RA, 40, 54 Risau W, 101, 102, 126, 127, 169, 176, 282, 285, 287, 293, 304, 305, 307, 309 Ritchie WA, 36, 53 Ritter DG, 242, 252 Robash M, 370, 371, 373, 381, 383, 385, 386, 387, 389, 390 Robb L, 63, 81, 184, 199, 216, 237 Roberson P, 358, 364 Robert B, 287, 307 Robertis EM, 270, 279 Roberts AB, 289, 307 Roberts E, 248, 253 Robertson E, 246, 247, 248, 253 Robertson EJ, 35, 53, 220, 237, 249, 254 Robertson GS, 355, 363 Robertson JE, 60, 78, 79 Robey PG, 289, 307 Robinson CS, 282, 303 Robitaille L, 192, 197 Robitaille LL, 192, 197 Rochlin I, 50, 58

425 Rockwell S, 230, 238 Rodan A, 62, 79 Rodaway A, 205, 221, 237 Roder J, 46, 47, 57 Rodriguez I, 39, 54, 350, 360 Roebroek AJ, 220, 237 Roelen BA, 75, 88 Roe R, 107, 131, 155, 156, 171, 172 Roger G, 249, 254 Rojas J, 102, 127 Rolfs A, 106, 130, 154, 155, 170, 171 Romanic AM, 299, 312 Ronai Z, 112, 134 Ronanno E, 282, 303 Rooke J, 12, 19 Rooke JE, 2, 18 Rorth P, 12, 19 Rosa F, 217, 237, 261, 277 Rosbash M, 105, 129, 145, 150, 373, 377, 378, 380, 382, 387, 388, 389, 390 Rosen A, 348, 360 Rosenblatt M, 69, 84, 259, 276 Rosenfeld MG, 36, 46, 47, 53, 57, 72, 87, 250, 254 Rosenfield K, 104, 128 Rosenthal A, 2, 17 Rosenthal M, 141, 149 Rosenthal N, 180, 198, 227, 236 Roshbash M, 145, 150 Rossant J, 43, 56, 59, 62, 68, 79, 80, 101, 102, 126, 127, 214, 236, 238, 247, 253, 282, 283, 304 Ross CA, 225, 235, 355, 362 Ross R, 313, 320, 334, 335 Rotello RJ, 355, 363 Rothenfluh-Hilfiker A, 371, 373, 379, 386, 387, 388, 389 Roth KA, 350, 360, 361 Rothman JH, 64, 81 Rothman MB, 270, 279 Roth ME, 69, 84 Roth PH, 106, 130 Roth SI, 293, 309 Rouliot V, 295, 309 Rouyer F, 383, 390

426

Author Index

Rowitch DH, 48, 57 Rowland KJ, 355, 362 Roy N, 348, 360 Roy S, 348, 355, 360, 362, 363 Rozowski M, 266, 278 Ruberg FL, 296, 310 Ruberg M, 355, 363 Ruberte E, 67, 82, 96, 125 Ruberti G, 357, 363 Rubin EM, 45, 56, 314, 322, 323, 324, 334, 335, 336 Rubin GM, 9, 12, 13, 19, 73, 87, 194, 198 Rubinstein M, 157, 169, 173 Rubkun G, 62, 80 Ruby NF, 367, 384 Ruco L, 293, 309 Ruddle FH, 35, 36, 43, 52, 56 Ruddle RH, 41, 42, 48, 55, 57 Ruderman JV, 111, 132 Rue E, 155, 171 Rue EA, 105, 106, 129, 154, 156, 170 Ruffolo SC, 355, 363 Ruiz Gomez M, 192, 200 Ruiz I, 62, 68, 80 Ruiz i Altaba A, 68, 83 Ruiz-Lozano P, 248, 253 Ruiz P, 161, 174 Ruland SL, 326, 336 Rumsey Wl, 156, 170, 172 Rundstadler JA, 263, 277 Runyan RB, 229, 235, 289, 290, 308 Rush MG, 73, 88, 99, 126 Rusholme SA, 289, 307 Rushton E, 183, 200 Russell DW, 169, 176 Rutila JE, 370, 373, 386, 387 Rutter WJ, 216, 236 Ruvkun G, 27, 33 Ryan HE, 155, 160, 171 Ryan TE, 103, 127, 285, 306

S Sablitzky F, 47, 57 Sack RL, 368, 385

Saez E, 46, 47, 57 Saez L, 373, 378, 379, 386, 388, 389 Saffitz JE, 69, 85 Sagawa K, 296, 310 Saga Y, 220, 237, 238 Sage EH, 282, 303 Sage M, 376, 388 Sah VP, 356, 363 Saigo K, 66, 82, 93, 123, 187, 200 Saijoh Y, 60, 78, 79, 160, 174, 244, 253 Saji T, 263, 277 Sajjadi F, 35, 53 Sakakida Y, 376, 387 Sakamoto K, 382, 390 Sakata Y, 356, 363 Salceda S, 109, 119, 131, 136, 168, 169, 175, 177 Salghetti SE, 111, 133 Salvesen GS, 348, 355, 360, 362 Salvino R, 193, 201 Samakovlis C, 66, 81, 82, 93, 94, 95, 98, 99, 102, 124 Sampath K, 244, 248, 253 Sampedro J, 73, 87 Samper E, 228, 236 Sancar A, 380, 389 Sanders LC, 299, 311 Sandgren EP, 39, 54 Sandmeyer S, 112, 134 Sanford LP, 229, 238 Sang H, 37, 53 Sangoram AM, 45, 56, 375, 378, 387, 388 Sankar S, 282, 303 Sapir A, 95, 125 Sariola H, 229, 238 Sarma V, 287, 306 Sasai Y, 191, 198, 213, 238 Sasaki H, 62, 68, 69, 80, 83, 247, 253 Sassa S, 273, 274, 280 Sato H, 299, 311 Satomura K, 289, 307 Sato T, 76, 89 Sato TN, 102, 127, 285, 293, 305 Sauer B, 40, 48, 55

Author Index Savakis C, 266, 277 Savion N, 296, 310 Savitz SI, 358, 364 Sawa H, 350, 360 Sawaya PL, 69, 83, 84 Sawyer H, 271, 279 Saxena S, 356, 363 Scalzitti JM, 168, 176 Scandrett JM, 206, 236 Scangos GA, 35, 36, 52 Scanu AM, 324, 336 Scaramuzino D, 283, 305 Scardrett JM, 191, 198 Schaad O, 69, 84 Schachter N, 329, 338 Schachter NS, 326, 337 Schainfeld R, 104, 128 Schalet AP, 266, 277 Schau M, 107, 111, 131, 169, 177 Schechner D, 30, 33 Schedl A, 213, 233, 237 Scheffer IE, 352, 361 Scheffner M, 112, 133, 134 Schibler U, 40, 55, 382, 390 Schier A, 70, 85 Schier AF, 216, 219, 226, 236, 237, 238, 263, 267, 277, 278 Schierenberg E, 61, 63, 79, 81 Schilling TF, 261, 276 Schinke M, 179, 184, 201 Schlange T, 245, 253 Schleicher JR, 354, 362 Schmandt R, 193, 201 Schmid B, 213, 237, 263, 277 Schmidt H, 119, 136 Schmidt JV, 160, 174 Schmidt K, 122, 137 Schmidt O, 2, 18 Schmitt D, 63, 81 Schnabel H, 61, 79 Schnabel R, 61, 79 Schneider A, 245, 253 Schneider MD, 217, 219, 222, 229, 230, 235, 238 Schnieke AE, 36, 53 Schoeckimann HO, 288, 307

427 Schoenwolf GC, 247, 253 Schoonbee HJ, 274, 280 Schottelius AJ, 112, 134 Schott JJ, 180, 185, 200 Schottler F, 352, 361 Schotz MC, 328, 338 Schrammel A, 122, 137 Schroff AD, 288, 307 Schrott HG, 326, 337 Schuchhardt S, 156, 170, 172 Schuchmann M, 353, 361 Schuerer-Maly CC, 120, 136 Schuh AC, 101, 126, 214, 238, 282, 304 Schuh R, 67, 82, 96, 125 Schulte-Merker S, 213, 236, 263, 277 Schultheiss TM, 63, 81, 180, 185, 200, 213, 238 Schultz G, 121, 137 Schulz RA, 192, 197, 208, 237 Schumacker PT, 120, 136, 158, 173 Schutz G, 62, 68, 69, 80, 84 Schwachtgen JL, 301, 312 Schwartz AL, 111, 133 Schwartzberg PL, 35, 53 Schwartz C, 111, 133 Schwartz L, 62, 80, 282, 304 Schwartz M, 282, 303 Schwartz RJ, 70, 85, 184, 192, 197, 200 Schwartz WJ, 368, 369, 385 Schwarz J, 215, 237 Schweitzer A, 295, 309 Schweitzer JB, 354, 362 Schweitzer R, 96, 125 Scott CJ, 322, 335 Scottgale TN, 13, 19 Scott IC, 228, 235 Scrable H, 352, 361 Scully AL, 104, 128 Scully SA, 76, 89, 99, 126 Sears R, 111, 133 Seashore C, 283, 305 Seawright A, 374, 387 Seemayer TA, 293, 309 Seger R, 95, 96, 125 Sehgal A, 145, 150, 371, 373, 386, 387

428 Seidman CE, 196, 180, 185, 200 Seidman JG, 180, 185, 196, 200 Seifert E, 62, 68, 80 Seiki M, 299, 311 Seipmann TJ, 115, 135 Seip RL, 326, 336 Sekine K, 76, 89 Seldin M, 287, 306 Selig L, 113, 134 Semanza Gl, 155, 171 Semenkovich CF, 326, 336 Semenza GL, 105, 106, 107, 118, 119, 120, 129, 130, 131, 135, 136, 147, 153, 154, 155, 156, 157, 158, 159, 161, 162, 163, 166, 167, 168, 169, 170, 170, 171, 172, 173, 174, 175, 176, 177 Semenza Gl, 154, 155, 156, 158, 159, 160, 161, 170, 171, 173 Semigran MJ, 356, 363 Semina EV, 72, 87, 244, 253 Senaldi G, 297, 311 Senju S, 350, 360 Sensabaugh SM, 98, 125 Seol JH, 112, 134 Sepulveda JL, 192, 200 Seroude L, 145, 150 Seung KJ, 75, 88 Shaham S, 343, 359 Shah AR, 167, 175 Shaharabany M, 96, 125 Shah PC, 230, 238 Shalaby F, 101, 126, 214, 238, 282, 304 Sham JS, 162, 163, 174 Shao H, 287, 306 Shapiro RA, 167, 175 Sharkey A, 180, 185, 196 Sharma K, 290, 308 Sharp FR, 166, 167, 175 Sharp R, 76, 89 Shashikant CS, 41, 42, 43, 55, 56 Shawver LK, 286, 306 Shaw-White JR, 69, 70, 72, 83, 85 Shearman LP, 376, 378, 383, 387, 388, 390 Shearman IP, 380, 389

Author Index Shearman MS, 355, 363 Shearn A, 16, 19, 20 Sheffield VC, 290, 308 Sheng X, 326, 337 Sheng Z, 328, 337 Shen HM, 232, 236, 293, 309 Shenkman B, 296, 310 Shen L, 69, 83 Shen Y, 296, 310 Shepherd S, 192, 201 Sherlock G, 25, 32 Sherman A, 37, 53 Shevchenko A, 112, 134 Shibuya M, 283, 284, 304, 305 Shigeyoshi Y, 375, 376, 387, 388 Shih C, 26, 33 Shih DM, 314, 334 Shi J, 120, 136 Shilo B, 95, 109, 120, 125, 131 Shilo BJ, 92, 97, 101, 123 Shilo BZ, 66, 67, 81, 82, 93, 95, 96, 97, 99, 105, 107, 108, 123, 124, 125, 131, 157, 169, 173, 183, 196 Shima A, 266, 278 Shimada A, 266, 278 Shimada M, 321, 335 Shima M, 289, 307 Shimano H, 321, 335 Shimoda LA, 154, 156, 162, 163, 170, 174 Shimono A, 60, 78, 79, 244, 253 Shindler KS, 350, 361 Shioi T, 192, 199 Shirane M, 111, 113, 133 Shirras AD, 98, 125 Shishido E, 187, 200 Shivdasani RA, 257, 275 Shizuya H, 41, 55 Shou W, 229, 238 Shows TB, 287, 306 Shrigley RJ, 93, 124 Shutter JR, 350, 360 Shwechuk BM, 40, 54 Shweiki D, 105, 128, 129, 168, 176 Sica A, 159, 173 Sick TJ, 141, 149

Author Index Siegbahn A, 284, 305 Siegel-Bartelt J, 72, 87 Siegfried E, 2, 17 Sie J, 2, 18 Siep M, 332, 339 Sigrist K, 69, 85, 192, 198, 199, 206, 219, 236 Sigworth FJ, 141, 149 Silberbach GM, 180, 185, 200 Silver LM, 316, 335 Silver M, 283, 305 Silverman HS, 167, 175 Simcox AA, 191, 197 Simmons DL, 296, 310 Simokat K, 104, 128 Simonet WS, 76, 89, 329, 338 Simon H, 230, 231, 236, 237 Simon M, 41, 55 Simon MC, 69, 84, 85, 106, 120, 131, 136, 158, 160, 168, 173, 174, 175, 191, 198, 260, 276 Simons JW, 168, 169, 176, 177 Simons M, 165, 174 Simpson P, 192, 200 Sinclair JF, 119, 136 Sinclair P, 119, 136 Singal A, 353, 361 Singaraja R, 355, 362 Singer A, 261, 277 Singer D, 140, 148 Singer F, 261, 276 Singh CM, 145, 150 Singleton K, 354, 361 Siraganian RP, 296, 310 Sisodia SS, 27, 32 Sith JD, 319, 335 Siwicki KK, 371, 386 Siwik E, 353, 361 Skaaer H, 99, 126 Skeath J, 193, 200 Skeath JB, 192, 197 Sklar J, 3, 18 Skowyra D, 112, 113, 134 Skuntz S, 46, 47, 57 Slavin BG, 331, 332, 339 Slepak T, 41, 55

429 Slonim DK, 330, 338 Slunt HH, 27, 32 Slusarski D, 74, 88 Smale G, 354, 361 Small KW, 72, 87 Smirnova O, 380, 389 Smit GL, 274, 280 Smith D, 355, 363 Smith DJ, 45, 56, 314, 334 Smith E, 68, 83, 289, 307 Smith GR, 42, 55, 268, 279 Smithies O, 35, 38, 40, 53 Smith JC, 257, 259, 275, 276 Smith JD, 314, 323, 329, 334, 336, 338 Smith M, 282, 303 Smith S, 76, 89, 101, 126 Smith SD, 3, 18 Smith T, 25, 32 Smith TJ, 326, 337 Smolik SM, 111, 132 Sneller M, 357, 363 Snider WD, 350, 360, 361 Snyder EY, 48, 57 Snyder SH, 225, 235 Sobe T, 353, 361 Soengas MS, 349, 360 Soffer D, 168, 176, 353, 361 Sogawa K, 107, 115, 131, 158, 160, 169, 173, 174, 176 Soker S, 284, 305 Solli T, 284, 305 Solnica-Krezel L, 216, 226, 237, 238, 263, 265, 277 Sommer B, 289, 307 Song EJ, 45, 56, 375, 387 Song JI, 257, 275 Song JJ, 257, 275 Song V, 111, 133 Soni B, 380, 389 Sonnenfeld M, 66, 82, 93, 123 Soreng AL, 3, 18 Sorensen LK, 289, 307 Sorenson CM, 350, 360 Sorescu D, 120, 136 Soriano P, 221, 228, 235, 236

430 Soucy JW, 145, 150 Soudais C, 69, 85, 192, 198, 206, 219, 236 Soulier S, 50, 57 Southgate E, 22, 33 Souza P, 76, 89 So WV, 145, 150, 377, 378, 381, 388, 389 Spadafora C, 39, 54 Spagnuolo R, 295, 309 Spana EP, 192, 193, 200 Spangler E, 322, 335 Spellberg J, 297, 311 Spellberg JP, 296, 310 Spencer E, 115, 117, 135 Spicer AP, 225, 235 Spoerke MJ, 24, 32 Sporn MB, 289, 307 Sprawson N, 161, 174 Springer JE, 354, 362 Squarcia U, 242, 252 Srinivas V, 118, 119, 135, 136 Sriram S, 380, 389 Srivastava D, 179, 184, 199, 208, 216, 217, 236, 237, 238 St. Clair RW, 314, 334 St. John MA, 3, 18 St. John PL, 287, 307 Staahl BT, 120, 137 Staehling-Hampton K, 183, 200 Stahl S, 66, 82, 93, 123 Stahl SJ, 69, 84 Stainier DY, 63, 81, 204, 205, 214, 216, 218, 220, 221, 222, 226, 235, 237, 238, 270, 271, 279 Staknis D, 378, 381, 388, 389 Stalfors P, 328, 338 Stambrook PJ, 43, 55 Stamler JS, 348, 360 Stanbridge E, 348, 360 Stanewsky R, 381, 383, 389, 390 Stanfield GM, 145, 151 Stanfield GS, 25, 26, 32 Stanford WL, 282, 304 Stanier DY, 271, 279 Stankis D, 378, 388

Author Index Staple JK, 350, 360 Stassi G, 357, 363 Stebbins CE, 115, 135, 169, 176 Steeg PS, 16, 20 Steen RG, 330, 338 Steeves TDL, 378, 388 Steinberg D, 315, 328, 334, 338 Steindler DA, 354, 362 Stein E, 104, 128, 287, 288, 307 Steiner M, 161, 174 Steingrimsson E, 216, 236 Steinmann K, 301, 312 Stemple DL, 216, 226, 237, 238 Stenmark K, 154, 156, 170 Stennicke HR, 348, 360 Stephan FK, 369, 385 Sternberg NL, 41, 55 Sternberg PW, 26, 27, 32, 33 Stern CD, 244, 253 Sternglanz R, 216, 236 Stern MD, 167, 175 Sterrer S, 293, 309 Stetler-Stevenson WG, 16, 19, 298, 299, 301, 311, 312 Stevens ME, 45, 56 Stewart CL, 232, 236, 293, 309 Stewart-Phillips J, 319, 335 Stewart RA, 3, 12, 16, 18 Stewart WB, 283, 284, 305 Stieg PE, 355, 362 Stilletti S, 299, 311 Stinnakre MG, 50, 57 Stitelman D, 301, 312 St-Jacques B, 69, 83 St Johnston RD, 192, 200 Stoesser KL, 270, 279 Stokowski RP, 326, 336 Stoler M, 155, 159, 167, 171 Stoltzfus LJ, 323, 336 Stone EM, 290, 308 Stone J, 166, 175, 283, 305 Stone JR, 121, 137 Storey KB, 142, 149 Strahle U, 214, 217, 219, 235, 237, 238 Strakc P, 111, 133 Strand D, 15, 19

Author Index Strasser A, 357, 364 Stratford IJ, 168, 176 Strauss AW, 180, 185, 196 Straus SE, 357, 363 Strawn LM, 286, 306 Strebel K, 113, 134 Streisininger G, 265, 277 Stresinger G, 261, 276 Stricker RB, 358, 364 Strieter RM, 282, 303 Stripp BR, 69, 83 Strober W, 357, 363 Stromblad S, 299, 311, 312 Strongin AY, 299, 311 Struhl G, 8, 18, 27, 33, 73, 88, 111, 115, 132 Sturgess JM, 248, 254 Sturtevant MA, 96, 125 Subbarayan V, 215, 217, 237 Subramani S, 35, 53 Sucov HM, 227, 236, 238 Sudanagunta SP, 45, 56 Suehiro A, 72, 87 Suen LF, 155, 171 Su G, 71, 86 Sugawara J, 155, 171 Sugaya K, 354, 361 Sugihara M, 244, 253 Sugimoto A, 64, 81 Su HC, 289, 307 Suh H, 72, 87 Suh YA, 120, 136 Sukhatme V, 301, 312 Sulik K, 247, 253 Sulston JE, 22, 33, 61, 63, 79 Sultana C, 296, 310 Su MSS, 349, 360 Su MT, 145, 150, 179, 180, 182, 183, 194, 195, 196, 200 Sun FL, 40, 54 Sun L, 24, 32 Sun X, 78, 89, 162, 163, 174 Sun YA, 143, 144, 146, 150 Sun YH, 192, 199 Sun Z, 267, 278 Sun ZS, 375, 376, 387, 388

431 Supp DM, 244, 248, 253, 254 Suri C, 102, 103, 127, 285, 305 Suri V, 380, 389 Susin SA, 348, 360 Sutherland D, 66, 67, 82, 93, 94, 95, 99, 102, 124, 125 Sutherland DC, 66, 81, 93, 95, 98, 124 Sutherland DJ, 2, 17 Suzuki A, 222, 238, 257, 275 Suzuki R, 72, 87 Suzuki T, 302, 312 Swieter M, 296, 310 Sylvester JT, 162, 163, 174 Symes JF, 104, 128 Szabo E, 208, 210, 238 Szabo K, 12, 19 Szeto D, 72, 87, 250, 254 Szuka SM, 50, 58

T Tabara H, 29, 33 Tabin C, 244, 253 Tabin CJ, 73, 88 Tacchini L, 155, 172 Tachiiri Y, 41, 55 Tada M, 263, 277 Taguchi K, 376, 387, 388 Taigen T, 70, 85 Taipale J, 102, 127, 283, 304 Tajima M, 167, 175 Takagi A, 220, 238 Takahashi JS, 45, 56, 380, 389 Takahashi T, 283, 305 Takahata S, 160, 174 Takano H, 167, 175 Takano R, 380, 389 Takao M, 380, 389 Takashima N, 376, 387 Takashima S, 284, 305 Takeda S, 248, 253, 254 Takeichi M, 98, 125, 222, 237 Takekida S, 376, 388 Taketo MM, 289, 307 Takeuchi T, 51, 53 Takimoto H, 228, 236, 297, 311

432 Takino T, 299, 311 Takumi T, 376, 387 Talbot D, 40, 55, 350, 360 Talbot WS, 219, 236, 238, 261, 263, 267, 277, 278 Tall AR, 324, 336 Tamamaki N, 72, 87 Tamarin I, 296, 310 Tam PP, 204, 205, 238 Tanaka H, 117, 135, 382, 390 Tanaka M, 179, 184, 201, 216, 238, 375, 387 Tanaka-Matakatsu M, 98, 125 Tanaka T, 73, 87 Tanaka Y, 248, 253, 254 Tan CC, 156, 172 Tanganelli P, 293, 309 Tang XL, 167, 175 Tang Z, 69, 70, 85, 191, 192, 198, 199, 206, 237 Tanioka M, 302, 312 Tannenbaum A, 140, 144, 148 Tan P, 112, 134 Tansey WP, 111, 133 Tanswell AK, 76, 89 Tao W, 3, 18 Tao WF, 68, 83 Tarzami S, 69, 85, 192, 198 Tateno H, 39, 54 Tatton WG, 354, 362 Taussig MJ, 45, 56 Taya C, 349, 360 Taya S, 169, 176 Taylor BL, 105, 129 Taylor BS, 167, 175 Taylor DG, 289, 307 Taylor JM, 329, 338 Taylor LS, 106, 130, 159, 173 Taylor S, 329, 338 Tazuke SI, 155, 171 Tease C, 266, 278 Tefft JD, 76, 89, 101, 126 Tei H, 375, 376, 387, 388 ten Dijke P, 289, 308 Tendler DS, 106, 131 Terwilliger JD, 331, 338

Author Index Testi R, 357, 363 Tetsu O, 111, 133 Theodosiou NA, 12, 19, 115, 135 Therrien M, 13, 19 Thiery JP, 76, 89, 99, 126 Thies RS, 257, 275 Thinakaran G, 27, 32 Thisjssen KL, 24, 28, 30, 32 Thomas D, 113, 134 Thomas JB, 104, 128, 370, 386 Thomas KH, 368, 385 Thomas KR, 35, 53, 282, 303, 304 Thomas LB, 354, 362 Thomason A, 76, 89 Thomas PQ, 72, 87 Thomas T, 216, 238 Thompson CB, 341, 359 Thompson EW, 299, 312 Thompson JS, 315, 334 Thompson MA, 263, 277 Thompson MW, 248, 254 Thompson RD, 297, 311 Thompson S, 35, 53 Thomson JN, 22, 33, 61, 63, 79 Thomson M, 261, 277 Thom SR, 296, 310 Thornberry N, 355, 362 Thornberry NA, 343, 344, 346, 348, 355, 359, 360, 363 Thresher RJ, 380, 389 Thurston G, 102, 127 Tian H, 169, 176 Tian J, 323, 336 Tian YM, 109, 131 Tichelaar JW, 69, 84 Tick G, 15, 19 Timmerman LA, 228, 236 Ting CN, 191, 198 Tischkau SA, 145, 150 Tissenbaum HA, 62, 80 Tissot C, 111, 132 Tobias KE, 111, 133 Todaro M, 357, 363 Todorov AT, 140, 144, 148 Toft N, 357, 363 Tolleshaug H, 313, 334

Author Index Tonissen KF, 63, 81, 179, 184, 201 Torok T, 15, 19 Toti P, 293, 309 Tournier-Lasserve E, 2, 17 Townsend K, 267, 278 Townsley FM, 111, 132 Toyoda Y, 39, 54, 244, 253 Tozawa Y, 285, 293, 305 Trainor C, 69, 84 Tran CM, 227, 238 Trapley JE, 76, 89 Treisman R, 92, 97, 123 Tremblay M, 71, 86 Tremble PM, 299, 312 Tremp G, 327, 337 Trifiro MA, 355, 362 Trogan E, 323, 336 Trojanowski JQ, 354, 362 Troost D, 354, 362 Trout J, 213, 237, 263, 277 Troutt AB, 357, 363 Trumbauer ME, 46, 47, 57 Tryggvason K, 301, 312 Trygstad O, 332, 339 Tsai FY, 69, 84, 259, 276 Tsai MJ, 217, 237 Tsai SF, 192, 199, 259, 260, 276 Tsai SY, 217, 237 Tsang PA, 260, 276 Tseu I, 76, 89 Tsinoremas NF, 382, 390 Tsuchiya T, 106, 131 Tsuda L, 115, 135 Tsuzuki T, 71, 86 Tucker A, 289, 308 Tucker AM, 282, 303 Tuddenham EGD, 70, 86 Tully T, 145, 150 Tuomanen EI, 358, 364 Turbay D, 145, 150, 185, 199 Turek FW, 375, 387 Turenchalk GS, 3, 15, 18, 19 Turki J, 45, 56 Turner D, 224, 237 Turner JA, 248, 254 Turrens JF, 158, 173

433 Tu Y, 23, 32 Tyers M, 112, 113, 134

U Uemura T, 98, 125, 192, 201 Ueno N, 257, 263, 275, 277 Uhl GR, 383, 390 Ulich TR, 76, 89 Ullmann B, 261, 276 Umans L, 220, 237 Unthan-Fechner K, 119, 136 Urase K, 73, 88 Urness LD, 289, 307 Uzu T, 368, 385

V Vaandrager AB, 121, 137 Vail B, 263, 277 Vaillancourt JP, 355, 363 Valenzuela DM, 102, 127 Vale PR, 165, 174 Valge-Archer VE, 259, 276 van Assendelft GB, 40, 54 van Buul PP, 266, 278 Vandenhoeck A, 102, 127, 282, 304 van den Maagdenberg AMJM, 326, 337 van der Horst GT, 380, 389 VanderPlas-de Vries I, 207, 208, 236 Van Der Ploeg LHT, 355, 363 van der Putten H, 36, 53 Vanderslice R, 265, 277 Van de Ven WJ, 220, 237 van Eeden FJ, 213, 217, 219, 222, 235, 236, 237, 243, 252, 271, 279 van Eeden FJM, 272, 279 Van Etten RL, 288, 307 Van Etten WJ, 330, 338 Van Gelder RN, 382, 390 Van Lenten BJ, 327, 337 van Leuven F, 220, 237 Van Mierop LHS, 243, 252 Van Parijs L, 356, 363 Van Praagh R, 242, 252 Van Praagh S, 242, 252

434 Van Vlaenderen I, 293, 309 van Vliet KJ, 274, 280 van Vlijmen BJM, 326, 337 Van Vuren JH, 274, 280 Vaporciyan AA, 296, 310 Varesio L, 106, 130, 159, 173 Varfolomeev EE, 353, 361 Varlet I, 246, 247, 248, 253 Varon D, 296, 310 Vasavada H, 70, 86 Vatti R, 293, 309 Vaux DL, 341, 359 Vaux EC, 113, 117, 135, 169, 176 Vecchi M, 285, 306 Veis DJ, 350, 360 Velculescu VE, 169, 177 Vencent JP, 183, 196 Venkatesh T, 179, 180, 182, 183, 196 Venkatesh TV, 183, 184, 185, 199, 201 Vera S, 291, 308 Verberne ME, 161, 174 Verdi JM, 193, 201 Verman IM, 36, 53 Vernet M, 295, 309 Vernon RB, 282, 303 Verschueren K, 222, 235, 293, 308 Verstuyft J, 319, 335 Verstuyft S, 322, 335 Vertes R, 140, 144, 148 Vervoort M, 192, 201 Vesselinovitch D, 315, 334 Vidal M, 2, 18 Vierstra RD, 112, 134 Vikkula M, 286, 293, 306 Vilalta A, 40, 55 Vilotte JL, 50, 57 Vincent EB, 229, 235, 290, 308 Vincent JP, 194, 198 Vincent S, 67, 82, 96, 125 Vinos J, 95, 125 Viragh S, 208, 210, 238 Virmani R, 356, 363 Visintin I, 357, 363 Visvader J, 257, 275, 276 Visvader JE, 257, 276

Author Index Vitaterna MH, 45, 56, 375, 380, 387, 389 Vittet D, 295, 309 Vocero-Akbani AM, 358, 364 Vogelsang E, 217, 219, 222, 235, 272, 279 Vogelsang L, 271, 279 Vogelstein B, 111, 133, 169, 177 Vogt C, 342, 359 Volk T, 183, 196 Vollberg TM, 71, 86 Volpe MV, 71, 86 von Boehmer H, 352, 361 von Ehrenstein G, 63, 81 Vonesch JL, 227, 236 von Reutern M, 169, 176 Vonsattel JPG, 355, 362 von Schalscha TL, 299, 311 von Schantz M, 380, 389 Vosatka RJ, 71, 86 Vosshall LB, 373, 386 Vrablic T, 247, 253 Vreeken C, 266, 277 Vreviario F, 295, 309 Vyas S, 354, 362

W Wade DP, 324, 336 Wager-Smith K, 373, 378, 381, 387, 388, 389 Wagner GP, 43, 56 Wahren B, 50, 58 Wakayama T, 39, 54 Wakeman A, 228, 236, 297, 311 Waldo KL, 161, 174 Walker C, 261, 266, 270, 276, 278, 279 Walker CS, 266, 278 Walker D, 329, 338 Walker EB, 354, 361 Wallach D, 353, 361 Walldorf U, 97, 125, 374, 387 Walmsley ME, 259, 276 Walsh A, 324, 326, 329, 330, 336, 337, 338 Walsh CA, 352, 361

Author Index Waltenberger J, 284, 305 Walters MC, 40, 54 Wang F, 350, 360 Wang GL, 105, 106, 118, 129, 130, 135, 154, 155, 156, 157, 170, 171, 172 Wang H, 274, 280 Wang HU, 104, 128, 287, 288, 293, 307 Wang J, 76, 89, 354, 362 Wang K, 329, 338 Wang Q, 167, 175 Wang T, 230, 238, 296, 310 Wang W, 3, 12, 16, 18, 43, 56 Wang WY, 12, 19, 115, 135 Wang X, 343, 359 Wang XF, 222, 238 Wang XP, 332, 339 Wang Y, 288, 291, 293, 307, 356, 357, 363 Wang YJ, 296, 310 Wani MA, 43, 55 Wappner P, 66, 67, 82, 92, 93, 95, 96, 97, 101, 105, 107, 108, 109, 120, 123, 124, 131 Warburton D, 67, 76, 82, 89, 101, 126 Warden C, 314, 334 Warden CH, 322, 335 Ward M, 66, 82, 93, 123 Ward PA, 296, 310 Ward WW, 23, 32 Warga RM, 204, 219, 236, 238, 270, 271, 279 Warman ML, 286, 293, 306 Warren AJ, 259, 276 Warren KS, 216, 238, 243, 252 Warren R, 70, 85, 192, 197 Waslh K, 104, 128 Wassarman DA, 13, 19 Wasserman PM, 317, 335 Watabe AM, 45, 56 Watanabe M, 353, 361 Watkins SC, 296, 310 Watson KL, 15, 19 Wawersik S, 2, 17 Weaver DR, 376, 378, 380, 383, 387, 388, 389, 390

435 Webb S, 357, 363 Webster KA, 155, 171 Weeks DL, 229, 235, 290, 308 Weier HU, 45, 56 Weigel D, 62, 68, 80, 83 Weigel M, 354, 361 Weigmann K, 12, 19 Weinberger RP, 204, 205, 238 Weinberg PM, 242, 252 Weinberg RA, 26, 33 Weinmann R, 156, 172 Weinreb A, 330, 338 Weinstein BM, 206, 213, 216, 238, 270, 279 Weinstein DC, 62, 68, 80 Weinstein M, 105, 118, 130, 136, 289, 308 Weinstein RS, 358, 364 Weinstock PH, 325, 328, 336, 337 Weintraub H, 216, 236 Weis-Garcia F, 75, 88 Weiss C, 287, 307 Weiss M, 69, 84, 259, 276 Weiss MJ, 260, 276 Weiss SJ, 299, 311 Weitz CJ, 381, 389 Weizman I, 66, 81, 93, 123 Welch C, 314, 332, 334, 339 Welch PJ, 49, 57 Welke FW, 165, 174 Wellington CL, 355, 362 Welsh DK, 369, 385 Wemple M, 183, 184, 201 Wendel DP, 289, 308 Wenger RH, 106, 120, 130, 136, 154, 155, 160, 161, 170, 171 Weng S, 25, 32 Weng W, 327, 337 Werb Z, 216, 236, 299, 312 Werner S, 76, 89, 99, 126 Wert S, 76, 89, 99, 126 Wert SE, 69, 72, 76, 83, 84, 89 Wesley CS, 2, 17, 379, 389 West A, 46, 47, 57 West DC, 3, 18 Westerfield M, 261, 266, 277, 278

436 Westerman CJJ, 291, 308 Westermarck J, 301, 312 Westin J, 210, 215, 238 Westlund B, 24, 32 Weston JA, 270, 279 Weston KM, 259, 276 Westphal H, 71, 74, 86 Wever R, 365, 384 Wexler RS, 282, 303 Wharton KA, 105, 129 White FA, 350, 361 White JG, 22, 33, 61, 63, 79 Whitelaw ML, 371, 386 Whiteley A, 378, 388 White NE, 145, 150, 377, 388 White RA, 69, 84 White WB, 367, 385 Whitey M, 98, 125 Whitfield TT, 272, 280 Whitsett J, 76, 89, 99, 126 Whitsett JA, 69, 70, 72, 76, 83, 84, 85, 89 Wickline L, 96, 125 Wiedemann LM, 72, 87 Wiegand SJ, 103, 127, 285, 305 Wiener CM, 154, 156, 162, 163, 170, 174 Wieschaus E, 265, 277 Wiesener MS, 113, 117, 135, 155, 169, 171, 176 Wiese RJ, 192, 199 Wiggelsworth VB, 101, 126 Wight D, 43, 55 Wilhelmi H, 245, 253 Wilkinson GA, 287, 307 Wilk R, 66, 67, 81, 82, 92, 93, 97, 101, 123 Wilks AF, 270, 279 Willems AR, 112, 134 Willett CE, 263, 264, 268, 277, 279 Williams AJ, 301, 312 Williams DL, 327, 337 Williams L, 76, 89, 99, 126 Williams SK, 282, 303 Willie AH, 342, 359 Willmore WG, 158, 173 Willnow TE, 328, 337

Author Index Willoughby DA, 40, 55 Wilmore WG, 142, 149 Wilmut I, 36, 53 Wilsbacher LD, 45, 56, 375, 378, 387, 388 Wilson CM, 70, 86 Wilson DB, 69, 85, 191, 198, 206, 208, 219, 236, 237 Wilson DF, 156, 170, 172 Wilson R, 192, 200 Wilson WJ, 111, 132, 169, 177 Wilting J, 284, 305 Wing JP, 112, 134 Wingrove JA, 120, 137 Winnier G, 73, 75, 88, 99, 126, 225, 228, 235, 238 Winoto A, 353, 361 Winston JT, 111, 133 Winters RB, 319, 335 Wirbelauer C, 113, 135, 169, 176 Wi SK, 167, 175 Wisniewski HM, 354, 361 Wissler RW, 315, 334 Witkowska HE, 273, 280 Witte DP, 248, 254 Witzenbichler B, 103, 127 Witztum JL, 315, 334 Woddarz A, 2, 17 Woessner JF, 298, 311 Woffer D, 105, 128 Wolburg-Buchholz K, 285, 293, 305 Wolburg H, 285, 293, 305 Wolf DA, 113, 134 Wolff T, 73, 87 Wolf FW, 287, 306 Wolpert L, 245, 253 Wolway KE, 166, 167, 175 Wong EA, 282, 304 Wong FS, 357, 363 Wong GG, 259, 276 Wong-Staal F, 49, 50, 57, 58 Woodhouse E, 16, 19 Woods DF, 15, 19 Wood SM, 105, 129, 155, 168, 171, 176 Wood WB, 22, 33 Woo M, 349, 360

Author Index

437

Wozney JM, 191, 199, 257, 275 Wright CV, 75, 88 Wright DVE, 244, 248, 253 Wuerffel MK, 69, 84 Wu H, 225, 238 Wu K, 112, 134 Wu L, 293, 309 Wu T, 143, 144, 146, 150 Wutz A, 39, 54 Wu X, 183, 184, 187, 195, 199, 201, 282, 304 Wu XF, 101, 126, 214, 238 Wyckoff CC, 113, 117, 135, 169, 176 Wyllie AH, 341, 342, 357, 359, 363 Wyman RJ, 143, 144, 146, 150 Wynder C, 45, 56

X Xanthoudakis S, 355, 363 Xian J, 45, 56 Xiao J, 104, 128 Xia Y, 140, 141, 144, 148 Xia YP, 293, 309 Xie PZ, 316, 334 Xin X, 155, 171 Xiong Y, 112, 134 Xu D, 355, 363 Xu P, 332, 339 Xu Q, 63, 81, 184, 199, 204, 237 Xu T, 2, 3, 9, 12, 13, 15, 16, 18, 19, 115, 135, 143, 150, 289, 307 Xu W, 225, 238 Xu X, 120, 136, 183, 201, 289, 308 Xu XL, 183, 201 Xydas S, 63, 81, 180, 185, 200

Y Yabkowitz RR, 285, 306 Yagishita N, 76, 89 Yaich LE, 193, 200 Yamada H, 166, 175 Yamada K, 73, 87 Yamada M, 184, 200 Yamada S, 43, 55

Yamada T, 354, 362 Yamagishi T, 229, 237 Yamaguchi TP, 101, 126, 214, 238, 282, 304 Yamaji N, 257, 275 Yamamoto A, 270, 279 Yamamoto H, 352, 361 Yamamoto S, 376, 388 Yamamura H, 224, 237, 238 Yamane HK, 285, 306 Yamasaki M, 76, 89 Yamasaki N, 179, 184, 201, 216, 238 Yamazaki S, 380, 389 Yanagimachi R, 39, 54 Yanazawa M, 72, 87 Yancopoulos G, 285, 305 Yancopoulos GD, 102, 103, 104, 126, 127, 128, 282, 285, 287, 288, 304, 306, 307 Yang D, 349, 360 Yangimachi R, 39, 54 Yang JT, 206, 217, 238, 293, 308 Yang N, 155, 171 Yang X, 289, 308 Yang XW, 42, 45, 55, 56 Yan HC, 296, 310 Yanker BA, 111, 133 Yan L, 376, 388 Yan LG, 296, 310 Yan M, 284, 305 Yan W, 184, 197 Yan XH, 69, 84 Yan YL, 261, 276 Yao TP, 47, 57 Yashar M, 323, 336 Yashiro K, 60, 78, 79, 244, 253 Yasugi S, 73, 88 Yatskievych TA, 72, 87 Yeates KM, 155, 171 Yee JK, 266, 267, 278 Yee K, 260, 263, 276, 277 Ye H, 69, 83 Yeh JY, 192, 199 Yelick P, 261, 277 Yelon D, 205, 221, 222, 237, 238 Yenush L, 2, 18

438

Author Index

Yew J, 354, 361 Yingling JM, 65, 81, 92, 99, 123, 222, 238 Yin Z, 183, 190, 201 Yoder B, 63, 81 Yodoi J, 107, 115, 131 Yokotani N, 169, 176 Yokoyama M, 244, 253 Yokoyama T, 244, 248, 253 Yonekawa Y, 248, 253 Yoshida H, 228, 236, 297, 302, 311, 312 Yoshida R, 349, 360 Yoshida T, 302, 312 Yoshioka H, 244, 253 Yoshiyama Y, 354, 362 Yoshizawa T, 76, 89 Young AB, 355, 362 Young MF, 289, 307 Young MW, 2, 17, 370, 371, 373, 379, 385, 386, 387, 388 Young RSK, 140, 148 Young SG, 326, 329, 336, 337, 338 Yuan J, 343, 354, 355, 359, 362, 363 Yu AS, 50, 57 Yu AY, 154, 155, 156, 160, 161, 162, 163, 164, 166, 167, 170, 171, 174, 175 Yueh YG, 48, 57 Yuen PS, 49, 57 Yun SH, 370, 386 Yu Q, 370, 382, 385, 390 Yu W, 3, 12, 16, 18

Z Zabel BU, 72, 87 Zachariae W, 112, 134 Zachariae Z, 112, 134 Zack DJ, 166, 175 Zagzag D, 102, 127, 168, 169, 176 Zaletayev D, 374, 387 Zamboni RJ, 355, 363 Zambrowicz BP, 38, 51, 54 Zamzami N, 348, 360 Zanetti A, 295, 309 Zapata AG, 263, 264, 277

Zaret K, 68, 83 Zavitz KH, 115, 135 Zeinstra LM, 75, 88 Zelzer E, 66, 67, 82, 92, 93, 97, 101, 105, 123, 157, 169, 173 Zeng H, 373, 387 Zeng M, 348, 360 Zerr DM, 371, 386 Zgleszewski SE, 71, 86 Zhang C, 260, 276 Zhang H, 111, 132, 222, 238 Zhang HY, 222, 238 Zhang J, 219, 236, 238, 353, 361 Zhang L, 169, 177, 283, 304 Zhang IP, 118, 135 Zhang M, 224, 238 Zhang Q, 167, 175, 350, 360 Zhang S, 3, 12, 16, 18, 19, 115, 135 Zhang SH, 319, 326, 335, 337 Zhang T, 355, 362 Zhang X, 383, 390 Zhang Y, 47, 57 Zhang Z, 111, 133 Zhao C, 289, 307 Zhao J, 67, 76, 82, 89, 101, 126 Zhao Y, 45, 56, 375, 387 Zheng B, 376, 388 Zheng H, 355, 363 Zheng JZ, 107, 131, 156, 172 Zheng TS, 349, 350, 360, 361 Zhong H, 168, 169, 176, 177 Zhong W, 193, 201 Zhou G, 217, 237 Zhou H, 102, 127 Zhou L, 69, 76, 83, 89 Zhou LI, 214, 237 Zhou W, 169, 177 Zhou X, 247, 253 Zhou Y, 184, 201 Zhuchenko O, 375, 376, 387 Zhu J, 64, 81 Zhulin IB, 105, 129 Zhu X, 119, 136 Zhu XH, 118, 135 Zhu Y, 329, 338, 355, 363 Zijlstra M, 35, 53

Author Index Zimmerman WF, 380, 389 Zinn K, 104, 128 Zipursky S, 115, 135 Zipursky SL, 104, 128 Ziyadeh FN, 290, 308 Zocchi MR, 296, 310 Zone SE, 368, 385 Zon L, 213, 236, 261, 263, 273, 277, 280 Zon LI, 42, 55, 191, 198, 206, 213, 238, 257, 259, 260, 261, 268, 270, 275, 276, 279

439 Zou H, 343, 359 Zou X, 45, 56 Zuasti A, 263, 264, 277 Zucker I, 369, 385 Zucker S, 299, 311 Zukowski MM, 296, 297, 310, 311 Zwartkruis F, 216, 226, 237, 238 Zwiebel LJ, 371, 386 Zwijsen A, 222, 235, 293, 308 Zychlinsky A, 358, 364 Zylka MJ, 376, 378, 380, 383, 387, 388, 389, 390

SUBJECT INDEX

A Abnormal left-right asymmetry pathology, 241 Abnormal wing discs, 16 ActIIRB asymmetrical signaling developing lungs, 250–251 Activated growth factor receptors HIF-1, 157 Activation-induced cell death (AICD) lymphocytes, 357 Activin2B receptor (ActIIRB) asymmetrical signaling developing lungs, 250–251 Adult hermaphrodites Caenorhabditis elegans, 23 Advance sleep phase syndrome (ASPS) circadian clocks, 368 AICD lymphocytes, 357 ALDA, 106 Aldolase A (ALDA), 106 Alpha4-integrin, 232–234 Alzheimer’s disease, 27 2-aminopurine HIF-1, 157 Anaphase-promoting (APC) complex, 112

Anatomy Caenorhabditis elegans model, 22– 24 Anemia cardiovascular response shep, 165 Anf, 216 Angiogenesis, 281–302. See also Vasculogenesis Drosophila melanogaster tracheal system development, 101–104 Angiopoietins, 102–103 Anoxia-tolerant organisms adaptive strategy, 142 unresolved questions, 142–143 Antennapedia (Antp) gene, 70–71 Anterior visceral endoderm (AVE), 60 Antisense strategy loss of function, 49 transgenic mice, 49 Antp gene, 70–71 APC complex, 112 ApoB gene, 323–324 ApoE-deficient mice, 319–321 Apolipoproteins genetically engineered mouse models, 318

441

442 Apoptosis, 25–26, 341–359 autoimmune disease, 356–357 cancer, 357–358 cardiac development, 353 cardiac disease, 356 caspases, 344–348 historical perspective, 342 hypoxia-mediated HIF-1, 167–168 immune system, 352 neural development, 349–352 neurodegenerative disease, 353– 356 pathways, 342–344 vertebrate development, 348–353 Apoxia tolerance, 139–147 electrophysiological aspects, 140– 141 metabolic considerations, 141–142 neurotransmitters, 141 oxidative stress, 142 between species and phyla, 140– 142 turtles, 139–142 ARNT, 105 Aryl hydrocarbon nuclear translocator (ARNT), 105 ASPS circadian clocks, 368 Asthma circadian clocks, 368 Asymmetrical signaling activin2B receptor developing lungs, 250–251 ATF-1, 106 Atherosclerosis induced mouse mutants, 318–330 mouse genetic models, 313–333 mouse models increased resistance, 322–323 novel genes, 330–333 Atrial natriuretic factor (anf ), 216 Autoimmune disease apoptosis, 356–357 AVE, 60 Axial mesoderm, 60

Subject Index B BACs, 41 Bacterial artificial chromosomes (BACs), 41 Bacterial inducible systems, 47 Bagpipe (bap) gene, 68 Balancer chromosome, 4 Balancer strain, 10 Bap gene, 68 Basic fibroblast growth factor (bFGF) angiogenesis, 102 Basic helix loop helix-PAS (bHLHPAS) family, 105, 110 Bcl-2 superfamily, 349–350 Behavioral genetics circadian rhythms, 369–370 period mutants, 369–370 Behavioral testing, 146 Beta-adrenergic receptor gene, 45 Beta-galactosidase gene, 43 BFGF angiogenesis, 102 BHLH-PAS family, 105, 110 Bicistronic reporter gene, 43 Binary systems transgenic mouse models, 47–48 Bipolar disorder circadian rhythm, 368 Blastomeres, 61 Bloodless mutants, 271–272 Blood pressure circadian clocks, 367–368 Blue light-sensitive cryptochrome genes, 379–381 Bmp4, 71, 75, 78 valve formation, 225–226 Bnf, 216 Bnl gene, 66, 93–95, 107 Bone morphogenetic protein 4 (Bmp4) valve formation, 225–226 Bonnie and clyde heart defects, 222 Brain natriuretic factor (bnf), 216 Brain tumors HIF-1 alpha, 168–169

Subject Index

443

Branching morphogenesis Drosophila melanogaster tracheal system development, 98–104 Branchless (bnl) gene, 66, 93–95, 107 Breast ductal carcinoma in situ HIF-1 alpha, 168–169 Breathless gene, 93–94 Bronchial cell migration FGF, 100 Buds early mouse lung, 77

C CAD, 109 Caenorhabditis elegans, 144 adult hermaphrodites, 23 endoderm development, 61–65 gene deletion project, 30 pharynx development, 61–65 Caenorhabditis elegans model, 21–31 anatomy, 22–24 gene knockout technology human disease gene homologs, 28– 30 genetics, 22 human/worm conservation, 25–28 molecular genetics, 24 prospects, 31 Caenorhabditis elegans pharyngeal development ceh-22 gene, 63 elt-2, 63–65 end-1, 63–65 pha-4, 62 Cancer apoptosis, 357–358 Cancer biology research, 15–17 fly and human homologs, 15 Candidate gene approach, 267 Carbon dioxide HIF-1, 158–160 Cardia bifida mutants, 219 Cardiac cell specification tinman, 186 wg gene, 186

Cardiac cell type-specific differentiation zfh-1, 194 Cardiac development apoptosis, 353 vertebrates. See Vertebrate cardiac development Cardiac differentiation genetic loci, 213–215 Cardiac disease apoptosis, 356 Cardiac induction, 185–191 Cardiac looping, 243 Cardiac mesoderm alternative cell fates, 192–194 Cardiac morphogenesis insects and vertebrates, 180–181 Cardiac septation, 210–211, 227–232 Cardiovascular malformations heterotaxia, 241–243 Casanova mutations endodermal defects, 219–220 Caspase-3, 349 Caspase-9, 349, 351 Caspase activation cascade, 344–348 initiation, 345–346 regulation, 346–348 Caspases, 344–348 CBP, 115 Ced-3 gene, 343 Ced-4 gene, 343 Ced-9 gene, 343 Ceh-22 gene, 63 Cell death. See Apoptosis Cell fate determination, 2–3 Cell-mediated gene transfers transgenic mouse models, 37–39 Cell proliferation, 2 Cellular response hypoxia, 104–122 Cerebral ischemia HIF-1, 166–167 CETP, 324 Chablis mutants, 272 Chamber differentiation, 215–218 Channel arrest, 142 Chardonnay mutants, 272

444 Chianti mutants, 272 Cholesterol ester transfer protein (CETP), 324 Ci gene, 73–74 Circadian clocks, 367–368 mammalian, 377 period gene, 370–371 Circadian dysrhythmias, 368 Circadian photoreception, 379–381 Circadian rhythms, 365–367 behavioral genetics, 369–370 Circadian transcriptional feedback loops, 370–373 timeless gene, 371–373 Circadian transcription factors flies, 376–378 mammals, 376–378 Cis-regulatory analysis transgenic mouse models, 42–44 Cld mutation, 331 Clear cell renal carcinoma HIF-1 alpha, 169 Cloche, 270–271 Clock mammalian, 368–369 postranslational, 378–379 Clock gene, 375 Clock gene homologs mammalian, 374 Clocks fly, 365–384 outputs, 381–382 Clockwork blue, 379–381 Coagulation factor VIII, 274 Cobalt hypoxic response, 119 Conventional screen vs. F1 mosaic screen, 10–11 Coup-tfll chamber differentiation, 217 CREB-1, 106 CREB-binding protein (CBP), 115 Cre recombinase, 48 Cropped gene, 67 Cryptochrome gene blue light-sensitive, 379–381

Subject Index Cspg2 valve formation, 224 C terminal transactivation domain (CAD), 109 Cubitus interruptus (Ci) gene, 73–74 Cushion overgrowth, 231–232 Cytochrome c oxidase, 120

D Dbt gene, 379 Dcry, 381 Decapentaplegic (dpp) gene, 66–67, 96, 183 ectopic heart specification, 189–190 vertebrate cardiogenesis, 191 Deferoxamine (DFO) hypoxic response, 118–119 Definitive hematopoiesis, 256 Delayed sleep phase syndrome (DSPS) circadian clocks, 368 Delta, 98 Desferrioxamine HIF-1, 157–158 Developing lungs asymmetrical signaling activin2B receptor, 250–251 Dextrocardia with situs solitus, 242 DFO hypoxic response, 118–119 Dhand chamber differentiation, 216–217 Diet-induced atherosclerosis mouse, 315 Diphenylene iodonium (DPI) HIF-1, 158 DOCK, 104 Dominant negative mutations, 49 Doubletime (dbt) gene, 379 Doubletime kinase, 378–379 Downregulation metabolism, 142 Dper, 375 DPI HIF-1, 158 Dpp gene, 66–67, 96, 183

Subject Index ectopic heart specification, 189–190 vertebrate cardiogenesis, 191 Dpp signaling pathway, 67, 97 Drifter gene, 93 Drosophila alternative cell fates, 192–194 cardiac induction, 185–191 cardiac morphogenesis, 180–181 circadian rhythms, 365–367 gastrulation, 181 hypoxia, 143–147 tinman expression, 182–185 Drosophila cryptochrome gene (dcry), 381 Drosophila epidermal growth factor receptor (EGFR) pathway, 93 Drosophila genetic model system, 1–17 cancer biology research, 15–17 elevating gene expression, 5–8 genetic research, 4 genetic screens, 9–15 inactivating gene expression, 8–9 P-element-mediated gene transfer, 4–5 techniques, 3–15 Drosophila genome mutagenizing, 146 Drosophila gut development, 67–68 Drosophila heart development, 179–180 Drosophila hedgehog (hh) gene, 73 Drosophila hypoxic response, 120–121 Drosophila melanogaster ubiquitin-mediated proteasomal pathway, 115 Drosophila melanogaster tracheal system development, 91–98 branching morphogenesis, 98–104 FGF receptor pathway, 93–95 tracheal branching, 96–98 tracheal cell fate differentiation, 93 tracheal placode subdivision, 95–96 vasculogenesis and angiogenesis, 101–104 Drosophila molting hormone ecdysone, 47 Drosophila per (dper), 375 Drosophila Schneider (SL2) cells, 106

445 Drosophila tracheal development, 65– 66 Drosophila tracheal network, 92 Drosophila tumor suppressers, 5 DSPS circadian clocks, 368 Dual cell autonomy, 220–222

E Ectopic activation screens, 12 Ectopic heart specification dpp gene, 189–190 tinman, 190–191 wg gene, 189–190 EGFR pathway, 93, 97 TGF-B, 95–96 Ehand chamber differentiation, 216–217 Electromobility shift assay (EMSA), 106–108 Electronic resources mouse genetics, 317 Electron transport chain complex IV, 120 Elevating gene expression, 5–8 ectopic expression endogenous promoters, 5–6 FLP-out system, 6–8 GAL4/UAS system, 6 Elt-2, 63–65 Embryogenesis HIF-1, 160–161 Embryonic stem (ES) cells, 35 EMSA, 106–108 End-1, 63–65 Endocardial differentiation, 213–214 Endoderm, 60 Endodermal defects, 219–220 Endoderm development Caenorhabditis elegans, 61–65 Endogenous promoters elevating gene expression ectopic expression, 5–6 Endoglin vasculogenesis, 291–293

446

Subject Index

ENO1, 106 Enolase 1 (ENO1), 106 EPAS-1, 110 Eph/Ephrin family vasculogenesis, 286–289 Ephrins, 103–104 Epiblast, 59 Epicardial development, 212, 232–234 Epistasis genetic pathway building inactivating gene expression, 8–9 EPO, 233–234 HIF-1, 155 Epor, 233–234 ErbB2 trabeculation defects, 230–231 ErbB3 trabeculation defects, 230–231 ErbB4 trabeculation defects, 230–231 Erythropoietin (EPO), 233–234 HIF-1, 155 Erythropoietin receptor (epor), 233–234 ES cells, 35 gene-targeting strategy, 50 Eukaryotic cells ubiquitin-mediated proteasomal pathway, 111 Eve gene, 187, 194 Even-skipped (eve) gene, 187, 194 Expression loss transgenic mouse models, 49–51 External light-dark cycles molecular clocks, 373–374 Eyeless gene, 374

F Familial combined hyperlipidemia (FCHL) mouse model, 326–327 Familial hypercholesterolemia, 325 Fatty liver dystrophy gene, 330–333 Faust heart defects, 221 FCHL

mouse model, 326–327 Fenton reaction, 119 Fgf8 left-right signaling, 249–250 FGF family, 75–78 Fgf10 gene, 76, 78, 99 FGF receptor pathway Drosophila melanogaster tracheal system development, 93–95 Fibroblast growth factor (FGF) family, 75–78 Fibronectin heart defects, 220 FixJ, 118 FixL, 118 Fk506 binding protein 12 (FKBP12) ventricular septal defects, 229 FkBP12 ventricular septal defects, 229 Fkh genes, 64, 68 Fld gene, 330–333 Fld mutation, 331–332 Flies. See also Drosophila biological similarity to humans, 2 circadian transcription factors, 376– 378 Flk-1 endocardial differentiation, 214 Floxed gene, 48 FLP-out system elevating gene expression ectopic expression, 6–8 FLP recombination target (FRT) sites, 8 Flt-1 endocardial differentiation, 214 Fly and human homologs cancer biology research, 15 Fly circadian clock, 372 Fly clocks, 365–384 F1 mosaic screen vs. conventional screen, 10–11 Foam cells, 314 Foregut development fork head/HNF-3 gene, 68–69 Fork head/HNF-3 gene, 62, 68 foregut development, 68–69

Subject Index

447

Forward gene, 383–384 Forward genetics, 265–266 mammalian, 374–375 Founders apoptotic defects, 350–352 Frascati mutants, 272 Freixenet mutants, 274 FRT sites, 8 Fruit flies mutant actogram, 366 Functional genomics transgenic mouse models, 51–52 Furin heart defects, 220–221 Fusion fate, 98 Fusion proteins, 47

G GAL4/UAS system, 48 elevating gene expression ectopic expression, 6 Gamma irradiation, 265–266 Gastrulation, 59, 204–216 Gata4 endodermal defects, 219 GATA family, 69–78 heart development, 191–192 hematopoiesis, 259–260 Gene activation hypoxia HIF-1, 154–156 Gene deletion project Caenorhabditis elegans, 30 Gene encoding green fluorescent protein (GFP), 43 Gene knockout, 35–36 human disease gene homologs Caenorhabditis elegans model, 28– 30 steps, 50 Gene regulatory elements in vivo mapping, 329–330 Gene-targeting strategy ES cells, 50

Genetically engineered mouse models lipid metabolism proteins, 318 Genetic analysis mouse, 316–317 zebrafish development, 264–265 Genetic defects left-right development, 245–251 Genetic invertebrate models hypoxia, 143–147 Genetic loci cardiac differentiation, 213–215 Genetic mosaics loss of function mutations inactivating gene expression, 9 Genetic pathway building epistasis inactivating gene expression, 8–9 Genetic research Drosophila, 4 Genetics behavioral circadian rhythms, 369–370 period mutants, 369–370 Caenorhabditis elegans model, 22 forward, 265–266 hematopoiesis, 258 mammalian forward, 374–375 molecular Caenorhabditis elegans model, 24 mouse electronic resources, 317 Genetic screens, 9–15 Gene-trapped ES cell clones commercial availability, 51 Genistein HIF-1, 157 Genome ping-ponging, 261–262 GFP, 43 Gli genes, 73–75 Glioblastoma multiforme HIF-1 alpha, 168–169 Glp-1, 61 Gut development Drosophila, 67–68 Gut endoderm, 59

448

Subject Index H

Hairpin ribozymes loss of function, 49–50 Hammerhead ribozymes loss of function, 49–50 Handed asymmetry development, 245 Has2 valve formation, 224–225 HD, 355 Heart normal left-right asymmetry, 240–241 trabeculation, 212 Heart-and-soul tube assembly, 223 Heart cell type diversification, 192– 195 Heart development GATA, 191–192 pannier, 191–192 tinman, 182–185 Heart morphogenesis, 204–213 epicardial development, 212 gastrulation, 204–216 heart tube assembly, 206–208 looping morphogenesis, 208 secondary convergence, 206–208 septation, 210–211 trabeculation, 212 valve formation, 208–210 Heart tube assembly, 206–208 Heat-shock protein 68 (hsp68) promoter, 43 Hedgehog (hh) gene, 66, 68, 73, 187 Hemangioblastoma HIF-1 alpha, 168–169 Hematopoiesis factors required, 256–260 genetics, 258 zebrafish, 255–274 Hematopoietic disorders zebrafish, 269–274 Hematopoietic mutations zebrafish, 271–274

Hematopoietic stem cells (HSCs), 256– 259 Hemophilia model, 274 HEP, 274 Hepatoerythropoietic porphyria (HEP), 274 Hermaphrodites Caenorhabditis elegans, 23 Heterotaxia cardiovascular malformations, 241– 243 Hex.Pitx2 gene, 72 Hh gene, 66, 68, 73, 187 HIF-1. See Hypoxia inducible factor-1 HIF-1 alpha brain tumors, 168–169 HIF-1 alpha-deficient embryonic stem cells, 160 HIF-1 alpha degradation ubiquitin-mediated proteasomal pathway, 111–115 Hnf-3beta, 247 Homeobox genes, 70–73 Homologous recombination, 30 Hormone-sensitive lipase (HSL), 328 Hox genes, 70–71, 99 Hper, 375 Hrs heart defects, 221 HSCs, 256–259 HSL, 328 Hsp68 promoter, 43 Human disease gene homologs gene knockout technology Caenorhabditis elegans model, 28– 30 Humanization mouse lipid profile transgene expression, 323–324 Human lipoprotein disorders murine models, 325–327 Human ortholog (hper), 375 Human pathology transgenic mouse models, 51–52 Human umbilical vein endothelial cell cultures (HUVEC), 284

Subject Index Human/worm conservation, 25–28 Huntington gene, 355 Huntington’s disease (HD), 355 HUVEC, 284 Hyaluronic acid synthase 2 (has2) valve formation, 224–225 Hydrogen peroxide hypoxic response, 118–119 Hyplip 1 gene, 330–333 Hypobetalipoproteinemia mouse model, 326 Hypoblast, 59 Hypochromic mutants, 272–273 Hypoxia cellular response, 104–122 disorders associated with, 153 Drosophila, 143–147 gene activation HIF-1, 154–156 genetic invertebrate models, 143–147 sensing pathways, 142 Hypoxia inducible factor-1 (HIF-1), 105–109 activated growth factor receptors, 157 carbon dioxide, 158–160 cerebral ischemia, 166–167 embryogenesis, 160–161 expression, 156–160 gene activation hypoxia, 154–156 hypoxia-mediated apoptosis, 167–168 hypoxic pulmonary hypertension, 161–164 ischemic neovascularization, 164–166 ischemic preconditioning, 167 mitochondrial electron transport chain inhibitors, 157–158 nitric oxide, 158–160 nonproteolytic oxygen regulation, 115–118 oxygen concentration, 156 oxygen homeostasis, 153–169 oxygen levels, 109 pharmacological inducers/inhibitors, 156–157 retinal vascularization, 166

449 [Hypoxia inducible factor-1 (HIF-1)] signal transduction pathways, 159 structure, 154 target genes, 155 tumor progression, 168–169 Hypoxia-mediated apoptosis HIF-1, 167–168 Hypoxia signal-transduction pathway, 156–157 Hypoxic pulmonary hypertension HIF-1, 161–164

I IDL, 314 Ig, 352 Imaginal disk cells cell fate determination, 2–3 Immune system apoptosis, 352 Immunoglobulin (Ig), 352 Inactivating gene expression epistasis genetic pathway building, 8–9 genetic mosaics loss of function mutations, 9 Induced mouse mutants atherosclerosis, 318–330 Induced Mutant Resource Database, 317 Inducible gene expression transgenic mouse models, 46–57 Insects cardiac morphogenesis, 180–181 oxygen delivery, 91–92 Insertional mutagenesis, 266–267 transgenic mouse models, 51 Insert size transgenic mouse models, 40–41 Integration site transgenic mouse models, 39–40 Intermediate-density lipoprotein (IDL), 314 Inversus viscerum (iv) mutation, 248 In vivo protein function, 327–328

450

Subject Index

In vivo mapping gene regulatory elements, 329–330 Ionic mechanisms apoxia tolerance, 141 Ischemic neovascularization HIF-1, 164–166 Ischemic preconditioning HIF-1, 167 Isolated dextrocardia, 242 Isolated levocardia, 242 Isomerism, 241 Iv mutation, 248

J Jekyll mutants, 226–227 Jet lag, 368 Jrk flies, 376–378

K Kartagener’s syndrome, 248–249 Knockout models atherosclerosis susceptibility, 319– 323 vasculogenesis and angiogenesis, 102

L Lac operon, 47 Lactate dehydrogenase A (LDH-A), 106 Lark gene, 383 Lateral signaling pathway defects, 249–251 Lats gene, 16 LDH-A, 106 LDL, 313–314 Left isomerism, 243 Left-right asymmetry abnormal, 241 development, 243–245 molecular development, 243–244 morphological development, 243 normal heart and lungs, 240–241 vertebrate development, 239–240

Left-right development genetic defects, 245–251 Left-right positional information, 245– 246 Left-right signaling fgf8, 249–250 Levocardia with situs inversus, 242 Light circadian rhythm, 367 Light-dark cycles external molecular clocks, 373–374 Lipid metabolism novel genes, 330–333 Lipoprotein lipase deficiency, 325 Lipoprotein receptors genetically engineered mouse models, 318 Lipoprotein remodeling enzymes genetically engineered mouse models, 318 Lonely atrium chamber differentiation, 217 Looping, 224 Looping morphogenesis, 208 Loss of function antisense strategy, 49 ribozymes, 49 Loss of function mutations genetic mosaics inactivating gene expression, 9 Low-density lipoprotein (LDL), 313– 314 Lungs development, 59–60 activin2B receptor, 250–251 Drosophila melanogaster tracheal system development, 99– 101 mouse buds, 77 normal left-right asymmetry, 240–241 Lymphocytes AICD, 357

Subject Index

451 M

Mammalian circadian clock, 377 Mammalian clock, 368–369 Mammalian clock gene homologs, 374 Mammalian forward genetics, 374–375 Mammalian respiratory system, 59 Mammals circadian transcription factors, 376– 378 MAPK, 284 MAP kinase activation pattern, 96 Matrix metalloproteases vasculogenesis, 298–302 MCry, 380 Mef2c chamber differentiation, 215–216 Merlot mutants, 272 Mersalyl HIF-1, 157 Mesodermal defects, 220 MesP1 mesodermal defects, 220 Metabolism downregulation, 142 Metal-induced growth hormone mice, 46 MGD, 317 Middle cerebral artery occlusion HIF-1, 166–167 Midline, 245–246 Miles apart heart defects, 222 Misexpression transgenic mouse models, 44–46 Mitochondrial electron transport chain inhibitors HIF-1, 157–158 Mitogen-activated protein kinase (MAPK), 284 Modifier screens, 12–14 Molecular clocks external light-dark cycles, 373–374 Molecular genetics Caenorhabditis elegans model, 24 Moonshine mutants, 271–272

Mosaic genetic screens, 9–12 Mouse diet-induced atherosclerosis, 315 genetic analysis, 316–317 Mouse cloning nuclear transplantation, 38 Mouse foregut and lung development, 68–78 FGF family, 75–78 GATA family, 69–78 Mouse genetic models atherosclerosis, 313–333 Mouse genetics electronic resources, 317 Mouse Genome Database (MGD), 317 Mouse lipid profile humanization transgene expression, 323–324 Mouse lung buds, 77 Mouse models atherosclerosis increased resistance, 322–323 FCHL, 326–327 hypobetalipoproteinemia, 326 protein function in vivo, 327–328 protein structure/function studies, 328–329 Mouse mutants web site for, 51 MPer, 376 Multicellularity human/worm conservation, 25 Murine models human lipoprotein disorders, 325–327 Mutant fruit flies actogram, 366 Mutated genes cloning, 267 Mutations epicardial development, 232–234 Myocardial differentiation swirl gene, 213 Myocyte enhancer factor 2c (mef2c) chamber differentiation, 215–216

452

Subject Index

Myxothiazol HIF-1, 158

N NAD, 109 Natter heart defects, 222 N-cadherin tube assembly, 223 Nematode. See also Caenorhabditis elegans pharynx, 61 NeoR, 50 N-ethyl-N-nitrosourea, 265 Neural development apoptosis, 349–352 Neuregulin trabeculation defects, 230–231 Neurodegenerative disease apoptosis, 353–356 Neurofibromatosis 1 (Nf1) cushion overgrowth, 231–232 Neurotransmitters apoxia tolerance, 141 Nf1 cushion overgrowth, 231–232 NF-ATc ventricular septal defects, 228–229 Nitric oxide, 120–121 HIF-1, 158–160 NK-2 genes, 64 Nkx2.1, 69, 71 Nkx2.3, 63 Nkx2.5, 184, 191 chamber differentiation, 216 Nkx2.7, 63 Nodal, 247 Node, 245–246 Node momocilium, 247–249 Nonproteolytic oxygen regulation HIF-1, 115–118 Normal heart left-right asymmetry, 240–241 Normal lungs left-right asymmetry, 240–241

NOS/PKG pathway, 121 Notch gene alternative cell fates, 192–193 Notch pathway, 98 Notch receptor, 3 Novel genes lipid metabolism, 330–333 N terminal transactivation domain (NAD), 109 Nuclear transplantation mouse cloning, 38 Numb gene alternative cell fates, 192–193

O One-eyed pinhead mutations endodermal defects, 219–220 Osteoporosis PTH, 358 Outputs clocks, 381–382 Overexpression transgenic mouse models, 44–46 Oxidative stress apoxia tolerance, 142 reoxygenation, 142 Oxygen concentration HIF-1, 156 Oxygen delivery insects, 91–92 Oxygen homeostasis HIF-1, 153–169 Oxygen levels HIF-1, 109 Oxygen sensing signal transduction, 118–122

P P300, 115 Pandora chamber differentiation, 217 Pannier heart development, 191–192 Pantip markers, 98

Subject Index Parathyroid hormone (PTH) osteoporosis, 358 PAS sensor domain, 118 Patched (ptc) gene, 73–74 PECAM-1 vasculogenesis, 296–298 P-element-mediated gene transfer Drosophila, 4–5 Per, 105 Period gene, 376–377 circadian clocks, 370–371 evolutionarily conserved, 375–376 Period mutants behavioral genetics, 369–370 Period (Per), 105 PGK1, 106 Pha-4 pharyngeal development, 62 Pharyngeal development Caenorhabditis elegans, 61–65 Phosphofructokinase (PKF), 106 Phosphogliceratokinase 1 (PGK1), 106 Photosensitive mutants, 274 Pinotage mutants, 272 Pitx2 gene, 72 PKF, 106 Pointed gene, 67 Positional cloning, 267–268 Postmitotic neurons apoptotic defects, 350–352 Postranslational clock, 378–379 Prehematopoietic mutations zebrafish, 269–271 Presinilins, 27 Programmed cell death, 25–26. See also Apoptosis Pronuclear injection transgenic mouse models, 36 Protein function in vivo, 327–328 Protein structure/function studies mouse model, 328–329 Pruned gene, 67 Pseudemys scripta apoxia tolerance, 140 Ptc gene, 73–74

453 PTH osteoporosis, 358

R Raldh2 chamber differentiation, 217 Random transgene integration, 51 Ras signaling, 26–27 Recombinational cloning transgenic mouse models, 42 Reoxygenation oxidative stress, 142 Respiratory system mammalian, 59 Respiratory tract development genetic models, 59–78 Caenorhabditis elegans pharyngeal development, 61–65 Drosophila tracheal and gut development, 65–68 mouse foregut and lung development, 68–78 Ret gene, 6 Retinal vascularization HIF-1, 166 Retinoblastoma gene, 28 Retinoid receptor (RXR), 47 Retroviral vectors transgenic mouse models, 36–37 Retsina mutants, 272 Rhizobium meliloti, 105 Rhodnius prolixus, 101, 107 Ribozymes loss of function, 49 Right isomerism, 242–243 RNA interference technique, 29 Rotenone HIF-1, 158 RXR, 47 RXRalpha gene ventricular septal defects, 227–228

S Sanpodo (spdo) gene alternative cell fates, 192–193

454 Sauternes mutants, 272 SCF VCB E3 ligase complexes, 114 SCF complex, 112–113 SCN, 368–369 Seasonal affective disorder circadian rhythm, 368 Secondary convergence, 218–223 Sel-12, 27 Sensing pathways hypoxia, 142 Serpent, 64 Severe blood deficiency mutants, 272 SGC, 120–121 Shh gene, 68–69, 71, 73–75, 99 Signal transduction oxygen sensing, 118–122 Signal transduction pathways HIF-1, 159 Sim, 105, 370 Sima, 110 Similar (Sima), 110 Single-minded (Sim), 105, 370 Situs inversus, 242 Situs solitus, 240 Skn-1, 61 Sloppy-paired gene, 187 SMAD5 vasculogenesis, 291–293 Sodium fluoride HIF-1, 157 Soluble guanylate cyclase (sGC), 120– 121 Sonic hedgehog (shh) gene, 68–69, 71, 73–75, 99 Spadetail, 270 Spatial promoters elevating gene expression ectopic expression, 5–6 Spdo gene alternative cell fates, 192–193 Sperm-mediated gene transfers transgenic mouse models, 39 Spiracles, 91 Spitz, 95–96 Spitz signaling pathway, 67

Subject Index Sprouty gene, 67, 76, 95, 99 Steroid response elements, 46–47 Swirl, 269–270 Swirl gene myocardial differentiation, 213 Syprachiasmatic nucleus (SCN), 368– 369

T Tango gene, 93 Target genes HIF-1, 155 TBASE, 317 T-cell receptor (TCR), 352 TCR, 352 T/ebp-1. See Nkx2.1 Tef-1 ventricular septal defects, 227–228 Temperature circadian rhythm, 367 Temporal promoters elevating gene expression ectopic expression, 5–6 Terminal branches, 91–93 Tet operon, 47 TGF-B EGFR pathway, 95–96 TGF-B2 ventricular septal defects, 229 TGF-BR vasculogenesis, 289 TGF-B type III receptor, 290 Therapeutic angiogenesis, 165–166 Thunderbird mutants, 272 Thymidine kinase (tk) gene, 50 Tie-1, 103 Tie-2, 103 Tie-angiopoietin family vasculogenesis, 285–286 Timeless (tim) gene circadian transcriptional feedback loops, 371–373 Tim gene circadian transcriptional feedback loops, 371–373

Subject Index Tinman gene cardiogenic role, 181–185 Drosophila, 182–185 ectopic heart specification, 190–191 heart cell specification, 186 Tissue oxygen deprivation, 139 Tk gene, 50 Tll1 ventricular septal defects, 228 Tolloid-like-1 (tll1) ventricular septal defects, 228 Trabeculation defects, 230–231 Trachea development, 59–60 Drosophila, 65–66 Tracheal branching Drosophila melanogaster tracheal system development, 96–98 Tracheal cell fate differentiation Drosophila melanogaster tracheal system development, 93 Tracheal cell migration FGF, 100 Trachealess gene, 66, 93, 105 Tracheal identity, 93 Tracheal placodes, 66, 93 Drosophila melanogaster tracheal system development, 95–96 Tracheal system development Drosophila melanogaster, 91–98 Transcription enhancer factor-1 (tef-1) ventricular septal defects, 227–228 Transferable degradation domains, 109– 114 Transforming growth factor-B2 (TGF-B2) ventricular septal defects, 229 Transforming growth factor receptorbeta (TGF-BR) vasculogenesis, 289 Transgene expression humanization mouse lipid profile, 323–324 Transgene models atherosclerosis susceptibility, 319–323 Transgenic analysis, 268–269 zebrafish, 262

455 Transgenic mouse models, 35–52 antisense strategy, 49 binary systems, 47–48 cis-regulatory analysis, 42–44 expression loss, 49–51 functional genomics, 51–52 gene transfer methods, 36–39 inducible gene expression, 46–57 insertional mutagenesis, 51 insert size, 40–41 integration site, 39–40 overexpression, 44–46 problems, 52 recombinational cloning, 42 Transgenic/Targeted Mutation Database (TBASE), 317 Transposons transgenic mouse models, 36–37 Trh, 105 Trimmed gene, 67 TTF-1. See Nkx2.1 Tumor growth factor-beta (TGF-B) EGFR pathway, 95–96 Tumor progression HIF-1, 168–169 Turtles apoxia tolerance, 139–142 Twist gene, 194 Type I hyperlipoproteinemia, 325 Type III hyperlipoproteinemia, 326

U Ubiquitin-mediated proteasomal pathway Drosophila melanogaster, 115 eukaryotic cells, 111 HIF-1alpha degradation, 111–115 UROD, 274 Uroporphyrinogen decarboxylase (UROD), 274

V VAB-1, 104 Valve formation, 208–210, 224–227

456 Vascular cell adhesion molecule-1 (vcam-1), 232–234 Vascular endothelial growth factor (VEGF) HIF-1, 155 Vasculogenesis, 281–302 Drosophila melanogaster tracheal system development, 101–104 endoglin, 291–293 Eph/Ephrin family, 286–289 matrix metalloproteases, 298–302 PECAM-1, 296–298 SMAD5, 291–293 TGF-betaR, 289 Tie-angiopoietin family, 285–286 VE-cadherin, 293–296 VEGFR/VEGF, 282–294 Vcam-1, 232–234 VE-cadherin vasculogenesis, 293–296 VEGF HIF-1, 105, 155 vasculogenesis, 102 VEGFR/VEGF vasculogenesis, 282–294 Ventricular septal defects, 227–230 Versican (cspg2) valve formation, 224 Vertebrate cardiac development, 203–235 cardiac differentiation, 213–215 cardiac septation, 227–232 chamber differentiation, 215–218 epicardial development, 232–234 heart morphogenesis, 204–213 looping, 224 secondary convergence, 218–223 tube assembly, 223 valve formation, 224–227 Vertebrate cardiogenesis dpp gene, 191 wg gene, 191 Vertebrate development left-right asymmetry, 239–240 Vertebrate hematopoiesis, 255–256 Vertebrates cardiac morphogenesis, 180–181

Subject Index VHL tumor suppresser protein, 113 Vinculin valve formation, 225 Viral infection transgenic mouse models, 36–37 Visceral endoderm, 59 Von-Hippel-Lindau (VHL) tumor suppresser protein, 113 VP16, 48 VP16-ecdysone receptor, 47

W Weibherbst mutants, 272 Wg gene, 66, 73, 183–184 ectopic heart specification, 189–190 heart cell specification, 186 vertebrate cardiogenesis, 191 Whole Mouse Catalog, 317 Wilms’ tumor (wt-1), 233 Wing discs abnormal, 16 Wingless (wg) gene, 66, 73, 183–184 ectopic heart specification, 189–190 heart cell specification, 186 vertebrate cardiogenesis, 191 Winter depression, 368 Wt-1, 233

X Xnoggin, 75

Y YACs, 41 Yeast artificial chromosomes (YACs), 41 Yquem mutants, 274

Z Zebrafish cardiac development, 203–235 cardiac septation, 210–211, 227–232 chamber differentiation, 215–218

Subject Index [Zebrafish] cloche mutation, 206 endocardial differentiation, 213– 214 development, 261–262, 264–265, 264–269 epicardial development, 232–234 gastrulation, 213–216 heart tube assembly, 206–208 hematopoietic development, 263– 264 hematopoietic disorders, 269–274 hematopoietic mutations, 271–274 looping, 224 looping morphogenesis, 208 mutations, 265

457 [Zebrafish] myocardial differentiation swirl gene, 213 prehematopoietic mutations, 269–271 secondary convergence, 206–208, 218–223 transgenic analysis, 262 tube assembly, 223 useful characteristics, 261–262 valve formation, 208–210, 224–227 valve mutants, 226–227 Zfh-1 cardiac cell type-specific differentiation, 194 Zinfandel mutants, 272 Ziprol1 gene, 45