Molecular Hematology [4th Edition] 9781119252931

Table of contents :
Contributors ix

Preface to the fourth edition xiii

Further reading xv

Acknowledgments xvi

1 Beginnings: the molecular pathology of hemoglobin 1
David Weatherall

2 Stem cells 21
David T. Scadden

3 The genetics of acute myeloid leukemias 37
Amy M. Trottier & Carolyn J. Owen

4 Molecular diagnostics and risk assessment in myeloid malignancies 49
Christian Scharenberg & Torsten Haferlach

5 Molecular basis of acute lymphoblastic leukemia 59
Bela Patel & Fiona Fernando

6 Chronic myeloid leukemia 71
Hagop Kantarjian, Jorge Cortes, Elias Jabbour & Susan O’Brien

7 Myeloproliferative neoplasms 87
Jyoti Nangalia, Anthony J. Bench, Anthony R. Green & Anna L. Godfrey

8 Lymphoma genetics 101
Jennifer L. Crombie, Anthony Letai & John G. Gribben

9 The molecular biology of chronic lymphocytic leukemia 111
John G. Gribben

10 The molecular biology of multiple myeloma 121
Wee Joo Chng & P. Leif Bergsagel

11 The molecular basis of bone marrow failure syndromes and red cell enzymopathies 131
Deena Iskander, Lucio Luzzatto & Anastasios Karadimitris

12 Anemia of chronic disease 155
Tomas Ganz

13 The molecular basis of iron metabolism 161
Nancy C. Andrews & Tomas Ganz

14 Hemoglobinopathies due to structural mutations 173
D. Mark Layton & Steven Okoli

15 Molecular pathogenesis of malaria 193
David J. Roberts, Arnab Pain & Chetan E. Chitnis

16 Molecular coagulation and thrombophilia 207
Björn Dahlbäck & Andreas Hillarp

17 The molecular basis of hemophilia 221
Daniel P. Hart & Paul L.F. Giangrande

18 The molecular basis of von Willebrand disease 235
Luciano Baronciani

19 Platelet disorders 251
Kenneth J. Clemetson

20 The molecular basis of blood cell alloantigens 267
Cristina Navarrete, Louise Tilley, Winnie Chong & Colin J. Brown

21 Functions of blood group antigens 285
Jonathan S. Stamler & Marilyn J. Telen

22 Autoimmune hematological disorders 297
Drew Provan & John W. Semple

23 Molecular therapeutics in hematology: gene therapy 319
William M. McKillop & Jeffrey A. Medin

24 Pharmacogenomics 339
Leo Kager & William E. Evans

25 History and development of molecular biology 353
Paul Moss

26 Cancer stem cells 363
Sara Ali & Dominique Bonnet

27 Molecular basis of transplantation 373
Francesco Dazzi & Antonio Galleu

Index 389

Citation preview

Molecular Hematology

Molecular Hematology FOURTH EDITION Edited by

Drew Provan MD FRCP FRCPath Emeritus Reader in Autoimmune Haematology Department of Haematology Barts and The London School of Medicine and Dentistry Queen Mary University of London, UK

John G. Gribben MD DSc FRCP FRCPath FMedSci Professor of Medical Oncology Barts Cancer Institute Barts and The London School of Medicine and Dentistry Queen Mary University of London, UK

This edition first published 2020 © 2020 by John Wiley & Sons Edition History [3rd edition, 2010] All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions. The right of Drew Provan to be identified as the author of editorial in this work has been asserted in accordance with law. Registered Office(s) John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK Editorial Office 9600 Garsington Road, Oxford, OX4 2DQ, UK For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com. Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in standard print versions of this book may not be available in other formats. Limit of Liability/Disclaimer of Warranty The contents of this work are intended to further general scientific research, understanding, and discussion only and are not intended and should not be relied upon as recommending or promoting scientific method, diagnosis, or treatment by physicians for any particular patient. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of medicines, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each medicine, equipment, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. While the publisher and authors have used their best efforts in preparing this work, they make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. Library of Congress Cataloging-in-Publication Data Names: Provan, Drew, editor. | Gribben, John editor. Title: Molecular hematology / edited by Drew Provan, John G. Gribben. Description: Fourth edition. | Hoboken, NJ : Wiley-Blackwell, 2020. | Includes bibliographical references and index. Identifiers: LCCN 2019034376 (print) | LCCN 2019034377 (ebook) | ISBN 9781119252870 (hardback) | ISBN 9781119252955 (adobe pdf) | ISBN 9781119252931 (epub) Subjects: MESH: Hematologic Diseases | Molecular Biology–methods Classification: LCC RC636 (print) | LCC RC636 (ebook) | NLM WH 120 | DDC 616.1/5–dc23 LC record available at https://lccn.loc.gov/2019034376 LC ebook record available at https://lccn.loc.gov/2019034377 Cover Design: Wiley Cover Image: © JUAN GAERTNER/SCIENCE PHOTO LIBRARY/Getty Images Set in 10/12pt MinionPro by Aptara Inc., New Delhi, India 10 9 8 7 6 5 4 3 2 1

Dedication

We would like to dedicate this book to two people: our dear friend and colleague, Professor Sir David Weatherall, who sadly passed away on 8 December 2018. He was truly a pioneer of molecular biology and was the first physician scientist to use molecular techniques to study hematological disease. We will all miss him very much. In addition, we would like to dedicate the book to Val Provan. Always in our thoughts and much missed.

Contents

Contributors, ix Preface to the fourth edition, xiii Further reading, xv Acknowledgments, xvi 1 Beginnings: the molecular pathology of hemoglobin, 1 David Weatherall 2 Stem cells, 21 David T. Scadden 3 The genetics of acute myeloid leukemias, 37 Amy M. Trottier & Carolyn J. Owen

13 The molecular basis of iron metabolism, 161 Nancy C. Andrews & Tomas Ganz 14 Hemoglobinopathies due to structural mutations, 173 D. Mark Layton & Steven Okoli 15 Molecular pathogenesis of malaria, 193 David J. Roberts, Arnab Pain & Chetan E. Chitnis 16 Molecular coagulation and thrombophilia, 207 Bj¨orn Dahlb¨ack & Andreas Hillarp 17 The molecular basis of hemophilia, 221 Daniel P. Hart & Paul L.F. Giangrande 18 The molecular basis of von Willebrand disease, 235 Luciano Baronciani

4 Molecular diagnostics and risk assessment in myeloid malignancies, 49 Christian Scharenberg & Torsten Haferlach

19 Platelet disorders, 251 Kenneth J. Clemetson

5 Molecular basis of acute lymphoblastic leukemia, 59 Bela Patel & Fiona Fernando

20 The molecular basis of blood cell alloantigens, 267 Cristina Navarrete, Louise Tilley, Winnie Chong & Colin J. Brown

6 Chronic myeloid leukemia, 71 Hagop Kantarjian, Jorge Cortes, Elias Jabbour & Susan O’Brien

21 Functions of blood group antigens, 285 Jonathan S. Stamler & Marilyn J. Telen

7 Myeloproliferative neoplasms, 87 Jyoti Nangalia, Anthony J. Bench, Anthony R. Green & Anna L. Godfrey 8 Lymphoma genetics, 101 Jennifer L. Crombie, Anthony Letai & John G. Gribben 9 The molecular biology of chronic lymphocytic leukemia, 111 John G. Gribben 10 The molecular biology of multiple myeloma, 121 Wee Joo Chng & P. Leif Bergsagel 11 The molecular basis of bone marrow failure syndromes and red cell enzymopathies, 131 Deena Iskander, Lucio Luzzatto & Anastasios Karadimitris 12 Anemia of chronic disease, 155 Tomas Ganz

22 Autoimmune hematological disorders, 297 Drew Provan & John W. Semple 23 Molecular therapeutics in hematology: gene therapy, 319 William M. McKillop & Jeffrey A. Medin 24 Pharmacogenomics, 339 Leo Kager & William E. Evans 25 History and development of molecular biology, 353 Paul Moss 26 Cancer stem cells, 363 Sara Ali & Dominique Bonnet 27 Molecular basis of transplantation, 373 Francesco Dazzi & Antonio Galleu Index, 389

vii

Contributors

Sara Ali MD Haematopoietic Stem Cell Laboratory, The Francis Crick Institute, London, UK

Kenneth J. Clemetson PhD, ScD, CChem, FRSC Department of Haematology, Inselspital, University of Berne, Berne, Switzerland

Nancy C. Andrews MD, PhD Duke University School of Medicine, Durham, NC, USA

Jorge Cortes MD Department of Leukemia, University of Texas MD Anderson Cancer Center, Houston, TX, USA

Luciano Baronciani PhD Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico, Angelo Bianchi Bonomi Hemophilia and Thrombosis Center, Milan, Italy Anthony J. Bench MA, PhD Laboratory Medicine, NHS Lothian, Edinburgh, UK P. Leif Bergsagel MD Division of Hematology-Oncology, Comprehensive Cancer Center, Mayo Clinic Arizona, Scottsdale, AZ, USA Dominique Bonnet PhD Haematopoietic Stem Cell Laboratory, The Francis Crick Institute, London, UK Colin J. Brown PhD, FRCPath Histocompatibility and Immunogenetics Laboratory, NHS Blood and Transplant; Faculty of Life Sciences & Medicine, King’s College London, London, UK Chetan E. Chitnis MSc, MA, PhD Malaria Group, Pasteur Institute, Paris, France Wee Joo Chng MB ChB, MRCP, FRCPath National University Cancer Institute, National University Health System of Singapore; University of Singapore, National University Hospital, Singapore Winnie Chong PhD Histocompatibility and Immunogenetics Service Development Laboratory, NHS Blood and Transplant, London, UK

Jennifer L. Crombie MD Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA ¨ Dahlback ¨ MD, PhD Bjorn Department of Translational Medicine, Section of Clinical Chemistry, Lund University, University Hospital, Malm¨o, Sweden Francesco Dazzi MD, PhD School of Cancer & Pharmaceutical Sciences, King’s College London; King’s Health Partners Cancer Research UK Centre, London, UK William E. Evans PharmD St Jude Children’s Research Hospital, Memphis, TN, USA Fiona Fernando MD Centre of Haemato-Oncology, Barts Cancer Institute, Queen Mary University of London, London, UK Antonio Galleu MD, PhD School of Cancer & Pharmaceutical Sciences, King’s College London; King’s Health Partners Cancer Research UK Centre, London, UK Tomas Ganz PhD, MD Department of Medicine, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA

ix

x Contributors

Paul L.F. Giangrande MD, FRCP, FRCPath, FRCPCH Formerly of Oxford Haemophilia and Thrombosis Centre, Churchill Hospital, Oxford, UK

Anthony Letai MD, PhD Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA

Anna L. Godfrey PhD, MRCP, FRCPath Department of Haematology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK

Lucio Luzzatto MD Department of Haematology and Blood Transfusion, Muhimbili University College of Health Sciences, Dar-esSalaam, Tanzania

Anthony R. Green PhD, FRCP, FRCPath, FMedSci Department of Haematology, Cambridge Institute for Medical Research; Wellcome Medical Research Council Stem Cell Institute, Cambridge, UK

William M. McKillop PhD Department of Pediatrics, Medical College of Wisconsin, Milwaukee, WI, USA

John G. Gribben MD, DSc, FRCP, FRCPath, FMedSci Barts Cancer Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

Jeffrey A. Medin PhD Departments of Pediatrics and Biochemistry, Medical College of Wisconsin, Milwaukee, WI, USA

Torsten Haferlach MD MLL Munich Leukemia Laboratory, Munich, Germany

Paul Moss MD, PhD School of Cancer Sciences, University of Birmingham, Birmingham, UK

Daniel P. Hart FRCP, FRCPath, PhD The Royal London Hospital Haemophilia Centre, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK Andreas Hillarp PhD Department of Clinical Chemistry and Transfusion Medicine, Halland County Hospital, Halmstad, Sweden Deena Iskander MD, PhD, MRCP Centre for Haematology, Imperial College London, Hammersmith Hospital, London, UK

Jyoti Nangalia PhD, MRCP, FRCPath Welcome Sanger Institute, Hinxton; Department of Haematology, University of Cambridge; Wellcome-Medical Research Council Cambridge Stem Cell Institute, Cambridge, UK Cristina Navarrete PhD, FRCPath Histocompatibility and Immunogenetics Service Development Department, NHS Blood and Transplant; Department of Immunology and Molecular Pathology, University College London, London, UK

Elias Jabbour MD Department of Leukemia, University of Texas MD Anderson Cancer Center, Houston, TX, USA

Susan O’Brien MD Department of Leukemia, University of Texas MD Anderson Cancer Center, Houston, TX, USA

Leo Kager MD Department of Pediatrics, St. Anna Children’s Hospital, Medical University Vienna, Austria

Steven Okoli MBChB, FRCP, FRCPath Center for Hematology, Imperial College London, London, UK

Hagop Kantarjian MD Department of Leukemia, University of Texas MD Anderson Cancer Center, Houston, TX, USA

Carolyn J. Owen MD, MDres(UK), FRCPC Division of Hematology and Hematological Malignancies, University of Calgary, Foothills Medical Centre, Calgary, Canada

Anastasios Karadimitris PhD, MRCP, FRCPath Department of Haematology and Blood Transfusion, Muhimbili University College of Health Sciences, Dar-esSalaam, Tanzania D. Mark Layton MB BS, FRCP, FRCPCH Center for Hematology, Imperial College London, London, UK

Arnab Pain PhD Biological and Environmental Sciences and Engineering (BESE) Division, King Abdullah University of Science and Technology, Jeddah, Saudi Arabia; Nuffield Division of Clinical Laboratory Sciences (NDCLS), The John Radcliffe Hospital, University of Oxford, Headington, Oxford, UK

Contributors xi

Bela Patel MD, FRCPath, MD(res) Centre of Haemato-Oncology, Barts Cancer Institute, Queen Mary University of London, London, UK Drew Provan MD, FRCP, FRCPath Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK David J. Roberts DPhil, MRCP, FRCPath National Health Service Blood and Transplant (Oxford), The John Radcliffe Hospital, Oxford, UK David T. Scadden MD Department of Stem Cell and Regenerative Biology, Harvard Stem Cell Institute, Harvard University; Center for Regenerative Medicine, Massachusetts General Hospital, Boston, MA, USA Christian Scharenberg MD, PhD Department of Hematology, Skaraborgs Hospital, Sk¨ovde; Department of Cell and Molecular Biology, Karolinska Institute, Stockholm, Sweden John W. Semple PhD Division of Hematology and Transfusion Medicine, Lund University, Lund, Sweden

Jonathan S. Stamler MD Harrington Discovery Institute and Institute of Transformative Molecular Medicine, University Hospitals Cleveland Medical Center and Case Western Reserve University, Cleveland, OH, USA Marilyn J. Telen MD Department of Medicine, Division of Hematology, Duke University Medical Center, Durham, NC, USA Louise Tilley PhD International Blood Group Reference Laboratory, NHS Blood and Transplant, Bristol, UK Amy M. Trottier MSc, MD, FRCPC Division of Hematology and Hematological Malignancies, University of Calgary, Foothills Medical Centre, Calgary, Canada David Weatherall MD, FRCP, FRS Formerly of Weatherall Institute of Molecular Medicine, The John Radcliffe Hospital, Oxford, UK

Preface to the fourth edition

Hematology is a fast-moving discipline with innovation both diagnostically and therapeutically. In the 19 years since the first edition of Molecular Hematology was published, many advances have been made. Molecular techniques have helped explain the basis of many diseases, starting initially with red cell disorders and hemostasis. However, thanks to the use of molecular biology we can now diagnose and stratify patients with diseases such as leukemia, myeloma, myeloproliferative neoplasms, and others. Such advances in technology have not only helped explain the underlying basis of the diseases, but have also provided targets for treatment. The world of red cells started the whole specialty of molecular medicine and, using molecular biology techniques, many of the phenotypic features of red cell disorders have been explained. This is discussed eloquently by the late Professor Sir David Weatherall at the beginning of the book. Other non-malignant areas which have been updated include the Functions of Blood Group Antigens, von Willebrand Disease, and Platelet Disorders. Undoubtedly, hemato-oncology has seen the biggest explosion in terms of understanding the molecular basis of diseases such as leukemias, lymphomas, the myeloprolifer-

ative neoplasms, myeloma, and myelodysplastic syndromes. The huge array of biological markers makes stratification and treatment much more sophisticated than ever before. All the chapters dealing with malignant blood diseases have been thoroughly revised and brought up to date. However, despite the growing complexity in terms of pathogenesis, diagnosis, and management of patients with blood diseases, the ethos of the book remains the same – namely, to provide a succinct account of the molecular biology of hematological disease written at a level where it should be of benefit to both the seasoned molecular biologist and the practicing clinician alike. We have retained the original structure for the chapters, high-quality artwork, and Further Reading sections in order to make the book visually appealing and relevant to modern hematology practice. We very much hope you enjoy this edition and, as always, we welcome any comments or suggestions from readers, which we will attempt to incorporate into the next edition. Drew Provan John Gribben

xiii

Further reading

Anderson, K.C. and Ness, P.M. (eds.) (2000). Scientific Basis of Transfusion Medicine: Implications for Clinical Practice, 2e. Philadelphia, PA: WB Saunders. Beutler, E. and Lichtman, M.A. (eds.). Williams’ Hematology, 6e. New York: McGraw-Hill. Cooper, G.M. (1997). The Cell: A Molecular Approach. Washington, DC: ASM Press. Cox, T.M. and Sinclair, J. (1997). Molecular Biology in Medicine. Oxford: Blackwell Science. Jameson, J.L. (ed.) (1998). Principles of Molecular Medicine. New York: Humana Press.

Mullis, K.B. (1990). The unusual origin of the polymerase chain reaction. Scientific American 262: 56–65. Roitt, I. (2001). Roitt’s Essential Immunology, 10e. Oxford: Blackwell Science. Stamatoyannopoulos, G., Nienhuis, A.W., Majerus, P.W., and Varmus, H. (eds.) (2000). The Molecular Basise of Blood Diseases, 2e. Philadelphia, PA: W.B. Saunders. Watson, J.D., Gilman, M., Witkowski, J., and Zoller, M. (eds.) (1992). Recombinant DNA, 2e. New York: Scientific American Books.

xv

Acknowledgments

We would like to express thanks to Claire Bonnett, Publisher, and Deirdre Barry, Senior Editorial Assistant, for their help with this work.

xvi

Chapter 1 Beginnings: the molecular pathology of hemoglobin David Weatherall Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Oxford, UK

Historical background, 1 The structure, genetic control, and synthesis of normal hemoglobin, 2 The molecular pathology of hemoglobin, 6 Genotype–phenotype relationships in the thalassemias, 12 Structural hemoglobin variants, 16

Historical background Linus Pauling first used the term “molecular disease” in 1949, after the discovery that the structure of sickle cell hemoglobin differed from that of normal hemoglobin. Indeed, it was this seminal observation that led to the concept of molecular medicine, the description of disease mechanisms at the level of cells and molecules. However, until the development of recombinant DNA technology in the mid-1970s, knowledge of events inside the cell nucleus, notably how genes function, could only be the subject of guesswork based on the structure and function of their protein products. However, as soon as it became possible to isolate human genes and to study their properties, the picture changed dramatically. Progress over the last 30 years has been driven by technological advances in molecular biology. At first it was possible only to obtain indirect information about the structure and function of genes by DNA/DNA and DNA/RNA hybridization; that is, by probing the quantity or structure of RNA or DNA by annealing reactions with molecular probes. The next major advance was the ability to fractionate DNA into pieces of predictable size with bacterial restriction enzymes. This led to the invention of a technique that played a central role in the early development of human molecular genetics, called Southern blotting after the name of its developer, Edwin Southern. This method allowed the structure and organization of genes to be studied directly for the first time and led to the definition of a number of different forms of molecular pathology. Once it was possible to fractionate DNA, it soon became feasible to insert the pieces into vectors able to divide Molecular Hematology, 4th Edition. Edited by Drew Provan and John G. Gribben. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

Molecular aspects of the high frequency of the hemoglobin variants, 17 Molecular aspects of the prevention and management of the hemoglobin disorders, 18 Postscript, 18 Further reading, 18

within bacteria. The steady improvement in the properties of cloning vectors made it possible to generate libraries of human DNA growing in bacterial cultures. Ingenious approaches were developed to scan the libraries to detect genes of interest; once pinpointed, the appropriate bacterial colonies could be grown to generate larger quantities of DNA carrying a particular gene. Later it became possible to sequence these genes, persuade them to synthesize their products in microorganisms, cultured cells, or even other species, and hence to define their key regulatory regions. The early work in the field of human molecular genetics focused on diseases in which there was some knowledge of the genetic defect at the protein or biochemical level. However, once linkage maps of the human genome became available, following the identification of highly polymorphic regions of DNA, it was possible to search for any gene for a disease, even where the cause was completely unknown. This approach, first called reverse genetics and later rechristened positional cloning, led to the discovery of genes for many important diseases. As methods for sequencing were improved and automated, thoughts turned to the next major goal in this field, which was to determine the complete sequence of the bases that constitute our genes and all that lies between them: the Human Genome Project. This remarkable endeavor was finally completed in 2006. The further understanding of the functions and regulation of our genes will require multidisciplinary research encompassing many different fields. The next stage in the Human Genome Project, called genome annotation, entails analyzing the raw DNA sequence in order to determine its biological significance. One of the main ventures in the era of functional genomics will be in what is termed proteomics, the large-scale analysis of the protein products of genes. The ultimate goal will be to try to define the protein

1

2 Molecular Hematology

complement, or proteome, of cells and how the many different proteins interact with one another. To this end, largescale facilities are being established for isolating and purifying the protein products of genes that have been expressed in bacteria. Their structure can then be studied by a variety of different techniques, notably X-ray crystallography and nuclear magnetic resonance spectroscopy. The crystallographic analysis of proteins is being greatly facilitated by the use of X-ray beams from a synchrotron radiation source. In the last few years both the utility and extreme complexity of the fruits of the genome project have become apparent. The existence of thousands of single-nucleotide polymorphisms (SNPs) has made it possible to search for genes of biological or medical significance. The discovery of families of regulatory RNAs and proteins is starting to shed light on how the functions of the genome are controlled, and studies of acquired changes in its structure, epigenetics, promise to provide similar information. Recent developments in new-generation sequencing of DNA and RNA are also providing invaluable information about many aspects of gene regulation. During this remarkable period of technical advance, considerable progress has been made toward an understanding of the pathology of disease at the molecular level. This has had a particular impact on hematology, leading to advances in the understanding of gene function and disease mechanisms in almost every aspect of the field. The inherited disorders of hemoglobin – the thalassemias and structural hemoglobin variants, the commonest human monogenic diseases – were the first to be studied systematically at the molecular level and a great deal is known about their genotype–phenotype relationships. This field led the way to molecular hematology and, indeed, to the development of molecular medicine. Thus, even though the genetics of hemoglobin is complicated by the fact that different varieties are produced at particular stages of human development, the molecular pathology of the hemoglobinopathies provides an excellent model system for understanding any monogenic disease and the complex interactions between genotype and environment that underlie many multigenic disorders. In this chapter I consider the structure, synthesis, and genetic control of the human hemoglobins, describe the molecular pathology of the thalassemias, and discuss briefly how the complex interactions of their different genotypes produce a remarkably diverse family of clinical phenotypes; the structural hemoglobin variants are discussed in more detail in Chapter 14. Readers who wish to learn more about the methods of molecular genetics, particularly as applied to the study of hemoglobin disorders, are referred to the reviews cited at the end of this chapter.

The structure, genetic control, and synthesis of normal hemoglobin Structure and function The varying oxygen requirements during embryonic, fetal, and adult life are reflected in the synthesis of different structural hemoglobins at each stage of human development. However, they all have the same general tetrameric structure, consisting of two different pairs of globin chains, each attached to one heme molecule. Adult and fetal hemoglobins have α chains combined with β chains (Hb A, α2 β2 ), δ chains (Hb A2 , α2 δ2 ), and γ chains (Hb F, α2 γ2 ). In embryos, α-like chains called ζ chains combine with γ chains to produce Hb Portland (ζ2 γ2 ), or with ε chains to make Hb Gower 1 (ζ2 ε2 ), while α and ε chains form Hb Gower 2 (α2 ε2 ). Fetal hemoglobin is heterogeneous; there are two varieties of γ chain that differ only in their amino acid composition at position 136, which may be occupied by either glycine or alanine; γ chains containing glycine at this position are called G γ chains, those with alanine A γ chains (Figure 1.1). The synthesis of hemoglobin tetramers consisting of two unlike pairs of globin chains is absolutely essential for the effective function of hemoglobin as an oxygen carrier. The classical sigmoid shape of the oxygen dissociation curve, which reflects the allosteric properties of the hemoglobin molecule, ensures that, at high oxygen tensions in the lungs, oxygen is readily taken up and later released effectively at the lower tensions encountered in the tissues. The shape of the curve is quite different to that of myoglobin, a molecule that consists of a single globin chain with heme attached to it, which, like abnormal hemoglobins that consist of homotetramers of like chains, has a hyperbolic oxygen dissociation curve. The transition from a hyperbolic to a sigmoid oxygen dissociation curve, which is absolutely critical for normal oxygen delivery, reflects cooperativity between the four heme molecules and their globin subunits. When one of them takes on oxygen, the affinity of the remaining three increases markedly; this happens because hemoglobin can exist in two configurations, deoxy(T) and oxy(R), where T and R represent the tight and relaxed states, respectively. The T configuration has a lower affinity than the R for ligands such as oxygen. At some point during the addition of oxygen to the hemes, the transition from the T to the R configuration occurs and the oxygen affinity of the partially liganded molecule increases dramatically. These allosteric changes result from interactions between the iron of the heme groups and various bonds within the hemoglobin tetramer, which lead to subtle spatial changes as oxygen is taken on or given up.

Beginnings: the molecular pathology of hemoglobin 3

1 Kb

31 32 99 100

ζ

ψζ

ψα2

ψα1

α2

α1

θ1

30 31

ε

Gγ

Aγ

ψβ

104

105

δ

β

16

11

ζ2ε2 Hb Gower 1

ζ2γ2 Hb Portland

α2ε2 Hb Gower 2

Embryo

α2γ2 HbF Fetus

α2δ2 HbA2

α2β2 HbA Adult

Fig. 1.1 The genetic control of human hemoglobin production in embryonic, fetal, and adult life. The standard names for these genes are as follows: Alpha genes HBA1 and HBA2, Beta gene HBB, Gamma genes HBG1 and HBG2, Delta gene HBD, and the embryonic genes HBE1 and HBZ.

The precise tetrameric structures of the different human hemoglobins, which reflect the primary amino acid sequences of their individual globin chains, are also vital for the various adaptive changes that are required to ensure adequate tissue oxygenation. The position of the oxygen dissociation curve can be modified in several ways. For example, oxygen affinity decreases with increasing CO2 tension (the Bohr effect). This facilitates oxygen loading to the tissues, where a drop in pH due to CO2 influx lowers oxygen affinity; the opposite effect occurs in the lungs. Oxygen affinity is also modified by the level of 2,3-diphosphoglycerate (2,3-BPG) in the red cell. Increasing concentrations shift the oxygen dissociation curve to the right (i.e. they reduce oxygen affinity), while diminishing concentrations have the opposite effect. 2,3-BPG fits into the gap between the two β chains when it widens during deoxygenation, and interacts with several specific binding sites in the central cavity of the molecule. In the deoxy configuration the gap between the two β chains narrows and the molecule cannot be accommodated. With increasing concentrations of 2,3-BPG, which are found in various hypoxic and anemic states, more hemoglobin molecules tend to be held in the deoxy configuration and the oxygen dissociation curve is therefore shifted to the right, with more effective release of oxygen. Fetal red cells have greater oxygen affinity than adult red cells, although, interestingly, purified fetal hemoglobin has an oxygen dissociation curve similar to that of adult hemoglobin. These differences, which are adapted to the oxygen requirements of fetal life, reflect the relative inability of Hb F to interact with 2,3-BPG compared with Hb A. This is because the γ chains of Hb F lack specific binding sites for 2,3-BPG.

In short, oxygen transport can be modified by a variety of adaptive features in the red cell that include interactions between the different heme molecules, the effects of CO2 , and differential affinities for 2,3-BPG. These changes, together with more general mechanisms involving the cardiorespiratory system, provide the main basis for physiological adaptation to anemia.

Genetic control of hemoglobin The α- and β-like globin chains are the products of two different gene families which are found on different chromosomes (Figure 1.1). The β-like globin genes form a linked cluster on chromosome 11, spread over approximately 60 kb (kilobase or 1000 nucleotide bases). The different genes that form this cluster are arranged in the order 5′ –ε–G γ–A γ– ψβ–δ–β–3′ . The α-like genes also form a linked cluster, in this case on chromosome 16, in the order 5′ –ζ–ψζ–ψα1– α2–α1–3′ . The ψβ, ψζ, and ψα genes are pseudogenes; that is, they have strong sequence homology with the β, ζ, and α genes, but contain a number of differences that prevent them from directing the synthesis of any products. They may reflect remnants of genes that were functional at an earlier stage of human evolution. The structure of the human globin genes is, in essence, similar to that of all mammalian genes. They consist of long strings of nucleotides that are divided into coding regions, or exons, and non-coding inserts called intervening sequences (IVSs) or introns. The β-like globin genes contain two introns, one of 122–130 base pairs between codons 30 and 31 and one of 850–900 base pairs between codons 104 and 105 (the exon codons are numbered sequentially from the 5′ to

4 Molecular Hematology

the 3′ end of the gene, i.e. from left to right). Similar, though smaller, introns are found in the α and ζ globin genes. These introns and exons, together with short non-coding sequences at the 5′ and 3′ ends of the genes, represent the major functional regions of the particular genes. However, there are also extremely important regulatory sequences which subserve these functions that lie outside the genes themselves. At the 5′ non-coding (flanking) regions of the globin genes, as in all mammalian genes, there are blocks of nucleotide homology. The first, the ATA box, is about 30 bases upstream (to the left) of the initiation codon; that is, the start word for the beginning of protein synthesis (see later). The second, the CCAAT box, is about 70 base pairs upstream from the 5′ end of the genes. About 80–100 bases further upstream there is the sequence GGGGTG, or CACCC, which may be inverted or duplicated. These three highly conserved DNA sequences, called promoter elements, are involved in the initiation of transcription of the individual genes. Finally, in the 3′ non-coding region of all the globin genes there is the sequence AATAAA, which is the signal for cleavage and polyA addition to RNA transcripts (see Gene Action and Globin Synthesis). The globin gene clusters also contain several sequences that constitute regulatory elements, which interact to promote erythroid-specific gene expression and coordination of the changes in globin gene activity during development. These include the globin genes themselves and their

C A C C C

C C A A T

T A T A

Gene action and globin synthesis The flow of information between DNA and protein is summarized in Figure 1.2. When a globin gene is transcribed, messenger RNA (mRNA) is synthesized from one of its strands, a process which begins with the formation of a transcription complex consisting of a variety of regulatory proteins together with an enzyme called RNA polymerase (see later). The primary transcript is a large mRNA precursor which contains both intron and exon sequences. While in the nucleus, this molecule undergoes a variety of modifications. First, the introns are removed and the exons are spliced together. The intron/exon junctions always have the same sequence: GT at their 5′ end, and AG at their 3′ end. This appears to be essential for accurate splicing; if there is a mutation at these sites this process does not occur. Splicing reflects a complex series of intermediary stages and the interaction

A A T T A A A A A

A T G

Flanking

IVS 1

NC

promoter elements: enhancers (regulatory sequences that increase gene expression despite being located at a considerable distance from the genes) and “master” regulatory sequences called, in the case of the β globin gene cluster, the locus control region (LCR) and, in the case of the α genes, HS40 (a nuclease-hypersensitive site in DNA 40 kb from the α globin genes). Each of these sequences has a modular structure made up of an array of short motifs that represent the binding sites for transcriptional activators or repressors.

Flanking

IVS 2

GT AG

GT

AG

5'

3'

5' CAP

AAAA-A AAAA-A

Nucleus

Gene

NC mRNA precursor

Excision of introns Splicing of exons Processed mRNA

Cytoplasm

Ribosome

AUG U AC

UG ACC UUC G AAG

AAAA-A

Translation

UAA

Transfer RNA Amino acid Growing chain Processed chain

Finished chain

Fig. 1.2 The mechanisms of globin gene transcription and translation.

Beginnings: the molecular pathology of hemoglobin 5

of a number of different nuclear proteins. After the exons are joined, the mRNAs are modified and stabilized; at their 5′ end a complex CAP structure is formed, while at their 3′ end a string of adenylic acid residues (polyA) is added. The mRNA processed in this way moves into the cytoplasm, where it acts as a template for globin chain production. Because of the rules of base pairing – that is, cytosine always pairs with thymine, and guanine with adenine – the structure of the mRNA reflects a faithful copy of the DNA codons from which it is synthesized; the only difference is that, in RNA, uracil (U) replaces thymine (T). Amino acids are transported to the mRNA template on carriers called transfer RNAs (tRNAs); there are specific tRNAs for each amino acid. Furthermore, because the genetic code is redundant (i.e. more than one codon can encode a particular amino acid), for some of the amino acids there are several different individual tRNAs. Their order in the globin chain is determined by the order of codons in the mRNA. The tRNAs contain three bases, which together constitute an anticodon; these anticodons are complementary to mRNA codons for particular amino acids. They carry amino acids to the template, where they find the appropriate positioning by codon–anticodon base pairing. When the first tRNA is in position, an initiation complex is formed between several protein initiation factors together with the two subunits that constitute the ribosomes. A second tRNA moves in alongside and the two amino acids that they are carrying form a peptide bond between them; the globin chain is now two amino acid residues long. This process is continued along –20

the mRNA from left to right, and the growing peptide chain is transferred from one incoming tRNA to the next; that is, the mRNA is translated from 5′ to 3′ . During this time the tRNAs are held in appropriate steric configuration with the mRNA by the two ribosomal subunits. There are specific initiation (AUG) and termination (UAA, UAG, and UGA) codons. When the ribosomes reach the termination codon, translation ceases, the completed globin chains are released, and the ribosomal subunits are recycled. Individual globin chains combine with heme, which has been synthesized through a separate pathway, and then interact with one like chain and two unlike chains to form a complete hemoglobin tetramer.

Regulation of hemoglobin synthesis The regulation of globin gene expression is mediated mainly at the transcriptional level, with some fine tuning during translation and post-translational modification of the gene products. DNA that is not involved in transcription is held tightly packaged in a compact, chemically modified form that is inaccessible to transcription factors and polymerases and is heavily methylated. Activation of a particular gene is reflected by changes in the structure of the surrounding chromatin, which can be identified by enhanced sensitivity to nucleases. Erythroid lineage-specific nuclease-hypersensitive sites are found at several locations in the β globin gene cluster. Four are distributed over 20 kb upstream from the ε globin gene in the region of the β globin LCR (Figure 1.3). This vital

0

20

40

60 kb

5' HS 4

3

2

3' HS 1

1

ε

Chromosome 11

Gγ

Aγ

δ

β

LCR

–40

–20

0

20

40 kb

HS–40

ζ

α2

α1

Chromosome 16 Fig. 1.3 The positions of the major regulatory regions in the β and α globin gene clusters. The arrows indicate the position of the erythroid lineage-specific nuclease-hypersensitive sites. HS, hypersensitive.

6 Molecular Hematology

regulatory region is able to establish a transcriptionally active domain spanning the entire β globin gene cluster. Several enhancer sequences have been identified in this cluster. A variety of regulatory proteins bind to the LCR, and to the promoter regions of the globin genes and to the enhancer sequences. It is thought that the LCR and other enhancer regions become opposed to the promoters to increase the rate of transcription of the genes to which they are related. These regulatory regions contain sequence motifs for various ubiquitous and erythroid-restricted transcription factors. Binding sites for these factors have been identified in each of the globin gene promoters and at the hypersensitive site regions of the various regulatory elements. A number of the factors which bind to these areas are found in all cell types. They include Sp1, Yy1, and Usf. In contrast, a number of transcription factors have been identified, including GATA-1, EKLF, and NF-E2, which are restricted in their distribution to erythroid cells and, in some cases, megakaryocytes, and mast cells. The overlapping of erythroid-specific and ubiquitous-factor binding sites in several cases suggests that competitive binding may play an important part in the regulation of erythroid-specific genes. Another binding factor, SSP, the stage selector protein, appears to interact specifically with ε and γ genes. Several elements involving the chromatin and histone acetylation required for access of these regulatory proteins have been identified. The binding of hematopoietic-specific factors activates the LCR, which renders the entire β globin gene cluster transcriptionally active. These factors also bind to the enhancer and promoter sequences, which work in tandem to regulate the expression of the individual genes in the clusters. It is likely that some of the transcriptional factors are developmental stage specific, and hence may be responsible for the differential expression of the embryonic, fetal, and adult globin genes. The α globin gene cluster also contains an element, HS40, which has some structural features in common with the β LCR, although it is different in aspects of its structure. A number of enhancer-like sequences have been identified, although it is becoming clear that there are fundamental differences in the pattern of regulation of the two globin gene clusters. In addition to the different regulatory sequences outlined, there are also sequences which may be involved specifically with “silencing” of genes, notably those for the embryonic hemoglobins, during development. Some degree of regulation is mediated by differences in the rates of initiation and translation of the different mRNAs, and at the post-transcriptional level by differential affinity for different protein subunits. However, this kind of posttranscriptional fine tuning probably plays a relatively small role in determining the overall output of the globin gene products.

Regulation of developmental changes in globin gene expression During development, the site of red cell production moves from the yolk sac to the fetal liver and spleen, and thence to bone marrow in the adult. Embryonic, fetal, and adult hemoglobin synthesis is approximately related in time to these changes in the site of erythropoiesis, although it is quite clear that the various switches, between embryonic and fetal and between fetal and adult hemoglobin synthesis, are beautifully synchronized throughout these different sites. Fetal hemoglobin synthesis declines during the later months of gestation, and Hb F is replaced by Hb A and Hb A2 by the end of the first year of life. Although the exact mechanism of the switch from fetal to adult hemoglobin is still not understood, recent studies of patients with unusually high levels of HbF and genome-wide association studies (GWAS) have yielded extremely promising information about some of the regulatory genes involved. They include BCL11A, MYB, and KLF1. It is clear from studies of these and related genes that they are involved directly in the regulation of hemoglobin switching, work which is yielding great promise for the development of future technology for increasing HbF synthesis to modify the phenotype of thalassemias and sickle cell anemia.

The molecular pathology of hemoglobin As is the case for many monogenic diseases, the inherited disorders of hemoglobin fall into two major classes. First, there are those that result from reduced output of one or other globin genes, the thalassemias. Second, there is a wide range of conditions that result from the production of structurally abnormal globin chains; the type of disease depends on how the particular alteration in protein structure interferes with its stability or function. Of course, no biological classification is entirely satisfactory and those which attempt to define the hemoglobin disorders are no exception. There are some structural hemoglobin variants which happen to be synthesized at a reduced rate and hence are associated with a clinical picture similar to thalassemia. And there are other classes of mutations which simply interfere with the normal transition from fetal to adult hemoglobin synthesis, a family of conditions given the general title hereditary persistence of fetal hemoglobin (HPFH). Furthermore, because these diseases are all so common and occur together in particular populations, it is not uncommon for an individual to inherit a gene for one or other form of thalassemia and a structural hemoglobin variant. The heterogeneous group of conditions that results from these different mutations and interactions is summarized in Table 1.1.

Beginnings: the molecular pathology of hemoglobin 7

Defective β globin gene transcription

Table 1.1 The thalassemias and related disorders α Thalassemia α0 α+ Deletion (−α) Non-deletion (αT )

γ Thalassemia

β Thalassemia β0 β+ Normal Hb A2 “Silent” Dominant

Hereditary persistence of fetal hemoglobin Deletion (δβ)0 Non-deletion Linked to β globin genes

δ Thalassemia εγδβ Thalassemia

G γβ+

δβ Thalassemia (δβ)+ (δβ)0 (A γδβ)0

A γβ+

Unlinked to β globin genes

Over recent years, determination of the molecular pathology of the two common forms of thalassemia, α and β, has provided a remarkable picture of the repertoire of mutations that can underlie human monogenic disease. In the sections that follow I describe, in outline, the different forms of molecular pathology that underlie these conditions.

The β thalassemias There are two main classes of β thalassemia, β0 thalassemia, in which there is an absence of β globin chain production, and β+ thalassemia, in which there is a variable reduction in the output of β globin chains. As shown in Figure 1.4, mutations of the β globin genes may cause a reduced output of gene product at the level of transcription or mRNA processing, or translation, or through the stability of the globin gene product.

There are a variety of mechanisms that interfere with normal transcription of the β globin genes. First, the genes may be either completely or partially deleted. Overall, deletions of the β globin genes are not commonly found in patients with β thalassemia, with one exception: a 619-bp deletion involving the 3′ end of the gene is found frequently in the Sind populations of India and Pakistan, where it constitutes about 30% of the β thalassemia alleles. Other deletions are extremely rare. A much more common group of mutations, which results in a moderate decrease in the rate of transcription of the β globin genes, involves single-nucleotide substitutions in or near the TATA box at about −30 nucleotides (nt) from the transcription start site, or in the proximal or distal promoter elements at −90 and −105 nt. These mutations result in decreased β globin mRNA production, ranging from 10% to 25% of the normal output. Thus, they are usually associated with the mild forms of β+ thalassemia. They are particularly common in African populations, an observation which explains the unusual mildness of β thalassemia in this racial group. One particular mutation, C → T at position −101 nt to the β globin gene, causes an extremely mild deficit of β globin mRNA. Indeed, this allele is so mild that it is completely silent in carriers and can only be identified by its interaction with more severe β thalassemia alleles in compound heterozygotes.

Mutations that cause abnormal processing of mRNA As mentioned earlier, the boundaries between exons and introns are marked by the invariant dinucleotides GT at the donor (5′ ) site and AG at the acceptor (3′ ) site. Mutations that affect either of these sites completely abolish normal splicing

Deletions

I

PR

C

I FS SPL NS

IVSI

SPL

2

FS NS

3

IVS 2

SPL

Point mutations

SPL

FS NS

Poly A

100 bp

Fig. 1.4 The mutations of the β globin gene that underlie β thalassemia. The heavy black lines indicate the length of the deletions. The point mutations are designated as follows: PR, promoter; C, CAP site; I, initiation codon; FS, frameshift and nonsense mutations; SPL, splice mutations; Poly A, poly A addition site mutations.

8 Molecular Hematology

and produce the phenotype of β0 thalassemia. The transcription of genes carrying these mutations appears to be normal, but there is complete inactivation of splicing at the altered junction. Another family of mutations involves what are called splice site consensus sequences. Although only the GT dinucleotide is invariant at the donor splice site, there is conservation of adjacent nucleotides and a common, or consensus, sequence of these regions can be identified. Mutations within this sequence can reduce the efficiency of splicing to varying degrees, because they lead to alternate splicing at the surrounding cryptic sites. For example, mutations of the nucleotide at position 5 of IVS-1 (the first intervening sequence), G → C or T, result in a marked reduction of β chain production and in the phenotype of severe β+ thalassemia. On the other hand, the substitution of C for T at position 6 in IVS-1 leads to only a mild reduction in the output of β chains. Another mechanism that leads to abnormal splicing involves cryptic splice sites. These are regions of DNA which, if mutated, assume the function of a splice site at an inappropriate region of the mRNA precursor. For example, a variety of mutations activate a cryptic site which spans codons 24– 27 of exon 1 of the β globin gene. This site contains a GT dinucleotide, and adjacent substitutions that alter it so that it more closely resembles the consensus donor splice site result in its activation, even though the normal splice site is intact. A mutation at codon 24 GGT → GGA, though it does not alter the amino acid which is normally found in this position in the β globin chain (glycine), allows some splicing to occur at this site instead of the exon–intron boundary. This results in the production of both normal and abnormally spliced β

globin mRNA and hence in the clinical phenotype of severe β thalassemia. Interestingly, mutations at codons 19, 26, and 27 result in both reduced production of normal mRNA (due to abnormal splicing) and an amino acid substitution when the mRNA which is spliced normally is translated into protein. The abnormal hemoglobins produced are Hb Malay, Hb E, and Hb Knossos, respectively. All these variants are associated with a mild β+ thalassemia-like phenotype. These mutations illustrate how sequence changes in coding rather than IVS influence RNA processing, and underline the importance of competition between potential splice site sequences in generating both normal and abnormal varieties of β globin mRNA. Cryptic splice sites in introns may also carry mutations that activate them even though the normal splice sites remain intact. A common mutation of this kind in Mediterranean populations involves a base substitution at position 110 in IVS-1. This region contains a sequence similar to a 3′ acceptor site, though it lacks the invariant AG dinucleotide. The change of the G to A at position 110 creates this dinucleotide. The result is that about 90% of the RNA transcript splices to this particular site and only 10% to the normal site, again producing the phenotype of severe β+ thalassemia (Figure 1.5). Several other β thalassemia mutations have been described which generate new donor sites within IVS-2 of the β globin gene. Another family of mutations that interferes with β globin gene processing involves the sequence AAUAAA in the 3′ untranslated regions, which is the signal for cleavage and polyadenylation of the β globin gene transcript. Somehow, these mutations destabilize the transcript. For example, a T → C substitution in this sequence leads to only one-tenth of

Normal splicing

β gene

GT

IVS1

AG

GT

IVS2

AG

TTGGTCT A

10% β+ thalassemia 90%

Fig. 1.5 The generation of a new splice site in an intron as the mechanism for a form of β+ thalassemia. For details see text.

Beginnings: the molecular pathology of hemoglobin 9

the normal amount of β globin mRNA transcript, and hence to the phenotype of a moderately severe β+ thalassemia. Another example of a mutation which probably leads to defective processing of the function of β globin mRNA is the single-base substitution A → C in the CAP site. It is not yet understood how this mutation causes a reduced rate of transcription of the β globin gene. There is another small subset of rare mutations that involve the 3′ untranslated region of the β globin gene, and these are associated with relatively mild forms of β thalassemia. It is thought that these interfere in some way with transcription, but the mechanism is unknown.

Mutations that result in abnormal translation of β globin mRNA There are three main classes of mutations of this kind. Base substitutions that change an amino acid codon to a chain termination codon prevent the translation of β globin mRNA and result in the phenotype of β0 thalassemia. Several mutations of this kind have been described; the commonest, involving codon 17, occurs widely throughout Southeast Asia. Similarly, a codon 39 mutation is encountered frequently in the Mediterranean region. The second class involves the insertion or deletion of one, two, or four nucleotides in the coding region of the β globin gene. These disrupt the normal reading frame, cause a frameshift, and hence interfere with the translation of β globin mRNA. The end result is the insertion of anomalous amino acids after the frameshift until a termination codon is reached in the new reading frame. This type of mutation always leads to the phenotype of β0 thalassemia. Finally, there are several mutations which involve the β globin gene initiation codon and which, presumably, reduce the efficiency of translation.

absence of mRNA from the cytoplasm of red cell precursors. This appears to be an adaptive mechanism, called nonsensemediated decay, whereby abnormal mRNA of this type is not transported to the cytoplasm, where it would act as a template for the production of truncated gene products. However, in the case of exon III mutations, apparently because this process requires the presence of an intact upstream exon, the abnormal mRNA is transported into the cytoplasm and hence can act as a template for the production of unstable β globin chains. The latter precipitate in the red cell precursors together with excess α chains to form large inclusion bodies, and hence there is enough globin chain imbalance in heterozygotes to produce a moderately severe degree of anemia.

The α thalassemias The molecular pathology of the α thalassemias is more complicated than that of the β thalassemias, simply because there are two α globin genes per haploid genome. Thus, the normal α globin genotype can be written αα/αα. As in the case of β thalassemia, there are two major varieties of α thalassemia, α+ and α0 thalassemia. In α+ thalassemia one of the linked α globin genes is lost, either by deletion (−) or mutation (T); the heterozygous genotype can be written –α/αα or αT α/αα. In α0 thalassemia the loss of both α globin genes nearly always results from a deletion; the heterozygous genotype is therefore written − −/αα. In populations where specific deletions are particularly common, Southeast Asia (SEA) or the Mediterranean region (MED), it is useful to add the appropriate superscript as follows: – –SEA /αα or – –MED /αα. It follows that when we speak of an “α thalassemia gene,” what we are really referring to is a haplotype; that is, the state and function of both of the linked α globin genes. α0 thalassemia

Unstable β globin chain variants Some forms of β thalassemia result from the synthesis of highly unstable β globin chains which are incapable of forming hemoglobin tetramers and which are rapidly degraded, leading to the phenotype of β0 thalassemia. Indeed, in many of these conditions no abnormal globin chain product can be demonstrated by protein analysis, and the molecular pathology has to be interpreted simply on the basis of a derived sequence of the variant β chain obtained by DNA analysis. Recent studies have provided some interesting insights into how complex clinical phenotypes may result from the synthesis of unstable β globin products. For example, there is a spectrum of disorders that result from mutations in exon 3 which give rise to a moderately severe form of β thalassemia in heterozygotes. It has been found that nonsense or frameshift mutations in exons I and II are associated with the

Three main molecular pathologies, all involving deletions, have been found to underlie the α0 thalassemia phenotype. The majority of cases result from deletions that remove both α globin genes and a varying length of the α globin gene cluster (Figure 1.6). Occasionally, however, the α globin gene cluster is intact, but is inactivated by a deletion which involves the major regulatory region HS40, 40 kb upstream from the α globin genes, or the α globin genes may be lost as part of a truncation of the tip of the short arm of chromosome 16. As well as providing us with an understanding of the molecular basis for α0 thalassemia, detailed studies of these deletions have yielded more general information about the mechanisms that underlie this form of molecular pathology. For example, it has been found that the 5′ breakpoints of a number of deletions of the α globin gene cluster are located approximately the same distance apart and in the same order

10 Molecular Hematology

–50

–10

0 ζ2

10 ψζ1 ψα2 ψα1

20 α2

30 α1

Inter-ζHVR

θ1 3'HVR

Fig. 1.6 Some of the deletions that underlie α0 and α+ thalassemia. The colored rectangles beneath the α globin gene cluster indicate the lengths of the deletions. The unshaded regions indicate uncertainty about the precise breakpoints. The three small deletions at the bottom of the figure represent the common α+ thalassemia deletions. HVR, highly variable regions.

along the chromosome as their respective 3′ breakpoints; similar findings have been observed in deletions of the β globin gene cluster. These deletions seem to have resulted from illegitimate recombination events, which have led to the deletion of an integral number of chromatin loops as they pass through their nuclear attachment points during chromosomal replication. Another long deletion has been characterized in which a new piece of DNA bridges the two breakpoints in the α globin gene cluster. The inserted sequence originates upstream from the α globin gene cluster, where normally it is found in an inverted orientation with respect to that found between the breakpoints of the deletion. Thus it appears to have been incorporated into the junction in a way that reflects its close proximity to the deletion breakpoint region during replication. Other deletions seem to be related to the family of Alu-repeats, simple repeat sequences that are widely dispersed throughout the genome; one deletion appears to have resulted from a simple homologous recombination between two repeats of this kind that are usually 62 kb apart. A number of forms of α0 thalassemia result from terminal truncations of the short arm of chromosome 16 to a site about 50 kb distal to the α globin genes. The telomeric consensus sequence TTAGGGn has been added directly to the site of the break. Since these mutations are stably inherited, it appears that telomeric DNA alone is sufficient to stabilize the ends of broken chromosomes. Quite recently, two other molecular mechanisms have been identified as the cause of α0 thalassemia which, though rare, may have important implications for an understanding of the molecular pathology of other genetic diseases. In one case, a deletion in the α globin gene cluster resulted in a widely expressed gene (LUC7L) becoming juxtaposed to a structurally normal α globin gene. Although the latter retained all its important regulatory elements, its expression was silenced. It was found in a transgenic mouse model that transcription of antisense RNA mediated the silencing of the

α globin gene region, a finding that provides a completely new mechanism for genetic disease. In another case of α0 thalassemia, in which no molecular defects could be detected in the α globin gene cluster, a gain-of-function regulatory polymorphism was found in the region between the α globin genes and their upstream regulatory elements. This alteration creates a new promoter-like element that interferes with the normal activation of all downstream α-like globin genes. In short, detailed analysis of the molecular pathology of the α0 thalassemias has provided valuable evidence not only about how large deletions of gene clusters are caused, but also about some of the complex mechanisms that may underlie cases in which the α gene clusters remain intact, but in which their function is completely suppressed. α+ thalassemia

As mentioned earlier, the α+ thalassemias result from the inactivation of one of the duplicated α globin genes, by either deletion or point mutation. 𝛼 + Thalassemia due to gene deletions. There are two common forms of α+ thalassemia that are due to loss of one or other of the duplicated α globin genes, −α3.7 and −α4.2 , where 3.7 and 4.2 indicate the size of the deletions. The way in which these deletions have been generated, approximately 4 kb long, was probably generated by an ancient duplication event. The homologous regions, which are divided by small inserts, are designated X, Y, and Z. The duplicated Z boxes are 3.7 kb apart and the X boxes are 4.2 kb apart. The result of misalignment reflects the underlying structure of the α globin gene complex (Figure 1.7). Each α gene lies within a boundary of homology and reciprocal crossover between these segments at meiosis, a chromosome is produced with either a single (−α) or triplicated (ααα) α globin gene. As shown in Figure 1.7, if a crossover occurs between homologous Z boxes 3.7 kb of DNA are lost, an event which is described as a rightward deletion, −α3.7 . A similar crossover

Beginnings: the molecular pathology of hemoglobin 11

ψα1

α2

X

(a)

Y

Z

ψα1

ψα1

X

α2

α2

α1

Y α1

Z 3.7 αααanti

α1

3.7

–α (b) Rightward crossover ψα1 Fig. 1.7 Mechanisms of the generation of the common deletion forms of α+ thalassemia. (a) The normal arrangement of the α globin genes, with the regions of homology X, Y, and Z. (b) The crossover that generates the −α3.7 deletion. (c) The crossover that generates the −α4.2 deletion.

ψα1

α2

α2

α1

4.2 αααanti

α1

4.2

–α (c) Leftward crossover

between the two X boxes deletes 4.2 kb, the leftward deletion −α4.2 . The corresponding triplicated α gene arrangements are called αααanti 3.7 and αααanti 4.2 . A variety of different points of crossing over within the Z boxes give rise to different length deletions, still involving 3.7 kb. Non-deletion types of 𝛼 + thalassemia. These disorders result from single or oligonucleotide mutations of the particular α globin gene. Most of them involve the α2 gene but, since the output from this locus is two to three times greater than that from the α1 gene, this may simply reflect ascertainment bias due to the greater phenotypic effect and, possibly, a greater selective advantage. Overall, these mutations interfere with α globin gene function in a similar way to those that affect the β globin genes. They affect the transcription, translation, or posttranslational stability of the gene product. Since the principles are the same as for β thalassemia, we do not need to describe them in detail, with one exception, a mutation which has not been observed in the β globin gene cluster. It turns out that there is a family of mutations that involves the α2 globin gene termination codon, TAA. Each specifically changes this codon so that an amino acid is inserted instead of the chain terminating. This is followed by “readthrough” of α globin mRNA, which is not normally translated until another in-phase termination codon is reached. The result is an elongated α chain with 31 additional residues at the C-terminal end. Five hemoglobin variants of this type have been identified. The commonest, Hb Constant Spring, occurs at a high frequency in many parts of Southeast Asia. It is not absolutely clear why the read-through of normally untranslated mRNAs leads to a reduced output from the α2

gene, although there is considerable evidence that it in some way destabilizes the mRNA. α thalassemia/mental retardation syndromes

There is a family of mild forms of α thalassemia which is quite different to that described in the previous section and which is associated with varying degrees of mental retardation. Recent studies indicate that there are two quite different varieties of this condition, one encoded on chromosome 16 (ATR-16) and the other on the X chromosome (ATR-X). The ATR-16 syndrome is characterized by relatively mild mental disability with a variable constellation of facial and skeletal dysmorphisms. These individuals have long deletions involving the α globin gene cluster, but removing at least 1–2 Mb. This condition can arise in several ways, including unbalanced translocation involving chromosome 16, truncation of the tip of chromosome 16, and the loss of the α globin gene cluster and parts of its flanking regions by other mechanisms. The ATR-X syndrome results from mutations in a gene on the X chromosome, Xq13.1–q21.1. The product of this gene is one of a family of proteins involved in chromatin-mediated transcriptional regulation. It is expressed ubiquitously during development and at interphase it is found entirely within the nucleus in association with pericentromeric heterochromatin. In metaphase, it is similarly found close to the centromeres of many chromosomes but, in addition, occurs at the stalks of acrocentric chromosomes, where the sequences for ribosomal RNA are located. These locations provide important clues to the potential role of this protein in the

12 Molecular Hematology

establishment and/or maintenance of methylation of the genome. Although it is clear that ATR-X is involved in α globin transcription, it also must be an important player in early fetal development, particularly of the urogenital system and brain. Many different mutations of this gene have been discovered in association with the widespread morphological and developmental abnormalities which characterize the ATR-X syndrome. α thalassemia and the myelodysplastic syndrome

Since the first description of Hb H (see later section) in the red cells of a patient with leukemia, many examples of this association have been reported. The condition usually is reflected in a mild form of Hb H disease, with typical Hb H inclusions in a proportion of the red cells and varying amounts of Hb H demonstrable by hemoglobin electrophoresis. The hematological findings are usually those of one or other form of the myelodysplastic syndrome. The condition occurs predominantly in males in older age groups. Very recently it has been found that some patients with this condition have mutations involving ATR-X. The relationship of these mutations to the associated myelodysplasia remains to be determined.

Rarer forms of thalassemia and related disorders There are a variety of other conditions that involve the β globin gene cluster which, although less common than the β thalassemias, provide some important information about mechanisms of molecular pathology and therefore should be mentioned briefly. The δβ thalassemias

Like the β thalassemias, the δβ thalassemias, which result from defective δ and β chain synthesis, are subdivided into the (δβ)+ and (δβ)0 forms. The (δβ)+ thalassemias result from unequal crossover between the δ and β globin gene loci at meiosis with the production of δβ fusion genes. The resulting δβ fusion chain products combine with α chains to form a family of hemoglobin variants called the hemoglobin Lepores, after the family name of the first patient of this kind to be discovered. Because the synthesis of these variants is directed by genes with the 5′ sequences of the δ globin genes, which have defective promoters, they are synthesized at a reduced rate and result in the phenotype of a moderately severe form of δβ thalassemia. The (δβ)0 thalassemias nearly all result from long deletions involving the β globin gene complex. Sometimes they involve the A γ globin chains and hence the only active locus remaining is the G γ locus. In other cases the G γ and A γ loci

are left intact and the deletion simply removes the δ and β globin genes; in these cases both the G γ and the A γ globin gene remain functional. For some reason, these long deletions allow persistent synthesis of the γ globin genes at a relatively high level during adult life, which helps to compensate for the absence of β and δ globin chain production. They are classified according to the kind of fetal hemoglobin that is produced, and hence into two varieties, G γ(A γδβ)0 and G γA γ(δβ)0 thalassemia; in line with other forms of thalassemia, they are best described by what is not produced: (A γδβ)0 and (δβ)0 thalassemia, respectively. Homozygotes produce only fetal hemoglobin, while heterozygotes have a thalassemic blood picture together with about 5–15% Hb F. Hereditary persistence of fetal hemoglobin

Genetically determined persistent fetal hemoglobin synthesis in adult life is of no clinical importance, except that its genetic determinants can interact with the β thalassemias or structural hemoglobin variants; the resulting high level of Hb F production often ameliorates these conditions. The different forms of HPFH result from either long deletions involving the δβ globin gene cluster, similar to those that cause (δβ)0 thalassemia, or point mutations that involve the promoters of the G γ or A γ globin gene. In the former case there is no β globin chain synthesis and therefore these conditions are classified as (δβ)0 HPFH. In cases in which there are promoter mutations involving the γ globin genes, there is increased γ globin chain production in adult life associated with some β and δ chain synthesis in cis (i.e. directed by the same chromosome) to the HPFH mutations. Thus, depending on whether the point mutations involve the promoter of the G γ or A γ globin gene, these conditions are called G γ β+ HPFH and A γ β+ HPFH, respectively. There is another family of HPFH-like disorders in which the genetic determinant is not encoded in the β chain cluster. In one case the determinant encodes on chromosome 6, although its nature has not yet been determined. It should be pointed out that all these conditions are very heterogeneous and that many different deletions or point mutations have been discovered that produce the rather similar phenotypes of (δβ)0 or G γ or A γ β+ HPFH.

Genotype–phenotype relationships in the thalassemias It is now necessary briefly to relate the remarkably diverse molecular pathology described in the previous sections to the phenotypes observed in patients with these diseases. It is not possible to describe all these complex issues here. Rather, I shall focus on those aspects that illustrate the more general principles of how abnormal gene action is reflected

Beginnings: the molecular pathology of hemoglobin 13

in a particular clinical picture. Perhaps the most important question that I will address is why patients with apparently identical genetic lesions have widely differing disorders, a problem that still bedevils the whole field of medical genetics, even in the molecular era.

The β thalassemias As we have seen, the basic defect that results from the 200 or more different mutations that underlie these conditions is reduced β globin chain production. Synthesis of the α globin chain proceeds normally and hence there is imbalanced globin chain output with an excess of α chains (Figure 1.8). Unpaired α chains precipitate in both red cell precursors and their progeny with the production of inclusion bodies. These interfere with normal red cell maturation and survival in a variety of complex ways. Their attachment to the red cell

membrane causes alterations in its structure, and their degradation products, notably heme, hemin (oxidized heme), and iron, result in oxidative damage to the red cell contents and membrane. These interactions result in intramedullary destruction of red cell precursors and in shortened survival of such cells as they reach the peripheral blood. The end result is anemia of varying severity. This, in turn, causes tissue hypoxia and the production of relatively large amounts of erythropoietin; this leads to a massive expansion of the ineffective bone marrow, resulting in bone deformity, a hypermetabolic state with wasting and malaise, and bone fragility. A large proportion of hemoglobin in the blood of β thalassemics is of the fetal variety. Normal individuals produce about 1% of Hb F, unevenly distributed among their red cells. In the bone marrow of β thalassemics, any red cell precursors that synthesize γ chains come under strong selection because they combine with α chains to produce fetal hemoglobin, and α

Ex ce ss

γ α2γ2 HbF

Denaturation Degradation

Selective survival of HbF-containing precursors

Increased levels of HbF in red cells

β

Hemolysis

Destruction of RBC precursors

Splenomegaly (pooling, plasma volume expansion)

Ineffective erythropoiesis

High oxygen affinity of red cells R O ed u c 2 d eli ed ve ry

Anemia

Erythropoietin

e su a Tis poxi y h Transfusion

Marrow expansion

Skeletal deformity Increased metabolic rate Wasting Gout Folate deficiency

Fig. 1.8 The pathophysiology of β thalassemia.

Inc r ab ease so d rp iro tio n n Iron loading

Endocrine deficiencies Cirrhosis Cardiac failure

14 Molecular Hematology

therefore the degree of globin chain imbalance is reduced. Furthermore, the likelihood of γ chain production seems to be increased in a highly stimulated erythroid bone marrow. It seems likely that these two factors combine to increase the relative output of Hb F in this disorder. However, it has a higher oxygen affinity than Hb A, and hence patients with β thalassemia are not able to adapt to low hemoglobin levels as well as those who have adult hemoglobin. The greatly expanded, ineffective erythron leads to an increased rate of iron absorption; this, combined with iron received by blood transfusion, leads to progressive iron loading of the tissues, with subsequent liver, cardiac, and endocrine damage. The constant bombardment of the spleen with abnormal red cells leads to its hypertrophy. Hence there is progressive splenomegaly with an increased plasma volume and trapping of part of the circulating red cell mass in the spleen. This leads to worsening of the anemia. All these pathophysiological mechanisms, except for iron loading, can be reversed by regular blood transfusion which, in effect, shuts off the ineffective bone marrow and its consequences. Thus it is possible to relate nearly all the important features of the severe forms of β thalassemia to the primary defect in globin gene action. However, can we also explain their remarkable clinical diversity?

Phenotypic diversity Although the bulk of patients who are homozygous for β thalassemia mutations or compound heterozygotes for two different mutations have a severe transfusion-dependent phenotype, there are many exceptions. Some patients of this type have a milder course, requiring few or even no transfusions, a condition called β thalassemia intermedia. A particularly important example of this condition is illustrated by the clinical findings in those who inherit β thalassemia from one parent and Hb E from the other, a disorder called Hb E/β thalassemia. Because the mutation that produces Hb E also opens up an alternative splice site in the first exon of the β globin gene, it is synthesized at a reduced rate and therefore behaves like a mild form of β thalassemia. It is the commonest hemoglobin variant globally and Hb E/β thalassemia is the commonest form of severe thalassemia in many Asian countries. It has an extraordinarily variable phenotype, ranging from a condition indistinguishable from β thalassemia major to one of such mildness that patients grow and develop quite normally and never require transfusion. Over recent years a great deal has been learned about some of the mechanisms involved in this remarkable phenotypic variability. In short, it reflects both the action of modifying genes and variability in adaptation to anemia and, almost certainly, the effects of the environment. Given the complexity of these interactions, it is helpful to divide the genetic modifiers

Table 1.2 Mechanisms for the phenotypic diversity of the β thalassemias Genetic modifiers Primary: alleles of varying severity Secondary: modifiers of globin chain imbalance α Thalassemia Increased α globin genes: ααα or αααα Genes involved in unusually high Hb F response Tertiary: modifiers of complications Iron absorption, bone disease, jaundice, infection Adaptation to anemiaa Variation in oxygen affinity (P50 ) of hemoglobin Variation in erythropoietin response to anemia Environmental Nutrition Infection Others a There

may be genetic variation in the adaptive mechanisms.

of the β thalassemia phenotype into primary, secondary, and tertiary classes (Table 1.2). The primary modifiers are the different β thalassemia alleles that can interact together. For example, compound heterozygotes for a severe β0 thalassemia mutation and a milder one may have an intermediate form of β thalassemia of varying severity, depending on the degree of reduction in β globin synthesis under the action of the milder allele. This is undoubtedly one mechanism for the varying severity of Hb E/β thalassemia; it simply reflects the variable action of the β thalassemia mutation that is inherited together with Hb E. However, this explanation is not relevant in cases in which patients with identical β thalassemia mutations have widely disparate phenotypes. The secondary modifiers are those which directly affect the degree of globin chain imbalance. Patients with β thalassemia who also inherit one or other form of α thalassemia tend to have a milder phenotype because of the reduction in the excess of α globin genes caused by the coexistent α thalassemia allele. Similarly, patients with severe forms of thalassemia who inherit more α genes than normal because their parents have triplicated or quadruplicated α gene arrangements tend to have more severe phenotypes. Other patients with severe thalassemia alleles appear to run a milder course because of a genetically determined ability to produce more γ globin chains and hence fetal hemoglobin, a mechanism that also results in a reduced degree of globin chain imbalance. It is now clear that several gene loci are involved in this mechanism; the best characterized is a polymorphism in the promoter region of the G γ globin gene that appears to increase the output from this locus under

Beginnings: the molecular pathology of hemoglobin 15

conditions of hemopoietic stress. However, there are clearly other genes involved in increasing the output of Hb F. Recent genome-wide linkage studies have shown clear evidence that there are determinants on chromosomes 6 and 8 and a particularly strong association has been found with BCL11A, a transcription factor known to be involved in hematopoiesis. The exact mechanism for the associated increase in Hb F in β thalassemia and in sickle cell anemia remains to be determined. The tertiary modifiers are those that have no effect on hemoglobin synthesis, but modify the many different complications of the β thalassemias, including osteoporosis, iron absorption, jaundice, and susceptibility to infection. Although neglected until recently, it is also becoming apparent that variation in adaptation to anemia and the environment may also play a role in phenotypic modification of the β thalassemias. For example, patients with Hb E/β thalassemia have relatively low levels of Hb F and hence their oxygen dissociation curves are more right-shifted than patients with other forms of β thalassemia intermedia with significantly higher levels of Hb F. Very recent studies also suggest that the erythropoietin response to severe anemia for a given hemoglobin level varies considerably with age; patients during the first years of life have significantly higher responses to the same hemoglobin level than those who are older. This observation may go some way to explaining the variation in phenotype at different ages that has been observed in children with Hb E/β thalassemia. Finally, it is clear that further studies are required to dissociate the effects on the phenotype of genetic modifiers and environmental factors. Thus, the phenotypic variability of the β thalassemias reflects several layers of complex interactions involving genetic modifiers together with variation in adaptation and, almost certainly, the environment. These complex interactions are summarized in Table 1.2.

The α thalassemias The pathophysiology of the α thalassemias differs from that of the β thalassemias mainly because of the properties of the excess globin chains that are produced as a result of defective α chain synthesis. While the excess α chains produced in β thalassemia are unstable and precipitate, this is not the case in the α thalassemias, in which excess γ chains or β chains are able to form the soluble homotetramers γ4 (Hb Bart’s) and β4 (Hb H) (Figure 1.9). Although these variants, particularly Hb H, are unstable and precipitate in older red cell populations, they remain soluble sufficiently long for the red cells to mature and develop relatively normally. Hence there is far less ineffective erythropoiesis in the α thalassemias and the main cause of the anemia is hemolysis associated with the precipitation of Hb H in older red cells. In addition, of

Fetus

Adult

γ

β α

α2γ2

α2β2

Excess

Excess

γ4 Hb Bart's

β4 Hb H

High oxygen affinity—hypoxia Instability of homotetramers Inclusion bodies. Membrane damage Shortened red cell survival—hemolysis Splenomegaly—hypersplenism Fig. 1.9 The pathophysiology of α thalassemia.

course, there is a reduction in normal hemoglobin synthesis, which results in hypochromic, microcytic erythrocytes. Another important factor in the pathophysiology of the α thalassemias is the fact that Hb Bart’s and Hb H are useless oxygen carriers, having an oxygen dissociation curve similar to that of myoglobin. Thus the circulating hemoglobin level may give a false impression of the oxygen-delivering capacity of the blood, and patients may be symptomatic at relatively high hemoglobin levels. The different clinical phenotypes of the α thalassemias are an elegant example of the effects of gene dosage (Figure 1.10). The heterozygous state for α+ thalassemia is associated with minimal hematological changes. That for α0 thalassemia (the loss of two α globin genes) is characterized by moderate hypochromia and microcytosis, similar to that of the β thalassemia trait. It does not matter whether the α genes are lost on the same chromosome or on opposite pairs of homologous chromosomes. Hence the homozygous state for α+ thalassemia, −α/−α, has a similar phenotype to the heterozygous state for α0 thalassemia (− −/αα). The loss of three α globin genes, which usually results from the compound heterozygous states for α0 and α+ thalassemia, is associated with a moderately severe anemia with the production of varying levels of Hb H. This condition, hemoglobin H disease, is characterized by varying anemia and splenomegaly with a marked shortening of red cell survival. Finally, the homozygous state for α0 thalassemia (− −/ − −) is characterized by death in utero or just after birth, with the clinical picture of hydrops fetalis. These babies produce no α chains and their hemoglobin consists mainly of Hb Bart’s with variable persistence of embryonic hemoglobin. This is reflected in gross intrauterine hypoxia; although these babies may have hemoglobin values as high as 8–9 g/dL,

16 Molecular Hematology

α0 Thal. trait

α0 Thal. trait

X

Normal

α0 Thal. trait

α0 Thal. trait

Hb Bart's hydrops

different amino acid in the affected globin chain. Rarely, these variants result from more subtle alterations in the structure of the α/β globin chains. For example, shortened chains may result from internal deletions of their particular genes, while elongated chains result from either duplications within genes or frameshift mutations, which allow the chain termination codon to be read through and in which additional amino acids are added to the C-terminal end. The majority of the 700 or more structural hemoglobin variants are of no clinical significance but a few, because they interfere with the stability or functions of the hemoglobin molecule, are associated with a clinical phenotype of varying severity.

Genotype–phenotype relationships α0

α+

Thal. trait

Thal. trait

X

Normal

α0 Thal. trait

α+ Thal. trait

Hb H disease

Fig. 1.10 The genetics of the common forms of α thalassemia. The open boxes represent normal α genes and the green boxes deleted α genes. The mating shown at the top shows how two α0 thalassemia heterozygotes can produce a baby with the Hb Bart’s hydrops syndrome. In the mating at the bottom, between individuals with α0 and α+ thalassemia, one in four of the offspring will have Hb H disease.

most of it is unable to release its oxygen. This is reflected in the hydropic changes, a massive outpouring of nucleated red cells, and hepatosplenomegaly with persistent hematopoiesis in the liver and spleen.

Structural hemoglobin variants The structural hemoglobin variants are described in detail in Chapter 14. Here, their molecular pathology and genotype– phenotype relationships are briefly outlined.

Molecular pathology The molecular pathology of the structural hemoglobin variants is much less complex than that of the thalassemias. The majority result from missense mutations – base substitutions that produce a codon change which encodes a

The sickling disorders

The sickling disorders represent the homozygous state for the sickle cell gene, sickle cell anemia, and the compound heterozygous state for the sickle cell gene and various structural hemoglobin variants, or β thalassemia. The chronic hemolysis and episodes of vascular occlusion and red cell sequestration that characterize sickle cell anemia can all be related to the replacement of the normal β6 glutamic acid by valine in Hb S. This causes a hydrophobic interaction with another hemoglobin molecule, triggering aggregation into large polymers. It is this change that causes the sickling distortion of the red blood cell and hence a marked decrease in its deformability. The resulting rigidity of the red cells is responsible for the vaso-occlusive changes that lead to many of the most serious aspects of all the sickling disorders. The different conformations of sickle cells (banana shaped or resembling a holly leaf) reflect different orientations of bundles of fibers along the long axis of the cell, the threedimensional structure of which is constituted by a rope-like polymer composed of 14 strands. The rate and extent of polymer formation depend on the degree of oxygenation, the cellular hemoglobin concentration, and the presence or absence of Hb F. The latter inhibits polymerization and hence tends to ameliorate sickling. Polymerization of Hb S causes damage to the red cell membrane, the result of which is an irreversibly sickled cell. Probably the most important mechanism is cellular dehydration resulting from abnormalities of potassium/chloride cotransport and Ca2+ -activated potassium efflux. This is sufficient to trigger the Ca2+ -dependent (Gardos) potassium channel, providing a mechanism for the loss of potassium and water and leading to cellular dehydration. However, the vascular pathology of the sickling disorders is not entirely related to the rigidity of sickled red cells. There is now a wealth of evidence that abnormal interactions between sickled cells and the vascular endothelium play a major role in the pathophysiology of the sickling disorders. Recently it has been demonstrated that nitric oxide may also play a role in some of the vascular complications of

Beginnings: the molecular pathology of hemoglobin 17

this disease. It has been found that nitric oxide reacts much more rapidly with free hemoglobin than with hemoglobin in erythrocytes, and therefore it is possible that such decompartmentalization of hemoglobin into plasma, as occurs in sickle cell disease and other hemolytic anemias, diverts nitric oxide from its homeostatic vascular function. The sickling disorders are discussed in detail in Chapter 14, Hemoglobinopathies due to Structural Mutations. Unstable hemoglobin variants

There is a variety of different mechanisms underlying hemoglobin stability resulting from amino acid substitutions in different parts of the molecule. The first is typified by amino acid substitutions in the vicinity of the heme pocket, all of which lead to a decrease in stability of the binding of heme to globin. A second group of unstable variants results from amino acids that simply disrupt the secondary structure of the globin chains. About 75% of globin is in the form of α helix, in which proline cannot participate except as part of one of the initial three residues. At least 11 unstable hemoglobin variants have been described that result from the substitution of proline for leucine, five that are caused by the substitution of alanine by proline, and three in which proline is substituted for histidine. Another group of variants that causes disruption of the normal configuration of the hemoglobin molecule involves internal substitutions that somehow interfere with its stabilization by hydrophobic interactions. Finally, there are two groups of unstable hemoglobins that result from gross structural abnormalities of the globin subunits; many are due to deletions involving regions at or near interhelical corners. A few of the elongated globin chain variants are also unstable. Abnormal oxygen transport

There is a family of hemoglobin variants associated with high oxygen affinity and hereditary polycythemia. Most result from amino acid substitutions that affect the equilibrium between the R and T states (see Structure and Function). Thus, many of them result from amino acid substitutions at the α1 – β2 interface, the C-terminal end of the β chain, and at the 2,3-BPG binding sites. Congenital cyanosis due to hemoglobin variants

There is a family of structural hemoglobin variants that is designated Hb M, to indicate congenital methemoglobinemia, and is further defined by their place of discovery. The iron atom of heme is normally linked to the imidazole group of the proximal histidine residue of the α and β chains. There is another histidine residue on the opposite side, near the sixth coordination position of the heme iron; this, the so-called distal histidine residue, is the normal site of binding of oxygen. Several M hemoglobins result from the

substitution of a tyrosine for either the proximal or distal histidine residue in the α or β chain.

Molecular aspects of the high frequency of the hemoglobin variants The inherited disorders of hemoglobin are by far the commonest monogenic diseases. It is estimated that between 300 000 and 400 000 babies are born with these conditions, mainly across the tropical belt or in countries with large numbers of immigrants from this region. Work over many years has shown that the major reason for this high frequency is heterozygote protection against malaria infection, particularly that due to Plasmodium falciparum. Heterozygote protection raises the frequency of births of heterozygotes until an equilibrium is achieved by the early loss of the homozygotes for many of these diseases. In the case of α+ thalassemia, presumably because homozygotes have an extremely mild phenotype, the gene frequency is by far the commonest for any monogenic disorder, in some parts of the world almost reaching fixation. The mechanism for the protective effect of heterozygotes against malaria is still not completely understood, although there is extensive evidence that in most cases it may be multifactorial. In the case of the sickle cell trait, there is evidence for impairment of invasion and growth of the malarial parasites under conditions of low oxygen tension and enhanced removal of the parasite from infected red cells. There is also evidence for a reduced display of the parasite-encoded protein P. falciparum Erythrocyte Protein1 (PfEMP1), which is involved in the sequestration of mature parasites in small blood vessels and hence damage to tissues, particularly the brain. There is also evidence for improved acquisition of malaria-specific immunity in heterozygotes. Rather similar findings have been observed in the case of carriers for HbC and for impairment of P. falciparum red cell invasion and growth in heterozygotes for HbE. In the case of the α thalassemias, there appears to be reduced pathogenicity due to reduced cytoadherence or rosetting, a process whereby uninfected red cells aggregate round cells that are infected with P. falciparum. There is also evidence of immunological priming, whereby babies are more prone to infection by Plasmodium vivax early in life, which may result in later protection because of cross immunity between P. falciparum and P. vivax. There are some remarkable epistatic interactions between the hemoglobin disorders in response to malaria. For example, while heterozygotes for HbS or α+ thalassemia are protected in those who inherit the genes for both conditions, this protection is completely canceled out. Although the exact mechanism is not understood, this interaction reduces the amount of HbS in heterozygotes, suggesting that a critical mass of this variant is required for protection against

18 Molecular Hematology

malaria. Several other epistatic interactions of this type have been described recently and they are providing invaluable information about the variable distribution of these variants in different populations.

Molecular aspects of the prevention and management of the hemoglobin disorders Over the years there have been increasingly sophisticated approaches applied for screening and prenatal diagnosis of the hemoglobin variants, particularly the thalassemias, in an attempt to reduce the numbers of births of homozygotes or compound heterozygotes. At first this was carried out by fetal blood sampling in the mid-trimester of pregnancy, followed by hemoglobin analysis by globin chain synthesis. When DNA analysis became possible, this approach was replaced by DNA studies of material obtained by chorion villous sampling. Both these approaches were invasive, of course, but they have led to a dramatic reduction in the births of affected babies in many high-frequency populations. Very recently a non-invasive approach has been established by attempting to detect homozygotes for monogenic diseases by nextgeneration sequencing of fetal blood in maternal plasma. This approach is still in its early days, but some promising results have already been obtained. Although for many years there has been a gradual improvement in the symptomatic management of the hemoglobin disorders, the only approach for a cure has been the application of bone marrow transplantation, which of course has the disadvantage of requiring matched donors. More recently there has been a major effort directed at the development of somatic-cell gene therapy, particularly for the thalassemias. There has been considerable progress toward this process using retroviral vectors and related approaches. Although progress has been slow, successes have been reported and by the use of new gene-editing techniques further progress in the near future seems very likely. The other area of molecular therapy – that is, attempts to increase the level of fetal hemoglobin in patients with different forms of β thalassemia or sickle cell anemia – was discussed earlier in this chapter.

Postscript In this short account of the molecular pathology of hemoglobin, we have considered how mutations at or close to the α or β globin genes result in a diverse family of clinical disorders due to the defective synthesis of hemoglobin or its abnormal structure. Work in this field over the last 30 years has given us a fairly good idea of the repertoire of different mutations that underlie single-gene disorders and how

these are expressed as discrete clinical phenotypes. Perhaps more importantly, however, the globin field has taught us how the interaction of a limited number of genes can produce a remarkably diverse series of clinical pictures, and something of the basis for how monogenic diseases due to the same mutation may vary widely in their clinical expression.

Further reading General background Peltonen, L. and McKusick, V.A. (2001). Genomics and medicine. Dissecting human disease in the postgenomic era. Science 291: 1224– 1229. Weatherall, D.J. (2013). The role of the inherited disorders of hemoglobin, the first “molecular diseases,” in the future of human genetics. Annu. Rev. Genomics Hum. Genet. 14: 1–24. Weatherall, D.J. and Clegg, J.B. (2001). The Thalassaemia Syndromes, 4e. Oxford: Blackwell Science. Weatherall, D.J., Schechter, A.N., and Nathan, D.G. (eds.) (2013). Hemoglobin and Its Diseases. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.

Hemoglobin genetics and structural variants Steinberg, M.H., Forget, B.G., Higgs, D.R., and Weatherall, D.J. (eds.) (2009). Disorders of Hemoglobin, 2e. New York: Cambridge University Press.

Hemoglobin switching Forget, B.G. (2011). Progress in understanding the hemoglobin switch. N. Engl. J. Med. 365: 852–854. Sankaran, V.G. and Nathan, D.G. (2010). Reversing the hemoglobin switch. N. Engl. J. Med. 363: 2258–2260. Sankaran, V.G. and Orkin, S.H. (2013). The switch from fetal to adult hemoglobin. Cold Spring Harb. Perspect. Med. https://doi.org/ 10.1101/cshperspect.a011643. Thein, S.L. and Menzel, S. (2009). Discovering the genetics underlying foetal haemoglobin production in adults. Br. J. Haematol. 145: 455– 467.

The β thalassemias Fucharoen, S. and Weatherall, D.J. (2013). The hemoglobin E thalassemias. Cold Spring Harb. Perspect. Med. https://doi.org/10.1101/ cshperspect.a011734. Lettre, G. (2013). The Search for Genetic Modifiers of Disease Severity in the beta-Hemoglobinopathies. Cold Spring Harb. Perspect. Med. https://doi.org/10.1101/cshperspect.a015032. Nienhuis, A.W. and Nathan, D.G. (2013). Pathophysiology and clinical manifestations of the beta thalassemias. Cold Spring Harb. Perspect. Med. https://doi.org/10.1101/cshperspect.a011726. Olivieri, N.F., Muraca, G.M., O’Donnell, A. et al. (2008). Studies in haemoglobin E beta-thalassaemia. Br. J. Haematol. 141: 388–397.

Beginnings: the molecular pathology of hemoglobin 19

Quek, L. and Thein, S.L. (2007). Molecular therapies in betathalassaemia. Br. J. Haematol. 136: 353–365. Thein, S.L. (2008). Genetic modifiers of the beta-haemoglobinopathies. Br. J. Haematol. 141: 357–366.

The α thalassemias Higgs, D.R. (2013). The molecular basis of alpha thalassemia. Cold Spring Harb. Perspect. Med. https://doi.org/10.1101/cshperspect. a011718. Higgs, D.R. and Gibbons, R.J. (2010). The molecular basis of alphathalassemia: a model for understanding human molecular genetics. Hematol. Oncol. Clin. North Am. 24: 1033–1054. Vichinsky, E. (2013). Natural history and clinical manifestations of the alpha thalassemias. Cold Spring Harb. Perspect. Med. https://doi.org/10.1101/cshperspect.a011742.

Evolutionary background of hemoglobin disorders Penman, B.S., Pybus, O.G., Weatherall, D.J., and Gupta, S. (2009). Epistatic interactions between genetic disorders of hemoglobin can explain why the sickle-cell gene is uncommon in the Mediterranean. Proc. Natl. Acad. Sci. U.S.A. 106: 21242–21246.

Weatherall, D.J., Williams, T.N., Allen, S.J., and O’Donnell, A. (2010). The population genetics and dynamics of the thalassemias. Hematol. Oncol. Clin. North Am. 24: 1021–1031. Williams, T.N., Mwangi, T.W., Wambua, S. et al. (2005). Negative epistasis between the malaria-protective effects of alpha+ − thalassemia and the sickle cell trait. Nat. Genet. 37: 1253–1257.

Molecular basis of prevention and management of hemoglobin disorders Cao, A. and Kan, Y.W. (2013). The prevention of thalassemia. Cold Spring Harb. Perspect. Med. https://doi.org/10.1101/cshperspect. a011775. Lo, Y.M. and Chiu, R.W. (2010). Noninvasive approaches to prenatal diagnosis of hemoglobinopathies using fetal DNA in maternal plasma. Hematol. Oncol. Clin. North Am. 24: 1179–1186. Nienhuis, A.W. and Persons, D.A. (2013). Development of gene therapy for thalassemia. Cold Spring Harb. Perspect. Med. https://doi.org/ 10.1101/cshperspect.a011833. Orkin, S.H. and Reilly, P. (2016). MEDICINE. Paying for future success in gene therapy. Science 352: 1059–1061.

Chapter 2 Stem cells David T Scadden1,2 1 2

Department of Stem Cell and Regenerative Biology, Harvard Stem Cell Institute, Harvard University, Boston MA, USA Center for Regenerative Medicine, Massachusetts General Hospital, Boston MA, USA

Introduction, 21 Stem cell definitions and distinctions, 21 Hematopoietic stem cell concepts and their origin, 21 Molecular regulation of hematopoiesis, 24

Introduction The generation of sufficient numbers of blood cells to maintain homeostasis requires the sustained production of mature cells. This process, called hematopoiesis, yields approximately 1011 blood cells daily, with the capability for dramatic increases in the number and subsets of cells in response to physiological stress. Hematopoiesis is therefore a highly dynamic process dependent upon numerous modulating factors. Its prodigious production capability derives from the sustained presence of a cell type which is generally quiescent, but the descendants of which proliferate vigorously. This cell is the hematopoietic stem cell (HSC).

Stem cell definitions and distinctions Stem cells derive their name from their ability to produce daughter cells of different types. Stem cells are defined by a combination of the traits of self-maintenance and the ability to produce multiple, varied offspring. Putting this in more biological terms, stem cells have the unique and defining characteristics of self-renewal and differentiation into multiple cell types. Thus, with each cell division there is an inherent asymmetry in stem cells that is generally not found with other cell types. While their name implies that stem cells have specific intrinsic characteristics, there are multiple different types of stem cells, each defined by their production ability. Totipotent stem cells are capable of generating any type of cell in the body, including those of the extra-embryonic tissues, such as Molecular Hematology, 4th Edition. Edited by Drew Provan and John G. Gribben. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

Trafficking of primitive hematopoietic cells, 29 Manipulating HSCs for clinical use, 31 Summary, 34 Further reading, 34

the placental tissues (Figure 2.1). Pluripotent stem cells may give rise to any type of cell found in the body except those of the extra-embryonic membranes. They can produce ectoderm, mesoderm, or endoderm cells. It has also become possible to create pluripotent cells by “reprogramming” mature cells. This allows for pluripotent cells to be made from any individual, a powerful tool for basic biology, disease modeling, and possible future cell therapies. These are called induced pluripotent stem cells (iPS). Pluripotent stem cells not made by reprogramming include embryonic stem cells, isolated from the inner cell mass of the blastocyst; embryonic germ cells, isolated from embryonic gonad precursors; and embryonic carcinoma cells, isolated from teratocarcinomas. Pluripotent stem cells may be maintained indefinitely in culture under specialized conditions that prevent differentiation. In particular, embryonic stem cells have been used to generate “knockout” mice, animals harboring targeted gene disruptions via homologous recombination that permit the in vivo study of individual gene function. Lastly, multipotent stem cells, such as the HSCs of the bone marrow, are capable of giving rise to multiple mature cell types, but only those of a particular tissue, such as blood. Multipotent stem cells are found in adults, perhaps in all tissue, and function to replace dead or damaged tissue. Such stem cells are commonly referred to as “adult” stem cells.

Hematopoietic stem cell concepts and their origin The cellular compartment model The short-lived nature of most blood cells was first deduced in the 1960s using thymidine labeling of reinfused blood. These studies demonstrated that the maintenance of normal

21

22 Molecular Hematology

Inner cell mass

Fertilization

Totipotent cells

Blastocyst

Fetal tissues

Adult tissues

Pluripotent cells

Multipotent cells

Embryonic stem (ES) cells

“Tissue” or “adult” stem cells

Fig. 2.1 Sources and types of stem cells. Source: Adapted with gratitude from the National Institutes of Health Stem Cell Information website.

numbers of blood cells in the adult requires a process with the capacity to briskly generate large numbers of mature cells along multiple blood lineages. The early history of HSC research was largely shaped by cellular biology and animal transplantation experiments. It was advanced by experiments in the early 1960s demonstrating that injection of marrow cells could generate large hematopoietic colonies in the spleen of irradiated mice. Such colonies were the clonal progeny of single initiating cells, termed colonyforming units, spleen (CFU-S), and contained hematopoietic populations of multiple lineages. CFU-S were further transplantable, demonstrating their self-renewing nature. HSCs are a minor component of marrow cells, able both to generate large numbers of progeny differentiated along multiple lines and to renew themselves. The field was further advanced by the use of in vitro cell culture techniques; in particular, solid-state cultures of marrow and spleen cells furthered understanding of the colonyforming capacity of individual hematopoietic cells. The original technique demonstrated clonal colonies of granulocytes and/or macrophages, termed in vitro colony-forming cells (CFCs), which are now considered lineage-committed progenitor cells. These cells could be separated from whole marrow cells and from CFU-S, were more numerous than CFU-S, and could be detected in splenic colonies as the progeny of CFU-S. These observations gave rise to the concept of the three-compartment model of hematopoiesis, the compartments being stem cells, progenitor cells, and dividing mature cells in increasing numbers; each compartment consists of the amplified progeny of cells in the preceding compartment. Subsequent analyses have added further complexity to the compartment model of hematopoiesis. The term CFU-S describes at least two groups of precursor cells. One group,

arising from committed progenitors with little capacity for self-renewal, gives rise to colonies that peak in size by day 8, while a second, arising from a more primitive cell that is capable of self-renewal, yields colonies that peak in size at day 12. To further highlight the complexity of the hematopoietic hierarchy, a rarer population of hematopoietic cells provides longer-term repopulation of an irradiated host than CFU-S. These long-term repopulating cells have the capacity for sustained self-renewal and were considered the true adult stem cells. The presence of stromal cells in the cultures is important for the long-term culture of CFU-S and repopulating cells. Cells capable of long-term survival in culture on stroma were termed long-term culture-initiating cells (LTC-ICs) and cobblestone area-forming cells (CAFCs). These multipotential cell types were considered more primitive than lineage-committed progenitor cells, but more mature than long-term repopulating cells. Thus, a more complex version of the compartmental model has emerged. This provides a model with two populations of stem cells, the most immature group consisting of long-term repopulating cells and a more mature group of short-term repopulating cells. These groupings are simple models for what is likely to be a continuum of cells with selfrenewal and multipotent differentiation capacity. The longterm repopulating cells also tend to be more quiescent, serving as a deep reserve for blood cell production in times of physiological stress. More mature progenitors (also called colony-forming cells or CFCs) have the capacity to give rise to colonies of clonal origin in semisolid media containing fully mature cells, permitting their analysis. A more mature set of precursor cells constitutes the bulk of bone marrow cells and has unique, identifiable features by light microscopy. Rapid division of precursor cells culminates in the production of

Stem cells 23

Hematopoiesis Progenitor cells

Precursor cells

Mature cells

Microenvironment

Stem cells

Long-term Fig. 2.2 Schematic view of hematopoiesis. See text for definition of abbreviations. Source: Modified from Figure 12.1 in Hematology: Basic Principles and Practice, 3rd edn (ed. R. Hoffman), 2000, with permission from Elsevier.

Short-term

LTC-IC/CAFC

CFU-S

HPP-CFC CFU-F/RF

mature cells. Although hematopoiesis proceeds according to this orderly scheme (Figure 2.2), special consideration must be given to the development of T and B lymphocytes. These cells are generated in the thymus and bone marrow, respectively, by a similar hierarchical process. Mature T and B lymphocytes enter peripheral lymphoid organs, where they encounter relevant antigens, leading to the production of new cells from reactivated mature cells. This process amplifies the de novo bone marrow formation of T and B lymphocytes. In addition, some members of this type of cell, memory T or B lymphocytes, are capable of sustained self-renewal, serving, in effect, as unipotent stem cells. Their inability to produce multiple different types of daughter cells distinguishes them from multipotent HSCs. In summary, the compartment model has given rise to terms that are generally applied to cells of hematopoietic origin. HSCs are those that are multipotent and self-renewing. Progenitor cells have limited ability to self-renew and are likely to be unipotential or of very limited multipotential. Precursor cells are restricted to a single lineage, such as neutrophil precursors, and are the immediate precursors of the mature cells found in the blood. The mature cells are generally short-lived and preprogrammed to be highly responsive to cytokines, while the stem cells are long-lived, cytokine resistant, and generally quiescent.

Models of lineage commitment Several theories have emerged to describe the manner by which HSCs undergo lineage commitment and differentiate. Some studies support a deterministic theory whereby the stem cell compartment encompasses a series of closely related cells maturing in a stepwise process. Other studies

CFU-C CFU-Mix

suggest that hematopoiesis is a random, stochastic process. The stochastic theory is based on in vitro observations that multilineage colonies develop variable combinations of lineages, and that such lineage choices occur independently of external influences. Similar controversy exists regarding the role of cytokines in cell lineage determination. An instructive model suggests that cytokine signaling forces the commitment of primitive cells along a particular lineage. Ectopic expression of the granulocyte macrophage colony-stimulating factor (GM-CSF) receptor in a common lymphoid progenitor (CLP) population was capable of converting the cells from a lymphoid to a myeloid lineage. The influence of the GM-CSF receptor was sufficiently dominant to change the entire differentiation program of cells, but only the CLP stage of development. A permissive model postulates that decisions about cell fate occur independently of extracellular signals. This model suggests that cytokines serve only to allow certain lineages to survive and proliferate. Evidence supporting this model is provided by the ectopic expression of growth receptors in progenitor cells. Expression of the erythropoietin receptor in a macrophage progenitor results in macrophage colony formation, whereas expression of the macrophage colony-stimulating factor (M-CSF) receptor in an erythroid progenitor results in erythroid rather than macrophage colony formation. Replacing the thrombopoietin receptor (c-mpl) with a chimeric receptor consisting of the extracellular domain of c-mpl with the cytoplasmic domain of the granulocyte colony-stimulating factor (G-CSF) receptor results in normal platelet counts in homozygous “knock-in” mice. Therefore, the instructive and permissive models may both be correct, but at different stages of hematopoietic differentiation. Cells at earlier points in the differentiation

24 Molecular Hematology

cascade may be more susceptible to fate-altering stimuli, while more committed cells may be irreversibly determined, with only proliferation, cell death, or the rate of differentiation susceptible to influence by external signals.

Molecular regulation of hematopoiesis The molecular nature of stem cell regulatory pathways has been determined using a variety of genetic approaches, including genetic loss-of-function and gain-of-function studies. These have provided several important concepts regarding the molecular control of hematopoiesis. First, some genes have binary functions and exert an effect by being above a threshold of expression, rendering their targets “on.” Other genes function in a continuum and have different effects at different levels. Secondly, while perturbations in single genes may have dramatic cellular effects, gene products often function in complexes and, in the case of transcription regulators, are only active if they have access to sites of open chromatin. Finally, signal integration often depends on the assembly of large signaling complexes and the spatial proximity of molecules to facilitate interaction is therefore important.

Cell-intrinsic regulators of hematopoiesis Cell cycle control

The quiescent nature of HSCs is supported by their low level of staining with DNA and RNA nucleic acid dyes, which is consistent with low metabolic activity. These studies have indicated a heterogeneity among stem cells with a subgroup that is deeply quiescent. Various studies have sought to determine the cell-intrinsic regulators of hematopoiesis involved in HSC cycle control. RNA analysis has been used to profile pertinent transcription factors and other molecules in HSCs induced to differentiate along various lineages by the application of cytokines. Elevated levels of cyclin-dependent kinase inhibitors (CDKIs) have been observed, suggesting that CDKIs present in HSCs function to exert a dominant inhibitory tone on HSC cell cycling. The bone marrow of some mouse strains deficient in CDKI p57(CDKN1C), p21(CDKN1A), or p18(CDKN2C) have increased HSC cell cycling, suggesting that these CDKIs function as dominant negative regulators of HSC proliferation. Other CDKIs, such as p27(CDKN1B), may serve as negative regulators of hematopoietic progenitor cells.

to lose the function of specific genes have been used to define the impact of those genes on hematopoiesis. Lossof-function studies involving the transcription factors c-Myb, AML1 (CBF2), SCL (tal-1), LMO2 (Rbtn2), GATA-2, and TEL/ETV6 have demonstrated global effects on all hematopoietic lineages. Stem cells in animals deficient in these molecules fail to establish definitive hematopoiesis. Similar methods have indicated that some regulators have different roles at different times in development. For example, the gene product of transcription factor SCL is absolutely required for establishing HSCs. Unexpectedly, there is not a requirement for SCL once the stem cell pool is present in the adult. Rather, SCL is required only for erythroid and megakaryocytic homeostasis. Therefore, transcription factor regulation of the stem cell compartment can be highly dependent on the stage of development of the organism. Loss-of-function studies have also proved useful in identifying lineage-specific transcription factors. Mice genetically deficient in the transcription factor Ikaros lack T and B lymphocytes and natural killer cells, but maintain erythropoiesis and myelopoiesis. Notch is required for T lineage induction and shifts differentiation away from B cell development. Further, losing expression of some genes can enable cells to revert in differentiation. For example the Pax-5 transcription factor is essential for B cell maturation, and loss of it results in cells that have already rearranged their B cell receptor or immunoglobulin locus showing up in the T and NK cell compartment. The critical roles for specific transcription factors in establishing and maintaining cell fate are shown in Figure 2.3. The ability of transcription factors to affect gene expression is also highly dependent upon epigenetic features of a particular gene locus. The histone marks and chromatin accessibility regulators that participate in the epigenome are beyond the scope of this chapter but should be recognized as critical aspects of gene expression and therefore, cell state.

Cell-extrinsic regulators Ultimately, hematopoietic stem and progenitor cell decisions are regulated by the coordinated action of transcription factors as modified by extracellular signals. Extracellular signals in the form of hematopoietic growth factors are mediated via cell surface hematopoietic growth factor receptors. Hematopoietic growth factors exert specific effects when acting alone and may have different effects when combined with other cytokines. There are at least six receptor superfamilies, and most growth factors are members of the type I cytokine receptor family. The effects of various cytokines during myelopoiesis are illustrated in Figure 2.4.

Self-renewal, commitment, and lineage determination

Experimental results involving transcription factors have demonstrated cell-intrinsic roles in both global and lineagespecific hematopoietic development. Mice engineered

Type I cytokine receptors

Type I receptors do not possess intrinsic kinase activity, but lead to phosphorylation of cellular substrates by serving

Stem cells 25

NK cell

T cell

Lymphoid pathway

IL-7R+

CLP

SCL (–) GATA-2 (–) NF-E2 (–) GATA-1 (–)

IL-7R+ c-mpl– SCL (–) GATA-2 (–) NF-E2 (–) GATA-1 (–)

HSC

LT-HSC

ST-HSC

SCL (++) GATA-2 (++) NF-E2 (–) GATA-1 (±)

C/EBPα (±) PU.1 (±) Aiolos (±) GATA-3 (±)

Pro-T

C/EBPα (–) PU.1 (+) Aiolos (+) GATA-3 (+)

Pro-B B cell IL-7R+ SCL (–) GATA-2 (–) NF-E2 (–) GATA-1 (–)

C/EBPα (–) PU.1 (+) Aiolos (++) GATA-3 (–)

GMP

Myeloid pathway IL-7R– c-mpl+

C/EBPα (–) PU.1 (–) Aiolos (++) GATA-3 (++)

Monocyte

CMP Epo-R–

SCL (++) C/EBPα (±) GATA-2 (+) PU.1 (±) NF-E2 (+) Aiolos (±) GATA-1 (+) GATA-3 (–)

Granulocyte

SCL (+) C/EBPα (++) GATA-2 (–) PU.1 (±) NF-E2 (–) Aiolos (–) GATA-1 (–) GATA-3 (–) MEP Megakaryocyte

Epo-R+

Erythrocyte SCL (++) GATA-2 (++) NF-E2 (++) GATA-1 (++)

C/EBPα (–) PU.1 (±) Aiolos (–) GATA-3 (–)

Fig. 2.3 Transcription factors active at various stages of hematopoiesis. CLP, common lymphoid progenitor; CMP, common myeloid progenitor; GMP, granulocyte monocyte progenitor; MEP, megakaryocyte erythrocyte progenitor; NK, natural killer. Source: Redrawn from Akashi, K., Traver, D., Miyamoto, T., Weissman, I.L. (2000). A clonogenic common myeloid progenitor that gives rise to all myeloid lineages. Nature 404: 193–197, with permission of Spring Nature.

as docking sites for adapter molecules with kinase activity. Examples of receptors in this family include leukemia inhibitory factor (LIF), interleukin (IL)-1, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-9, IL-13, IL-18, GM-CSF, G-CSF, erythropoietin, prolactin, growth hormone, ciliary neurotrophic factor, and c-mpl. These receptors share several features, including enhanced binding and/or signal transduction when expressed as heterodimers or homodimers, four cysteine residues and fibronectin type III domains in the extracellular domain, WSXWS ligand-binding sequence in the extracellular cytokine receptor domains, and lack of a known catalytic domain in the cytoplasmic portion. Another shared feature of receptors in this family is the ability to transduce signals that prevent programmed cell death (apoptosis).

Receptor tyrosine kinases

Receptors with intrinsic kinase activity are of considerable relevance to hematology because their ligands are growth factors, but also because abnormalities of them can result in unregulated activation. For example, internal tandem duplications of the Flt3 receptor are associated with acute myeloid leukemia (AML) and activating mutations of c-kit play a key role in systemic mastocytosis and core binding factor AMLs. Loss of c-fms, the receptor for M-CSF, has been associated with myelodysplasia and a predisposition to AML. The availability of agents targeting tyrosine kinases also makes these receptors of particular interest in hematologic disease. Details of receptor–ligand pairs are provided in Table 2.1. Protein serine–threonine kinase receptors

Type II cytokine receptors

This class includes the receptors for tissue factor, IL-10, and interferon (IFN)-γ. This family contains a type III fibronectin domain in the extracellular domain, like the type I family.

This family includes the 30 members of the transforming growth factor (TGF)-β superfamily, which bind to their receptors as homodimers. Members of this family include the three TGF-β receptors: type I (TbRI, 53 kDa), type II

26 Molecular Hematology

G-CSF, IL-1, IL-6, IL-10, IL-11, IL-12, IL-13 HPP-CFC b-FGF, HGF, LIF, SCF/KL, FLT3 Ligand, TPO Mixed progenitor cell

Pluripotent stem cell

CFU-GEMM

FLT3, SCF, IL-3, IL-6, GM-CSF, G-CSF Myelomonocytic progenitor

SCF G-CSF

CFU-GM

IL-4 IL-5

IL-3 GM-CSF

SCF, IL-3, IL-6, G-CSF GM-CSF, CSF-1

IL-5 CFU-G

CFU-M

CFU-Baso

CFU-Eos NGF

CSF-1, IL-3 GM-CSF Monoblast

CSF-1 GM-CSF Promonocyte

CSF-1 Monocyte

G-CSF GM-CSF IL-4 Myeloblast

IL-5 GM-CSF

IL-10, IL-9, CSF

IL-5, GM-CSF

IL-3, IL-4

G-CSF GM-CSF Myelocyte

G-CSF Neutrophil

Eosinophil

NGF

Basophil, mast cell

Chemokines MCP-1-3, IP-10, Rantes, MIP-1α

IL-8, NAP-2, Gro-α

Eotaxin, MCP-4

Rantes

IL-8, MCP-1, 3

Fig. 2.4 Cytokines active at various stages of hematopoiesis. See text for definition of abbreviations. Source: Modified from Figure 16.3 in Hematology: Basic Principles and Practice, 3rd edn (ed. R. Hoffman), 2000, with permission from Elsevier.

(TbRII, 75 kDa), and type III (TbRIII, 200 kDa). Members of this family have a profound inhibitory effect on the growth and differentiation of hematopoietic cells and on auxiliary hematopoietic cells. Binding of TFG-β requires TbRII. After binding, signal transduction occurs via activation of serine–threonine kinase cytoplasmic domains of the receptor chains, which results in the phosphorylation of Smad molecules on serines. Phosphorylated Smad complexes translocate to the nucleus, where they induce or repress gene transcription. TGF-β is the best-characterized negative regulator of hematopoiesis. It inhibits mitosis by inducing cell cycle inhibitors such as p21(CDKN1A), p27(CDKN1C), and p16(CDKN2A), inhibiting the cyclin-dependent kinases Cdk4 and Cdk6, and inducing phosphorylation of the retinoblastoma protein. The TGF-β receptor family and its

downstream mediators act as braking factors for a number of cell types and are frequently inactivated by somatic mutation in a number of cancers. Chemokine receptors

This family comprises seven transmembrane-spanning G-protein-coupled receptors that influence both cell cycle and cellular movement, or chemotaxis. These receptors are divided into three families, α or CXC, β or CC, and γ or C, on the basis of variability in cysteine residues. The best characterized is CXCR4, which mediates homing and engraftment of HSCs in bone marrow and is critical to hematopoietic development. IL-8 and macrophage inflammatory protein (MIP)-1α act as inhibitors of progenitor cell proliferation.

Stem cells 27

Table 2.1 Factors affecting hematopoietic control

Growth factor

Growth factor receptor

Erythropoiesis EPO (erythropoietin)

EPO-R

Produced by

Bioactivity

Deficient states

Adult kidney Liver during development

Stimulates clonal growth of CFU-E and BFU-E subsets Suppresses erythroid progenitor cell apoptosis Induces bone marrow release of reticulocytes Induces erythroid globin synthesis Promotes proliferation and differentiation of pre-CFC cells Acts synergistically with IL-3, GM-CSF, and TPO to support growth of CFU-GEMM, BFU-E, and CFU-Mk Expansion of committed progenitor cells in vivo Stimulates mast cell hyperplasia, degranulation, and IgE-dependent mediator release Induces DNA synthesis and has anti-apoptotic effects in erythroid progenitors Simulates erythroid colony growth in the absence of EPO at high doses

Anemia

SF (steel factor), kit ligand, mast cell growth factor

c-kit (CD117) Bone marrow perivascular mesenchymal cells Endothelial cells

IGF-1 (insulin-like growth factor, somatomedin C)

IGF-1R

Granulopoiesis G-CSF (granulocyte G-CSFR colony-stimulating factor)

GM-CSFR GM-CSF (granulocyte macrophage colony-stimulating factor) M-CSF (macrophage c-fms colony-stimulating factor)

Thrombopoietin

c-mpl

IL-5

IL-5R

IL-11

IL-11R

Liver

Anemia Mast cell deficiency

Growth retardation, neurological defects, homozygous deficiency lethal

Stimulates growth of progenitors committed to neutrophil Neutropenia, failure to differentiation develop neutrophilic Activates neutrophil phagocytosis leukocytosis in Stimulates quiescent HPCs to enter G1 /S response to infection Stimulates mobilization of HSCs and HPCs from bone marrow to periphery Susceptibility to Stimulates multilineage hematopoietic progenitor cells Mast cells, T infections caused by Stimulates BFU-E and granulocyte, macrophage, and lymphocytes, obligate eosinophil colony growth endothelial cells, intracellular fibroblasts, thymic organisms epithelial cells Induces monocyte/macrophage growth and differentiation Macrophage and Monocytes, and activation osteoclast macrophages, deficiency, fibroblasts, hematopoietic epithelial cells, failure vascular endothelium, osteoblasts Stimulates in vitro growth of CFU-Mk, megakaryocytes, Bone marrow Thrombocytopenia and platelets stroma, spleen, renal tubule, liver, Stimulates clonal growth of individual CD34+ CD38− cells muscle, brain Synergizes with SF, IL-3, and FL Primes response to platelet activators ADP, epinephrine, and thrombin, but no effect on aggregation Inability to mount T lymphocytes Stimulates eosinophil production and activation eosinophilic Activates cytotoxic T cells response Induces immunoglobulin secretion Fibroblasts, bone Acts synergistically with IL-3 or SF to stimulate the clonal No hematological marrow stroma growth of erythroid (BFU-E and CFU-E) and primitive defect megakaryocytic (BFU-Mk) progenitors Shortens duration of G0 of HPCs Quickens hematopoietic recovery after chemotherapy and radiation Monocytes, macrophages, endothelial cells, fibroblasts

(continued)

28 Molecular Hematology

Table 2.1 (Continued)

Growth factor

Growth factor receptor

Lymphopoiesis IL-7

IL-7R

IL-2

IL-2R

IL-15

IL-15R

IL-4

IL-4R

IL-10 Early-acting factors IL-3

IL-3R

FLT3-ligand (FL)

FLT-3R, flk2

IL-9 (T-cell growth factor) IL-6

IL-9R IL-6R

Produced by

Bioactivity

Deficient states

Bone marrow stroma, spleen, thymus T lymphocytes

Induces clonal growth of pre-B cells Induces growth of pre-T cells

B- and T-cell lymphopenia

Induces proliferation and activation of T cells, B cells, and NK cells

Fatal immunoproliferative disorder, loss of self-tolerance

Monocytes, macrophages, epithelial cells, skeletal muscle cells, bone marrow and thymic stroma T lymphocytes

Induces proliferation and activation of T cells, B cells, and NK cells

Induces proliferation of activated B cells Inhibits IL-2-stimulated proliferation of B cells Induces T-cell proliferation Inhibits monocyte/macrophage-dependent synthesis of Th1- and Th2-derived cytokines

Defective T helper cell responses

No hematopoietic defect in steady state, deficient delayed-type hypersensitivity Reduction in pro-B Weak colony-stimulating activity alone, but synergizes Most tissues, cells, pre-B cells, with IL-3, GM-CSF, SF, IL-11, IL-6, G-CSF, IL-7, and others including spleen, B-cell Augments retroviral transduction of HSCs when added to lung, stromal colony-forming cytokine cocktails cells, peripheral potential, reduced Mobilizes HSCs to periphery weakly alone, but adds blood repopulating greatly to G-CSF mononuclear cells capacity of stem cells T lymphocytes Stimulates growth of BFU-E when combined with EPO Stimulates clonal growth of fetal CFU-Mix and CFU-GM Reduced HSC and Synergistic with IL-3 for CFU-GEMM colony growth Macrophages, progenitor cell endothelial cells, Synergistic with IL-4 in inducing T-cell proliferation and survival, reduced colony growth fibroblasts, T T-cell numbers, Synergistic with M-CSF in macrophage colony growth lymphocytes reduced Synergistic with GM-CSF in granulocyte colony growth proliferation and Co-induces differentiation of B cells maturation of erythroid and myeloid cells T lymphocytes, mast Stimulates multilineage colony growth and growth of cells primitive cell lines with multilineage potential Stimulates BFU-E proliferation

BFU-E, burst-forming unit, erythroid; CFU-mix, colony-forming unit, mix; CFU-Mk, colony-forming unit, megakaryocyte; CFU-GM, colony-forming unit, granulocyte/macrophage; CFU-GEMM, colony-forming unit, granulocyte, erythroid, monocyte, megakaryocyte.

Stem cells 29

Members of this receptor family have also been implicated in cancer metastasis and the entry of HIV-1 into cells. Tumor necrosis factor receptor family

Members of the tumor necrosis factor receptor (TNFR) family have varied effects, some having the ability to induce programmed cell death and others stimulating mesenchymal cells to secrete hematopoietic growth factors. These receptors contain Cys-rich extracellular domains and 80-amino acid cytoplasmic “death domains,” which are required for transducing the apoptotic signal and inducing NF-κB activation. Members of this family include TNFR1, TNFR2, fas, CD40, nerve growth factor (NGF) receptor, CD27, CD30, and OX40, each with at least one distinct biological effect.

Components of the hematopoietic microenvironmental niche While soluble factors influence stem cell fate, these factors are seen by the cell in the context of the cell–cell contact among heterologous cell types and cell–matrix contact that comprise the three-dimensional setting of the bone marrow. What actually constitutes the critical microenvironment for hematopoiesis is surprisingly poorly defined. The ability of primitive cells to mature in vitro in complex stromal cultures suggests that at least some elements of the regulatory milieu of the bone marrow can be recapitulated ex vivo. Studies based solely on ex vivo systems are suspect, however, as no fully satisfactory re-creation of stem cell expansion or selfrenewal has been defined. Recognizing this limitation, it has been determined that mesodermal cells of multiple types are needed to enable hematopoietic support. These include multiple mesenchymal stromal cells, adipocytes, hematopoietic cells, and endothelium. HSCs tend to be perivascular in location endogenously and in periendosteal regions when transplanted. The most immature hematopoietic stem/progenitor cells (HSPCs) have been reported to be adjacent to periarteriolar mesenchymal cells, while more mature cells of lymphoid lineage are nearby to more mature osteolineage cells. There appears to be spatial organization of the bone marrow with multiple niches for particular subsets of HSPC, though the precise details of the marrow microenvironment are still being defined.

Trafficking of primitive hematopoietic cells The migratory behavior characteristic of primitive hematopoietic cells is an area of intense research because of its relationship to bone marrow transplantation. Trafficking of HSCs can be divided into the components of homing, retention, and engraftment. Homing describes the tendency

of cells to arrive at a particular environment, while retention is their ability to remain in such an environment after arrival. Lastly, engraftment reflects the ability of cells to divide and form functional progeny in a given microenvironment. Much has been learned about trafficking from the ontogeny of mouse and human HSCs.

Hematopoietic ontogeny In both humans and mice, hematopoiesis occurs sequentially in distinct anatomical locations during development. These shifts in location are accompanied by changes in the functional status of the stem cells and reflect the changing needs of the developing organism. These are relevant for adult hematopoiesis, since they offer insight into how the blood production process can be located in different places with distinct regulation. There are essentially five sites of blood cell formation recognized in mammalian development, and these are best defined in the mouse. At about embryonic day 7.5 (E7.5), blood and endothelial progenitors emerge in the extra-embryonic yolk sac blood islands. The yolk sac supports the generation of primitive hematopoietic cells, which are primarily composed of nucleated erythrocytes and macrophages. The macrophages generated in the yolk sac are generally tissue-resident cells that durably populate tissues throughout adult life. Examples are the microglia of the brain and the Kupffer cells of the liver. More sustained or definitive hematopoiesis derives from the aorta–gonad–mesonephros (AGM) region, as has been defined in both the mouse (E8.5) and the human. Current data suggest that yolk sac cells do not seed the AGM region, but rather hematopoietic cells arise there de novo. The placenta may also be a site of de novo HSC generation. By E10 in the mouse, the fetal liver assumes the primary role of cell production. By E14 in the mouse and the second trimester of human gestation, the bone marrow becomes populated with HSCs and it takes over blood cell production, along with the spleen and thymus. The spleen remains a more active hematopoietic organ in the mouse than in the human. Stem cell proliferation is very active in the AGM region and fetal liver. While it is also initially robust in the bone marrow, there is a dramatic shift to quiescence shortly after residence in the bone marrow, experimentally defined as after four weeks in the mouse.

Homing and engraftment of HSCs following infusion Despite the use of HSC transplantation for over three decades, the exact mechanisms whereby bone marrow cells home to the bone marrow are not fully understood. Other than lectins, no adhesion receptors have been identified that are exclusively present on HSCs. Furthermore, no adhesion

30 Molecular Hematology

ligands, other than hemonectin, have been identified that are exclusively present in the bone marrow microenvironment. When first infused, HSCs lodge in the microvasculature of the lung and liver; they then colonize the bone marrow, first passing through marrow sinusoids, migrating through the extracellular space of the bone marrow, and ultimately settling in stem cell niches. Passage through the endothelial barriers at first requires tethering, through endotheliumexpressed addressins that bind hematopoietic cell selectins, and this is followed by firm attachment mediated by integrins. Selectins are receptors expressed on hematopoietic cells (L- and P-selectins) and endothelium (E- and P-selectins). They have long extracellular domains containing an aminoterminal Ca2+ -binding domain, an epidermal growth factor domain, and a series of consensus repeats similar to those present in complement regulatory molecules. Ligands for selectins are sialylated fucosyl glucoconjugates present on endothelium. L-selectin is present on CD34+ hematopoietic progenitors, while L-selectin and P-selectin are present on more mature myeloid and lymphoid cells. Tethering by selectins allows integrin-mediated adhesion to the endothelium. Integrins, a family of glycoproteins composed of α and β chains responsible for cell–extracellular matrix and cell–cell adhesion, not only provide firm attachment, but also allow migration of hematopoietic cells through the endothelium and bone marrow extracellular space. The functional state of integrins is only loosely tied to their expression level and depends on ligand affinity modulation, regulated by the β subunit in response to cytokines and other stimuli. The process of migration depends on the establishment of adhesion at the leading edge of the cell and simultaneous release at the trailing edge. The rate of migration depends on dynamic changes in the strength of the cell–ligand interactions, which is dictated by the number of receptors and their affinity state and the strength of the adhesion receptor– cytoskeleton interactions. Cell–ligand interaction strength may also be modulated by chemokines that provide directional cues for cell migration based on ligand gradients. Thus, successful engraftment relies not only on the presence of several different adhesion receptors, but also on the presence of chemokine gradients that encourage directional movement of the cells. Chemokines such as CXCL12, the ligand for the receptor CXCR4, are localized discontinuously in the bone marrow microvasculature and provide site-specific guidance for HSPC migration and localization.

Egress of HSCs from bone marrow under physiological conditions The majority of primitive HSCs are resident within the bone marrow space under steady-state physiological conditions.

However, a population of CD34+ cells capable of forming CFCs and LTC-ICs and of long-term repopulation may be found circulating in the peripheral blood, and these may increase after physiological stressors such as exercise, stress, and infection. Animal studies have suggested that a relatively large number of bone marrow–derived stem cells circulates during the course of a day, and that these cells periodically transit back into an engraftable niche to establish hematopoiesis. Defining the processes involved is important in guiding new approaches to peripheral blood stem cell mobilization for transplantation. Examining mice in which specific adhesion molecules have been deleted has revealed several key molecular determinants of stem cell localization in the bone marrow. Among these, the chemokine receptor CXCR4 has perhaps the most striking phenotype. In the absence of this receptor, stem cells fail to traffic from the fetal liver to the bone marrow in development. Partly because of these studies, others have defined that CXCR4 is relevant for the engraftment of transplanted stem cells, and that the modulation of CXCR4 signaling can affect adult stem cell localization in the bone marrow versus peripheral blood. As described, the integrin and selectin families are also important molecular participants in stem cell location. For example, HSCs from animals that are heterozygous deficient for β1 integrin cannot compete with wild-type cells for the colonization of hematopoietic organs. Pre-incubation of HSCs with α4 integrin antibodies prior to transplantation results in decreased bone marrow and increased peripheral recovery of cells, while the continued presence of α4 antibodies prevents engraftment. Evidence for selectin involvement has been demonstrated in animals deficient of single selectins or combinations of selectins. Endothelial P-selectin mediates leukocyte rolling in the absence of inflammation, while L-, P-, and E-selectins contribute to leukocyte rolling in the setting of inflammation. L-selectin is important in lymphocyte homing. Transplantation studies performed in animals deficient in P- and E-selectins demonstrate severely decreased engraftment due to impaired homing, an effect that is further compromised by blocking vascular cell adhesion molecule (VCAM)-1. Mature hematopoietic cells are thought to migrate from the marrow to the blood by similar mechanisms, though these are not well defined. One purported mechanism is a shift in expression from molecules thought to interact with stromal proteins to those that interact with endothelium. For example, myeloid progenitors express functional α4β1 and α5β1 integrins that act to ensure that these progenitors are retained in the bone marrow through interactions with VCAM and fibronectin. Mature neutrophils, in contrast, express β2 integrins that permit interaction with ligands, such as intercellular adhesion molecule, expressed

Stem cells 31

by endothelial cells. Mature neutrophils also express β1 integrins that permit interaction with collagen and laminin present in basal membranes, perhaps regulating a progressive shift in cell affinities for specific microenvironmental determinants that ultimately results in cell egress into the blood. Mobilization of murine HSCs induced by cyclophosphamide or G-CSF is accompanied by changes in integrin expression levels and functional changes in homing, thus linking cellular localization with adhesion molecule receptor expression.

Manipulating HSCs for clinical use Mobilization of HSCs Mobilization of HSCs in response to chemotherapy or cytokines was first documented in the 1970s and 1980s. This process may be induced by a variety of molecules, including cytokines such as G-CSF, GM-CSF, IL-7, IL-3, IL-12, SCF, and flt-3 ligand; and chemokines such as IL-8, MIP-1α, Gro-β, and SDF-1. The one that is most often used clinically is G-CSF, which may be combined with chemotherapeutic agents for added benefit. This mobilizing capability has resulted in a dramatic change in the manner by which HSCs are harvested for transplantation. Up to 25% of candidates for autologous transplantation are unable to mobilize sufficient cells to enable the procedure to be safely performed. The study of mobilization and its counterpart, engraftment, has implications of great significance for patient care. The ability of G-CSF to mobilize bone marrow HSCs has several apparent mechanisms. The first is reported to be the activation of neutrophils, causing the release of neutrophil elastases capable of cleaving CXCR4 on HSCs, thus reducing HSC–bone marrow interaction. Other receptors that undergo cleavage are VCAM-1 and c-kit. A second mechanism of G-CSF-induced mobilization is via CD26, an extracellular dipeptidase present on primitive HSCs that is able to cleave SDF-1 to an inactive form. Other options for improving mobilization include co-administration of G-CSF and inhibitors directed against the CXCR4–CXCL12 interaction. The inhibition of the CXCR4 receptor by a small molecule has been shown to effectively mobilize HSCs into the blood of patients, including those with poor G-CSF-induced mobilization. Adding this compound to G-CSF enhances HSPC mobilization and leukopheresis yields. G-CSF does invoke an inflammatory response and alternatives for its use are actively sought. Recent pre-clinical data suggest that CXCR4 antagonism plus CXCR2 agonism with agents such as Gro-b or CXCR4 plus VLA-4 antagonists may provide efficient mobilization alternatives.

Isolating stem cells for manipulation Characteristics of HSCs used for isolation

Physical Early attempts to isolate HSCs were based on cell size and density. In order to clarify whether the heterogeneity of CFU-S was due to differences in the input cells used, velocity sedimentation was performed to separate cells by size, demonstrating that smaller cells were more likely to produce secondary CFU-S than larger cells. HSCs are similar in size to mature lymphocytes and, when flow cytometry is performed, overlap the lymphocyte region on plots of forward and side scatter. Using cell-cycle-active drugs Because HSCs are largely in a quiescent portion of the cell cycle (G0 or G1 ), investigators have used cell-cycle-active drugs to deplete bone marrow populations of cycling cells and thereby enrich for primitive HSCs. Treatment of mice with nitrogen mustard resulted in a 30-fold enrichment in CFU-S. HSCs may be isolated by in vitro treatment with 5-fluorouracil, and this remains the most commonly used agent. In addition, HSC populations may be further enriched by first stimulating cells to enter the cell cycle with the early-acting cytokines c-kit ligand and IL-3, before forcing them to metabolic death. This strategy is useful for human cells but not murine cells, probably because of different cycling characteristics. It should be noted that these techniques may affect the quality of HSCs obtained. Markers of primitive HSCs A variety of strategies have been used to identify HSC surface markers. In general, these have depended on excluding cells that express antigens known to be on more mature cells. For example, B cell, T cell, monocyte, granulocyte, and erythroid markers are combined in a so-called lineage cocktail to remove cells that bind any of these antibodies. The remainder of the cells have been sequentially tested for binding of particular antibodies and then transplanted to test the bound cells’ function as stem cells. This iterative process has enabled the definition of cells enriched for reconstituting function in irradiated mice. In the case of human cells, the mice are multi-immune deficient, so they are tolerant of human hematopoietic cell engraftment. Using such strategies, antibody combinations are now defined that enrich for long-term reconstituting HSCs. The panels of markers used are provided in Table 2.2. Supravital stains Since HSCs are inherently quiescent, spend most of their time in inactive portions of the cell cycle, and are resistant to toxins, exclusion of dyes has been used as a method of isolation. The DNA dye Hoechst 33342 was first used to separate quiescent cells from the bone marrow. Cells with low-intensity staining were enriched for high proliferative potential (HPP)-CFC and day-12 CFU-S. The red and

32 Molecular Hematology

Table 2.2 Proposed surface markers of primitive hematopoietic stem cells Mouse

Human

CD34low/− Sca-1+ C-kit+ CD150+ CD48− lin−

CD34+ CD49f+ Thy1+ (CD90)+ CD38low/− CD45RA−/low lin−

blue emissions from this dye have been used to define a small subset of bone marrow cells known as the side population (SP). SP cells have extremely low fluorescence emission in these channels, resulting from efflux of Hoechst 33342 by multidrug resistance pumps that are highly expressed on HSCs. SP cells constitute approximately 0.1% of the bone marrow and are highly enriched in reconstitution potential. The mitochondrial dye rhodamine-123 (Rh-123) has also been used to subdivide primitive stem cells. Mitochondria in quiescent cells bind low levels of Rh-123 and fluorescenceactivated cell sorting (FACS) can be used to separate Rh123low cells. These cells were enriched for day-13 CFU-S and multilineage reconstituting potential. While supravital stains have been useful, the simplicity of analysis with fluorescent antibodies against cell surface markers has generally supplanted them. Using a combination of antibody staining features, it is possible to enrich for HSCs such that fewer than 10 cells are required to reconstitute hematopoiesis. Methods of isolation of HSCs Fluorescence-activated cell sorting

While the flow cytometer may be used for the analysis of cells, the apparatus may also physically sort cells of desired fluorescence or fluorescence pattern, size, and granularity characteristics. Sorting is both expensive and labor intensive, as it requires costly machines, a high degree of expertise, and time to sort samples consisting of single-cell suspensions. However, FACS is the most commonly used method to isolate highly purified HSCs using both positive and negative selection strategies, with fluorescence-labeled antibodies directed against primitive hematopoietic cell antigens, as described earlier. Magnetic bead columns Large-volume isolation of HSC subsets has been facilitated by the use of magnetic bead columns. Using this system, cells are incubated with antibodies directed against primitive hematopoietic cells. These

antibodies are typically coupled to a hapten. A second-step incubation is then performed using a magnetic microbead conjugated to a hapten that is able to bind the first-step hapten. The effect is to label HSCs with a magnetic bead. Cells are then passed through a column mounted adjacent to a magnet. Labeled cells are retained within the column and unbound cells can be washed through. Then, the column is removed from the magnet and the desired cells may be eluted. Alternatively, negative selection may be performed by capturing only the cells that pass through the column. For example, a sample may be depleted of mature cells by labeling with antibodies directed against mature blood cell antigens (Linpos ). Cells can then be passed over a column in which the mature cells adhere and immature cells pass through and may be isolated. Systems of these types permit rapid isolation of large numbers of primitive cells of relatively high purity. Bead-based methods sacrifice cell purity for higher cell yields compared with FACS. Ex Vivo expansion

Given the possible clinical applications of HSCs for such uses as bone marrow transplantation, there is increasing interest in strategies that both result in an increase in the quantity of HSCs and the ability to manipulate HSCs ex vivo. Thus, ex vivo expansion of HSCs represents a highly prioritized goal of clinically oriented HSC research. The first benefit of expanding HSCs is to provide sufficient cells for transplantation when insufficient numbers exist. For example, cord blood represents a rich source of primitive CD34+ cells that are less immunocompetent and are therefore transplantable across partial human leukocyte antigen (HLA) disparity barriers. However, the absolute quantity of HSCs within a single cord blood is low and transplantation is followed by periods of aplasia. Ex vivo expansion would thereby facilitate cord blood transplantation. Similarly, selective expansion of HSC subsets would permit the extension of tumor-free cells from patients with limited quantities of normal bone marrow due to bone marrow–infiltrating diseases, such as leukemia, for the purpose of autologous transplantation. The second benefit of ex vivo manipulation is that HSCs have a relative growth advantage over other cell types, such as tumor cells. Therefore, ex vivo growth provides a purging effect. Furthermore, specific tumor cell purging may be achieved via the application of certain cytokines (IL-2, IFNγ), antitumor agents such as 5-fluorouracil or cyclophosphamide, tumor-specific antibodies combined with complement-mediated lysis, and oncogene-specific tyrosine kinase inhibitors, in addition to other targeted therapies, such as antisense oligonucleotides, prior to use of the graft.

Stem cells 33

The third benefit is the support of gene transfer into HSCs for the purpose of gene therapy or gene editing. A variety of gene transfer and gene editing mechanisms are improved with cell cycling. Further, low efficiency events in gene modification may be made clinically relevant if HSC expansion can be achieved. Strategies to expand HSCs ex vivo have used a wide variety of cytokine cocktails. These have often resulted in more cells, but not more stem cells. Progress has been made with small molecules added to cytokine combinations. The cytokines generally involve kit ligand, thrombopoietin, IL-6, and flt-3 ligand. Adding aryl hydrocarbon receptor antagonist led to improved outcome in umbilical cord blood transplant in one clinical study. Also, early results look encouraging for nicotinaminde or the small molecule UM171 currently in clinical testing. Other results indicate that culture of cells with Notch ligands can reduce intervals of myeloid cell cytopenia and brief exposure to prostaglandinE2 analogues may affect cord blood transplant. This area of research has historically been one of frustration, but these developments indicate that progress is being made. (Note to reader: the chapter author does have involvement in companies focused on this area, so interpret with caution.) Functional analysis of HSCs

Functional assays for HSCs are critical, as immunophenotypic and other isolation methods depend on characteristics that are not necessarily consistent with stem cell function. This is particularly true when assessing cells after in vitro manipulation or in older subjects. Immunophenotypes have been defined by transplantation, generally using young donors. Only functional validation should be considered a bona fide measure of stem cells. However, even in the case of most assays, the setting is transplantation, which is an extreme state and does not necessarily reflect stem cells under homeostasis. In Vitro assays The CFU-C measures hematopoietic progenitor function and is performed by plating cells in semisolid media containing methylcellulose and one or more cytokines. After 5–14 days, colonies comprising mature cell populations committed to either myeloid or lymphoid lineages may be observed. While most colonies obtained using this assay are composed of cells of a single lineage, less frequently multipotent progenitors can yield colonies containing multiple lineages. Another type of primitive cell, known as the HPP-CFC, which possesses a high degree of proliferative and multilineage potential, may be detected in this culture system. Formation of HPP-CFC colonies, characterized by size greater than 0.5 mm and multilineage composition, requires the use of multiple cytokines in order

to proliferate. These are more primitive progenitors, but they cannot be equated with HSCs. The LTC-IC assay correlates more closely to HSCs. Here, hematopoietic cells are plated on top of stromal cell lines or irradiated primary bone marrow stroma. Primitive HSCs are able to initiate growth and generate progeny in vitro for up to 12 weeks. Progenitor cells and mature myeloid cells are removed weekly to prevent overgrowth. Ultimately, HSCs, characterized by high proliferative and self-renewal capabilities, are able to sustain long-term culture and may be enumerated at the conclusion of the assay. The CAFC assay represents a type of LTC-IC that similarly measures the ability of cells to initiate growth and generate progeny in vitro for up to 12 weeks. However, the readout is slightly different. Hematopoietic cells are plated at limiting dilution on top of a monolayer consisting of irradiated bone marrow stroma or a stromal cell line. The growth of colonies consisting of at least five small non-refractile cells reminiscent of cobblestones, found underneath the stromal layer, is counted. Such cultures are maintained using weekly half-media changes until up to 5 weeks after seeding. In this assay, more primitive cells appear later, and day-35 CAFCs represent a close correlate of a cell with in vivo long-term multilineage repopulating potential. LTC-ICs may be enumerated after day 35 by completely removing the CAFC medium, overlaying methylcellulose, and counting the number of colonies produced after 8–10 days. In Vivo assays The CFU-S assay, first developed by Till and McCulloch in 1961, is described earlier in this chapter (see Hematopoietic stem cell concepts and their origin). Bone marrow or spleen cells are transplanted to irradiated recipients and animals are killed after 8 or 12 days for analysis of spleen colonies, termed CFU-S8 and CFU-S12 , respectively. Cells that give rise to CFU-S8 are predominantly unipotential and produce erythroid colonies. CFU-S12 colonies consist of several types of myeloid cells, including erythrocytes, megakaryocytes, macrophages, and granulocytes. Cells giving rise to CFU-S12 represent a more primitive population of multipotent cells than those that result in CFU-S8 . However, these assays measure primitive progenitors, not HSCs. The long-term repopulation assay is a more accurate measure of HSC activity. Whole collections of hematopoietic cells or fractionated subpopulations are transplanted to lethally irradiated syngeneic mice, typically by tail vein injection. Recipients are screened for ongoing hematopoiesis 8– 10 weeks after transplantation. By this time, hematopoiesis is firmly established and donor-derived blood is produced by transplanted HSCs. This assay requires that cells fulfill the two central features of HSCs: multilineage reconstitution, consistent with multipotentiality, and indefinite hematopoiesis, indicative of self-renewal.

34 Molecular Hematology

Assess percentage of multilineage bone marrow engraftment using flow cytometry with antibodies directed against Ly5.1 and Ly5.2

Lethal irradiation with 9-10Gy Ly5.2 test cells

>16 weeks

104

Ly5.1 mouse

Ly5.1 PE

103 102 101

Ly5.1 whole bone marrow competitor cells

100

Tracking of transplanted cells was originally conducted using radiation-induced chromosomal abnormalities or by retrovirally marking donor cells. However, a major advance in the ability to track transplanted cells has been the development of congenic mice with minor allelic differences in the leukocyte common antigen Ly5, which is expressed on all nucleated blood cells. The C57/BL6 (“black-6”) strain contains the Ly5.2 antigen, while the BL6/SJL strain contains a separate allele, Ly5.1. However, these syngeneic strains may be transplanted interchangeably. Both antibodies are available with distinct fluorescent labels. FACS analysis using these antibodies permits measurement of donor-derived reconstitution of the nucleated blood lineages. However, erythrocytes and platelets do not express the Ly5 antigen and cannot be tracked using this technique. Instead, investigators use congenic strains with allelic variants of hemoglobin and glucose phosphate isomerase to track erythroid and platelet engraftment, respectively. A modification of this assay permits quantitation of HSCs within the graft. Here, HSCs are quantified by transplanting limiting-dilution numbers of bone marrow into lethally irradiated recipients. Each recipient also receives 1 × 105 cells of the host’s marrow to ensure survival during the period of pancytopenia immediately after irradiation. At 10–12 weeks, host peripheral blood is assessed to determine whether donor-derived reconstitution has occurred. Donor cells must constitute at least 1% of the peripheral blood to contend that at least one HSC was present in the donor population. Also, both lymphoid and myeloid lineages must demonstrate at least 1% donor derivations. The percentage of reconstituted animals in each group may be plotted against the number of input cells to determine a limitingdilution estimate of the frequency of HSCs within the donor population. This assay is termed a competitive repopulation assay, as transplanted HSCs compete with the host’s HSCs that survive irradiation-induced death, in addition to host cells transplanted with the graft. The HSCs detected are

101

102

103

104

Ly5.2 FITC

Fig. 2.5 Competitive repopulation assay.

termed competitive repopulation units. The competitive repopulation assay using congenic mouse strains is depicted in Figure 2.5.

Summary Investigation of HSCs has been facilitated by the development of in vitro and in vivo assays of hematopoietic cell function, followed by the identification of molecular cell surface markers that permit the isolation of purified subsets of cells with defined characteristics. Studies in this field have contributed greatly to the understanding of both general stem cell biology and hematopoiesis. Further investigation of cellintrinsic and cell-extrinsic regulators of hematopoiesis will enable rational manipulation of HSCs and thereby extend the current uses of stem cells in clinical practice.

Further reading Regulation of hematopoiesis Chai-Ho, W. and Chute, J.P. (2017). Paracrine regulation of normal and malignant hematopoiesis. Curr. Opin. Hematol. 24 (4): 329–335. Gottgens, B. (2015). Regulatory network control of blood stem cells. Blood 125 (17): 2614–2620. Hu, D. and Shilatifard, A. (2016). Epigenetics of hematopoiesis and hematological malignancies. Genes Dev. 30 (18): 2021–2041.

Hematopoietic stem cell niche Gao, X., Xu, C., Asada, N., and Frenette, P.S. (2018). The hematopoietic stem cell niche: from embryo to adult. Development 145 (2), dev139691. Hoggatt, J., Kfoury, Y., and Scadden, D.T. (2016). Hematopoietic stem cell niche in health and disease. Annu. Rev. Pathol. 11: 555–581.

Stem cells 35

Hematopoietic stem cell quiescence Foudi, A., Hochedlinger, K., Van Buren, D. et al. (2009). Analysis of histone 2B-GFP retention reveals slowly cycling hematopoietic stem cells. Nat. Biotechnol. 27 (1): 84–90. Wilson, A., Laurenti, E., and Trumpp, A. (2009). Balancing dormant and self-renewing hematopoietic stem cells. Curr. Opin. Genet. Dev. 19 (5): 461–468.

Hematopoietic stem/progenitor cell expansion Boitano, A.E., Wang, J., Romeo, R. et al. (2010). Aryl hydrocarbon receptor antagonists promote the expansion of human hematopoietic stem cells. Science 329 (5997): 1345–1348. Lund, T.C., Boitano, A.E., Delaney, C.S. et al. (2015). Advances in umbilical cord blood manipulation – from niche to bedside. Nat. Rev. Clin. Oncol. 12 (3): 163–174.

Isolation of hematopoietic stem cells Majeti, R., Park, C.Y., and Weissman, I.L. (2007). Identification of a hierarchy of multipotent hematopoietic progenitors in human cord blood. Cell Stem Cell 1 (6): 635–645. Oguro, H., Ding, L., and Morrison, S.J. (2013). SLAM family markers resolve functionally distinct subpopulations of hematopoietic stem cells and multipotent progenitors. Cell Stem Cell 13 (1): 102– 116.

Hematopoietic stem cell engraftment Broxmeyer, H.E. (2016). Enhancing the efficacy of engraftment of cord blood for hematopoietic cell transplantation. Transfus. Apher. Sci. 54 (3): 364–372. Hoggatt, J., Speth, J.M., and Pelus, L.M. (2013). Concise review: sowing the seeds of a fruitful harvest: hematopoietic stem cell mobilization. Stem Cells 31 (12): 2599–2606.

Chapter 3 The genetics of acute myeloid leukemias Amy M. Trottier & Carolyn J. Owen Division of Hematology and Hematological Malignancies, University of Calgary, Foothills Medical Centre, Calgary, Canada

Introduction, 37 AML with recurrent cytogenetic abnormalities, 37 Molecular genetic aberrations not detectable by conventional cytogenetics, 40

Introduction Acute myeloid leukemia (AML) is a heterogeneous disease with respect to clinical and morphological features as well as acquired genetic aberrations. Traditionally AML patients were divided into three broad risk groups based on cytogenetic abnormalities: favorable, intermediate, and adverse, with each having different cure rates. However, approximately 50% of patients have a normal karyotype (normal-karyotype AML). These patients have varied clinical outcomes, with some attaining a good response to conventional chemotherapy and others having high rates of relapse and disease-related mortality. This led to the search for recurrent genetic aberrations that are not detectable by conventional cytogenetics. With the advent and wider availability of next-generation sequencing methods such as whole-genome sequencing and gene expression profiling, significant progress has been made in the identification of recurrent genetic mutations at the molecular level. These “molecular markers” include mutations in specific genes as well as altered gene expression. The importance of these new molecular markers was manifest in the 2008 update of the World Health Organization (WHO) classification with the addition of the provisional entities AML with mutated NPM1 and AML with mutated CEBPA. Verification of the prognostic significance of these molecular aberrations and the discovery of numerous other leukemogenic driver mutations via whole-genome sequencing are reflected in the recent 2016 revision to the WHO classification with the formal incorporation (no longer provisional entities) of AML with mutated NPM1 and AML with

Molecular Hematology, 4th Edition. Edited by Drew Provan and John G. Gribben. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

AML therapies and MRD monitoring targeted by genetics, 45 Summary, 46 Further reading, 46

biallelic mutations of CEBPA, as well as the addition of a new provisional entity, AML with mutated RUNX1. Molecular analysis for mutations of CEBPA, NPM1, and FLT3 is now standard of care in newly diagnosed AML and the importance of many other gene mutations is being increasingly recognized. These insights have led to the incorporation of molecular aberrations into validated risk stratification models, such as that of the European LeukemiaNet (ELN), and recommendations for risk-adapted treatment strategies, aiming to improve treatment outcomes and minimize toxicity. Unfortunately, the overall outcomes in AML remain poor, with most patients succumbing to their disease. Outcomes are particularly poor for older adults (>65 years). This chapter reviews the prognostically important recurrent genetic aberrations observed in AML. As most currently available data on the impact of genetic abnormalities in AML are derived from large trials enrolling younger patients (50% Down syndrome ~50% JAK/2 mutated

-Activating mutations

Tumour suppressor

~13% near haploid ALL

Kinase-activating mutations

>90% of Ph-like

RTK/RAS

~70% near haploid ALL

TP53

90% low hypodiploid ALL

RB1

~40% low hypodiploid ALL

ERG

~7% –15% of B other

Kinase-activating mutations in Ph-like ALL CRLF2, tyrosine kinase genes (ABL1/2, CSFR1, PDGFRB), other JAKSTAT targets (e.g. JAK2, EPOR) and RAS

Inactivating mutations

Transcriptional regulator Inactivating deletions

Fig. 5.3 Mechanisms of aberrant gene activities by chromosomal translocations.

genome-wide understanding of secondary oncogenetic events in ALL, has defined new molecular entities while also enhancing our understanding of disease pathogenesis in well-established cytogenetic groups. These are now further discussed and summarized in Figure 5.3. Philadelphia chromosome positive t(9;22) (q34;q11) ALL (Ph-pos)

The pathogenic consequence of balanced translocation between chromosomes 9 and 22, t(9;22)(q34;q11), leads to a shortened chromosome 22, also known as the Philadelphia chromosome. Ph-pos ALL is perhaps the best-elucidated molecular entity of ALL, owing partly to the enhanced molecular pathophysiological understanding of chronic myeloid leukemia (CML), which similarly is characterized by t(9;22) genetic rearrangement. Creation of the Philadelphia chromosome (Ph) 22q leads to fusion of the breakpoint cluster region gene (BCR) at chromosome 22q11 to the Abelson gene (ABL) protein-tyrosine kinase gene at chromosome 9q23 during t(9;22) recombination, with resultant

BCR-ABL1 oncogene creation. Depending on the breakpoint, either the p210 or p190 BCR/ABL isoform results, which can readily be detected ty reverse transcription-polymerase chain reaction (RT-PCR). The resulting BCR/ABL1 fusion gene encodes a chimeric BCR/ABL1 oncoprotein that has constitutively activated ABL1 tyrosine kinase enzyme activity, an activator of cellular signaling. Tyrosine kinases initiate cellular signal transduction cascades by catalyzing the transfer of the phosphate of adenosine triphosphate (ATP) to tyrosine residues on target protein (signaling) substrates, which drives a signal transduction program to enhance cell proliferation, self-renewal, and reduce apoptosis – behaviors which typically contribute to malignant transformation. Molecular elucidation of BCR-ABL1-mediated oncogenesis has blazed the trail for molecularly targeted therapies in the form of tyrosine kinase inhibitors (TKIs). Imatinib, a first-generation BCR/ABL1 TKI, blocks the binding of ATP to the kinase domain of the BCR-ABL1 oncoprotein, which in turn blocks the phosphorylation activity of BCR/ABL1 oncoprotein, thereby extinguishing oncogenic cell signaling.

64 Molecular Hematology

Subsequent-generation TKIs nilotinib, dasatinib and ponatinib have increased potency for kinase inhibition. Although TKIs broadly share the same mechanism of action, namely competing with ATP, the phosphorylating entity, at the catalytic binding site of tyrosine kinase, they differ from each other most notably in potency and in their interaction with either the closed/non-functional (imatinib and nilotinib) or open/functional conformation of the kinase domain (dasatinib). Despite the proven efficacy of TKIs in Ph ALL, resistance can occur by a multitude of mechanisms, including genetic mutations caused by acquired mutations in the ABL-Kinase domain, which prevent drug binding. The most intractable ABL1 kinase mutation is the T315I, which confers pan-TKI resistance with sensitivity remaining to ponatinib alone. Overall, the molecular elucidation of the BCR-ABL oncogene is a shining example of how molecular scrutiny can directly inform underlying ALL biology, yielding novel predictors of response to targeted treatments and facilitate individualization of treatment. However, unlike CML, where TKI inhibition effects a functional cure, Ph-pos ALL cannot be cured through TKI inhibition alone, pointing to important biological differences between these diseases. One key difference is the target cell for transformation. ALL represents malignant transformation of a lymphoid progenitor, as opposed to the hematopoietic stem cell (HSC), which is the cellular target of malignant transformation in CML. Furthermore, additional acquired defects in several tumor suppressor genes typify Ph-associated ALL versus CML, strongly implicating the disruption of these genes as contributory to Ph ALL pathogenesis and aggressiveness. One of the key recurring genetic inactivations associated with Ph ALL involves the IKZF1 gene encoding the B cell-specific transcription factor Ikaros, leading to severely reduced expression of the functional protein. Over 60% of Ph ALL exhibit focal deletions of IKZF1, making this the most prevalent co-occurring genetic abnormality associated with Ph ALL; point mutations also contribute to IKZF1 haploinsufficiency, although these are less frequently observed. Recurring loss of function in CDKN2A/B cell cycle control together with PAX5 and EBF1 B cell development genes are also prevalent (∼50%) in the Ph ALL subtype and likely contribute to BCR-ABL1 leukemogenesis. Loss of Ikaros has been shown by some studies to confer a poor prognosis, yielding this as a potential marker for clinical risk stratification. MLL-rearranged ALL

Chromosomal translocations involving the KMT2 gene (also known as MLL) represent one of the most aggressive forms of ALL. A large number of partner genes, AFF1 (AF4), MLLT1 (ENL), MLLT4 (AF6), MLLT3 (AF9), and MLLT10 (AF10), have been shown to recombine with mixed-lineage leukemia

(MLL), but by far the most common is AF4, located on chromosome 4, which undergoes rearrangement with MLL on chromosome 11, resulting in t(4;11)(q21;q23) and MLL-AF4 gene fusion. A typical MLL fusion protein contains the N terminus of MLL encoded by the first 8–13 exons and the C terminus of one of over 50 fusion partner genes. MLL rearrangement mediated by t(4;11) (q21;q23) is frequently present in infant leukemia, accounting for up to 50% of cases, with a diminishing incidence (2% in children, 7–10% in adults) thereafter. A number of features define MLL-rearranged ALL as a distinct clinicobiological subtype, namely poor outcome, unique association with a pro-B immunophenotype, and excessive use of VH6 immunoglobulin gene – an early event in immunoglobulin gene rearrangement as well as a distint gene expression profile. The major mechanisms underpinning MLL-rearranged leukemogenesis have now been clarified and point to disordered epigenetics as a major mechanism of malignancy. The MLL gene encodes a large multifunctional epigenetic regulator (DNA methyltransferase and histone H3K4 methyltransferase – a positive global regulator of gene transcription providing essential epigenetic maintenance of the HOX family of genes, master regulators of hematopoiesis. MLL fusion proteins corrupt the intrinsic epigenetic regulatory activity, leading to global changes in gene transcription and in particular inappropriate expression of MLL target genes such as HOXA9 and MEIS1, which act downstream to suppress differentiation and induce aberrant self-renewal. Extensive studies have characterized the mediators of transcriptional and epigenetic perturbations caused by MLL mutations. These studies have identified co-factor epigenetic enzymes (LEDGF, DOT1L, CBX8, KDM1A, BRD4, P-TEFb, BET) that assemble with MLL fusion proteins to form multiprotein complexes to mediate the epigenetic disturbance. The elucidation of these protein associates of MLL fusions has opened up several therapeutic opportunities. For example, identification of the DOT1L mediator of H3K79 histone modification, which is significantly enriched at MLL target genes, has led to the targeting of histone methylation induced by DOT1L. Similarly, the BET family of chromatin adapter proteins, which associate with oncogenic MLL fusions, can be targeted by novel small-molecule inhibitors that displace BET proteins from chromatin. Both strategies are currently in early-phase clinical trials and results are eagerly awaited. Targeting of epigenetic enzymes that associate with MLL fusions to induce leukemogenic suppression is therefore a promising therapeutic avenue. High hyperdiploidy

The hyperdiploid karyotype is characterized by massive aneuploidy. The formal definition is an increase in

Molecular basis of acute lymphoblastic leukemia 65

chromosomal copy numbers, with gains of between 51 and 65 chromosomes. Hyperdiploid karyotypes constitute 25–30% of pediatric ALL, making this one of the largest cytogenetic subsets in children. The prognosis of highhyperdiploid ALL in children is excellent, with overall survival exceeding 90%. In contrast, high hyperdiploidy occurs less frequently in adult ALL and does not confer a similar magnitude of excellent prognosis, although certain chromosomal amplifications may be associated with increased survival gain. Chromosomal gains in hyperdiploid ALL typically include +X, +4, +6, +10, +14, +17, +18, and +21, and can also include trisomies (e.g. +4, +10, +17, +18), which may be specified as additional biomarkers for improved survival. The pathogenic consequences of chromosomal hyperdiploidy are poorly understood, as the specific target genes affected are not known, but it is generally believed that gene dosage effects contribute to the oncogenic role of a hyperdiploid karyotype. Secondary genetic events (IKZF1 deletions, ETV6, and RAS pathway mutation) have been identified using high-resolution techniques, suggesting that high hyperdiploidy also arises through a multistep process. The presence of these secondary abnormalities has not been shown to confer prognostic significance. t(12;21) (p13;q22) ALL

The t(12;21)(p13;q22) translocation is the most common chromosomal abnormality in pediatric ALL. The rearrangement creates a fusion gene encompassing the non-DNA binding region of TEL (ETV6) and almost the entire locus of AML1 (RUNX1), which retains its coding sequence. ETV6RUNX1 resulting from t(12;21)(p13;q22) is cytogenetically cryptic and cannot be discerned reliably in G-banded preparations, therefore FISH or RT-PCR is required for its accurate detection. TEL and AML1 are two well-characterized transcription factors which are required for development and definitive hematopoiesis. The TEL-AML1 fusion causes perturbations in normal hematopoiesis, enhancing self-renewal of B cell progenitors and altering HSC homeostasis, but is insufficient on its own to induce the full leukemia phenotype. The multistep model of leukemogenesis which premises a pre-leukemia initiating clone has been best demonstrated for this unique ALL subtype. Cytokine receptor-like factor 2 deregulated ALL

Deregulated expression of CRLF2, resulting from juxtapositioning of the CRLF2 gene to either the Immunoglobulin heavy gene enhancer (IGH@-CRLF2) or the pseudoautosomal region or the P2RY8 promoter (P2RY8-CRLF2) of chromosomes X or Y, occurs in up 5–7% of B-ALL, with an increased incidence in B-other (30%) and in Down syndrome–associated ALL (>50%). The biochemical

consequences of chromosomal activation of CRLF2 appear to be activation of the JAK-STAT, ERK, and mTOR/PI3K cellular signaling pathways. Furthermore, JAK2 gene mutations are concomitantly present in ∼50% of CRLF2-rearranged ALL, which contributes to hyperactivation of this signaling axis. Although an association of CRLF2-deregulated ALL with clinical outcome has not been reproducibly confirmed, some studies in adults demonstrate an association with higher rates of treatment failure. More significantly, the discovery of CRLF2 molecular alteration provides a rationale target for signal transduction inhibitors of the JAK/STAT and PI3K pathways in CRLF2-overexpressing ALL. Hypodiploid ALL

Karyotypes with 90% of cases by TP53 alterations, with further abnormalities of CDKN2A/B, RB1, and inactivation of IKZF2 (∼50%). Interestingly, TP53 mutations are constitutional in a significant proportion of childhood LH cases, indicative of a possible underlying Li–Fraumeni syndrome. The finding of hypodiploid cells harboring activating mutations in RAS signaling pathways provides a rationale for blockade therapies targeting MAPK/MEK/ERK signaling in this poor-risk ALL subset.

T ALL T ALL accounts for ∼20% of adult and 8–15% of childhood ALL. Gross chromosomal alteration is less frequently observed in T- versus B-lineage ALL, with just 50% of cases harboring a primary lesion. A further contrast is that T ALL– specific abnormalities are not clearly prognostic for relapse or overall survival, and are therefore not clinically applied for risk assignment. The nature of chromosomal aberrancy in T ALL is dominated by balanced translocations, which occur in up to 20–25% of cases, with numerical changes being exceedingly rare. Although the participating chromosome regions undergoing rearrangement can vary, the unifying outcome of these events is induction of aberrant expression

66 Molecular Hematology

of hematopoietic transcription factor oncogenes, due to inappropriate positioning of these genes near constitutively active T-cell-specific enhancers/regulatory regions within the TCR beta (7q32-q36) or TCRA–TCRD (14q11) 7q32-q36 loci. These oncogenic transcription factors include basic helixloop-helix (bHLH) family members such as TAL1, LYL1, TAL2, and BHLHB1, LIM domain only (LMO) genes (LMO1 and LMO2), and HOX transcription factor oncogenes TLX1, TLX3, and HOXA. Activation of these transcription factor genes or transcription factor accessory proteins (LMO1 and 2) leads to enhanced activity of the transcriptional regulatory protein at the promoter or enhancer elements of a specific set of target genes involved in the development of hematopoietic lineages. A prototypical example of chromosomal translocation/ transcription factor dysregulation in T ALL is the disruption of TAL1 (T-cell acute leukemia), also known as the SCL or TCL5 gene. The genetic events leading to TAL1 overexpression occur by a variety of mechanisms. Most commonly, TAL1 is activated by a cryptic interstitial deletion, TALd, which results in the deletion of a ∼100 kb DNA fragment next to the TAL1 locus and fusion with the 5′ part of the SIL gene, leading to formation of the SIL-TAL1 chimeric gene. This deletion occurs in up to 25% of T ALL in children, but is much less common (∼5–10%) in adults. TAL1 overexpression is also brought about by t(1;14)(p33;q11), which leads to inappropriate TAL1 activation, due to its relocation close to the TCRAD locus. A recently discovered genetic mechanism delineating a novel basis for TAL1 dysregulation in T ALL independent of chromosomal translocation is the occurrence of somatic mutation in non-coding regions downstream of TAL1, leading to the formation of an oncogenic “super-enhancer” that drives transcription and overexpression of TAL1. Thus, multiple genetic defects in T ALL converge on TAL-1 oncogene overexpression. The second largest cytogenetic subgroup in T ALL, representing ∼20% of cases in childhood, is characterized by t(5;14)(q35;q32) which juxtapositions the TLX3/HOX11L2 gene to BCL11B, leading to TLX3 transcription factor dysregulation via a non-TCR enhancer-driven mechanism. Occasionally, chromosomal translocations in T ALL are seen that produce fusion genes encoding chimeric transcription factor oncogenes, such as t(11;19)(q23;p13) leading to MLL-ENL, and also t(10;11)(p13;q14) resulting in PICALM-MLLT10 (CALM-AF10). A common feature of leukemias harboring these fusion transcription factor oncogenes is the aberrant expression of developmentally important HOXA genes. Recently, a rare cryptic 9q34 deletion generating the SETNUP214 fusion gene has been described in T ALL which, similar to BCR-AB1, has constitutively activated tyrosine kinase activity and is associated with sensitivity to inhibition with tyrosine kinase inhibitors, especially nilotinib and dasatinib.

Somatic mutations involving the NOTCH1 gene lead to constitutive activation of Notch1 signaling, which is a prominent oncogenic pathway in T ALL. In both childhood and adult ALL, the incidence of activating NOTCH1 mutations is between 50% and 70%. In addition, FBXW7 mutations, present in about 15% of T ALL cases, contribute to Notch activation by impairing the proteasomal degradation of activated Notch1 in the nucleus. The NOTCH1 gene encodes a type I transmembrane heterodimeric receptor, which functions as a ligand-activated transcription factor. Upon ligand binding, a series of proteolytic cleavages of the receptor, the last of which is catalyzed by gamma secretase, releases intracellular Notch1, which acts downstream to form a transcriptional activator complex for Notch1 target genes influencing differentiation, proliferation, apoptotic events, and in-cell fate determination/choices. Crucial effectors of the oncogenic program downstream of NOTCH1mutated T ALL include MYC and the PI3K–AKT signaling pathway. A rare translocation t(7;9) (q34;q34.3) leads to aberrant expression of a number of truncated Notch1 isoforms, referred to as TAN (for translocation-associated Notch homolog). These truncated Notch1 isoforms (for example TAN1) have a dominant oncoprotein effect, possessing constitutive, ligand-independent activity. However, the major cause of Notch1 dysregulation in T ALL is somatic mutations of the extracellular Notch1 heterodimerization domain or C terminal PEST domain, which confer constitutive, ligand-independent activity to the Notch1 receptor. Discovery of molecularly altered Notch1 has led to targeted therapies that therapeutically terminate Notch1 signaling using γ-secretase inhibitors. Studies have shown that combined with steroids, γ-secretase inhibitors can lead to clinically significant leukemia regressions, demonstrating the promising potential for Notch1 blockade therapies in T ALL. A further frequent target of recurrent mutation in T ALL is CDKN2A deletions, occurring in 70% of cases. Multiple other components of cell cycle control are also targets of damaging mutations. These include chromosomal deletion of RB1 and CDKN1B genes and CCND, all seen in up to 15% of cases. Thus cell cycle dysregulation plays a prominent role T ALL pathogenesis. The MYC oncogene is one of the most frequently activated oncogenes in T ALL. The mechanism of MYC upregulation is mostly post-translational, although it can result from t(8;14)(q24;q11) chromosomal translocation in 1% of cases. Altered signal transduction programs also contribute to T ALL oncogenesis. Gain-of-function mutations within the JAK-STAT signaling axis are also a typical feature of T ALL. Activating mutations target the IL7R gene (10%) encoding the IL-7 receptor, JAK1, and JAK3 (10%) genes, as well as more downstream effectors STAT5B (5–10%). Activating mutations that predict hyperactivity of the MAPK pathway are also reported and result from PTEN tumor suppressor

Molecular basis of acute lymphoblastic leukemia 67

T-cell maturity Early T precursor cell

Early cortical

Late cortical

– – – – Weak + +

+/– + + +

+ + – –

+ +/–

+++ +++

++ ++

Myeloid tumor suppressors Hematopoietic genes Epigenetic regulators Cell signaling

NUP214-ABL1 Homebox genes PHD finger protein 6 Wilms tumor 1 Protein tyrosine phosphatase Non-receptor type 2

TAL1 activation PTEN deletion

Immunophenotype sCD3 sCD4 sCD8 sCD1a sCD5 Myeloid marker Stem cell marker Genetic alterations Notch1/FBXW7 CDKN2A

Fig. 5.4 Genetic abnormalities across T acute lymphoblastic leukemia (ALL) subtypes.

loss, as well as activation of HRAS, KRAS, and PTPN11 oncogenes. The finding of these molecular alterations has clearly opened up opportunities for signal blockade therapies which, although still in the early phase of pre-clinical and clinical testing, hold tremendous promise. Further insights into the molecular pathogenesis of T ALL have been garnered from recent genome-wide and transcriptomic studies. These efforts have led to the recognition of novel molecular oncogenic subtypes corresponding to developmentally distinct stages of T ALL which are underpinned by unique biology (Figure 5.4). A major T ALL entity identified is the early T-cell precursor (ETP) subgroup, which exhibits a unique constellation of clinical, cellular, and molecular features. ETP ALL comprises 10–15% of childhood T ALL and approximately 40–50% of adult T ALL. ETP ALL is defined by characteristic immunophenotypic features typical of early immature (early cortical) thymocytes: absence of CD1a, CD4, and CD8 expression, weak CD5, and aberrant expression of at least one myeloid or stem-cell marker such as CD34 and myeloid surface antigens (CD13 and CD33). Compared with more developmentally mature T ALL (e.g. cortical T ALL), the clinical outcome of this group is poor. Large-scale sequencing data have provided major insights into the molecular mechanisms underlying this unique clinicobiological subgroup. Transcriptionally, ETP T ALLs closely relate to HSCs and myeloid progenitors. The spectrum of acquired somatic alterations are distinct from classical T ALL

genetic alterations of transcription factor oncogene activation. The mutations identified affect three principal pathways, namely loss-of-function mutations in transcription factor–encoding genes involved in hematopoietic development (e.g. RUNX1 IKZF1, ETV6 GATA3 EP300; ∼50%), activating mutations in factors mediating cytokine receptor and cellular signaling axis (NRAS, KRAS JAK1, JAK3, IL7R FLT3), and inactivating mutations in several epigenetic modifiers (PRC2, a H3K27 trimethylase, and loss of function of EZH2, IDH1, IDH2, DNMT3A). Notably, some of the reported genetic alterations correspond to mutations in myeloidspecific oncogenes and tumor suppressor genes, supporting the demonstration of a myeloid gene expression program in ETP ALL. A particular prominence of aberrant activation of JAK-STAT signaling in ETP tumors due to IL7 receptor and JAK1 and JAK3 as well as NRAS/KRAS pathway mutations suggests that these are rational targets for therapeutic intervention. Crucially, genome-wide studies of T ALL have highlighted the increasing importance of the non-protein coding regions in leukemogenesis. Several T ALL–specific noncoding RNAs which are under the direct control of NOTCH1 have been identified which can contribute to the oncogenic state of T ALL. In addition, recurring chromosomal duplications distal to MYC have also been reported. Such alterations create an enhancer for oncogenic NOTCH1, which in turns drives MYC expression. Therefore, there is a broadening

68 Molecular Hematology

genomic basis for the core oncogenic programs implicated in T ALL oncogenesis, namely NOTCH1 and MYC, which extends beyond just an intergenic cause. Recently, genomic profiling studies have identified recurrent mutations in ribosomal protein genes, in particular RPL10, RPL5, and RPL11, and CCR4–NOT transcription complex subunit 3 (CNOT3), which are thought to contribute to T ALL development by increasing basal levels of the protein synthesis that facilitates downstream malignant transformation.

Ph-like ALL Ph-like ALL is a recently characterized molecular subgroup in which the underlying gene expression signature is similar to that of Ph-positive ALL, but where a detectable t(9;22) chromosomal rearrangement or BCRABL1 oncogene is lacking. The incidence of Ph-like ALL increases with age, from 10% in children to up to 27% in young adults. The common features of this uniquely identified subgroup are poor clinical outcome and enrichment of CRLF2 rearrangement and overexpression (∼50%), plus IKZF1 gene deletion (∼30%). The identification of Ph-like ALL has led to fervent studies into the underlying signaling pathway alterations mediating the Ph-like gene expression signature. Apart from identification of JAK1/2 mutations and rearrangement of CRLF2, both of which are highly prevalent in Ph-like ALL, a number of novel kinase-activating mutations have been identified. These include kinaseactivating gene fusions (∼12%) involving PDGFRB, ABL1, ABL2, and CSF1R; gene rearrangement of JAK2 (∼7%) and EPOR (∼3%); other mutations in JAK-STAT signaling (∼12%, FLT3, IL7R, SH2B3, JAK1/3, TYK2, IL2RB, and TSLP); in addition to Ras pathway mutations (5%, KRAS, NRAS, NF1, PTPN11, and BRAF). The importance of delineating the underlying activating mutations of Phlike ALL is underscored by the potential tractability of these pathways to treatment intervention. For instance, Phlike ALL would theoretically be amenable to inhibitors of the JAK pathway (CRLF2, JAK2, and EPOR) or tyrosine kinases such as dasatinib (ABL1, ABL2, PDGFRB, and CSF1R). Although large-scale data demonstrating the effectiveness of genetics-based treatment intervention in Phlike ALL are currently lacking, the identification of Phlike ALL offers real potential for novel treatments in this highly distinct poor-risk group. Standardized diagnostics to facilitate Ph-like ALL identification will be a necessary development to realize the potential for optimizing therapy in these patients.

Biological basis of relapsed ALL Recurrence of ALL during or following primary therapy carries an extremely dismal prognosis, with long-term

survival estimated at just ∼7%. The biological mechanisms giving rise to treatment resistance and eventual therapy failure have been greatly informed by increasing in-depth molecular characterization of progressive ALL. In particular, the analysis of matched diagnosis and relapse pairs and of xenografted primary ALL tumors has defined key causal mechanisms contributing to leukemia clone evolution and the specific genomic drivers of relapse (Figure 5.5). These studies reveal that in the majority of cases, relapses are not direct descendants of the original tumor, but are rather the evolved products of a pre-ancestral clone that was concurrently present (at a low level) in the primary tumor. This is consistent with a model of intratumoral clonal diversity, where, rather than a single genome or clone, there are multiple tumor genomes/clones present within a tumor. Detailed evaluation of genetic patterning in sequential samples together with xenografted primary ALL tumors reveals that instead of a linear sequence of mutational events, ALL clones appear to have evolved by a more branching subclonal diversification, resulting in a highly variegated subclonal architecture comprising multiple variant subclones which may share only minimal genetic relatedness. This branching clonal architecture, realized from the molecular genetics of ALL, has now become a major tenet across the wider cancer field. This model predicts that under certain selective pressures – that is, those imposed for instance by cytotoxic therapy – variant subclones possessing a particular fitness advantage will survive to further diversify and/or cause relapsed disease. The precise determinants of “fitness” – that is, whether it reflects an advantage conferred by a particular mutant genotype versus epigenetic or microenvironmental interaction, or indeed whether it is entirely stochastic – continue to be debated, with evidence currently for all of these. From the point of view of the key genomic drivers of disease relapse, several have so far been identified. The main driver mutations emerging at relapse include deletions of cell cycle regulator CDKN2A/B and lymphoid transcription factors ETV-6, IZKF1 ERG, SPI1, TCF4, and TCF7L2, tumor suppressors (TP53), RAS signaling, plus chromatin modifiers SETD2 (histone methyltransferase) and CREBBP (histone acetyltransferase), the latter impairing response to glucocorticoid response genes as well as nucleoside response genes, NCOR1. These lesions can be subclonal at diagnosis, in keeping with the evolutionary theory of cancer, where mutant-bearing fitter clones survive chemotherapeutic targeting. Potentially, the a priori knowledge of relapse-specific mutations at diagnosis may in the future inform the application of upfront therapies (cytokine blockade and drugs that modulate histone marks, e.g. histone demethylases, histone deacetylases, and epigenetic readers targeting bromodomain proteins) aimed at eliminating relapse-initiating clones which may mitigate against ALL recurrence.

Molecular basis of acute lymphoblastic leukemia 69

Therapy

Branching clonal architecture

Emergence of minor (ancestral) subclone Relapse

Diagnosis

Gene alterations Major clone

Ancestral clone

CREBBP NTC52 NCOR1 TP53 IKZF1

Minor subclones

Drug resistance

Fig. 5.5 Molecular mechanisms underlying acute lymphoblastic leukemia relapse.

Molecular evaluation of Minimal Residual Disease A principal feature of ALL reflecting the cellular origin of the disease – a B- or T-cell precursor – is rearrangement of genes within the Ig and TCR locus. This is a normal physiological process active in developing lymphocytes and leads to antibody diversity. While Ig/TCR gene rearrangements in ALL carry no pathophysiological consequence, their occurrence can be exploited for diagnostic purposes to measure for the presence of minimal residual disease (MRD). The concept of MRD arises from known limitations in the conventional assessment of disease response, where despite 90%) of precursor B ALL (pre-B ALL) and T ALL patients, this approach is widely applicable. Ig/TCR MRD quantification relies on the rearrangement and subsequent joining of gene components within the V, D, and J regions of the immunoglobulin and T-cell receptor locus. The joining of component V, D, and J genes results in loss of original nucleotides as well as the de novo insertion of nucleotides (N region insertion) at the junctional regions. The co-occurring insertion and loss of nucleotides at the junctional sites are unique to each lymphocyte or lymphocyte clone. In the case of a leukemic clone, these junctional regions serve as leukemic signatures or unique fingerprints of the patient’s leukemia cell, which can be amplified by PCR using breakpoint-specific primers for MRD assessment. The aim is to identify at least two MRD PCR targets per patient. This is achieved by first determining the presence of clonal Ig and TCR gene rearrangements in a diagnostic sample by heteroduplex analysis of PCR products. This method relies on denaturation of PCR products, followed by rapid cooling to induce homoduplex (bearing PCR

70 Molecular Hematology

products with identical junctional regions) and heteroduplex (bearing PCR products with heterogeneous junctional regions) formation, corresponding to clonal and polyclonal cell populations, respectively. Homoduplexes and heteroduplexes are separated on a polyacrylamide gel, where the former migrate rapidly and the latter more slowly, forming a background smear. Homoduplexes corresponding to clonal IG/TCR rearrangements are then sequenced to identify junctional regions. Thereafter, PCR primers can be designed complementary to the junctional regions and tested for sensitivity and specificity for clone amplification in serially diluted diagnostic DNA. Detection and quantitation of MRD after treatment can then be performed by standard curve generation from the diagnostic dilution series using a real-time fluorescence reporter PCR system. The main drawback of Ig/TCR-based MRD quantification is the occurrence of continuing rearrangements of the original clone during the disease course, which can lead to false negative results. This can be reduced by careful primer design, which avoids amplification of gene segments involved in continuing gene rearrangements and the use of at least two MRD targets per patient. In the case of chromosomal aberrations, the resultant fusion genes can be used as reverse transcriptase PCR (RT PCR) targets for MRD monitoring. The Philadelphia chromosome, t(9;22), and resultant BCR/ABL fusion gene is by far the most commonly assessed target for MRD monitoring in this category. In contrast to antigen gene receptor targets, these targets are not subject to clonal evolution and remain stable throughout disease course. Excellent sensitivities of 10−4 –10−6 can be reached. A disadvantage of this method is

false positive results due to contamination. In addition, only a minority (0.48 (women) or Elevated red cell mass >25% above mean normal predicted value 2. BM biopsy features: hypercellularity; trilineage proliferation (erythroid, granulocytic, and megakaryocytic); pleomorphic, mature megakaryocytes 3. JAK2V617F or JAK2 exon 12 mutation Minor Subnormal serum erythropoietin level Essential thrombocythemia Major 1. 2. 3. 4.

Platelet count ≥450 × 109 l−1 BM biopsy features: predominantly megakaryocytic hyperplasia; large, mature megakaryocytes Not meeting criteria for another myeloid neoplasm (e.g. CML, PV, PMF, MDS) JAK2, CALR, or MPL mutation

Minor Another clonal marker; or absence of reactive thrombocytosis Primary myelofibrosis (overt) Major 1. Megakaryocytic hyperplasia and atypia; reticulin and/or collagen fibrosis 2. Not meeting criteria for another myeloid neoplasm (e.g. CML, ET, PV, MDS) 3. JAK2, CALR, or MPL mutation; in the absence of these mutations, another clonal marker (e.g. ASXL1, EZH2, TET2, IDH1/2 mutation) or absence of reactive myelofibrosis Minor 1. Unexplained anemia 2. Leukocytosis ≥11 × 109 l−1 3. Splenomegaly (palpable) 4. Increased serum LDH 5. Leucoerythroblastic blood film BM, bone marrow; CML, chronic myeloid leukemia; Hb, hemoglobin; Hct, hematocrit; LDH, lactate dehydrogenase; MDS, myelodysplastic syndrome. a Diagnosis of polycythemia vera (PV) requires either all three major criteria, or the first two major criteria and the minor criterion. Diagnosis of essential thrombocythemia requires all four major criteria or the first three major criteria and the minor criterion. Diagnosis of overt primary myelofibrosis (PMF) requires all three major criteria and at least one minor criterion. Separate criteria exist for the diagnosis of pre-fibrotic primary myelofibrosis, the principal difference being the absence of reticulin fibrosis >grade 1.

Molecular markers do not only offer a simplified strategy for diagnosis, but also add prognostic information that can guide management. For example, mutations in ASXL1, SRSF2, EZH2, and IDH1/2 are associated with an adverse prognosis in PMF, and in the case of ASXL1 this is independent of the International Prognostic Scoring System. Mutational analysis may therefore aid in decisions such as the appropriateness of allogeneic stem cell transplantation. In the future, molecular information also may increasingly

inform the potential benefit of targeted therapies, an example being the use of JAK inhibitors in patients with CNL and a membrane-proximal CSF3R mutation.

Conclusions and future directions Identification of JAK2, MPL, and CALR mutations that are relatively specific to PV, ET, and PMF has reinforced

Myeloproliferative neoplasms 97

Table 7.2 British Committee for Standards in Haematology (BCSH) diagnostic criteria for polycythemia vera, essential thrombocythemia, and primary myelofibrosisa Polycythemia vera A1 High hematocrit (men >52%, women >48%) or an increased red cell mass (>25% above predicted value) A2 Mutation in JAK2 Essential thrombocythemia A1 A2 A3 A4 A5

Platelet count >450 × 109 l−1 Presence of an acquired pathogenetic mutation (e.g. in JAK2, CALR, or MPL genes) No other myeloid malignancy, especially JAK2-positive PV, PMF, CML, or MDS No reactive cause for thrombocytosis and normal iron stores Bone marrow aspirate and trephine biopsy showing increased megakaryocytes with a spectrum of morphology, predominantly large with hyperlobated nuclei and abundant cytoplasm. Reticulin generally not increased.

Primary myelofibrosis A1 A2 B1 B2 B3 B4 B5 B6

Reticulin grade 3 or higher (on a 0–4 scale) Pathogenetic mutation or absence of both BCR-ABL1 and reactive causes of bone marrow fibrosis Palpable splenomegaly Unexplained anemia Leucoerythroblastosis Teardrop red cells Constitutional symptoms (drenching night sweats, weight loss >10% over six months, unexplained fever, or diffuse bone pain) Histological evidence of extramedullary hematopoiesis

Diagnosis of essential thrombocythemia requires A1 − A3, or A1 + A3 − A5. Diagnosis of primary myelofibrosis (PMF) requires A1 + A2 and any two of the B criteria. CML, chronic myeloid leukemia; MDS, myelodysplastic syndrome. a Diagnosis of polycythemia vera (PV) requires both criteria to be present. Separate criteria exist for the diagnosis of JAK2-negative PV. Source: Adapted from McMullin M.F., Harrison C.N., Ali S. et al. (2019). A guideline for the diagnosis and management of polycythaemia vera. A British Society for Haematology Guideline. Br. J. Haematol. 184: 176–191. Harrison C.N., Butt N., Campbell P., et al. (2014). Modification of British Committee for Standards in Haematology diagnostic criteria for essential thrombocythaemia. Br. J. Haematol. 167: 421–3; Reilly J.T., McMullin M.F., Beer P., et al. (2012). Guideline for the diagnosis and management of myelofibrosis. Br. J. Haematol. 158: 453–71, with permission.

Dameshek’s dictum that these disorders are closely linked. It has also revolutionized all aspects of the MPNs, including our understanding of pathogenetic mechanisms, clarification of diagnostic criteria, and, importantly, development of new therapies. The discovery of the JAK2V617F mutation led to the rapid development of JAK inhibitors, the first of which was approved for use in PMF just seven years later in 2012, and was subsequently also licensed for some patients with PV. A greater understanding of the additional mutations in MPNs, especially those affecting histone modifications, has prompted clinical trials of other novel drugs such as histone deacetylase inhibitors. In spite of these advances, important challenges remain in the understanding of MPN disease biology and its translation into improving clinical outcomes. Studies of mutational allele burdens in patients taking pharmaceutical agents demonstrate that none of these have a reliable, selective, inhibitory effect on the HSCs that drive disease. This issue must be tackled if therapy is to reverse the underlying disease process and reduce the risk of disease transformation, particularly to

AML with its poor prognosis. It is also increasingly evident that clinical heterogeneity in all of the chronic myeloid disorders reflects significant underlying mutational heterogeneity, and in future more comprehensive mutational analysis might be able to facilitate a more “personalized” approach to therapy.

Further reading Introduction Arber, D.A., Orazi, A., Hasserjian, R. et al. (2016). The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia. Blood 127: 2391–2405. Bench, A.J. and Pahl, H.L. (2005). Chromosomal abnormalities and molecular markers in myeloproliferative disorders. Semin. Hematol. 42: 196–205. Campbell, P.J. and Green, A.R. (2006). The myeloproliferative disorders. N. Engl. J. Med. 355: 2452–2466.

98 Molecular Hematology

Dameshek, W. (1951). Some speculations on the myeloproliferative syndromes. Blood 6: 372–375. Prchal, J.F. and Axelrad, A.A. (1974). Letter: Bone-marrow responses in polycythemia vera. N. Engl. J. Med. 290: 1382.

The JAK2V617F mutation in Philadelphia-negative MPNs Anand, S. and Huntly, B.J. (2012). Disordered signaling in myeloproliferative neoplasms. Hematol. Oncol. Clin. North Am. 26: 1017– 1035. Baxter, E.J., Scott, L.M., Campbell, P.J. et al. (2005). Acquired mutation of the tyrosine kinase JAK2 in human myeloproliferative disorders. Lancet 365: 1054–1061. Godfrey, A.L., Chen, E., Pagano, F. et al. (2012). JAK2V617F homozygosity arises commonly and recurrently in PV and ET, but PV is characterized by expansion of a dominant homozygous subclone. Blood 120: 2704–2707. Godfrey, A.L. and Green, A.R. (2012). Genotype–phenotype interactions in the myeloproliferative neoplasms. Hematol. Oncol. Clin. North Am. 26: 993–1015. James, C., Mazurier, F., Dupont, S. et al. (2008). The hematopoietic stem cell compartment of JAK2V617F-positive myeloproliferative disorders is a reflection of disease heterogeneity. Blood 112: 2429– 2438. James, C., Ugo, V., Le Couedic, J.P. et al. (2005). A unique clonal JAK2 mutation leading to constitutive signalling causes polycythaemia vera. Nature 434: 1144–1148. Jones, A.V. and Cross, N.C. (2013). Inherited predisposition to myeloproliferative neoplasms. Ther. Adv. Hematol. 4: 237–253. Jutzi, J.S., Bogeska, R., Nikoloski, G. et al. (2013). MPN patients harbor recurrent truncating mutations in transcription factor NF-E2. J. Exp. Med. 210: 1003–1019. Keil, E., Finkenstadt, D., Wufka, C. et al. (2014). Important scaffold function of the Janus kinase 2 uncovered by a novel mouse model harboring a Jak2 activation-loop mutation. Blood 123: 520–529. Kralovics, R., Passamonti, F., Buser, A.S. et al. (2005). A gain-of-function mutation of JAK2 in myeloproliferative disorders. N. Engl. J. Med. 352: 1779–1790. Levine, R.L., Wadleigh, M., Cools, J. et al. (2005). Activating mutation in the tyrosine kinase JAK2 in polycythemia vera, essential thrombocythemia, and myeloid metaplasia with myelofibrosis. Cancer Cell 7: 387–397. Li, J., Kent, D.G., Chen, E., and Green, A.R. (2011). Mouse models of myeloproliferative neoplasms: JAK of all grades. Dis. Model Mech. 4: 311–317. Ortmann, C.A., Kent, D.G., Nangalia, J. et al. (2015). Effect of mutation order on myeloproliferative neoplasms. N. Engl. J. Med. 372: 601–612. Tapper, W., Jones, A.V., Kralovics, R. et al. (2015). Genetic variation at MECOM, TERT, JAK2 and HBS1L-MYB predisposes to myeloproliferative neoplasms. Nat. Commun. 6: 6691. Vainchenker, W., Delhommeau, F., Constantinescu, S.N., and Bernard, O.A. (2011). New mutations and pathogenesis of myeloproliferative neoplasms. Blood 118: 1723–1735. Zhao, R., Follows, G.A., Beer, P.A. et al. (2008). Inhibition of the Bcl-xL deamidation pathway in myeloproliferative disorders. N. Engl. J. Med. 359: 2778–2789.

Other JAK2 mutations in MPNs Grisouard, J., Li, S., Kubovcakova, L. et al. (2016). JAK2 exon 12 mutant mice display isolated erythrocytosis and changes in iron metabolism favoring increased erythropoiesis. Blood 128: 839–851. Kapralova, K., Horvathova, M., Pecquet, C. et al. (2016). Cooperation of germ line JAK2 mutations E846D and R1063H in hereditary erythrocytosis with megakaryocytic atypia. Blood 128: 1418–1423. Mead, A.J., Rugless, M.J., Jacobsen, S.E., and Schuh, A. (2012). Germline JAK2 mutation in a family with hereditary thrombocytosis. N. Engl. J. Med. 366: 967–969. Milosevic Feenstra, J.D., Nivarthi, H., Gisslinger, H. et al. (2016). Whole-exome sequencing identifies novel MPL and JAK2 mutations in triple-negative myeloproliferative neoplasms. Blood 127: 325–332. Passamonti, F., Elena, C., Schnittger, S. et al. (2011). Molecular and clinical features of the myeloproliferative neoplasm associated with JAK2 exon 12 mutations. Blood 117: 2813–2816. Scott, L.M., Tong, W., Levine, R.L. et al. (2007). JAK2 exon 12 mutations in polycythemia vera and idiopathic erythrocytosis. N. Engl. J. Med. 356: 459–468.

MPL mutations in ET and PMF Beer, P.A., Campbell, P.J., Scott, L.M. et al. (2008). MPL mutations in myeloproliferative disorders: analysis of the PT-1 cohort. Blood 112: 141–149. Guglielmelli, P., Pancrazzi, A., Bergamaschi, G. et al. (2007). Anaemia characterises patients with myelofibrosis harbouring Mpl mutation. Br. J. Haematol. 137: 244–247. Pecquet, C., Staerk, J., Chaligne, R. et al. (2010). Induction of myeloproliferative disorder and myelofibrosis by thrombopoietin receptor W515 mutants is mediated by cytosolic tyrosine 112 of the receptor. Blood 115: 1037–1048. Pikman, Y., Lee, B.H., Mercher, T. et al. (2006). MPLW515L is a novel somatic activating mutation in myelofibrosis with myeloid metaplasia. PLoS Med. 3: e270. Vannucchi, A.M., Antonioli, E., Guglielmelli, P. et al. (2008). Characteristics and clinical correlates of MPL 515W>L/K mutation in essential thrombocythemia. Blood 112: 844–847.

CALR mutations in ET and PMF Araki, M., Yang, Y., Masubuchi, N. et al. (2016). Activation of the thrombopoietin receptor by mutant calreticulin in CALR-mutant myeloproliferative neoplasms. Blood 127: 1307–1316. Chachoua, I., Pecquet, C., El-Khoury, M. et al. (2016). Thrombopoietin receptor activation by myeloproliferative neoplasm associated calreticulin mutants. Blood 127: 1325–1335. Klampfl, T., Gisslinger, H., Harutyunyan, A.S. et al. (2013). Somatic mutations of calreticulin in myeloproliferative neoplasms. N. Engl. J. Med. 369: 2379–2390. Kollmann, K., Warsch, W., Gonzalez-Arias, C. et al. (2017). A novel signalling screen demonstrates that CALR mutations activate essential MAPK signalling and facilitate megakaryocyte differentiation. Leukemia 31: 934–944. Marty, C., Harini, N., Pecquet, C. et al. (2014). Calr mutants retroviral mouse models lead to a myeloproliferative neoplasm mimicking an

Myeloproliferative neoplasms 99

essential thrombocythemia progressing to a myelofibrosis. Blood 124: 157–157. Nangalia, J., Massie, C.E., Baxter, E.J. et al. (2013). Somatic CALR mutations in myeloproliferative neoplasms with nonmutated JAK2. N. Engl. J. Med. 369: 2391–2405.

Pardanani, A., Lasho, T.L., Laborde, R.R. et al. (2013). CSF3R T618I is a highly prevalent and specific mutation in chronic neutrophilic leukemia. Leukemia 27: 1870–1873. Schwaab, J., Umbach, R., Metzgeroth, G. et al. (2015). KIT D816V and JAK2 V617F mutations are seen recurrently in hypereosinophilia of unknown significance. Am. J. Hematol. 90: 774–777.

Other somatic mutations in PV, ET, and PM Grinfeld, J., Nangalia, J., and Green, A.R. (2017). Molecular determinants of pathogenesis and clinical phenotype in myeloproliferative neoplasms. Haematologica 102: 7–17. Guglielmelli, P., Lasho, T.L., Rotunno, G. et al. (2014). The number of prognostically detrimental mutations and prognosis in primary myelofibrosis: an international study of 797 patients. Leukemia 28: 1804–1810. Vannucchi, A.M., Lasho, T.L., Guglielmelli, P. et al. (2013). Mutations and prognosis in primary myelofibrosis. Leukemia 27: 1861–1869.

Chronic neutrophilic leukemia and chronic eosinophilic leukemia Cools, J., DeAngelo, D.J., Gotlib, J. et al. (2003). A tyrosine kinase created by fusion of the PDGFRA and FIP1L1 genes as a therapeutic target of imatinib in idiopathic hypereosinophilic syndrome. https://doi.org/10.1056/NEJMoa025217. N. Engl. J. Med. 348: 1201– 1214. Maxson, J.E., Gotlib, J., Pollyea, D.A. et al. (2013). Oncogenic CSF3R mutations in chronic neutrophilic leukemia and atypical CML. N. Engl. J. Med. 368: 1781–1790. Meggendorfer, M., Haferlach, T., Alpermann, T. et al. (2014). Specific molecular mutation patterns delineate chronic neutrophilic leukemia, atypical chronic myeloid leukemia, and chronic myelomonocytic leukemia. Haematologica 99: e244–e246.

Integration of molecular analysis into diagnostic and management pathways Harrison, C.N., Butt, N., Campbell, P. et al. (2014). Modification of British Committee for Standards in Haematology diagnostic criteria for essential thrombocythaemia. Br. J. Haematol. 167: 421–423. McMullin M.F., Harrison C.N., Ali S. et al. (2019). A guideline for the diagnosis and management of polycythaemia vera. A British Society for Haematology Guideline. Br. J. Haematol. 184: 176–191. Reilly, J.T., McMullin, M.F., Beer, P.A. et al. (2012). Guideline for the diagnosis and management of myelofibrosis. Br. J. Haematol. 158: 453–471. Vannucchi, A.M., Lasho, T.L., Guglielmelli, P. et al. (2013). Mutations and prognosis in primary myelofibrosis. Leukemia 27: 1861–1869.

Conclusions and future directions Andersen, C.L., McMullin, M.F., Ejerblad, E. et al. (2013). A phase II study of vorinostat (MK-0683) in patients with polycythaemia vera and essential thrombocythaemia. Br. J. Haematol. 162: 498–508. Harrison, C., Kiladjian, J.J., Al-Ali, H.K. et al. (2012). JAK inhibition with ruxolitinib versus best available therapy for myelofibrosis. N. Engl. J. Med. 366: 787–798. Verstovsek, S., Mesa, R.A., Gotlib, J. et al. (2012). A double-blind, placebo-controlled trial of ruxolitinib for myelofibrosis. N. Engl. J. Med. 366: 799–807.

Chapter 8 Lymphoma genetics Jennifer L. Crombie1 , Anthony Letai1 & John G. Gribben2 1 2

Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston MA, USA Barts Cancer Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

Introduction, 101 Techniques, 101 Commonly occurring translocations in lymphoma, 103 Burkitt lymphoma, 104 Diffuse large B-cell lymphoma, 105 High-grade B-cell lymphomas with MYC and BCL2 or BCL6 translocations, 105 Mantle cell lymphoma, 106

Introduction As with all cancers, lymphomas were originally categorized based primarily on morphology and clinical behavior. The use of antibodies against cell surface markers allowed the study of lymphoma specimens with antibody panels that could, along with morphological criteria, usually place a given lymphoma into a diagnostic category. Even within a given lymphoma category, however, there is considerable heterogeneity of clinical behavior. A prominent example is diffuse large B-cell lymphoma (DLBCL), in which over twothirds of patients are cured with conventional chemotherapy, whereas those with relapsed/refractory disease have dismal outcomes. In this case, as with many of the lymphomas, a prognostic index (the International Prognostic Index or IPI) based on pre-treatment clinical criteria is able to provide very useful prognostic information. However, even among patients with the same IPI score, significant clinical heterogeneity persists. Furthermore, it is likely that the IPI defines subclasses of lymphomas based upon biological differences among these lymphomas. Studies of genetic abnormalities have proved to be important tools to further classify and prognosticate lymphomas. In addition, a better understanding of the molecular pathophysiology has led to improvements in therapeutic options.

Follicular lymphoma, 106 Lymphoplasmacytic lymphoma, 106 Mucosa-associated lymphoid tissue lymphomas, 107 Anaplastic large cell lymphoma, 107 Chronic lymphocytic leukemia, 107 Hodgkin lymphoma, 108 Conclusions, 108 Further reading, 109

Techniques Techniques for studying genetic abnormalities in tumor specimens have undergone a revolution in the past two decades (Table 8.1). Initial genetic analyses were based on the technique of chromosomal study by Giemsa-trypsin banding. In these studies, cells are grown in short-term culture, usually in the presence of mitogens. Colcemid treatment results in cell accumulation in metaphase, at which point the cells are fixed and dropped on glass slides. The slides are treated with trypsin followed by Giemsa to give a banding pattern. An experienced cytogenetic technician can then identify normal chromosomes, translocations, numerical abnormalities, and sometimes more subtle deletions. The technique can identify only genetic changes large enough to disrupt a Giemsastained band, requiring a change of many megabases. More modern techniques are able to detect abnormalities with greater sensitivity. Southern hybridization starts with the electrophoretic separation of tumor DNA on a gel, followed by transfer to a membrane. This membrane is then probed with radioactively labeled polynucleotide probes specific for certain genes of interest. Changes in the expected size or intensity of the band of interest can indicate mutation, translocation, amplification, or deletion of the gene of interest. It can also be used to evaluate the presence of clonal

Molecular Hematology, 4th Edition. Edited by Drew Provan and John G. Gribben. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

101

102 Molecular Hematology

Table 8.1 Techniques to study lymphoma genetics r Cytogenetic analysis r Southern blot analysis

r Polymerase chain reaction (PCR) analysis r Fluorescent in situ hybridization (FISH)

r Comparative genomic hybridization (CGH) r CGH microarray

r Gene expression profiling r Single-nucleotide polymorphism (SNP) array r microRNA expression profiling

r Next-generation sequencing (NGS): whole-genome sequencing,

whole-exome sequencing, RNA sequencing, and ChIP sequencing r Clustered regularly interspersed short palindromic repeats (CRISPR)-Cas9 functional genetic screening

rearrangements of the immunoglobulin loci in B cells or the T-cell receptor locus in T cells. Polymerase chain reaction (PCR) technology has allowed for the detection of genetic abnormalities using only a small amount of tumor DNA. Using primers designed to flank the genomic region of interest, repetitive cycles of annealing, DNA polymerization, and thermal melting eventually yield a PCR product. The presence and size of this product may be analyzed by gel electrophoresis to determine the presence of a translocation. Furthermore, a PCR product may be sequenced to look for point mutations. Like Southern blotting, PCR can also be used to evaluate the presence of clonal rearrangements of the immunoglobulin loci in B cells or the T-cell receptor locus in T cells. Fluorescent in situ hybridization (FISH) uses fluorescently labeled DNA probes to bind to specific regions of genomic DNA. Images are then analyzed under a fluorescent microscope. Numerical chromosomal abnormalities may be detected by simply counting the number of signals per cell: more than two indicates the addition of a chromosome, whereas less than two indicates a deletion. To investigate a potential translocation, two probes are used, one to detect the genomic DNA on each side of the known translocation. If the two probes are consistently approximated, this indicates the presence of a translocation. FISH may be performed on interphase cells, so that growth in culture is not a requirement for this type of analysis as it is for conventional cytogenetics. Tests for small deletions or other subtle abnormalities may be better performed on metaphase cells. FISH requires knowledge of the area to be labeled. Since they rely on the annealing of a labeled specific DNA probe or primer, Southern hybridization, PCR, and FISH are techniques to determine the presence or absence of a known genetic abnormality. Modern techniques which provide a

genome-wide scan for abnormalities, requiring no prior suspicion of a particular abnormality which can have a clinical impact, are comparative genomic hybridization (CGH), gene expression profiling (GEP), single-nucleotide polymorphism (SNP) arrays, micro-RNA arrays, next-generation sequencing (NGS), and clustered regularly interspersed short palindromic repeats (CRISPR)-Cas9 based lethality screens. In original CGH techniques, DNA is isolated from the tumor sample and a normal control sample. The DNA in each is labeled with a different fluorescent dye, for example green for the tumor DNA and red for the normal DNA. These samples are mixed and hybridized onto slides of metaphase spreads of normal cells. Images of metaphase spreads are then analyzed for the green-to-red color ratio. Regions of chromosomes that have a high green-to-red ratio contain a putative area of amplification. Regions that have a low green-to-red ratio contain a putative deletion. In this way, the entire genome may be examined for abnormalities. Other techniques, generally beyond what is performed in clinical laboratories, are required to determine the critical genes involved in areas of amplification and deletion. Small abnormalities and balanced translocations cannot be observed using this technique. Currently in more common use is a variation of the CGH technique (known as “array CGH”) which hybridizes the DNA to defined arrays of genomic DNA fragments. These arrays can cover the genome and improve resolution to the size of the DNA fragments within the microarray, currently down to the level of 1 MB. GEP, which allows for a comprehensive, quantitative examination of the mRNA transcripts of a tumor sample, a group of molecules that has been termed the “transcriptome,” was one of the original tools to provide a deeper insight into the molecular underpinnings of lymphoma. In this technique, mRNA is purified from a fresh or frozen tumor sample, amplified, and fluorescently labeled. The labeled RNA is then hybridized to oligonucleotide probes, which can be quantitated by microscopy and image analysis software. Comparison of different tumor samples, and comparison with wild type, can then allow the determination of transcripts that are over- or underrepresented in certain conditions or tumors. Tumors may be categorized using these profiles, and subgroups of messages may be used to create predictors of clinical behavior. For example, GEP has been used use to characterize two molecular subgroups of DLBCL, germinal center B-cell (GCB) subtype and activated B-cell (ABC) subtype, which have both prognostic and therapeutic implications. In recent years, RNA sequencing using NGS, which is discussed in greater detail shortly, has been replacing microarray assays for evaluation of RNA expression. Whole-genome analyses of SNPs have also been facilitated by the development of SNP arrays. DNA fragments are digested and annealed to oligonucleotide linkers, amplified, and hybridized to an SNP array plate. Hundreds of

Lymphoma genetics 103

thousands of oligonucleotides, representing all known SNPs, can be assessed for using this method. This can identify which SNPs are present in a given tumor sample. Similarly, consistent loss of all SNPs along a region indicates a region of chromosomal loss, whereas consistent increase in SNP signal at a particular region indicates amplification. As with GEP, newer NGS-based methods are replacing this technology. miRNAs are small non-coding RNA transcripts that act primarily by modulating the expression of other genes. miRNAs exert this function by annealing to mRNAs, the consequence of which can be shortened mRNA half-life or decreased translation. Any single miRNA can modulate the expression at many different genes. It has been difficult, however, to use the primary sequence of the miRNA to predict the genes at which it acts. Several studies have demonstrated that by measuring the levels of the hundreds of known miRNAs in the genome, one can segregate cancers into different groups. Furthermore, such segregation may provide information about prognosis, progression, and response to therapy. Supporting the concept that miRNAs play an important role in determining cancer behavior, non-coding regions of the genome that are frequently deleted in cancer often contain miRNA genes. Chronic lymphocytic leukemia (CLL) is an example of a lymphoid cancer for which miRNA biology has been informative. For instance, it has been found that in many cases of CLL, particularly those of indolent behavior, there is downregulation or deletion of mir-15-A and mir-161, located at 13q14.3. Both of these miRNAs have the ability to decrease BCL-2 levels, so their deletion may explain the high levels of BCL-2 found in nearly all CLL cells. In addition, expression levels of a limited number of miRNAs may distinguish between indolent and aggressive clinical subtypes in CLL. More recently, NGS has revolutionized DNA and RNA analysis. Whole-exome sequencing (WES), for example, has been utilized on paired tumor and germline DNA samples to identify recurrently, somatically mutated proteincoding genes in lymphoma, allowing for the identification of genes not previously recognized as drivers of cancer. Wholegenome sequencing (WGS), which as the name implies results in sequencing of the entire genome, has the ability also to interrogate variants that may be important for controlling gene transcriptional regulation or splicing. Results from these sequencing tools have allowed for the development of clinical gene panels able to identify frequently occurring mutations with potential therapeutic impact. These panels are transforming the management of hematological malignancies, such as leukemia, providing physicians with rapid knowledge of mutations that can guide the use of targeted therapy and risk stratify patients to various treatment approaches. While these panels are not yet in clinical use for lymphoma, they will likely be incorporated in the near future,

allowing physicians access to the genomic landscape and the ability to personalize management. Similarly, high-throughput RNA sequencing (RNA-Seq), which involves NGS of cDNA, is able to measure the presence and quantity of RNA in a biological sample. This technique can identify alternatively spliced gene transcripts, gene fusions, and post-transcriptional modifications. RNA-Seq can also identify different populations of RNA in a sample, including total RNA, miRNA, and tRNA. Lastly, technologies such as chromatin immunoprecipitation (ChIP) assays can be combined with NGS in a process known as chromatin immunoprecipitation–sequencing (ChIP-Seq). This is a powerful method to determine genomewide binding sites for transcription factors and other proteins. Together, these technologies allow for high-throughput sequencing of thousands of genes from small clinical samples, readily providing massive amounts of data on lymphoma genomics and transcriptomics. Genome editing technology based on the prokaryotic CRISPR–Cas9 system has also revolutionized genome engineering and our ability to identify genes that are critical for survival. In this technique, libraries of small-guide RNAs (sgRNAs) can guide the Cas9 enzyme to target a host of genes in lymphoma models. Next-generation DNA sequencing can subsequently be used to identify genes of interest. This functional genomics approach identifies mutations with functional consequence and thus has the potential to identify optimal therapeutic targets.

Commonly occurring translocations in lymphoma Commonly occurring translocations in lymphoma may be crudely divided into two main classes: those that foster increased proliferation, and those that inhibit programmed cell death, or apoptosis. The classical gene in lymphomagenesis that induces proliferation is c-MYC. Burkitt’s lymphoma (BL), one of the most rapidly dividing lymphomas, is the archetype of a lymphoma that overexpresses c-MYC by the t(8;14). c-MYC is a helix-loop-helix leucine zipper transcription factor, which requires heterodimerization with the protein MAX to activate transcription and induce proliferation. Targets of this dimer include genes controlling cell cycle progression, cell growth, metabolism, differentiation, and apoptosis. The net effect of c-MYC expression is generally an increase in proliferation; however, this effect is context specific. In some cells, c-MYC overexpression can induce cell cycle arrest or apoptosis via p53. Therefore, it may require a concurrent apoptotic defect to permit c-MYC overexpression. BCL-2 is an oncogene that does not directly foster increased proliferation, but rather opposes apoptosis. It does

104 Molecular Hematology

Table 8.2 Chromosomal translocations in non-Hodgkin lymphomas (NHL)

NHL histological type

Translocation

% of cases involved

Protooncogene

Burkitt lymphoma

80% 15% 5% 35%

c-MYC

Cell proliferation and growth

Transcriptional deregulation

Diffuse large cell lymphoma

t(8;14) t(2;8) t(8;22) der(3)

BCL-6

Transcriptional deregulation

Mantle cell lymphoma

t(11;14)

>70%

Transcriptional repressor, required for germinal center formation Cell cycle regulator

Follicular lymphoma Lymphoplasmacytic lymphoma

t(14;18) t(9;14)

90% 50%

Mucosa-associated lymphoid tissue (MALT) lymphoma Anaplastic large cell lymphoma

t(11;18) t(1;14) t(2;5)

50% rare 60% in adults 85% in children

this at least in part by binding and sequestering pro-apoptotic BCL-2 family members, preventing them from communicating or executing death signals, especially at the mitochondrion. It is classically overexpressed in follicular lymphoma due to the t(14;18). This is also the most commonly occurring translocation to occur in conjunction with t(8;14) in doublehit lymphomas, an aggressive subtype of lymphoma that is often refractory to conventional chemotherapy, occurring in approximately 80% of cases. A frequent hallmark of translocations found in B-cell lymphomas is their exploitation of immunoglobulin gene regulatory elements to drive expression of an oncogene in a malignant B-cell or B-cell precursor. Burkitt’s lymphoma is an example of a lymphoma characterized by the overexpression of c-MYC. While the most common translocation is t(8;14), which puts c-MYC under the control of the immunoglobulin heavy chain (IgH) transcription elements, the less common t(2;8) and t(8;22) are also found, putting c-MYC transcription under the control of the light chain κ and λ transcription elements, respectively. The t(14;18) found in follicular lymphoma drives BCL-2 expression using IgH transcription elements. The BCL-6 expression found in many DLCBL cases is often driven by IgH, Igκ, and Igλ elements in the t(3;14), t(2;3), and t(3;22), respectively. PAX5 expression in lymphoplasmacytic lymphoma and cyclin D1 expression in mantle cell lymphoma are likewise driven by the t(9;14) and t(11;14) which exploit the IgH locus. Improved techniques of genetic study have allowed for the identification of a large number of chromosomal translocations, the most common of which are shown in Table 8.2. Those abnormalities which are the most common,

Cyclin D1 (BCL-1) BCL-2 PAX-5

API-2-MLT BCL-10 NPM/ALK

Function

Anti-apoptotic Transcription factor regulating B cell proliferation API-2 is anti-apoptotic ? Anti-apoptotic Anaplastic lymphoma kinase (ALK) is a tyrosine kinase

Mechanism of activation of oncogene

Transcriptional deregulation Transcriptional deregulation Transcriptional deregulation

Fusion protein Transcriptional deregulation Fusion protein

or which have been demonstrated to have the greatest impact on prognosis or treatment, are described there. While these translocations are commonly seen in lymphoma, and can represent the defining feature of subtypes such as follicular lymphoma, other lymphomas are less commonly associated with chromosomal translocations. GEP and other sequencing techniques have been used to identify the drivers of lymphogenesis in many of these cases. For example, while lymphoplasmacytic lymphoma (LPL) was previously thought to have no clear translocation of oncogene, we now know that over 90% of LPLs have MYD88 L265P mutations. Similarly, BRAF mutations have been found in nearly all cases of hair cell lymphoma.

Burkitt lymphoma Burkitt lymphoma (BL) is a very high-grade B-cell malignancy. Pathologically it is characterized by small, noncleaved cells. The presence of many apoptotic malignant cells gives rise to tingible body macrophages and the “starry sky” appearance characteristic of this and other very rapidly dividing tumors. Frequent mitotic figures also demonstrate the rapid cell division typical of this tumor. Though rapidly dividing, it is one of the most curable lymphomas, with >90% of adults enjoying long-term survival when treated with a regimen similar to that proposed by MacGrath. As the MacGrath regimen is quite different, and yields much improved results when compared to the R-CHOP regimen that is used for other aggressive B-cell lymphomas, it is

Lymphoma genetics 105

important to make the diagnostic distinction between BL and DLBCL. Genetic testing plays a key role in making the diagnosis. The genetic hallmark of BL is overexpression of the c-MYC oncogene due to a translocation which places c-MYC transcription under the control of elements at an immunoglobulin locus. The most common translocation, t(8;14), is a chromosomal rearrangement involving c-MYC and the IgH locus. Other translocations involve c-MYC with the κ(t(2;8)) or λ(t(8;22)) light chain loci. It is difficult to make the diagnosis of BL in the absence of a c-MYC translocation by either cytogenetics, FISH, or PCR. The c-MYC (myelocytomatosis) oncogene is a helix-loop-helix, zinc finger containing transcription factor that is associated with a proliferative phenotype. There have been reports of lymphomas that resemble BL, though they lack MYC rearrangements. These lymphomas are associated by 11q alterations, and are now recognized as a new provisional entity in the revised WHO classification of 2016. NGS has further improved our understanding of the underlying biology of BL. For example, 70% of sporadic and 40% of endemic BLs harbor mutations in TCF3, a transcription factor that regulates the B-cell receptor and phosphoinositide 3-kinase (PI3K), or their negative regulator ID3. The mechanism through which these mutations cooperate with MYC and the potential to target these mutations therapeutically remain to be explored. Evidence for latent Epstein–Barr virus (EBV) infection is also found in nearly all of the African endemic BL, but in only 20% of the sporadic form found outside Africa. It has been suggested that EBV plays a causative role by opposing apoptosis, though the exact mechanism is unknown.

Diffuse large B-cell lymphoma DLBCLs, the most common subtype of non-Hodgkin lymphoma (NHL), are a heterogeneous group of lymphomas of aggressive clinical behavior. The majority likely derive from follicular center cells, and roughly one-fifth of DLBCLs derive from transformation of a pre-existing follicular lymphoma. As their name suggests, DLBCLs have a diffuse histological pattern of large lymphoid cells. Approximately two-thirds of patients with this disease will be cured with combination chemotherapy and the CD20 antibody rituximab. While approximately half of patients with relapsed/refractory disease can be rescued by autologous stem cell transplant following high-dose therapy, the remaining patients have limited therapeutic options and dismal outcomes. In an attempt to better divide the heterogeneous group of diseases encompassed by the label DLCBL, an IPI was created. The IPI uses just five pieces of clinical and laboratory

data to further subclassify DLBCLs into four groups. While this formulation does provide a useful refinement of prognosis, it still falls short of the ideal predictor: one that would definitively determine, prior to a particular therapy, whether that therapy would work. As this scoring system was developed in the pre-rituximab era, its ability to predict outcomes with novel therapy remains unknown. While the ideal predictor of response to therapy may not yet be available in practice, progress has been made in improving prognosis using the wealth of molecular data provided by GEP. Two groups, one based at the US National Cancer Institute and one based at Dana-Farber Cancer Institute, have published results of applying GEP to lymphoma samples for which clinical data were available. In both cases, predictors generated by GEP were able to identify new subclasses of lymphomas and also to further refine prognosis even within IPI subgroups. The two main molecular subgroups that define cell of origin (COO) by GEP are GCB-like and ABC-like lymphomas. There are somatic mutations common to both subtypes, including activating mutations of TP53, genes involved in immunosurveillance (B2M, CD58), epigenetic regulators (CREBBP/EP300, KMT2D/C, MEF2B), and BCL6. Other mutations appear to cluster by COO. In particular, mutations affecting genes involved in histone modification (EZH2) and translocations involving BCL-2 are more common in the GCB subtype, while genes involved in B-cell receptor signaling (MYD88, CD79A, CARD11, TNFAIP3) and the NF–κB pathway are common in the ABC subtype. These classifications also have prognostic significance, with patients with ABC subtype DLBCL having worse outcomes. The Hans algorithm can be used to identify these subgroups immunohistochemically, using antibodies to CD10, BCL6, and IRF4/MUM1, though categorization does not exactly mirror the findings detected by GEP. Newer methods in which RNA is extracted from formalin-fixed paraffinembedded tissue have the potential to improve classification of COO, though they are not routinely available. As the molecular signature of DLBCL become more clearly defined, the molecules involved in that signature can be immediately identified as potential targets of anti-cancer therapy, a feat not possible when prognosis is determined by purely clinical criteria. Although the clinical significance of many of the mutations remains unknown, there are ongoing efforts to exploit these mutations with targeted therapies and to tailor therapy by COO.

High-grade B-cell lymphomas with MYC and BCL2 or BCL6 translocations While there was previously an entity known as DLBCL unclassifiable, with features intermediate between DLBCL

106 Molecular Hematology

and BL, the criteria was imprecise and not uniformly used. As DLBCL unclassifiable was found to be overlap with DLBCL with MYC and BCL-2 and/or BCL-6 translocations by morphology and GEP, a new category known as high-grade B-cell lymphoma with MYC and BCL2 or BCL6 translocations, also known as “double-hit” and “triple-hit” lymphomas, has been incorporated into the WHO criteria. These patients often present at an advanced stage and have rapidly progressive disease that is refractory to therapy. GEP has identified a host of mutations in epigenetic mediators, genes involved in signaling pathways, and tumor suppressors.

Mantle cell lymphoma Mantle cell lymphoma is a B-cell lymphoma thought to be the malignant counterpart to the memory B cells found in the mantle zone of lymphoid follicles. It has characteristic cell surface markings of CD5+, CD10−, and CD23−. While it is classically thought of as an aggressive lymphoma, clinically indolent variants with low proliferation indexes have been identified. These indolent variants are often characterized by a mutated IGHV heavy chain and negative staining for SOX11. While mantle cell lymphoma often responds to cytotoxic chemotherapy, it has frustrated attempts at cure with chemotherapy, though there are reports of long-term survivors following allogeneic stem cell transplantation. As in DLBCL, a mantle cell IPI score has been developed, known as the MIPI score, which places patients into one of three risk groups, with low-risk scores correlating with a five-year overall survival (OS) of 60% and high-risk scores correlating with an OS of approximately two years. Mantle cell lymphoma is almost uniformly characterized via classical cytogenetics or PCR by a t(11;14) which puts the cyclin D1 (also known as BCL-1, or B-cell leukemia/ lymphoma 1) gene under control of the igH transcription control elements. Cyclin D1 binds to and activates cyclindependent kinases (CDKs). An important target of this activated CDK complex is the retinoblastoma (RB) gene product. In its hypophosphorylated state, RB inhibits entry into the S phase of the cell cycle by binding the transcription factor E2F. When RB is phosphorylated, E2F is freed to activate the transcription of genes which propel the cell into S phase. Therefore, overexpression of cyclin D1 acts to overcome this late G1 phase checkpoint and maintain continuous proliferation.

is characterized by a follicular pattern in the lymph node. The appearance can be similar to that of non-malignant follicular hyperplasia. Light chain restriction can be useful in suggesting the clonality of the tumor, which distinguishes it from benign hyperplasia. A t(14;18) translocation is found in >85% of follicular lymphomas. This rearrangement puts the BCL-2 gene under the transcriptional control of elements from the igH locus. The BCL-2 protein functions to oppose programmed cell death. It is presumed that BCL-2 expression in malignancies such as follicular lymphoma permits survival of cancer cells under conditions (cell cycle checkpoint violation, metastatic location, genomic instability) which would otherwise trigger programmed cell death. The cloning of BCL-2 led to the identification of a family of related proteins. While some are anti-apoptotic like BCL-2, many are pro-apoptotic, but all function in the control of apoptosis. While follicular lymphoma can be cured by local therapy in very localized disease, it is more usually diagnosed in an advanced stage where cure is exceedingly rare. It is generally quite responsive to chemotherapy, but almost always relapses. The clinical course is commonly marked by a series of chemotherapy-induced remissions followed by relapses, with the interval between these decreasing over time. The end stage of the disease may be characterized by insuperable resistance to chemotherapy or by transformation to an aggressive large B-cell phenotype. Despite the very low cure rate, many patients nonetheless survive for more than a decade due to the indolent nature of the disease. Follicular lymphoma can transform into a higher-grade lymphoma with DLBCL morphology. Numerous genetic changes have been associated with this transformation, including trisomy 7, loss of p53, and c-MYC rearrangements. As outcomes in follicular lymphoma are heterogeneous, a variety of prognostic scoring systems have been developed. The FLIPI score and the more recent FLIPI2 score incorporate clinical criteria to predict OS and progression-free survival (PFS), respectively. There has also been integration of gene mutations into a prognostic model known as the m7-FLIPI score. This scoring system incorporates the mutational status of seven genes (EZH2, ARID1A, MEF2B, EP300, FOXO1, CREBBP, and CARD11), along with the FLIPI score and ECOG performance status, to create a more predictive prognostic tool.

Lymphoplasmacytic lymphoma Follicular lymphoma Follicular lymphoma is an indolent lymphoma. The COO is thought to be the follicular center B cell. Histologically it

Lymphoplasmacytic lymphoma (LPL) is an indolent lymphoma. The cells of this lymphoma have a phenotype that lies midway between mature lymphocytes and plasma cells, and for this reason they are often nicknamed “lymphocytes.” This

Lymphoma genetics 107

lymphoma commonly expresses IgM, which can lead to the syndrome of Waldenstrom macroglobulinemia (WM). WM is characterized by IgM expression, hyperviscosity, bleeding, Raynaud’s phenomenon, visual disturbances, and other neurological symptoms. Roughly half of all LPL cases will demonstrate the t(9;14) which juxtaposes the PAX5 gene and the IgH locus. PAX5 encodes the BSAP (B-cell specific activator protein), which is a transcription factor. Its expression is associated with increased expression of genes important in early B cell development and decreased expression of the p53 tumor suppressor. WGS of patients with WM has also identified the somatic mutation MYD88 L265P to occur in more than 90% of patients. Mutated MYD88 triggers tumor growth through the activation of NF-κB by Bruton’s tyrosine kinase (BTK). Mutations in MYD88 suggest susceptibility to ibrutinib, a targeted inhibitor of BTK, while CXCR4 mutations alternatively convey resistance.

Mucosa-associated lymphoid tissue lymphomas The mucosa-associated lymphoid tissue (MALT) lymphomas are thought to arise from the extranodal counterpart to postfollicular memory B cells found in the marginal zone of lymph node follicles. These tumors are often localized, and their behavior is generally indolent. At least some have a dependence on continued antigen stimulation for survival, as demonstrated by the prolonged complete responses that are seen when early-stage gastric MALT lymphoma is treated with an antibiotic regimen to eradicate chronic helicobacter pylori infection. t(11;18)(q21;q21) is found in more than half of all lowgrade MALT lymphomas, with a preference for gastric lymphomas. The translocation is not typically found in highgrade MALT lymphomas. API-2-MALT 1 fusion protein is expressed from the mutant locus. API-2 (also known as IAP-2) belongs to a family of inhibitors of apoptosis that prevent death likely due to their direct interaction with caspases, the proteases activated by programmed cell death. The physiological function of the MALT 1 protein is less well understood, though it possesses a caspase-like domain at its C terminus. The function of the fusion protein is unclear, though there is some evidence that it activates NF-κB, perhaps leading to inhibition of apoptosis. BCL-10 is overexpressed in a minority of MALT lymphoma cases via the t(1;14)(p22;q32), putting the coding region of BCL-10 under the influence of the IgH enhancer. The function of this protein is unclear, but some researchers have suggested an interaction between BCL-10 and MALT-1,

leading to NF-κB activation. Others have shown that API2MALT 1 correlates with the nuclear location of BCL-10. These findings suggest that these two translocations may be involved in activating the same pathway. Trisomy 3 is observed in 20–60% of all MALT lymphomas. The oncogenic properties of this numerical chromosomal abnormality are not understood.

Anaplastic large cell lymphoma Anaplastic large cell lymphoma (ALCL) is characterized by strong surface expression of the CD30 (Ki-1) antigen, a cytokine receptor in the tumor necrosis factor receptor family. The majority of ALCLs demonstrate T cell surface markers and/or clonal rearrangements of the T-cell receptor locus. There are two main clinical forms, systemic and cutaneous. The cutaneous form is particularly indolent. While it is clinically aggressive, systemic ALCL is generally sensitive to chemotherapy. Approximately 50% of systemic ALCLs carry the t(2;5), which confers good prognosis. The t(2;5)(p23;q35) results in a chimeric gene encoding a fusion of the nucleophosmin (NPM) and anaplastic lymphoma kinase (ALK) proteins. NPM is a multifunctional protein which has been implicated in ribosome assembly, control of centrosome duplication, and nuclear transport as a shuttle protein; it also possesses chaperonin and ribonuclease activities. ALK is a member of the insulin family of receptor tyrosine kinases. Its natural ligand is unknown. The NPM–ALK fusion contains the oligomerization domain of NPM and the tyrosine kinase domain of ALK. It results in a self-oligomerizing, constitutively active tyrosine kinase with transforming properties. NPM–ALK can activate numerous downstream effectors, including phospholipase C-γ PI3K, and RAS. Patients with ALK mutations may also derive benefit from the targeted inhibitor of ALK, crizotinib.

Chronic lymphocytic leukemia CLL is a low-grade lymphoma marked by a peripheral lymphocytosis of CD5+, CD20+, and CD23+ small lymphocytes, similar in morphology to normal lymphocytes. Staging based on the presence of lymphadenopathy, organomegaly, anemia, or thrombocytopenia can provide prognostic information, with those cases in the best prognostic groups enjoying near normal mean survival times. Prognosis can also be estimated by purely molecular criteria. In about half of CLL cases, VH genes exhibit somatic hypermutation, while the others lack VH mutations. CLL with IgVH genes that exhibit more than 2% somatic

108 Molecular Hematology

hypermutation, or mutated IGHV status, have significantly better responses to chemoimmunotherapy and OS than the latter. Results of FISH cytogenetics, which detect genomic aberrations in over 80% of cases, also have prognostic implications. As FISH is a directed rather than screening technique, these abnormalities have previously been demonstrated by conventional banding techniques. The abnormalities include del 13q (50%), del 11q (18%), +12q (16%), del 17p (7%), and del 6q (7%). Regression analysis allowed the assignment of 90% of these cases to one of five prognostic classes based on genetic abnormalities. The best prognostic group included those that had 13q deletion as their sole abnormality, whereas the worst prognostic group contained those with a 17p deletion. CLL with 17p and 11q deletions was more likely to have extensive lymphadenopathy, splenomegaly, cytopenias, and B symptoms. These data raise the question of whether the classical clinical staging, which can be used to predict survival, is partly just a surrogate for particular genetic abnormalities, and it is the genetic abnormalities and resulting expression patterns that are more important in determining prognosis. While there are no disease-defining mutations in CLL, sequencing has revealed a number of mutations that occur at low frequency. Mutations in TP53, NOTCH1, SF3B1, and BIRC3, for example, are of clinical interest because of their prognostic and therapeutic implications. There are ongoing efforts to further understand how these mutations and others impact pathogenesis and whether they can be targeted therapeutically. While genetic changes in CLL have yet to result in major therapeutic advances, further understanding of the signaling pathways implicated in CLL has recently yielded a pharmacological revolution, with the approval of several new targeted agents. B-cell receptor signaling, for example, has emerged as a driving factor in CLL survival. Downstream of the B-cell receptor is BTK, which is essential for activating additional downstream targets such as Akt, extracellular signal–regulated kinase (ERK), and NF-κB. BTK has also been implicated in B cell homing and adhesion. The targeted inhibitor of BTK, ibrutinib, has resulted in high rates of durable responses in patients with relapsed/refractory CLL and those with high-risk features, such as deletion 17p, leading to its Food and Drug Administration (FDA) approval in both the frontline and relapsed/refractory settings. PI3K, another key component of the breakpoint cluster region (BCR) signaling pathway, has also been identified as a therapeutic target, with idelalisib, a PI3Kδ, being FDA approved for relapsed/refractory disease in combination with rituximab. BCL-2, which expressed at high levels in more than 70% of CLL cases, is also a target in CLL. While BCL-2

overexpression in CLL is rarely if ever due to a t(14;18), it is thought that the deletion of two miRNA genes, mir-15-A and mir-16-1, which downregulate BCL-2 levels, are responsible for overexpression. Nearly all CLL cases are dependent on BCL-2 for survival, which is largely due to the requirement for BCL-2 to tonically sequester the large amounts of the pro-death molecule BIM that are generated in CLL cells. When BCL-2 function is abrogated, BIM is released, the mitochondria are permeabilized, and the cell dies. Venetoclax, a targeted BCL-2 inhibitor, has had dramatic clinical activity in patients with relapsed/refractory CLL, including those with high-risk biology, and is now a fundamental therapeutic option for patients with CLL.

Hodgkin lymphoma Classical Hodgkin lymphoma (cHL), which arises from germinal center or post-germinal center B cells, is characterized histologically by small numbers of malignant Reed–Sternberg (RS) cells surrounded by an extensive but ineffective inflammatory cell infiltrate. It has a bimodal distribution, with one peak occurring in young adults and another in adults of older age. While HL is curable in majority of patients treated with conventional chemotherapy with or without radiation, there is a subset of patients with relapsed/refractory disease who are more difficult to treat. cHL lack surface immunoglobulin expression and B-cell receptor signaling, and rely on alternative survival pathways, such as constitutive activation of NF-κB. Despite advances in understanding the genomics of cHL, there had been few therapies aimed at targeting known mutations. More recently, however, high-density SNP arrays with paired transcriptional profiles have identified 9p24.1 amplifications, which result in increased expression of PD-1 ligands, PDL1 and PDL-2, and subsequently increased immune evasion. Based on this discovery, checkpoint blockade has since emerged as an effective therapy for cHL, with excellent responses in patients with relapsed/refractory disease.

Conclusions Identification of the genes involved in lymphoma pathogenesis has allowed for better characterization of the disease. This has also led the discovery of novel, targeted agents and immunotherapies which have dramatically improved outcomes for patients. While great progress has been made, there are many questions remaining about the wealth of genetic information we now possess, and further investigation on how best to utilize this information is required.

Lymphoma genetics 109

Further reading Introduction Sehn, L.H., Donaldson, J., Chhanabhai, M. et al. (2005). Introduction of combined CHOP plus rituximab therapy dramatically improved outcome of diffuse large B-cell lymphoma in British Columbia. J. Clin. Oncol. 23 (22): 5027–5033.

Techniques Alizadeh, A.A., Eisen, M.B., Davis, R.E. et al. (2000). Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403 (6769): 503–511. Barrans, S.L., Crouch, S., Care, M.A. et al. (2012). Whole genome expression profiling based on paraffin embedded tissue can be used to classify diffuse large B-cell lymphoma and predict clinical outcome. Br. J. Haematol. 159 (4): 441–453. Dubois, S., Viailly, P.J., Mareschal, S. et al. (2016). Next-generation sequencing in diffuse large B-cell lymphoma highlights molecular divergence and therapeutic opportunities: a LYSA study. Clin. Cancer Res. 22 (12): 2919–2928. Lohr, J.G., Stojanov, P., Lawrence, M.S. et al. (2012). Discovery and prioritization of somatic mutations in diffuse large B-cell lymphoma (DLBCL) by whole-exome sequencing. Proc. Natl. Acad. Sci. U S A 109 (10): 3879–3884.

Commonly occurring translocations in lymphoma Treon, S.P., Xu, L., Yang, G. et al. (2012). MYD88 L265P somatic mutation in Waldenstrom’s macroglobulinemia. N. Engl. J. Med. 367 (9): 826–833.

Burkitt lymphoma Love, C., Sun, Z., Jima, D. et al. (2012). The genetic landscape of mutations in Burkitt lymphoma. Nat. Genet. 44 (12): 1321–1325. Magrath, I., Adde, M., Shad, A. et al. (1996). Adults and children with small non-cleaved-cell lymphoma have a similar excellent outcome when treated with the same chemotherapy regimen. J. Clin. Oncol. 14 (3): 925–934. Richter, J., Schlesner, M., Hoffmann, S. et al. (2012). Recurrent mutation of the ID3 gene in Burkitt lymphoma identified by integrated genome, exome and transcriptome sequencing. Nat. Genet. 44 (12): 1316–1320. Salaverria, I., Martin-Guerrero, I., Wagener, R. et al. (2014). A recurrent 11q aberration pattern characterizes a subset of MYC-negative high-grade B-cell lymphomas resembling Burkitt lymphoma. Blood 123 (8): 1187–1198. Schmitz, R., Young, R.M., Ceribelli, M. et al. (2012). Burkitt lymphoma pathogenesis and therapeutic targets from structural and functional genomics. Nature 490 (7418): 116–120. Swerdlow, S.H., Campo, E., Pileri, S.A. et al. (2016). The 2016 revision of the World Health Organization classification of lymphoid neoplasms. Blood 127 (20): 2375–2390.

Diffuse large B-cell lymphoma Davis, R.E., Ngo, V.N., Lenz, G. et al. (2010). Chronic active B-cellreceptor signalling in diffuse large B-cell lymphoma. Nature 463 (7277): 88–92. International Non-Hodgkin’s Lymphoma Prognostic Factors Project (1993). A predictive model for aggressive non-Hodgkin’s lymphoma. N. Engl. J. Med. 329 (14): 987–994. Morin, R.D., Johnson, N.A., Severson, T.M. et al. (2010). Somatic mutations altering EZH2 (Tyr641) in follicular and diffuse large B-cell lymphomas of germinal-center origin. Nat. Genet. 42 (2): 181–185. Morin, R.D., Mendez-Lago, M., Mungall, A.J. et al. (2011). Frequent mutation of histone-modifying genes in non-Hodgkin lymphoma. Nature 476 (7360): 298–303. Scott, D.W., Mottok, A., Ennishi, D. et al. (2015). Prognostic significance of diffuse large B-cell lymphoma cell of origin determined by digital gene expression in formalin-fixed paraffin-embedded tissue biopsies. J. Clin. Oncol. 33 (26): 2848–2856. Sehn, L.H., Berry, B., Chhanabhai, M. et al. (2007). The revised international prognostic index (R-IPI) is a better predictor of outcome than the standard IPI for patients with diffuse large B-cell lymphoma treated with R-CHOP. Blood 109 (5): 1857–1861. Shipp, M.A., Ross, K.N., Tamayo, P. et al. (2002). Diffuse large Bcell lymphoma outcome prediction by gene-expression profiling and supervised machine learning. Nat. Med. 8 (1): 68–74. Teras, L.R., DeSantis, C.E., Cerhan, J.R. et al. (2016). 2016 US lymphoid malignancy statistics by World Health Organization subtypes. CA Cancer J. Clin. 66 (6): 443–459. Vose, J.M., Bierman, P.J., Anderson, J.R. et al. (1992). Progressive disease after high-dose therapy and autologous transplantation for lymphoid malignancy: clinical course and patient follow-up. Blood 80 (8): 2142–2148.

Mantle cell lymphoma Hoster, E., Dreyling, M., Klapper, W. et al. (2008). A new prognostic index (MIPI) for patients with advanced-stage mantle cell lymphoma. Blood 111 (2): 558–565.

Follicular lymphoma Bernell, P., Jacobsson, B., Liliemark, J. et al. (1998). Gain of chromosome 7 marks the progression from indolent to aggressive follicle centre lymphoma and is a common finding in patients with diffuse large B-cell lymphoma: a study by FISH. Br. J. Haematol. 101 (3): 487– 491. Certo, M., Del Gaizo Moore, V., Nishino, M. et al. (2006). Mitochondria primed by death signals determine cellular addiction to antiapoptotic BCL-2 family members. Cancer Cell 9 (5): 351–365. Federico, M., Bellei, M., Marcheselli, L. et al. (2009). Follicular lymphoma international prognostic index 2: a new prognostic index for follicular lymphoma developed by the international follicular lymphoma prognostic factor project. J. Clin. Oncol. 27 (27): 4555– 4562. Lo Coco, F., Gaidano, G., Louie, D.C. et al. (1993). p53 mutations are associated with histologic transformation of follicular lymphoma. Blood 82 (8): 2289–2295.

110 Molecular Hematology

McDonnell, T.J., Deane, N., Platt, F.M. et al. (1989). bcl-2immunoglobulin transgenic mice demonstrate extended B cell survival and follicular lymphoproliferation. Cell 57 (1): 79–88. Pastore, A., Jurinovic, V., Kridel, R. et al. (2015). Integration of gene mutations in risk prognostication for patients receiving first-line immunochemotherapy for follicular lymphoma: a retrospective analysis of a prospective clinical trial and validation in a population-based registry. Lancet Oncol. 16 (9): 1111–1122. Solal-Celigny, P., Roy, P., Colombat, P. et al. (2004). Follicular lymphoma international prognostic index. Blood 104 (5): 1258–1265. Tsujimoto, Y., Yunis, J., Onorato-Showe, L. et al. (1984). Molecular cloning of the chromosomal breakpoint of B-cell lymphomas and leukemias with the t(11;14) chromosome translocation. Science 224 (4656): 1403–1406.

Lymphoplasmacytic lymphoma Iida, S., Rao, P.H., Nallasivam, P. et al. (1996). The t(9;14)(p13;q32) chromosomal translocation associated with lymphoplasmacytoid lymphoma involves the PAX-5 gene. Blood 88 (11): 4110–4117. Treon, S.P., Tripsas, C.K., Meid, K. et al. (2015). Ibrutinib in previously treated Waldenstrom’s macroglobulinemia. N. Engl. J. Med. 372 (15): 1430–1440.

Mucosa-associated lymphoid tissue lymphomas Lucas, P.C., Yonezumi, M., Inohara, N. et al. (2001). Bcl10 and MALT1, independent targets of chromosomal translocation in MALT lymphoma, cooperate in a novel NF-kappa B signaling pathway. J. Biolumin. Chemilumin. 276 (22): 19012–19019. Maes, B., Demunter, A., Peeters, B., and De Wolf-Peeters, C. (2002). BCL10 mutation does not represent an important pathogenic mechanism in gastric MALT-type lymphoma, and the presence of the API2MLT fusion is associated with aberrant nuclear BCL10 expression. Blood 99 (4): 1398–1404. Uren, A.G., O’Rourke, K., Aravind, L.A. et al. (2000). Identification of paracaspases and metacaspases: two ancient families of caspase-like proteins, one of which plays a key role in MALT lymphoma. Mol. Cell 6 (4): 961–967.

Anaplastic large cell lymphoma Gambacorti-Passerini, C., Messa, C., and Pogliani, E.M. (2011). Crizotinib in anaplastic large-cell lymphoma. N. Engl. J. Med. 364 (8): 775– 776. Kutok, J.L. and Aster, J.C. (2002). Molecular biology of anaplastic lymphoma kinase-positive anaplastic large-cell lymphoma. J. Clin. Oncol. 20 (17): 3691–3702. Morris, S.W., Kirstein, M.N., Valentine, M.B. et al. (1994). Fusion of a kinase gene, ALK, to a nucleolar protein gene, NPM, in nonHodgkin’s lymphoma. Science 263 (5151): 1281–1284.

Vose, J., Armitage, J., Weisenburger, D., and International TCLP (2008). International peripheral T-cell and natural killer/T-cell lymphoma study: pathology findings and clinical outcomes. J. Clin. Oncol. 26 (25): 4124–4130.

Chronic lymphocytic leukemia Byrd, J.C., O’Brien, S., and James, D.F. (2013). Ibrutinib in relapsed chronic lymphocytic leukemia. N. Engl. J. Med. 369 (13): 1278–1279. Chiaretti, S., Marinelli, M., Del Giudice, I. et al. (2014). NOTCH1, SF3B1, BIRC3 and TP53 mutations in patients with chronic lymphocytic leukemia undergoing first-line treatment: correlation with biological parameters and response to treatment. Leuk. Lymphoma 55 (12): 2785–2792. Cimmino, A., Calin, G.A., Fabbri, M. et al. (2005). miR-15 and miR-16 induce apoptosis by targeting BCL2. Proc. Natl. Acad. Sci. U S A 102 (39): 13944–13949. Crombie, J. and Davids, M.S. (2017). IGHV mutational status testing in chronic lymphocytic leukemia. Am. J. Hematol. 92 (12): 1393–1397. Damle, R.N., Wasil, T., Fais, F. et al. (1999). Ig V gene mutation status and CD38 expression as novel prognostic indicators in chronic lymphocytic leukemia. Blood 94 (6): 1840–1847. Del Gaizo Moore, V., Brown, J.R., Certo, M. et al. (2007). Chronic lymphocytic leukemia requires BCL2 to sequester prodeath BIM, explaining sensitivity to BCL2 antagonist ABT-737. J. Clin. Invest. 117 (1): 112–121. Dohner, H., Stilgenbauer, S., Benner, A. et al. (2000). Genomic aberrations and survival in chronic lymphocytic leukemia. N. Engl. J. Med. 343 (26): 1910–1916. Furman, R.R., Sharman, J.P., Coutre, S.E. et al. (2014). Idelalisib and rituximab in relapsed chronic lymphocytic leukemia. N. Engl. J. Med. 370 (11): 997–1007. Hamblin, T.J., Davis, Z., Gardiner, A. et al. (1999). Unmutated Ig V(H) genes are associated with a more aggressive form of chronic lymphocytic leukemia. Blood 94 (6): 1848–1854. Roberts, A.W., Davids, M.S., Pagel, J.M. et al. (2016). Targeting BCL2 with venetoclax in relapsed chronic lymphocytic leukemia. N. Engl. J. Med. 374 (4): 311–322. Wang, L., Lawrence, M.S., Wan, Y. et al. (2011). SF3B1 and other novel cancer genes in chronic lymphocytic leukemia. N. Engl. J. Med. 365 (26): 2497–2506.

Hodgkin lymphoma Ansell, S.M., Lesokhin, A.M., Borrello, I. et al. (2015). PD-1 blockade with nivolumab in relapsed or refractory Hodgkin’s lymphoma. N. Engl. J. Med. 372 (4): 311–319. Green, M.R., Monti, S., Rodig, S.J. et al. (2010). Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 ligand expression, and further induction via JAK2 in nodular sclerosing Hodgkin lymphoma and primary mediastinal large B-cell lymphoma. Blood 116 (17): 3268–3277.

Chapter 9 The molecular biology of chronic lymphocytic leukemia John G. Gribben Barts Cancer Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK

Introduction and definition, 111 CLL pathogenesis, 111 Epidemiology, 111 CLL cell of origin, 112 Stereotype receptors, 113 Molecular genetics of CLL, 113 The B-cell receptor and B-cell signaling, 115

Introduction and definition Chronic lymphocytic leukemia (CLL) is characterized by the clonal expansion of small mature-appearing malignant B cells with a distinct immunophenotype, with co-expression of CD19, CD5, and CD23, coupled with low levels of CD20 and surface immunoglobulins. This immunophenotype is different from any normal B cell, making it more difficult to determine the cell of origin of this malignancy. Although CLL often has an indolent clinical course, there is great heterogeneity in its clinical behavior, with some patients rapidly progressing to need for therapy, while other patients live long term with the disease with few symptoms and might never require treatment. Research in CLL has identified a large number of prognostic factors that can be used to help identify those patients who are at high risk for disease progression. Because of its indolent course, it was previously felt that CLL cells resembled and arose from long-lived resting lymphocytes, but this has been challenged by heavy water labeling experiments demonstrating that CLL often contains fractions of cells with rapid turnover. This disease has therefore become a paradigm for the concept of a cancer living in a near-equilibrium of cells driven by proliferation through the B-cell receptor (BCR) and ongoing cell death, driven by the expression of both pro- and anti-apoptotic members of the BCL2 family of proteins. An understanding of the pathophysiology of this disease has led to great advances in its treatment by the approval of agents such as ibrutinib and acalabrutinib targeting BCR

BCL2 family member expression in CLL, 116 The CLL microenvironment, 116 CLL and impaired immune responses, 116 Prognostic factors and their clinical utility, 117 Richter transformation, 118 Conclusions, 119 Further reading, 119

signaling and the BCL2 inhibitor venetoclax. The activity of agents targeting these pathways provides strong evidence for their importance in the pathogenesis and maintenance of this disease.

CLL pathogenesis The molecular pathogenesis of CLL is complex and incompletely understood. It appears to be a complex interaction of genetic predisposition, poorly understood cancer-initiating events in mature B cells, and the subsequent acquisition of a variety of genetic and epigenetic changes, shaped by interaction with the host microenvironment (Figure 9.1).

Epidemiology CLL is the most common leukemia in the Western world, but its incidence varies between individuals in different geographical regions, and ranges from 0.06% of individuals in Europe and the USA to less than 0.01% of individuals in eastern Asia. The risk of developing CLL is almost twice as high for men than for women. The risk of developing CLL increases with age, with the median age at diagnosis of around 72 years.

Hereditary factors The finding that strong geographical variability does not alter in patients of Asian origin who live in Europe or the USA

Molecular Hematology, 4th Edition. Edited by Drew Provan and John G. Gribben. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

111

112 Molecular Hematology

Leukemia initiating event

Microenvironment interactions

Genetic predisposition

Initiation

Promotion/accumulation/ loss of immune control

Polygenic IRF4 IRF8 BCL2 miR-15/16

del13q +12

Signaling pathways BCR NF-kB TLR CD38 VLA-4 Integrins NOTCH1 CXCR4

Secondary lesion

Disease progression Chemorefractoriness Richter transformation TP53 NOTCH1 SF3B1 BIRC3 ATM MYC CDKN2A

Fig. 9.1 Chronic lymphocytic leukemia (CLL) pathogenesis and progression. CLL arises in the setting of polygenic predisposition and this differs among geographical locations. Leukemia-initiating genetic events permit longer survival of B cells, which leads to subsequent accumulation, resulting in subclonal genetic complexity.

suggests that genetic and not environmental factors are important. This is further supported by the familial preponderance also seen in CLL. First-degree relatives of patients with CLL have an 8.5-fold elevated risk of developing this disease, and the concordance of CLL is higher among monozygotic than dizygotic twins. Genome-wide association studies in familial CLL have shown that this susceptibility is polygenic, and single-nucleotide polymorphisms (SNPs) have been identified in over 30 loci associated with familial CLL, suggesting that common genetic variations contribute to heritable risk and that these genetic variations differ in diverse populations around the world. A number of these loci occur at sites of genes implicated in CLL pathogenesis. For example, CLL-associated SNPs have been found at the BCL2 locus and also at IRF4, leading to reduced expression of interferon regulatory factor 4, implicated in Notch signaling. An SNP associated with reduced expression of miR-15a and miR161 is associated with familial CLL, and reduced expression of these microRNA allows for increased expression of BCL2 and ZAP70, proteins that confer increased resistance to cell death or enhanced BCR-mediated signaling.

Environmental factors Pesticide exposure might be a risk factor for CLL, and the US Department of Veterans Affairs has accepted that exposure to Agent Orange is a risk factor for the disease. There is little evidence to suggest that ionizing radiation or viral infections

can increase the risk of CLL, and no evidence that dietary or lifestyle factors increase risk.

CLL cell of origin CLL can be divided into two major subsets, distinguished by whether CLL cells express immunoglobulin heavy chain variable (IGHV) genes which have undergone somatic hypermutation in their variable regions. Those cases that have less then 98% homology with germline are designated mutated chronic lymphocytic leukemia (M-CLL), and those that have 98% or higher are designated unmutated chronic lymphocytic leukemia (UM-CLL). These are felt to reflect the stage of normal B cell differentiation from which the CLL cells originate. At diagnosis, some 65% of CLL cases have mutated IGHV, which are therefore proposed to arise from post– germinal center B cells expressing immunoglobulin that has undergone somatic hypermutation and, in some cases, also immunoglobulin isotype switching (Figure 9.2). The other 35 of cases express unmutated IGHV, suggesting origin from a B cell that has not undergone germinal center–driven differentiation. Patients with CLL cells with mutated IGHV typically have a more indolent disease course than CLL patients whose cells express unmutated IGHV. Some CLL cases have been described that are similar to unmutated IGHV CLL, but

The molecular biology of chronic lymphocytic leukemia 113

Germinal center

Selection Low/lost affinity Apoptosis

CD5+ B Cell

Proliferation and SHM

FDC

Improved affinity

Class-switched Memory B cell

Class switching

TH Cell Dark Zone

Light Zone Constrained SHM

?

IgM+IgD+ Memory B cell

IGHV-mutated Ig+ CLL

IGHV-mutated IgM+IgD+ CLL IGHV-unmutated CLL

CLL with limited SHM (e.g. IGHV3-21/IGLV3-21)

Fig. 9.2 Chronic lymphocytic leukemia (CLL) cell of origin. A CD5+ B cell enters the germinal center and undergoes a process of affinity maturation and class switching. Leukemogenesis events arising in cells at different stages within this pathway explain the mutated and unmutated subtypes of CLL seen. SHM, somatic hypermutation.

originate from B cells with limited somatic mutation, such as those encoded by mutated IGHV3-21.

Stereotype receptors The repertoire of immunoglobulin (Ig) molecules produced by CLL cells is considerably more limited than the repertoire of Ig molecules in normal B cells, reflecting biased use of IGHV genes that have restricted somatic mutation and limited junctional and heavy–light chain combinatorial diversity. As many as 30% of patients have CLL cells that express Ig “stereotypes,” and more than 1% of cases have near-identity in their primary structure in the variable region among patients. This limited Ig diversity provides compelling evidence that CLL B cells are selected by the binding activity of their expressed surface Ig, suggesting that B-cell receptor signaling plays a critical role in CLL pathogenesis.

Molecular genetics of CLL Unlike other B-cell malignancies, chromosomal translocations are rare in CLL, but there are a number of

well-characterized genetic alterations, including chromosomal alterations, mutations, alterations in miRNA expression, and epigenetic modifications.

Chromosomal alterations More than 80% of CLL patients carry at least one of four common chromosomal alterations. The most common is a deletion in chromosome 13q14.3 (del13q), occurring in more than 50% of patients and associated with a good prognosis. Deletion in chromosome 11 (del11q) is found in 18% of patients, often associated with alterations in ATM, which encodes a protein involved in DNA repair. Del 11q is associated with more rapid disease progression, but has a good response to modern therapy. Trisomy 12 is found in 16% of patients with CLL and is associated with an intermediate prognosis. Deletion in chromosome 17 (del17p) is found in 5–7% of patients, is associated with loss of the tumor suppressor gene TP53, and has the worst prognostic impact.

Somatic mutations The application of whole-exome sequencing (WES) and whole-genome sequencing (WGS) has transformed our

114 Molecular Hematology

* Q < 0.05

25

• Q < 0.10

Subclonal-low

Subclonal-high

Clonal

* 86

% cases based on mutations: 36% (unmutated)

14%

% cases

20 • 51

15 10

22% 10%

* 43 36 27

5

* 27

• 27

22

20

* 19

18

* 17

17

15

15

14

11

* 7

* 6

6

6

5

5

3

3

18% (clonal + subclonal)

2

2

2

AF NX F1 DT X1 BC OR CC ND 2 KR AS IR F M 4 ED 1 ZM 2 YM 3 NR AS TR AF 3 PI M 1

BR

NO T

CH 1 SF 3B 1 TP 53 AT M PO T1 NF KB I ZN E F2 92 XP O1 EG R2 FB XW 7 M GA BI RC 3 KL HL 6 RP S1 M 5 YD 88 DD X3 X

0

Fig. 9.3 Frequency of genetic events seen in chronic lymphocytic leukemia and demonstrated in a word cloud. Source: Modified from Landau, D.A., Tausch, E., Taylor-Weiner, A.N. et al. (2015). Mutations driving CLL and their evolution in progression and relapse. Nature 526(7574): 525–530.

understanding of CLL pathophysiology, and sequences of more than 500 cases have been reported to date. CLL is one of the tumors with the lowest background mutational load (0.6 per Mb). It is genetically heterogeneous and there are no unifying or diagnostic gene mutations (Figure 9.3). The most frequently identified mutations were found in SF3B1 (21%), ATM (15%), TP53 (7%), NOTCH1 (6%), and BIRC3 (4%). Recurrent somatic mutations have been observed in genes that have a role in mRNA processing (SF3B1 and XPO1), DNA damage (TP53 and ATM), Notch signaling (NOTCH1), B-cell signaling (EGR2 or BRAF), chromatin modification (HIST1H1E, CHD2 and ZMYM3), Wnt signaling, and inflammatory pathways (MYD88). Next-generation sequencing has also revealed considerable intratumoral heterogeneity in CLL. Some somatic mutations, such as those in MYD88, or chromosomal abnormalities, such as trisomy 12 or del13q, are usually found in almost all CLL cells in any one patient, suggesting that these genetic alterations occurred early in disease evolution. Other mutations, notable those found in SF3B1 or NOTCH1, TP53,

or del17p, are found in only a fraction of the CLL cells and must therefore represent subclonal events which occur later in disease evolution. These subclonal driver mutations are associated with more aggressive disease, particularly when two or more mutations are found concurrently. Single-cell analysis is being used to determine whether these events are occurring in the same cells or whether there are different clonal subpopulations, and whether mutations occurring within the same cell are required to affect prognosis. Of considerable interest are studies examining the impact of chemotherapy and how this can induce large clonal shifts due to increases in the proportions of CLL cells that have mutations in genes conferring resistance to therapy, notable TP53 mutation or del17p.

MicroRNA alterations CLL was the first human disease found to be associated with alterations in microRNA (miRNA), with the identification that miR-15a and miR-16-1 are deleted, altered, or

The molecular biology of chronic lymphocytic leukemia 115

downregulated in up to 60% of cases of CLL and are dysfunctional in some cases of familial CLL. Reduced expression or loss of miR-15a and miR-16-1 can enhance the expression of target genes such as BCL2 and MCL1, which encode anti-apoptotic proteins of the BCL2 family. A number of other miRNAs are dysregulated or differentially expressed in subgroups with distinctive biological features, including miR-29a/b, miR-29c, miR-34b, miR-181b, and miR-3676, which target TCL1A. Notably, there is transgenic expression of TCL1 in B cells which promotes the development of CLL in mice. Increased expression of miR-155 is associated with enhanced BCR signaling, B cell proliferation, and lymphomagenesis.

Epigenetic changes CLL cells exhibit global hypomethylation combined with local hypermethylation. Methylation profiling has demonstrated substantial heterogeneity of intratumoral methylation. Increasing methylation heterogeneity is associated with increased genetic complexity, likely due to the acquisition of subclonal mutations. There is therefore a link between genetic and epigenetic evolution in CLL. UM-CLL and MCLL cells have distinctive methylation patterns, similar to those of pre–germinal center versus post–germinal center memory B cells, providing further evidence for the cell of origin, as shown in Figure 9.2. However, when methylation signatures are used to classify distinct clinical CLL subgroups,

the data suggest that CLL cells derive from a continuum of B cell maturation states.

The B-cell receptor and B-cell signaling CLL cells typically co-express low levels of expression of surface IgM and IgD, and relatively few cases of CLL express surface IgG. The BCR in CLL comprises a heterodimer of the transmembrane Ig molecule of IgM or IgD and the signaling components CD79A (Igα) and CD79B (Igβ). Functional BCR is required for CLL cell survival. Crosslinking of surface Ig phosphorylates the immunoreceptor tyrosine-based activation motifs of CD79A and CD79B molecules and triggers BCR-mediated signaling. The Ig of CLL B cells most likely engages with auto-antigens, leading to constitutive BCR signaling in vivo (Figure 9.4). Enhanced BCR-mediated signaling is more commonly observed in UM-CLL, whereas an anergic response predominates in M-CLL, perhaps accounting for the more indolent behavior of M-CLL cases. BCR signaling also has impacts on other signaling pathways, including integrins and chemokines. The chemokine receptor CXCR4 is downmodulated by BCR signaling, altering CLL cell adhesion and migration. This intrinsic dependence on BCR-mediated signaling explains the great sensitivity of CLL to BCR inhibitors such as Bruton’s tyrosine kinase (BTK) or phosphoinositide 3-kinase (PI3K) inhibitors. Indeed, the success of treatment with the

Antigen BCR

CD79

Extracellular

A B P P LYN

PI3K δ

P P

S Y K

LYN P

S Y K

PIP 3

PIP 2

P

P BTK P

P

Intracellular

AKT mTOR

P PLC γ2 PKC β

Idelalisib IKK

Ibrutinib Acalabrutinib

NF- K B

Fig. 9.4 Simplified B-cell receptor (BCR) signaling pathways and inhibitors of the pathway available to treat chronic lymphocytic leukemia.

116 Molecular Hematology

BTK inhibitor ibrutinib provides perhaps the strongest evidence for the importance of BCR-mediated signaling in the survival of CLL cells, and has activity in CLL cells with TP53 abnormalities that do not respond well to conventional chemotherapy.

venetoclax with or without anti-CD20 monoclonal antibody, and with or without ibrutinib, targeting both the BCR and BCL2 pathways to provide perhaps the most effective treatments currently available.

The CLL microenvironment BCL2 family member expression in CLL The BCL2 protein family consists of at least 30 related proteins characterized by the expression of BCL2 homology domains. BCL2 family members are divided into three different subclasses based on structural and functional features. Pro-survival or anti-apoptotic proteins include BCL2, BCL-XL, and MCL1 proteins, which are all overexpressed in CLL and are upregulated by contact of CLL cells with the tumor microenvironment. The mechanism of expression of BCL2 in CLL likely involves miRNA-mediated changes in the malignant cells, and the promoter region of CBCL2 is hypomethylated in CLL, leading to increased expression. The pro-apoptotic family members include BAX and BAD, and the BH3-only protein members include BID, BIM, BAD, NOVA, and PUMA. CLL cells express increased levels of both pro- and anti-apoptotic proteins, so that these cells exist in a complex relationship which is now able to be exploited in the clinic (Figure 9.5). Venetoclax is a small-molecule BH3 mimetic that inhibits BCL-2. This drug is highly potent in inducing apoptosis in CLL cells, by diminishing the capacity of BCL-2 to sequester BIM. Venetoclax is effective in patients with relapsed or refractory disease and in patients with relapsed disease and TP53 mutations or loss. Studies are examining the use of

BCL-2

CLL depends upon survival signals that the cells receive in the tumor microenvironment. Within lymph nodes, CLL form “proliferation centers,” where they interact with non-malignant stromal cells, CLL-associated macrophages (Nurse-like cells), T cells, and mesenchymal-derived stromal cells, among others (Figure 9.6). Within proliferation centers, all CLL cells are exposed to chemokines, integrins, cytokines, and survival factors including BAFF and APRIL, which activate canonical NF-κB, and induce the expression of miR-155, thereby enhancing BCR-mediated signaling. Cytokines such as IL-4 upregulate surface IgM, potentially facilitating interactions of the CLL cell with auto-antigens. Notch or Hedgehog signaling provides pro-survival stimulation, particularly in those cases with trisomy 12.

CLL and impaired immune responses CLL is associated with significant immune suppression and infectious complications are a major cause of morbidity and mortality in this disease. An important aspect of CLL is the development of hypogammaglobulinemia, with the consequent increased risk of infection and poor response to vaccination. The mechanism involved is unclear, but likely

Venetoclax

Pro-apoptotic protein BCL-2

h toc

BIM BAX

on dri

Pro-apoptotic protein

Mi

Apoptosis initiation

on

Cancer cell survival

BAK BAX

Cancer cell death Activation of caspases

Cytochrome c

BCL-2 overexpression allows cancer cells to evade apoptosis by sequestering pro-apoptotic proteins.

Venetoclax binds selectively to BCL-2, freeing pro-apoptotic proteins that initiate programmed cell death (apoptosis).

Fig. 9.5 BCL2 family member expression in chronic lymphocytic leukemia and impact of sequestering BCL2 by venetoclax.

The molecular biology of chronic lymphocytic leukemia 117

Antigen

CCR1/3

IL-10 CCL3/4 CCR4

T cell

Nurse-like cell LAM

IL-4

L CC 17

slg

/22

CCL 3/4

CD267 (TACI) CD269 (BCMA) CD268 (BAFF-R)

FF) (BA L) 7 5 I CD2 6 (APR 5 2 1 CD CD3

CCR1/3

CD40L

IL-4

CD40

Cell B-cell

CCL21 CD38 CD44 VLA-4

HA

ROR2

HEV VCA M-1

ROR1

Frz

CD38 5

CXCL12

12

Wnt

Wnt5a

CXCL13

CXCR

CX CR 4

CCR7

CCL19

CXCL

Strom al cell

Fig. 9.6 Interactions with chronic lymphocytic leukemia (CLL) cells within the microenvironment. HEV, high endothelial venules; LAM, lymphoma-associated macrophage.

represents replacement of CLL cells for normal B cells within the immune microenvironment, as well as impairment of T cell help. CLL cells express high levels of programmed cell death 1 ligand 1 (PD-L1) and programmed cell death 1 ligand 2 (PD-L2), which suppress the effector responses of T cells expressing programmed cell death protein 1 (PD-1) and other immune checkpoint inhibitors, leading to an “exhausted” T cell and NK cell phenotype and impaired cellular immune function.

Prognostic factors and their clinical utility The clinical course of newly diagnosed CLL is extremely variable, with some patients being rapidly symptomatic or developing high-risk disease, which requires treatment soon after diagnosis, while others remain free of symptoms for decades. Prognostic factors that can be used to identify patients who require earlier treatment include clinical features, and

genetic, molecular, and biochemical characteristics of the CLL cells. Multivariable models, prognostic indices, and nomograms have been developed to consolidate the prognostic factors. Commonly used parameters associated with poorer outcome are late-stage disease at initial presentation, male sex, increased age, poor performance status (largely due to co-morbidities), and CLL cell characteristics, including unmutated IGHV, increased expression of ZAP-70, CD49d, or CD38, high serum beta-2-microglobulin, and/or the presence of complex karyotype, del11q, del17p, or TP53 mutations. The most reliable prognostic models are those developed to predict the interval from diagnosis to time of first treatment. There is also increasing interest in predictive models to define outcome with specific types of therapy. Prognostic markers that help predict response to specific treatments are termed predictive biomarkers. TP53 and IGHV mutational status are powerful predictive biomarkers for response to chemoimmunotherapy. These markers are often specific for the type of treatment being investigated, and both of them

118 Molecular Hematology

Relapse

Diagnosis Chemotherapy

Refractoriness Chemotherapy

TP53 mutated CLL cell Small TP53 mutated subclone mixed with TP53 wt clones

Chemotherapy removes TP53 wt clones and selects for the TP53 mutated subclone

Subsequent expansion of the TP53 mutated clone

lose their predictive value when treatment with ibrutinib or venetoclax is used. One of the most powerful predictive biomarkers has been the emergence of del17p to TP53 mutations (Figure 9.7). These changes are relatively rare at diagnosis, but recent WES data has suggested that small subclones harboring these changes might more frequently be present at low frequency. Subsequent treatment with chemotherapy removes the sensitive CLL cells, leaving behind and enriching for cells with mutations conferring resistance to chemotherapy, so that over time the disease becomes more refractory to treatment. Treatment options are changing, and novel agents, notably BCR-signaling inhibitors and BCL2 inhibitors, are highly effective agents that have activity among patients who would have been considered high risk when the only option was conventional chemotherapy. These agents both have activity in TP53 mutated cells. Ongoing clinical research is investigating whether earlier use of these agents will prevent the emergence of chemorefractory clones.

Richter transformation Richter transformation (RT) represents the clinicopathological transformation of CLL into aggressive lymphoma, most commonly diffuse large B-cell lymphoma (DLBCL), or more rarely Hodgkin lymphoma (Figure 9.8). About 7% of CLL patients develop RT, with an incidence rate of approximately 0.5% per year. In over half the patients with RT, the lymphoma cells with Richter transformation do not express CD5 or CD23, which are almost invariably expressed by CLL cells. The DLBCL-like lymphomas in RT most often share the same IGHV-D-J rearrangement as the original CLL clone and as such may express unmutated IGHV, unlike de novo DLBCLs,

Fig. 9.7 Molecular mechanism of chemoresistance in chronic lymphocytic leukemia. Chemotherapy selects for expansion of subclones, such as TP53 mutant cells, with change in subclonal architecture over time and after chemotherapy.

which virtually always express IGHV with somatic mutations as a post–germinal center malignancy. RT cells have other distinctive genetic differences from de novo DLBCL. About 60% of RT lymphomas have inactivating mutations and/or deletions in TP53, often with deregulation of MYC, which is observed in around 40% of cases. MYC deregulation is usually caused by translocations juxtaposing MYC to Ig loci, gene amplification of MYC at 8q24, or somatic mutations affecting MYC transregulatory factors, such as NOTCH1, which is mutated in around 30% of cases. NOTCH1 mutations are largely mutually exclusive with MYC oncogenic activation, consistent with the observation that NOTCH1 directly stimulates MYC transcription and suggesting that activation of oncogenic MYC may be a common pathway in RS transformation. CDKN2A encodes the cell cycle progression inhibitor p16, and is mutated or deleted in roughly 30–50% of cases, but this rarely occurs in CLL or de novo DLBCL. The high prevalence of TP53 disruption in RS reflects the selection of a chemorefractory clone under the pressure of previous CLL treatments, thus suggesting an explanation for the poor outcome and the limited sensitivity to conventional drugs. On the other hand, RT lymphomas typically do not have mutations in a number of genes which are commonly seen in de novo DLBCL, including those encoding proteins involved in nuclear factor-κB signaling or in transcriptional repressors such PRDM1/BLIMP1 or BCL6. Up to 20% of the DLBCL-type Richter transformation and ∼50% of Hodgkin lymphoma–type Richter transformation have IGHV rearrangements that differ from that of the original CLL clone, suggesting that these lymphomas represent a de novo secondary malignancy; some of these seem to be associated with Epstein–Barr virus infection, particularly in patients with severe disease-related immune dysfunction and/or treatment-related immune suppression. Cases

The molecular biology of chronic lymphocytic leukemia 119

DLBCL

(a) MYC translocation amplification

NOTCH1 mutations

MYC activation

CDKN2A loss

MGA mutations

Transformation CLL TP53 disruption

Richter transformation

(b)

Clonally related Richter

80%

V4-39 D6 J4

CLL V4-39 D6 J4

Clonally unrelated Richter V4-34 D2-2J3 20%

Fig. 9.8 Molecular mechanisms of Richter transformation in chronic lymphocytic leukemia (CLL). (a) Changes in TP53, activation of MYC, and loss of CDKN2A are among the most frequent changes seen with diffuse large B-cell lymphoma (DLBCL) transformation in CLL. (b) Representation of clonally related and de novo clonally unrelated DLBCL arising in CLL.

with unrelated IGHV rearrangements have better response to chemotherapy and improved outcome.

Conclusions Advances in our understanding of CLL pathophysiology help explain the clinical heterogeneity that we see in this disease. With the identification of pathways involved in the development of CLL, an increasing array of prognostic and predictive biomarkers are increasingly used in clinical practice. The importance of BCR signaling in driving CLL cell proliferation and BCL2 overexpression to prevent apoptosis explains the long survival of CLL cells. Identification of the importance of BCR signaling and BCL2 pathways to CLL survival has led to the development of the most effective agents that we have in this disease, and these novel agents are able to overcome the poor prognostic markers that predict poor

response to conventional chemotherapy. These novel agents targeting these pathways are very well tolerated and suitable for more elderly patients with CLL. With increased survival in CLL with these new agents, Richter transformation is emerging as a major unmet need in CLL, and identification of agents that will improve the outcome of these patients is the focus of much ongoing research.

Further reading Burger, J.A., Tedeschi, A., Barr, P.M. et al. (2015). Ibrutinib as initial therapy for patients with chronic lymphocytic leukemia. N. Engl. J. Med. 373 (25): 2425–2437. Byrd, J.C., Brown, J.R., O’Brien, S. et al. (2014). Ibrutinib versus ofatumumab in previously treated chronic lymphoid leukemia. N. Engl. J. Med. 371 (3): 213–223. Fischer, K., Bahlo, J., Fink, A.M. et al. (2016). Long-term remissions after FCR chemoimmunotherapy in previously untreated patients

120 Molecular Hematology

with CLL: updated results of the CLL8 trial. Blood 127 (2): 208– 215. Hallek, M., Cheson, B.D., Catovsky, D. et al. (2018). iwCLL guidelines for diagnosis, indications for treatment, response assessment, and supportive management of CLL. Blood 131 (25): 2745–2760. Landau, D.A., Sun, C., Rosebrock, D. et al. (2017). The evolutionary landscape of chronic lymphocytic leukemia treated with ibrutinib targeted therapy. Nat. Commun. 8: 2185.

Landau, D.A., Tausch, E., Taylor-Weiner, A.N. et al. (2015). Mutations driving CLL and their evolution in progression and relapse. Nature 526 (7574): 525–530. Stilgenbauer, S., Eichhorst, B., Schetelig, J. et al. (2018). Venetoclax for patients with chronic lymphocytic leukemia with 17p deletion: results from the full population of a phase II pivotal trial. J. Clin. Oncol. 36 (19): 1973–1980.

Chapter 10 The molecular biology of multiple myeloma Wee Joo Chng1,2 & P. Leif Bergsagel3 1

National University Cancer Institute, National University Health System of Singapore, Singapore University of Singapore, National University Hospital, Singapore 3 Division of Hematology-Oncology, Comprehensive Cancer Center, Mayo Clinic Arizona, Scottsdale AZ, USA 2

Introduction, 121 A plasmablast/plasma cell tumor of post–germinal center B cells, 121 Stages of plasma cell neoplasms, 121 Immunoglobulin translocations present in the majority of MM tumors, 122 Marked karyotypic complexity in MM, 122 Chromosome content associated with at least two different pathogenic pathways, 122 Recurrent IgH translocations and recurrent trisomies representing primary oncogenic events, 123

Introduction Multiple myeloma (MM) is an incurable post–germinal center B-cell malignancy. In 2018, it is estimated that in the USA 30 770 new cases will have been diagnosed, with 12 770 patients succumbing to the disease. It is almost always preceded by a pre-malignant tumor called monoclonal gammopathy of undetermined significance (MGUS), which is the most common lymphoid tumor in humans, occurring in approximately 4% of individuals over the age of 50. The prevalence of both MM and MGUS increases with age, and is about twofold higher in African Americans than in Caucasians, although the rate of progression from MGUS to MM is similar in these two populations.

A plasmablast/plasma cell tumor of post–germinal center B cells Post–germinal center B cells that have undergone productive somatic hypermutation, antigen selection, and immunoglobulin heavy chain (IgH) switching can generate plasmablasts, which typically migrate to the bone marrow, where the microenvironment enables differentiation into long-lived plasma cells. Importantly, MGUS and MM are monoclonal tumors that are phenotypically similar to plasmablasts/ Molecular Hematology, 4th Edition. Edited by Drew Provan and John G. Gribben. © 2020 John Wiley & Sons Ltd. Published 2020 by John Wiley & Sons Ltd.

Universal cyclin D dysregulation, 123 Molecular classification of MM, 124 Prognostic and therapeutic implications of molecular classifications, 124 Possible events mediating transformation of MGUS to MM, 125 Model of molecular pathogenesis of MM, 126 Critical role of the bone marrow microenvironment and novel therapeutic strategies targeting the bone marrow milieu, 127 Conclusion, 128 Further reading, 128

long-lived plasma cells, including a strong dependence on the bone marrow microenvironment for survival and growth. In contrast to normal long-lived plasma cells, MGUS and MM tumors retain some potential for an extremely low rate of proliferation, usually with no more than a few percent of cycling cells until advanced stages of MM.

Stages of plasma cell neoplasms Based on the current diagnostic criteria proposed by the International Myeloma Working Group, four stages of plasma cell neoplasms can be identified based on the presence of myeloma-defining events (hypercalcemia, renal impairment, anemia, bone lesions, >60% bone marrow plasma cells, involved/uninvolved free light chain ratio >100, more than one focal lesion on magnetic resonance imaging [MRI]), level of bone marrow plasma cell infiltration and serum or urine monoclonal immunoglobulins, and extramedullary involvement. The stages include MGUS, smoldering multiple myeloma (SMM), symptomatic myeloma, and plasmacytoma (PCT) (Table 10.1). MGUS can progress sporadically to MM expressing the same monoclonal immunoglobulins with a probability of about 0.6–3% per year. Through the analysis of several large prospective cohort studies, three predictive factors for progression of MGUS to MM were identified, including M-protein greater than 15 g l−1 , IgM or IgA M-protein, and presence of an abnormal free light chain ratio. In the presence of all three risk factors,

121

122 Molecular Hematology

Table 10.1 Features used to diagnose monoclonal gammopathy of undetermined significance (MGUS), smoldering myeloma (SMM), and multiple myeloma (MM) Serum M-protein

Bone marrow plasma cells (BMPC)

MGUS SMM MM

1 focal lesion on magnetic resonance imaging None, with biopsy-proven plasmacytoma (PCT)

the risk of progression at 20 years is 58%, compared with 5% in the absence of any of these risk factors. The risk of progression to MM is higher for SMM than for MGUS, with a median time to progression ranging from one to five years. In one study, M-protein in excess of 30 g l−1 , presence of IgA subtype, and urinary M-protein excretion above 30 g l−1 were factors associated with early progression to MM. Extramedullary MM is a more aggressive tumor that can present as secondary or primary plasma cell leukemia, depending on whether or not preceding intramedullary MM has been recognized. Human multiple myeloma cell lines (HMCLs), which are presumed to include most oncogenic events involved in tumor initiation and progression of the corresponding tumor, have been generated mainly from a subset of extramedullary MM tumors.

Immunoglobulin translocations present in the majority of MM tumors Like other post–germinal center B cell tumors, translocations involving the IgH locus (14q32) or one of the immunoglobulin lambda (IgL) loci (κ, 2p12 or λ, 22q11) are common. In general, these events are mediated by errors in one of the three B-cell-specific DNA modification mechanisms: VDJ recombination, IgH switch recombination, or somatic hypermutation. With rare exceptions, these translocations result in dysregulated or increased expression of an oncogene that is positioned near one or more of the strong immunoglobulin enhancers on the translocated derivative chromosome 14. However, translocations involving an IgH switch region uniquely dissociate the intronic from one or both 3′ IgH enhancers, so that an oncogene might be juxtaposed to an IgH enhancer on either or both of the derivative chromosomes, as first demonstrated for FGFR3 on der(14) and WHSC1 on der(4) in MM. These IgH translocations are efficiently detected by fluorescence in situ hybridization (FISH) analyses. Large studies from several groups show that the prevalence of IgH translocations increases with disease stage: about 50% in

MGUS or SMM, 55–70% for intramedullary MM, 85% in plasma cell leukemia (PCL), and above 90% in HMCLs. Limited studies indicate that IgL translocations are present in about 10% of MGUS/SMM tumors, and in about 15–20% of intramedullary MM tumors and HMCLs. Translocations involving an Igκ locus are rare, occurring in only 1–2% of MM tumors and HMCLs.

Marked karyotypic complexity in MM The karyotypes of MM are characterized by complex abnormalities, both structural and numerical. Numerical chromosomal abnormalities are present in virtually all MM tumors and most, if not all, MGUS tumors. There is nonrandom involvement of different chromosomes in different myeloma tumors, and often heterogeneity among cells within a tumor. The mechanism underlying this karyotypic instability is not fully understood. Centrosome abnormalities, one of the mechanisms mediating chromosomal instability in solid tumors, have been identified in MGUS and about one-third of MM tumors. However, mutations of genes involved in the mitotic spindle checkpoint, another mechanism leading to genomic instability in solid tumors, have not been identified in MM. It is thought that karyotypic complexity increases during tumor progression, although karyotypic progression has not been well documented.

Chromosome content associated with at least two different pathogenic pathways There is a clear consensus that chromosome content reflects at least two pathways of pathogenesis. Approximately half of tumors are hyperdiploid (HRD; 48–75 chromosomes), and typically have multiple trisomies involving chromosomes 3, 5, 7, 9, 11, 15, 19, and 21, but only infrequently (